Heterotypic 3D Model of Breast Cancer Based on Tumor, Stromal and Endothelial Cells: Cytokines Interaction in the Tumor Microenvironment

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

We have carefully reviewed the manuscript Cells-4032312 entitled “Heterotypic 3D Model of Breast Cancer Based on Tumor, Stromal and Endothelial Cells: Cytokines Interaction in the Tumor Microenvironment” by Anastasia Leonteva’s group.

The topic is highly relevant, as understanding the physiology and dynamics of cancerous tissues is a major challenge that can no longer be addressed using simplistic models such as 2D cultures. Many research groups have invested significant effort in developing more representative models, and this study contributes to that effort by focusing on breast cancer.

I would like to emphasize the quality and innovative nature of the work, which aligns with the direction taken by an increasing number of studies in cancer research and beyond. There are several points that could further strengthen the manuscript.

 

Major Suggestions

Expand Biological Characterization; Cancer hallmarks include significant metabolic changes and heterogeneity within tumors; Stiffness, particularly that contributed by the stroma, plays an important role in tumor progression and therapy resistance; Consider adding basic assays such as: Immunofluorescence for key markers; Metabolic tests; Mechanobiological evaluations. These would provide valuable complementary data.

 

Therapeutic Relevance: Including a treatment with a chemotherapeutic agent could demonstrate the model’s utility. It would be interesting to assess whether treatment efficacy varies according to spheroid composition. The effect of adding estrogen could also be explored.

 

Dynamic Monitoring and Molecular Characterization: A dynamic follow-up of spheroids, particularly regarding cytokine secretion, would add depth.

Consider a more detailed molecular characterization, for example: Basement membrane components; Vessel markers (CD31, VWF, NG2); Hypoxia indicators

 

Have the authors considered adding ascorbic acid to enhance extracellular matrix deposition?

 

Specific Comments on Figures and Text

Introduction (Line 81)

This section reads somewhat like results. It might be clearer to move it to the Results section and include Figure 1 there.

Line 297

Suggest replacing “cells” with “other cells” for clarity.

 

Figures

Figure 2a: Lacks detail—please clarify what is shown.

Figure 2d: Image appears slightly blurry.

Figure 3 (Line 349): “Vessel-like structures” are not easily visible. Molecular characterization and lumen diameter measurements would help.

Note: Since there is no circulation inside these vessels, oxygen and nutrients are likely supplied by the surrounding medium, which is an important consideration for tumor physiology.

Figure 4: Consider moving the bottom-left panel fully to the left and placing the legend at the bottom right for easier reading.

Western Blot

Could different fibronectin isoforms (e.g., EDA) also be examined?

 

Overall Recommendation

The manuscript addresses an important topic and presents promising results. Incorporating some of the above suggestions—particularly additional characterization and minor figure adjustments—would significantly enhance the impact and clarity of the study.

Author Response

Reviewer 1.

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files. 

Comments for Reviewer 1:

Point 1. Expand Biological Characterization; Cancer hallmarks include significant metabolic changes and heterogeneity within tumors; Stiffness, particularly that contributed by the stroma, plays an important role in tumor progression and therapy resistance; Consider adding basic assays such as: Immunofluorescence for key markers; Metabolic tests; Mechanobiological evaluations. These would provide valuable complementary data.

Response 1: It is acknowledged that the suggestions put forward are to be welcomed, and it is believed that the expansion of the biological characteristics of spheroid models in the directions proposed will result in a significant increase in their value and translational potential. In the new version of the article, a concise account of the biological characteristics of tumours has been incorporated, along with pertinent details regarding cell models, within the introduction section. Furthermore, an array of mechanobiological studies was conducted on matrix gel and gelatin, alongside an analysis of the cancer stem cell population within the model. Concomitant with this, the discussion has undergone an expansion in scope. See line (56-75, 81-102, 452-508, 580-601,691-708, 740-750, 846-858, 902-922).

Point 2. Therapeutic Relevance: Including a treatment with a chemotherapeutic agent could demonstrate the model’s utility. It would be interesting to assess whether treatment efficacy varies according to spheroid composition. The effect of adding estrogen could also be explored.

Response 2: We warmly welcome your ideas and plan to include the following research in our future work. The incorporation of these prospective studies within the designated сonclusions section will be facilitated by the establishment of direct links to the relevant comments, the purpose of which is to provide a visual representation of the project's developmental trajectory. These experiments will form the basis for a separate, more applied study on personalised drug testing and the role of the microenvironment. See line  81-102, 846-858, 888-890. Moreover, it should be noted that comparisons between the sensitivity of cells in two and three dimensions have been made on numerous occasions. For instance, in a recent publication by the authors https://doi.org/10.1038/s41598-025-12556-9. Consequently, in order to produce higher-quality work that truly demonstrates the potential of heterotypic spheroid models as a platform for screening, significantly more time and careful refinement of screening and dose selection methods are required. This is our priority task in future research.

Point 3. Dynamic Monitoring and Molecular Characterization: A dynamic follow-up of spheroids, particularly regarding cytokine secretion, would add depth.

Response 3: We fully agree with you regarding the scientific value of this approach. However, within the scope of this study, which focused on optimising the creation and basic morphological and phenotypic validation of the 3D model, we evaluated cytokines as a whole. In the future, we will focus on determining which cytokines are more important for our research, including from a therapeutic point of view, and will target the evaluation of dynamic changes. However, we have added a dynamics analysis to study cell migration and invasion. See line 452-510 and Figure 4,5 and S1, S5.

Point 4. Consider a more detailed molecular characterization, for example: Basement membrane components; Vessel markers (CD31, VWF, NG2); Hypoxia indicators

Response 4: We are pleased to acknowledge your contribution to the development of a comprehensive plan for the second, molecularly oriented phase of our project. The present article demonstrates that the presence and mechanobiological role of labelled ECs and various types of fibroblasts are confirmed by data on changes in the secretion of factors such as VEGF, PDGFR, and FGF (See Figure 7b and S7). Furthermore, it has been posited that a comprehensive molecular characterisation of their status should be accorded a high priority for future publications.

Point 5. Have the authors considered adding ascorbic acid to enhance extracellular matrix deposition?

Response 5: While this is not the case in the present work, the comment is of great value in terms of informing future research planning.

Point 6. Introduction (Line 81). This section reads somewhat like results. It might be clearer to move it to the Results section and include Figure 1 there.

Response 6: We concur with the sentiments expressed in your commentary and have accordingly relocated this block to the results section. See section of 3.4. Protein-protein interaction network of cytokines and growth factors via Search Tool for the Retrieval of Interacting Genes/Proteins of database in current version manuscript. 

Point 7. Line 297. Suggest replacing “cells” with “other cells” for clarity.

Response 7: It has been corrected.

Point 8. Figure 2a: Lacks detail—please clarify what is shown.

Response 8: It has been corrected.

Point 9. Figure 2d: Image appears slightly blurry.

Response 9: Thank you for carefully studying the micrographs and providing this valuable technical note. We fully agree that achieving high image resolution in three-dimensional structures such as spheroids is a challenging task. Fluorescence microscopy was used to confirm this. Confocal microscopy would have yielded significantly better results. However, due to the characteristics of the FDA staining protocol and the rapid release of dye from living cells, we used fluorescence microscopy. We repeated the experiment to improve image quality. See Figure 1d. Blurring can occur due to a number of factors characteristic of thick, light-scattering samples. A spheroid is a dense three-dimensional sphere of cells and extracellular matrix. Passing through different layers with different refractive indices, the light beam is deflected and scattered, reducing contrast and resolution, especially at the depth of the sample. Focusing at the depth of the spheroid also causes spherical and chromatic aberrations.

Point 10. Figure 3 (Line 349): “Vessel-like structures” are not easily visible. Molecular characterization and lumen diameter measurements would help.Note: Since there is no circulation inside these vessels, oxygen and nutrients are likely supplied by the surrounding medium, which is an important consideration for tumor physiology.

Response 10: When developing approaches to produce 3D heterotypic spheroid models of human tumours, it is important to consider the specifics of vascularisation. To this end, we incorporate endothelial cells into the 3D models. While these cells can form 2D capillary-like structures in Matrigel, these “classical” structures cannot be formed in 3D models. However, we observed the distribution of endothelial cells during the self-assembly and lifespan of the 3D spheroid. We therefore interpreted the clusters of endothelial cells that gathered at one pole of the spheroid as possibly representing the initial stage of capillary-like structure formation in tumor. We also reanalysed our confocal data and created a new figure. See Figure 2.

 

Point 11. Figure 4: Consider moving the bottom-left panel fully to the left and placing the legend at the bottom right for easier reading.

Response 11: It has been corrected.

Point 12. Western Blot. Could different fibronectin isoforms (e.g., EDA) also be examined?

Response 12: Indeed, two isoforms of fibronectin are observed. The 250 kDa subunit has been identified as fibronectin monomer https://www.abcam.com/en-us/technical-resources/target-tips/fibronectinfn1. It is the primary constituent of both soluble plasma fibronectin and insoluble cellular fibronectin. This larger form as 272 kD is one of numerous isoforms produced by the process of alternative splicing of a single gene. The WB results have been recalculated. See Figure 3b.

Point 13. Overall Recommendation. The manuscript addresses an important topic and presents promising results. Incorporating some of the above suggestions—particularly additional characterization and minor figure adjustments—would significantly enhance the impact and clarity of the study.

Response 13: We are immensely grateful for your overall positive assessment of the significance of our research and the results presented, and for the detailed and constructive suggestions you have provided. We are in full agreement with the recommendation as a whole. The incorporation of the aforementioned improvements, particularly those pertaining to the expanded biological and molecular characterisation of the models and the technical refinement of the graphical materials, is indeed a necessary step to enhance the scientific rigour, clarity, and, consequently, impact of our work.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In the manuscript titled “Heterotypic 3D Model of Breast Cancer Based on Tumor, Stromal and Endothelial Cells: Cytokines Interaction in the Tumor Microenvironment”, the authors present a co-culture model incorporating breast cancer, stromal, and endothelial cells. While the recreation of the tumour microenvironment remains a critical challenge in experimental oncology, this study aims to elucidate the interplay among these cell types within heterotypic spheroids. The hypothesis is of interest; however, several major concerns must be addressed to strengthen the manuscript:

Major Concerns:

1. The authors state, “Following a five-day culture period, the spheroids were predominantly composed of viable cells, with a minor presence of dead cells and an active proliferating edge.” However, no proliferation marker appears to have been employed. The authors should clarify how proliferation was assessed and provide supporting data accordingly.
2. In Section 3.1, the manuscript refers to the formation of “distinct capillary-like structures.” The structures shown, however, do not convincingly resemble classical capillary-like networks as described in the literature. Rather, they appear to represent cellular aggregation. The authors are encouraged to consider plating the cultures on a matrix such as Matrigel to better demonstrate capillary-like organization.
3. The cell lines used in Figure 2a are not identified. This information is essential and should be clearly stated in the figure legend or main text.
4. The manuscript would benefit from a discussion of how the observed aggregation of different cell types within the heterotypic spheroids reflects the in vivo tumour microenvironment. The physiological relevance of this model should be more thoroughly addressed.
5. The methodology for analyzing the secretory profile of the heterotypic spheroids is unclear. Specifically:

1. Were the cells cultured in suspension (liquid overlay)?
2. Were they transferred to a matrix?
3. Were they plated on plastic?

The authors should explicitly describe the conditions under which the secretome was collected. Furthermore, if conditions (a) or (c) were used, the authors should justify their relevance to in vivo physiology. If condition (b) applies, the specific matrix used should be detailed.

1. Regarding Figure 4b, it is unclear whether lysates were prepared from intact spheroids or if cell types were separated prior to lysis. The observed increase in vimentin or fibronectin may simply reflect the inclusion of fibroblasts. A more appropriate control would be a mixed-cell population (in the same ratio used to generate the spheroids) cultured without spheroid formation. This would help determine whether the observed changes are due to heterotypic interactions rather than the mere presence of fibroblasts.

< !--StartFragment -->

Minor Comments:

• Line 144: The word “dinamics” appears to be a typographical error and should be corrected to “dynamics.”
• Similar spelling and typographical errors are present throughout the manuscript. A thorough language and grammar review is strongly recommended to improve clarity and readability.
• Lines 363, 376, 421, 496, 592, 593, 757, and 758: These lines contain non-standard characters (“« »”) that appear to be formatting artifacts. The authors should review and correct these errors.

 

Author Response

Reviewer 2.

We thank the reviewer for the positive feedback on our manuscript.

Reviewer 2.

We thank the reviewer for the positive feedback on our manuscript.

 

Point 1. The authors state, “Following a five-day culture period, the spheroids were predominantly composed of viable cells, with a minor presence of dead cells and an active proliferating edge.” However, no proliferation marker appears to have been employed. The authors should clarify how proliferation was assessed and provide supporting data accordingly.

Response 1: The conclusion regarding the presence of a proliferation zone was based on an assessment using FDA/PI/Hoechst 33342 staining, which revealed a phenotype indicating actively growing cell aggregates with a dense nucleus and a peripheral halo of cells demonstrating reduced intercellular adhesion.  We repeated the experiment to improve image quality. See Figure 1d. This technique is employed for indirect assessment of cell proliferation, in which the spheroid cell zones are differentiated based on differential dye penetration. This process is associated with the presence of nutrient and oxygen gradients, which in turn are linked to proliferative activity. It is acknowledged that the incorporation of a standard proliferation marker, such as Ki-67, in forthcoming experiments would be a valuable addition. The present study placed emphasis on functional tests (i.e. the reattachment test) that serve to indirectly confirm the proliferative potential of cells post-spheroid dissociation. See Figure S5. In the current version, the phrase referring to cell proliferation has been removed from the sentence. See line 323-324.

 

Point 2. In Section 3.1, the manuscript refers to the formation of “distinct capillary-like structures.” The structures shown, however, do not convincingly resemble classical capillary-like networks as described in the literature. Rather, they appear to represent cellular aggregation. The authors are encouraged to consider plating the cultures on a matrix such as Matrigel to better demonstrate capillary-like organization.

Response 2: The creation of 3D cultures comprising multiple cell types is achieved through the utilisation of the TIME endothelial cell line, which has the capacity to generate capillary-like structures within Matrigel. This is one of its key functional characteristics. However, Matrigel is not suitable for the purpose of obtaining spheroids. See Figure 5. Conversely, due to the considerably reduced size of spheroids relative to conventional 2D capillary-like structures, it is physically unfeasible to establish complete vascular networks within the confined space of spheroids. However, it was observed that the distribution of endothelial cells during the process of self-organisation and the life cycle of a spheroid exhibited certain peculiarities. The resulting disordered clusters of endothelial cells, which lack a clear hierarchy, morphologically correspond to the early, immature stages of tumour angiogenesis. In this regard, the clusters of endothelial cells concentrated at one pole of the spheroid were interpreted as a potential initial stage in the formation of capillary-like structures. We also reanalysed our confocal data and created a new figure. See Figure 2.

 

Point 3. The cell lines used in Figure 2a are not identified. This information is essential and should be clearly stated in the figure legend or main text.

Response 3: It has been corrected.

 

Point 4. The manuscript would benefit from a discussion of how the observed aggregation of different cell types within the heterotypic spheroids reflects the in vivo tumour microenvironment. The physiological relevance of this model should be more thoroughly addressed.

Response 4: We would like to express our gratitude to the reviewer for this valuable and constructive feedback. A new paragraph has been added to the revised version of the "Discussion" section that directly addresses this question. See line 691-708. In this study, we emphasise that the spatial segregation observed in our model is of significance. This is characterised by the formation of a dense tumour core surrounded by stromal cells and endothelium in fibroblast strands, which manifest as pathologically branched vessels. This is a key histological feature of many carcinomas in vivo, known as "tumor-stromal separation." It is also observed that the differentiated influence of the microenvironment on aggressive (TNBC, HER2+) and less aggressive (ER+) tumour cell subtypes in the model accurately reproduces known clinical patterns of tumour behaviour, thus confirming the physiological relevance of the system for studying specific intercellular interactions.

 

Point 5. The methodology for analyzing the secretory profile of the heterotypic spheroids is unclear. Specifically:

1. Were the cells cultured in suspension (liquid overlay)?
2. Were they transferred to a matrix?
3. Were they plated on plastic?

The authors should explicitly describe the conditions under which the secretome was collected. Furthermore, if conditions (a) or (c) were used, the authors should justify their relevance to in vivo physiology. If condition (b) applies, the specific matrix used should be detailed.

Response 5: A review of the methodology for analysing the secretory profile was conducted, resulting in its subsequent revision. See line 286-288. The 3D models of breast cancer were created using the liquid layering method in 24-well Nunclon™ Sphera™ plates with a U-shaped bottom at a rate of 200,000 cells/well. Following a five-day cultivation period, the spheroids were subjected to a centrifugation process, accompanied by the culture medium from each well (1000g, 15 min, 4°C). The supernatant was then utilized for the assessment of cytokine and chemokine levels through the implementation of the xMAP technique, in conjunction with the Bio-Plex Pro Human Cytokine Screening 48-Plex Panel (#12007283, Bio-Rad Laboratories, Hercules, USA) and QuattroPlex Lab (Dia-M, Moscow, Russia). The study was conducted in triplicates. In the present study, the spheroids were not transferred to a matrix or adhesive plastic for the study of secreted cytokines/chemokines.

 

Point 6. Regarding Figure 4b, it is unclear whether lysates were prepared from intact spheroids or if cell types were separated prior to lysis. The observed increase in vimentin or fibronectin may simply reflect the inclusion of fibroblasts. A more appropriate control would be a mixed-cell population (in the same ratio used to generate the spheroids) cultured without spheroid formation. This would help determine whether the observed changes are due to heterotypic interactions rather than the mere presence of fibroblasts.

 

Response 6: A review of the methodology for analysing the WB was conducted, resulting in its subsequent revision. See line 255-257. Following a five-day cultivation period of the spheroids, the samples from each well were transferred to centrifuge tubes. Subsequently, the samples were subjected to a centrifugal process at a speed of 1000g for a duration of 15 minutes at a temperature of 4°C. The sediment of spheroids was subjected to lysis in RIPA buffer. It is acknowledged that the expression of diverse proteins exhibits variation between two-dimensional (2D) and three-dimensional (3D) cultivation methods. To illustrate this point, we may consider the findings of Hwa-Jeong Park and David M. Helfman, who demonstrated that levels of both mRNA and FN protein increased in MDA-MB-231 cells when cultured in a 3D environment. Evidence has emerged from studies utilising 3D suspension culture and fibroblast cells, which are mesenchymal in nature, demonstrating an augmentation in FN levels. Furthermore, TNBC cells do not express FN under normal 2D conditions, but do so when cultured in a 3D environment https://doi.org/10.1038/s41598-019-56276-3. As a control, we utilise mono-cultivation of cells in 3D, as opposed to monolayer or cell suspension methods. Indeed, the results of this study demonstrate that fibroblasts contribute significantly to the modification of extracellular matrix proteins in the model. However, as the cell model becomes more complex, there is an increase in the level of fibronectin and vimentin synthesis, which is most likely due to heterotypic interactions between cells in the 3D model. We detected two principal fibronectin monomeric isoforms at ~250 kDa and ~272 kDa (Figure 3), reflecting differential alternative splicing (https://www.abcam.com/en-us/technical-resources/target-tips/fibronectinfn1). The 250 kDa form, devoid of the EDA and EDB extra domains. The predominant contribution to this process is most likely derived from fibroblast populations (Figure S4). This larger form as 272 kDa is one of numerous isoforms produced by the process of alternative splicing of a single gene. The discrepancy in size is typically attributable to the incorporation of specific additional domains (EDA and EDB). As demonstrated in Figure 3b, the highest level of large fibronectin was observed in the sample obtained from CAFs spheroids (at twice the level of the sample from normal fibroblasts). The level of fibronectin protein, such as 272 kDa and 250 kDa, in homotypic and heterotypic spheroids (tumour-endothelial cells) from MCF7 cells was found to be low, and low levels were also identified in spheroids from MDA-MB-231 and SK-BR-3.

Point 7. Line 144: The word “dinamics” appears to be a typographical error and should be corrected to “dynamics.”

Response 7: It has been corrected.

 

Point 8. Similar spelling and typographical errors are present throughout the manuscript. A thorough language and grammar review is strongly recommended to improve clarity and readability.

Response 8: Thank you for your kind consideration of our work. In the current version of the manuscript, we corrected minor spelling and typo errors. Current version of the manuscript was proofed by English speaker.

 

Point 9. Lines 363, 376, 421, 496, 592, 593, 757, and 758: These lines contain non-standard characters (“« »”) that appear to be formatting artifacts. The authors should review and correct these errors.

Response 9: It has been corrected.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

After thoroughly reviewing the authors’ response and the revisions implemented in the manuscript, I am satisfied that the work now meets the standards required for publication. I extend my best wishes to the authors in the continuation of their research.

 

 

Author Response

Point 1. After thoroughly reviewing the authors’ response and the revisions implemented in the manuscript, I am satisfied that the work now meets the standards required for publication. I extend my best wishes to the authors in the continuation of their research.

 Response 1: We would like to thank you sincerely for confirming that our manuscript has been accepted for publication. We are also grateful to the reviewer for their diligent review, as well as for their encouraging final remarks and best wishes for our future research. Their guidance has been invaluable.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors efforts to address the concerns raised initially are appreciated. However, their responses to points 5 and 6 remain unconvincing.

Point 5: The physiological relevance of spheroids cultured at a liquid interface is still unclear. The explanation provided does not adequately resolve this issue.

Point 6: Although the authors cite relevant literature, their experimental design does not convincingly demonstrate that the increased expression of fibronectin and other markers arises from the 3D architecture itself rather than from the inherent expression profile of fibroblasts. This concern remains insufficiently addressed.

Author Response

Point 1. The authors efforts to address the concerns raised initially are appreciated. However, their responses to points 5 and 6 remain unconvincing.

Response 1: Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the resubmitted files.

Point 2. The physiological relevance of spheroids cultured at a liquid interface is still unclear. The explanation provided does not adequately resolve this issue.

Response 2: We would like to express our gratitude to the reviewer for highlighting this critical issue concerning the physiological relevance of our spheroid model. It is acknowledged that demonstrating translational value is of the utmost importance for any in vitro system. Up to now, different techniques (scaffold-based and -free) have been used for spheroid formation, being the Liquid Overlay Technique (LOT) one of the most explored methodologies, due to its low cost and easy handling. Additionally, during the last few decades, this technique has been widely investigated in order to enhance its potential for being applied in high-throughput analysis 10.3389/fbioe.2023.1260049 10.3390/pharmaceutics15030806 10.1038/s41598-025-26985-z 10.1002/biot.201700417. The secretome can be defined as both soluble and insoluble factors that are released or secreted into the extracellular environment. These include chemokines, cytokines, growth factors, coagulation factors, hormones, enzymes, glycoproteins, and nucleic acids 10.3390/biomedicines12081842 10.1016/j.jddst.2024.106033. The most basic and extensively researched secretome type is the cancer cell-derived conditioned medium (CM) of cancer cells grown in 2D or 3D culture 10.1002/pmic.201000530 10.1007/s13402-018-00418-8. Typically, culturing the cancer cells in serum-starved media for experiment. After the medium is collected and centrifuged to remove apoptotic bodies, concentrated, and further subjected to secretome identification. In this study, a heterotypic spheroid model was developed in a low-adhesion 96-well format without serum. The standard spheroid formation protocol, as outlined in the 'Materials and Methods' section of the manuscript, was employed, and subsequently, the only culture medium (without spheroids) was collected for analysis by xMAP. We added a graphic abstract with methodology in the current version manuscript. This enabled controlled, high-throughput for future drug screening within a more physiologically relevant context. The incorporation of endothelial and stromal cells enables the model to replicate essential tumour-stroma interactions. The incorporation of endothelial and stromal cells into the model enhances the analysis of aggressive tumour characteristics, including alterations in cytokine and chemokine secretion, augmented invasive and migratory activity, preservation of viable cells, and the formation of endothelial and stromal components. To elaborate, it can be observed that the incorporation of microenvironmental cells into the modelling paradigm results in an augmentation of both the invasive and migratory capabilities of the cells (Figure 4-5). The integration of 2.5% Matrigel during the process of spheroid formation resulted in enhanced compaction. However, this did not induce the formation of pseudovessels; instead, it led to an augmentation in the disability of endothelial cells (Figure S2). It is acknowledged that the model is a simplification and does not capture the full complexity of the tumour microenvironment (e.g. vascular fluid flow, adaptive immune components). The primary value of this platform lies in its capacity to provide a standardised, reproducible system that serves as a crucial bridge between two-dimensional (2D) culture models and in vivo models. This platform is specifically designed for the study of stromal-mediated drug resistance and invasion. The current manuscript has been augmented by a comparative analysis between 2D and 3D cell cultures. Please see Figure S4 for visual representation. Comparative analysis of the PD-L1-positive population in 2D and 3D cell models demonstrated that the level of positive cells did not correlate between 2D and 3D cultures and that this population increased in heterotypic spheroids, whereas in monolayer fibroblasts it was only weakly positive. The highest level was observed in the MDA-MB-231 cell line, but in a heterotypic 3D culture from this tumour line, this level decreased. Concurrently, despite a parallel tendency in CD24 and CD44 content between 2D and 3D models, it is evident that the proportion of such cells escalates in 3D cultivation. These alterations are not contingent on the content of CD24 and CD44 in stromal cell cultures. Additionally, the manuscript has been expanded to include additional information on preclinical cancer models, which has been integrated into the introduction and discussion sections (See line 81-83, 88-117, 147-150, 193-195, 223-224, 378-381, 398, 639-644, 646-650, 736-752, 927).

Point 3. Although the authors cite relevant literature, their experimental design does not convincingly demonstrate that the increased expression of fibronectin and other markers arises from the 3D architecture itself rather than from the inherent expression profile of fibroblasts. This concern remains insufficiently addressed.

Response 3: We would like to express our gratitude for this invaluable and perspicacious contribution. As you have correctly identified, a significant methodological challenge arises in the field of 3D culture research, namely the necessity to distinctly differentiate between effects that are unique to the three-dimensional architectural context and those that are intrinsic to the inherent properties of cells. A comparative analysis was conducted between two dimensions (2D) and three dimensions (3D) of cultures. As illustrated in Figure S4a, the analysis of tumour cells revealed low fibronectin expression in 2D, with the MDA-MB-231 and SK-BR-3 cell lines demonstrating negative expression. However, when monocultured in 3D, it can be seen that two forms of fibronectin protein are synthesised, with the highest levels observed among tumour cells in MDA-MB-231 and SK-BR-3 cultures. This finding is consistent with the existing literature and is specifically related to the culture conditions. We added it in manuscript (See line 454-459, 474-480). It is also observed that stromal cells exhibit a positive reaction to fibronectin protein, both in 2D and 3D culture conditions.  Furthermore, it is observed that the 250-kDa protein manifests exclusively in the presence of stromal cells within the heterotypic model. However, this form of the protein is deemed to be of negligible significance in the monotypic 3D from fibroblasts (Figure 3). This observation suggests that heterotypic interactions within the model may be a more pertinent factor in determining the protein's behaviour. Concurrently, it is evident that the 272-kDa protein is exclusively observed in fibroblasts. This  suggests that fibroblasts likely constitute a substantial proportion of fibronectin synthesis in the model. Moreover, the general trend of fibronectin protein changes in the model as it becomes more complex shows an increase in fibronectin protein levels, and fibroblasts, including СAFs, which were previously described as matrix-modelling, play a key role in these changes 10.3390/ijms26167789. Consequently, the observed phenotype cannot be explained solely by the 'intrinsic profile' of fibroblasts, because: The following two factors must be considered in order to understand the process under investigation: (a) The profile of the tumour cells is subject to radical change. (b) The appearance of a qualitatively new isoform is observed, for which intercellular cross-flow is a prerequisite. The combination of three-dimensional architecture and heterotypic signals functions as an inducer, with fibroblasts playing a key role in this process.