Wnt signalling controls many steps in embryonic development. For example, it sets aside certain cells to do specific things - it actually specifies 'cell fates'. Also, Wnt signalling operates in the stem cell compartments of adult tissues, where the pathway is needed to maintain their homeostasis. Importantly, the Wnt pathway does different things, or has different outputs, in different cellular contexts. What determines these outputs is how it cooperates with other incoming signals.
Dr. Mariann Bienz | After her PhD at the University of Zurich, Switzerland, she started working at the MRC Laboratory of Molecular Biologoy in Cambridge in 1981. Five years later, she became Assistant Professor in Zurich, and in 1991 returned to MRC-LMB as a senior staff member, where she now is joint head of the Protein and Nucleic Acid Chemistry Division.
So the signalling within cells isn't as straightforward as one signal switching on one gene: several signals will interact with each other to determine the output?
Yes. Signalling in general, but Wnt signalling in particular, does not activate the same set of genes in every cell. That wouldn't work, because you want the signalling process to do something different in each cell. That is done by different signals acting together in a combinatorial fashion. With Wnt signalling, the cooperation with other signals is executed by an integrating module called the enhanceosome complex. The ability of Wnt signalling to activate a specific set of genes depends on which inputs act on this complex.
We have just published a paper that offers a mechanistic basis for how this enhanceosome works. We discovered an important module of the enhanceosome called the ChiLS - Chip/LDB-SSDP - complex, which can integrate several different signalling inputs by directly binding to or interacting simultaneously with several proteins.
It sounds like this enhanceosome complex acts almost like an interchange for several different molecules that control Wnt signalling. How did you discover this complex?
Well we've been trying to find this for about eight years or so. We've been working for a long time on a protein called Pygo that we discovered earlier in a genetic screen in flies. We knew that something needed to bind to it to help it to make β-catenin active. β-catenin is the main effector of what we call the canonical Wnt pathway, and it switches on gene transcription in response to Wnt.
We've been looking for many years to find out what binds to Pygo, it has been very technically challenging. But we finally found that this ChiLS complex binds to Pygo, and this has led us to define the enhanceosome complex. Through this discovery, we now know how Pygo works. And it basically explains how Wnt signals can be context-dependent and cooperate with other signalling inputs.
The enhanceosome complex | acts like an 'interchange' for the Wnt signalling pathway, pointing to genes that should be activated.
Beta-catenin is an important part of Wnt signalling because it has been implicated in cancer. What role does β-catenin play in causing cancer?
Without a Wnt signal, β-catenin is normally destabilised by a protein called the adenomatous polyposis coli APC protein, or APC. APC therefore stops β-catenin from relaying the Wnt signal to the nucleus, where it activates gene transcription. So if you mutate APC or it loses its function, you get hyperactive β-catenin.
Hyperactive β-catenin can make the Wnt pathway active in the wrong cells in the intestine to cause colon cancer. Once the Wnt pathway is active when it shouldn't be, cells become stem-cell-like and they begin to divide to create a polyp. The polyps are only benign - you need other mutations to get full-blown cancer. But almost every colorectal cancer in humans starts this way, so it must be massively important to keep β-catenin inactive.
Some of your earlier work was looking at a molecule called Dishevelled that has a key role in Wnt signalling. What did your research into Dishevelled find?
Each Dishevelled molecule on its own has a very low ability to bind to its partners. But when Dishevelled molecules come together in a signalosome blob, there is a very high local concentration that enables it to bind to other proteins very efficiently. This principle of localisation that we discovered is an important step in understanding how Dishevelled works within cells. And it's very very striking that Dishevelled forms these blobs that then enable it to do its job.
It sounds like the way your research works is you start looking into something, and it leads to something else and then to something else, so you're piecing together these puzzles of pathways. How do you choose which direction to go into at any one time?
Well, I think what you've just described is generally how research goes. You get interested in a question that you think is important for some reason. Or in this case you latch onto a striking observation that you think doesn't make sense or you don't quite understand. You just have a gut feeling that it must be meaningful because maybe it's unusual or there isn't a good explanation for it.
This gut feeling is hard to describe and I guess partly what makes scientists successful. You find something that doesn't quite fit what you know and pick it up to see if you can answer how it fits in. That's always the hard bit, actually finding the answer!
Do you really relish the complexity of the signalling systems you work with?
Not really! Most people don't. But we need to know how these complicated systems work because they are relevant for disease. Part of doing research is to contribute to knowledge per se, but another important part of it is to try to find new ways of tackling disease or preventing it or diagnosing it.
But what we are constantly trying to do is wade through a morass of complicated findings to try to construct a simple model that we can then test. And that is often quite challenging. Invariably most of these models are hopelessly oversimplified. But if we don't do that, we can never really draw conclusions. It's hard to make progress if you don't try to build a model that is at least testable.