We demonstrate an essential role for endocytic trafficking in Wnt signaling. Through the use of various small molecule, dominant-interfering, and loss of function inhibitors of distinct stages of endocytosis, we define an explicit requirement for internalization in Wnt signaling. Our observation that β-catenin stabilization by both mammalian (Wnt-3A) and Drosophila (Wg) Wnt proteins is sensitive to blockade of endocytosis suggests that this requirement may be universal for the Wnt pathway, and not merely limited to one particular Wnt. After submission of this manuscript, Seto and Bellen  showed that Wg signaling in vivo is also dependent on internalization. The differential sensitivity of β-catenin stabilization to blockade of endocytosis in L cells treated with lithium and the APC mutant SW480 cells indicates that the key endocytic step may reside between GSK3β and APC in the pathway. Indeed, it was quite surprising that the stabilization of β-catenin effected by lithium stimulation (presumably through inhibition of GSK3β) required an intact endocytic pathway, given traditional models of the Wnt pathway scheme. However, recent studies have implicated GSK3β in the regulation of endocytic trafficking, perhaps independent of its role in the Wnt pathway [17,18]. Doronin et al.  demonstrates that stimulation of cells with lithium causes rapid endocytosis of the β2-adrenergic receptor, and Pelkmans et al.  report that siRNA-mediated silencing of GSK3β causes internalized transferrin to accumulate in early endosomes. Incorporating these recent findings [17,18] with our epistasis data could suggest the following scenario: inhibition of GSK3β in response to a Wnt signal (or directly via lithium) facilitates rapid internalization of the signaling machinery in a manner which sequesters it from the APC protein, thereby leading to the stabilization and nuclear translocation of β-catenin.
Over the past few years, a number of studies have suggested that endocytosis may have a facilitating role in signal transduction. For example, activation of Erk1/2 directly through RTKs, or through certain GPCRs, has been shown to require endocytosis [19,20,8]. Indeed, activated RTKs and GPCRs have been found to be present on populations of clathrin-coated vesicles containing a number of intermediate proteins in the Erk1/2 pathway including Ras, Raf, MEK, and Shc [21,19]. Further, it was recently discovered that clathrin-mediated endocytosis and subsequent trafficking of TGFβ receptors to endosomes is required for productive signaling . These studies have contributed to a "signaling endosome hypothesis," suggesting that this organelle may serve as a nexus for a number of signal transduction pathways. Indeed, partial co-localization of internalized Wnt with transferrin further implicates the clathrin-mediated pathway in this process and suggests that a perinuclear recycling endosome functions as a compartment for Wnt signal transduction.
In contrast to the RTK, GPCR, and TGFβ pathways, very little is known about the subcellular localization of Wnt signaling. The few studies examining the potential endocytosis and trafficking of Wnt proteins have considered this as a mechanism for the formation of the Wnt morphogen gradient in the embryo or the termination of the pathway [5,6]. To date, no previous studies have suggested that internalization is explicitly required for propagation of Wnt signaling cascades. In an elegant study in the Drosophila embryo, a horseradish peroxidase-conjugated Wg was shown to be degraded by receptive cells in a manner sensitive to chloroquine and dependent upon deep orange , suggesting that the ligand is trafficked to the lysosomes .