|Home | About | Journals | Submit | Contact Us | Français|
A new study in Nature has identified a small molecule inhibitor for the oncogenic Wnt/β-catenin pathway that is responsible for many human cancers. This inhibitor targets an unsuspected cellular enzyme, Tankyrase, which controls the destruction of a β-catenin destructor.
Signaling by the Wnt family of secreted proteins plays essential roles in embryonic development and adult tissue homeostasis, and its deregulation has been implicated in many human diseases including cancers (Clevers, 2006; MacDonald et al., 2009). Indeed, a majority of colorectal cancers are associated with and likely caused by too much Wnt signaling (Kinzler and Vogelstein, 1996; Clevers, 2006). Inhibition of Wnt signaling has become an attractive strategy for cancer therapeutics (Barker and Clevers, 2006). A new study published recently in Nature (Huang et al., 2009), together with an earlier one (Chen et al., 2009), has identified a new class of small molecule inhibitors that blocks Wnt signaling in colon cancer cells, and along the way has uncovered a previously unknown mechanism of Wnt signaling regulation through an unsuspected enzyme, Tankyrase (TNKS).
The Wnt pathway that is relevant to this discussion is the canonical Wnt/β-catenin pathway, which regulates the stability of transcriptional co-activator β-catenin (Clevers, 2006; MacDonald et al., 2009). In the absence of Wnt stimulation, β-catenin is phosphorylated by the so-called Axin complex, which is assembled by the scaffolding protein Axin and consists of kinases GSK3 and CKIα, and the tumor suppressor protein APC (the adenomatous polyposis coli gene product) (Liu et al., 2002). Phosphorylated β-catenin is recognized and degraded by the ubiquitin/proteasome pathway, keeping cytosolic β-catenin at a low level (Figure 1A). Thus, the Axin complex is often referred to as the ‘β-catenin destruction complex’. Wnt binding to cell surface receptors leads to, via mechanisms that are yet to be fully understood, inhibition of β-catenin phosphorylation by the Axin complex, thereby stabilizing β-catenin (Figure 1B). Accumulated β-catenin enters the nucleus and binds to the TCF/LEF family of DNA-binding factors for activation of gene transcription (Clevers, 2006; MacDonald et al., 2009). Because Axin is a concentration-limiting component for the assembly of the destruction complex, Axin abundance has profound influence on Wnt/β-catenin signaling (Lee et al., 2003; Tolwinski et al., 2003).
Wnt/β-catenin signaling is of particularly relevance to colorectal cancers, which are the second leading cause of cancer death in Western societies and are becoming more prevalent in developing countries. APC mutations, which result in a lack of β-catenin phosphorylation/degradation, are responsible for Familial Adenomous Polyposis and more than 80% of sporadic colorectal cancers (Kinzler and Vogelstein, 1996), and β-catenin mutations, which directly allow β-catenin to evade phopshorylation/destruction, are detected in most of remaining colorectal cancer cases (Polakis, 2000). Because Wnt/β-catenin signaling is required for intestinal stem cell renewal and homeostasis, deregulated/elevated β-catenin signaling leads to excessive progenitor cell production without differentiation, predispositing tumorigenesis (Clevers, 2006). Given the constitutive β-catenin accumulation in colon cancers, it was generally believed that blocking agents targeting the Wnt pathway downstream of β-catenin are critical for therapeutics. Several small molecule inhibitors were identified based on this notion. For example, CPG049090 inhibits β-catenin/TCF interaction, and IGC-001 inhibits β-catenin interaction with the CBP transcription co-activator (Emami et al., 2004; Lepourcelet et al., 2004). These inhibitors block β-catenin functions in the nucleus without decreasing β-catenin levels. However, caution has been raised because β-catenin interacts with many proteins via similar or overlapping interfaces. For example, TCF and E-cadherin bind to the same β-catenin interface. Therefore, small molecules that disrupt β-catenin-TCF interaction will likely also interfere with β-catenin–E-cadherin interaction and thus epithelial tissue integrity, and will have severe therapeutic limitation (Barker and Clevers, 2006).
Enter the two new studies. Using a standard TCF/β-catenin-dependent reporter assay in mouse L cells, Chen et al. (2009) performed a high-throughput screening of synthetic compound libraries and identified several small molecules including IWR-1 that inhibit Wnt signaling. In a similar reporter-based screen of compound libraries in HEK293 cells, Huang et al. (2009) identified a structurally distinct small molecule Wnt inhibitor, XAV939. Intriguingly and somewhat unexpectedly, both compounds not only inhibit Wnt signaling in cells that have the unperturbed Wnt pathway, but also in colon cancer cell lines DLD-1 and SW480, which have APC mutations and thus constitutively high β-catenin signaling (Chen et al., 2009; Huang et al., 2009). Both groups found that their compounds result in dramatic stabilization of the Axin protein, thereby leading to increased β-catenin destruction even in the absence of APC function (Chen et al., 2009; Huang et al., 2009). Although previous studies have shown that Axin overexpression is able to promote β-catenin degradation in APC mutant cancer cells (Behrens et al., 1998), the fact that the endogenous Axin protein can be stabilized via IWR-1 and XAV939 to antagonize β-catenin signaling is quite striking. Indeed both IWR-1 and XAV939 increase β-catenin phosphorylation, and their effects depend on Axin and are abolished by siRNA against Axin (Chen et al., 2009; Huang et al., 2009). Both groups further showed that the two compounds inhibit colon cancer cell proliferation in vitro and tailfin and intestinal tissue regeneration in zebrafish in vivo, all of which represent processes known to require β-catenin signaling (Chen et al., 2009; Huang et al., 2009).
To identify the cellular target(s) of these compounds, Chen et al. (2009) found that biotinylated IWR-1 can pull down Axin (overexpressed) from cell extracts, leading to a simple interpretation that IWR-1 binds to and somehow stabilizes Axin. Using a similar pull down approach in combination with subtractive/quantitative mass spectrometry, Huang et al. (2009) were able to narrow down their candidate XAV939-binding proteins to an unexpected enzyme, TNKS. TNKS is a member of the poly-ADP-ribose polymerase (PARP) family, which uses NAD+ as a substrate to generate ADP-ribose polymers onto target proteins, resulting in a post-translational modification referred to as PARsylation (Hsiao and Smith, 2008). TNKS was first identified as a binding partner for telomerase repeat binding factor 1 (TRF1), which is a key player in the regulation of telomere length at the chromosome ends. TNKS has been shown to have roles in telomere maintenance and sister chromatid separation during mitosis through PARsylation of TRF1, as well as in vesicular trafficking of the glucose transporter in adipocytes (Hsiao and Smith, 2008). There are two TNKS genes, TNKS1 and TNKS2, in the human and mouse genome. Individual and double-knockout of Tnks1 and Tnks2 in mice suggest that they share significant functional redundancy (Chiang et al., 2008). Indeed, Huang et al. identified both TNKS1 and TNKS2 in XAV939-pull down experiments, and showed that the two TNKS proteins have indistinguishable properties in all assays they performed. Thus we will refer them together as TNKS.
Is TNKS the target through which XAV939 stabilizes Axin? Huang et al.'s experiments provided quite unambiguous answer (Huang et al., 2009). (i) XAV939 binds to TNKS catalytic (PARP) domain with rather high affinity (Kd = 93–99 nM), which is about 10 times stronger than its binding to other PARP proteins; (ii) XAV939 inhibits the enzymatic activity of TNKS but not of other PARP proteins in auto-PARsylation reactions; (iii) knockdown of TNKS by siRNA or RNAi increases Axin levels and decreases Wnt/β-catenin signaling in both mammalian and Drosophila cell cultures, showing a conserved role of TNKS in β-catenin signaling; (iv) TNKS binds to the Axin amino-terminal region via a nine-amino acid motif (called TBD for TNKS-binding domain), which is invariable among Axin proteins from Drosophila to humans. Indeed, the TBD destabilizes Axin protein and is sufficient to mediate the Axin stabilizing effect of XAV939—it is ironic that this conserved TBD in control of Axin stability has evaded detection, despite extensive studies of the Axin protein, likely due to the fact that Axin overexpression employed in most studies has masked the intrinsic instability of Axin. Remarkably, a heterologous TNKS-binding motif can functionally replace the Axin TBD in mediating Axin instability and Axin stabilization by XAV939; (v) Axin is PARsylated in a TBD-dependent manner by TNKS, both in vitro and in vivo, and Axin PARsylation appears to be required for its subsequent ubiquitination and degradation by the proteasome; (vi) Finally, further icing the case, the IWR-1 compound was also shown to bind to and inhibit TNKS. Huang et al. (2009) have thus provided strong evidence that XAV939 and IWR-1 inhibit TNKS, thereby stabilizing Axin and turning off β-catenin signaling (Figure 1C and D). It remains possible that IWR-1 pulled down Axin (Chen et al., 2009) indirectly via TNKS.
Huang et al. have offered an interesting showcase for chemical biology in cell signaling, from discovery of a lead compound for potential therapeutic development to identification of its molecular target and the elucidation of a novel regulatory mechanism. Like any exciting discovery, their findings have raised many interesting questions for further investigation. (i) Axin stability regulation remains poorly understood. For example, how does PARsylation by TNKS destabilize Axin? Why is Axin PARsylation required for its ubiquitination? Axin is phosphorylated by CK1 and GSK3 kinases, and Axin phosphorylation apparently increases its stability (MacDonald et al., 2009). What is the relationship between Axin phosphorylation and PARsylation? Wnt stimulation has been shown to lead to Axin destruction (MacDonald et al., 2009). Does Wnt signaling regulate TNKS activity/function? (ii) The role of TNKS in Wnt/β-catenin signaling in vivo remains to be clarified. It is somewhat puzzling that in zebrafish depletion of both TNKS gene products or application of XAV939, while significantly decreasing Wnt/β-catenin reporter expression, has little effect on early embryonic patterning that is known to be regulated by Wnt/β-catenin signaling (Huang et al., 2009). In the same vein, Tnks1 and Tnks2 double-knockout mice exhibit lethality at embryonic day 10 (Chiang et al., 2008), but the limited analysis thus far does not permit phenotypic correlation with known Wnt/β-catenin signaling defects. Thus, further genetic studies of TNKS genes in mice and also in Drosophila will be important to uncover whether TNKS has a global or tissue-restricted role in Axin destruction. (iii) Some glaring discrepancies between the results of Huang et al. and earlier studies on TNKS function during mitosis need to be resolved. Neither XAV939 nor siRNA against TNKS causes mitotic arrest in DLD-1 colon cancer cells, in sharp contrast with observations in other studies (Hsiao and Smith, 2008). Resolving roles of TNKS in Wnt/β-catenin signaling versus general cell cycle progression in different cell types studied will be critical for further understanding of TNKS functions. (iv) A recent study has demonstrated that telomerase may be a co-activator for β-catenin signaling in skin and intestinal stem/progenitor cells (Park et al., 2009). TNKS is another example of a protein that participates in both telomere regulation and Wnt/β-catenin signaling, although the mechanisms involved are quite distinct between TNKS and telomerase. As telomere maintenance and Wnt/β-catenin signaling are essential features of stem cell regulation, aging and senescence, the connection between the two processes is perhaps more intimate than we currently understand. (v) Finally, any therapeutic development targeting the Wnt pathway will likely have a long winding road ahead. The precise mechanism of XAV939 and IWR-1 will require co-crystal structures of the compound-TNKS to be determined, which will help rational designs for future generations of TNKS inhibitors that are more effective and specific. Given that Wnt signaling is required for normal tissue homeostasis, a major challenge, common to strategies that target any developmental signaling pathways, is how to develop therapeutic Wnt/β-catenin signaling inhibitors that suppress cancer formation/progression while leaving normal tissues minimally perturbed? The observations that both XAV939 and IWR-1 inhibit tailfin and intestinal epithelial regeneration in zebrafish call for caution (Chen et al., 2009; Huang et al., 2009). In any event, these new findings (Huang et al., 2009), together with the earlier one (Chen et al., 2009), represent a major advance in both chemical biological study of Wnt signaling and mechanistic understanding of this pivotal regulatory pathway.