Josquin Courte, PhD

#HelloQuitteX #20Janvier

#syntheticbiology #synbio #devbio #axialelongation #morphogenesis #preprint

@steffengrosser.bsky.social, who performed tissue fluidity analysis.

and Calvin Lam, who performed, ages ago, the first simulations of a potential circuit for tissue elongation!

Thanks to the whole team!

@leonardomorsut.bsky.social, our beloved PI,

Christian Chung, Naisargee Jain, Catcher Salazar: wonderful students I had the pleasure to manage for this project, their work was critical!

This concludes this "skytorial" ("blueditorial"?), we are looking forward to reviewers comments, and yours!

Any feedback will help us reach this milestone of synthetic developmental engineering: the programming of any tissue shapes through guided self-organization.

We need better cell engineering methods for inserting larger payloads with many genes induced by the same transcription factor.

And/or methods to avoid gene silencing, so we can do the same in a stepwise manner without losing the induction of the fist gene by the time we add the third!

For the 2nd and 3rd goals, we think the way is to screen more effectors, alone and in combination.

We tried to, but were faced with the common bane of synthetic biologists: gene silencing!

For the 1st goal, we have an idea of how to proceed:

Have the OFF transceivers express P-cad, and the ON transceiver N-cad.

This just requires a NOT gate downstream of synNotch.

...We have the plasmids already cloned, now we just need a brave student or postdoc to use them!

What we learn from this visualization method is that we need to:

1st: improve segregation of the “growing tip” and “structural support” regions,

2nd: further decrease tissue fluidity.

3rd: further control tissue growth speed.

... still a lot of work!

We then simulated elongation circuits in this “realistically bound” parameter space.

To better understand how our in vitro results compared to in silico ones, we used a visualization approach with a “morphospace” framework, inspired by @ricardsole.bsky.social

At this point, we have obtained the first elongating phenotype guided by synthetic gene circuits…

But we were still far of what the computational model promised… Probably in part because of its incomplete parametrization.

=> We re-parametrized the model with all our growth and fluidity data.

In the second class, we induced N-cad (+/- p21) in activated transceivers, and this too led to tissue elongation.

We combined our findings into 2 classes of “complete” circuits.

In the first class, we did not induce N-cad but had it constitutively expressed (its induction increased growth...)

And, lo and behold! When inducing the most promising fluidity control effectors, we obtained elongating structures!

At this point, we implemented 2/3 target design principles.

So, what about enforcing the segregation of the “growing tip” region?

=> We found that inducing N-cad upon transceivers activation (so, in the “structural support” region) would lead to better segregation! (We are not sure why)

Coming back to our design: we then screened candidate effectors for controlling tissue fluidity.

Here, we believe we share the first screen of this type, as we had to look for inspiration in many publications!

Our two best hits targeted actomyosin contractility: constitutively active RHOA & MLCK.

Side note: there still is a need for better and more specific genes controlling tissue growth.

We probably need to:

1) screen a large amount of (mutant) genes (directed evolution?)

2) simultaneously induce multiple effectors. We have prelim. data showing this works, the limit is cell engineering

… The most efficient growth inhibitors also increased tissue rounding speed!

(We do not yet know why.)

p21 was the only effector which both decreased tissue growth and fluidity, so we decided to build upon its induction.

With this method, we first screened effectors to inhibit tissue growth.

We found some promising ones, like p53.

… but there was a hidden issue with those…

We then screened effectors that could control tissue physics (growth and fluidity)

To test them in the most relevant context, we devised a pipeline where their impact could be evaluated in the right context: under the control of the transceiver circuit itself.

For more details, check Supp Fig 4 😉

We first identified a transceiver clone with a fate propagation speed in the right range. (the "Fast" clone here)

Now that we had a proof of concept in a computational model, we decided to go forward and implement this in real cells (mammalian fibroblasts: L929).

As we were trying to build a complicated circuit, we decided to proceed in a stepwise manner, splitting our work in semi-independent modules.

We used this model to identify 3 important design principles:

1) strong difference in growth rate between the “growing tip” and “structural support” regions.

2) strong segregation of the “growing tip”, through the right adhesion matrix

3) high rigidity of the “structural support” region.

To verify if this would work, we first implemented the circuit in our own CompuCell3D framework: pubs.acs.org/doi/abs/10.1021/acssynbio.0c00369

… and this worked great! (...before physics were completely parametrized!)

(color code mistake here, the growing tip is blue instead of gray!)

So, how to exploit this circuit design?

We got the idea that if transceivers changed their properties based on their activation status, this could result in tissue elongation.

For that, the cells must only proliferate when OFF (grey), and become collectively very rigid when ON (red).

This circuit can be used to propagate a change in cell fate in 2D in a wave-like pattern… And also in 3D!

In this animation, a spheroid of “sender” cells (red+green=yellow) is fused with a spheroid of “transceiver” cells. Activated transceivers become green.

But, how to translate this concept to a synthetic gene circuit?

We were very lucky… We could just take advantage of another circuit we developed in the lab, the “transceiver” circuit!

(here: www.nature.com/articles/s41467-024-53078-8)

Looking at other (very different!) examples, one can notice a common design principle:

A combination of local material addition (to the tip) with mechanical resistance to rounding forces (of the bulk).

So, how did we do that?

First, of course, we got our initial inspiration from the literature on axial elongation.

Notably this wonderful 2018 publication from @campaslab.bsky.social where the team demonstrated how tissue fluidity supports axial elongation in the zebrafish embryo!

While we were aiming to generate baguettes, we obtained something which is more in the range of olives / cocktail sausages, but, well, you have to start somewhere!

Here is a 20-ish posts summary, check my repost of Leonardo’s recap for a shorter version 😉

Delighted to share our last preprint!!! 🎉🎉🎉
biorxiv.org/content/10.1101/2024.12.11.627621v1

We share the first synthetic gene circuit guiding the self-organization of mammalian tissues in elongating structures by dynamically tuning growth, viscosity and adhesion.