WNTers in La Jolla

Wnts are a family of signaling proteins that exert a profound influence on cell fate and behavior in all animals thus far studied(Clevers, 2006; Nusse, 2005). Chosen from among nearly 200 abstracts, the presentations were roughly split between talks that probed mechanisms by which Wnt signals activate effector pathways or target genes, and those that explored how Wnt signaling governs higher-order phenomena, including development, regeneration and disease.

Several presentations were devoted to the mechanisms of signal generation by Wnt-secreting cells and signal presentation to responding cells. Recent work from the laboratory of Hendrik Korswagen (Hubrecht Institute, Utrecht,Netherlands) has demonstrated that efficient Wnt production in Caenorhabditis elegans requires retromer, an intracellular protein-sorting complex (Coudreuse et al.,2006; Prasad and Clark,2006). Korswagen and Xinhua Lin (Cincinnati Children's Hospital,OH, USA) reported that C. elegans MIG-14 and its Drosophilahomolog, Wntless (Wls, also known as Sprinter and Evenness interrupted)(Banziger et al., 2006; Bartscherer et al., 2006; Ching and Nusse, 2006; Hausmann et al., 2007), which are conserved transmembrane proteins that are essential for Wnt secretion,require components of the retromer complex for activity. MIG-14 protein levels and Wnt signaling were reduced in the absence of retromer function. Both groups found that the requirement for retromer in Wnt-producing cells in Drosophila is bypassed by Wls overexpression. These findings suggest a model in which the rate of Wnt secretion is governed by retromer-dependent recycling of Wls from the plasma membrane back to the Golgi(Fig. 1A).

Lipid modification of Wnt proteins and their low solubility in aqueous environments has raised the issue of how such aggregation-prone molecules transit through tissue to reach distant responding cells. The identification of Wnt-associated proteins, including lipophorin and proteoglycans(Eaton, 2006; Lin, 2004; Panakova et al., 2005),suggests that Wnts are not alone in this challenging task. `Contaminating'some of the biochemical preparations of a Drosophila Wnt protein, as reported by Kim Harnish in Roel Nusse's laboratory (Stanford University, CA,USA), was a member of the Lipocalin family of lipid transport proteins, which are best-known for their roles in transporting small hydrophobic molecules in plants (Grzyb et al., 2006). Harnish reported that Lipocalin-conditioned medium potentiated the activity of Wingless (Wg), a Drosophila homolog of the vertebrate protein Wnt1,in cultured cells, whereas Lipocalin depletion had the opposite effect. It will be of great interest to establish a role for lipocalins in regulating Wnt signaling in vivo.

How Wnt ligand-receptor interactions elicit intracellular signaling remains a subject of intense study. `Canonical' Wnt signaling leads to the accumulation of β-catenin and the altered transcription of target genes,but many `non-canonical' Wnt pathways are being increasingly recognized. In current models, canonical signaling is initiated by Wnt jointly engaging Frizzled (Fz) and low density lipoprotein-related proteins 5/6 (Lrp5/6, also known as Arrow - FlyBase) receptors, leading to Lrp5/6 phosphorylation(Cadigan and Liu, 2006; He et al., 2004). Dishevelled(Dsh, Dvl in vertebrates) is a family of scaffolding proteins that bind several Wnt signaling components and are required for Wnt signal transmission,but how Dsh inhibits the complex that promotes β-catenin destruction remains enigmatic (Wallingford and Habas,2005). Xi He (Children's Hospital, MA, USA) reported that mammalian cells that lack Dvl function had reduced levels of Wnt-induced Lrp6 phosphorylation, indicating that Dvl might act upstream of Lrp5/6 in canonical signaling. Together with work recently published by the group of Marcel Wehrli(Oregon Health Sciences University, OR, USA)(Baig-Lewis et al., 2007), the current data suggest that canonical signaling proceeds through discrete initiation and amplification steps mediated by the stepwise assembly of macromolecular complexes that include Wnt ligands, their receptors, Dvl, Axin and the protein kinases Casein kinase I (CK1) and Glycogen Synthase Kinase 3(GSK3, also known as Shaggy - FlyBase) that phosphorylate Lrp5/6(Fig. 1B). These findings extend the model for signaling via Lrp5/6-mediated Axin sequestration(Mao et al., 2001; Tamai et al., 2004) and further support the idea that efficient Wnt signaling requires the assembly of a large protein signaling complex, the Wnt `signalosome'(Bilic et al., 2007)(Fig. 1B).

Attracting much recent attention are a family of four secreted proteins termed R-spondins (RSpo), which can regulate canonical Wnt signaling in several tissues and which have been implicated in human disease(Kazanskaya et al., 2004; Kim et al., 2006; Wei et al., 2007). One possibility is that R-Spondins represent a class of non-Wnt ligands that bind and activate Wnt receptors in a fashion analogous to Wnt proteins themselves(Nam et al., 2006; Wei et al., 2007), but Minke Binnerts (Nuvelo, CA, USA) presented evidence that RSpo1 activates the canonical signaling pathway by antagonizing the activity of the extracellular Wnt inhibitor dickkopf 1 (Dkk1) in mammalian cells. Binding of RSpo1 to kringle containing transmembrane protein 1 (Kremen1), a transmembrane protein that associates with Dkk1 (Mao et al.,2002), results in increased Lrp6 levels at the cell surface. Thus,RSpo1 might lower the activation threshold for Wnt proteins by relieving the inhibition imposed on Wnt signaling by Dkk1(Binnerts et al., 2007).

Non-canonical Wnt pathways mediate important steps in morphogenesis, such as convergent extension during vertebrate gastrulation, but their relationship to the canonical signaling, which is often triggered by the same ligand,remains understudied. Alexandra Schambony in Doris Wedlich's laboratory(University of Karlsruhe, Karlsruhe, Germany) discussed the involvement of vertebrate Ror2, a receptor tyrosine kinase, in Wnt5a-mediated non-canonical signaling. Schambony reported that depletion of either Wnt5a or Ror2 resulted in defective convergent extension movements in Xenopus embryos(Fig. 2A,B), a finding that might be related to the dependence on Ror2 for Wnt5a-induced cell migration and filopodia dynamics in cultured cells(Nishita et al., 2006). Schambony demonstrated that a major target of Wnt5a/Ror2 signaling was transcriptional upregulation of the Xpapc (paraxial protocadherin) gene via a pathway involving PI3-kinase, cdc42 and Map-kinase-dependent activation of Atf2/Jun transcriptional complexes(Schambony and Wedlich, 2007). This report raises the possibility that targeting transcription might be a general property of noncanonical Wnt pathways. Wnt5a has long been recognized for its ability to inhibit canonical Wnt signaling(Torres et al., 1996). Amanda Mikels from the Nusse laboratory reported that Ror2 mediates Wnt5a-dependent inhibition of canonical β-catenin signaling(Mikels and Nusse, 2006), and suggested that the molecular basis of some cases of human brachydactyly(congenitally short digits) might be due to reduced Wnt5a/Ror2 signaling and hence elevated canonical Wnt signaling. It remains to be demonstrated whether Wnt5a-mediated inhibition of canonical signaling occurs via a signaling crosstalk or whether it requires transcriptional responses.

Recent studies in mice have revealed an unsuspected link between the cilium, a microtubule-based cell protrusion, and Hedgehog (Hh) signaling(Eggenschwiler and Anderson,2006; Huangfu and Anderson,2005). Functional interactions between Dvl and the ciliary protein inversin (Invs) (Simons et al.,2005) suggest that similar links might exist between cilia and Wnt signaling (Oishi et al., 2006; Park et al., 2006). Kevin Corbit from Jeremy Reiter's laboratory [University of California (UC) at San Francisco, CA, USA] observed upregulated canonical Wnt signaling in mouse embryos and cells lacking the kinesin Kif3a, a motor protein required for cilia formation (Marszalek et al.,1999). Cells lacking cilia contained stabilized β-catenin and showed increased CK1-dependent phosphorylation of Dvl, in support of a model in which cilia normally restrain the ability of CK1 to activate Wnt signaling. In a related presentation, Vera Voronina from Randy Moon's laboratory(University of Washington, WA, USA) reported that mice deficient in Chibby(Cby), an antagonist of β-catenin(Takemaru et al., 2003), lack mucociliary transport in sinuses, suggesting that hyperactive Wnt signaling caused by Cby deficiency might interfere with cilia function in vivo.

Precise regulation of Wnt signal-dependent gene expression requires a carefully orchestrated balance between activation and repression. Although complexes between β-catenin and the Tcf family of high mobility group(HMG)-domain transcription factors remain the commonly accepted link between Wnt signals and target gene activation, some developmental functions of Tcf proteins might be independent of β-catenin. Mice mutant in one of the four Tcf paralogs, Tcf3, die early in embryogenesis with partial axis duplications and gastrulation defects, presumably due to derepression of the crucial target gene Nanog in cells fated to give rise to the primitive streak (Merrill et al.,2004; Pereira et al.,2006). Bradley Merrill (University of Illinois, IL, USA) showed that mice homozygous for a Tcf3 allele that lacks the β-catenin binding sequence form a normal primitive streak, suggesting that Tcf3-β-catenin interactions are not required for Nanogrepression and primitive streak formation. Because Wnt proteins are expressed in preimplantation-stage embryos and are essential for primitive streak formation following implantation (Kemp et al., 2005; Liu et al.,1999; Wang et al.,2004), the question of which Tcf proteins transduce Wnt signals in early mouse embryos remains important.

Non-overlapping roles for Tcf proteins in different tissues might in principle be due to tissue-specific expression patterns or due to intrinsic differences in protein function. Two Tcf proteins, Tcf1 and Tcf4, exist as C-terminal splice variants (known as the E-tail) with distinct DNA-binding and functional properties (Arce et al.,2006). Marian Waterman (UC Irvine, CA, USA) showed that the E-tail encodes a novel sequence-specific DNA-binding domain that recognizes a GC-rich consensus sequence, thereby imposing additional specificity upon recognition of the core Wnt response element by the Tcf HMG domain. Her group found that the E-tail is essential for Tcf1 and Tcf4 regulation of Lef1transcription and for colon cancer cell proliferation, and thus might be involved in the activation of a subset of Wnt target genes that control cell growth.

Although initial studies suggested that Cby represses Wnt signaling by competing with β-catenin for Tcf binding(Takemaru et al., 2003),Ken-Ichi Takemaru (State University of New York at Stony Brook, NY, USA)proposed an additional mechanism for Cby action. Using mass spectrometry, his group identified 14-3-3 proteins as candidate Cby-binding partners. 14-3-3 proteins are a family of small, highly conserved proteins that are best-known for their roles in binding phosphorylated protein substrates and promoting their nuclear export (Dougherty and Morrison, 2004). Cby-14-3-3 binding enhances cytoplasmic sequestration of β-catenin in a manner dependent on the phosphorylation of a serine residue in the 14-3-3-binding domain of Cby. These observations suggest that Cby in concert with 14-3-3 inhibits signaling by facilitating nuclear export of β-catenin.

Wnt signals regulate cell polarity and asymmetric cell division, and some signaling components distribute in an asymmetric fashion during signaling events (Fig. 2C,D), but whether such localizations are a cause or a consequence of signaling remains controversial. Initial studies of Wnt signaling in C. elegansconcerned the Wnt-dependent nuclear export of the Tcf homolog POP-1 from the E blastomere, closer to the source of Wnt, but not from the more distant sister MS blastomere, at the eight-cell stage(Lin et al., 1998). Intriguingly, transiently reduced levels of nuclear POP-1 are observed in posterior daughter cells of nearly all subsequent cell divisions that are oriented along the anteroposterior (AP) axis, whereas SYS-1, a β-catenin homolog, is distributed in a complementary pattern(Huang et al., 2007; Phillips et al., 2007). By contrast, in vertebrate development, as originally demonstrated in Xenopus, nuclear localization of β-catenin is first observed in a population of dorsal cells prior to gastrulation(Larabell et al., 1997; Schneider et al., 1996). Similarly, nuclear β-catenin is found in specific groups of cells in early sea urchin (Logan et al.,1999; Wikramanayake et al.,1998) and sea anemone(Wikramanayake et al., 2003)embryos. Stephan Schneider from Bruce Bowerman's laboratory (University of Oregon, OR, USA) examined the localization of β-catenin during early animal-vegetal axis-oriented cell divisions in the polychaete annelid Platynereis dumerilii (Schneider and Bowerman, 2007). Remarkably, nuclear β-catenin levels were higher in vegetal daughter cells as compared with animal daughter cells(Fig. 2D). Treatment with GSK3 inhibitors that block β-catenin degradation promoted nuclear accumulation of β-catenin in animal daughter cells and resulted in animal-to-vegetal cell fate transformations. The observed asymmetries are proposed to constitute an ancient binary cell fate-specification mechanism, retained in annelid and nematode worms, that has been modified during evolution to regulate cell fate in populations of cells in other metazoans.

The dual functions of β-catenin as a Wnt effector and a component of the adherens junction have raised the question of how the different pools ofβ-catenin affect signaling (Nelson and Nusse, 2004). In C. elegans, the four knownβ-catenin homologs have evolved distinct signaling and adhesion functions, some of which appear to be crucial for Wnt-directed asymmetric cell divisions. Kota Mizumoto from Hitoshi Sawa's laboratory (RIKEN, Kobe, Japan)showed that forced expression of a membrane-anchored form of theβ-catenin homolog WRM-1 inhibits Wnt signaling and localization of endogenous WRM-1 to the nucleus in a manner dependent on APR-1, a C. elegans homolog of the vertebrate adenomatous polyposis coli(APC) gene (Mizumoto and Sawa,2007). Their data suggest a model whereby Wnt-dependent asymmetric cortical and nuclear localizations of crucial signaling components direct polarized cell divisions and subsequent fate determination of each daughter cell (Fig. 2C). Although a role for Wnt signaling in the asymmetric division of vertebrate cells remains largely inferred, a poster by Christophe Marcelle (Developmental Biology Institute of Marseille, Luminy, France) showed that Wnt11 acts as a directional cue to orient embryonic myocytes during skeletal muscle morphogenesis in chick embryos.

During mouse development, Wnt3a is expressed in the posterior region of the early- to mid-gestation embryo and is essential for mesoderm specification and somitogenesis (Takada et al., 1994). Bill Dunty from Terry Yamaguchi's group (National Cancer Institute, MD, USA) used microarrays to identify a Wnt3a-target gene in mouse termed mesogenin 1 (Yoon and Wold,2000), which encodes a basic helix-loop-helix (bHLH) transcription factor that promotes mesoderm maturation by acting as a feedback suppressor to limit the domain of Wnt3a expression.

Wnt signaling plays multiple sequential roles in cardiogenesis(Hamblet et al., 2002; Tzahor, 2007). Alexandra Klaus from Walter Birchmeier's laboratory (Max-Delbrueck Center, Berlin, Germany)demonstrated that Bmp and Wnt signaling have to be precisely timed for induction of the mouse primary and secondary heart fields, respectively. Jianbo Wang and Leah Etheridge (Wynshaw-Boris laboratory, UC San Diego, CA,USA) discussed the role of the three mouse Dvl genes in cardiac development. Using a BAC transgenic rescue assay, Wang showed that the conotruncal heart defects in Dvl1^-/- Dvl2^-/- mice are due to defective non-canonical Wnt signaling in the secondary heart field (which normally gives rise to the right atrium, ventricle and outflow tract), but not in the cardiac neural crest. Etheridge used green fluorescent protein(GFP)-tagged Dvl transgenes to rescue Dvl-mutant phenotypes and observed distinct subcellular localizations for the different Dvl proteins in the cochlea, an organ that requires non-canonical Wnt signaling for the proper alignment of stereocilia. These observations suggest that the three mammalian Dvl proteins might play unique roles in specific signaling events during vertebrate development.

In the nervous system, Wnts have been shown to regulate neuronal polarity and migration, axon pathfinding, and synaptogenesis(Hilliard and Bargmann, 2006; Lyuksyutova et al., 2003; Pan et al., 2006; Salinas, 2005; Speese and Budnik, 2007; Zhang et al., 2007; Zou, 2004). Jason Kennerdell from Cori Bargmann's laboratory (Rockefeller University, NY, USA) showed that CWN-2, one of the five C. elegans Wnts, acts with both Fz and ROR(receptor tyrosine kinase-like orphan receptor) receptors to regulate the position of the pharyngeal nerve ring. Previous work has suggested that Wnt signaling promotes synaptogenesis(Salinas, 2005; Speese and Budnik, 2007), but Kang Shen (Stanford University, CA, USA) showed that Wnt signaling inhibits en passant synaptogenesis of the DA9 neuron in C. elegans(Klassen and Shen, 2007). Scott Clark (New York University, NY, USA) reported that the RING-finger protein PLR-1 downregulates Wnt signaling by reducing the cell surface levels of Frizzled during neuronal development. Yimin Zou (UC San Diego, CA, USA)presented data suggesting that PI3 kinase and atypical protein kinase C (aPKC)mediate Wnt4 signaling in migrating commissural axon growth cones after crossing the midline in the vertebrate neural tube. In support of these observations, a recent study has suggested that the interaction of Dvl and aPKC mediates Wnt signaling in neuronal polarity(Zhang et al., 2007).

Several presentations explored the role of Wnt signaling in vertebrate eye development. Fumi Kubo from Shinichi Nakagawa's laboratory (RIKEN, Wako,Japan) identified Hairy1, a bHLH transcription factor whose transcription can be regulated by the Notch pathway (Davis and Turner, 2001), as a target of Wnt2b that maintains stem cells in the ciliary marginal zone of the developing eye(Kubo et al., 2005). Pygopus(Pygo) is required for all β-catenin-dependent transcription in Drosophila via bridging interactions with Bcl9 (also known as legless, Lgs) (Belenkaya et al.,2002; Kramps et al.,2002; Parker et al.,2002; Thompson et al.,2002). By contrast, many tissues that require Wnt signaling develop normally in mice that lack the two pygo homologs, pygo1 and pygo2 (Li et al., 2007; Schwab et al.,2007). Birchmeier's and Richard Lang's (University of Cincinnati,OH, USA) groups demonstrated that the microphthalmia seen in pygo2^-/- mouse embryos correlates with a lack of Pax6 expression and this function of Pygo2 appears to be Wnt-independent (Song et al.,2007). Li Li in Birchmeier's laboratory suggested that the small-eye phenotype of pygo2^-/- mutants is due to a failure of pygo2 to activate the extracellular Wnt inhibitor Sfrp2 in the eye field.

Wnts are implicated in stem cell maintenance and are required for tissue regeneration. In the keynote address, Hans Clevers (Hubrecht Institute,Utrecht, Netherlands) described his laboratory's characterization of Wnt/Tcf-regulated target genes in intestinal development and homeostasis using a combination of chromatin immunoprecipitation and microarray approaches. Although thousands of Tcf-binding sites exist on chromatin, only a few of them correspond to bona fide target genes. One target, Gpr49 (also known as Lgr5 - Mouse Genome Informatics), encodes an orphan G-protein-coupled receptor that is expressed in colon cancers and appears to mark intestinal stem cells. Lineage tracing of Gpr49-expressing cells using Cre-based recombination technology resulted in the long-term labeling of whole intestinal crypts, supporting a model in which individual crypts can be maintained and renewed by the progeny of a single stem cell at the base of the crypt. In a related technical advance, Calvin Kuo (Stanford University, CA, USA) described his laboratory's long-term culture (>3 months) of intestinal stem cells.

Several presentations concerned the Wnt-dependent regulation of stem cell self-renewal and tissue regeneration. Arial Zeng from the Nusse laboratory showed that reporter activity for the Wnt target Axin2^lacZenriches for self-renewal in mammary stem cell preparations. Presentations from Wolfram Goessling and Trista North from Len Zon's laboratory (Children's Hospital, MA, USA) discussed the synergy between the prostaglandin and Wnt/β-catenin pathways in hematopoietic stem cell renewal(North et al., 2007) and in liver regeneration. The data suggest that the reported effects of prostaglandins on morphogenesis and inhibition of cancer might be mediated by Wnt signaling (Castellone et al.,2005; Cha et al.,2006). Yasuhiko Kawakami from the Izpisua-Belmonte laboratory(Salk Institute, CA, USA) discussed his studies that demonstrate a crucial role for Wnt/β-catenin signaling in limb regeneration across vertebrate evolution (Kawakami et al.,2006). Growth and regeneration in vertebrate limbs require interactions between ectoderm and mesenchyme that limit chondrogenesis to those cells distant from surface ectoderm. Lilia Topol from Yingzi Yang's laboratory [National Institutes of Health (NIH), MD, USA] discussed a dual mechanism used by the pro-chondrogenic HMG-domain transcription factor Sox9 to inhibit the anti-chondrogenic effects of the Wnt/β-catenin pathway(Akiyama et al., 2004; Hill et al., 2005). Derk Ten Berge from the Nusse laboratory proposed that an ectodermal Wnt signal suppresses Sox9 expression in underlying mesenchyme, and hypothesized that the extent of growth and differentiation during limb development might be predicted from the range of Wnt3a activity in cultured limb explants.

Despite the growing list of Wnt signaling-associated diseases, therapies that rationally target Wnt signaling or its known targets have yet to make a clinical impact. Several presentations and posters described screens for small molecules that affect Wnt signaling. Michael Kahn (University of Southern California, CA, USA) described two inhibitors that disrupt the interaction between β-catenin and the CREB-binding protein (CBP or p300) histone acetyl transferase coactivators, the choice of which determines whether embryonic stem (ES) cells differentiate or proliferate, respectively(Ma et al., 2005). Interestingly, blocking the p300/β-catenin interaction with one of these molecules supported the proliferation of ES cells in culture, without serum or added growth factors, in a Wnt3a-dependent fashion(Miyabayashi et al., 2007). To avoid tumor formation following therapeutic stem cell implantation, the balance between proliferation and differentiation of implanted cells needs to be highly coordinated. In this regard, Ernest Arenas (Karolinska Institute,Stockholm, Sweden) successfully differentiated fetal mesencephalic stem cells into dopaminergic (DA) neurons in a Wnt5a-dependent manner(Bryja et al., 2007) and demonstrated their efficacy when transplanted into a mouse model of Parkinson's disease. These proof-of-principle studies are required to justify future stem cell transplants intended to replace lost DA neurons in Parkinson's patients (Parish and Arenas,2007).

Vanessa Bres in Katherine Jones' laboratory (UC San Diego, CA, USA)discussed the cooperation of Wnt and Notch signaling pathways in human breast and colon cancer. Charlotta Lindvall in Bart Williams' laboratory (Van Andel Research Institute, MI, USA) showed that rapamycin-dependent inhibition of the target of rapamycin (TOR) pathway, as well as loss of Lrp5, inhibits tumor cell growth in a mouse mammary tumor virus (MMTV)-Wnt mouse breast cancer model, in line with the known tumorigenic properties of the TOR pathway(Bjornsti and Houghton, 2004; Hay and Sonenberg, 2004). Predisposition to Wilms tumor, a childhood renal cancer with elevated Wnt signaling, is often due to mutant WT1 or WTX (FAM123B) proteins, the latter of which was recently shown to associate with the β-catenin `destruction complex' (Major et al., 2007; Rivera and Haber, 2005; Rivera et al., 2007). However,the mechanism by which WT1 suppresses Wilms tumor formation or canonical Wnt signaling remains unclear. Myoung Shin Kim in Sean Lee's laboratory (NIH, MD,USA) searched for WT1 targets (Kim et al.,2007) and identified a novel gene product similar to the Dvl-binding protein IDAX (CXXC4) (Hino et al., 2001), which might mediate WT1-dependent suppression of Wnt/β-catenin signaling. Catriona Jamieson (UC San Diego, CA, USA)described a population of cancer stem cells in blast-phase chronic myeloid leukemia (CML) with elevated Wnt/β-catenin signaling, presumably due to secondary mutations in the machinery that promote β-catenin degradation(Jamieson et al., 2004). In some CML patients in blast crisis, she identified unique splice mutations in the GSK3B gene that might be responsible for elevated Wnt/β-catenin signaling, but whether the mutations act in a loss-of-function or dominant-negative manner remains to be determined.

Given the widespread role of Wnt/β-catenin signaling in stem cell and tissue homeostasis, specificity of therapeutic targeting remains a formidable challenge. Venita DeAlmeida from Paul Polakis' group at Genentech (CA, USA)discussed their efforts to generate an extracellular Wnt inhibitor with optimal pharmacokinetic properties, efficacy and minimal side effects(DeAlmeida et al., 2007). The extracellular domain of the Fz8 receptor fused to the Fc portion of human immunoglobulin remains relatively stable when injected into mice(t_½∼4 days). The fusion protein significantly inhibited growth of MMTV-Wnt1 mammary tumors, as well as tumors derived from PA-1 and NTera2 teratoma cell lines, whose growth in vitro requires autocrine Wnt signaling. Remarkably, at doses in which the chimeric molecule suppressed the growth of MMTV-Wnt1 and NTera2 tumors in nude mice, there were no adverse effects on β-catenin levels or on the histology of intestine or hair follicles, two tissues whose constant turnover requires Wnt signaling. These data provide hope that Wnt signaling might someday be selectively targeted in diseased cells and tissues with minimal host toxicity.

As co-discoverer of Wnt proteins a quarter century ago, Roel Nusse closed the meeting by commenting that the Wnt field continues to be characterized by a `vigorous and rich' array of science. The meeting demonstrated the need to further understand the context-dependence of diverse Wnt pathways and the molecular bases by which they integrate with other signaling processes in the cell. The breadth of Wnt-related investigations and their impact in modern biomedical science is truly amazing. We anticipate that the growing significance of `all things Wnt', as evidenced by over 6,000 Wnt-related references in PubMed, ensures that future Wnt meetings will be as exciting as the one in La Jolla.

We thank Karl Willert, Anthony Wynshaw-Boris and Katherine Jones for a flawless meeting organization and a stimulating scientific program;individuals mentioned in the review for their feedback on the manuscript and for their permission to discuss unpublished data; and all meeting attendees for lively scientific discussion. We regret that due to space constraints we were unable to cover many interesting presentations. We acknowledge funding by March of Dimes Birth Defects Foundation (to S.Y.S.) and the NIH (to S.Y.S. and K.A.W.). K.A.W. is a W. W. Caruth, Jr, Scholar in Biomedical Science at UT Southwestern.