Author's response

Bruno C. Vellutini shared about Patterned embryonic invagination evolved in response to mechanical instability

1. Would it make sense to use spatial transcriptomics to dive into detail of what’s happening at the population level at the head-trunk interface during gastrulation?

Yes. Spatial transcriptomics could help us uncover the identity of the “initiator cells”, for example. These are the cells that initiate the invagination of the cephalic furrow by shortening along the apical-basal axis at the onset of gastrulation. This initiator behavior depends on the co-expression of buttonhead and even skipped (upstream) and on the activation of lateral myosin (downstream). But the signaling cascade between these upstream and downstream factors remains unclear. We suspect that initiator cells must express a different, probably unique, combination of genes compared to their immediate neighbors. However, we haven’t been able to pinpoint these genes using the available single-cell transcriptome datasets in Drosophila. Being able to analyze the combinatorial expression of thousands of genes in situ with spatial information would be a promising approach to uncovering the transcriptional fingerprints of cells at the head-trunk boundary. Nevertheless, it may take a few years before spatial technologies with superior resolution and depth become more affordable and widely available.

2. Is there a connection between developmental time and furrow formation? What is the time component of this process (going into spatio-temporal processes/balance)?

Timing is essential to most developmental processes, and the cephalic furrow is no exception. Like the ventral furrow, the cephalic furrow is one of the first morphogenetic movements of gastrulation and begins to invaginate as soon as the embryo becomes cellularized. By the time the mitotic domains appear, most of the tissue is already invaginated. This quickness is important because cell divisions can disrupt morphogenesis—a cell cannot divide and invaginate at the same time. In addition to this, our work reveals another potentially relevant factor. Using simulations, we found that the earlier the cephalic furrow forms relative to the mitotic domains or germ band, the more effective it is at preventing ectopic buckling of the tissue. Therefore, we believe that the timing of cephalic furrow formation is also critical for its function as a mechanical buffer.Another interesting factor regarding timing is developmental speed. One consequence of faster embryogenesis is that complex morphogenetic processes get concatenated into shorter and shorter periods of time. This will inevitably lead to some kind of developmental conflict. And, as mentioned above, a dividing cell cannot invaginate. Interestingly, we found that mechanical instability in cephalic furrow mutants only occurs when both the mitotic domains and germ band extension are present (neither process alone can induce ectopic buckling). Their concomitant formation is a determinant factor for the increase in compressive stresses and the triggering of mechanical instabilities at the head-trunk boundary. This temporal overlap between morphogenetic processes is probably a consequence of Drosophila’s incredibly fast embryogenesis. And it makes us wonder whether developmental speed might have been a relevant factor associated with the evolution of the cephalic furrow itself.

3. Are similar mechanical buffer processes observed in other species and/or at another scale?

Great question. Epithelial folding as a result of mechanical instability is common; intestinal villi and brain folds are known examples. And when they buckle (or wrinkle), the tissue strain is relieved, which also, in a sense, buffers mechanical stress. But in this case, the folding happens through passive forces. The cephalic furrow is unique because it is patterned by genes and forms autonomously using active cellular mechanisms. Most patterned embryonic invaginations that form through intrinsic forces give rise to something obviously important. Like the famous ventral furrow in Drosophila that gives rise to mesoderm—without it, development fails. However, what we’re searching for are folds that likely have no associated cell fates, but that are patterned and active and that appear in contexts where there’s some kind of mechanical challenge. This could hint at a more physical role in development. Although it’s tempting to point to some folds out there, it’s a challenging question to answer. It depends on understanding the patterning, formation mechanisms, and tissue mechanics, as well as some comparative data, to better understand the evolutionary context. But once we start looking from this perspective, I believe that more examples of mechanical buffering folds will be discovered.

4. Can the CF also act as a chemical buffer (e.g. for signaling molecules), in addition to being a mechanical one?

That’s a good point for further exploration. The cells invaginated in the cephalic furrow become tightly juxtaposed with their apical sides facing each other, creating a transient niche environment. So it’s not so far-fetched to think this might facilitate some kind of cell-to-cell communication or isolate these cells from other extracellular signals. On an evolutionary scale, such a niche could even become a fertile ground for further cell fate innovation. The fact that the cephalic furrow is a relatively novel trait (~150 MYA) might explain why no associated fates have been identified—perhaps there wasn’t enough time for them to evolve. But testing whether the cephalic furrow can influence signaling is probably possible without having to wait a million years for the next experimental data point. One could check whether the invaginated cells are exposed or not to signaling molecules during development, or if the cell-to-cell contact has any direct effect on their identities. Another approach would be to investigate these invaginated cells in other species that have the cephalic furrow. If the invagination can influence cell signaling and fate, it might well have already happened in other fly lineages.

Yu-Chiun Wang and Steffen Lemke shared about Divergent evolutionary strategies preempt tissue collision in fly gastrulation

Thoughts

How do the mutations introduced in the preprints affect later developmental stages? i.e are they embryonic lethal?  

YCW: This I got asked a lot. To rephrase the question, what one would really want to know is what’s the late-stage phenotype or long-term consequence of loss of CF. Of the two methods that we developed to surgically remove CF, we hesitate to use evo1KO to assess this as we were not sure whether the enhancers that we removed from the eve genomic construct confer any late-stage expression pattern. If this indeed is the case, one cannot rule out the possibility that the late-stage phenotypes may be attributed to the loss of such late-stage expression. The other method – optogenetic inhibition of myosin contractility to block CF – is a more promising method and we are in the process of trying these experiments. We are also developing a second optogenetics-based method to manipulate Eve expression. This could have the advantage over the optogenetic inhibition of myosin contractility for lesser degree of mechanical (myosin inhibition) perturbation.

SL: Same here. The challenge is that we rarely introduce mutations in a way that do not affect anything else than the structure we like to manipulate. Simple mutations are very likely to have pleiotropic effects, but I agree with YCW: the use of optogenetics might be a promising path to overcome those constraints eventually.

Is CF evolution linked to a more complex life cycle? different environmental niches?

YCW: Not sure what “a more complex life cycle” really means. Whether the need to have a CF is linked to a particular environmental niche is an interesting question. Judged by the likelihood that CF is a conserved feature among cyclorrhaphan flies and cyclorrhaphan flies are present in a broad spectrum of niches, the answer to this question is probably a ‘no’.

SL: I would definitely agree with YCW – cyclorrhaphan flies can be found in too many different environments for the CF to be a likely adaptation for a particular niche. So the question remains: why did the CF evolve in the stem group of Cyclorrhapha? This is a question I get asked a lot, too. And, as YCW already outlines, this could have something to do with how other aspects of early embryonic development changed, like extra-embryonic development or maybe properties of the blastoderm epithelium. We are currently looking further into this, with one of the possibilities being that the stem group of Cyclorrhapha represent a time of major evolutionary innovations: from all what we know, the origin of biciod as the anterior determinant in head-to-tail axis specification also falls into this window!

Why do only a subset of flies, but not a broader range of insects, require a CF?

YCW: From the point-of-view of CF as mechanical sink that preemptively releases mechanical stress that can rise due to tissue expansion along the AP axis, I would surmise that this is a mechanical constraint that only long germ insects would like to be facing as the morphogenetic program of short germ insects along the AP axis is radically different. In some ways, this might be related to the organization of the extraembryonic tissues, which tend to be more pliable. Intuitively, it seems likely that the relatively large embryonic proportion of the extraembryonic tissues in the short germ insets could suggest that they may function as a major ‘shock absorber’ as the embryo proper grows and deforms.

SL: Just to add to what YCW already said about the differences in short and long germ insects: the way we tend to think about the requirements of a CF is a function of space, potentially colliding morphogenetic programs, and time. If there is in-plane division of cells in the head, cell intercalation by convergent extension in the trunk, AND all of this is happening at the same time and in a blastoderm where there is little else than head and trunk (read: little relative amount of extraembryonic tissue), then probably the insects requires some kind of mechanical sink that preemptively releases mechanical stress. Change any one of these parameters – and that could already be the relative timing of morphogenetic programs – then the need for a mechanical sink may be less pronounced.

How does this finding relate to mammalian developmental biology? 

YCW: I got asked about this recently as well. The mouse embryologist I spoke to told me that they were not aware of transient epithelial folds. Overall, it remains to be seen when and where epithelial folds form and retract outside cyclorrhaphan flies, be it mammalian or else. From the point-of-view of the need to evolve dedicated morphogenetic program to manage mechanical stress and avoid unwanted deformation, we hope that mammalian developmental biologists will begin to examine morphogenesis through the lens of this concept and they might be something conceptually similar.

SL: I agree: we have to keep in mind that the significance of the CF as a transient epithelial fold only started to become an exciting research question when we realised that it was a recent innovation. I am excited for some of the comparative breadth we currently gaining through in vivo and ex vivo developmental biology in mammals, and it will be certainly interesting to see how potential solutions for stress management could look like. As Chironomus taught us already: to dissipate mechanical stress, folds are probably not the only option.

Is the CF just a buffer?

YCW: In the context of early gastrulation morphogenesis, it does seem as if that is what the CF is.

After the CF unfolds, what do these cells become?

YCW: This is indeed what we are working on as we speak. Stay tuned!

Questions

1. Would it make sense to use spatial transcriptomics to dive into detail of what’s happening at the population level at the head-trunk interface during gastrulation?

YCW: This is an interesting question. We based our analysis on isogenized strains of Drosophila melanogaster. As such, we do not know whether there are phenotypic heterogeneities of head-trunk morphogenesis at the population level. Hypothetically, one would want to first survey wild populations to see whether there is phenotypic variability, and if so, to consider the possibility that such a variability may be due to genetic/transcriptomic heterogeneities. My hunch, though, is the CF is highly canalized in Drosophila melanogaster that one might not necessarily expect a large degree of phenotypic or transcriptomic variation.

SL: Very interesting question, yes! Not sure, though, whether I would want to restrict any hypothetical analysis to just spatial transcriptomics. Considering the CF as a means to dissipate stress, and stress at least in part being a function of the egg boundary as shown so beautifully by Vellutini et al., then probably you would want to include possible variation in egg shape and its genetic encoding in the mum as part of your analysis.

2. Is there a connection between developmental time and furrow formation? What is the time component of this process (going into spatio-temporal processes/balance)?

YCW: This is also a very interesting question. The CF obviously needs to be formed well ahead of the rise of compressive stress to function effectively as a mechanical sink. How the timing of its formation is controlled is certainly of great interest to us.

SL: Very interesting point indeed. As indicated already in the response to your question #3 in the section above: the timing of relative morphogenetic events is certainly a parameter that we are interested in.

3. Are similar mechanical buffer processes observed in other species and/or at another scale?

YCW: See the answer to the #4 question in the section above. This is indeed what we hope to see when more people begin to look into morphogenesis from the angle of mechanical stress management.

SL: Yes, that’s what I would think as well. In a way, the concept has been out there since Newmann’s initial proposal, but finding a proper framework to address it was not so easy. Maybe this is also why the two studies complement each other so nicely: Vellutini et al. looked at Drosophila wildtype embryos using high resolution SPIM imaging and found indications for buckling next to the CF. We have looked at many different fly species and found it was a recent innovation without apparent physiological function. The conclusion from these two very different starting points then were very similar: maybe we are looking at a mechanical element. Now that we know how such mechanical buffer processes can be detected in either wildtype data or via interspecies comparisons, I would hope to see more such phenomena to be reported.

4. Can the CF also act as a chemical buffer (e.g. for signaling molecules), in addition to being a mechanical one?

YCW: We are indeed intrigued by such a possibility. The inspiration includes Darren Gilmour’s work on how the luminal space created by the formation of a mechanosensory organ during lateral line morphogenesis serves to concentration FGF ligands and Cliff Tabin’s work on how intestinal folds concentrate BMP ligand in the mesenchymal layer underneath the epithelium.