Spheroid culture remodels mitosis and the proteome in tumor cells

  1. Ana Petelinec
  2. Claudia Cavarischia-Rega
  3. Adrian Perhat
  4. Boris Maček
  5. Iva M. Tolić

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Mitosis depends on precise spindle assembly and positioning, processes influenced by cell shape, size, and microenvironment. Most mechanistic insights into mitosis come from two-dimensional (2D) monolayer cultures, which lack the spatial constraints and extracellular matrix found in tissues, leaving the influence of the tissue environment on mitosis poorly understood. Here, we combine high-resolution imaging and quantitative proteomics to compare mitosis in three-dimensional (3D) multicellular spheroids, generated by magnetic levitation, with that in 2D monolayers. Using a non-transformed cell line and three cancer cell lines from breast, bone, and ovary, we show that 3D culture reshapes mitotic cells and their spindles. Tumor spheroids exhibited a prometaphase delay together with minor chromosome alignment defects, yet chromosome segregation remained largely accurate. Cells in spheroids were rounder, and their spindles were smaller, with increased multipolarity and defects in orientation and position, which varied by cell line. Proteomic profiling revealed broad downregulation of mitotic regulators in spheroids, including kinesins (KIF11, KIF4A), spindle checkpoint proteins, and APC/C components, accompanied by enrichment of metabolic and mitochondrial pathways. Together, our results reveal both shared and cell line-specific modes of mitotic restructuring and establish a framework that connects proteome state to mitotic architecture in 3D environments.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    The authors do not wish to provide a response at this time.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary:

    In this manuscript, Petelinec et al compared mitotic characteristics of cells cultured in 2D versus 3D using RPE1 p53KD, MDA-MB-231 (p53 R380K), U2OS (p53 WT), and OVSAHO (p53 R342*) cells. Magnetic particles were added to cells and those grown in 3D were transferred to a cell-repellent plate with a magnetic lid. The fraction of cells in mitotic stages after anaphase onset was reduced in cancer cell lines grown in 3D. In 1 of 3 cancer cell lines, this correlated with a ~20% increase in uncongressed chromosomes, though uncongressed chromosomes were also elevated in RPE1 p53KD cells, which did not exhibit a significant difference in mitotic stages in 2D versus 3D culture. Cell height increased in all 4 cell lines grown in 3D, and RPE1 p53KD and U2OS cells were more likely to exhibit a round morphology. Spindles in 3D culture were smaller in all four cell lines. Proteomic analysis showed a decrease in expression of mitotic proteins in 3D culture. Overall, the authors conclude that 3D culture induces shared and cell line-specific differences and that they have established a framework that connects proteome state to mitotic architecture.

    Major comments:

    1. The figure legend for S1A indicates that MDA-MB-231 cells grown in 3D expressed H2B-GFP, while the cells grown in 3D did not. Was that the case for all experiments? If so, comparison of these two different populations of cells could account for some of the differences observed throughout.
    2. The number of biological replicates as well as the number of cells analyzed per biological replicate should be clearly stated in the figure legends. As presented, much of the data appear to come from a single biological replicate, which would be insufficiently rigorous.
    3. The downregulation of mitotic proteins that are cell cycle regulated (Aurora A, cyclin A2, cyclin B1, Bub1B, CDC20, KIF11; doi: 10.1091/mbc.E13-05-0264) strongly suggests that, rather than the proposed "global rewiring of cell-cycle regulation in 3D", the proliferation rate is lower in 3D. Ki67 expression is markedly lower in MDA-MB-231 and RPE1 p53KD cells grown in 3D (Fig S4J). Quantitation of mitotic index is only provided for MDA-MB-231 and OVSAHO cells, and the values for the different cell lines are combined (Fig S1A). This is an unusual way to present the data, and obscures any differences that may be occurring. Together with the reported p value of 0.052, this does not provide strong evidence that proliferation rate is not reduced in 3D culture. Reporting the mitotic index for each cell line in 2D and 3D is a rapid and straightforward way to address this issue.
    4. The manuscript concludes that 3D culture increases multipolar spindles. However, this only appears to be true in MDA-MB-231 cells. In the 3 other cell types examined, the incidence of multipolarity appears to be <5%.
    5. Similarly, spindles in 3D culture are reported to be prone to "misalignment", but there are no data reporting the incidence of misaligned chromosomes in this section. Perhaps this is meant to indicate that they are misoriented with respect to the long axis of the cell, but changes in this orientation were only observed for 2 of the 4 cell lines.

    Minor comments:

    1. Though the model is described as "spheroids", the example in Fig 1B is not spherical, nor are the measurements described. Based on this, the term "spheroid" seems like a misnomer and another term (perhaps "organoid") would be a better descriptor.
    2. It would be helpful to include measurements for spindle height (in addition to length and width) in Fig 3.
    3. Based on the p values, it seems like the comparisons in Fig 1D, F, 2B,C,E, 3B,C,F,G,I,K were done by comparing the total number of cells rather than comparing the average of each biological replicate, which would be more rigorous.
    4. It is stated that SAC proteins were generally downregulated in 3D culture, and data for BUB1B are shown, but data for MAD2 should also be shown.
    5. The images in Fig 1C are too small to readily show that cells are elongated in 2D and round in 3D. Insets/higher magnification views are warranted.
    6. In Fig 2A, it would be helpful to indicate what stain was used to demarcate the cell boundaries to measure length and width in the figure legend.
    7. In Fig 2F, it isn't possible to distinguish the dots from the 8 different groups. It would be helpful to have 2 different graphs, one showing the data for cells grown in 2D and the other for cells grown in 3D.
    8. In Fig 3G, were the "round" and "elongated" categories based on measurements in Fig 3B-C? Or was this a qualitative assessment? It would be helpful to clarify this in the figure legend.

    Significance

    This article will be of interest to a specialized audience. Its strengths are that it provides 1) measurements of changes that occur in cell and spindle size in four human cell types by varying growth conditions in 2D versus 3D and 2) matched proteomics analysis. Its limitations are that 1) it is descriptive and 2) the physiological relevance of growth in spheroids due to magnetic levitation is unclear. While it seems reasonable that 3D growth is more physiologic than growth on 2D, and there are certainly differences between 2D and 3D culture, it is not clear that the changes that occur in 3D magnetic spheroids hold true for spheroids grown using other methodologies. Importantly, while it is implied that the changes observed in 3D growth are more representative of what occurs in the body, evidence for that is lacking. Directly providing these comparisons to other 3D systems or to human tissues would be both challenging and time consuming and is not considered necessary for publication of this work. However, a thorough and well-cited discussion of previous studies with such quantitation and clear acknowledgement of the extent to which the similarities between 3D culture and in vivo tissue environments remain unknown would provide substantial benefit.

    The proposed framework connecting proteome state to mitotic architecture would be an additional strength of the manuscript, but the link is underdeveloped in the current version of the manuscript. It would be helpful to describe multiple examples in which differing protein expression in the various cell lines correlated with the differential phenotypes observed.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary:

    Petelinec et al. have documented thoroughly mitotic progression characteristics - mitotic timing, spindle assembly and chromosome segregation - in tumor cell lines from different organ origin. Taking advantage of a magnetic levitation approach to establish 3D cultures, they compared the same population of cells in this 3D setting with conventional 2D monolayers. This description is coupled with a proteomic profiling of mitotic cells that highlights differences in the level of key proteins involved in the regulation of the mitotic checkpoint and regulatory proteins essential for spindle assembly and mitotic timing.

    Major comments:

    The manuscript is well documented, explained and illustrated. Figures are self-explanatory and convincing.

    1. Nonetheless, the manuscript in its current state is clearly lacking validation of some hits identified after the comparative proteomic profiling to demonstrate that the differences observed between the 3D and 2D settings can readily be explained by these cell intrinsic factors. If the authors correlate in figure 5 spindle morphometrics and proteomics, this is clearly not sufficient to prove any causal relationship. Along this line, the authors report a global downregulation of mitotic proteins from 2D to 3D settings (Figure 5). Nonetheless, they also report a mitotic index of 2.97% in 2D versus 1.62% in 3D, which is not significant. If mitotic proteins are readily downregulated, the index should be significantly different. This justifies the necessity to further validate functionally the differences observed regarding the mitotic protein level between the two settings.
    2. The magnetic levitation approach is efficient to enable the organization of cells in multilayers and the establishment of 3D cell-cell contacts. Nonetheless, the flat appearance of the "spheroid" might reflect some stretching forces applied to the cells. Application of such forces might, on top of 3D cell-cell contacts impact mitotic progression and spindle organization. To address this point, comparison of mitotic characteristics of at least one cell line (MDA cells for example) cultured under magnetic levitation and in 3D round spheroid shape (which can be enabled by culturing the cells in suspension on a repulsive culture substrate) should be performed by the authors.
    3. The authors report minor chromosome misalignments, probably due to prometaphase delay, based on immunofluorescent approaches using fixed samples (Figure 1). It is always better to perform time-lapse experiments to confirm deviations in mitotic timing, time spent in the different phases of mitosis and to evaluate final chromosome alignment before anaphase onset.

    Minor comments:

    1. Multipolar spindles appear more frequent in 3D settings (Figure 3). Can the authors relate this increase to polyploidization after more frequent cytokinesis failures in the 3D setting for example? Or to defects in spindle assembly by spindle pole splitting for example? They could perform centrosome protein staining to address this question.
    2. Do the authors have any explanation regarding the increased frequency of off-centered spindles in the 3D setting? They propose in the discussion a link with NuMA. Can the authors verify this point by immunofluorescent staining?

    Significance

    This is an unprecedented descriptive study of the impact of 3D cell-cell contacts on mitotic progression and spindle assembly in tumor cell lines in relation with proteomic profiling. In its state, the limitation of the study is the lack of validation of differences between 3D and 2D settings in protein level and impact of these differences on mitotic entry, progression or spindle formation. These findings will further fuel the concept that studying cancer cells in 3D is a pre-requisite (at least for some cancer cells) to study and/or target mitotic processes. This study will be of interest for cell biologists and especially mechanobiologists, with a particular interest in cancer biology. I am expert in cell biology, especially in the regulation of cell cycle progression.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    The authors investigate the impact of 3D culture systems compared to traditional 2D models on mitosis, with a particular focus on spindle assembly. To address this, they combine live imaging and mass spectrometry to analyze M-phase progression in 2D monolayers versus 3D spheroid cultures.

    While the figures are visually compelling, I found it difficult, by the end of the manuscript, to distill the main finding into a single clear statement-an issue that raises concerns about the overall focus of the study. In its current form, the central message and the novelty of the work remain unclear.

    I also have a few technical comments:

    1/ The choice of the 3D model, flat Spheroids generated using magnetic cell levitation (Souza et al., 2010), is somewhat unsatisfactory. As stated in the manuscript, 2 to 4 cell layers encompassing 20 to 50 m in Z, does not constitute a true 3D model. Did the authors observe differences in behavior depending on the thickness in Z (20 versus 50 m-wide regions)?

    2/Related to this, one of the conclusions from the work is that: "while 3D culture reshapes interphase cells, mitotic entry and overall cell cycle progression appear largely similar." But maybe it is the case because the 3D model is not really 3D, but closer to a 2D-one.

    3/ I have a difficulty to understand the difference in the quantification of phenotypes between 3F and 3G. What exactly is the difference between multipolar or irregular spindle?

    4/ The rationale behind the mass spectrometry approach is not entirely clear, or at least it is not sufficiently explained. It is unclear whether the authors performed mass spectrometry on asynchronous cell populations in both 2D and 3D conditions. Moreover, the proportion of mitotic cells differs between these conditions (approximately 3% in 2D versus 1.6% in 3D) and remains low overall. As a result, the mass spectrometry samples are likely to be predominantly composed of interphase cells. This raises concerns about the ability to draw meaningful conclusions regarding differences in the mitotic proteome and to reliably link these differences to observed mitotic phenotypes.

    5/Some of the mass spectrometry conclusions put forward (such as levels of KIF11 or NUMA) should at least be verified using immunofluorescence of mitotic in the different culture conditions.

    6/ The scheme presented in Fig 4B is very difficult to read and should be simplified

    Significance

    It is somewhat surprising that the authors neither cite nor discuss prior work on spindle scaling derived from embryonic models. Indeed, numerous studies using embryonic systems-which represent physiologically relevant 3D contexts-have extensively characterized the scaling relationship between mitotic spindle size and blastomere (i.e., cell) size (e.g., Wühr et al., Curr Biol 2008; Greenan et al., Curr Biol 2010; Courtois et al., J Cell Biol 2012; Wilbur & Heald, eLife 2013). Incorporating and discussing this body of work might help better contextualize the present findings and clarify how they relate to established scaling principles.

  5. Version published to 10.64898/2026.03.12.711371 on bioRxiv