WNTer wonderland

Development. Disease. Regeneration. Aging. The Wnt/β-catenin signaling cascade has provided enough surprises and insight to fill my career many times over. It all started with a hunch. Like many newly minted assistant professors, I was working on too many things at once, and I didn't want to abandon any of them. I couldn't resist trying just one more. During Christmas break in 1988, I stole back to the laboratory while most of my students were home for the holidays, and I injected scores of tadpole embryos with the RNA of a gene called INT-1 that had piqued my interest.

It was one of those goofball experiments that you never expect to work. A friend of mine, Andy McMahon at the Roche Institute (and now at Harvard University) had been struggling to knock out the INT-1 gene in mice where it was known to cause mammary tumors. We wondered what would happen in the reverse situation - if we overexpressed the gene. We hypothesized that since the gene was expressed in the nervous system, then perhaps the embryos would develop with neural defects. I certainly wasn't expecting to come back the next day to find a hundred tadpoles with two heads.

The forked spines of these embryos reminded me of a famous experiment that every developmental biologist knows by heart. In the 1920s, German Nobel laureate Hans Spemann transplanted sections of one embryo into another, creating a secondary embryonic axis and thus a double headed tadpole. He called the transplanted section the organizing center. What was remarkable was that it looked as if I had replicated the effect with a single gene - the INT-1 oncogene.

It was clear that something really important had fallen into my lap. That very day I euthanized all the meandering projects in the lab (except for those that grad students depended on to finish their PhDs), and focused everyone on INT-1. Working closely with McMahon, we repeated and developed the finding. We eventually put together a manuscript describing the axis-inducing activity for INT-1 in vertebrate development. It was featured on the cover of Cell¹ and the experiments were quickly repeated by many people in the field, either out of a sense of disbelief or just because they were so easy and exciting to do.

Today INT-1 is called WNT-1. Researchers realized that INT-1 and the Wingless gene in fruit flies were the same, so they combined the names to WNT. In fact, an entire family of 19 Wnt glycoproteins has now been identified in mammals alone. Each of these glycoproteins is expressed in spatially and temporally restricted patterns in embryos and adults. We now know that the Wnts are secreted factors that activate multiple signaling pathways in cells by binding to receptors known as the Frizzled receptors, as well as to other transmembrane proteins.

In the 20 years that followed that first experiment, others and I have tried to understand how Wnt operates. We have learned that the normal roles of Wnt signaling include control of cell proliferation, differentiation, and movements in a variety of contexts. We have discovered a tremendous amount about the mechanisms by which these Wnt pathways work, and what happens when they fail. Wnt signaling is altered in cancer, bone density diseases, familial exudative vitreoretinopathy (a retinal disease), inflammatory bowel disease, and other conditions. This raises the possibility of new therapeutics, and indeed a number of large and small companies now work on Wnt signaling as a target for therapies in a range of diseases.

So, back to the two-headed tadpoles. Why do they form and what does this tell us about the normal roles of Wnt signaling in early development? Carolynn Larabell, David Kimelman, and I showed that in the first hours of normal development, components of the endogenous Wnt pathway, including β-catenin, move along microtubule tracks to the future dorsal side of the fertilized egg. These become operational and activate the homeobox genes that in turn tell the embryo where to form the Spemann or gastrula organizer. Thus there is a direct link between Wnt/ β-catenin signaling and gene expression in directing the embryo where to make its body axis. We showed that later in development the same Wnt signaling is involved in programming the different patterns of neural tissue as well as different patterns of mesodermal cell types. Wnt signaling is clearly important in an array of activities in development.

We and others used a range of model organisms to tease out the mechanisms by which Wnt pathways operate. It isn't difficult to determine the genes that were originally discovered in Drosophila: Fruit fly geneticists have a penchant for wacky names. The rest of us provided the gravitas!

The first connections were made in the 1990's, when other researchers worked out that the Wnt pathway in Drosophila activates the function of the cytoplasmic protein called Dishevelled. Dishevelled functions upstream of Armadillo, which is called β-catenin in vertebrates. It was not clear how the product of the Shaggy gene (called Glycogen Synthetase Kinase-3, or just GSK-3, in vertebrates) worked; it was operating downstream of Dishevelled and upstream of Armadillo. Roel Nusse found that the Wnt receptor in flies was Frizzled, and a few months later (darn those slow reviewers!) we showed that this was also true for vertebrate Frizzled proteins. This gave us an outline of the pathway starting with a ligand, the Wnts, the receptors, the Frizzleds, and some of the cytoplasmic components that transduced the signal.

More connections started to appear. Kimelman, a wonderful colleague at the University of Washington, and I thought that the GSK-3 and β-catenin proteins might interact to activate the Wnt pathway. Sure enough, when we injected frog eggs with β-catenin RNA, the double-headed phenotype reappeared. However, it was blocked by coinjection of GSK-3 RNA, suggesting that the genes were working antagonistically in vertebrates as they do in flies. The question was: How was this working?

Kimelman and I suspected that GSK-3 might directly phosphorylate β-catenin to alter its function. To test this we mutated the site on β-catenin that we thought was most likely to be phosphorylated by GSK-3. Not only was this β-catenin not phosphorylated at normal endogenous levels, but it also accumulated to very high levels, first in the cytoplasm and subsequently in the nucleus. Nuclear accumulation resulted in a kind of hyperactive induction of the two-headed phenotype in frogs. ² I believe that the publication of this work was a key paper for the field, for three reasons: It showed directly that GSK-3 phosphorylates β-catenin, that this phosphorylation regulates posttranslational stability, and that the Wnt pathway controls the accumulation of β-catenin in the nucleus.

The first inkling that GSK-3-mediated phosphorylation might be relevant to medicine and not just frogs came from papers in Science and elsewhere showing that these same phosphorylation sites were often mutated in cancers such as melanoma. We realized that modulators of the Wnt pathways might have an important therapeutic application, and we immediately started to think about experiments that could demonstrate this in animals.

Courtesy of Cristi Stoick-Cooper
Day 1	Day 2	Day 3

Since Wnts were important in embryonic cells, I wondered if they might also be important in the most undifferentiated cells of adults, namely stem or progenitor cells. In a chance meeting with Mickie Bhatia then at the Robarts Research Institute of the University of Western Ontario, we decided to undertake what seemed like a simple-minded experiment. It was the rough equivalent of a human bone-marrow or cord-blood transplant: Human blood stem cells were introduced into immunodeficient mice in the presence or absence of a GSK-3 inhibitor. The GSK-3 inhibitor turns on the Wnt pathway by preventing the inactivation of β-catenin by phosphorylation. The mice that received the GSK inhibitor were better able to reconstitute their immune systems, indicating that the Wnt pathway could have a positive impact on stem cell function.

That simple experiment convinced us that small molecules targeting Wnt and other signaling pathways could help patients. ³ Patients with various types of cancer undergo radiation or chemotherapy, followed by transplantation with bone-marrow stem cells or cord blood to replace the immune cells that were killed by the treatment along with the cancer cells. However, there aren't many stem cells in cord blood; in fact, when patients receive stem cells from two cords, their chances of survival increase greatly. Could GSK-3 blockers increase the efficiency of cord blood transplants, conserving this key resource?

This idea, and the broader notion that developmental signaling pathways such as Wnts, Hedgehog, and Notch might be greatly untapped sources for therapies, led to the formation of Fate Therapeutics. Four colleagues and I started the company, with timely backing and great scientific insight from four venture capital companies lead by Arch Venture Capital in Seattle and Polaris based in Boston. One major goal is to discover drugs that regulate endogenous and transplanted stem and progenitor cells, as well as drugs that reprogram cells to adopt other fates. The company was launched in 2007, and is off to a great start with a pipeline of products in development and testing.

While helping start the company was important to me, I wanted to keep investigating Wnt signaling in the lab with my students and postdocs. I still had many questions in my mind about how it worked. For example, Wnt signaling plays multiple important roles in embryos - could we identify further roles in adults? One possibility was regeneration, since that involves stem and progenitor cells, which we knew were sensitive to Wnts.

Cristi Stoick-Cooper, then a graduate student, took the project on and in an elegant series of experiments showed that the Wnt/ β-catenin pathway is central to regeneration in zebra-fish tails. ⁴ With help from Gilbert Weidinger, then a postdoc, she found that blocking Wnt signaling through β-catenin completely blocked regeneration. Conversely, enhancing β-catenin signaling enhanced regeneration. Another postdoc, Hitoshi Yokayama, found that hind limb regeneration in frog tadpoles also required Wnt signaling through β-catenin. Thus, Wnt signaling is necessary for regeneration, raising new possibilities for therapies in mammals.

Of course, mammals have at best, a limited capacity to regrow limbs. Certain organs, such as the liver, do have some regenerative properties. We studied liver regeneration in mice and found that Wnt signaling through β-catenin was activated. These were exciting findings, among the first to show that Wnt signaling was key to control of tissue regeneration in mammals.

The finding raised the possibility that the converse might also be true: If Wnts served a positive role in regenerative processes, might Wnt signaling be reduced in some degenerative diseases? The literature gave hints that Wnt signaling might be important in Alzheimer's disease. For example, GSK-3 activity levels are elevated in patients, but this was not tied to a definite genetic connection. A highly motivated Chilean postdoc, Giancarlo de Ferrari, decided to look for this missing link. ⁵

Aside from Frizzled, Wnt sometimes uses a coreceptor called LRP6 to initiate its signaling pathway. Giancarlo showed that a statistically significant number of patients with Alzheimer's disease had a single amino-acid change in the LRP6 protein. In some cells this change reduced its signaling ability as we had hypothesized.

An interesting follow-up to this is that lithium chloride, an approved drug used to treat patients with bipolar disorder, inhibits GSK-3. As described above, the consequence of GSK-3 inhibition is activation of β-catenin. Could an increase in β-catenin by lithium treatment help protect neurons of patients predisposed to Alzheimer's? De Ferrari had some evidence to support this in rats from his PhD work. Nicely enough, recent data from other labs support this happening in humans. So I am hopeful that understanding Wnt signaling or at least GSK-3 inhibition might contribute to therapies for some patients with this horrible disease. We are working on drugs that might help the lithium to work better and at lower doses than are needed at present.

For many, developmental biology - a field that often studies early development of fruit flies, worms, mice, sea urchins, fish, and frogs - seems like it should be on the top-ten list of stupid things on which to spend federal funds. However, many biotech companies, including Fate Therapeutics, are emerging, aiming to find ways to harness Wnt and other early developmental pathways for useful drugs. It looks like the investment by the National Institutes of Health in this field is going to pay off.

I can only hope that politicians, and the public, begin to understand the importance of investing in basic research at nonfluctuating levels, which would encourage students to invest decades of their lives in training, so that the basic research of today will have major impact on therapies in the future. Certainly no one foresaw that studying flies and two-headed frogs would save lives, but we and others are convinced that this will happen, and we are dedicating our careers to making it a reality.

Randall Moon is a Howard Hughes Medical Institute investigator at the University of Washington and director of the Institute for Stem Cell and Regenerative Medicine.