From tension to buckling - a mechanical transition underlies the puzzle-shape morphogenesis of histoblasts in the Drosophila epidermis

  1. Annafrancesca Rigato
  2. Huicheng Meng
  3. Claire Chardes
  4. Adam Runions
  5. Faris Abouakil
  6. Richard S Smith
  7. Loïc LeGoff

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

The polygonal shape of cells in proliferating epithelia is a result of the tensile forces of the cytoskeletal cortex and packing geometry set by the cell cycle 1,2 . In the larval Drosophila epidermis, two cell populations, histoblasts and larval epithelial cells, compete for space as they grow on a limited body surface. They do so in the absence of cell divisions.

We report a striking morphological transition of histoblasts during larval development, where they change from a tensed network configuration with straight cell outlines at the level of adherens junctions to a highly folded morphology. The apical surface of histoblasts shrinks while their growing adherens junctions fold, forming deep lobules. Volume increase of growing histoblasts is accommodated basally, compensating for the shrinking apical area. The folded geometry of apical junctions resembles elastic buckling, and we show that the imbalance between the shrinkage of the apical domain of histoblasts and the continuous growth of junctions triggers buckling. Our model is supported by laser dissections and optical tweezer experiments together with computer simulations. Our analysis pinpoints the ability of histoblasts to store mechanical energy to a much greater extent than most other epithelial cell types investigated so far, while retaining the ability to dissipate stress on the hours time scale. Finally, we propose a possible mechanism for size regulation of histoblast apical size through the lateral pressure of the epidermis, driven by the growth of cells on a limited surface.

Buckling effectively compacts histoblasts at their apical plane and may serve to avoid physical harm to these adult epidermis precursors during larval life. Our work indicates that in growing non-dividing cells, compressive forces, instead of tension, may drive cell morphology.

Article activity feed

  1. Version published to 10.1101/2022.02.10.479869v2 on bioRxiv
  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

    In the manuscript "Compaction of Drosophila histoblasts in a crowded epidermis is driven by buckling of their apical junctions", Rigato et al. study the behaviour of histoblasts during the third larval stage of Drosophila. Histoblasts are the progenitor cells of the adult abdominal epidermis. They are produced in the embryo and lie between the larval epithelial cells (LECs) in the larva before abdominal morphogenesis takes place during metamorphosis. The authors use this system to explore the role of forces during epithelial development.

    The authors show that histoblasts undergo a dramatic change in morphology between 90 and 115 h AEL (after egg laying), with adherens junctions changing from a straight to a highly folded appearance. They also show that, during this time, histoblast volume is increasing. Furthermore, both junctional actin and junctional myosin II levels decrease over time. Interestingly, reduction of contractility (Rok-RNAi) and increase of contractility (overexpression of Rok constitutively active) in both histoblasts or LECs has no significant effect on histoblast behaviour.

    The authors hypothesise that growth of histoblasts and LECs lead to a compression of the histoblasts, which causes the buckling of the histoblast junctions. To test this hypothesis, they perform three experiments and show that (1) overexpression of a dominant negative form of insulin receptor in LECs leads to a non-autonomous effect on histoblast cell size, (2) overexpression of a dominant negative Rab11 in histoblasts interferes with histoblast buckling, and (3) overexpression of String in histoblasts induces histoblast division and reduces histoblasts buckling.

    Major comments

    As mentioned in the significance box below, a major weakness of the manuscript is that the presented data does not sufficiently support the hypothesis that histoblasts are compressed by the LECs: Firstly, the manuscript does not provide evidence that the LECs expand and compress the histoblasts. Such an analysis needs to be included.

    Secondly, the authors present two pieces of evidence to support their hypothesis, which are not very convincing:

    1. In line 147ff, the authors imply that their experiments in Fig. S3D (Rok alterations in LECs) support a 'mechanical tug of war'. However, the authors point out that their analysis is not statistically significant, so their data does not actually support their hypothesis.
    2. The authors show that overexpression of a dominant negative form of insulin receptor results in a non-autonomous increase in apical histoblast cell area. Here, I wonder to which extent the change in histoblast cell size might be due to a change in histoblast cell number per nest rather than a change in LEC pressure. Furthermore, they state that junctions are straighter, however, it would be good to see a statistical analysis here. Importantly, the conclusion of the experiment is not very convincing as it has not been shown that LECs are smaller, or grow slower, or exert less pressure on the histoblasts due to the performed genetic manipulation. So, the reason for the observed histoblast phenotype is unclear. This needs to be explored further.

    Thirdly, there are a few observations, which further challenge the author's hypothesis:

    1. Buckling appears to happen in the plane of the tissue, but should it not happen orthogonal to it, as shown in the scheme in Fig. 2J, if LEC compression was the cause of buckling?
    2. In Fig. 1D, one can see buckling at interfaces between histoblasts and LECs - does this not suggest that histoblasts push against LECs (and that in that area LECs are not pushing against histoblasts)? I think this observation is very interesting and a more detailed analysis of this phenomenon at histoblast-LEC junctions could be included in the manuscript.

    Based on the above point, there are various sections of the manuscript that need to be adjusted in my opinion, for example:

    • Line 190ff. 'Combined, these perturbation experiments provide strong evidence that junctional buckling of the histoblasts is the result of a imbalance between the addition of junctional material in the histoblasts and mechanical constraints from the overcrowding of the epidermis.' - There is not sufficient evidence for this statement with respect to overcrowding. The presented data does however show convincingly that junctional remodelling is needed for buckling.
    • Line 201 ff. 'This experiments confirms that junctional buckling is a result of the combined overcrowding and absence of divisions.' - The presented experiments do not provide sufficient evidence for this statement. The presented data does however show convincingly that inducing cell division leads to less buckling. I wonder whether this result is counterintuitive with respect to the authors' compression hypothesis, as if increased compression from LECs would lead to buckling would then not also increased pushing by neighbouring histoblast lead to buckling?
    • Line 235 ff. 'We investigated the formation of histoblast junctional folding and found that it is a non-autonomous transition originated by the competition for space of the two cell populations.' - This needs to be rephrased are there is not sufficient evidence for this hypothesis.

    Further major comments

    Line 56: More details about which histoblasts were imaged would help the reader understanding the experiments better.

    • Which abdominal segment was imaged?
    • The authors state that they have imaged the "dorsal posterior nest, which has the largest number of cells (15-17)" - is this correct? In later stages, the anterior dorsal nest is larger than the posterior dorsal nest.
    • Do the different abdominal segments have different histoblast numbers/ histoblast nest sizes?
    • I think that it would be helpful to show the behaviour of the ventral nest - the authors' hypothesis suggests that all histoblasts behave similarly and the buckling should be observed in all nests.

    In the figures, the authors do not clearly present the statistics done. Here, giving the p-values in the graphs would be very helpful. There are instances where it is not sufficiently clear whether the authors present a significant finding or merely a non-significant tendency (e.g. in Fig. 6).

    Fig. 4CD. I do not agree with the authors' interpretation that Sqh::GFP is lost from the junctions. The figure shows that levels are reduced. Also, in line 157, it might be better to use 'reduction of junctional myosin' rather than 'loss of junctional myosin'.

    With a few exceptions, the authors do not show the fluorescent marker used in their experiments to report where and how strong the Gal4 is expressed, which they use to drive their RNAi or dominant negative constructs. For example, in Fig. 5, they say that they have co-expressed a cytosolic GFP, but they do not show it. To have this kind of information would be very useful when interpreting the data.

    In line 171 and in other places the authors talk about 'plastic remodelling'. It would be helpful if the authors could explain in more detail what they mean by this.

    Minor comments

    In my opinion, the paragraph about the 'Qualitative model of junctional buckling' would be better placed in the discussion.

    Fig. 6. Why was the insulin receptor experiment done in white pupa and not in wandering L3 larvae? This makes comparison of data more difficult.

    I wonder when are histoblasts stopping to show the buckling? I assume that it must be before the beginning of abdominal morphogenesis, as at the beginning of LEC replacement, the junctions are straight again?

    What is the morphology of histoblasts in late L1- and late L2-stage larvae? Potentially there might also be crowding during those stages?

    Significance

    The observations by Rigato et al. are very interesting. They present a novel model system that enables the study of the interaction of two epithelial cell types, which do not divide. The fact that the presented data suggests that neither histoblasts nor LECs are under tension, makes this an extremely interesting novel system to explore the forces involved. The presented results provide some interesting insights into histoblast biology. However, a major weakness of the manuscript is that the proposed hypothesis of histoblast compression by larval epithelial cells is not sufficiently supported by the results. So apart from the very interesting observations, the manuscript lacks insight into the mechanistical basis of junctional buckling.

  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 manuscript by Rigato and colleagues uses live imaging of Drosophila larvae to describe the transition of histoblasts from a configuration where junctions are straight to a configuration where junctions are folded. The emergence of folded junctions is concomitant with a decreasing apical area, an increasing cell height and decreasing junctional Myosin. Expression of Rab11-DN (reducing apical junctional trafficking) or Cdc25 (inducing proliferation) in histoblasts abolishes the emergence of folded junctions. Expression of InRdn (impairing insulin receptor pathway) in surrounding larval epidermal cells (LEC) is claimed to also reduce folding (but see comment below). Finally, laser ablation of folded junctions reveals little or no relaxation, indicating that these junctions are not under tension. The authors propose a model in which growth mediated cellular crowding generates mechanical constrain on histoblasts which, in conjunction with decreased junctional Myosin II, leads to the buckling of junctions.

    • The authors convincingly show that histoblast junctions buckle/fold during late larval stages. They also convincingly show that junctional folding requires Rab11 and is counteracted by cell proliferation. The conclusion that folded junctions are not under tension is less well supported. Ablation of folded junction results in little/no relaxation (Fig. 5), however, the positive control (late pupal histoblasts) did not show obvious relaxation either (in absence of quantification).
    • The key conclusion that growth mediated crowding drives junctional folding is not well supported. The authors claim that InRdn expression in LECs leads to straight junctions of histoblast, but that is not obvious from Fig. 6F. Additional experiments will be required to further test the role of growth mediated crowding in junctional folding. For example, the authors should further attempt to modulate growth of LECs and/or test whether overgrowth of larvae to increase epidermal surface (e.g. lgl mutants) would influence junctional folding. These experiments will likely require several months.
    • Moreover, the authors provide a 'qualitative model' (Fig. 7). The paper would be strengthened by providing a model based simulation (best based on quantitative data derived from experiment (e.g. growth rates of histoblasts, LECs, etc.) of the transition of straight to folded junctions. This simulation would test whether the authors' hypothesized mechanism is feasible. This work will likely require several months.
    • Statistical analysis to reveal significances between datasets is missing throughout the manuscript. The number of biological replicates (i.e. larvae) for each experiment are not provided. The authors should provide statistical analysis.
    • Fig. 3/S3. The authors observe a correlation between junctional folding and a decrease in F-actin and Myosin on junctions. However, changes in Rok activity do not alter junctional folding. It is therefore unclear whether the decrease of F-actin and Myosin is a cause or consequence of junctional folding. The authors should tune down their conclusion that junctions 'soften' through a loss of cytoskeletal components.
    • Fig. 5. The laser ablation experiments require a quantitative analysis.
    • Fig. 6. A control image (Cad;mKate without perturbation) is missing.
    • Fig. 7. The junctional buckling model is a hypothesis by the authors to explain their observation; it is not data and therefore should be removed from the 'results' section of the manuscript.

    Significance

    The manuscript describes an interesting behaviour of cells to change their junctional morphology from straight to folded during development. Previous work has mainly described epithelia as networks of cell junctions under tension. This work advances our understanding by providing evidence that epithelia can also be in a configuration that is not based on tension. The authors provide evidence that this cell configuration requires the absence of cell proliferation and trafficking of junctional material. However, beyond this, the mechanisms that drive epithelia into this configuration remain somewhat unclear. The manuscript would therefore likely target a specialist audience

  5. Version published to 10.1101/2022.02.10.479869v1 on bioRxiv