(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Polarised cell intercalation during Drosophila axis extension is robust to an orthogonal pull by the invaginating mesoderm [1]

['Claire M. Lye', 'Department Of Physiology', 'Development', 'Neuroscience', 'University Of Cambridge', 'Cambridge', 'United Kingdom', 'Guy B. Blanchard', 'Jenny Evans', 'Alexander Nestor-Bergmann']

Date: 2024-05

As tissues grow and change shape during animal development, they physically pull and push on each other, and these mechanical interactions can be important for morphogenesis. During Drosophila gastrulation, mesoderm invagination temporally overlaps with the convergence and extension of the ectodermal germband; the latter is caused primarily by Myosin II–driven polarised cell intercalation. Here, we investigate the impact of mesoderm invagination on ectoderm extension, examining possible mechanical and mechanotransductive effects on Myosin II recruitment and polarised cell intercalation. We find that the germband ectoderm is deformed by the mesoderm pulling in the orthogonal direction to germband extension (GBE), showing mechanical coupling between these tissues. However, we do not find a significant change in Myosin II planar polarisation in response to mesoderm invagination, nor in the rate of junction shrinkage leading to neighbour exchange events. We conclude that the main cellular mechanism of axis extension, polarised cell intercalation, is robust to the mesoderm invagination pull. We find, however, that mesoderm invagination slows down the rate of anterior-posterior cell elongation that contributes to axis extension, counteracting the tension from the endoderm invagination, which pulls along the direction of GBE.

The aim of this study is to evaluate systematically the different possible impacts of mesoderm invagination on GBE. Some of the discrepancies reported can be explained by the challenge of quantitatively comparing wild-type embryos with mutant embryos defective for mesoderm invagination. Here, we tackle this difficulty by acquiring imaging datasets that are as comparable as possible between wild-type and twist mutant embryos. We use these datasets to quantify different metrics, including Myosin II densities and junctional behaviours ( Fig 1F ). We first ask whether the mesoderm mechanically interacts with the extending ectoderm. We next investigate whether a DV-oriented pull by the mesoderm could augment Myosin II enrichment at DV-oriented cell junctions via a mechanotransduction mechanism, and if this could increase the rate of junctional shrinkage during neighbour exchange events. Alternatively, if no such mechanotransduction occurred, then as DV-oriented cell junctions stretch in response to mesoderm invagination, cortical Myosin II might become diluted. Additionally, changing the balance of extrinsic forces could have mechanical effects on the speed of junctional shrinkage and growth throughout the process of cell neighbour exchange. The change in boundary conditions could also change how much cells elongate along AP in response to endoderm invagination. Finally, mesoderm invagination might help align cells, and subsequent neighbour exchanges, with the embryonic axes, so that neighbour exchanges drive maximal tissue convergence and extension.

While the role of endoderm invagination in GBE is well established, it is unclear what the impact of mesoderm invagination is, if any, as the results reported so far are inconsistent [ 3 , 27 – 30 ]. The autonomous cell behaviour underlying mesoderm invagination (and endoderm invagination) is apical constriction, which requires the assembly of a contractile network of actomyosin cytoskeleton at the apex of the cells [ 31 , 32 ]. During mesoderm invagination, the ventral-most mesodermal cells contract the strongest, making a ventral furrow through which the whole of the mesoderm tissue invaginates ( Fig 1A, 1C and 1E ). Because of the geometry of the mesoderm (an AP elongated rectangle, Fig 1A ), these ventral cells contract predominantly in DV causing a DV-oriented pull, which stretches the more lateral mesodermal cells along DV [ 33 – 38 ]. It is unclear how much this DV-oriented tension propagates to the adjacent ectoderm. It has been proposed that stretching of the lateral mesoderm (and also dorsal tissue) accommodates the mechanical tension induced by the invaginating mesoderm, providing a mechanical buffer for the ectoderm [ 13 , 35 ]. However, contrary to the idea that the germband is buffered from a mechanical impact from mesoderm invagination, other studies suggest that mesoderm invagination speeds up GBE [ 3 , 28 – 30 ]. In particular, it has been proposed that the rate of polarised cell intercalation increases in response to mesoderm invagination leading to enhanced planar polarisation of Myosin II, through a mechanotransduction mechanism [ 28 , 30 ]. If the ectoderm is indeed pulled by the invaginating mesoderm, another possible outcome, which has not been proposed so far, is that mesoderm invagination could act a brake for GBE, since the mesoderm pulls against the direction of convergence and perpendicularly to the direction of extension.

Scanning electron micrographs from Flybase ( A - D ) and tranverse section micrograph (from Munoz [ 55 ]) ( E ) showing anatomy of Drosophila embryos from early to mid-GBE (stages 6 to 8). Coloured overlays highlight the converging and extending ectodermal germband (magenta), mesectoderm (cyan), and mesoderm (blue). Red arrows depict tissue movements. The mesoderm invaginate ventrally, while the ectodermal germband starts convergence and extension. The endoderm invaginates where the pole cells are located, at the posterior end of the germband. Anterior to the left in ( A - D ). ( F ) Workflow diagram summarising image acquisition and analysis. Panels are maximum intensity projections of live spinning-disc confocal imaging of Gap43Cherry (upper panel) and sqhGFP KI (lower panel), followed by segmentation and tracking of Gap43Cherry channel to extract cell behaviours and interface metrics and quantification of Myosin II channel to extract Myosin density on cell interfaces (blue, low; red, high) and bipolarity measures. ( G and H ) Example frames for defining cell types in wild-type ( G ) and twist ( H ) movies (mesoderm/mesectoderm in cyan, ectodermal germband in magenta) overlayed on Myosin II channel (maximum intensity projection). Unmarked cells are those poorly tracked and excluded from the analysis. See also S1 Fig and S3 and S4 Movies. ( G ’ and H ’) Examples of segmentation of ectodermal cells (grey) and tracking of cell centroids (coloured lines) in wild type ( G ’) and twist ( H ’), showing trajectories of cells over the previous 5 minutes. ant, anterior; GBE, germband extension; PC, pole cells; post, posterior.

Early embryogenesis in Drosophila has become an important paradigm for understanding how tissue morphogenesis is driven by a combination of forces generated directly by cells (intrinsic forces) and, indirectly, at the tissue or embryo scale (extrinsic forces), as several well-characterised morphogenetic movements occur within a short period of time [ 16 – 18 ]. Specifically, at the same time as the mesoderm invaginates ventrally and the endoderm invaginates posteriorly, the ectoderm initiates convergence and extension to extend the main body axis of the embryo (germband extension (GBE)) ( Fig 1A–1E ). GBE is caused primarily by polarised cell intercalation driven by Myosin II [ 19 – 22 ]. Antero-posterior (AP) patterning controls the planar polarised distribution of junctional cortical Myosin II, which leads to shrinkage of cell junctions parallel to the dorso-ventral (DV) axis (so called DV-oriented junctions), causing cells to exchange neighbours [ 23 – 26 ]. In addition to this primary mechanism, an extrinsic AP-oriented pull from the invaginating posterior endoderm contributes to axis extension by driving elongation of cells in AP [ 3 , 4 ]. We showed by quantitative analysis that the elongation of cells along AP contributes to about one-third of tissue extension during the first 30 minutes of GBE (the so-called “fast” phase of extension [ 23 ], while polarised cell intercalation contributes to the other two-third [ 3 ]. In addition to elongating cells, the AP-oriented extrinsic force produced by endoderm invagination orients growing junctions during polarised cell intercalation [ 9 ].

The generation of tissue shapes during animal development is complex, but we are beginning to understand how cell autonomous behaviours such as oriented cell division, cell shape changes, and cell intercalation contribute to tissue morphogenesis. In addition to intrinsic forces generated within the cell, it is becoming increasingly clear that developmental morphogenesis can also be influenced by extrinsic forces acting at the scale of the tissue, generated by the deforming tissue itself or by the movements of neighbouring tissues [ 1 , 2 ]. For example, extrinsic forces in developing tissues can cause cell shape changes, drive or reorient cell intercalations, reorganise planar polarity, or lead to changes in gene transcription [ 3 – 12 ]. In addition to examples where physical interactions between tissues are important for their morphogenesis, it is becoming apparent that mechanisms also exist to buffer physical forces and mechanically isolate tissues from one another [ 13 – 15 ].

Results

Comparing germband extension in wild-type and twist mutant embryos To investigate whether mesoderm invagination has an impact on GBE, we carried out a quantitative analysis of GBE in wild-type and twist mutants, which lack mesoderm invagination. We chose to analyse twist mutants rather than snail or twist snail mutants, because although ventral furrow formation fails in all these mutants, some contractility remains in mesodermal cells in twist mutants alone, which decreases the width of the mesoderm and makes the ventrolateral field of cells more comparable with wild type. In snail or twist snail mutants, no contractility remains and the uninvaginated mesoderm takes significant space at the surface of the embryo [31]. We acquired movies of the ventral side of embryos from before the onset of mesoderm invagination until the middle of GBE, for both wild-type and twist mutants (Fig 1F–1H’). Four wild-type and 6 twist movies were analysed throughout this study. We labelled cell membranes with Gap43mCherry, and Myosin II with GFP-tagged knock-in of MRLC to visualise all Myosin II molecules (called sqhGFPKI here; [39]). Strains were constructed to ensure that these labels were expressed at the same level in wild-type and twist embryos (Methods). We also controlled the temperature during imaging (20.5 +/− 1°C) to limit variability. Cell contours were segmented from the Gap43mCherry channel and tracked over time (Methods and S1 and S2 Movies) to calculate metrics describing key cell and interface behaviours (Fig 1F) [40]. Myosin II density and planar polarity were quantified from the Myosin II channel as before (Fig 1F) [26]. Our field of view captured the convergence and extension of the ectodermal tissue on the ventrolateral surface of the embryo but also included mesodermal cells and mesectodermal cells. The inclusion of the ventral midline enabled us to precisely measure orientations with respect to this landmark (Methods). To restrict our analysis to the extending ectoderm, we carefully excluded mesodermal and mesectodermal cells based on (i) whether they had invaginated (wild type only); (ii) their proximity to the midline; and (iii) the timing of their cell divisions [41,42] (Fig 1G and 1H’ and S1, S3 and S4 Movies). Next, we checked that the ectodermal cells analysed in wild type and twist are comparable in terms of DV and AP genetic patterning. We examined the expression patterns of key DV patterning genes: single-minded (sim) is expressed in the mesectoderm, ventral nervous system defective (vnd) is expressed in the ectoderm abutting sim, and intermediate neuroblast defective (ind) is expressed next, abutting vnd [43]. While there is some de-repression of both sim and ind in the non-invaginated mesoderm and in the mesectoderm in twist mutants, the expression of all 3 genes is similar in the ectoderm (S2 Fig). We also checked the expression patterns of the LRR cell surface receptors Tolls 2, 6, 8, and Tartan, which are required for the polarised distribution of Myosin II during GBE downstream of the AP patterning genes [22,44–47]. We confirmed that these genes are expressed with the same patterns in twist mutants as in wild type (S2 Fig). Therefore, the regions of the ectodermal germband we analyse here in wild type and twist mutants are indistinguishable in term of patterning. To compare averaged metrics between wild type and twist mutants, we synchronised the movies in time, using the start of tissue extension along the AP axis (measured as AP tissue strain rate) in each movie as time zero as before (Methods) (S3A and S3B Fig) [3,4,26,47]. To check that timelines of wild-type and twist mutant movies were comparable once synchronised, we checked the timings of 2 developmental events: (i) when Myosin II becomes detectable apically in the ectoderm; and (ii) when the first cell divisions occur (Methods). We found that appearance of apical ectodermal Myosin II (around 10 minutes before GBE) and of the first cell divisions in the lateral ectoderm (around 32 minutes after GBE) are grouped in time across movies (S3C Fig). We note that mesectoderm cell divisions are delayed by nearly 10 minutes in twist mutants compared to wild type, which might be a consequence of misregulation of mesodermal and mesectodermal patterning genes. We conclude that our movie synchronisation gives comparable developmental time windows for the 2 genotypes. We next determined the number of cells tracked at each time-point in wild-type and twist movies (S3D–S3F Fig). At early time points, we track a total of approximately 800 cells (from 4 movies) for wild type and a total of 2,000 cells (from 6 movies) for twist. By the end of the period of analysis, over 2,500 cells are included for both wild type and twist. In summary, the above characterisation demonstrates that we have acquired comparable live imaging datasets of thousands of tracked cells of the ectodermal germband for wild-type and twist embryos covering a developmental window of 15 minutes prior to 30 minutes after the onset of GBE. We confirmed that DV and AP patterning in the extending ectoderm is unchanged in twist mutant compared to wild type, allowing us to explore the mechanical impact of mesoderm invagination on GBE independently of patterning.

Comparing the rates of junctional shortening and neighbour exchange in wild-type and twist mutants In addition to intrinsic forces within the germband driving its convergence and extension, ectoderm extension is facilitated by an extrinsic pull (along AP) from the posterior endoderm [3,4,9]. Both the convergence and the extension in the germband could be mechanically counteracted by the pull along DV from the invaginating mesoderm. If so, in the absence of mesoderm invagination, we might expect junctional shortening to speed up due to the lack of a DV tension opposing junctional shrinkage. Thus, boundary conditions imposed by the endoderm and mesoderm invaginations could affect the behaviour of ectodermal cells during GBE. Junctional shortening rates could also be influenced by mesoderm invagination in the opposite way (i.e., junctional shortening being faster in wild type than in twist mutants) through mechanosensitive Myosin II recruitment, as considered above. Therefore, we examined whether mesoderm invagination impacts on the rate of junctional shrinkage of the extending ectoderm throughout early and mid-GBE (0 to 30 minutes). We find that while there is a consistent trend over time for the proportional rate of junctional shrinkage to be higher in wild-type interfaces compared to twist, this difference is statistically not significantly different (Fig 4A). Because wild-type cell interfaces have been stretched by mesoderm invagination, they will generally be longer in length when they initiate junctional shrinkage for intercalation. Therefore, we also compared the speed of junctional shrinkage in microns per minute. We find that wild-type cell junctions initially shrink significantly more quickly than twist in absolute terms (S4E Fig). In summary, although wild-type interfaces start longer and shrink more quickly in absolute terms, in proportional terms, they shrink at approximately the same rate (no significant difference) but with a tendency for wild-type junctions to shrink slightly quicker. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Junctional shortening speed and rate of neighbour exchange in wild-type and twist mutants. (A) Proportional rate of interface shrinkage plotted against time to T1 swap during GBE (data for 0–30 minutes of GBE) summarised for wild type (magenta) and twist (green). (B) Average density of Myosin II on shrinking interfaces, plotted against time to an intercalation event (swap) during GBE (data for 0–30 minutes of GBE) summarised for wild type (magenta) and twist (green). (C) Example of detection of neighbour exchange events in the germband in wild type at 20 minutes of GBE. Loss of neighbours is shown as blue lines between cell centroids, while gain of neighbours is shown in red. Cell centroids are shown as black dots. Underlaid is an image of Gap43Cherry signal at the level of adherens junctions extracted for tracking. (D) Rate of neighbour exchange. Number of interfaces involved in a T1 swap expressed as a proportion of the total number of DV-oriented interfaces for all tracked ectodermal germband cells, summarised for wild type (magenta) and twist (green) over time of GBE. Data associated with this figure can be found in S3 Data. DV, dorso-ventral; GBE, germband extension. https://doi.org/10.1371/journal.pbio.3002611.g004 Since Myosin II is proposed to be the main source of mechanical stress driving junction shrinkage [19,20], we next quantified the density of Myosin II specifically on shrinking interfaces (Fig 4B). We find that wild-type Myosin II densities tend to be higher than those of twist mutants, but with only a very brief period of statistically significant difference between the 2 genotypes, at 5 minutes prior to neighbour exchange. This is consistent with the very mild effect on Myosin II in response to mesoderm invagination described in the section above. The slight decrease in Myosin density on shrinking interfaces in the twist mutant compared to wild type could explain the slight decrease in their proportional contraction rate. The proportional speed in junctional shortening should influence the rate at which cells exchange neighbours, and from above our prediction would be that there should be no difference between wild type and twist. We tracked discrete cell neighbour exchanges (T1 swaps) over time (Fig 4C) (see Methods). We find that the number of T1 swaps increase from the start of GBE until about 20 minutes and that wild-type and twist embryos show a comparable proportion of T1 swaps throughout the course of GBE, with no discernible delay in twist compared to wild type (Fig 4D). In summary, we conclude that there is no significant effect of mesoderm invagination on the proportional rate of junction shrinkage. However, wild-type junctions start longer and shrink more quickly in absolute terms and therefore are not delayed in reaching the point of neighbour exchange compared to twist mutants. Consistent with these observations, we find no difference in the proportion of cells undergoing intercalation events between wild-type and twist mutants. Therefore, there is no significant positive or negative effect of mesoderm invagination on the process of junctional shortening leading to cell intercalation events. However, we do see a slightly (not significantly) higher proportional rate of junction shrinkage in wild type compared to twist, accompanied by a slightly increased Myosin II density on shrinking junctions. Therefore, junctional shrinkage appears to be mildly affected by a mechanosensitive response of Myosin II, and this may mask the effects of boundary conditions on this process.

Comparing the orientation of cell interfaces during cell intercalation in wild-type and twist mutants Next, we asked whether the observed changes in the alignment of cells and cell junctions with the embryonic axes, in twist compared to wild type, could impact tissue extension, since this could lead to polarised cell intercalation being less well aligned with the embryonic axes, which might lessen how effective convergence and extension is. First, we examined Myosin II bipolarity alignment with the embryonic axes in twist mutants compared to wild type. We projected our bipolarity cell measure along the anterior-posterior axis to assess the extent to which Myosin II is enriched at the anterior and posterior of each cell (Figs 6A, S6A and S6A’). We find that the rapid increase of AP-projected Myosin II polarity is delayed by approximately 5 minutes and significantly reduced in twist compared to wild type, until approximately 10 minutes into GBE. This difference is more pronounced than for the unprojected data (compare Fig 6A with Fig 3E), showing that the Myosin II bipolarity of cells is less well aligned with the embryonic axes early in GBE, in embryos where mesoderm invagination is defective. We conclude that the Myosin II–enriched junctions in twist mutants are less well aligned perpendicular to the AP axis than in wild type. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. Orientations of growing and shrinking junctions and their impact on AP intercalation strain rates in wild-type and twist mutants. (A) Myosin II bipolarity, projected along AP, of all tracked ectodermal cells, summarised for wild type (magenta) and twist (green) plotted against time of GBE. (B) Orientation (degrees from AP axis) of shrinking interfaces at 6.5 minutes prior to T1 swap of all tracked ectodermal germband cells, summarised for wild type and twist over time of GBE. (C) Orientation (degrees from AP axis) of growing interfaces in ectodermal germband cells, summarised over the first 30 minutes of GBE for wild type and twist plotted against time from neighbour exchange event. (D) Orientation (degrees from AP axis) between cell centroids of pairs of “new neighbours” in ectodermal germband cells, summarised over the first 30 minutes of GBE for wild type and twist, plotted against time from neighbour exchange event. (E) Rate of productive neighbour exchange. Rate of net gains along the AP axis, using a continuous angular measure for productivity of T1 swaps to axis extension (see Methods), expressed as a proportion of the total number of DV-oriented interfaces for all tracked ectodermal cells, summarised for wild type (magenta) and twist (green) over time of GBE. (F) AP-projected intercalation strain rate of all tracked ectodermal cells, summarised for wild type and twist plotted against time of GBE. Data associated with this figure can be found in S5 Data. AP, antero-posterior; DV, dorso-ventral; GBE, germband extension. https://doi.org/10.1371/journal.pbio.3002611.g006 While the number of intercalation events is not reduced in twist compared to wild type (Fig 4D), the changes in average cell junction orientation (Fig 2K) and AP-projected bipolarity of Myosin II (Fig 6A) suggest that shrinking interfaces involved in T1 transitions may be poorly aligned with the DV axis. So, we measured the orientation of shrinking interfaces shortly before neighbour exchange and throughout GBE (Fig 6B) (Methods). We find no statistically significant difference between the orientations of shrinking interfaces in wild type and twist. In the first 5 minutes of GBE, there is a trend for shrinking interfaces to be less well aligned with the DV-axis in twist mutants, but this difference does not exceed 15 degrees (Fig 6B). These results suggest that the orientation of actively shrinking interfaces is not predominantly controlled by the global tissue movements and tension induced by mesoderm invagination. Because the maximum deviation in shrinking junctions misalignment we observe is 15 degrees at the start of GBE, then the maximum expected difference in strain rate projected along the DV axis from this misalignment would be: 1 –cos(15) = 3.4%, predicting a very minor effect. Consistent with this, we found no significant reduction in DV intercalation rate contributing to convergence, in twist embryos compared to wild-type embryos (see S4B Fig, the negative values show the convergence rate in DV). Previously, in other mutants affecting the orientations of shrinking interfaces, we had seen that the angle between junction shrinkage and growth was maintained (approximately 90 degrees), so growing junctions were also misoriented [50]. So, we also assessed the orientation of new junction growth. We did not detect a statistically significant difference in orientation of growing interfaces in twist mutants compared to wild type, although the trend is for twist to be less well aligned with the AP axis compared to wild type (Fig 6C). When plotting the angles between the centroids of newly neighbouring cells, we do see that these centroid–centroid angles are significantly less well aligned with the DV axis in twist mutants than in wild type for the first few minutes after a neighbour exchange (Fig 6D). Therefore, although T1 neighbour exchanges occur at normal rates in twist mutants (Fig 4D), the resulting cell packing is mildly perturbed compared to wild type soon after the neighbour exchange, but this effect is short-lived. We also quantified the rate of “productive” neighbour exchanges contributing to tissue extension (that is net T1 gains along the AP axis, taking into account the varying angles of T1 events with respect to the embryonic axes, see Methods) and observe that it is very similar in twist compared to wild type (Fig 6E). This shows that minor observed differences in angles of T1 events between wild type and twist do not impact on the ability of T1 events to extend the ectodermal germband. Finally, we measured the intercalation strain rate in AP to ask whether the minor differences that we observe in junction growth speed and orientation cause any difference in the extension of the tissue by cell intercalation (Fig 6F). This is a continuous measure of how much cell intercalation is contributing to tissue extension in AP [3,40] and therefore will integrate effects on the numbers of T1 events, speed and orientation of junctional shrinkage, and junctional growth (see S6B Fig for the definition of intercalation strain rate). We find no convincing difference between wild-type and twist embryos, though there is a trend for twist intercalation strain rate to be slightly higher, with several short bursts of significant differences between wild-type and twist. This tendency for twist intercalation strain rate to be mildly higher is consistent with our findings that the junctional growth rate is increased after T1 events (see Fig 5A). In summary, although there are measurable differences in orientation of shrinking and growing interfaces, the effect on tissue extension is negligible. Considering all our results together, we conclude that the polarised cell intercalations causing GBE are not significantly augmented by mesoderm invagination and that they are robust to the extrinsic force that it exerts.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002611

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.