Thread by @wntlab

New pre-print!

A novel Axin2 knock-in mouse model for visualization and lineage tracing of WNT/CTNNB1 responsive cells

Shared 1st authors @AnoeskaMoosdijk & Yorick van de Grift, with critical contributions by @Saskia_deMan & Amber Zeeman.

biorxiv.org/content/10.110…

Details ⬇️

Who would be interested?

#Wnt aficionados, developmental biologists and stem cell biologists who are interested in visualizing and tracking populations of WNT/CTNNB1 responsive cells.

What's new?

This model combines a fluorescent reporter (so cells with active WNT/CTNNB1 signaling can be directly visualized) with a lineage tracing driver (so these cells can also be permanently labeled and tracked to see what happens to their offspring).

Why did we build it?

Axin2 is known to report active WNT/CTNNB1 signaling, but existing models had drawbacks. We wanted to keep expression of the Axin2 gene intact and make sure that the reporter and the driver mimicked this endogenous expression as well as possible.

In addition, we wanted to move away from using a tamoxifen-inducible lineage tracing allele because we knew from experience that tamoxifen can interfere with outgrowth of the mammary epithelium.

This gave us the opportunity to test how well a direct fluorescent reporter in the Axin2 locus would work and how the use of a 3' knock-in with both P2A and T2A sequences would work out in vivo. Plus we would enter the realm of doxycycline-inducible lineage tracing.

So lots of technical novelties to help us develop new expertise in this area - critical for building more refined models in the future.

In an ideal world, this new model would also be the perfect reporter & lineage tracing driver.

Did we succeed? Keep on reading to find out!

First of all, the fluorescent reporter, which is expressed in the nucleus of all WNT/CTNNB1 responsive cells, is beautiful.

Thanks @joachimgoedhart for all FP tips/trix.

We also tested if the reporter responds to changes in WNT/CTNNB1 signaling, which we did in intestinal organoid cultures. And the answer is yes: the fluorescent reporter signal can be tuned up and down depending on whether you activate or inhibit WNT/CTNNB1 signaling.

We can also lineage-trace WNT/CTNNB1 responsive cells in the intestinal epithelium, hair follicles and liver.

So is it all peaches and cream?

Well... no, not entirely. We really wanted to build this model for lineage tracing in the mammary gland, but we've run into some issues there.

That's a euphemism for "so far that hasn't worked".

Why? We're not entirely sure, to be honest.

Explanation I favor is that we've been too successful: we've mimicked the endogenous Axin2 expression levels so closely, that we're finding out how low those levels really are - so low, that the fluorescent signal & the genetic switch are simply not strong enough to their job

Fascinating, right? We now have the knowledge & tools to build genetic models that recapitulate what happens at endogenous levels inside cells and now biology is showing us that it can make do with levels that are so low, that we are running into technical limitations.

Bummer.

This is the 1st mouse model to come out of the lab & it has been a long time coming. Funded by @kwf_nl in 2015 & delayed between then & 2017 for all sorts of logistical reasons. Bit risky for me as junior PI at the time, definitely risky and stressful for @AnoeskaMoosdijk.

Especially when the model didn't quite work in the mammary gland as designed...

But.... I do think it can be a very useful tool for others - in particular embryologists, since embryonic Axin2 expression levels are much higher.

We will deposit this model in an international repository as soon as we can (breeding is ongoing, but everybody knows that Mendel's law tends to behave as Murphy's law when you most need it to just do its job), while simultaneously finding this study a peer-reviewed home.

On a broader note: I'm used to this type of research taking a long time, but it is definitely not entirely (entirely not?) compatible with an academic system that requires every individual to focus on their own immediate output (publications).

Building and improving models one at a time can take years of effort (as was the case here). I think it is useful to keep investing in this, because that is the only way we can improve and refine them. But it really requires a long-term and multi-person effort.

So thanks to everyone in the lab for doing their share and contributing all the bits and pieces for this manuscript.

On to the next!

We will celebrate with virtual drinks and real ones when we can hang out IRL again.