Identification of Key Bioactive Compounds of Medicine–Food Homologous Substances and Their Multi-Target Intervention Effects in Osteosarcoma Treatment

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study focuses on identifying key bioactive compounds derived from substances with a dual function as food and medicine for the treatment of osteosarcoma, a highly aggressive bone tumor. The researchers utilized transcriptomic data to identify differentially expressed genes, selected bioactive compounds through databases, and evaluated their interactions with target proteins. They identified several crucial molecules and genes, confirmed the stability of the molecular complexes with simulations, and validated the results with cellular experiments. In particular, ellagic acid dihydrate emerges as a potential multi-target agent to counteract osteosarcoma.

The conclusion that the effect observed is specifically due to ellagic acid is a deduction based on the conducted in vitro and  studies with Citrus aurantium extract. The evidence from computational biochemical and molecular analyses suggests that ellagic acid dihydrate has potential bioactive activity and shows strong binding to key molecular targets. This study highlights the presence of various bioactive compounds in Citrus aurantium extract, including ellagic acid; however, experiments using individual components were not performed. To obtain more definitive and clinically translatable results,  experimets should be repeated with purified single compounds, particularly ellagic acid, and assess whether their effects are additive or synergistic. This is essential for elucidating the precise mechanisms of action and optimizing therapeutic strategies.

However, without direct experimental studies isolating and testing ellagic acid alone, it cannot be definitively confirmed that all the biological effects of the extract are solely due to this molecule. The effects of a plant extract are often the result of complex interactions among multiple active compounds, which may act synergistically or additively.

Studies based solely on database research and molecular interaction simulations should not be published without experimental validation. Computational data should serve as a foundation for hypothesis generation rather than conclusive evidence. It is essential that these hypotheses be confirmed through in vitro or in vivo experiments to demonstrate the actual biological relevance and efficacy of the findings.

Moreover, while the text stresses the need for further validation and for considering the pharmacokinetic and bioavailability characteristics, it does not explicitly state that ellagic acid can harm healthy cells.

Furthermore, among the experiments, there is no MTT experiment on the osteosarcoma cell line indicating the degree of cytotoxicity of the extract.

Author Response

Comments1：The conclusion that the effect observed is specifically due to ellagic acid is a deduction based on the conducted in vitro and studies with Citrus aurantium extract. The evidence from computational biochemical and molecular analyses suggests that ellagic acid dihydrate has potential bioactive activity and shows strong binding to key molecular targets. This study highlights the presence of various bioactive compounds in Citrus aurantium extract, including ellagic acid; however, experiments using individual components were not performed. To obtain more definitive and clinically translatable results,  experimets should be repeated with purified single compounds, particularly ellagic acid, and assess whether their effects are additive or synergistic. This is essential for elucidating the precise mechanisms of action and optimizing therapeutic strategies.

However, without direct experimental studies isolating and testing ellagic acid alone, it cannot be definitively confirmed that all the biological effects of the extract are solely due to this molecule. The effects of a plant extract are often the result of complex interactions among multiple active compounds, which may act synergistically or additively.

 

Response1：Thank you for pointing this out.You rightly identified a limitation in our original experimental design—namely, that the use of an extract prevented precise attribution of the observed biological activity to specific single components identified through computational screening. To directly address and resolve this core concern, we have followed your suggestion by replacing the Citrus aurantium extract with the purified compound ellagic acid dihydrate in key cellular functional validation experiments. In the revised Methods section, we have explicitly stated that cell experiments were conducted by treating cells with different concentrations of ellagic acid dihydrate (lines 794–795). In the Results section, we report new experimental data demonstrating that ellagic acid dihydrate significantly inhibits the migration of 143B osteosarcoma cells in vitro, with effects consistent with the trends previously observed for the extract (lines 363–370).We fully acknowledge that while this modification directly addresses your main concern, the new experimental design itself still has limitations. Accordingly, we have expanded the Limitations section in the Discussion to reflect this.Once again, we sincerely thank you for your perceptive comment, which has directly prompted us to refine the experimental design of the study, thereby strengthening and making more direct the chain of evidence supporting our key conclusions. We believe these revisions have substantially enhanced the quality of the manuscript.

 

Comments2：Studies based solely on database research and molecular interaction simulations should not be published without experimental validation. Computational data should serve as a foundation for hypothesis generation rather than conclusive evidence. It is essential that these hypotheses be confirmed through in vitro or in vivo experiments to demonstrate the actual biological relevance and efficacy of the findings.

Response2：Thank you for raising this fundamental point, which is crucial for clarifying the value boundaries and evidence levels of different types of research. We fully agree with your core view: computational predictions without experimental validation have limited certainty and biological relevance, and the primary contribution of computational biology lies in providing high-value hypotheses for subsequent experiments. We would like to further elaborate on the positioning, design, and contributions of this study in the following aspects, hoping to gain your understanding.

1. Consensus and Clarification on the Nature of the Study
   We entirely agree that purely database mining or molecular simulation should not constitute the endpoint of a complete study. Therefore, in this work, we did not present the computational prediction results as final conclusions. Instead, we strictly followed a workflow of "computational screening → prioritization → preliminary validation." Our core effort was to develop and execute a systematic, multi-level integrative screening pipeline designed to filter out the "high-potential candidates" most worthy of experimental investment from a vast array of medicinal and edible ingredients through progressive steps. This methodological value is what our study aims to highlight.
2. Preliminary Experimental Validation Conducted
   In response to the scientific principle that "hypotheses must be validated," we have completed preliminary cellular validation of the core predictions within our means (as described in Sections 2.11 and 2.12 of the manuscript). Specifically, we selected ellagic acid dihydrate—the central node predicted to simultaneously target all five key genes—and performed preliminary functional and molecular tests on osteosarcoma 143B cells. Although these results are preliminary, they provide initial experimental support for our computational hypotheses, thereby closing the loop from "computational prediction" to "observation of preliminary biological effects." This goes beyond studies that remain purely at the computational level.
3. Acknowledgment of Limitations and Clarification of Future Directions
   We fully recognize that the current cell-based experiments remain limited in depth and mechanistic insight. To address your concern more rigorously, we have made important additions to the "Limitations and Future Perspectives" section in the Discussion (revised manuscript lines 549–558). We explicitly state that while the main findings of this study are based on bioinformatics and computational simulations and are supplemented by preliminary in vitro cellular validation, the conclusions still require confirmation through more in-depth functional gain/loss experiments in vitro and in vivo, as well as direct target-binding validation studies. At the same time, we have more clearly defined the positioning of this paper as a systematic hypothesis-generation and prioritization study. Its primary contribution lies in providing a rigorously computationally evaluated candidate list and molecular mechanism hypotheses with high potential for experimental validation. Finally, we have outlined plans and prospects for future research.

In summary, we understand the reviewer's strict emphasis on evidence standards. This study does not attempt to replace experiments with computation but aims to provide an efficient and reliable pipeline to reduce the blindness and cost of subsequent experimental research. We believe that this multi-level integrated analysis combined with preliminary validation represents a unique and meaningful research paradigm for exploring the potential value of complex systems, such as medicinal and edible substances.Thank you for your valuable feedback, which has prompted us to articulate these points more clearly. We hope that the above explanations and the corresponding revisions we have made to the manuscript will address your concerns.

 

Comments3：Moreover, while the text stresses the need for further validation and for considering the pharmacokinetic and bioavailability characteristics, it does not explicitly state that ellagic acid can harm healthy cells.

Response3:Thank you to the reviewer for raising the important issue regarding the safety of potential compounds. We fully agree with this concern.Firstly, we have replaced the Citrus aurantium extract with the purified compound ellagic acid dihydrate. Secondly, the concentrations used in our cellular functional validation experiments were selected within a safe range confirmed by preliminary tests, with the aim of evaluating their potential to inhibit tumor cell migration.We fully recognize the importance of systematic safety evaluation. To address this, we have explicitly added in the Discussion section of the manuscript that this study has not yet assessed the potential effects of these key active components on normal cells. Future research will prioritize selective toxicity testing on normal bone cells and evaluate their in vivo safety in animal models (lines 549-558、488-491 in the revised manuscript).Thank you for this valuable feedback, which has further refined our research planning.

 

Comments4：Furthermore, among the experiments, there is no MTT experiment on the osteosarcoma cell line indicating the degree of cytotoxicity of the extract.

Response4：Thank you for raising the important question regarding the quantitative method for assessing cytotoxicity. We understand that the MTT assay is one of the classic methods for evaluating compound cytotoxicity. In this study, we chose to use the CCK‑8 assay to measure cell viability, which operates on a similar principle to the MTT assay and is widely employed for assessing cell proliferation and activity. Through preliminary experiments, we determined the concentration range used (0 to 640 µM). Within this range, the extract effectively inhibited cell migration (scratch assay) without causing significant reduction in cell viability (CCK‑8 assay), suggesting that its effect may be more focused on inhibiting migration rather than direct cytotoxicity.We acknowledge that the current study does not provide a complete characterization of dose‑dependent cytotoxic effects. We have explicitly addressed this point in the "Limitations" section of the Discussion (lines 549–558 in the revised manuscript) and will conduct a comprehensive assessment of its cytotoxic profile in follow‑up studies.Thank you for your suggestion, which has helped us to more clearly outline the necessary subsequent experiments

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript of Jie Ren et al, describes the identification of some compounds which may be useful in treatment of osteosarcoma. 

A large part of paper presents in silico research (databases processing) and some in vitro assays. 

 

First, the paper is difficult to be read, a better workflow presentation may be useful.

Figures 2, 3, 4, 5, 6, 7 and 8 are too small and difficult to interpret. Authors replace the original images with larger ones.

The name of compounds or genes in the first column of Table 1 and  3, respectively,  are overlapping and difficult to read. A larger first column may correct the issue. 

The measurement unit in Table 3 is missing.

Figure 4 must be changed, because there is impossible to understand the binding of compounds. 

It is unclear the length of MD simulation - because in section 2.5 is written 20 ns, and in Figure 5 is written 100000 ps. 

Figure 5 must be changed, because the data series are overlapping. 

In the molecular docking study it is unclear what site was targeted for each protein. Was the active site of another site? A blind docking was performed?

The assay presented at 2.11 is not fitting with the present manuscript, because there is no characterization of the extract. The extract has a lot of compounds and the present paper focuses on specific compounds. The assay on cells must be performed on solutions of compounds, not a mixture of compounds.

Line 411: the authors are wrong, quercetin and luteolin are not polysaccharides.

Author Response

Comments1：The manuscript of Jie Ren et al, describes the identification of some compounds which may be useful in treatment of osteosarcoma. A large part of paper presents in silico research (databases processing) and some in vitro assays.

1.First, the paper is difficult to be read, a better workflow presentation may be useful.

Response1：Thank you very much for your valuable suggestions. We fully agree that a clear presentation of the workflow is crucial for improving the readability of the paper. In response to your comments, we have made three key improvements to the manuscript:

1. Optimized the Graphical Abstract: We have refined the graphical abstract to visually summarize the complete research workflow from data mining to experimental validation, incorporating stepwise labels such as "Step 1, Step 2," etc., to help readers quickly grasp the overall structure of the study.
2. Restructured the Workflow Description: In the second paragraph of the Discussion, we have systematically outlined and elaborated on the integrative strategy and steps adopted in this research to enhance the clarity of the methodological logic (lines 398–418 in the revised manuscript).
3. Added Transition Statements: At the beginning of Section 2.11 in the Results (lines 356–358 in the revised manuscript), we have included explicit introductory sentences to indicate that this section marks the transition from the "computational prediction" stage to the "preliminary experimental validation" stage.

 

Comments2：Figures 2, 3, 4, 5, 6, 7 and 8 are too small and difficult to interpret. Authors replace the original images with larger ones.

Response2：Thank you very much for your important feedback regarding the readability of the figures. Following your suggestion, we have uniformly revised all images, including increasing their dimensions and adjusting font sizes. The updated figures have been replaced with high-resolution versions to ensure all details are clearly legible.

 

Comments3：The name of compounds or genes in the first column of Table 1 and  3, respectively,  are overlapping and difficult to read. A larger first column may correct the issue.

Response3：Thank you for pointing out the formatting issues in the tables. We have carefully revised all tables in the manuscript (including Tables 1 and 3) in accordance with your suggestions. Specific improvements include adjusting column widths, optimizing text wrapping, and refining alignment to ensure that entries such as compound and gene names are clearly displayed without overlap. The readability of the revised tables has been significantly enhanced.

 

Comments4：The measurement unit in Table 3 is missing.

Response4：Thank you very much for your detailed feedback. We have clearly added the unit of binding free energy (kcal/mol) to the corresponding numerical column in Table 3. We sincerely apologize for any inconvenience caused and appreciate your help in improving the details of our manuscript.

Comments5：Figure 4 must be changed, because there is impossible to understand the binding of compounds.

It is unclear the length of MD simulation - because in section 2.5 is written 20 ns, and in Figure 5 is written 100000 ps.

Response5：Thank you very much for your valuable comments on the details of the figures and the rigor of the data. Following your suggestions, we have re-optimized the molecular docking result figures. The specific adjustments are as follows: the protein structure is rendered in translucent light gray, the ligand (key active compound) is highlighted with a bright yellow ball-and-stick model, and a scale bar has been added. This optimization clearly shows the spatial position and conformation of the ligand within the protein binding pocket, avoiding structural overlap. Additionally, we have included color definitions for the lines representing key interactions such as hydrogen bonds to clarify the binding mechanism.Regarding the MD simulation duration, thank you for pointing out this critical inconsistency. After verification, the molecular dynamics simulation duration for all systems is 100 ns (100,000 ps). The labeling in Figure 5 is correct. Accordingly, we have revised the description of the simulation duration in the Results section of the manuscript (lines 210-211 in the revised version) to ensure rigorous consistency of all data throughout the paper.

 

Comments6：Figure 5 must be changed, because the data series are overlapping.

Response6：Thank you very much for pointing out the issue of overlapping data lines in Figure 5. To clearly present the data from each group, we have optimized this figure: we attempted to adjust line styles and add data markers, but the overlapping issue persisted. Therefore, we ultimately plotted the data for different groups in separate subplots. This adjustment effectively resolves the problem of overlapping lines and significantly enhances the readability and data presentation of the figure.

 

Comments7：In the molecular docking study it is unclear what site was targeted for each protein. Was the active site of another site? A blind docking was performed?

Response7：Thank you very much for raising this important technical question. In the molecular docking step of this study, we employed a blind docking approach without predefined binding sites. We have clearly described and refined this procedure in the Methods section (Section 4.8, lines 664-665 of the revised manuscript) to ensure clarity in both the experimental design and its description.

 

Comments8：The assay presented at 2.11 is not fitting with the present manuscript, because there is no characterization of the extract. The extract has a lot of compounds and the present paper focuses on specific compounds. The assay on cells must be performed on solutions of compounds, not a mixture of compounds.

Response8：Thank you very much for raising this crucial point. You correctly identified the core limitation in our previous experimental design—that using a partially characterized crude extract could not provide direct evidence for the key conclusion that "a specific single compound (such as ellagic acid dihydrate) is the active component." This indeed weakened the persuasiveness of the study. To thoroughly address this issue and ensure precise alignment between the experimental evidence and the research focus, we have strictly followed your suggestion and made key revisions to the experimental section.Specifically, we have replaced the Citrus aurantium extract in Section 4.14 (Functional Cell Assays) with commercially purchased ellagic acid. This ensures that the experiments directly validate the target compound identified through computational screening. Accordingly, all functional data in the Results section (Sections 2.11 and 2.12), including the inhibition of cell migration in the scratch assay and changes in the expression of key target genes detected by qRT-PCR, have been updated to reflect results obtained using the pure compound. The data show that this single compound is sufficient to reproduce the previously observed trends of migration inhibition and regulation of key gene expression seen with the extract.This fundamental adjustment successfully bridges the logical gap between computational screening and experimental validation, ensuring high consistency throughout the manuscript around the central theme of "identifying and validating a specific lead compound." It significantly strengthens the certainty and persuasiveness of the core conclusions.

 

Comments9：Line 411: the authors are wrong, quercetin and luteolin are not polysaccharides.

Response9：Thank you very much for pointing out this important classification error. We sincerely apologize for this oversight. In line 385 of the manuscript (and related sections throughout the text), we have corrected the category of quercetin and luteolin to "flavonoids." We have also thoroughly reviewed the entire manuscript to ensure the accuracy of such factual descriptions.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have responded to all my comments, and the manuscript is now acceptable as it stands.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors provided answer in Chinese. English is mandatory for communication with MDPI

Comments for author File: Comments.zip

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

The authors made the requested changes.