Novel 2D/3D Hybrid Organoid System for High-Throughput Drug Screening in iPSC Cardiomyocytes

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The work "Novel 2D/3D Hybrid Organoid System for High-Throughput Drug Screening in iPSC Cardiomyocytes" by Lewis, J., and colleagues is highly intriguing, and the proposed model holds promise in the field. However, several questions arise when the study is closely analyzed, which the authors should address for it to be fully considered.

a) The authors propose a 2D/3D model and analyze calcium dynamics and contractile coupling after doxorubicin treatment. However, much of the results focus on comparing proximal and distal sections, rather than demonstrating that the cells behave similarly, regardless of their distance from the center, with or without doxorubicin. This comparison is critical to validate the model’s reliability and utility for drug screening. The statistical analysis presented compares the different sections with one another, but not with the center/organoid itself. This raises two concerns: First, it seems the objective of Figures 2 and 4 is to show that all sections behave differently compared to the center. Second, the pairwise comparisons make it difficult to determine whether the "far" areas are similar to the center. As a reader, I couldn’t find the relevance of these comparisons.

b) I also found it unclear why the authors used three concentric circles, each spaced 185 microns apart. Is this related to their methodology? Should someone using the same method always apply this specific distance to obtain comparable data? The authors should explain why this particular distance was chosen.

c) A critical aspect of the study is that much of the data comes from dissociated cells from the organoid. The authors should expand on this in the discussion. If the distal section shows lower cell density, it suggests that cardiomyocytes may not be forming a syncytial structure, potentially affecting results related to electromechanical coupling. Similarly, this could reduce the paracrine effects of cardiomyocytes on calcium transients. If this is influencing the results, would it be more reliable to use data from the close and intermediate sections? Do these sections behave similarly?

d) The data presentation could be improved for clarity. In the graphs, the authors should consider using asterisks or other symbols to indicate statistical significance. For the calcium-related dot plots, circles could be added to indicate data from close, mid, and far areas.

e) Regarding data analysis, I noticed that the error bars are quite long, yet statistical significance is still reported. For instance, in Figure 4h, the control center and close sections are significantly different, but the large error bars raise doubts about whether this difference is meaningful. A violin plot could better illustrate the data and address this concern. Additionally, the authors should clarify whether they performed normality tests before choosing to use ANOVA and t-tests.

f) The discussion is heavily focused on doxorubicin’s effects. However, comparisons against doxorubicin were not made for each section. The authors should be more careful with the affirmations or include these comparisons to support their claims.

g) If possible, the authors should include representative videos of the different conditions as supplementary information.

Author Response

Please see attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The study identifies numerous strengths of the 2D/3D hybrid organoid model for measuring cardiotoxicity. A crucial strength is the model's uniqueness, which successfully combines the benefits of both 2D and 3D cultures, allowing for a more realistic assessment of pharmacological effects. The full examination of calcium dynamics, contractility, and signal propagation provides a complete picture of the drug's action. Additionally, the model's high-throughput capabilities allows for efficient screening of many medicines, which is critical in drug development. The use of machine learning to identify calcium transients is quite novel and can help mproves accuracy and minimizes human error, resulting in greater analytical efficiency, while the use of hiPSC-CMs raises the findings' relevance to human cardiac health.

The study has a few details, including a narrow focus on doxorubicin, which limits the model's applicability to other drugs and necessitates additional validation with a broader range of compounds. Additionally, the sequential imaging of calcium transients and contractility may overlook important interactions that occur concurrently, would be advisable to view at a molecular level interactions.

 

Revise Conciseness: Some sentences are a bit lengthy and could be streamlined for clarity. For instance, breaking up complex sentences may enhance readability. Consider briefly explaining terms like "microelectrode arrays" or "optical analysis of calcium transients" for broader accessibility.

 

Consider Investigating a range of drug concentrations over multiple time points to better understand the dose-dependent effects and temporal dynamics of cardiotoxicity.

 

While study center on doxi consider other cardiotoxic drugs. This would help assess the model's generalizability across different pharmacological profiles.

Consider evaluating molecular level expression of pumps e.g., SERC2A

 

 

Comments on the Quality of English Language

Overall well written, just revise a few sentences

Author Response

Please see attachment

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The Manuscript «Novel 2D/3D Hybrid Organoid System for High-Throughput Drug Screening in iPSC Cardiomyocytes» is devoted to the current and rapidly developing topic of using human iPSC-derived cardiomyocytes to study the mechanism of action of pharmaceutical and toxic drugs. The authors present a methodological approach that should simplify the analysis of the contractile characteristics of cardiomyocytes from 3D culture. The paper presents the proposed approach and results in detail. There are a number of questions arise regarding the model, experimental methodology, and discussion of the data obtained.

In general

1. The main question concerns the experimental model itself. The authors write that they work with a 2D/3D Hybrid Organoid System, but 2D and 3D models have significant differences. The authors claim that the presented 2D/3D system is similar in its properties to 3D organoid: «Our model circumvents these discrepancies because the isolated single “2D” cells in our hybrid model are developed as part of the organoid, and not as monolayers, thus preserving properties of true cardiomyocytes and cell-cell interactions…» (P. 12, Line 306). Are the authors sure that after 10 days the detached cells retain properties similar to those of the cells in the organoids? First of all, authors should provide evidence that in the presented system, cells outside the organoid are similar to cells in the organoid. Meanwhile, in the discussion, the authors mainly talk about the fact that the presented 2D/3D Hybrid Organoid System reflects the effects of doxorubicin well.

2. It is not clear from the description of a 2D/3D Hybrid Organoid System whether cells outside the organoid interact with the organoid and with each other. If cells outside the organoid interact with each other, this must be indicated when describing the model. Otherwise, the term «signal propagation» is a misnomer.

3. Please explain the reason for choosing a distance of 185 microns.

4. Usually the calcium transition is described by such characteristics as time to 50% peak, time to 50% decay, etc., as done in the work (Gentillon et al. J Mol Cell Cardiol. 2019;132:120-135. doi: 10.1016/j.yjmcc.2019.05.003) with the participation of the authors of the presented Manuscript. In addition, it is necessary to indicate how characteristics of calcium transition and contraction presented in the Manuscript correlate with those previously obtained on iPSC-derived cardiomyocytes.

5. The discussion is quite extensive and needs to be structured to make it easier to understand.

6. The discussion is overloaded with descriptions of the mechanisms of action of doxorubicin described in previous works, but there is no explanation of the effects obtained in the presented work. For example, it is necessary to explain the spatial difference in the characteristics of the calcium transition and contractile properties.

7. There are several controversial points in the discussion. For example: «Our results demonstrated diastole prolongation with increased distance from the beating organoid, suggesting that Dox cardiotoxicity amplifies across a signal» (P. 12, Line 317). What is the mechanism of this signal?

Minor points

Line 78 and further Better to replace «degrees Celsius» with «°C».

Line 169 There is no period at the end of the sentence.

Line 287 Hyphen before the word calcium. In this sentence, you need to add references.

Fig 1-3. It is necessary to indicate in the figure caption which statistical method was used and in what form the experimental values ​​are presented. It is also necessary to indicate the number of cells examined.

It would be better to separate the Limitations and Conclusion into a separate section.

Author Response

Please see attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Despite the addition of videos addressing most of the initial inquiries related to the model, the authors have not sufficiently deepened their initial question: "We sought to design a novel 2D/3D hybrid organoid system relying on optical techniques and spatial segmentation to obtain electromechanical information on calcium dynamics, cardiomyocyte function, and signal propagation in a high-throughput manner that retains the complexity of a 3D organoid."

Most of the article focuses on demonstrating the functionality of their model using a well-known cardiotoxicity drug. While the presented information is essential to establish the model’s relevance, the described results are not enough to support the claim that the 3D/2D model is superior to others. For example, Figure 1 primarily illustrates how Dox treatment affects the model’s behavior, but the model itself had not been described previously.

Interestingly, the authors state: "Our untreated iPSC-organoid hybrids showed similar calcium transient and contractility properties as those reported for isolated cardiomyocytes in the literature" (lines 358–359). The data supporting this statement should be presented separately and incorporated into Figure 1. Additionally, details such as cell density and other section characteristics should be included to ensure reproducibility. Once this foundational description is provided, the authors could then proceed with their analysis of Dox effects at different levels, as they have already done.

Finally, after this description, the authors should summarize their findings in a table, categorizing features by section (close, mid, far). This would be a valuable reference for readers to identify key characteristics and determine which section could be most useful for assessing and predicting drug-induced cardiotoxicity.

These proposed changes should strengthen the results and enhance the model’s relevance in cardiovascular therapeutics.

Author Response

We thank the reviewers for their comments and have addressed them as noted below and highlighted in blue in the manuscript

Comment 1: Despite the addition of videos addressing most of the initial inquiries related to the model, the authors have not sufficiently deepened their initial question: "We sought to design a novel 2D/3D hybrid organoid system relying on optical techniques and spatial segmentation to obtain electromechanical information on calcium dynamics, cardiomyocyte function, and signal propagation in a high-throughput manner that retains the complexity of a 3D organoid."
Most of the article focuses on demonstrating the functionality of their model using a well-known cardiotoxicity drug. While the presented information is essential to establish the model’s relevance, the described results are not enough to support the claim that the 3D/2D model is superior to others. For example, Figure 1 primarily illustrates how Dox treatment affects the model’s behavior, but the model itself had not been described previously.

Response 1: We thank the reviewers for their comment. While we do believe our model offers a NOVEL innovative way to study cells that may improve some of the limitations of separate 2D cellular sheets and 3D organoids, we do not directly compare other models to make the claim of superiority.  The novelty of the model is a 2D/3D hybrid system that provides superior data than either 2D or 3D model alone, whereby the 3D organoid provides enhanced and coordinated contractility and the 2D model allows isolation of fluorescence signals away from interference of overlapping cells to capture calcium dynamics. We have added additional data (New Figure 1) that demonstrates the limitations of separate 2D cellular sheets and 3D organoids that we have noted. Additionally, we have now further described the limitations and advantages of 2D versus 3D models in isolation before the discussion of the toxicity of Doxorubicin and added that as Figure 1 (Results section also added Line 174-194). 

Comment 2: Interestingly, the authors state: "Our untreated iPSC-organoid hybrids showed similar calcium transient and contractility properties as those reported for isolated cardiomyocytes in the literature" (lines 358–359). The data supporting this statement should be presented separately and incorporated into Figure 1. Additionally, details such as cell density and other section characteristics should be included to ensure reproducibility. Once this foundational description is provided, the authors could then proceed with their analysis of Dox effects at different levels, as they have already done.
Response 2: The intention of this statement was originally misleading as written and we appreciate you pointing this out. Our intent was to point out that the hybrid model can capture calcium transients and contractility and the corresponding abnormalities after exposure to dox which is reflected in our statement “Our results reflect what has been elucidated from molecular studies of Dox cardiotoxicity and abnormal calcium handling”. Cell density data has been included in Figure 2c which provides the number of cells per square area in the visual field, broken down by treatment status and spatial region.

Comment 3: Finally, after this description, the authors should summarize their findings in a table, categorizing features by section (close, mid, far). This would be a valuable reference for readers to identify key characteristics and determine which section could be most useful for assessing and predicting drug-induced cardiotoxicity. These proposed changes should strengthen the results and enhance the model’s relevance in cardiovascular therapeutics.
Response 3: We thank the reviewers for their comment. We have added a supplemental table that summarizes the relationship of spatial features of our model in a table (supplemental table 1) 

Reviewer 3 Report

Comments and Suggestions for Authors

The authors took into account some of the comments, but, unfortunately, important questions remained unanswered.

Point 1: The authors did not answer the first question about the experimental model. The authors write that they work with a 2D/3D Hybrid Organoid System, but 2D and 3D models have significant differences. The authors claim that the presented 2D/3D system is similar in its properties to 3D organoid. But the authors did not provide definitive proof that the cells inside and outside the organoid are identical. Since the work is devoted to a new experimental model, it is necessary to prove its adequacy not only with doxorubicin and without its addition.

I think that the direct comparison of contractility and calcium transients for the three control group regions at the very beginning of the results before describing the effects of doxorubicin was such proof. This data is presented in the machine learning section however, they are not given special attention.

Page 287-289 The authors write «In the control group, durations for systole and diastole varied across the different spatial regions, but there were no consistent findings of prolongation or shortening based on distance from the organoid.»

I believe that the proposed model can be adequate if the characteristics of 3D organoid and cells outside are identical. What causes the difference in characteristics?

Point 2: It is not clear from the description of a 2D/3D Hybrid Organoid System whether cells outside the organoid interact with the organoid and with each other. If cells outside the organoid interact with each other, this must be indicated when describing the model. Otherwise, the term «signal propagation» is a misnomer.

Response: We thank the reviewer for their comment. The cells outside the organoid do interact with it and we have now addressed this in our results on page 8: “Despite signal duration changes on a millisecond scale in both control and Dox-treated organoids and cells, calcium fluorescence was grossly synchronized among Center, Close, Intermediate, and Far regions, indicating that the 2D cells retained connection to the 3D organoid and its properties, which provided validation to our hybrid model and methodology.”

The proposed answer is very circumstantial evidence. To prove connections of cells, you can provide an image of the cells.

Author Response

Thank you for your comments. All updated texts in the manuscript are highlighted in blue. 

Comment 1: The authors took into account some of the comments, but, unfortunately, important questions remained unanswered.
Point 1: The authors did not answer the first question about the experimental model. The authors write that they work with a 2D/3D Hybrid Organoid System, but 2D and 3D models have significant differences. The authors claim that the presented 2D/3D system is similar in its properties to 3D organoid. But the authors did not provide definitive proof that the cells inside and outside the organoid are identical. Since the work is devoted to a new experimental model, it is necessary to prove its adequacy not only with doxorubicin and without its addition. I think that the direct comparison of contractility and calcium transients for the three control group regions at the very beginning of the results before describing the effects of doxorubicin was such proof. This data is presented in the machine learning section however, they are not given special attention.

Response 1: We thank the reviewers for their comment. We have added a new figure 1 to account for the reviewers comments and provided additional data and videos that demonstrate the identical synchronicity of the cells inside and outside the organoid in Figure 1h and 1i with respect to contractility and fluorescence.  New Figure 1 has cross correlative analysis on the model without the addition of doxorubicin, as suggested by the reviewer to first prove its adequacy. For each model, the strongest peaks were at a lag of 0, indicating the signals had near perfect synchronicity for contractility and calcium fluorescence, thereby demonstrating their connectivity and identicality.  

Comment 2: Page 287-289 The authors write «In the control group, durations for systole and diastole varied across the different spatial regions, but there were no consistent findings of prolongation or shortening based on distance from the organoid.» I believe that the proposed model can be adequate if the characteristics of 3D organoid and cells outside are identical. What causes the difference in characteristics?
Response 2: We thank the reviewer for their comment. We have included additional data in Figure 1 to demonstrate the high degree of synchronicity of the cells inside and outside the organoid in the control group, from which we have concluded the cells are highly interconnected and identical. While the cells beat and fluoresce synchronously as is demonstrated by the videos as well as our data in the control samples, there can be slight changes in the characteristics of the beats/fluorescence from cell to cell. The statement was meant to capture that there may be slight differences in sub-contractile properties such as systole and diastole. 

Comment 3: Point 2: It is not clear from the description of a 2D/3D Hybrid Organoid System whether cells outside the organoid interact with the organoid and with each other. If cells outside the organoid interact with each other, this must be indicated when describing the model. Otherwise, the term «signal propagation» is a misnomer.
Response 3: We thank the reviewer for their comment. The cells outside the organoid do interact with it as they are extension of the organoid and we have now addressed this in our results on (lines 232-234): “Despite signal duration changes on a millisecond scale in both control and Dox-treated organoids and cells, calcium fluorescence was grossly synchronized among Center, Close, Intermediate, and Far regions, indicating that the 2D cells retained connection to the 3D organoid and its properties, which provided validation to our hybrid model and methodology.”

Comment 4: The proposed answer is very circumstantial evidence. To prove connections of cells, you can provide an image of the cells.

Response 4: We thank the reviewer for their comment. We have provided videos of our cells beating and fluorescing synchronously with the organoids as proof of connection of cells in our manuscript (Figure 2b and 5a). Furthermore, cross-correlative analysis demonstrated high synchronicity indicating the cells are activated by the same signal and therefore connected in our New Figure 1 and the corresponding results section (lines 174-193).