Three Dimensional Dynamics of Epithelial Monolayers

Article activity feed

1. 

   eLife

   Nov 13, 2026

   eLife Assessment

   This article by Lästad and colleagues proposes to reconsider some important aspects of single-cell volume and dry mass homeostasis in a model system of epithelial monolayer, based on a novel method relying on quantitative phase imaging to measure these parameters. The article proposes an in-depth analysis of a large dataset uniquely based on this method. This is the main current weakness of the manuscript, since the claims that differ from previous findings are not tested using other methods to exclude potential artifacts.

   Read the original source
2. 

   eLife

   Nov 13, 2026

   Reviewer #1 (Public review):

   Summary:

   The study uses Quantitative Phase Imaging to measure and infer cell shapes, areas, volumes, and masses during the evolution of a confluent epithelium. These are important physical parameters of a monolayer and help to better understand the dynamical evolution of an epithelium and collective cell behaviors.

   Strengths:

   Quantitative Phase Imaging allows simultaneous, live-cell measurements of mass density and height that are not easily achievable by other means.

   Weaknesses:

   The biological or biophysical conclusions drawn are not fully substantiated, as QPI is the sole experimental technique used, and the analysis relies heavily on indirect 2.5D assumptions that the authors themselves question. This makes the novelty questionable, especially since a technical paper establishing this QPI pipeline for …

   Reviewer #1 (Public review):

   Summary:

   The study uses Quantitative Phase Imaging to measure and infer cell shapes, areas, volumes, and masses during the evolution of a confluent epithelium. These are important physical parameters of a monolayer and help to better understand the dynamical evolution of an epithelium and collective cell behaviors.

   Strengths:

   Quantitative Phase Imaging allows simultaneous, live-cell measurements of mass density and height that are not easily achievable by other means.

   Weaknesses:

   The biological or biophysical conclusions drawn are not fully substantiated, as QPI is the sole experimental technique used, and the analysis relies heavily on indirect 2.5D assumptions that the authors themselves question. This makes the novelty questionable, especially since a technical paper establishing this QPI pipeline for epithelia has been published elsewhere by the same authors.

   To study the dynamics of confluent epithelium, the authors applied 2D and 3D quantitative phase imaging (QPI) to measure local dry mass densities and heights of MDCK monolayers. These techniques allow simultaneous, live-cell measurements of mass density and height that are not easily achievable by other means, enabling a direct test of dry mass conservation that enriches the coarse-grained description of epithelial tissues. The authors also use the mass/height information and apical segmentation to estimate cell areas, volumes, and dry masses, leading to deductions about non-prismatic cell shapes, mass shedding, and volume reduction.

   Overall, the study demonstrates an interesting technical application of QPI. However, while I find the methodology promising, I am not fully convinced by the biological or biophysical conclusions drawn, as QPI is the sole experimental technique used, and the analysis relies heavily on indirect 2.5D assumptions that the authors themselves question. Furthermore, given that the authors have already published a technical paper establishing this QPI pipeline for epithelia (Reference 22), the present manuscript cannot claim novelty on methodological grounds alone. Its contribution therefore hinges almost entirely on the biological insights it provides. However, as I outline below, those biological insights are either expected from previous literature (non-prismatic shapes in low-density MDCK), inferred indirectly from 2.5D projections (frustum shapes), or remain at best correlative without experimental validation (mass shedding). Thus, in its current form, I struggle to identify a clear advance. Below, I outline specific technical concerns and suggest ways to strengthen the study.

   (1) Indirect inference of non-prismatic cell shapes from 2.5D projections.

   The authors segment the projected (apical) area of the cells and assume prismatic, vertical-sided geometries to calculate cell volumes, and compare them to frustum-based cell models. They conclude that cells are non-prismatic if volumes are constant. While I appreciate the modeling effort, the evidence for non-prismatic shapes remains indirect, as it relies on comparing 2.5D projected correlations rather than a direct measurement of the 3D geometry. Since the authors already have full 3D refractive index tomograms (as shown in Figure 1A), a much more direct and definitive approach would be to perform true 3D volumetric segmentation along the z-axis to directly reconstruct actual cell shapes. Additionally, structures that lead to non-prismatic shapes (e.g., cryptic lamellipodia) are well-documented for low-density MDCK cells, so I do not find this conclusion particularly illuminating for epithelium biology. To improve on this analysis, tracking the evolution of directly segmented 3D shapes as density increases beyond the range they presented would be highly instructive, as one would expect the lateral sides to become progressively more vertical under compaction.

   (2) Apparent conflict between ensemble and single-cell dry mass concentration trends.

   The ensemble-averaged dry mass concentration is shown to increase with cell density (Figure 1D). However, the single-cell tracks in Figure 3B (which span approximately 10 hours of evolution) show dry mass density to be roughly constant. Given that the monolayer density is likely changing over these 10 hours, this seems like a conflict between ensemble and single-cell data. Can the authors clarify?

   (3) Breakdown of continuum mass conservation at the cellular scale.

   The authors showed a breakdown of the 2D continuity equation at cellular scales, which is not surprising. To move beyond this, I would suggest directly testing mass conservation at the cell scale by following mass evolution of each cell using integrated mass from 3D tomograms. This would provide a much stronger conclusion than the current 2D coarse-graining analysis, which is relevant for monolayer-scale, but not at the cell-scale.

   (4) The "mass shedding" hypothesis requires experimental validation.

   The claim that cells periodically shed 10-15% of their dry mass is an important biological statement, yet it remains unclear due to the 2.5D analysis and is, at best, correlative. Given that epithelial cells are known to secrete basement membrane components, this hypothesis is testable. I would suggest the authors perform appropriate pharmacological perturbations of secretion pathways, if not genetic perturbations, and repeat the inference from 3D tomograms to see if the apparent volume/mass loss is abolished.

   (5) Sensitivity of mass measurements to the specific refractive index increment.

   The conversion from refractive index contrast to dry mass concentration relies on the specific refractive index increment. While the authors use a standard average value of α ≈ 0.18−0.2mL/g, it is worth noting that this parameter varies between different biomolecular classes. The authors should discuss the potential sensitivity of their absolute mass and concentration values to compositional shifts (e.g., changes in the lipid-to-protein ratio) that might occur as cell density increases, and provide an estimate of the error bounds this could introduce into their conclusions.

   Read the original source
3. 

   eLife

   Nov 13, 2026

   Reviewer #2 (Public review):

   Summary:

   In this study by Lästad et al., the dry mass and mass density of confluent MDCK epithelia are carefully measured using 2D and 3D QPI microscopy. Extensive imaging experiments were performed. The single-cell-resolved data show that the cell mass density is relatively conserved, even though there is considerable cell height and cell volume variation. There appear to be transport processes happening between neighboring cells within the epithelium.

   Strengths:

   The data appears to be high quality, with good statistics. I think the conclusions are generally interesting and sound, with some questions noted below.

   Weaknesses:

   (1) I think the weakest part might be Figure 3 concerning cell volume and area. The authors refer to their previously published paper on their segmentation and height determination …

   Reviewer #2 (Public review):

   Summary:

   In this study by Lästad et al., the dry mass and mass density of confluent MDCK epithelia are carefully measured using 2D and 3D QPI microscopy. Extensive imaging experiments were performed. The single-cell-resolved data show that the cell mass density is relatively conserved, even though there is considerable cell height and cell volume variation. There appear to be transport processes happening between neighboring cells within the epithelium.

   Strengths:

   The data appears to be high quality, with good statistics. I think the conclusions are generally interesting and sound, with some questions noted below.

   Weaknesses:

   (1) I think the weakest part might be Figure 3 concerning cell volume and area. The authors refer to their previously published paper on their segmentation and height determination methods. It appears that some of these results may depend on the parameters of the imaging method, although I am not an expert. I am wondering to what extent the data depends on these image analysis methods. For example, there are some Gaussian blur parameters used. Do these parameters significantly affect the results? I suspect not, but I would be more convinced if additional validation is shown.

   (2) Along the same vein, I couldn't find how the velocity field in Equation 8 was determined. This seems to be a critical quantity for conclusions in Figure 5.

   (3) Equation 8 assumes that mass is conserved within the epithelium. Is this really the case in the experiment? If some cells are growing, then there should be a source term in Equation 8. Could some of the mass non-conservation come from growth fluctuations?

   (4) The paper is nearly entirely physical measurements. While I am fine with this, it would be nice to perform some biological manipulation to validate the results. For instance, if there is mass transport between cells, then blocking tight junctions might stop this. Or if serum is removed, then cells would be arrested, and mass would definitely (maybe?) be conserved?

   Read the original source
4. 

   eLife

   Nov 13, 2026

   Reviewer #3 (Public review):

   Summary:

   In this article, Lästad and colleagues propose a large dataset of combined 2D and 3D quantitative phase imaging (QPI), from which they extract the volume, dry mass, and height of single cells in a monolayer formed by MDCK epithelial cells at various cell densities. Some of these measures are at odds with previous reports or assumptions, mostly concerning the height of cells in the monolayer. As a consequence, the volume measures also differ, showing, like the height, larger fluctuations than expected. This raises a fundamental question, since the dry mass concentration seems to stay very homogenous in the monolayer, suggesting that dry mass of single cells also fluctuates to a rather large extent, both up and down (on top of the longer timescale growth trend). The possibility that dry mass decreases …

   Reviewer #3 (Public review):

   Summary:

   In this article, Lästad and colleagues propose a large dataset of combined 2D and 3D quantitative phase imaging (QPI), from which they extract the volume, dry mass, and height of single cells in a monolayer formed by MDCK epithelial cells at various cell densities. Some of these measures are at odds with previous reports or assumptions, mostly concerning the height of cells in the monolayer. As a consequence, the volume measures also differ, showing, like the height, larger fluctuations than expected. This raises a fundamental question, since the dry mass concentration seems to stay very homogenous in the monolayer, suggesting that dry mass of single cells also fluctuates to a rather large extent, both up and down (on top of the longer timescale growth trend). The possibility that dry mass decreases significantly over hour periods in a single cell seems novel and unexpected. The authors then propose that a large part of these apparent volume fluctuations and thus dry mass fluctuations (and loss in particular) might actually come from the wrong attribution of signal to individual single cells, in case the cell/cell interface is not as vertical as assumed in the largely 2D segmentation of the cell area. They nicely show that a large part of the volume behavior that they observe could be due to that. They do not consider other possible sources of error in the volume/height measures, but there are likely even more (see weaknesses below), so this could be the main reason for this apparent novel/unexpected result. This does not explain the strong discrepancy between their measures of cell height and other reports. This could be due either again to a segmentation error that has not been sufficiently checked, or to a different culture condition. This point is not really discussed by the authors. In conclusion, this article presents a potentially very powerful and novel approach to access single cell size and growth homeostasis in monolayers, with some surprising findings that, at least for some of them, raise a question about the validity of the approach (at least if no other method is used to correct it, as is the case here), and this is in part nicely acknowledged by the authors.

   Strengths:

   The main strength of the article is in the combination of the 2D and 3D QPI, which allows the extraction of several parameters that single measures would not allow, in addition to also providing cross-validation power for some of the measures - the authors had published some of that in a prior article, which explains well these strengths. The other interest of the article is that it questions some prior findings (or assumptions), mostly on the height of the cells in monolayers. This has always been the less quantitative aspect of studies on monolayer morphogenesis, due to the poor resolution of fluorescence microscopy in the z-axis. Finally, the authors propose a well-balanced interpretation of their data, including a clear exploration of a potential artifactual origin of the volume fluctuations they report, due to wrong assumptions on the segmentation of single cells. The article is also very well written and clear.

   Weaknesses:

   The major weakness of the article is the lack of cross-validation of the most surprising findings with other, more direct methods, which should be rather easy to perform, at least on some specific points (see below), to make sure that it is not all about artifacts of the QPI, in addition to other 3D-related limitations that the authors already explored. A potentially strong weakness is that the authors do not really propose a solution (and maybe there is no solution) to solve the issue of the complex 3D shape of single cells and the potential artifact it creates on volume measures.

   Specific weaknesses:

   (1) According to the authors, the main source of possible error in their volume measures would be caused by cell/cell junctions that would not be vertical. As the authors point out, since the cell/cell junctions in their monolayers seem rather thin in the z-axis, this might not explain all. The authors should consider other sources of error, also related to wrong segmentation, especially from the 3D QPI. This could be cell protrusions or other extensions that are not always segmented the same way from one image to another, or also issues in determining precisely the threshold for the height. To test these possible issues, it seems more appropriate to first validate the method on single isolated cells, which is not done here nor in the previous method article from the same authors. Single cells have many advantages, and it would be much easier to rule out all kinds of measurement errors related to segmentation by testing against other well-established methods. For example, the volume of cells reported by the authors seems too large compared to a number of other reports, which is typical of segmentation errors, especially in height. A way to test for that is to simply image single cells in a non-adherent state to get their diameter. This is not very precise, but it can easily rule out factors of 2 in the volume. FxM would be more precise, but might not be directly compatible with 3D QPI, so it could not be performed on the same cell, leaving some uncertainties (although at least it could be used to check the volume distribution of the MDCK cells when plated as single cells, and compare to the values obtained from 3D QPI). In general, data on single cells should be provided - the authors could perform the same set of analyses, but with much less ambiguity in their interpretation, and thus more easily identify possible sources of errors with their method.

   (2) The authors claim a very good precision on the height (5%, page 6), but this is not validated against a more direct method of imaging. It is hard to believe, as claimed on page 8, that other authors would have totally missed such a large amplitude of height fluctuations using confocal imaging - the authors report heights from 5 to 9 microns, so almost a factor of two, over a range of distances that can be easily resolved in confocal microscopy, even in the z-axis. Before making such a claim, it appears essential that the authors image their monolayers using up-to-date confocal microscopy, or even better, some microscopy with better z-axis resolution (there is a large variety of options). In case they find the same height variations (5 to 9 microns), then it is very likely that there is a difference in the cell culture conditions and that they are just not looking at the same type of monolayers as others. In case they do not see any height variation, they should strongly question their method and maybe try to find a way to address the origin of the error more directly. In case they really doubt the capacity to resolve the z-axis by confocal microscopy, they could also rather easily arrange a setup to look at their monolayer from the side (there are again a variety of options) and check the height this way.

   (3) The height profile provided in Figure 2 makes it hard to believe that this is a regular 'mature' monolayer with junctions, etc. I am not a specialist in these systems, so I might be wrong. But I would recommend that the authors check classical markers for cell/cell junctions, as others do, to at least assess whether they are dealing with the same type of monolayers that others have been investigating in the reports that the authors comment on; if it is not the case, then it might not be surprising that they find very different results.

   (4) The authors make a strong point about a very strict stable dry mass concentration, which then raises a lot of questions. They could also perform some simple tests to check if concentrations are constant in the cell. It is rather simple to mix labelled and not labelled cells to form a monolayer and use cytosolic labelling, either genetically expressed or using a dye (there are many possible ways to get molecules inside cells: membrane-permeable ones, or electropermeation, or using micro-injection). Then, using confocal microscopy, it is easy to assess the concentration while checking that the total quantity of fluorophore is constant, with widefield microscopy (or, if the labelling molecule is well characterized, like a large dextran, it can be assumed). This has been performed in numerous studies. It does not give the volume in absolute numbers, but gives the relative volume changes in time. For claims such as constant concentration but fluctuating volume, it is important to have at least one additional method to test the claim.

   (5) The hypothesis that cells shed some dry mass, discussed in many places in the article, seems contradicted by the fact that dry mass appears constant when the measure is averaged over a large enough area. It seems that the authors can discard this hypothesis fully. Instead, they might leave more open the possibility that cells exchange material, since this does not have to be only through gap junctions, but also, for example, cells uptaking fractions of neighboring cells (instead of shedding material in the extracellular medium), or exchanging material via other junctions between cells, such as membrane tubes. These have been reported in a variety of contexts, either as 'tunneling nanotubes', which have been shown to transport large cargo, or as 'unresolved cytokinetic bridges', with cells that seem to be already out of mitosis but are actually still connected by rather large tubes for hours. It is out of the scope of the article to test all these (although it can be tested rather easily with fluorescent reporters inside a subfraction of cells), but the authors could discuss these possibilities.

   (6) In their discussion p14 (point a.), the authors seem to leave aside the question of timescales. All these size changes they mention have very different timescales. To typically achieve a significant change in value (1 to 10 %), water can pass membranes in seconds, ions are rather pumped at the minute timescale, while dry mass accumulation is rather on the hour timescale. This usually also helps to tell apart the potential mechanisms involved.

   (7) The discussion, to a large extent (as well as the conclusion), is just rephrasing the results without adding much more. It is a place where the authors could propose more radical alternative interpretations of their data and explore other potential limitations of their method (see all the points above).

   Read the original source
5. 

   eLife

   Nov 13, 2026

   Author response:

   We thank the reviewers for constructive and thorough reports. We are encouraged that they find the data of high quality, the method promising, and the interpretation well balanced. We will revise the manuscript thoroughly. Here we respond to the main common concern and outline the planned revisions.

   Independent validation.

   The assessment and all three reviewers ask whether our results could be artefacts of QPI. We would like to clarify that the results rest on two instruments that work in fundamentally different ways. 2D QPI (digital holography or interferometry, Holomonitor) measures the integrated optical path length. Refractive index tomography (3D QPI, Tomocube) reconstructs the 3D refractive index from multiple illumination angles. The two use different light paths, reconstruction algorithms and post-processing. In …

   Author response:

   We thank the reviewers for constructive and thorough reports. We are encouraged that they find the data of high quality, the method promising, and the interpretation well balanced. We will revise the manuscript thoroughly. Here we respond to the main common concern and outline the planned revisions.

   Independent validation.

   The assessment and all three reviewers ask whether our results could be artefacts of QPI. We would like to clarify that the results rest on two instruments that work in fundamentally different ways. 2D QPI (digital holography or interferometry, Holomonitor) measures the integrated optical path length. Refractive index tomography (3D QPI, Tomocube) reconstructs the 3D refractive index from multiple illumination angles. The two use different light paths, reconstruction algorithms and post-processing. In particular, the tomographic cell height comes from the geometric position of the RI contrast along the optical axis. It therefore does not depend on the absolute refractive index of the cell which 2D QPI height does. What the two methods share is the relation between optical phase and dry mass, which has been established for more than seventy years [1,2]. This relation has repeatedly been cross-validated with independent methods. Both QPI modalities have several times been cross-validated against other methods for single cells. For example, dry mass concentration from refractive index tomography agrees with stimulated Raman scattering measurements on the same cells [3]. Optical measurements (fluorescence exclusion volume combined with QPI dry mass) and mechanical measurements (suspended microchannel resonator) independently find the same mitotic swelling at constant dry mass [4,5]. The two modalities agree on the density dependence of monolayer height and on the magnitude of its spatial and temporal fluctuations. In the revision we will present the two modalities separately rather than pooled, so that this agreement can be seen directly.

   We nevertheless agree that comparison with a non-optical-phase method would strengthen the conclusions. We are exploring, with collaborators, volume measurements on MDCK monolayers using fluorescence exclusion adapted to monolayers. This may require some additional time.

   Comparison with previous height measurements.

   Our cell densities (1300–3200 cells/mm²) overlap those of the key earlier studies of MDCK monolayer dynamics: about 1400 cells/mm² in Zehnder et al. [6,7] and 1480–2630 cells/mm² in Angelini et al. [8]. Zehnder et al. [6,7] use the same parental MDCK line as we do, ATCC NBL-2 CCL-34 (T. E. Angelini, personal communication). We therefore do not think the difference from earlier reports is explained by a different cell line or by studying monolayers at a different stage of confluence. Remaining differences in substrate and medium will be discussed in the revision.

   We also note that the confocal literature on MDCK monolayer height is itself divided. Zehnder et al. found a flat layer (7.1 ± 0.7 µm, 4.7 % spatial variation) over a 160 × 160 µm area of about 64 cells [6]. In contrast, Thiagarajan et al. found thickness changes up to twofold during pulsations, also by confocal microscopy [9]. Our measurements cover a field of view of about 570 × 570 µm and agree with the latter. Zehnder et al. also reported single-cell area oscillations with an amplitude of 20 % and a period of 4 h [6], similar to the area fluctuations we observe. In their own confocal tracking of individual cells, some cells conserved volume and others did not (Supporting Material of [6]). What is new in our data is that the dry mass concentration stays nearly constant during these fluctuations. We will expand the discussion of why membrane-label confocal measurements and phase measurements may differ: axial resolution, the criterion used to define the apical surface, and averaging over the cell versus at the cell centre.

   Non-prismatic geometry and apparent mass non-conservation.

   We agree with Reviewers 1 and 3 that the most debatable part of the manuscript is the interpretation of apparent dry mass loss in segmented cells. We always presented mass shedding as one of two non-exclusive interpretations, and we explicitly favoured non-prismatic geometry as the main contribution. In the revision we will strengthen this emphasis. We will also add the alternative mechanisms suggested by Reviewer 3: material exchange between neighbouring cells through tunnelling nanotubes or unresolved cytokinetic bridges, and partial uptake of neighbours. We agree with Reviewer 3 that the recovery of mass conservation after coarse-graining favours redistribution over net loss.

   We note, however, that the existence of non-prismatic cell shapes in MDCK monolayers is not our central point. What matters is that such shapes quantitatively bias widely used measures: segmented cell volume, cell velocity from apical PIV or tracking, and continuum descriptions based on plug flow. The breakdown of the continuity equation below about two cell diameters and 0.6 h is a direct test of this. It implies that velocity fields measured at cellular scales do not in general represent the motion of the cell's centre of mass. We will make this point more explicit.

   We will attempt 3D segmentation of the refractive index tomograms, as Reviewer 1 suggests. We note, though, that lateral membranes have low RI contrast and that the axial resolution of holotomography is limited by the missing-cone problem (Appendix B). Reliable 3D segmentation of confluent monolayers therefore remains a challenge.

   Timescales.

   We thank Reviewer 3 for raising the question of timescales. The volume and height fluctuations we report have periods of hours (≈5 h). That is much slower than water flux (seconds) or ion pumping (minutes), and matches the timescale of dry mass changes. Together with the near-constant dry mass concentration, this supports our conclusion that fluid transport is not the dominant driver of the pulsations. We will restructure the discussion around this argument and reduce the repetition of results that Reviewer 3 noted.

   Further revisions. We will:

   - Add a sensitivity analysis of the results to the image-analysis parameters (Gaussian filtering, segmentation and height thresholds) (Reviewer 2);

   - Expand the description of the velocity-field determination (Reviewer 2);

   - Discuss growth as a source term in the continuity equation (Reviewer 2);

   - Add an error estimate for the dry mass concentration under plausible compositional changes in the refraction increment (Reviewer 1);

   - Clarify how the weak density dependence of the dry mass concentration in the ensemble (Fig. 8G) is consistent with the near-constant concentration in single-cell tracks (Reviewer 1).

   References

   (1) Barer R. Interference microscopy and mass determination. Nature 169, 366–367 (1952). doi:10.1038/169366b0

   (2) Davies HG, Wilkins MHF. Interference microscopy and mass determination. Nature 169, 541 (1952). doi:10.1038/169541a0

   (3) Oh S, Lee C, Yang W, Li A, Mukherjee A, Basan M, et al., Kirschner MW. Protein and lipid mass concentration measurement in tissues by stimulated Raman scattering microscopy. Proc Natl Acad Sci USA 119, e2117938119 (2022). doi:10.1073/pnas.2117938119

   (4) Zlotek-Zlotkiewicz E, Monnier S, Cappello G, Le Berre M, Piel M. Optical volume and mass measurements show that mammalian cells swell during mitosis. J Cell Biol 211, 765–774 (2015). doi:10.1083/jcb.201505056

   (5) Son S, Kang JH, Oh S, Kirschner MW, Mitchison TJ, Manalis S. Resonant microchannel volume and mass measurements show that suspended cells swell during mitosis. J Cell Biol 211, 757–763 (2015). doi:10.1083/jcb.201505058

   (6) Zehnder SM, Suaris M, Bellaire MM, Angelini TE. Cell volume fluctuations in MDCK monolayers. Biophys J 108, 247–250 (2015). doi:10.1016/j.bpj.2014.11.1856

   (7) Zehnder SM, Wiatt MK, Uruena JM, Dunn AC, Sawyer WG, Angelini TE. Multicellular density fluctuations in epithelial monolayers. Phys Rev E 92, 032729 (2015). doi:10.1103/PhysRevE.92.032729

   (8) T. E. Angelini, E. Hannezo, X. Trepat, M. Marquez, J. J. Fredberg, and D. A. Weitz, Glass-like dynamics of collective cell migration, PNAS 108, 4714 (2011).

   (9) Thiagarajan R, Inamdar MM, Riveline D. Interplay between cell height variations and planar pulsations in epithelial monolayers. Eur Phys J E 45, 49 (2022). doi:10.1140/epje/s10189-022-00201-6

   Read the original source
6. 

   Version published to 10.64898/2026.03.10.710903v4 on bioRxiv

   Jun 25, 2026
7. 

   Version published to 10.64898/2026.03.10.710903v3 on bioRxiv

   Jun 19, 2026
8. 

   Version published to 10.64898/2026.03.10.710903v2 on bioRxiv

   Jun 10, 2026
9. Mentioned by @biorxiv-biophys.bsky.social on Bluesky

   Mar 13, 2026

   Three Dimensional Dynamics of Epithelial Monolayers https://www.biorxiv.org/content/10.64898/2026.03.10.710903v1

10. Mentioned by @biorxivpreprint.bsky.social on Bluesky

    Mar 13, 2026

    Three Dimensional Dynamics of Epithelial Monolayers https://www.biorxiv.org/content/10.64898/2026.03.10.710903v1

11. 

    Version published to 10.64898/2026.03.10.710903v1 on bioRxiv

    Mar 13, 2026