Activated Hedgehog-GLI Signaling Causes Congenital Ureteropelvic Junction Obstruction

BASIC RESEARCH www.jasn.org

Sepideh Sheybani-Deloui,1,2 Lijun Chi,1 Marian V. Staite,1,2 Jason E. Cain,1 Brian J. Nieman,3,4,5,6 R. Mark Henkelman,4,5 Brandon J. Wainwright,7 S. Steven Potter,8 Darius J. Bagli,1,2,9 Armando J. Lorenzo,9 and Norman D. Rosenblum1,2,10,11

1Program in Developmental and Stem Cell Biology, 10Division of Nephrology, 3Program in Physiology and Experimental Medicine, and 9Division of Urology, The Hospital for Sick Children, Toronto, Ontario, Canada; 2Departments of Physiology, 4Medical Biophysics and Medical Imaging, and 11Paediatrics, University of Toronto, Toronto, Ontario, Canada; 5Mouse Imaging Centre, Toronto Centre for Phenogenomics Toronto, Ontario, Canada; 6Ontario Institute for Cancer Research, Toronto, Ontario, Canada; 7Genomics of Development and Disease Division, Institute for Molecular Bioscience, University of Queensland, Brisbane, Queensland, Australia; and 8Department of Pediatrics, Cincinnati Children’s Hospital, Cincinnati, Ohio

ABSTRACT Intrinsic ureteropelvic junction obstruction is the most common cause of congenital hydronephrosis, yet the underlying pathogenesis is undefined. Hedgehog proteins control morphogenesis by promoting GLI- dependent transcriptional activation and inhibiting the formation of the GLI3 transcriptional repressor. Hedgehog regulates differentiation and proliferation of ureteric smooth muscle progenitor cells during murine kidney-ureter development. Histopathologic findings of smooth muscle cell hypertrophy and stroma-like cells, consistently observed in obstructing tissue at the time of surgical correction, suggest that Hedgehog signaling is abnormally regulated during the genesis of congenital intrinsic ureteropelvic junction obstruction. Here, we demonstrate that constitutively active Hedgehog signaling in murine intermediate mesoderm–derived renal progenitors results in hydronephrosis and failure to develop a patent pelvic-ureteric junction. Tissue obstructing the ureteropelvic junction was marked as early as E13.5 by an ectopic population of cells expressing Ptch2, a Hedgehog signaling target. Constitutive expression of GLI3 repressor in Ptch1-deficient mice rescued ectopic Ptch2 expression and obstructive hydronephrosis. + Whole transcriptome analysis of isolated Ptch2 cells revealed coexpression of genes characteristic of stromal progenitor cells. Genetic lineage tracing indicated that stromal cells blocking the ureteropelvic junction were derived from intermediate mesoderm–derived renal progenitors and were distinct from the smooth muscle or epithelial lineages. Analysis of obstructive ureteric tissue resected from children with congenital intrinsic ureteropelvic junction obstruction revealed a molecular signature similar to that observed in Ptch1-deficient mice. Together, these results demonstrate a Hedgehog-dependent mechanism underlying mammalian intrinsic ureteropelvic junction obstruction.

J Am Soc Nephrol 29: 532–544, 2018. doi: https://doi.org/10.1681/ASN.2017050482

Ureteropelvic junction obstruction (UPJO) is the Received May 4, 2017. Accepted October 5, 2017. most common form of congenital urinary tract ob- S.S.-D. and L.C. contributed equally to this work. struction. UPJO is also a major cause of antenatal fi Published online ahead of print. Publication date available at dilation of the renal pelvis, identi ed by ultrasound www.jasn.org. in 1% of fetuses.1 Most cases of UPJO are caused by Correspondence: Dr. Norman D. Rosenblum, The Hospital for intrinsic narrowing of the ureter at the pelvic-kidney Sick Children, Peter Gilgan Centre for Research and Learning, Rm junction, with deposition of extracellular matrix, in- 16-9-706, 686 Bay Street, Toronto, ON M5G 0A4, Canada. Email: filtration of inflammatory cells, and smooth muscle [email protected] – hypertrophy.2 4 Yet, underlying pathogenic Copyright © 2018 by the American Society of Nephrology

532 ISSN : 1046-6673/2902-532 J Am Soc Nephrol 29: 532–544, 2018 www.jasn.org BASIC RESEARCH mechanisms are largely undefined, particularly because of the Statement of Significance absence of experimental models of congenital UPJO. Lack of reli- able biomarkers to differentiate UPJO from other forms of kidney Ureteropelvic junction obstruction (UPJO) is a major cause of con- dilation limits the ability to predict a need for corrective surgery, genital dilation of the renal pelvis. The pathogenesis of UPJO is fi fi currently performed in approximately 50% of affected infants,5,6 unde ned, no biomarkers are identi ed to predict its natural history, and surgical correction is the only treatment. This study demon- fi 2–4 and the development of speci c therapeutic strategies. strates that increased Hedgehog signaling activity, caused by Ptch1 The ureteropelvic junction (UPJ) is a component of the deficiency within a narrow time window during mouse develop- renal collecting system, which consists of collecting ducts, ca- ment, results in ectopic localization of cortical stromal cells to the lyces, and the pelvis, all derived from the epithelial ureteric bud ureteropelvic junction, causing blockage of urinary outflow and (UB). After UB outgrowth from the Wolffian duct at 5 weeks dilation of the renal pelvis. Expression of genes characteristic of cortical stromal cells and Hedgehog signaling activity was increased human gestation (approximately E10.5 in mice), the distal UB in obstructive ureteric tissue resected from affected children. These elongates caudally and integrates with the bladder.7,8 The ros- results provide a basis for genetic studies in human UPJO and tral tip of the UB invades the metanephric mesenchyme (MM) therapeutic strategies targeting the Hedgehog signaling pathway. and undergoes repeated branching events regulated by recip- rocal inductive interactions between nephron progenitors sur- during murine renal development.26,28 Embryonic lethality rounding the UB tips, and Foxd1+ cells that surround the before metanephric kidney development in Ptch1-null mice nephron progenitors.9–12 Renal calyces and the pelvis are provides a basis to investigate Ptch1 functions in a lineage-specific formed from the first several ureteric branch generations.13 manner.29 We generated mice in which Ptch1 deficiency is re- Although the morphogenetic events in pelvicalyceal remodel- fl fl stricted to the metanephric lineage using Ptc1 ox/ ox mice and an ing are largely undefined, Wnt7b expression in the ureteric 2 2 2 2 MM-specific Rarb2Cre allele (Ptch1 / MM).30–32 Ptch1 / MM stalk is crucial in ureteric tree elongation and pelvis formation.14 Ureteric smooth muscle development begins at E15.5 mice showed a 60% decrease in expression of Ptch1 exon 3, the with differentiation of Tsh3+ and Tbx18+ smooth muscle pro- target of CRE-mediated excision, in metanephric rudiments at , genitors, both of which arise from tail-bud mesoderm.15–17 E11.5 (n=3; P 0.05; Figure 1A). qRT-PCR analysis of an undis- During renal development, the Hedgehog (HH) ligand, Sonic rupted exon of Ptch1 (exon 17) demonstrated a two-fold increase , hedgehog (SHH), and its receptor Patched 1 (Ptch1), are expressed as early as E11.5 (n=3; P 0.05; Figure 1B), consistent with in tubular epithelium and stromal cells adjacent to the pelvic- increased HH signaling activity. Compared with control mice, kidney junction. In vertebrates, HH ligand binds to PTCH and expression of the Ptc1-lacZ reporter allele was increased in the 2/2MM relieves PTCH-mediated inhibition of Smoothened, a transmem- medullary stroma and expanded to the renal cortex in Ptch1 brane protein expressed on the cell surface.18,19 In the presence of kidneys (Figure 1, C and D). 2/2MM HH, full-length GLI proteins translocate to the nucleus, acting as Histologic analysis of Ptch1 kidneys at E18.5 dem- transcriptional activators.20,21 In the absence of HH, full-length onstrated distention of the renal pelvis (i.e., hydronephrosis) GLI3 is processed to a truncated N-terminal form (GLI3R), which bilaterally and ablation of the renal medulla with no ureteric acts as a transcriptional repressor.22,23 In mice, GLI3R regulates dilation (80% penetrance, 29 males and 33 females) (Figure 1, ureteric branching, nephron progenitor proliferation, and urinary E and F). Measurement of pelvic volume in three-dimensional pacemaker activity.24–27 SHH acts as a paracrine signal to Tbx18+ images of E18.5 mutant kidneys generated by magnetic reso- mesenchymal cells to control smooth muscle differentiation and nance imaging (MRI) revealed a nine-fold increase (n=10; proliferation.16,28 P,0.05; Figure 1, G–I). Histologic and immunofluorescence Here, we identified increased HH-GLI signaling as a cause of analysis of the nascent epithelial collecting ducts and distal congenital UPJO in mice and a marker of UPJO in pediatric tubules marked, respectively, by calbindin and E-cadherin, patients. Ptch1 deficiency in the MM caused congenital intrinsic demonstrated comparable formation of the epithelial tubular UPJO because of an ectopic population of PTCH2+ stromal pro- network within the renal medulla in both wild-type (WT) and genitor cells arising from intermediate mesoderm (IM)–derived mutant kidneys at E15.5 (Supplemental Figure 1, A–F). Fur- kidney progenitors at the onset of MM specification. Remarkably, ther, expression of LEF1, a canonical WNT pathway effector expression of stromal and HH signaling-related genes was in- that is expressed in medullary interstitial cells and regulates creased in obstructing ureter tissue segments in 38% of children the collecting duct cortico-medullary axis,14 did not differ 2 2 with intrinsic UPJO. Together, these results describe a HH-GLI– between Ptch1 / MM kidneys and controls (Supplemental mediated pathogenic mechanism underlying congenital UPJO. Figure 1, G and H). Intrapelvic dye injection in E18.5 embryos demonstrated that injected dye failed to pass the UPJ in mutant mice (Figure 1, J and K). Together, these data indicate that RESULTS Rarb2Cre-mediated Ptch1 deletion in the MM causes congenital intrinsic UPJO. Ptch1 Deletion Causes Congenital Intrinsic UPJO HH signaling was previously shown to control cellular differ- HH signaling activity, as reported by expression of Ptch1-lacZ, entiation during ureter formation28 and defects in urothelial differ- is localized to the UB and the surrounding mesenchyme entiation or smooth muscle proliferation cause UPJO in mice.33,34

J Am Soc Nephrol 29: 532–544, 2018 Hedgehog and Ureteropelvic Junction Obstruction 533 BASIC RESEARCH www.jasn.org

Figure 1. Ptch1 deficiency targeted to the MM results in obstructive hydronephrosis. (A and B) qRT-PCR analysis on metanephric rudiment isolated from E11.5 embryos (n=3 per genotype) demonstrates a significant decrease in Exon3 (A), and a significant increase in Exon17 of the Ptch1 transcript (B) P,0.05. (C and D) b-galactosidase staining reveals expansion of HH signaling activity from the ureteric mesenchyme and medullary stroma to the more cortical stroma (black arrowheads) and the cap mesenchyme (white arrowhead) 2 in Rarb2Cre;Ptch1LacZ/flox at E13.5. Scale bars, 200 mm. (E and F). Histologic analysis shows hydronephrosis in Rarb2Cre;Ptch1flox/ kidney tissues at E18.5. Scale bars, 500 mm. (G–I). Three-dimensional reconstruction created by MRI imaging of the WT and mutant kidneys demonstrates an expanded renal pelvis and no pelvis exit. Scale bars, 1 mm. Measurement of the pelvis size demonstrates a nine-fold increase in mutants (n=10 mutants and six WT; P,0.05) (I). (J and K) Dye injected to the renal pelvis of the left control kidney passes through the ureter and accumulates in the bladder, but it does not traverse the UPJ (arrowhead) and accumulates in the pelvis in the mutants. Scale bars, 200 mm. Insets: original magnification, 335. (L–O) Immunofluorescence staining by anti-smooth muscle actin, anti-cytokeratin, E-cadherin, and anti-uroplakin antibodies shows no marked defect in the ureteric smooth muscle and epithelial layers in mutant kidneys. Scale bars, 100 mm. p, pelvis; u, ureter.

2 2 Yet, analysis of the ureter distal to UPJO in Ptch1 / MM mice failed to in the ureter after the full onset of hydronephrosis (Figure 1, demonstrate a difference in the expression of ureteric epithelial L–O). In situ hybridization analysis of Tbx18 expression at markers, including uroplakin, cytokeratin, and E-cadherin. Sim- E13.5 demonstrated no defect in the ureteric mesenchymal 2 2 ilarly,uretericsmoothmusclecells,markedbysmoothmuscle lineage in Ptch1 / MM mutants (Supplemental Figure 1, K actin, expressed transgelin, a differentiation marker, in both WT and L). Together, these data strongly suggest that UPJO in andmutanttissueatE17.5(SupplementalFigure1,IandJ).We Ptch1-deﬁcient mice is not associated with impaired urothe- observed no obvious difference in smooth muscle thickness lial differentiation or smooth muscle proliferation.

534 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 532–544, 2018 www.jasn.org BASIC RESEARCH

Increased HH Signaling Activity and Ectopic Ptch2 in the Presumptive UPJO We investigated the role of HH signaling in 2 2 UPJO in Ptch1 / MM mice. At E18.5, Ptch1-lacZ was strongly expressed in the obstructing tissue located between the renal pelvis and the ureter (Figure 2B). Fur- ther, expression of Ptch2,aPtch1 homolog, was assayed using a Ptch2-lacZ reporter,35 and was found only in the mutant kidneys in the presumptive UPJ as early as E12.5, with continued expression in UPJ obstructing tissue until E18.5 (Figure 2, D and F, Supple- mental Figure 2). To examine the functional contribution of Ptch2 to UPJO, we generated compound mutants with both Ptch1 and Ptch2 deficiency targeted to the metanephric lineage using Rarb2Cre. Histologic analysis at E18.5 revealed no difference between the Ptch1;Ptch2 compound mutants and the Ptch1 mutants alone (Figure 2, G and H). These results demonstrate that the UPJ is an active domain of HH signaling in kidneys 2 2 of Ptch1 / MM mice, and Ptch2 is a specific marker of UPJ obstructing cells. Next, we investigated the functional role of increased HH signaling in the pathogenesis of UPJO by decreasing HH signaling activity. Genetic analyses in mice have established the requirement for a balance of GLI3 repressor and activator during kidney development.36,37 Here, we introduced a D constitutive GLI3R allele (Gli3 699)inthe 2 2 Ptch1 / MM genetic background to re- duce HH signaling activity. In contrast to 2 2 Ptch1 / MM mice, Ptch2-lacZ expression was greatly diminished in the presumptive 2 2 D UPJ of Ptch1 / MM;Gli3 699/+ mice (Fig- ure 3, A–C). Moreover, histologic analysis suggested a partial restoration of normal pelvic morphology (Figure 3, D–F), intra- Figure 2. UPJO is characterized by strong HH signaling activity and ectopic Ptch2 ex- pelvic dye injection demonstrated rescue of b pression. (A and B) -galactosidase staining of kidney tissues isolated from E18.5 Ptch1-lacZ UPJO (Figure 3, G and H), and MRI reporter mice reveals localization of HH signaling activity to the ureter and the medullary revealed a marked reduction in pelvic vol- stroma surrounding the collecting ducts in controls. In the presence of Rarb2Cre, Ptch1-LacZ is strongly expressed in the obstructing tissue at the UPJ, indicated by arrowhead. Scale bars ume in seven out of ten mice (Figure 3I). 500 mm. The insets show ectopic HH signaling activity in the cortical stroma of the mutant These results demonstrate that increased 2/2MM kidney compared with control. Scale bars, 100 mm. (C) b-galactosidase staining of kidney HH signaling causes UPJO in Ptch1 tissues isolated from E13.5 Ptch2-LacZ reporter mice shows no LacZ expression in the mice. 2 2 2 presumptive UPJ in controls. (D) In Ptch1 / MM kidneys (Rarb2Cre;Ptch1flox/ ;Ptch2LacZ/+), there is strong LacZ expression specific to the presumptive UPJ. Scale bars, 200 mm. (E and PTCH2-Positive Cells Blocking the F) Ptch2-LacZ continues to be highly expressed in the obstructing tissue at the UPJ in UPJ Express Genes Characteristic of 2/2MM m Ptch1 embryos at E17.5. Scale bars, 500 m. (G and H) Histologic analysis at E18.5 Stromal Progenitor Cells 2/2MM flox/2 reveals that complete removal of PTCH2 in Ptch1 background (Rarb2Cre;Ptch1 ; We investigated the identity of Ptch2+ cells that LacZ/LacZ m 2 2 Ptch2 ) does not rescue the phenotype. Scale bars, 500 m. block the UPJ in Ptch1 / MM mice at E13.5, a

J Am Soc Nephrol 29: 532–544, 2018 Hedgehog and Ureteropelvic Junction Obstruction 535 BASIC RESEARCH www.jasn.org

in Ptch2+ cells by qRT-PCR (Figure 4D). Con- sistent with these results, in situ hybridization analysis demonstrated expansion and ectopic expression of Foxd1 and Raldh2 at the UPJ of 2 2 Ptch1 / MM mice at E13.5 (Figure 4, E–H). Next, we investigated whether Ptch2+ cells coexpress HH-related and stromal markers or, alternatively, are a heterogeneous group of cells that express either but not both of these markers. We analyzed gene expression at a single-cell resolution using double- ﬂorescence in situ hybridization and confocal microscopy of the pelvic-kidney junction in 2 2 Ptch1 / MM mice. As expected, Raldh2,and Foxd1 colocalized in the cells blocking the UPJ (Figure 4, I’–I’’’). Remarkably, the vast majority of Ptch2+ cells also expressed Raldh2 and Foxd1 (Figure 4, J’–K’’’). In contrast, coexpression of Bmp4 and Ptch2 was rarely observed (Figure 4, L’–L’’’). Together, these results indicate that Ptch2+ cells that 2 2 block the presumptive UPJ in Ptch1 / MM mice are ectopically located cortical stromal cells.

Ptch1 Deletion in IM-Derived Renal Figure 3. Addition of one Gli3R allele in the Ptch1-deficient mice rescues UPJO and Progenitors Causes UPJO pelvis size. (A2C) b-galactosidase staining shows very little Ptch2-LacZ expression in Renal stromal progenitors are derived from 2 2 D + the presumptive UPJ in Ptch1 / /MM;Ptch2LacZ/+;Gli3 699/+ compared with that of the Osr1 population of the IM and the spec- control mice at E14.5. Scale bars, 200 mm. (D–F) Histologic analysis of E18.5 kidney ification of the stromal lineage occurs be- 2 2 D tissues reveals a partial normalization of the phenotype in Ptch1 / MM;Gli3 699/+ mice tween E9.5 and E11.0.15,40,41 Because compared with Ptch1-deficient mutants. Scale bars, 500 mm. (G and H) Intrapelvic dye Rarb2-Cre promoter activity is evident at LacZ/flox injection demonstrates the restoration of UPJ connectivity in Rarb2Cre;Ptch1 ; E9.5,31 we examined whether the ectopic pop- D699/+ m Gli3 mice compared with WT kidneys. Scale bars, 75 m. (I) Measurement of the ulation of Foxd1+ cells emerges as a result of pelvis size using three-dimensional reconstructed images created by MRI confirms a + 2 2 D 2 2 increased HH signaling in Osr1 cells. To ad- significant reduction in pelvis size in Ptch1 / MM;Gli3 699/+ compared with Ptch1 / MM fl fl dress this possibility, we crossed a Ptch1 ox/ ox (n=10 per genotype; *P,0.05; **P,0.005). 1 2 dam with an Osr1EGFPCreERt2;Ptch1 / male, followed by tamoxifen injection at E9.5. Histo- day after the onset of ectopic Ptch2-lacZ expression. We micro- logic examination of kidneys isolated from all four Osr1EGFPCreERt2; 2 2 fl 2 dissected kidney-ureter junction tissue from Ptch1 / MM mice Ptch1 ox/ mice at E18.5 revealed hydronephrosis similar to that 2 2 obtained from two litters, purified Ptch2+ cells by FACS using observed in Ptch1 / MM mice (Figure 5, A and B). Further, in situ antibodies specific for PTCH2, and performed RNA sequencing hybridization confirmed ectopic Ptch1 expression in cells blocking on Ptch2-positive and -negative cells (Figure 4, A and B). To iden- theUPJ(Figure5,CandD). tify markers for Ptch2+ cells, we inspected differentially expressed Osr1 is expressed in both anterior and posterior IM, which genes (GEO accession no. GSE97126) for those that were upregu- give rise to the ureteric epithelium and metanephric lineages, lated and identified the HH signaling effector genes, Ptch1, Gli1, respectively.41 To exclude the possibility that increased HH Hhip,andBmp4. The expression of these genes was validated by signaling in the UB causes the observed phenotype, we acti- qRT-PCR (Figure 4, C and D). Surprisingly, RNA-seq data revealed vated HH signaling in Sall1+ daughters of Osr1+ cells found that Ptch2+ cellshighlyexpressgenes(Pdgfrß, Tnc,andSfrp1)that only in the posterior nephrogenic mesenchyme.42 We gener- fl 2 fl fl are normally expressed in the embryonic cortical interstitium and ated Sall1CreERt2;Ptch1 ox/ embryos by crossing a Ptch1 ox/ ox 1 2 capsule (Figure 4C).9,38,39 Because both cortical interstitium and dam with a Sall1CreERt2;Ptch1 / male and injected tamoxifen capsule are derived from Foxd1+ stromal progenitors,9 we hypoth- at E9.5.43 A total of 82% of the mutant embryos (n=9 out of esized Foxd1+ cells to be the origin of the cells blocking the UPJ. 11) demonstrated hydronephrosis with UPJO similar to that fl 2 Indeed, a significant increase in expression of both Foxd1 and observed in Osr1EGFPCreERt2;Ptch1 ox/ mice (Figure 5, E–H). Raldh2, another marker of stromal progenitor cells, was found Interestingly, we did not observe a similar phenotype after

536 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 532–544, 2018 www.jasn.org BASIC RESEARCH

Figure 4. UPJO cells are characterized by ectopic expression of Ptch2 and markers of stromal progenitors. (A and B) Approximately 2 2500 Ptch2+ and approximately 14,000 Ptch2 cells were isolated from presumptive UPJ of E13.5 kidneys using anti-PTCH2 antibodies followed by FACS. (C) RNA-seq analysis reveals that among the genes upregulated in the Ptch2+ cell population, there were specific markers of cortical stroma and HH signaling targets. (D) qRT-PCR analysis demonstrates a significant increase in Ptch1, Ptch2,and Bmp4 genes downstream of HH signaling, in addition to Foxd1 and Raldh2,specific markers of the cortical stroma (n=4; P,0.05). (E–H) In situ hybridization in E13.5 kidney tissues demonstrates an increase in Foxd1 and Raldh2 gene expression at the cortex and ectopic expression in the presumptive UPJO in Ptch1-deficient mice (arrowheads). Scale bars, 200 mm. (I’–L’) Double-fluorescence in situ hybridization on E13.5 demonstrates coexpression of Foxd1 and Raldh2, in addition to coexpression of Ptch2 with both Foxd1 and 2 Raldh2. Bmp4 does not fully colocalize with Ptch2 (asterisks in Ptch2+;Bmp4 cells). Scale bars, 15 mm. tamoxifen injection at E8.5 or E10.5 (n=4 per time point, data nephrogenic lineages. At E16.5, the funnel-shaped pelvis that not shown). Immunofluorescence confirmed that Raldh2+ normally opens into the ureter was blocked by tdTomato- cells block the UPJ as early as E16.5 when Ptch1 deletion is expressing cells in Ptch1-deficient kidneys (Figure 6, B and C). driven by either Sall1CreErt2 or Osr1CreErt2 (Figure 5, I–L). These blocking cells did not express smooth muscle actin, but To determine the embryonic origin of the cells that block expressed the mesenchymal protein vimentin (Figure 6, D– theUPJ,weperformedlineagetracingbyintroducinga F). Vimentin-expressing cells in the UPJO colocalize with fl 2 Rosa26-tdTomato reporter allele into a Sall1CreERt2;Ptch1 ox/ ge- ectopic expression of RALDH2 in the mutant kidneys (Sup- netic background. Cre-mediated recombination was induced at plemental Figure 3). We further confirmed coexpression of E9.5 by tamoxifen treatment of pregnant females. Sall1CreERt2; tdTomato and vimentin in individual UPJO cells using higher fl Ptch1 ox/+;Rosa26tdT/+ littermates were used as controls. Descen- resolution, three-dimensional reconstruction of stacks of sin- dants of Sall1+ cells, indelibly labeled with tdTomato, were iden- gle optical sections (Figure 6, G and H). Together, these data tified by high levels of PBX1 expression in stromal cells and low demonstrate that the ectopically located cortical stromal cells 2 2 PBX1 levels in nephrogenic cells (Figure 6A). Our results dem- observed in Ptch1 / MM kidneys are derived from Sall1+ pro- onstrated that at E9.5 Sall1+ cells contribute to both stromal and genitor cells.

J Am Soc Nephrol 29: 532–544, 2018 Hedgehog and Ureteropelvic Junction Obstruction 537 BASIC RESEARCH www.jasn.org

Figure 5. Induction of Ptch1 deletion in Sall1+ and Osr1+ renal precursor cells results in obstructive hydronephrosis. (A and B) 2 Osr1CreErt2;Ptch1flox/ kidneys exhibit hydronephrosis at E18.5 when dams are injected with tamoxifen at E9.5. (Scale bars: 500 mm). (C and D) Ptch1 is highly expressed in the medullary stroma in control kidneys at E18.5. Ptch1 expression is expanded to the cortex and 2 ectopically detected at the UPJO. Scale bars: 500 mm (Insets: original magnification, 320). (E and F) Sall1CreERt2;Ptch1flox/ kidneys exhibit hydronephrosis at E18.5 when dams are injected with tamoxifen at E9.5. (Scale bars: 500 mm). (G and H) Dye injected to the pelvis of the control kidneys (n=9) traverses the ureter and accumulates in the bladder, while it accumulates in the pelvis of the mutant kidney (n=4) and does not reach the bladder (asterisk) t: testis. (I–L) Immunostaining for RALDH2, a marker of cortical stromal cells demonstrates restricted expression of RALDH2 in the cortex at two stages of development in control kidneys. RALDH2 expression is 2 maintained in the cortex of the Osr1CreErt2;Ptch1flox/ and Sall1CreErt2;Ptch1flox/– kidneys, but is ectopically detected in the presumptive UPJ (arrowheads). u, ureter; p, pelvis; *, incomplete pelvis. (Scale bars: 100 mm).

FOXD1 and PTCH2 are Upregulated in Obstructive compared with contiguous normal ureter (Figure 7A). In con- Ureteric Tissue of a Subset of Children with Congenital trast to our findings in Ptch1-deficient mice, expression of Intrinsic UPJO RALDH2 was not significantly altered in affected patients (data Our findings in genetic mouse models, described above, raise the not shown). HH signaling activity in obstructing ureteric tissue possibility that aberrant HH signaling activity in Foxd1+ cells could was assessed by analyzing the expression of PTCH2, PTCH1, contribute to the genesis of human UPJO. To address this possi- GLI1,andHHIP. PTCH2 expression was increased by an average bility, we obtained obstructed and normal ureter tissue from eight of three-fold in four of the five patients with increased FOXD1 infants and young children at the time of surgical repair of con- (Figure 7B, Table 1). GLI1 and HHIP were significantly increased genital intrinsic UPJO. Children with a history of urinary tract in three of these four patients (three of the five patients with infection were excluded to control for secondary effects on gene increased FOXD1) (Figure 7, C and D). PTCH1 was increased expression. RNAwas isolated from a segment of both the obstruc- in a single patient (U5) who did not exhibit an increase in either tive and normal ureter and was subjected to qRT-PCR for gene PTCH2 or FOXD1 (summarized in Table 1). Patient U5 also expression analysis. Contiguous tissue was used for in situ exam- exhibited increased expression of GLI1 and HHIP (Figure 7, C ination of protein expression by immunohistochemistry. We and D). Immunohistochemical analysis confirmed increased ex- controlled for variations across individuals by normalizing gene pression of both FOXD1 and PTCH2 protein in the muscular expression in the obstructed UPJ segment to that in the adjacent layer of the obstructing tissue in patients U2, U4, and U7 (Figure normal ureter for each individual analyzed. 7, E–H, representative samples from patient U7). Overall, our We identified five patients in whom FOXD1 was increased by data supports a role for increased HH signaling and renal stro- an average of six-fold in the obstructive ureteric tissue segment mal cells in the pathogenesis of human UPJO.

538 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 532–544, 2018 www.jasn.org BASIC RESEARCH

DISCUSSION

The molecular mechanisms underlying congenital intrinsic UPJO are poorly un- derstood. Here, we demonstrated that constitutive activity of HH-GLI signaling in IM-derived renal progenitors induces formation of an ectopic population of Ptch2+ cortical stromal cells that obstruct the UPJ during embryogenesis. In addition, we identiﬁed a molecular signature characteristic of stromal progenitors and activated HH signaling in the obstructed UPJ in a subset of children with congenital UPJO. These results demonstrate a HH-GLI– dependent mechanism causing murine intrinsic UPJO and implicate dysregulated HH signaling in the pathogenesis of human UPJO.

Role of Stromal Cells in the Pathogenesis of UPJO Although it has been established that the majority of Foxd1+ stromal progenitors originate from the Osr1+ posterior IM, how stromal lineage segregation occurs is not clear.15,40 Previous studies have shown that neither Foxd1 nor Osr1 regulate specification and differentiation of the stromal lineage.9,40 Here, we provide the firstevidencethatHHsignal- ing may play a role in stromal cell specification from precursor cells. Induction of Ptch1 deficiency in both Osr1+ and Sall1+ cells at E9.5 caused obstructive hydronephrosis with ectopic Raldh2 expression in the presumptive UPJ.40,43 Lineage tracing analysis after activation of HH signaling in Sall1+ cells confirmed that the ectopically located stromal cells in the UPJ are direct descendants of IM-derived renal progenitors. Previous published work has

tdTomato+ cells are distinct from smooth muscle actin-expressing smooth muscle lineage (green). (F) Higher resolution confocal image showing vimentin (dark blue) in the UPJ 2 obstructing tdTomato+;SMA cells. Scale bar, 25 mm, inset shows a cross-section of z-stacks. Figure 6. Sall1+ renal precursor cells give rise to UPJ obstructing cells. (A) Immuno- (G and H) Three-dimensional reconstructed ﬂuorescence staining in Sall1CreERt2;Ptch1ﬂox/+;Rosa26tdT/+ kidneys at E16.5, showing image from stacks of single optical sections of high PBX1 expression (green) in stromal cells and low PBX1 expression in nephron the UPJ obstructing cells showing colocaliza- progenitors. After Cre-mediated recombination in Sall1+ cells at E9.5, tdTomato- tion of tdTomato and vimentin (dark blue) in labeled cells localize to both stromal and nephron progenitors, and are excluded from individual cells. *indicates UPJO; dashed lines the ureteric lineage (light blue). Scale bars, 25 mm. (B and C) tdTomato+ cells localize in indicate kidney periphery or individual cells; and block the UPJ in mutant mice. Scale bars, 200 mm. (D and E) UPJ obstructing arrows indicate obstructing tdTomato+ cells.

J Am Soc Nephrol 29: 532–544, 2018 Hedgehog and Ureteropelvic Junction Obstruction 539 BASIC RESEARCH www.jasn.org

progenitors are responsive to the effect of HH signaling before establishment of the nephrogenic and stromal lineage boundaries. HH signaling has been shown to regulate Foxd1 expression in the context of craniofacial and optic vesicle development.45,46 It is likely that constitutive HH signaling activity during stromal cell specification promotes stromalcellfatebyupregulatingFoxd1.Adif- ferential analysis of the genetic profile in Osr1+ or Sall1+ cellsislikelytorevealmechanisms involved in stromal lineage specification. What mechanism could underlie the ectopic location of Ptch1-deficient stromal progenitors? Foxd1+ cells initially form in a caplike pattern anterior to the MM, and subsequently integrate into the kidney periphery in a Hox10-dependent manner.47 As the kidney grows, a subset of dif- ferentiated stromal cells localize within the glomerular capillary loops and medulla, whereas self-renewing Foxd1+ cells are maintained in the cortex.9 Our data suggests that Ptch1 in the MM is not crucial for the initial integration of Foxd1+ cells into the cortex because renal capsule and cortical interstitium patterning was unaffected in Ptch1-deficient mice. RNA-seq of Ptch2+ cells demonstrated enrichment for a SHH receptor, Boc, which is essential in SHH-dependent commissural neuron guidance in the spinal cord.48,49 Thus, SHH generated in renal medulla may act as a chemoattractant for BOC-expressing stromal cells, resulting in ectopic localization in the UPJ.

HH-GLI Mutations and Control of Human Renal-Urinary Tract Figure 7. A subset of children with congenital UPJO demonstrates changes in ex- Development pression of FOXD1 and PTCH2. (A) The expression of FOXD1 in the obstructive UPJ Little is known about the role of PTCH2 in tissue resected from children with congenital UPJO relative to the nonobstructive development. Similar to Ptch1,transcriptional fi ureteric tissue is signi cantly higher in patients 2, 4, 6, 7, and 8. (B) In patients 2, 4, 7, activation of Ptch2 occurs in response to SHH, and 8, the expression of PTCH2 in the obstructive UPJ tissue relative to the non- and Ptch2 is downregulated in the absence of obstructive ureteric tissue is significantly higher than that of the rest of the patients. (C 50 and D) U2, 4, 6, 7, and 5 demonstrate significantly higher relative expression of GLI1 GLI. Although medullary stroma is a do- compared with the other patients. U2, 4, 7, and 5 demonstrate significantly higher main of high HH signaling activity, Ptch2 is relative expression of HHIP compared with the other patients. (*P,0.05; **P,0.005) notexpressedinthemedullarystromainWT (E–H) Immunohistochemical staining of the obstructive UPJ and nonobstructive ure- kidneys.UponremovaloffunctionalPTCH1 teric tissue from patient U7 using anti-FOXD1 and anti-PTCH2 antibodies shows a from the MM, Ptch2 is markedly upregulated higher intensity of expression of both FOXD1 and PTCH2 in the UPJO (Original in the obstructing cells at the presumptive UPJ magnification, 340). as early as E12.5. This finding illustrates a requirement for PTCH1 to downregulate Ptch2 shown that Ptch1 deletion in the MM using Tcf21Cre,whichisnot expression, which is consistent with increased expression of Ptch2 active at the IM stage, results in glomerular and tubular cysts but inbothhumanbasalcellcarcinoma and medulloblastoma with not hydronephrosis.44 Our findings further suggest that renal Ptch1 mutations.51,52 Loss of Ptch2 in Ptch1-null mice worsens the

540 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 532–544, 2018 www.jasn.org BASIC RESEARCH

Table 1. Summary of the transcriptional analysis in children Sall1Cre or Osr1Cre does not cause gross defects or overgrowth with congenital UPJO of the ureteric mesenchyme. In another UPJO model, targeted + Patient No. Foxd1 Ptch2 Gli1 Hhip Ptch1 inactivation of Sec10 in Ksp1.3 ureteric epithelium resulted in 2++++— decreased ureteric expression of uroplakin, followed by over- 4++++— growth of ureteric mesenchymal cells.34 In contrast, in Ptch1- 7++++— deficient mice reported here, the urothelial layer is physically 8++—— —intact and exhibits normal expression of urothelial markers 6+— + ——distal to the UPJ. This provides further evidence that the MM 5 ——++ +and UB behave as distinct lineages after E8.5.41,59 1 ————— 3 ————— Resected obstructive ureteric tissue from patients 2, 4, and 7 demonstrates CONCISE METHODS an increase in the levels of FOXD1, PTCH2, GLI1,andHHIP transcripts relative to the adjacent nonobstructive ureteric tissue. Patients 6 and 8 have increased FOXD1 along with GLI1 and PTCH2, respectively. Patient 5 Mice demonstrates increased PTCH1, GLI1,andHHIP but not FOXD1 or PTCH2. Rarb2Cre mice (provided by R. Behringer, University of Texas, Hous- In patients 1 and 3, expression of the analyzed genes were comparable in 30 1/2 LacZ/+ obstructive tissue to that of the adjacent nonobstructive ureter. +, upregu- ton, TX ) were mated with Ptch1 or Ptch1 (provided by C.C. lated; —, not upregulated. Hui, University of Toronto, Toronto, Canada) mice to generate 1 2 Rarb2Cre;Ptch1 / or Rarb2Cre;Ptch1LacZ/+ males. Ptch1loxP/loxP mutant phenotype, suggesting that the two homologs have (provided by B. Wainwright, University of Queensland, Queensland, 2 overlapping functions.53,54 However, genetic removal of Ptch2 Australia)60 females were used to generate Rarb2Cre;Ptch1loxP/ or 2 2 2 2 in our Ptch1 / MM mice did not affect the severity of the observed Rarb2Cre;Ptch1LoxP/LacZ (Ptch1 / MM) progeny. Sall1CreERT2; (a gift 2 phenotype. This suggests that Ptch2 does not play a functional role from R. Nishinakamura43)andOsr1EGFPCreERT2;Ptch1loxP/ embryos in regulating HH signaling activity during kidney development, were generated by a similar approach. Osr1EGFPCreERT2 and Rosa26- but serves as a specificmarkeroftheonsetofUPJO. tdTomato mice were purchased from Jackson Laboratories. For Inactivating mutations in Gli3 and Ptch1 have been identi- mouse breeding, noon of the day of a vaginal plug was considered fied in humans with Greig cephalopolysyndactyly syndrome as E0.5. To activate Cre recombinase in the ERT2-carrying embryos, andNevoidbasalcellcarcinomasyndrome(Gorlinsyn- pregnant dams were injected intraperitoneally with 3 mg of tamox- drome), respectively.55,56 Currently, no known association be- ifen dissolved in sesame oil. All of the animal procedures were ap- tween human Gli3 loss of function or Ptch1 mutations and proved by the Animal Ethics Committee at the Toronto Center for renal malformation exists. Previous studies have shown a sig- Phenogenomics (Toronto, Ontario, Canada) and in accordance with nificant decrease in Gli3 expression in UPJO specimens from the Canadian Council of Animal Care. neonatal patients compared with that in ureters from patients with Wilm tumor.57 Our data extend these observations and Patient Specimens demonstrate a significant increase in Ptch2 expression in the This study was approved by the Research Ethics Board of the Hospital obstructive UPJ segments in a subset of patients with congen- for Sick Children in Toronto, Canada (approval no. 1000043163). ital UPJO. After informed consent, excisional UPJ biopsies and a short segment of the adjacent unaffected ureter were obtained during reparatory Ptch1-Deficient Mice as a Novel Genetic Model of surgery at the Hospital for Sick Children. Congenital UPJO Our results uncover a mechanism distinct from previous Histologic Analyses studies that implicated abnormal ureteric smooth muscle or Paraffin-embedded kidney or ureter were prefixed in 4% paraformal- urothelial development in UPJO. Conditional deletion of the dehyde overnight at 4°C and sectioned at 5 mm. OCT-embedded frozen TGF-b/BMP signaling effectors, Smad4 and Id2,intheTbx18+ kidneys were prefixed in 4% paraformaldehyde, cryopreserved using ureteric mesenchyme has been shown to cause UPJO, second- sucrose gradient, and sectioned at 8 mm. Immunofluorescence and ary to a functional obstruction caused by a reduction in immunohistochemical staining was performed on paraffin-embedded smooth muscle cell number and ureteric contractility.58 or frozen sections using standard protocols (see Supplemental Material Although HH signaling has been shown to regulate smooth for protocol and antibodies used). For b-galactosidase staining, 10 mm muscle development,28 we did not observe disruption in ure- frozen sections were postfixed in 0.2% paraformaldehyde, washed in teric smooth muscle even after the onset of hydronephrosis in LacZ rinse buffer, and stained by immersion in b-galactosidase staining mice with Ptch1 deficiency driven by Rarb2Cre or Sall1Cre. solution overnight at 37°C in the dark (see Supplemental Material for This is consistent with previous studies showing intact smooth solution composition). The X-gal–stained sections were counterstained muscle cells in mice with targeted deletion of smoothened in with eosin. Immunofluorescence imaging was done using a Zeiss epi- the MM,25 and confirms the nonoverlapping domain of Tbx18 fluorescence microscope and Olympus spinning disk confocal micro- and Rarb2Cre.15 Although Osr1, Sall1,andTbx18 have over- scope, and Z-stack processing was done using Volocity software. Other lapping expression domains, Ptch1 deficiency induced by images were captured using Leica light microscope.

J Am Soc Nephrol 29: 532–544, 2018 Hedgehog and Ureteropelvic Junction Obstruction 541 BASIC RESEARCH www.jasn.org

Intrapelvic Dye Injections In Situ Hybridization Briefly, bromophenol blue dye was injected into the renal pelvis of In situ hybridizations were performed on paraffin-embedded kidney isolated whole urogenital systems at E18.5 using a pulled-out Pasteur sections using digitoxin-labeled RNA probes encoding Ptch1, Ptch2, glass pipette.16 Bmp4, Foxd1,andRaldh2 as described.64 Double-florescence in situ hybridization was carried out following the published protocols.65 MRI Preparation, Imaging, and Renal Pelvis Measurement Statistical Analyses Kidney specimens were fixed for 24 hours in MRI fixative solution Statistical analysis was performed using GraphPad Prism software containing 0.2 ml/50 ml Prohance (Bracco, Quebec, Canada)/4% (version 5.0c). Data were analyzed using two-tailed t test. A proba- paraformaldehyde. Kidneys were then placed in 0.2 ml/50 ml in Pro- bility of ,0.05 was considered statistically significant and is desig- hance/PBS for 1 week. Image acquisition was performed at the Mouse nated with an asterisk in all figures. Imaging Centre on a 7.0T Magnex magnet equipped with a multi- channel Varian DirectDrive Console (Varian, Inc., Palo Alto, CA) using multiple-mouse MRI with solenoid coils (14 mm diameter, ACKNOWLEDGMENTS 19 mm length).61 The imaging sequence was a gradient-echo (pa- – – rameters: TR= 50 milliseconds; TE= 3.2 4.8 milliseconds; 4 6 aver- This work was supported by operating grants by the Canadian Institutes m – ages; 35 m isotropic voxels; scan time, 10 hours 22 minutes 15 of Health Research and the Kidney Foundation of Canada (to N.D.R.), a hours 33 minutes). Renal pelvis measurements were obtained using TierICanadaResearchChairinDevelopmentalNephrology(toN.D.R.), MRI and three-dimensional volumetric visualization software a Tier I Canada Research Chair in Imaging (to R.M.H.), and an (Amira v5). Kidney volumes were serially segmented semiautomati- Alexander Graham Bell Canada Graduate Scholarship (to S.S.-D). cally, and then volumes of the renal pelvis were compared between groups.62,63 DISCLOSURES FACS and Quantitative RT–PCR None. Tissue from E13.5 mutant kidneys in each litter was pooled and dis- sociated into single cells using 1 mg/ml collagenase B (Sigma-Aldrich) at 37°C for 10 minutes. Cells were incubated with an anti-PTCH2 REFERENCES antibody (1:100, rabbit, G1191; Cell Signaling) followed by staining with Alexa-488 goat anti-rabbit secondary antibody (1:1000; 1. Mallik M, Watson AR: Antenatally detected urinary tract abnormalities: Molecular Probes, Invitrogen). Cells were sorted using FACS Aria More detection but less action. Pediatr Nephrol 23: 897–904, 2008 SORP (BD Biosciences) at Sickkids-UHN flow cytometry facility. 2. Zhang PL, Peters CA, Rosen S: Ureteropelvic junction obstruction: – Total RNA from sorted cells, and whole kidneys were isolated using Morphological and clinical studies. Pediatr Nephrol 14: 820 826, 2000 3. Özel SK, Emir H, Dervis¸oglu S, Akpolat N, S¸ enel B, Kazez A, Söylet Y, the Qiagen RNeasy Micro kit. RNA from human UPJ specimens Çetin G, Danis¸mend N, Büyükünal SNC: The roles of extracellular ma- was isolated using TRIzol RNA extraction protocol, followed by trix proteins, apoptosis and c-kit positive cells in the pathogenesis of purification with QIAGEN RNeasy micro kit. First-strand cDNA ureteropelvic junction obstruction. J Pediatr Urol 6: 125–129, 2010 was synthesized using SuperScript (II) First-Strand Synthesis Sys- 4. Hosgor M, Karaca I, Ulukus C, Ozer E, Ozkara E, Sam B, Ucan B, Kurtulus S, tem (Invitrogen). Primers for specific transcripts were designed Karkiner A, Temir G: Structural changes of smooth muscle in congenital ureteropelvic junction obstruction. JPediatrSurg40: 1632–1636, 2005 using RNA blast and verified through UCSC in silico PCR. Quan- 5. Oliveira EA, Oliveira MCL, Mak RH: Evaluation and management of titative PCR was carried out using Applied Biosystems 7900 HT fast hydronephrosis in the neonate. Curr Opin Pediatr 28: 195–201, 2016 RT-PCR system. 6. Chertin B, Pollack A, Koulikov D, Rabinowitz R, Hain D, Hadas-Halpren I, Farkas A: Conservative treatment of ureteropelvic junction obstruction in children with antenatal diagnosis of hydronephrosis: Lessons learned RNA Sequencing after 16 years of follow-up. Eur Urol 49: 734–738, 2006 RNA purity and yield was determined using Nanodrop analysis fol- 7. Batourina E, Tsai S, Lambert S, Sprenkle P, Viana R, Dutta S, Hensle T, lowed by integrity analysis with Agilent RNA 6000 Pico Lab Chip Wang F, Niederreither K, McMahon AP, Carroll TJ, Mendelsohn CL: analysis before cDNA library preparation. All samples had RIN.8. Apoptosis induced by vitamin A signaling is crucial for connecting the RNA samples were used for TruSeq cDNA library preparation and ureters to the bladder. Nat Genet 37: 1082–1089, 2005 sequencing on the Illumina HiSEquation 2500 system. The RNA-seq 8. Batourina E, Gim S, Bello N, Shy M, Clagett-Dame M, Srinivas S, Costantini F, Mendelsohn C: Vitamin A controls epithelial/mesenchy- datasets comprised 17,000 paired-end reads. Reads were aligned to mal interactions through Ret expression. Nat Genet 27: 74–78, 2001 fi the mouse genome (mm9) using TopHat v2.05, and quanti ed with 9. Kobayashi A, Mugford JW, Krautzberger AM, Naiman N, Liao J, cufflinks v2.0.0, using ENSEMBL 65 genes. Subsequent analysis of McMahon AP: Identification of a multipotent self-renewing stromal differentially expressed genes was carried out using GeneSpring soft- progenitor population during mammalian kidney organogenesis. Stem – ware, filtering out duplicate reads, reads that align to more than one Cell Reports 3: 650 662, 2014 10. Hatini V, Huh SO, Herzlinger D, Soares VC, Lai E: Essential role of place in the genome, and removing genes that did not have at least stromal mesenchyme in kidney morphogenesis revealed by targeted three RPKM expression in two samples. The raw data are deposited in disruption of Winged Helix transcription factor BF-2. Genes Dev 10: GEO with accession number GSE97126. 1467–1478, 1996

542 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 532–544, 2018 www.jasn.org BASIC RESEARCH

11. Kobayashi A, Valerius MT, Mugford JW, Carroll TJ, Self M, Oliver G, 31. Di Giovanni V, Alday A, Chi L, Mishina Y, Rosenblum ND: Alk3 con- McMahon AP: Six2 defines and regulates a multipotent self-renewing trols nephron number and androgen production via lineage-specificef- nephron progenitor population throughout mammalian kidney devel- fects in intermediate mesoderm. Development 138: 2717–2727, 2011 opment. Cell Stem Cell 3: 169–181, 2008 32. Mendelsohn C, Ruberte E, LeMeur M, Morriss-Kay G, Chambon P: 12. Mugford JW, Yu J, Kobayashi A, McMahon AP: High-resolution gene Developmental analysis of the retinoic acid-inducible RAR-beta 2 expression analysis of the developing mouse kidney defines novel promoter in transgenic animals. Development 113: 723–734, 1991 cellular compartments within the nephron progenitor population. Dev 33. Yan J, Zhang L, Xu J, Sultana N, Hu J, Cai X, Li J, Xu PX, Cai CL: Smad4 Biol 333: 312–323, 2009 regulates ureteral smooth muscle cell differentiation during mouse 13. Blake J, Rosenblum ND: Renal branching morphogenesis: Morpho- embryogenesis. PLoS One 9: e104503, 2014 genetic and signaling mechanisms. Semin Cell Dev Biol 36: 2–12, 2014 34. Fogelgren B, Polgar N, Lui VH, Lee AJ, Tamashiro KK, Napoli JA, 14. Yu J, Carroll TJ, Rajagopal J, Kobayashi A, Ren Q, McMahon AP: A Walton CB, Zuo X, Lipschutz JH: Urothelial defects from targeted in- Wnt7b-dependent pathway regulates the orientation of epithelial cell activation of Exocyst Sec10 in mice cause ureteropelvic junction ob- division and establishes the cortico-medullary axis of the mammalian structions. PLoS One 10: e0129346, 2015 kidney. Development 136: 161–171, 2009 35. Motoyama J, Takabatake T, Takeshima K, Hui C: Ptch2, a second 15. Bohnenpoll T, Bettenhausen E, Weiss A-C, Foik AB, Trowe M-O, Blank mouse Patched gene is co-expressed with Sonic hedgehog. Nat Genet P, Airik R, Kispert A: Tbx18 expression demarcates multipotent pre- 18: 104–106, 1998 cursor populations in the developing urogenital system but is exclu- 36. Bai CB, Auerbach W, Lee JS, Stephen D, Joyner AL: Gli2, but not Gli1, is sively required within the ureteric mesenchymal lineage to suppress a required for initial Shh signaling and ectopic activation of the Shh renal stromal fate. Dev Biol 380: 25–36, 2013 pathway. Development 129: 4753–4761, 2002 16. Airik R, Bussen M, Singh MK, Petry M, Kispert A: Tbx18 regulates the 37.ParkHL,BaiC,PlattKA,MatiseMP,BeeghlyA,HuiCC,NakashimaM, development of the ureteral mesenchyme. JClinInvest116: 663–674, Joyner AL: Mouse Gli1 mutants are viable but have defects in SHH signaling 2006 in combination with a Gli2 mutation. Development 127: 1593–1605, 2000 17. Caubit X, Lye CM, Martin E, Coré N, Long DA, Vola C, Jenkins D, Garratt 38. Levinson RS, Batourina E, Choi C, Vorontchikhina M, Kitajewski J, AN, Skaer H, Woolf AS, Fasano L: Teashirt 3 is necessary for ureteral Mendelsohn CL: Foxd1-dependent signals control cellularity in the smooth muscle differentiation downstream of SHH and BMP4. Devel- renal capsule, a structure required for normal renal development. De- opment 135: 3301–3310, 2008 velopment 132: 529–539, 2005 18. Hooper JE, Scott MP: The Drosophila patched gene encodes a putative 39. Bagherie-Lachidan M, Reginensi A, Pan Q, Zaveri HP, Scott DA, Blencowe membrane protein required for segmental patterning. Cell 59: 751– BJ, Helmbacher F, McNeill H: Stromal Fat4 acts non-autonomously with 765, 1989 Dchs1/2 to restrict the nephron progenitor pool. Development 142: 19. Rohatgi R, Milenkovic L, Scott MP: Patched1 regulates hedgehog sig- 2564– 2573, 2015 naling at the primary cilium. Science 317: 372–376, 2007 40. Mugford JW, Sipilä P, McMahon JA, McMahon AP: Osr1 expression 20. Nieuwenhuis E, Hui CC: Hedgehog signaling and congenital malfor- demarcates a multi-potent population of intermediate mesoderm that mations. Clin Genet 67: 193–208, 2005 undergoes progressive restriction to an Osr1-dependent nephron progen- 21. Kim J, Kim P, Hui CC: The VACTERL association: Lessons from the Sonic itor compartment within the mammalian kidney. Dev Biol 324: 88–98, 2008 hedgehog pathway. Clin Genet 59: 306–315, 2001 41. Taguchi A, Kaku Y, Ohmori T, Sharmin S, Ogawa M, Sasaki H, 22. Persson M, Stamataki D, te Welscher P, Andersson E, Böse J, Rüther U, Nishinakamura R: Redefining the in vivo origin of metanephric nephron Ericson J, Briscoe J: Dorsal-ventral patterning of the spinal cord re- progenitors enables generation of complex kidney structures from quires Gli3 transcriptional repressor activity. Genes Dev 16: 2865– pluripotent stem cells. Cell Stem Cell 14: 53–67, 2014 2878, 2002 42. Buck A, Kispert A, Kohlhase J: Embryonic expression of the murine 23. Buttitta L, Mo R, Hui C-C, Fan C-M: Interplays of Gli2 and Gli3 and their homologue of SALL1, the gene mutated in Townes–Brocks syndrome. requirement in mediating Shh-dependent sclerotome induction. De- Mech Dev 104: 143–146, 2001 velopment 130: 6233–6243, 2003 43. Inoue S, Inoue M, Fujimura S, Nishinakamura R: A mouse line ex- 24. Blake J, Hu D, Cain JE, Rosenblum ND: Urogenital development in pressing Sall1-driven inducible Cre recombinase in the kidney mes- Pallister-Hall syndrome is disrupted in a cell-lineage-specificmannerby enchyme. Genesis 48: 207–212, 2010 constitutive expression of GLI3 repressor. Hum Mol Genet 25: 437– 44. Maezawa Y, Binnie M, Li C, Thorner P, Hui CC, Alman B, Taketo MM, 447, 2016 Quaggin SE: A new Cre driver mouse line, Tcf21/Pod1-Cre, targets 25. Cain JE, Islam E, Haxho F, Blake J, Rosenblum ND: GLI3 repressor metanephric mesenchyme. PLoS One 7: e40547, 2012 controls functional development of the mouse ureter. JClinInvest121: 45. Jeong J, Mao J, Tenzen T, Kottmann AH, McMahon AP: Hedgehog 1199–1206, 2011 signaling in the neural crest cells regulates the patterning and growth of 26. Cain JE, Islam E, Haxho F, Chen L, Bridgewater D, Nieuwenhuis E, Hui facial primordia. Genes Dev 18: 937–951, 2004 C, Rosenblum ND: GLI3 repressor controls nephron number via regu- 46. Hernández-Bejarano M, Gestri G, Spawls L, Nieto-López F, Picker A, lation of Wnt11 and Ret in ureteric tip cells. PLoS One 4: e7313, 2009 Tada M, Brand M, Bovolenta P, Wilson SW, Cavodeassi F: Opposing 27. Hu MC, Mo R, Bhella S, Wilson CW, Chuang P-T, Hui C-C, Rosenblum Shh and Fgf signals initiate nasotemporal patterning of the zebrafish ND: GLI3-dependent transcriptional repression of Gli1, Gli2 and kidney retina. Development 142: 3933–3942, 2015 patterning genes disrupts renal morphogenesis. Development 133: 47. Yallowitz AR, Hrycaj SM, Short KM, Smyth IM, Wellik DM: Hox10 genes 569–578, 2006 function in kidney development in the differentiation and integration of 28. Yu J, Carroll TJ, McMahon AP: Sonic hedgehog regulates proliferation the cortical stroma. PLoS One 6: e23410, 2011 and differentiation of mesenchymal cells in the mouse metanephric 48. Xavier GM, Seppala M, Barrell W, Birjandi AA, Geoghegan F, Cobourne kidney. Development 129: 5301–5312, 2002 MT: Hedgehog receptor function during craniofacial development. 29. Goodrich LV, Milenkovic L, Higgins KM, Scott MP: Altered neural cell Dev Biol 415: 198–215, 2016 fates and medulloblastoma in mouse patched mutants. Science 277: 49. Okada A, Charron F, Morin S, Shin DS, Wong K, Fabre PJ, Tessier- 1109–1113, 1997 Lavigne M, McConnell SK: Boc is a receptor for sonic hedgehog in the 30. Kobayashi A, Kwan K-M, Carroll TJ, McMahon AP, Mendelsohn CL, guidance of commissural axons. Nature 444: 369–373, 2006 Behringer RR: Distinct and sequential tissue-specific activities of the 50. Pearse RV 2nd, Vogan KJ, Tabin CJ: Ptc1 and Ptc2 transcripts provide LIM-class homeobox gene Lim1 for tubular morphogenesis during distinct readouts of Hedgehog signaling activity during chick em- kidney development. Development 132: 2809–2823, 2005 bryogenesis. Dev Biol 239: 15–29, 2001

J Am Soc Nephrol 29: 532–544, 2018 Hedgehog and Ureteropelvic Junction Obstruction 543 BASIC RESEARCH www.jasn.org

51. Zaphiropoulos PG: Distinct roles of PTCH2 splice variants in Hedgehog 58. Tripathi P, Wang Y, Casey AM, Chen F: Absence of canonical Smad signalling. Biochem J 378: 325–334, 2004 signaling in ureteral and bladder mesenchyme causes ureteropelvic 52. Lee Y, Miller HL, Russell HR, Boyd K, Curran T, McKinnon PJ: Patched2 junction obstruction. JAmSocNephrol23: 618–628, 2012 modulates tumorigenesis in patched1 heterozygous mice. Cancer Res 59. Xu J, Wong EYM, Cheng C, Li J, Sharkar MTK, Xu CY, Chen B, Sun J, 66: 6964–6971, 2006 Jing D, Xu P-X: Eya1 interacts with Six2 and Myc to regulate expansion 53. Nieuwenhuis E, Motoyama J, Barnﬁeld PC, Yoshikawa Y, Zhang X, Mo of the nephron progenitor pool during nephrogenesis. Dev Cell 31: R, Crackower MA, Hui CC: Mice with a targeted mutation of patched2 434–447, 2014 are viable but develop alopecia and epidermal hyperplasia. Mol Cell 60. Ellis T, Smyth I, Riley E, Graham S, Elliot K, Narang M, Kay GF, Wicking Biol 26: 6609–6622, 2006 C, Wainwright B: Patched 1 conditional null allele in mice. Genesis 36: 54. Zhulyn O, Nieuwenhuis E, Liu YC, Angers S, Hui C-C: Ptch2 shares 158–161, 2003 overlapping functions with Ptch1 in Smo regulation and limb devel- 61. Bock NA, Nieman BJ, Bishop JB, Mark Henkelman R: In vivo multiple- opment. Dev Biol 397: 191–202, 2015 mouse MRI at 7 Tesla. Magn Reson Med 54: 1311–1316, 2005 55. Gorlin RJ: Nevoid basal-cell carcinoma syndrome. Medicine (Balti- 62. Nieman BJ, Bock NA, Bishop J, Chen XJ, Sled JG, Rossant J, more) 66: 98–113, 1987 Henkelman RM: Magnetic resonance imaging for detection and anal- 56. Johnston JJ, Olivos-Glander I, Killoran C, Elson E, Turner JT, Peters KF, ysis of mouse phenotypes. NMR Biomed 18: 447–468, 2005 Abbott MH, Aughton DJ, Aylsworth AS, Bamshad MJ, Booth C, Curry 63. Nieman BJ, Wong MD, Henkelman RM: Genes into geometry: Imag- CJ, David A, Dinulos MB, Flannery DB, Fox MA, Graham JM, Grange ing for mouse development in 3D. Curr Opin Genet Dev 21: 638–646, DK, Guttmacher AE, Hannibal MC, Henn W, Hennekam RC, Holmes LB, 2011 Hoyme HE, Leppig KA, Lin AE, Macleod P, Manchester DK, Marcelis C, 64. Hui CC, Joyner AL: A mouse model of greig cephalopolysyndactyly Mazzanti L, McCann E, McDonald MT, Mendelsohn NJ, Moeschler JB, syndrome: The extra-toesJ mutation contains an intragenic deletion of Moghaddam B, Neri G, Newbury-Ecob R, Pagon RA, Phillips JA III, the Gli3 gene. Nat Genet 3: 241–246, 1993 Sadler LS, Stoler JM, Tilstra D, Walsh Vockley CM, Zackai EH, Zadeh 65. Pinaud R, Mello CV, Velho TA, Wynne RD, Tremere LA: Detection of TM, Brueton L, Black GCM, Biesecker LG: Molecular and clinical anal- two mRNA species at single-cell resolution by double-ﬂuorescence in yses of Greig cephalopolysyndactyly and Pallister-Hall syndromes: situ hybridization. Nat Protoc 3: 1370–1379, 2008 Robust phenotype prediction from the type and position of GLI3 mutations. Am J Hum Genet 76: 609–622, 2005 57. Chen H, Ji H-Y, Yang Y: The expression of Gli3 and Teashirt3 in the stenotic tissue of congenital pelvi-ureteric junction obstruction in chil- This article contains supplemental material online at http://jasn.asnjournals. dren. Int J Med Sci 13: 412–417, 2016 org/lookup/suppl/doi:10.1681/ASN.2017050482/-/DCSupplemental.

544 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 532–544, 2018