GLI3 Repressor Controls Nephron Number via Regulation of Wnt11 and Ret in Ureteric Tip Cells

Truncating GLI3 mutations in Pallister-Hall Syndrome with renal malformation suggests a requirement for Hedgehog signaling during renal development. HH-dependent signaling increases levels of GLI transcriptional activators and decreases processing of GLI3 to a shorter transcriptional repressor. Previously, we showed that Shh-deficiency interrupts early inductive events during renal development in a manner dependent on GLI3 repressor. Here we identify a novel function for GLI3 repressor in controlling nephron number. During renal morphogenesis, HH signaling activity, assayed by expression of Ptc1-lacZ, is localized to ureteric cells of the medulla, but is undetectable in the cortex. Targeted inactivation of Smo, the HH effector, in the ureteric cell lineage causes no detectable abnormality in renal morphogenesis. The functional significance of absent HH signaling activity in cortical ureteric cells was determined by targeted deletion of Ptc1, the SMO inhibitor, in the ureteric cell lineage. Ptc1−/−UB mice demonstrate ectopic Ptc1-lacZ expression in ureteric branch tips and renal hypoplasia characterized by reduced kidney size and a paucity of mature and intermediate nephrogenic structures. Ureteric tip cells are remarkable for abnormal morphology and impaired expression of Ret and Wnt11, markers of tip cell differentiation. A finding of renal hypoplasia in Gli3 −/− mice suggests a pathogenic role for reduced GLI3 repressor in the Ptc1−/−UB mice. Indeed, constitutive expression of GLI3 repressor via the Gli3Δ699 allele in Ptc1−/−UB mice restores the normal pattern of HH signaling, and expression of Ret and Wnt11 and rescued the renal phenotype. Thus, GLI3 repressor controls nephron number by regulating ureteric tip cell expression of Wnt11 and Ret.


Introduction
Development of the permanent mammalian kidney (the metanephros) is dependent on growth and branching of the ureteric bud and its daughter branches, a process termed renal branching morphogenesis. At the onset of this process, the ureteric bud elongates towards and invades the metanephric mesenchyme before undergoing spatial specification into 'ureteric stalk' and 'ureteric tip' domains. Reciprocal inductive interactions between the ureteric tip and surrounding metanephric mesenchyme results in division of the ureteric tip, forming the first of a series of ureteric branches, which ultimately constitute the mature collecting duct system. Simultaneously, each ureteric bud tip induces adjacent metanephric mesenchyme cells to undergo a mesenchymeepithelial transformation and form the epithelial components extending from the glomerulus to the distal tubule, a process known as nephrogenesis. The number of nephrons formed is directly related to the number of ureteric branches and their inductive capacity. Severe reductions in nephron number, characteristic of renal hypoplasia/dysplasia, are the leading cause of childhood renal failure. More subtle defects in nephron number have been associated with the development of adult-onset essential hypertension and chronic renal failure [1,2,3,4,5].
The Hedgehog-GLI signaling pathway plays critical roles during mammalian kidney development. GLI proteins function downstream of Hedgehog's (HH), extracellular proteins, and Patched (PTC) and Smoothened (SMO), cell surface proteins. HH ligand signals upon binding to its receptor, PTC, relieving PTC-mediated inhibition of SMO, a transmembrane protein. In this state, SMO interacts with a molecular complex including Costal-2 (Cos2), Fused (Fu) and Suppressor of Fused (SuFu), ultimately resulting in translocation of GLI protein family members into the nucleus where they act as transcriptional activators. In the absence of HH ligand, PTC inhibits SMO and prevents its interaction with the Cos2-Fu-SuFu complex resulting in C-terminal cleavage of GLI protein, which translocates to the nucleus and acts as a transcriptional repressor. In vertebrates, three GLI family members, GLI1, GLI2 and GLI3, mediate HH signals. During murine embryogenesis, GLI1 and GLI2 are believed to function primarily as transcriptional activators while GLI3 is believed to function primarily as a transcriptional repressor [6,7].
The balance of GLI activator and repressor activities is critical during renal morphogenesis. Mutations that are predicted to generate a truncated protein similar in size to GLI3 repressor are observed in humans with Pallister-Hall Syndrome (PHS) and renal dysplasia [8]. The pathogenic role of constitutive GLI3 repressor activity during renal morphogenesis is further demonstrated by the renal dysplastic phenotype in mice engineered to express GLI3 repressor in a dominant manner [9] and in Shh-deficient mice [10]. Dysplastic kidney tissue in Shh-deficient mice is characterized by sustained GLI3 repressor expression in the face of decreased levels of GLI activators (GLI1, GLI2 and full-length GLI3), resulting in a shift in the balance of GLI activators and GLI repressors in favor of repressor [10]. Remarkably, genetic elimination of Gli3 in the Shh null background restores expression of GLI activators and normalizes renal morphogenesis [10]. The expression of Shh in ureteric cells suggests that it may control renal development via direct effects in the ureteric cell lineage. While conditional inactivation of Shh in ureteric cells results in renal hypoplasia, characterized by reduced kidney size and glomerular number [11], the dependency of this pathogenic phenotype on Shh signaling in ureteric cells is unknown.
Here we define the specific function of HH signaling in the ureteric cell lineage during murine kidney development, in genetic models of deficient or constitutively active signaling. HH signaling activity is specifically restricted to the ureteric cells of the medulla and ureter but is absent from the ureteric cell tips of the renal cortex. Genetic inactivation of Smo in the ureteric cell lineage exerted no deleterious effects on renal morphogenesis. In contrast, genetic inactivation of Ptc1 in the ureteric cell lineage caused ectopic HH signaling activity in ureteric tip cells, impaired ureteric tip cell-specific gene expression and renal hypoplasia. Genetic inactivation of Gli3 alone, the primary GLI repressor, resulted in a similar phenotype suggesting a critical role for GLI3 repressor. Indeed, introduction of a constitutively active GLI3 repressor in a Ptc1-deficienct background normalized the renal phenotype, restored the normal domain of HH signaling activity and rescued expression of genes specific to ureteric tip cells and required for their functions. We propose a model in which SHH-SMO signaling controls the spatial generation of GLI3 repressor, which is required in the cortical ureteric cells for ureteric tip cell-specific gene expression and cell function.

Spatial Restriction of HH Signaling Activity to the Renal Medulla and Ureter
To begin to further investigate the role of SHH signaling during renal morphogenesis, we examined the expression of Ptc1 utilizing the Ptc1-lacZ reporter mouse [12]. Since Ptc1 is a downstream target of HH signaling, Ptc1-lacZ expression is indicative of the site of HH signaling activity [12,13,14]. In the WT (Ptc1 lacZ/+ ) kidney at E13.5, Ptc1-lacZ is strongly localized to cells surrounding the presumptive ureter and the presumptive medullary stroma ( Figure 1A,B), consistent with the pattern of Ptc1 mRNA expression [11]. Ptc1-lacZ is also weakly localized to the epithelium of the presumptive ureter and the distal or medullary collecting ducts ( Figure 1A-C). Interestingly, Ptc1-lacZ expression is not observed in any structures of the presumptive renal cortex, suggesting that HH signaling activity is restricted to the ureter and medullary regions of the developing kidney ( Figure 1B,D). At a later stage (E18.5) of kidney development, a similar pattern of expression is maintained in the cells surrounding the ureter and medullary stroma (Figure S1A-C). However, at E18.5, Ptc1-lacZ expression is not observed in any epithelial structures (Figure S1A-C). Taken together, Ptc1-lacZ expression in both ureteric and metanephric mesenchymederived structures suggests a role for SHH function in both the ureteric bud and metanephric mesenchyme lineages of the early developing kidney but only in the presumptive ureter and medullary regions.

SHH-SMO-Dependent Signaling is not Required in the Ureteric Cell Lineage
We began to investigate the possible autocrine functions of SHH-SMO-dependent signaling by generating a loss of function model for HH signaling in the ureteric cell lineage.
Smoothened is required for the transduction of all HH signals. Similar to Shh inactivation, inhibition of SMO with a steroidal alkaloid, cyclopamine, results in sustained GLI3 repressor in the absence of GLI activators [10]. Homozygous germline deficiency of Smo in the mouse results in embryonic lethality prior to the commencement of metanephric development [15]. Therefore, we utilized Hoxb7creEGFP mice to generate a murine model in which Smo is genetically inactivated in the ureteric cell lineage [16,17], thereby eliminating SHH-SMO-dependant signaling.
Targeted deficiency of Smo in the ureteric cell lineage did not adversely effect survival since mutants were recovered in the near expected Mendelian ratios (  Figure 1E). Consistent with a loss of Smo, analyses of Ptc1-lacZ in Smo 2/2UB kidneys at E13.5 revealed a marked reduction in HH signaling activity in the ureteric cells of the ureter and medullary collecting ducts ( Figure 1F,G). Examination of the gross anatomical and histological features of newborn Smo 2/2UB kidneys revealed no major differences compared to WT kidneys ( Figure 1H,I,L,M). Consistent with this observation, a more detailed analysis of nephrogenic and ureteric structures using immunofluorescence microscopy and RNA in situ hybridization revealed that glomerulogenesis, nephrogenesis, nephron segmentation, ureteric branching morphogenesis and ureteric tip cell-specific gene expression was normal in Smo 2/2UB kidneys ( Figure S2). In order to determine if HH-SMO-dependant signaling is required in the mature kidney, we also examined Smo 2/2UB kidneys at PN30. No histological abnormalities were detected in the Smo 2/2UB kidneys ( Figure 1J,K,N,O). Taken together, these results demonstrate that HH-SMO-dependent signaling is not required in the ureteric cell lineage and suggest that SHH has no autocrine function [in the ureteric cells] during renal morphogenesis.

A Model of Cortical HH Signaling Activity in the Embryonic Kidney
Our results demonstrate that HH signaling activity is absent from the renal cortex. We determined the importance of this spatial restriction of HH signaling activity by generating a gain-offunction model in which HH signaling is ectopically activated in the cortical ureteric epithelium.
Patched1 is a negative regulator of the HH signaling pathway. In the absence of Ptc1, repression of HH target genes is alleviated, even in the absence of SHH ligand. Homozygous germline deficiency of Ptc1 in the mouse results in embryonic lethality prior to the commencement of metanephric development [12]. Therefore, we utilized Hoxb7creEGFP mice to generate a murine model in which Ptc1 is genetically inactivated in the ureteric cell lineage [16,17] (refer to Methods). To confirm that inactivation of Ptc1 results in a constitutively active HH signaling pathway in the ureteric cell lineage, we analyzed Ptc1-lacZ expression in Ptc1 2/2UB kidneys. While Ptc1-lacZ expression is maintained in the cells surrounding the presumptive ureter and presumptive medullary stroma (Figure 2A,B), Ptc1-lacZ is markedly upregulated in the epithelium of the presumptive ureter and distal collecting ducts in Ptc1 2/2UB kidneys ( Figure 1C vs. Figure 2C). Remarkably, Ptc1-lacZ is ectopically expressed throughout the cortical or proximal collecting ducts and in the ureteric bud tips (Figure 2A,B,D), albeit in a mosaic pattern ( Figure S1F). A similar pattern of upregulated and ectopic ureteric bud epithelial Ptc1-lacZ expression was also observed at a later stage (E18.5) of kidney development ( Figure  S1D-F). To confirm that Ptc1 inactivation occurs by an early stage in renal morphogenesis, we assayed expression of Ptc1 mRNA, a surrogate marker of HH signaling activity, in ureteric bud tissue isolated from E11.5 kidney. Quantitative real-time PCR using primers designed for an undisrupted region of the mutant Ptc1 transcript demonstrated a 50-fold increase in Ptc1 mRNA transcripts in Ptc1 2/2UB compared to WT ureteric cells (WT vs. Ptc1 2/2UB : 2.2260.02 vs. 108.5166.47, p,0.001) ( Figure S3H). Together, these results show that genetic elimination of PTC1 in the ureteric cell lineage leads to increased and ectopic HH signaling activity in the developing kidney.  Ectopic HH Signaling Activity in the Proximal Epithelium Causes Renal Hypoplasia Viable neonatal PTC1-deficient pups could not be recovered (Table S2). However, live Ptc1 2/2UB mutant embryos were recovered in expected Mendelian ratios at all embryonic time points analyzed (Table S2). Macroscopic analyses of Ptc1 2/2UB embryos revealed severe exencephaly, with 100% penetrance. Otherwise, Ptc1 2/2UB embryos were indistinguishable from WT littermates ( Figure S3). It is therefore likely that Ptc1 2/2UB embryos reach term and die and/or are cannibalized immediately following birth.
Analysis of Ptc1 2/2UB kidneys at E18.5 revealed renal hypoplasia, characterized by a 45% reduction in kidney volume (WT vs.  Figure 2O,P). However, no difference in renal histology or the number of nephrogenic intermediate structures was observed between WT and Ptc1 2/2UB kidneys at E13.5 suggesting that nephrogenesis is unaffected at that stage ( Figure S4). Together, these results demonstrate that PTC1-deficiency leads to deficient nephrogenesis.
Ureteric branching is required for nephron formation. Since ureteric branch tips induce contiguous metanephric mesenchyme cells to engage in nephron formation, the number of ureteric branches is considered to be a critical determinant of the number of nephrons generated. To determine the effect of Ptc1-deficiency on early ureteric branching morphogenesis we quantified ureteric branching at E12.5 by whole mount immunofluorescence using Calbindin-D 28K , a marker of ureteric bud epithelium. No significant difference in the number of ureteric branches was observed between WT and Ptc1 2/2UB kidneys ( Figure 2Q,R and Figure S4F). However, we did observe abnormalities in the ureteric branch pattern. In contrast to WT kidneys, several irregularly shaped and dilated ureteric bud tips were observed in Ptc1 2/2UB kidneys ( Figure 2Q,R). Analyses of ureteric branching morphogenesis a day later at E13.5 revealed a mild reduction in ureteric branch number in Ptc1 2/2UB kidneys compared to controls ( Figure 2S,T). At E15.5, this reduction was more even more pronounced ( Figure 2U,V). Importantly, normal renal architecture is maintained in Ptc1 2/2UB kidneys. Taken together, these results demonstrate that ectopic HH signaling activity in proximal ureteric cells causes renal hypoplasia, characterized by marked defects in the formation of nephrogenic structures, reduced ureteric branching and a qualitative defect in early stage ureteric branch tips.

Effect of Ptc1-Deficiency on Proliferation and Apoptosis
Cell proliferation is crucial for ureteric branching morphogenesis and nephrogenesis. Since SHH signaling controls renal cell proliferation [10] we investigated the possible contributions of abnormal cell proliferation to the hypoplastic phenotype in PTC1deficient kidneys at E13.5, a time point prior to the onset of the phenotype. Analysis of cell proliferation, using an in situ BrdU incorporation assay, revealed a 40% decrease in ureteric bud cell proliferation in Ptc1 2/2UB kidneys at E13.5 (WT vs. Ptc1 2/2UB : 38.465.6 vs. 23.169.0, p = 0.05) ( Figure 3A-C). In contrast, no significant difference in mesenchymal cell proliferation was detected (WT vs. Ptc1 2/2UB : 2.6860.7 vs. 2.0560.6, p = 0.25) ( Figure 3A,B,D).
Mesenchyme cell survival is critical to nephrogenesis. Moreover, elevated ureteric bud apoptosis has been implicated in the pathogenesis of renal hypoplasia [18]. Accordingly, we determined whether cell survival was impaired in PTC1-deficient kidneys at E13.5 ( Figure S5). Analysis of apoptosis, using a TUNEL assay, revealed no significant difference in ureteric bud or metanephric mesenchyme cell death in Ptc1 2/2UB kidneys.
Taken together, these results suggest that enhanced and ectopic HH signaling activity in the ureteric cell lineage decreases ureteric cell proliferation but does not affect apoptosis.

Ectopic HH Signaling Activity in the Proximal Ureteric Epithelium Impairs Expression of Wnt11 and Ret in Ureteric Bud Tip Cells
Our findings of renal hypoplasia, abnormal ureteric tip morphology and decreased ureteric cell proliferation (in Ptc1 2/2UB kidneys) suggested that ectopic HH signaling activity in ureteric tips disrupts normal ureteric tip function.
Specification of ureteric bud tip cells distinct from ureteric stalk cells is essential for ureteric branching morphogenesis and nephrogenesis. Determination of ureteric tip cell fate is dependant on Gdnf/Ret signaling [19]. Furthermore, Gdnf/Ret signaling is required for the maintenance of Wnt11 expression, also in the ureteric bud tips cells [20,21]. Conversely, Wnt11 promotes Gdnf expression in the surrounding metanephric mesenchyme suggesting that Gdnf, Ret and Wnt11 participate in an autoregulatory feedback loop to regulate ureteric branching morphogenesis [20]. We examined the effect of ectopic HH signaling activity on ureteric tip cell-specific gene expression using in situ hybridization. In contrast to WT kidneys, expression of Ret and Wnt11 was markedly reduced in the majority of ureteric bud tips in Ptc1 2/2UB kidneys at E13.5 ( Figure 3E-H). Consistent with the autoregulatory feedback loop, expression of Gdnf in the surrounding metanephric mesenchyme was also markedly reduced in the Ptc1 2/2UB kidneys ( Figure 3I,J). The specificity of decreased Gdnf expression in Ptc1 2/2UB kidneys is demonstrated by normal expression of Osr1, Six2, and CITED1, each of which marks mesenchymal precursor cells [22,23,24,25,26]; Wnt4, a marker of pretubular aggregates [27]; and Wnt9b which is required for the earliest inductive response in metanephric mesenchyme and acts upstream of Wnt4 [28] ( Figure S6). To further investigate the ontogeny of abnormal ureteric tip cell gene expression in Ptc1 2/2UB mice, we assayed Ret and Wnt11 expression in isolated ureteric bud tissue using quantitative real-time PCR. At E11.5, a stage that immediately follows induction of the metanephric mesenchyme by the ureteric bud, expression of Ret was comparable between WT and Ptc1 2/2UB ureteric cells (WT vs. Ptc1 2/2UB : 1.6560.37 vs. 1.5060.61, p = 0.78) ( Figure 3M). In contrast, Wnt11 expression was reduced by ,70% in Ptc1 2/2UB ureteric cells (WT vs. Ptc1 2/2UB : 1.7960.26 vs. 0.5160.27, p,0.05) ( Figure 3N). Taken together, these results demonstrate that ectopic HH signaling activity in the proximal ureteric epithelium specifically impairs the expression of Wnt11 and Ret in ureteric tip cells.
Our results indicate that HH signaling activity is normally restricted to the distal ureteric cells of the ureter and medulla.
Given the impairment of tip cell gene expression, we investigated the possibility that HH signaling activity biases ureteric cells towards a distal cell fate. To determine whether Ptc1 2/2UB tips cells had adopted characteristics of the ureteric stalk we performed DBA-lectin whole mount immunofluorescence microscopy on WT and Ptc1 2/2UB kidneys at E13.5. DBA is a marker of the ureteric stalk but not the ureteric tip [29]. DBA expression was observed in ureteric tips in Ptc1 2/2UB kidneys in a mosaic pattern but rarely in WT kidneys ( Figure 3K,L and Figure S5E-H). The expression of Wnt7b, another marker of ureteric stalk that is absent from the ureteric tips [30], was comparable between WT and Ptc1 2/2UB kidneys. Since ectopic HH signaling activity is also observed in proximal collecting ducts in mutant kidneys, we next investigated the possibility that HH signaling activity biases cortical collecting ducts towards a more medullary/distal cell fate. Expression of uroplakin III, a marker of the transitional epithelium of the ureter and renal pelvis, was unchanged in the Ptc1 2/2UB kidneys ( Figure  S7A-F). Similarly, the localization of aSMA, a marker of the smooth muscle population surrounding the ureter, was also unaltered ( Figure S7G-L). Together, these results suggest that ectopic HH signaling activity in the proximal epithelium does not induce a distal ureteric bud cell fate. Furthermore, increased HH signaling activity in the distal epithelium (ureter and distal collecting ducts) has no deleterious effects on these structures.

Reduced Levels of GLI Repressor in the Renal Cortex Result in Renal Hypoplasia
The spatial restriction of HH signaling activity from the renal cortex during normal renal morphogenesis suggests that the cortex is a GLI repressor-dominant domain. We hypothesized that the deleterious effects of ectopic HH signaling in the proximal ureteric cells in Ptc1 2/2UB kidneys is due to a reduction in local GLI repressor levels. We addressed this hypothesis by analyzing kidneys in Gli3-deficient embryos.
GLI3 is the primary source of GLI repressor in mammalian cells [31]. Mice with homozygous deficiency in Gli3 die soon after birth or in late gestation [32]. We were able to recover viable Gli3deficient embryos as late as E18.5. Gli3-deficient embryos exhibited polysyndactyly and occasional exencephaly but were similar in size to WT littermates (data not shown). Gli3 +/XtJ mice kidneys were indistinguishable from WT littermates (data not shown). Analysis of Gli3 XtJ/XtJ kidneys at E18.5 revealed mild hypoplasia, characterized by a 15% decrease in kidney volume (p,0.05) and 15% reduction in glomerular number (p,0.05) compared to WT littermates ( Figure 4). Otherwise, glomerulogenesis, nephron segmentation, and smooth muscle and urothelium differentiation was normal in Gli3 XtJ/XtJ kidneys ( Figure S8). These results are consistent with a functional role for GLI repressor in the renal cortex during renal morphogenesis.

Discussion
Disruption of renal development in humans with Pallister-Hall Syndrome and truncating GLI3 mutations [33] and mice with elevated levels of GLI3 repressor [9,10] provides compelling evidence in favor of a critical role for GLI3-dependent signaling during mesenchymal-epithelial interactions during early stages of  Here, we demonstrate that domains of GLI-dependent activator and repressor function are spatially patterned during renal morphogenesis. We investigated the functional significance of these domains in the ureteric cell lineage using genetic murine models of deficient or constitutively active HH signaling. Smodeficiency targeted to the ureteric cell lineage does not disrupt kidney development, demonstrating HH-dependent GLI activators are not required for ureteric cell function. The absence of HH signaling in the renal cortex of WT mice suggests that the cortex is a zone of low GLI activator and high GLI repressor levels. We determined the importance of exclusion of HH signaling activity from the cortical collecting ducts in mice with Ptc1-deficiency targeted to the UB lineage. Absence of ureteric cell Ptc1, a negative regulator of the HH signaling pathway, results in ectopic HH signaling activity in the cortical collecting ducts and ureteric bud tips. Ectopic HH signaling activity in the ureteric bud tips, which under normal circumstances is a domain of GLI repressor function, leads to decreased expression of Ret and Wnt11. These changes result in disruption of ureteric branching morphogenesis and nephrogenesis resulting in renal hypoplasia, likely due to impaired tip function. Remarkably, constitutive GLI3 repressor expression in the Ptc1 2/2UB background, restores ureteric tip cellspecific gene expression and normalizes renal morphogenesis, demonstrating a spatial requirement for GLI3 repressor in the proximal ureteric cells. Together, these results demonstrate a requirement for GLI3 repressor-dependent regulation of nephron number via ureteric tip cell Wnt11and Ret-dependent functions.

Regulation of Ureteric Tip Cell Differentiation by GLI3 Repressor
We have established that loss of GLI3 repressor impairs ureteric tip cells by reducing Ret and Wnt11 expression. The precise mechanism by which GLI-dependent signaling may control Ret and Wnt11 is unclear. Decreased expression of Wnt11 precedes a decrease in Ret expression in the ureteric bud tips of Ptc1 2/2UB kidneys. Wnt11 maintains Gdnf expression in the mesenchyme [20]. Thus, it is probable that reduced Gdnf expression in Ptc1 2/2UB mice is secondary to decreased Wnt11 expression in ureteric tip cells. Consistent with this, Wnt11 null mice exhibit mild renal hypoplasia and a reduction in Gdnf expression [20] almost identical to that observed in Ptc1 2/2UB mutants.
The mosaic expression of Ptc1-lacZ expression in the ureteric tips suggests that not all tip cells were exposed to ectopic HH signaling. This explains, in part, why Ret and Wnt11 expression is not completely abolished from the ureteric tips and why expression levels are variable among tips within the same kidney. It is tempting to speculate that the extent of 'chimerism' of a ureteric tip may influence the efficiency of its function. The ability of 'mutant' tips to confer some function in Ptc1 2/2UB kidneys could account for the mild phenotype since complete abolishment of Gdnf/Ret signaling would result in renal agenesis [34,35,36]. Impairment of tip cell gene expression has also been reported in mice with deletion of bcatenin targeted to the ureteric cells [37]. Consistent with our findings, b-catenin-deficient mice demonstrate reduced expression of Ret, Wnt11 and Gdnf and dilated ureteric tips. However, while early ureteric branching is normal in Ptc1 2/2UB kidneys, it is arrested in b-catenin mutants resulting in severe renal dysplasia. Thus, it is likely that some residual inducing function is maintained in the ureteric tips in Ptc1 2/2UB mice during early stages of kidney development. However, the participation of Gdnf, Ret and Wnt11 in an autoregulatory feedback loop to regulate ureteric branching morphogenesis is likely to exacerbate the early decrement in Wnt11, Ret, and Gdnf contributing to the progressively more severe branching phenotype we observed in Ptc1 2/2UB mice.
Gli2 and Gli3 can serve both redundant activator and repressor functions in a number of mammalian tissues [31,38,39]. Therefore, it is not surprising that the effect of Gli3-deficiency alone is less severe than that observed in conditional deletion of Ptc1. Constitutive activity of the HH signaling pathway, such as that in Ptc1 2/2UB kidneys, is expected to inhibit the formation of both Gli2 and Gli3 repressor isoforms. However, the capacity for the formation of Gli2 repressor isoforms remains intact in Gli3deficient kidneys allowing for partial redundancy. Early lethality prior to the onset of metanephric development in the majority of Gli2;Gli3 compound null mutants [31] limits the ability to investigate this functional redundancy in greater detail.
HH signaling is a known regulator of proliferation in mammalian tissues. Indeed, mutations in Ptc1 are associated with increased incidence of tumorigenesis [12]. In contrast, we observed a moderate decreased in ureteric cell proliferation in Ptc1-deficient kidneys. Gdnf/Ret signaling has been shown to activate intracellular signaling pathways including the ERK MAP kinase pathway [40]. In the kidney, activation of this pathway leads to cellular events, including cell proliferation, cell survival and migration, and inhibition of this pathway results in reduced ureteric branching [41]. Therefore, reduced Ret and Gdnf expression in Ptc1 2/2UB kidneys could result in the observed reduction in proliferation. Consistent with this, the absence of a phenotype in Smo 2/2UB kidneys suggests HH signaling is not required for ureteric cell proliferation.
Inactivating mutations in Gli3 and Ptc1 have been identified in humans with Greig Cephalopolysyndactyly Syndrome (GCPS) and Nevoid Basal Cell Carcinoma Syndrome (NBCCS, also known as Gorlin Syndrome) respectively [32,42]. Pertinent to the phenotype, mutations would lead to loss of GLI3 repressor and PTC1-deficiency respectively. Yet, currently no known association between human Gli3 inactivation or Ptc1 mutations and renal malformation exists. This is likely due to a number of reasons. Firstly, analysis of kidneys in patients with GCPS or NBCCS has not been performed. Our results provide a basis for analyses of kidney size in affected individuals. Secondly, the degree of haploinsufficiency in humans with Ptc1 mutations may be insufficient to result in a renal phenotype. Ptc1 heterozygous mutant mice exhibit phenotypes similar to those observed in humans with BNCCS [12], yet we did not observe any renal abnormalities in these mice. Furthermore, Ptc1 homozygous null mice are embryonic lethal suggesting that Ptc1 homozygous null mutations in humans are also incompatible with life [12]. Further, mutational analysis in humans exhibiting sporadic renal hypoplasia for genes involved in HH signaling, including Gli3 and Ptc1, is also warranted, as done for other genes [43,44].

SHH Does Not Signal in a Autocrine Manner During Renal Morphogenesis
We show that HH signaling activity is not required in the ureteric bud lineage for normal renal morphogenesis. This is consistent with previous results demonstrating metanephric mesenchyme as the primary target of SHH signaling [11]. Analysis of mice deficient in Shh the ureteric bud lineage revealed that Shh secreted by the epithelium of the ureter and distal collecting ducts acts on the surrounding mesenchyme to promote cell proliferation and regulate the timing and patterning of smooth muscle progenitor differentiation [11]. Since we have genetically eliminated Smo in the ureteric bud only, Shh is still capable of paracrine signaling, acting on the surrounding stroma and mesenchyme. In addition, recent in vitro and in vivo data in the pancreas, has suggested additional, non-canonical, mechanisms of GLI activation, downstream of SMO, via two HH-unrelated pathways, RAS and TGFb [45,46,47]. Whether non-canonical GLI activation occurs during renal morphogenesis remains unknown.

A Model of Spatial GLI Activator and Repressor Functional Domains
We propose a model whereby distinct SHH-dependent medullary GLI activator domain and cortical repressor domain functions are critical for normal renal ureteric patterning and function ( Figure 6). It is likely these domains are established by a SHH gradient, emitted by the ureteric cells of the medullary collecting ducts. While SHH signaling is not required in the medullary collecting ducts themselves, the absence of signal in cortical ureteric cells, contributed to by diminishing SHH concentration and/or pathway inhibitors, is critical for ureteric tip cell gene expression required for ureteric branching and nephron induction. Identification of GLI3 repressor gene targets will provide novel insights into this novel mechanism of ureteric tip cell regulation and function. The presence of renal agenesis/ dysplasia in humans with Pallister-Hall syndrome and GLI3 repressor dominant murine models [9,10] suggests that a fine spatial and lineage-specific balance of GLI activator and GLI repressor must be maintained for normal renal morphogenesis. Furthermore, these results implicate misregulation of HH signaling as a possible underlying mechanism in unexplained human renal dysplasia.

Ethics Statement
Experiments using mice were approved in advance by the Animal Ethics Committee at The Hospital for Sick Children and were carried out in accordance with the 'Canadian Council of Animal Care.' Mice Embryonic day 0.5 (E0.5) was considered to be noon on the day of the plug. Littermates were used for all experiments in which normal and mutant embryos were compared.

b-galactosidase staining
Whole kidneys were briefly fixed in lacZ fix solution (25% gluteraldehyde, 100 nM EGTA, 1 M MgCl 2 , 0.1 M sodium phosphate) and rinsed in wash buffer (0.1 M sodium phosphate buffer, 2% nonidet-P40), 1 M MgCl 2 ). Kidneys were then placed in lacZ staining solution (25 mg/ml X-gal, potassium ferrocyanide, potassium ferricyanide) at 37uC overnight in the dark. Once staining had occurred the reaction was terminated in wash buffer and post-fixed in 10% buffered formalin at 4uC. Whole kidneys were photographed using a Lieca EZ4D dissecting microscope and processed for embedding in paraffin wax and sectioned at 5 mm. Sections were counterstained with eosin or nuclear fast red.

Histology and immunohistochemistry
Paraffin-embedded kidney sections were analyzed by histology after generating 4 mm tissue sections and staining with haematoxylin and eosin. Immunofluorescence was performed on formalinfixed, paraffin-embedded kidney sections using anti-Pax2 (Covance

Ureteric bud isolation and real-time reverse transcriptase-PCR
Kidneys from E11.5 WT, Smo 2/2UB and Ptc1 lacZ/2UB mice were dissected and incubated in DMEM/Hams-F12 culture media (GIBCO) containing 10% FCS and 0.2% Collagenase-B (Roche Diagnostics) for 10 min at 37uC. Kidneys were washed in ice-cold media containing 10% FCS and ureteric buds and metanephric mesenchyme were isolated by microdissection using 30 g needles. Ureteric buds were stored in RNAlater RNA stabilization reagent (Qiagen) and RNA was then isolated using the RNAqueous-Micro RNA Isolation kit (Ambion Inc.). cDNA was generated using First Strand cDNA Synthesis (Invitrogen) from total RNA. Real-time PCR reaction mix contained 1 ng of each cDNA sample, SYBR green PCR Mix (Applied Biosystems) and 300 nM of each primer to a total volume of 25 ml. Primers for Smo (Exon 1), Ptc1 (Exon 3), Ret and Wnt11 were designed using Primer 3 software and verified using the UCSC genome bioinformatics website (genome.ucsc.edu). Real-time PCR Amplification was performed using the Applied Biosystems 7900 HT fast RT-PCR system. Relative levels of mRNA expression were determined using the standard curve method. Individual expression values were normalized by comparison to b-2 Microglobulin.

Calculation of kidney volume and the number of glomeruli
Kidney volume and glomerular number was measured according to Bertram et al. [54] with the following modifications: kidneys embedded in paraffin were exhaustively sectioned at 5 mm, collected at 100 mm intervals and stained with Haematoxylin and Eosin. The area of the tissue section was measured with AxioVision 4.6.3-SP1 (Zeiss) and multiplied by the section thickness. Total kidney volume is the sum of volumes for each section. Glomeruli were identified by the presence of both a podocyte layer and Bowman's capsule.

Data analysis
Statistical analysis was performed using GraphPad Prism software (Version 3.01) (GraphPad Software Inc., San Diego, CA). Data were analyzed using a Student's t-test (two tailed). A probability of less than 0.05 was considered to indicate statistical significance. Values are given as means6SD or SEM. Figure S1 HH signaling activity in developing murine kidney. Ptc1-lacZ expression and thereby HH signaling activity at E18.5.

Supporting Information
(A-C) In WT kidneys, Ptc1-lacZ is strongly localized cells surrounding the ureter (not shown) and the medullary stroma (ms). No Ptc1-lacZ activity is observed in the distal collecting ducts (dc) or any structures of the renal cortex. (D-F) In Ptc1 2/2UB kidneys, in addition to strong localization of Ptc1-lacZ to the cells surrounding the ureter and the medullary stroma, Ptc1-lacZ is ectopically expressed in the epithelium of the distal collecting ducts, proximal collecting ducts (pc) and in a mosaic pattern in the ureteric bud tips (tip). n = nephrogenic structure.