Review Reports - Modulation of ESKAPE Bacteria Properties by NK-92 and NK-92-Derived LEVs: First Insights

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses a pressing issue: the search for new approaches to combating multidrug-resistant bacteria of the ESKAPE group. The authors demonstrate for the first time that large extracellular vesicles (LEVs) derived from NK-92 cells exhibit direct antibacterial activity and are capable of modulating the sensitivity of S. aureus to clindamycin. The study was performed at a high methodological level, and the results are of interest for molecular immunology and infectology. The authors utilize a wide range of methods (flow cytometry, confocal microscopy, ELISA, and microbiological assays), making their conclusions quite compelling. It is also worth noting that the use of the NK-92 cell line, already used in oncology clinical trials, increases the translational potential of the study.

Comments:
1. Confusion in section numbering. Results:
The text contains two sections: 2.1. The first (p. 2) is "Evaluation of TLR2 and TLR5 expression...", the second (p. 3) is "Quantification of α-defensin-1...". Section 2.8 is missing. Check the consecutive numbering.
2. Section 2.4 (Confocal microscopy):
The results are presented very briefly. There is only a mention of a video in the Supplementary Material. It would be great to add at least one representative micrograph (static z-stack images) to the main text so the reader can assess the nature of the interaction (adhesion, phagocytosis?). In general, this section should at least provide some conclusion.
3. Sections 2.9 and 2.10 (Antibiotics):
Table 1 presents essentially raw data. A graph or statistical analysis could be used for clarity. Mean values and standard deviations are provided, but the number of experiments is not specified. You must specify n (number of replicates) and p-value.
Also, a question: Why is the Etest result (Figure 13) only shown for clindamycin, although Table 1 shows a small effect for erythromycin and cefoxitin? Show the MICs (by the way, there is no transcript) for these antibiotics, or explain in the text why only clindamycin was chosen?
4. Page 12: "Gam-negative bacteria" is a typo; it should be Gram-negative.
Page 12: "...that effect can be mediated..." is grammatically incorrect.
Page 13 (Discussion): "NK cells exert cytotoxicity primarily through granularity and granzymes" — granularity is not the mechanism. They probably meant granules (perforin/granzymes). Please correct.
5. Materials and Methods (4.1):
A reference to "regulatory documents on Sanitary and epidemiological requirements..." for bacterial cultivation is provided. In an international journal, this reference is not always informative. Replace it with standard protocols (or specific environmental conditions, temperature, incubation time).
6. Format:
References:
Reference #44 is in Russian, which is unacceptable. It is better to replace it with a peer-reviewed publication, if possible.
Unfortunately, the article itself is extremely uneven in formatting, making it difficult to read and navigate. Also, the figures and captions are presented in different styles (captions for panels A, B, etc.).

Author Response

Dear Reviewer,

Thank you for taking the time to review our work. We appreciate your attention to detail have carefully considered all of your comments.

This manuscript addresses a pressing issue: the search for new approaches to combating multidrug-resistant bacteria of the ESKAPE group. The authors demonstrate for the first time that large extracellular vesicles (LEVs) derived from NK-92 cells exhibit direct antibacterial activity and are capable of modulating the sensitivity of S. aureus to clindamycin. The study was performed at a high methodological level, and the results are of interest for molecular immunology and infectology. The authors utilize a wide range of methods (flow cytometry, confocal microscopy, ELISA, and microbiological assays), making their conclusions quite compelling. It is also worth noting that the use of the NK-92 cell line, already used in oncology clinical trials, increases the translational potential of the study.

Confusion in section numbering. Results: The text contains two sections: 2.1. The first (p. 2) is "Evaluation of TLR2 and TLR5 expression...", the second (p. 3) is "Quantification of α-defensin-1...". Section 2.8 is missing. Check the consecutive numbering.

We thank the reviewer for this comment, we fixed this, see lines 90, 120,127,142,153,164,177.

Section 2.4 (Confocal microscopy): The results are presented very briefly. There is only a mention of a video in the Supplementary Material. It would be great to add at least one representative micrograph (static z-stack images) to the main text so the reader can assess the nature of the interaction (adhesion, phagocytosis?). In general, this section should at least provide some conclusion.

Dear Reviewer, thank you for your valuable comment. In the original manuscript, the acquisition of this image was described in the Results section, and the resulting video was referenced in the Discussion (lines 228–229) as evidence of NK cell–bacterium interaction. Following your suggestion, supplementary z-stack images with detailed annotations have been added to the Supplementary Materials. A clarifying conclusion summarizing these observations has also been incorporated into the revised text (lines 124–126). We believe these adjustments significantly improve the clarity and interpretability of our findings.

Sections 2.9 and 2.10 (Antibiotics): Table 1 presents essentially raw data. A graph or statistical analysis could be used for clarity. Mean values and standard deviations are provided, but the number of experiments is not specified. You must specify n (number of replicates) and p-value. Also, a question: Why is the Etest result (Figure 13) only shown for clindamycin, although Table 1 shows a small effect for erythromycin and cefoxitin? Show the MICs (by the way, there is no transcript) for these antibiotics, or explain in the text why only clindamycin was chosen?

Dear Reviewer, thank you for your comment. We have clarified the experimental design (lines 174–176): all experiments were performed with n = 3 biological replicates, each with two technical repeats. Statistical analysis was not applied, as the disk diffusion assay served solely as a preliminary screening tool to evaluate LEVs' potential to modulate antibiotic susceptibility. Clindamycin was selected for MIC testing based on: the most pronounced susceptibility increase during screening (+5.1 ± 3.0% vs. control), and its clinical relevance for ESKAPE pathogens. This rationale is now detailed in lines 179–181 and 304–308.

Page 12: "Gam-negative bacteria" is a typo; it should be Gram-negative.
Page 12: "...that effect can be mediated..." is grammatically incorrect.
Page 13 (Discussion): "NK cells exert cytotoxicity primarily through granularity and granzymes" — granularity is not the mechanism. They probably meant granules (perforin/granzymes). Please correct.

We thank the reviewer for this comment, we fixed it! (lines 244, 262, 197). We meant that cytotoxicity is mediated by granulysin and granzymes as main proteins.

Materials and Methods (4.1):A reference to "regulatory documents on Sanitary and epidemiological requirements..." for bacterial cultivation is provided. In an international journal, this reference is not always informative. Replace it with standard protocols (or specific environmental conditions, temperature, incubation time).

We thank the reviewer for this suggestion. We agree that this reference could be not informative/ Now we changed it with “All strains were cultured on agarose medium under appropriate biosafety containment, in accordance with institutional safety protocols for handling pathogenic microorganisms.” See lines 357-359. As it was mentioned in our work https://doi.org/10.3390/ijms26178449

Format: References Reference #44 is in Russian, which is unacceptable. It is better to replace it with a peer-reviewed publication, if possible.

We thank the reviewer for this suggestion! We replaced this source (which is a PhD manuscript of our colleague) with her peer- reviewed publication describing the same method.

Unfortunately, the article itself is extremely uneven in formatting, making it difficult to read and navigate. Also, the figures and captions are presented in different styles (captions for panels A, B, etc.).

Dear Reviewer, we sincerely appreciate the time and effort you dedicated to reviewing our manuscript. We have carefully revised the formatting to align with the journal's requirements and the provided template. We hope that our corrections and point-by-point responses satisfactorily address all your comments.

Please also accept our apologies for the delay in submitting this revision, which was due to previously scheduled team vacations. We thank you for your patience and understanding.

Reviewer 2 Report

Comments and Suggestions for Authors

What is meant by “Ablation of ESKAPE bacteria properties”? Does it refer to complete bacterial killing, elimination of drug resistance, virulence factors, or biofilm‑forming ability, or merely growth inhibition? Have the authors confused bactericidal/bacteriostatic effects with “ablating properties”? Please provide a clear definition and distinction.
Please supply complete MIC, MBC (minimum bactericidal concentration), and time‑kill curves.
The unit for clindamycin MIC is reported as mg/ml in the manuscript. However, the susceptibility breakpoint for S. aureus ATCC 29213 (the standard strain used) is ≤5 μg/ml.
Peptide/lipopeptide compounds tend to aggregate or degrade in physiological saline. Have the authors determined their half‑life in culture medium or serum? If not, claims about in vivo application are unsupported.

Author Response

Dear Reviewer,

Thank you for taking the time to review our work. We appreciate your attention to detail have carefully considered all of your comments.

What is meant by “Ablation of ESKAPE bacteria properties”? Does it refer to complete bacterial killing, elimination of drug resistance, virulence factors, or biofilm‑forming ability, or merely growth inhibition? Have the authors confused bactericidal/bacteriostatic effects with “ablating properties”? Please provide a clear definition and distinction.

We thank you for this insightful question, which prompted us to reconsider our work. We would like to clarify that the term "ablation" was not used in the original manuscript; however, we fully agree on the importance of precise terminology when describing the properties under investigation. As this is a pilot study, it aimed to detect phenomenon, we employed a multi-method approach to characterize diverse bacterial responses to LEVs, including: colony growth suppression, changes in antibiotic susceptibility, alterations in cell surface permeability, and interactions with membrane vesicles. We acknowledge that the observed effects warrant further mechanistic investigation, which we propose as a direction for future research.

Please supply complete MIC, MBC (minimum bactericidal concentration), and time‑kill curves.

Time-kill curves added in Supplementary part. Information about concentration of NK-92 derived LEVs added in material and methods part (391-392 line). Because their concentration can be figured out in protein content it was 1 µg/10⁶ parental cells, we used 16*10⁶ cells which means that extracted concentration of LEVs which we used was 640 µg/10⁶ bacteria.

The unit for clindamycin MIC is reported as mg/ml in the manuscript. However, the susceptibility breakpoint for S. aureus ATCC 29213 (the standard strain used) is ≤5 μg/ml.

We are really sorry for this typo. As mentioned in Material and Methods we used Etest (BIOMÉRIEUX, France) for clindamycin its 0.016-256 μg/mL. We changed units in text and figures (see line 183).

Peptide/lipopeptide compounds tend to aggregate or degrade in physiological saline. Have the authors determined their half‑life in culture medium or serum? If not, claims about in vivo application are unsupported.

Thank you for this comment. We have added the requested details to the manuscript (lines 368–369). LEVs were isolated and characterized by zeta-potential analysis (Zetasizer, Malvern Instruments, UK); the results confirmed suspension stability under the experimental conditions. All procedures involving LEVs were performed in Ca/Mg-free Hanks' balanced salt solution to prevent ion-mediated aggregation of peptides/lipopeptides and preserve vesicle integrity. We acknowledge that the limited stability of LEVs and the requirement for prompt use pose challenges for therapeutic translation. However, given that their modulatory effects on bacterial susceptibility are shown in 60 minutes of interaction, this time may be compatible with applications.

We hope that our revisions and point-by-point responses satisfactorily address all your comments. Please also accept our sincere apologies for the delay in submitting this revision. The extended timeline was due to previously scheduled team vacations and the time required to prepare the additional review report as requested. We greatly appreciate your patience and continued consideration of our manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript addresses an interesting and timely topic by exploring the antimicrobial potential of NK-92 cells and their derived large extracellular vesicles (LEVs) against ESKAPE pathogens, which is conceptually appealing given the growing interest in immune cell–derived vesicles as antibacterial agents. However, the study has several important limitations that need to be addressed.

Moderate Comments

Abstract

Lines 16–18: “This interaction leads to inhibition of colony formation…”. The wording implies causality, but only correlation is shown. Replace with “is associated with inhibition…”

Lines 18–20: “mediated by defensin-α1”. This is not experimentally demonstrated and should be revised to reflect a hypothesis rather than a conclusion.

Introduction

Lines 31–33: The statement regarding NK receptors interacting with bacteria is too general and lacks mechanistic specificity.

Lines 37–38: Grammatical error: “purposes Moreover,” requires proper punctuation.

Lines 45–48: The knowledge gap is well stated, but the manuscript does not fully address receptor functionality.

Lines 49–54: The use of “Probably” is not appropriate for scientific writing. Replace with “It is hypothesized that…”

Aim

Lines 60–61: The objective is poorly structured. It should be rewritten to clearly state the mechanistic focus, e.g., “to characterize the mechanisms underlying…”

Results

TLR Expression

Lines 65–67: Grammatical error: “they are lack detectable…” should be “they lack detectable…”

Conceptual issue: the manuscript reports interaction with bacteria but shows absence of TLR2 and minimal TLR5. This contradiction is not adequately addressed.

Defensin Analysis

Line 73: Missing verb: “α-Defensins and β-defensins cationic antimicrobial peptides…” should include “are.”

Lines 77–79: Results are presented, but the number of replicates (n) is not specified.

Colony Formation Assays

Lines 89–94: Major issue: the results rely heavily on images (Figures 3–8; see pages 4–7), without proper quantification. There is no presentation of mean values, standard deviations, or statistical tests.

Confocal Microscopy

Lines 119–123: The analysis is descriptive only. There is no quantification of interaction (e.g., colocalization metrics), limiting interpretability.

LEVs Antimicrobial Effects

Lines 126–130: The rationale for selecting S. aureus and K. pneumoniae is insufficiently justified. The authors should clarify the selection criteria.

CFSE Transfer

Lines 141–144: The conclusion that this represents “cargo delivery” is overstated. CFSE transfer indicates association but not functional delivery.

Viability Assay

Lines 154–156: While the result is interesting, the methodology lacks detail, including gating strategy and appropriate controls for membrane integrity.

Antibiotic Susceptibility

Lines 163–165: Reported increases in inhibition zones (1–5%) are small and may not be biologically meaningful. Statistical analysis is required.

Discussion

Lines 226–229: The statement “we can expect the presence of TLRs” is speculative and unsupported.

Line 240: There is a contradiction between the lack of TLR4 expression and the observed response to LPS. This requires clarification.

Lines 260–263: Sentence structure is incorrect and should be revised.

Lines 273–275: Literature comparison is useful, but experimental conditions differ significantly.

Lines 285–288: The hypothesis regarding capsule-mediated resistance is plausible but not experimentally tested.

Lines 304–306: The proposed mechanism (membrane permeabilization) is speculative and lacks direct evidence.

Materials and Methods

Lines 335–344: LEV isolation relies only on differential centrifugation, which may result in contamination. Additional purification steps and validation are needed.

Lines 360–370: The number of biological replicates is not specified.

Line 368: Use of a microbiological loop for plating is not quantitative and reduces reproducibility.

Line 401: Typographical error: “S. auereus”

Conclusions

Lines 411–420: The conclusions are overstated. The evidence supports preliminary observations rather than definitive antimicrobial activity.

Minor Comments

Inconsistent bacterial nomenclature

Potential duplicate references (e.g., 19 and 25)

Author Response

Dear Reviewer,

Thank you for taking the time to review our work. We appreciate your attention to detail have carefully considered all of your comments.

Abstract

Lines 16–18: “This interaction leads to inhibition of colony formation…”. The wording implies causality, but only correlation is shown. Replace with “is associated with inhibition…”

Thank you for this comment. Done (line 18).

Lines 18–20: “mediated by defensin-α1”. This is not experimentally demonstrated and should be revised to reflect a hypothesis rather than a conclusion.

Replaced with: «Possible mechanism can involve defensin-α» (line 19).

Introduction

Lines 37–38: Grammatical error: “purposes Moreover,” requires proper punctuation.

Done.

Lines 31–33: The statement regarding NK receptors interacting with bacteria is too general and lacks mechanistic specificity.

Lines 45–48: The knowledge gap is well stated, but the manuscript does not fully address receptor functionality.

Dear reviewer, as this study aims to provide first insights into NK/LEVs/bacteria interactions, we did not investigate receptor functionality in mechanistic detail. However, based on surface expression of TLRs and defensins on both NK-92 cells and their LEVs, we hypothesize their involvement in mediating binding to bacterial targets.

Lines 49–54: The use of “Probably” is not appropriate for scientific writing. Replace with “It is hypothesized that…”

Done (line 49).

Aim

Lines 60–61: The objective is poorly structured. It should be rewritten to clearly state the mechanistic focus, e.g., “to characterize the mechanisms underlying…”

Replaced with “This study aims to elucidate the antimicrobial effects of NK-92 cells and their derived LEVs, focusing on their characteristics that mediate bacterial interaction.” (lines 60-61).

Results

TLR Expression

Lines 65–67: Grammatical error: “they are lack detectable…” should be “they lack detectable…”

Done (line 66).

Conceptual issue: the manuscript reports interaction with bacteria but shows absence of TLR2 and minimal TLR5. This contradiction is not adequately addressed.

Dear reviewer, we have previously demonstrated the presence of both receptors on the surface of parental NK-92 cells 10.3390/ijms26178449. We acknowledge the apparent contradiction noted in the Discussion (lines 231–237). This discrepancy is intentional and reflects a key observation of our pilot study: the lower surface density of TLRs on LEVs may contribute to their comparatively modest modulatory effect relative to parental NK-92 cells.

Defensin Analysis

Line 73: Missing verb: “α-Defensins and β-defensins cationic antimicrobial peptides…” should include “are.”

Done (Line 73)

Lines 77–79: Results are presented, but the number of replicates (n) is not specified.

Done. Lines 87-89.

Colony Formation Assays

Figures 3–10 are presented as representative micrographs to visually document the phenomenon of interaction between NK-92/NK-92-derived LEVs and ESKAPE bacteria. Our intention was not to derive quantitative metrics from these specific images, but rather to provide qualitative evidence of possibility of colony growth inhibition. We wish to highlight that also our work provides information characterizing the properties of the NK-92/LEVs as the mediators of the antibacterial effect. We conducted different experiments with wide range of methods. So, we believe that this pilot study's value lies in its integrated dataset, which provides a foundation for future work by our group and others on LEVs-based therapy.

Confocal Microscopy

Lines 119–123: The analysis is descriptive only. There is no quantification of interaction (e.g., colocalization metrics), limiting interpretability.

Dear Reviewer, we intentionally employed confocal microscopy as a qualitative imaging approach in this study to provide direct visual evidence of NK-92/bacteria interactions. Advanced sample preparation techniques would be required to find mechanisms of molecular dynamics. Such methodological deepening is planned for follow-up studies building on the foundational observations reported here.

LEVs Antimicrobial Effects

Lines 126–130: The rationale for selecting S. aureus and K. pneumoniae is insufficiently justified. The authors should clarify the selection criteria.

Dear reviewer! We selected these bacteria due to several reasons:

LEVs isolation is an expensive and time-/labor-consuming process. So we were interested in choosing limited number of bacteria.
aureus and K. pneumoniae. Coculturing with NK-92 had an effect on them so we suggested that LEVs also will have an effect.
Also, these 2 bacteria represent Gram- and Gram – so we could evaluate effect of LEVs on 2 different structures of the bacterial cell.

CFSE Transfer

Lines 141–144: The conclusion that this represents “cargo delivery” is overstated. CFSE transfer indicates association but not functional delivery.

Done 146-147.

Viability Assay

Lines 154–156: While the result is interesting, the methodology lacks detail, including gating strategy and appropriate controls for membrane integrity.

Dear reviewer, gating strategy is shown here. We used LIVE/DEAD™ BacLight™ Bacterial Viability Kit by Termo Fisher. It includes 2 dyes – PI and SYTO-9. While SYTO-9 stains all cells, dead cells concentrate PI and its intensity of fluorescence rises. We used unstained control (first picture in first line, controls stained with a single dye – SYTO-9 (2 picture) or PI (3 picture in first line).

Antibiotic Susceptibility

Lines 163–165: Reported increases in inhibition zones (1–5%) are small and may not be biologically meaningful. Statistical analysis is required.

Dear reviewer, weadded a clarification regarding the number of experiments with technical replicates; line 174-176). We also note that these data were not subjected to formal statistical analysis, as the disk diffusion assay was employed strictly as a preliminary screening step to assess the potential of LEVs to modulate bacterial susceptibility to antibiotics. After this step clindamycin was selected for subsequent MIC evaluation.

Discussion

Lines 226–229: The statement “we can expect the presence of TLRs” is speculative and unsupported.

It is shown that markers of parental cells are also expressed on their LEVs 10.3390/ijms222413663; also, TLR5 are detected both on NK-92 and their LEVs so we added this sentence in text.

Line 240: There is a contradiction between the lack of TLR4 expression and the observed response to LPS. This requires clarification.

In this study it was not planned to find mechanism of this phenomenon. But we can suggest that minor expression of TLR4 exists, but difficult for detection.

Lines 260–263: Sentence structure is incorrect and should be revised.

Done. 265-268.

Lines 273–275: Literature comparison is useful, but experimental conditions differ significantly.

Yes, this research differs but it also highlights antibacterial properties of NK-92 cells. As we found it is the only study in this area. Further comparison is needed.

Lines 285–288: The hypothesis regarding capsule-mediated resistance is plausible but not experimentally tested.

We added “But this suggestion needs more experimental evidences”, see lines 293-294

Lines 304–306: The proposed mechanism (membrane permeabilization) is speculative and lacks direct evidence.

Our study did not have an aim to explain the mechanism, so we had suggestion in Discussion part. Also we added “Further studies are required to elucidate the underlying mechanisms.” (313-314).

Materials and Methods

Lines 335–344: LEV isolation relies only on differential centrifugation, which may result in contamination. Additional purification steps and validation are needed.

Differential centrifugation is a main method of LEVs isolation. Due to “Minimal information for studies of extracellular vesicles (MISEV2023): From basic to advanced approaches” 10.1002/jev2.12404 it is allowed to use this method together with controlling methods. So, we controlled the size and stability of isolated LEVs used a Zetasizer NanoZS laser correlation spec-trometer (Malvern Instruments, Great Britain), a granulometric analysis was performed as well as zeta potential measurement. See lines 366-369.

Lines 360–370: The number of biological replicates is not specified.

We provided this information in lines 87-89.

Line 368: Use of a microbiological loop for plating is not quantitative and reduces reproducibility.

Dear reviewer, using of microbiological loop for plating is a widely accepted practice in microbiological workflows and aligns with GMP guidelines for single-use equipment (https://www.pharmaceuticalmicrobiology.in/2016/12/inoculation-loop.html#types). We also perform gravimetric validation using distilled water. Importantly, loops were used for even distribution of the bacterial suspension across the agar; not for volumetric transfer or quantitative inoculation. For this step we we used calibrated automated pipettes.

Line 401: Typographical error: “S. auereus”

Done.

Conclusions

Lines 411–420: The conclusions are overstated. The evidence supports preliminary observations rather than definitive antimicrobial activity.

Dear reviewer, we sincerely thank you for the time and thoughtful attention dedicated to reviewing our manuscript. During the revision process, we recognized "Limitations of the Study" part is necessary (lines 321–340). We believe this addition significantly improves the interpretability and balance of our conclusions.

Minor Comments

Inconsistent bacterial nomenclature

Here we use ATCC lines of ESKAPE bacteria, their strains are provided, possible typos are corrected.

Potential duplicate references (e.g., 19 and 25)

Done.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

I have evaluated the revised manuscript and the authors’ responses to the previous comments. The authors have satisfactorily addressed all concerns, and the manuscript has improved accordingly.