|Home | About | Journals | Submit | Contact Us | Français|
WNT signaling is a fundamental molecular pathway in both embryogenesis and disease. Nephron development is dependent on WNT signaling. The nephron epithelia proximal to the collecting duct develop from progenitor cells in the metanephric mesenchyme. The process involves formation of proto-epithelial cell aggregates, conversion into epithelia, and proximal-distal patterning of the nephron. Two ligands from the WNT family, namely Wnt9b and Wnt4, are required for nephron differentiation. Recent studies have addressed the downstream targets of these WNT ligands and delineated the role of the canonical WNT signaling pathway. This pathway depends on the intracellular protein β-catenin and the T cell-specific transcription factor/lymphoid enhancer factor-1 (TCF/Lef1) family of transcription factors. Selective β-catenin signaling antagonism inhibits differentiation of metanephric mesenchymal progenitor cells, while forced activation triggers a stage progression towards proto-epithelial aggregates. Nonetheless, activation of the pathway is transient during epithelial differentiation and titration of pathway activity may be central for the proper coordination of differentiation and morphogenesis. We review current evidence on the WNT/β-catenin/TCF/Lef1 signaling pathway in kidney epithelial development and discuss the potential implication of non-canonical WNT signaling and WNT-independent events.
The Wolffian duct sends off a dorsal branch called the ureteric bud to initiate mammalian kidney development. This epithelial structure invades an adjacent blastemal progenitor cell population called the metanephric mesenchyme. The ureteric bud receives signals from the meta-nephric mesenchyme to undergo branching morphogenesis and develop into the ureter and collecting duct system. Reciprocally, the metanephric mesenchyme receives signals from the branching tips of the ureteric bud, which facilitate its survival, cellular proliferation, and condensation into proto-epithelial clusters termed pretubular aggregates, the direct renal epithelial precursors. Additional events mediate proximal–distal patterning and terminal differentiation of these progenitors into all nephron epithelia proximal to the collecting duct.1,2 The process of nephron differentiation from epithelial progenitors is recapitulated in vitro when the metanephric mesenchyme is isolated from rat or mouse embryos and placed in an organ culture in the presence of inductive factors. These inductive factors that were originally derived from the ureteric bud or embryonic spinal cord have now been characterized more closely on a molecular level.2–8 Remarkably, an activation of the WNT (after the Drosophila homolog Wingless and the Int-1 [Wnt1] integration site in virally induced murine breast tumors) signaling pathway appears to be common to all these inducers.2,9,10
Since their discovery in the 1980s, WNT proteins have been found to be involved in a number of cell type-specific processes governing development and homeostasis. WNT proteins belong to a highly conserved family of secreted growth factors that contain roughly 20 members in mammals. Although different members of the WNT family have been identified based on their amino-acid homology, they display a substantial divergence in their biological effects on target cells. Intracellular WNT signaling is classified into ‘canonical’ and ‘non-canonical’ pathways. In this context, these terms refer to whether or not the signaling pathway is β-catenin-dependent (canonical) or β-catenin-independent (non-canonical).
β-Catenin is a multifunctional intracellular protein that is characterized by a core domain containing 12 copies of a 42-amino-acid ‘armadillo’ motif, named after the Drosophila ortholog of β-catenin. These repeats generate a positively charged groove that facilitates β-catenin interaction with several negatively charged ligands, including adenomatous polyposis coli and TCF/Lef1 transcription factors.11 In addition to its intracellular signaling functions, β-catenin constitutes a central component of adherens junctions, where it interacts with cadherins and α-catenin.12 A multiprotein complex called the ‘β-catenin destruction complex’ can capture intracellular β-catenin by default (Figure 1). This complex contains the proteins axin, adenomatous polyposis coli, and glycogen synthase kinase 3β (GSK3β), which phosphorylates β-catenin on several N-terminal serine/threonine residues. Phosphorylated β-catenin is then ubiquitinated and targeted for proteasomal degradation. Therefore, in the absence of WNT signaling, intracellular β-catenin levels are kept low through continuous degradation. Once a canonical WNT ligand binds to a FZD (Frizzled)/LRP (lipoprotein receptor-related protein) receptor complex on the cell surface, the β-catenin destruction complex is inhibited through mechanisms that involve the intracellular protein Dishevelled. Consequently, β-catenin is no longer phosphorylated by GSK3β, escapes degradation, accumulates in the cytoplasm, and is translocated to the nucleus. There β-catenin interacts with transcription factors of the TCF/Lef1 family displacing transcriptional repressors of the TLE/Groucho family. Thereby, β-catenin converts TCF/Lef1 factors from transcriptional repressors into transcriptional activators. By this rather linear transduction cascade, WNT ligands act as extracellular switches that turn on cell type-specific transcriptional programs in their target cells, which in turn mediate the biological effects of canonical WNT signaling.12,13
The non-canonical branches of WNT signaling have been identified more recently.14 Notably, WNT signals have been implicated in the organization of proximal–distal epithelial cell polarity across a sheet or tube, a process termed planar cell polarity (PCP).14,15 Components of the WNT/PCP pathway also include Frizzled receptors and Dishevelled, but they are independent of β-catenin stabilization and TCF/Lef1 signaling. Instead, WNT/PCP signaling involves the GTPases RhoA and Rac1 as well as Jun N-terminal kinase. Different WNT ligands have been reported to elicit differential activity on these signaling branches. For instance, Wnt3a and Wnt1 signal through the β-catenin/TCF/Lef1 pathway in most cell types, whereas Wnt5a and Wnt11 have been associated with non-canonical pathways. However, even for these well-characterized family members, the activity appears to be context- and cell type-dependent.16
At least two WNT ligands, Wnt9b and Wnt4, are intimately involved in the epithelial differentiation of metanephric mesenchymal progenitors. Wnt9b is produced by the ureteric bud and its genetic deletion in mice leads to severe kidney hypoplasia, which is related, at least in part, to an arrest of progenitor cell differentiation in the metanephric mesenchyme.7 In Wnt9b mutant mice, the metanephric mesenchyme is present; however, it fails to undergo the first step of differentiation, namely formation of the pretubular aggregate. As a result, the characteristic marker molecules of this structure, such as Pax8, Fgf8, and Lhx1, are absent.
A second WNT ligand, Wnt4, is placed downstream of Wnt9b signaling. The Wnt4 expression domain is confined to the pretubular aggregate and early-stage epithelia derived subsequently. Similar to Wnt9b mutants, Wnt4 gene-deleted mouse embryos display pronounced kidney hypoplasia and a near-complete arrest of tubulogenesis at the stage of the pretubular aggregate with only few progenitor cells converting into more advanced stages of nephron formation.6,17 Remarkably, a loss-of-function WNT4 mutation has recently been implicated in a severe renal hypoplasia syndrome in humans, indicating that this molecule has a conserved role in mammalian kidney development.18 More generally, this observation is the first piece of evidence supporting the notion that defective WNT signaling in epithelial progenitors participates in human congenital diseases of the kidney.
When metanephric mesenchymes from mouse or rat embryos are isolated from the ureteric bud and placed in organ culture, they undergo rapid apoptosis and fail to convert into epithelia unless they are recombined with either the ureteric bud or surrogate stimuli. Remarkably, WNT ligands (including Wnt4 and Wnt9b), when expressed from co-cultured fibroblasts placed adjacent to the metanephric mesenchyme provide such surrogate stimuli. As a result, the mesenchyme survives and undergoes complete conversion into nephron-like epithelia in their presence.5,7,19 Indeed, these data suggest that Wnt9b fulfills the criteria of the ureteric bud-derived metanephric mesenchyme inducer that was proposed over half a century ago.20 Yet, a note of caution should be made when interpreting the available data. First, the epithelial progenitors in the metanephric mesenchyme remain viable and detectable in Wnt9b mutant mice, as determined by Pax2 and Gdnf staining.7 This state of affairs contrasts to what is observed in organ culture after removal of the ureteric bud. This finding suggests that additional factors emanating from the ureteric bud ensure survival of these progenitors. Also, the data obtained by use of WNT-expressing fibroblasts as a source of the WNT ligands leaves open questions. For instance, to elicit their anti-apoptotic and differentiation-inducing activity WNT ligands must be expressed from co-cultured fibroblasts that are in direct contact with the isolated metanephric mesenchymal progenitors.5 Supernatants from WNT-expressing cells alone are ineffective as are recombinant WNT proteins.21 Furthermore, the type of cell line used for WNT protein expression in co-culture experiments appears to affect the outcome.7 These observations suggest that WNT ligands may require additional co-factors to elicit their full survival and differentiation-inducing activity. Direct overexpression of WNT ligands in the isolated metanephric mesenchyme in a feeder-free system may resolve these issues, but has not been attempted. Despite these reservations, WNT ligands clearly constitute central players in the course of nephron development.
Is β-catenin/TCF/Lef1 signaling necessary for WNT-dependent effects in the metanephric mesenchyme? Conditional β-catenin gene deletion in the metanephric mesenchyme in mice leads to marked renal hypoplasia and a reduction in several direct or indirect target genes, including Pax8, Wnt4, Fgf8, and Lhx1. 22 Nephron numbers are markedly reduced in these mutant kidneys and the few nascent nephrons observed in these animals may form as a result of incomplete β-catenin gene deletion.22 Overall, the phenotype of metanephric mesenchymal β-catenin mutants resembles that of Wnt9b mutants suggesting an involvement of β-catenin downstream of Wnt9b. Is this effect mediated by TCF/Lef1 transcription factors? Although direct in vivo data are missing, rat metanephric mesenchymal cells induced to express a dominant-negative N-terminal truncation mutant of TCF undergo apoptosis and are progressively depleted in the organ culture model of metanephric mesenchymal differentiation.9 Taken together, these data suggest a model in which a WNT/β-catenin/TCF/Lef1 signaling axis in the metanephric mesenchyme is required for the formation of the pretubular aggregate and thus a prerequisite for nephron development.
Is β-catenin/TCF/Lef1 signaling sufficient to trigger epithelial conversion from metanephric mesenchymal progenitors? Lithium chloride and bromoindirubin-3′-oxime, both GSK3β inhibitors, are each sufficient to prevent apoptosis and induce the appearance of nephron-like epithelia expressing polarity markers such as E-cadherin when applied to isolated metanephric mesenchyme. The inhibitors appear to ‘phenocopy’ the effect of WNT-expressing cell lines and other surrogate inducers.9,23,24 This effect is associated with increased levels of cytoplasmic β-catenin and the induction of two putative β-catenin/TCF/Lef1 target genes in the metanephric mesenchyme, namely Tcf7 (often referred to as Tcf-1) and Lef1.23 These data suggest that WNT-induced GSK3β inhibition may be a key mechanism in nephron differentiation and that stabilization of β-catenin may be involved in this activity. In addition, combinations of non-WNT growth factors, including either leukemia inhibitory factor or transforming growth factor-β2, are sufficient to activate nephron differentiation in isolated rat metanephric mesenchyme, which is accompanied by binding of TCF/Lef1 consensus oligos by nuclear extracts of these progenitors.8 Furthermore, leukemia inhibitory factor activates a set of direct β-catenin/TCF/Lef1 target genes in rat metanephric mesenchyme, including Pax8, Emx2, and Ccnd1, which may be secondary to a leukemia inhibitory factor-induced upregulation of Wnt4 expression.9 These data suggest that WNT/β-catenin/TCF/Lef1 signaling may be a common pathway activated by all nephron-inducing growth factors thus far identified.
However, WNT ligands, inhibitors of GSK3β, and inductive growth factors may trigger important intracellular events in epithelial progenitors other than the stabilization of β-catenin. Importantly, the WNT/GSK3β axis has recently been shown to activate intracellular signaling independent of β-catenin stabilization.25 To differentiate β-catenin-dependent from β-catenin-independent effects, genetically stabilized β-catenin mutants are used to facilitate a pure activation of the β-catenin pathway and to omit effects on other GSK3β targets.26 Interestingly, the conditional β-catenin stabilization in metanephric mesenchymal epithelial progenitor cells in vivo leads to the formation of ectopic cell aggregates that express markers of the pretubular aggregate, including Pax8, Wnt4, Fgf8, and Lhx1.22 This effect is also observed on a Wnt9b- or Wnt4-deficient genetic background.22 The finding suggests that β-catenin stabilization is a sufficient mechanism by which these WNT ligands initiate renal epithelial cell lineage progression up to the pretubular aggregate stage. Similarly, when β-catenin is stabilized in the isolated metanephric mesenchyme placed in organ culture, the cells escape apoptosis, proliferate, and initiate pretubular aggregate marker expression.9,22,23 Remarkably, polarized epithelia are observed in neither setting. In fact, E-cadherin expression is completely absent in the cell aggregates induced by stabilization of β-catenin.9,22 Moreover, when β-catenin is stabilized in the metanephric mesenchyme in vivo, the cellular transition to epithelial structures is completely blocked, even in the presumed presence of endogenous ureteric bud-derived inductive signals.22 In addition, the differentiation-inducing effect of GSK-3β inhibitors in the isolated metanephric mesenchyme is only present when these agents are applied transiently or at low dosages, whereas high dosages block differentiation.9,23 Together, these data suggest that β-catenin/TCF/Lef1 signaling may in fact be antagonistic to the epithelial conversion of pretubular aggregates. This model is consistent with the observation that β-catenin/TCF/Lef1 signaling inhibits differentiation features in cultured epithelial cells from other lineages.27,28
These observations may be relevant not only in the setting of normal renal organogenesis, but also during tumorigenesis. Notably, stabilization of β-catenin is frequently observed in Wilms tumor, a common pediatric kidney malignancy, which is thought to be initiated in ‘nephrogenic rests’, that is, islands of persistent blastema in the otherwise fully developed kidney.29 Ten to fifteen percent of Wilms tumors display oncogenic mutations of β-catenin.30 Another one third of Wilms tumors are related to inactivating mutations of the X-linked tumor suppressor gene WTX,31 which normally associates with the β-catenin degradation complex and promotes β-catenin disposal (Figure 1).32 Blastemal regions of Wilms tumors express a series of genes also expressed in the metanephric mesenchyme.2,33 Suggesting that the response of these cells to an activation of β-catenin/TCF/Lef1 signaling may be similar. Thus, in analogy to its effects in the metanephric mesenchyme, β-catenin signaling may contribute to the differentiation arrest in Wilms tumors and—at the same time—promote cellular proliferation.
Interestingly, β-catenin/TCF/Lef1 signaling has also recently been implicated in ureteric bud epithelial differentiation. In contrast to the differentiation of the proximal nephron from mesenchymal progenitors, the ureteric bud is formed by budding from a pre-existing epithelial tube, the Wolffian duct. Although all cells of the ureteric bud have already undergone epithelial conversion, the more immature distal branches express several genes associated with branching morphogenesis, such as Ret and Wnt11,34 but lack several markers of terminal collecting duct differentiation, namely aquaporin 3 or ZO-1α+.35 A conditional β-catenin gene deletion in the ureteric bud causes premature terminal epithelial differentiation even in the distal-most tips, which is associated with a defect of ureteric branching morphogenesis and severe renal hypoplasia/dysplasia.35,36 Furthermore, conditional β-catenin stabilization in the ureteric bud antagonizes terminal differentiation of collecting duct epithelia.35 These data suggest a more general model, where β-catenin/TCF/Lef1 signaling in the renal epithelial lineage is associated with ongoing morphogenesis but inhibitory to terminal differentiation.
These bimodal β-catenin signaling effects are consistent with studies of in vivo TCF/Lef1 signaling activity in mice that express β-galactosidase reporters driven by TCF/Lef1-responsive promoters.9,37,38 Their activity is confined to the immature regions of the ureteric bud and to early-stage metanephric mesenchymal-derived epithelia, but becomes undetectable in the more mature portions of the nephron. Whereas all in vivo TCF/Lef1 reporters appear to display some context-dependent differences in sensitivity and specificity,39 this model is also supported by the expression patterns of β-catenin/TCF/Lef1 target genes, including Pax8, Emx2, and Ccnd1.9,36
One challenge for future studies will be to more thoroughly elucidate the consequences of this temporal and spatial β-catenin/TCF/Lef1 signaling activity window in the course of nephron formation. An intriguing hypothesis is that the appropriate balance of β-catenin/TCF/Lef1 signaling and non-canonical WNT/PCP signaling is critical for terminal epithelial differentiation, as well as proper tubular elongation and nephron patterning. Defects in epithelial differentiation in nephron progenitors that express stabilized β-catenin may possibly be related to a disrupted balance of the two WNT signal transduction branches. This idea is supported by experiments carried out in a recently developed colony-forming assay system of clonally isolated metanephric mesenchymal progenitors that are co-cultured with Wnt4-expressing fibroblasts.21 In these progenitors, inhibition of Jun N-terminal kinase, a late step in PCP signaling, impairs E-cadherin-positive epithelial cell differentiation suggesting a requirement of PCP signaling for terminal epithelial differentiation. Furthermore, in epithelia that have completed terminal differentiation, constitutive β-catenin signaling activation by conditional adenomatous polyposis coli inactivation or overexpression of stabilized β-catenin promotes the development of epithelial cysts. These cysts display characteristic apical–basal polarity defects, which may be indicative of a differentiation defect.40,41 Consistently, several molecules that inhibit cystogenesis, including the ciliary proteins inversin, Bbs4, and Kif3a, have since been identified as negative β-catenin/TCF/Lef1 signaling regulators.42–44 In the case of inversin and Bbs4, a reciprocal modulation of PCP signaling has been demonstrated.42,43 These data point to an intricately coordinated molecular machinery in developing and mature renal epithelia that controls the balance of the two WNT signaling branches. However, a more direct characterization of PCP signaling and β-catenin/TCF/Lef1 signaling and their interaction is warranted to definitively establish their individual contributions.
Although many open questions remain, the fact that WNT/β-catenin signaling is a necessary and sufficient trigger of the early stages of nephron differentiation is now clear. The perspectives for the future will be threefold: (1) to delineate the role of β-catenin at later stages of epithelial development; (2) to elucidate the detailed roles of additional WNT/GSK3β-dependent and non-canonical WNT pathways; and, (3) to dissect WNT-dependent from WNT-independent mechanisms by additional methods.
We are grateful to Friedrich C. Luft for the insightful comments concerning the manuscript. Kai M. Schmidt-Ott is an Emmy Noether Fellow of the Deutsche Forschungsgemeinschaft, Germany (SCHM 1730/2-1). Jonathan Barasch is supported by grants from the March of Dimes, the Emerald Foundation, and the National Institute of Diabetes and Digestive and Kidney Diseases (Grants DK-55388 and DK-58872).
All the authors declared no competing interests.