Downregulating Hedgehog Signaling Reduces Renal Cystogenic Potential of Mouse Models

BASIC RESEARCH www.jasn.org

† † Pamela V. Tran,* George C. Talbott,* Annick Turbe-Doan,* Damon T. Jacobs, † † † Michael P. Schonfeld, Luciane M. Silva, Anindita Chatterjee, Mary Prysak,* † ‡ Bailey A. Allard, and David R. Beier*

*Genetics Division, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts; †Department of Anatomy and Cell Biology and the Kidney Institute, University of Kansas Medical Center, Kansas City, Kansas; and ‡Center for Developmental Biology and Regenerative Medicine, Seattle Children’s Research Institute, Seattle, Washington

ABSTRACT Renal cystic diseases are a leading cause of renal failure. Mutations associated with renal cystic diseases reside in genes encoding proteins that localize to primary cilia. These cystoproteins can disrupt ciliary structure or cilia-mediated signaling, although molecular mechanisms connecting cilia function to renal cystogenesis remain unclear. The ciliary gene, Thm1(Ttc21b), negatively regulates Hedgehog signaling and is most commonly mutated in ciliopathies. We report that loss of murine Thm1 causes cystic kidney disease, with persistent proliferation of renal cells, elevated cAMP levels, and enhanced expression of Hedgehog signaling genes. Notably, the cAMP-mediated cystogenic potential of Thm1-null kidney explants was reduced by genetically deleting Gli2, a major transcriptional activator of the Hedgehog pathway, or by culturing with small molecule Hedgehog inhibitors. These Hedgehog inhibitors acted independently of protein kinase A and Wnt inhibitors. Furthermore, simultaneous deletion of Gli2 atten- uated the renal cystic disease associated with deletion of Thm1. Finally, transcripts of Hedgehog target genes increased in cystic kidneys of two other orthologous mouse mutants, jck and Pkd1, and Hedgehog inhibitors reduced cystogenesis in jck and Pkd1 cultured kidneys. Thus, enhanced Hedgehog activity may have a general role in renal cystogenesis and thereby present a novel therapeutic target. J Am Soc Nephrol 25: 2201–2212, 2014. doi: 10.1681/ASN.2013070735

Cystic kidney disease represents a wide disease Primary cilia project from the apical surface of spectrum characterized by fluid-filled cysts, which renal tubular epithelial cells and have been proposed destroy surrounding renal parenchyma. The spec- tofunction asmechanosensors, bending in response trum affects adults in the most common life- to urine flow and initiating signaling cascades.4 In- threatening hereditary disease, autosomal dominant traflagellar transport (IFT), the bidirectional trans- polycystic kidney disease (ADPKD), as well as chil- port of multiprotein complexes along the ciliary dren in the form of autosomal recessive PKD microtubule-based core, builds and maintains cilia (reviewed by Torres and Harris1). Infantile or juvenile and is integral to signaling.5 Anterograde IFT cystic kidney disease is also manifested in disease syndromes caused by mutation of ciliary genes that are collectively termed ciliopathies. These include Received July 15, 2013. Accepted January 9, 2014. nephronophthisis, Bardet-Biedl syndrome, Meckel- Published online ahead of print. Publication date available at Gruber syndrome, Jeune syndrome, and Joubert www.jasn.org. syndrome.2 PKD genes encode proteins that also Correspondence: Dr. Pamela V. Tran, Department of Anatomy 3 fi localize to primary cilia. Together, these ndings and Cell Biology and The Kidney Institute, University of Kansas have led to the proposition that perturbed cilia func- Medical Center, 3901 Rainbow Boulevard, MS #3038, Kansas tion may be a unifying etiologic basis for cystic kid- City, KS 66160. Email: [email protected] ney disease. Copyright © 2014 by the American Society of Nephrology

J Am Soc Nephrol 25: 2201–2212, 2014 ISSN : 1046-6673/2510-2201 2201 BASIC RESEARCH www.jasn.org delivers proteins to the ciliary distal tip and is mediated largely complex B genes, including a hypomorphic allele of Ift8819 by the kinesin motor and IFT complex B proteins, while ret- and renal-specificablationofIft20 or of Kif3a, which rograde IFTreturns proteins to the base and is mediated by the encodes a kinesin subunit, causes renal cysts.20–22 Because cytoplasmic dynein-2 motor and IFT complex A proteins. loss of complex B proteins results in loss of Hh response, In mice, loss of IFT proteins causes cilia structural defects, this may account in part for why Hh signaling has not been which affect regulation of Hh signaling,6 a pathway fundamental studied extensively in cystic kidney disease. to proper development and tissue maintenance.7 Three mam- One report has characterized the role of IFT complex A malian Hh ligands—Sonic Hh, Indian Hh, and Desert Hh— proteins in renal function; kidney-specific deletion of Ift140 can initiate signaling upon binding the 12-transmembrane re- caused renal cysts.23 However, THM1 mutations have been ceptor, Patched, in the cilium. Upon being bound, Patched reported in 5% of patients with ciliopathies, rendering exits the cilium, enabling ciliary translocation of the seven- THM1 the most commonly mutated gene in ciliopathies24 transmembrane signal transducer, Smoothened (SMO).8,9 and thereby implicating this gene in renal disease. We inves- Within the cilium, SMO is activated, which culminates in ac- tigated the consequences of Thm1 deletion in renal cystogen- tivation of full-length glioblastoma (GLI1, GLI2, GLI3) tran- esis using the Thm1aln/aln developmental mutant and a newly scription factors. In most tissues, GLI2 acts as the primary developed Thm1 conditional knockout mouse. We further transcriptional activator (reviewed by Eggenschwiler and examined a direct role of Hh signaling in this and two other Anderson10), although in certain tissues, GLI3 activator orthologous cystic kidney disease mouse mutants, in part by (GLI3A) or GLI3 repressor (GLI3R), which is formed by examining effects of Hh inhibitors on cystogenic potential of cleavage of the full-length GLI3 protein, plays a predominant mutant kidneys cultured with cAMP, an assay proposed as a role because of redundancy and/or tissue specificity of the GLI general model of cystic kidney disease25 and useful for testing proteins.11,12 Nonetheless, the balance between activity of GLI efficacy of potential pharmaceutical therapies.26 activators and GLI3 repressor determines level of Hh signaling output within a cell.10,13 Despite the connection between primary cilia and Hh RESULTS signaling, a role for Hh signaling in cystic kidneydisease has not been studied extensively. Yet the few studies implicating Hh Loss of THM1 Causes Renal Cysts signaling in cystic kidney disease are quite compelling. In the To explore a role for Thm1 deletion in renal cystogenesis, we first, 50% of Shh-null mouse embryos formed a single, ectopic, examined the kidneys in Thm1aln/aln mice during late embryo- dysplastic kidney.14 In the second, loss of the transcription genesis. At E16.5 and P0, the last day on which Thm1aln/aln factor, Glis2, a member of the Kruppel-like C2H2 zinc finger mutants survive, histologic analysis of the kidneys revealed protein subfamily, which includes the GLI proteins, resulted in cystic dilations of glomeruli and surrounding tubules (Figure nephronophthisis in humans and mice.15 Expression profiling 1A). Coimmunostaining of kidney sections for Lotus tetrago- 2 2 of Glis2 / kidneys revealed an upregulation of Gli1, a direct nolobus lectin (LTL) and Na+K+ ATPase (a6F) demonstrated target of the pathway, suggesting upregulated Hh signaling. that Thm1aln/aln renal tubular dilations originate in proximal Finally, a study examining the effects of corticosteroid over- tubules and ascending loops of Henle (Figure 1B). exposure on metanephric development found that adding hy- Because Thm1aln/aln mutants die shortly after birth, we gen- drocortisone to a metanephros organ culture unexpectedly erated Thm1 conditional knockout (cko) mice to analyze the caused renal cysts as well as upregulated Ihh transcripts.16 role of THM1 in maintaining renal tubular integrity. Ubiqui- Addition of Hh inhibitor, cyclopamine, reduced hydrocortisone- tous deletion of Thm1 at E17.5 resulted in cystic kidney disease induced cysts without affecting organ growth, implicating in- in 6-week-old Thm1 cko mice (Figure 1C), with elevated ratios creased Hh signaling in this mechanism of cystogenesis. of percentage kidney weight to body weight (%KW/BW) and We identified an N-ethyl-N-nitrosourea–derived, develop- BUN levels (Figure 1, D and E). Levels of cAMP were also mental mouse mutant, alien (aln),asthefirst IFT complex A higher in Thm1 cko cystic kidneys (Figure 1F), similar to other mammalian mutant.17 This defect impaired retrograde IFT, mouse models, such as the cpk autosomal recessive PKD causing accumulation of proteins in bulb-like structures at the model and the Pkd1 orthologous ADPKD model. Thm1 cko ciliary distal tips and increased Hh signaling mediated by en- renal cysts labeled positively for LTL, Tamms–Horsfall protein hanced GLI2 and GLI3 activator transcriptional activities re- (THP), and Dolichus biflorus agglutinin (DBA), indicating that sulting in limb and neural tube patterning defects. The genetic cysts originate from proximal tubules, loops of Henle, and lesion in alien results in the absence of Thm1 (TPR-containing collecting ducts, respectively (Figure 2, B–D). In kidneys of Hh modulator 1; also termed Ttc21b), an ortholog of Chlamy- Thm1 cko mice, primary cilia were stunted and showed accu- domonas complex A protein, IFT139. Another IFT complex A mulation of IFT88 protein in bulb-like structures at the distal mouse mutant, sob/IFT122, also exhibits upregulated Hh sig- tips (Figure 3, A–C), characteristic of the IFT complex A mu- naling18; this contrasts with most IFTand cilia mutants, which tant phenotype. Scanning electron microscopy revealed that in lose the capacity to respond to the Hh signal (reviewed by collecting ducts, Thm1 cko primary cilia length is 60% that of Eggenschwiler and Anderson10). Deficiency of murine IFT wild-type (wt) (Figure 3D).

2202 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2201–2212, 2014 www.jasn.org BASIC RESEARCH

served markedly higher levels of PCNA+ cells in Thm1 cko cystic kidneys (Figure 2A). A similar percentage of PCNA+ cells was observed across Thm1 cko LTL, THP, and cortical DBA–labeled tubules (Figure 2, B–E), suggesting that increased proliferation did not correlate with a speciﬁc renal tubule. Todetermine when cystic kidney disease initiates, we examined Thm1 cko kidneys at earlier time points and observed that by P15, dilations of proximal tubules and loops of Henle in the cortex were present (Figure 2, G and H, Supplemental Figure 1). Because proliferation has been proposed as an initiating factor in renal cystogenesis, we immunostained P15 and P20 kidneys for PCNA. At P15, the percentage of PCNA+ cells was similar between wt and Thm1 cko kidneys (Figure 2J). From P15 to P20, PCNA+ cells were markedly reduced in wt renal medulla, but not in Thm1 cko medulla (Supplemental Figure 2). These data suggest that the ciliary defect may de- lay maturation of the kidney, prolonging the kidney in a developing state. The developmental state of a kidney has been proposed to have a signiﬁcant role in predisposing to cystogenesis because loss of cystoproteins before P12–P14 results in se- vere cystic kidney disease, while loss of cystoproteins after this window causes very mild cystic kidney disease evident only after 6 months.27,28 To determine whether THM1 loss is similarly sensitive to the developmental state, we ablated Thm1 at 5 weeks of age. As with all other cystoproteins reported, Thm1 deletion at a mature stage did not result in cysts after 3 months (Supplemental Figure 3).

Genetic and Pharmacologic Inhibition Figure 1. Loss of THM1 causes renal cysts. (A) Hematoxylin and eosin images of E16.5 aln/aln of Hh Signaling Reduces Renal and P0 wt and Thm1 kidneys. Black dotted squares on first-row panels are shown aln/aln Cystogenic Potential of Thm1 at higher magnification than in the second-row panels. Arrows point to dilated tubules aln/aln aln/aln Cultured Kidneys in Thm1 kidneys. (B) Marker analysis of wt and Thm1 nephron segments. fl Kidney sections were co-immunostained for fluorescein-conjugated LTL (green) and Previously, we observed that the oor plate aln/aln for Na+K+ adenosine triphosphatase (red). (C) Whole-mount and histologic sections of of the Thm1 neural tube is dorsally 6-week-old wt and Thm1 cko kidneys. (D) %KW/BW, (E) BUN levels, and (F) cAMP expanded, reflecting increased GLI2 activ- levels of 6-week-old wt and Thm1 cko kidneys. Bars represent mean6SEM of 4 wt ity.17 Consistent with this, Gli2 deletion in mice and 3 Thm1 cko mice. *P,0.05. Thm1aln/aln embryos reversed the effect of the aln mutation in the floor plate, estab- lishing THM1 as a negative regulator of Hh Acellular hallmark of PKD is increased proliferation of cyst- signaling. We queried whether Gli2 deficiency could similarly lining epithelial cells. Using proliferating cell nuclear antigen rescue the Thm1aln/aln kidney phenotype and examined Thm1, (PCNA) immunostaining to evaluate proliferation, we ob- Gli2 genetic interaction using an embryonic kidney culture

J Am Soc Nephrol 25: 2201–2212, 2014 Hedgehog Signaling in Renal Cysts 2203 BASIC RESEARCH www.jasn.org

levels between wt and Thm1aln/aln kidneys (Supplemental Figure 4). Topharmacologically replicate the Thm1, Gli2 genetic interaction, we assessed the effects of small molecule GLI antagonist, Gant61, on wt and Thm1aln/aln kidneys cultured in the presence of 8-Br-cAMP. At a concentration of 10 mM, Gant61 prevented tubular dilations in Thm1aln/aln cultured kidneys (Figure 4B). To further explore the role of Hh signaling in renal cystogenesis, we also examined the effect of the small molecule SMO antagonist, Sant2. Similarly to Gant61, 5 mM Sant2 prevented cysts in mutant kidneys cultured with 8-Br-cAMP (Figure 4B). Thus, Hh inhibitors acting at two different steps in the signaling pathway could counteract the effects of cAMP and abrogate cyst formation. To assess a functional role for GLI2 in renal cystogenesis in vivo,wedeletedThm1 and Gli2 simultaneously at E17.5. At 6 weeks of age, Thm1; Gli2 double cko mice showed milder cystic kidney disease than did Thm1 cko mice (Figure 4C), with reduced KW/BW ratios (Figure 4D) and BUN levels (Figure 4E), supporting a role for increased GLI2 activity in Thm1 cko renal cystogenesis in vivo.

Small Molecule Hh Inhibitors Do Not Act Through Protein Kinase A or Wnt Signaling to Reduce Cystogenic Potential Figure 2. Proliferation is increased in Thm1 cko cystic kidneys. (A) Immunostaining of A role for the Hh pathway in cAMP signaling 6-week-old wt and Thm1 cko kidneys for PCNA alone or in combination with (B) LTL, leading to cyst formation is unknown. Levels (C) THP, or (D) DBA. (E) Proliferation rates of 6-week-old wt and Thm1 cko kidneys. of cAMP are elevated in PKD-affected kid- + Number of PCNA cells was divided over number of DAPI cells for a particular tubule. neys,29–31 which leads to activation of protein (F) Immunostaining of P15 wt and Thm1 cko kidneys for PCNA alone or in combi- kinase A (PKA). This in turn phosphorylates nation with (G) LTL, (H) THP, or (I) DBA. (J) Proliferation rates of P15 wt and Thm1 cko and activates cystic ﬁbrosis transmembrane kidneys. All scale bars represent 500 mm. n=3 wt; n=3 Thm1 cko for each time point. 2 conductance regulator (CFTR), driving Cl ions out into the lumen, which is accompa- 2 2 assay.25 E13.5 wt, Thm1aln/aln and Thm1aln/aln,Gli2 / kid- nied by ﬂuid secretion, a major cellular abnormality in PKD. neys were cultured in the presence of cystogenic reagent, 8- Accordingly, addition of the PKA inhibitor H89 or of the CFTR Br-cAMP, for 4 days. Thm1aln/aln kidneys showed a 3-fold inhibitor 172 to kidneys cultured with 8-Br-cAMP abrogated greater cystogenic potential than wt (Figure 4A), consistent cyst formation.25 To determine whether the effect of Hh inhib- with a role for THM1 loss in renal cystogenesis. Importantly, itors was mediated through this pathway, we examined the effect 2 2 Thm1aln/aln,Gli2 / kidneys showed a 2-fold lower cystogenic of the PKA inhibitor H89 relative to Gant61 and Sant2 on wt potential than Thm1aln/aln kidneys, suggesting that increased CD1 kidneys cultured in the presence of 8-Br-cAMP. We found GLI2 activity mediates Thm1aln/aln renal cystogenesis. that the Hh inhibitors prevented cysts to an extent similar to that Todetermine the contribution of GLI2 and GLI3 proteins to of H89 (Figure 5, A and B). To examine the effect of the Hh Thm1aln/aln renal cystogenesis, we assayed protein extracts of inhibitors on PKA activity, we analyzed treated kidneys for levels E13.5 Thm1aln/aln cultured kidneys and E18.5 Thm1aln/aln kid- of cAMP response element-binding protein (CREB) and phos- neys for GLI1, GLI2, and GLI3. However, examination of these pho-CREB (P-CREB), a target of PKA. As expected, we found whole kidney extracts did not show obvious differences in GLI high levels of P-CREB in kidneys cultured with cAMP, while P-

2204 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2201–2212, 2014 www.jasn.org BASIC RESEARCH

Sant2-treated kidneys, GLI1 was markedly reduced, GLI2 was slightly reduced, and GLI3R was increased (Figure 5C). These changes in GLI protein levels are consistent with downregulation of the Hh pathway at the level of SMO. Increased GLI3R is also consistent with the notion that higher levels of PKA activity enable enhanced GLI3 processing. Gant61, which inhibits at the level of GLI1- and GLI2-mediated transcriptional activity,32 did not show changes in GLI protein levels (Figure 5C). Interestingly, GLI1 and GLI3 activator levels were reduced in protein extracts of kidneys treated with the PKA inhibitor H89 (Figure 5C). Thus, in addition to H89 inhibiting the CFTR molecule responsi- ble for fluid secretion,25 H89 may also partially inhibit Hh activity. In contrast to Hh signaling, Wnt signaling has been studied extensively in cystic kidney disease, with the suggestion that overactive canonical Wnt signaling contributes to renal cystogenesis.22,33 Disruption of IFT in 2 2 Kif3a / mutants or in cells of hypomorphic complex B IFT88orpk/orpk mutants can result in upregulated canonical Wnt signaling.34 However, the role of cilia in mediating Wnt signaling remains controversial because IFT mutant embryos show Hh mutant phenotypes rather than phenotypes characteristic of misregulated Wnt signaling (reviewed by Eggenschwiler and Anderson10). Currently, a role for an IFT complex A protein in Wnt signaling has not been reported, and the con- trasting Hh phenotype of Thm1aln/aln from Figure 3. Thm1 cko renal primary cilia show IFT complex A mutant phenotype. (A) IFT complex B and motor mouse mutants Immunostaining of wt and Thm1 cko renal primary cilia for acetylated a-tubulin (red) further compelled us to examine the effect and IFT88 (green). Scale bar represents 25 mm. (B) Scanning electron micrographs of of THM1 on canonical Wnt signaling. We aln/aln wt and Thm1 cko collecting ducts (original magnification, 32000). Scale bar repre- generated wt and Thm1 mice that sents 15 mm. (C) Scanning electron micrographs of wt and Thm1 cko collecting duct harbor the b-catenin activated transgene primary cilia (original magnification, 315,000). Scale bar represents 3 mm. (D) Lengths (BAT)-gal Wnt reporter allele and created of wt and Thm1 cko collecting duct primary cilia. Bars represent mean6SEM of n=23 mouse embryonic fibroblast (MEF) lines. 28 cilia from three wt mice and n=25 cilia from two Thm1 cko mice. *P,5310 . In the presence of Wnt3a ligand, Thm1aln/aln MEF showed higher b-galactosidase activity than wt cells (Figure 5D). Thus, cells CREB levels were dramatically reduced in extracts of kid- lacking THM1 respond to Wnt ligand in a fashion similar neys cultured with both cAMP and H89. In contrast, P-CREB to that of cells deficient in Kif3a and IFT88. levels were not reduced in kidneys treated with Gant61 or Sant2 We next examined whether small molecule Wnt inhibitors (Figure 5C). Thus these data suggest that, despite reduced cyst might reduce Thm1aln/aln renal cystogenic potential. In cul- formation, Gant61- and Sant2-treated kidneys have high levels tured cells, nanomolar amounts of inhibitor of Wnt pro- of PKA activity. duction-2 (IWP-2) and inhibitor of Wnt response-1 (IWR-1) PKA activity is required to process full-length GLI3 proteins inhibited production of the Wnt ligand and the Wnt re- to form GLI3 repressor (GLI3R).11 To determine whether the sponse, respectively, while 10 mM IWR-1 suppressed tailfin Hh inhibitors were acting by increasing GLI3R levels, we regeneration in zebrafish, which requires canonical Wnt ac- examined GLI protein levels in treated cultured kidneys. In tivity.35 In our initial analysis, 10 mM, 20 mM, and 40 mMof

J Am Soc Nephrol 25: 2201–2212, 2014 Hedgehog Signaling in Renal Cysts 2205 BASIC RESEARCH www.jasn.org

kidneys as well (Figure 4B, Supplemental Figure 6 [Sant1]), suggesting a general role for Hh activity in cAMP-induced cystogenesis. We therefore questioned whether the Hh pathway could be im- plicated in other models of cystic kidney disease. Using quantitative RT-PCR, we observed increased expression of Ptch2, Gli1, Gli2, and Gli3 in kidneys of 6-week- old Thm1 cko mice (Figure 6A). Similarly, we examined Gli expression in jck and Pkd1 mutants. The jck mutant phenotypically resembles ADPKD29 and carries a mutation in the Nek8 kinase.36 Mutations in NEK8 have been identiﬁed in patients with nephronophthisis-937 and in the Lewis PKD rat.38 At 7 weeks of age, jck/jck mice showed a mean (6SEM) %KW/BW of 4.960.33 (n=11) compared with 1.360.04 in wt mice (n=19). Quantitative RT-PCR using RNA lysates from 7-week- old wt and jck/jck kidneys showed increased expression of Hh target genes, Ptch2, Gli1, and Gli3, in jck/jck cystic kidneys (Figure 6B). Using a conditional allele of Pkd1, together Figure 4. Genetic or pharmacologic inhibition of Hh signaling reduces cystogenesis in with a ubiquitous tamoxifen-inducible Cre aln/aln aln/aln Thm1 kidney explants and Thm1 cko kidneys. (A) E13.5 wt, Thm1 ,and recombinase, we deleted Pkd1, ortholog of aln/aln 2/2 m Thm1 ,Gli2 kidneys were incubated with 100 M of 8-bromo-cAMP for 4 the most commonly mutated gene in days. Quantitative assessment of kidney images showed a 3-fold greater cystogenic ADPKD, at P2. At P23, %KW/BW of wt potential in Thm1aln/aln kidneys than wt and a 2-fold decrease in cystogenic potential 2 2 (n=11) and Pkd1 cko mice (n=4) were of Thm1aln/aln,Gli2/ kidneys relative to Thm1aln/aln kidneys. Bars represent 2 2 6 6 mean6SEM of 4 wt, 6 Thm1aln/aln,and8Thm1aln/aln,Gli2 / kidneys from two ex- 1.80 0.21 and 7.60 1.55, respectively. periments. (B) E13.5 wt and Thm1aln/aln kidneys were cultured in presence of 100 mM Quantitative RT-PCR analysis showed in- 8-bromo-cAMP, in combination with 10 mM Gant61, 5 mM Sant2, or control DMSO for creased expression of Gli1, Gli2, and Gli3 4 days. Graphs represent quantitative assessment of kidney images following 4-day in Pkd1 mutant kidneys (Figure 6C), sug- culture. Gant61 or Sant2 reduces cystogenic potential of both normal and aln kidneys. gesting upregulated Hh signaling. Bars represent mean6SEM of 11 control and 7 Thm1aln/aln kidneys from 3 experiments Next we examined whether Hh inhib- for Gant61, and 9 control and 3 Thm1aln/aln kidneys from 3 experiments for Sant2. (C) itors could reduce cystogenic potential of Hematoxylin and eosin staining of kidney sections of 6-week-old wt, Thm1 cko, and cultured jck/jck mutant kidneys, which Thm1;Gli2 double cko mice. (D) KW/BW fold difference of 24 wt, 11 Thm1 cko, and 7 have shown increased cyst formation in Thm1;Gli2 double cko mice. (E) BUN of 8 wt, 6 Thm1 cko, and 3 Thm1;Gli2 double the cAMP kidney explant assay.26 Similarly, cko mice. *P,0.05; **P,0.005; ***P,0.0005. we observed 2- and 3-fold higher cystogenic potentials in cultured jck/+ and jck/jck kidneys, compared with wt (Figure 7A). Im- IWR-1orIWP-2didnotreduce,butrather increased,tubular portantly, this increased cystogenesis was markedly reduced by dilations in CD1 kidneys cultured with 8-Br-cAMP (Supple- treatment with Gant61 or Sant2. mental Figure 5). Moreover, 40 mM of IWR-1 did not reduce To determine the effect of Hh inhibitors on cultured Pkd1 cystogenic potential of cultured Thm1aln/aln kidneys (Figure mutant kidneys, we used the N-ethyl-N-nitrosourea–derived 5, E and F). Pkd1m1Bei/m1Bei mouse mutant, which carries a missense mu- 2 2 tation in Pkd1.39 Like Pkd1 / mutants, Pkd1m1Bei/m1Bei em- Small Molecule Hh Inhibitors Reduce Renal Cystogenic bryos are edemic with renal tubular dilations and die during m1Bei/m1Bei Potential of Cultured jck/jck and Pkd1 late development. In the cAMP cystogenic assay, the Pkd1m1Bei Kidneys allele increases cystogenic potential.25 We observed 2- and 3- Intriguingly, Hh inhibitors Gant61 and Sant2 prevented fold higher cystogenic potentials in cultured Pkd1m1Bei/+ and tubular dilations not only in Thm1aln/aln kidneys but in wt Pkd1m1Bei/m1Bei kidneys, respectively, relative to wt (Figure 7B).

2206 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2201–2212, 2014 www.jasn.org BASIC RESEARCH

As seen with Thm1aln/aln and jck/jck kidneys, treatment with Gant61 or Sant2 dramatically reduced cystogenic potential of cultured Pkd1m1Bei/m1Bei kidneys.

DISCUSSION

In this report, we demonstrate a role for THM1 loss inrenalcystogenesis.TheThm1aln allele increases renal cystogenic potential and causes renal tubular dilations in homozygous embryos. This renal phenotype combined with the polydactyly and exencephaly of Thm1aln/aln mutants17 models the clinical tri- al of Meckel-Gruber syndrome.40 The clinical relevance of THM1 is underscored by the finding that THM1 contributes the most pathogenic alleles in patients with infantile and pediatric ciliopathies.24 Further, Thm1 ablation during late embryogenesis results in cystic kidney disease in adulthood, show- ing that THM1 loss can cause both pediatric and adult forms of the cystic kidney disease spectrum. Because THM1 negatively regulates Hh signaling,17 we investigated a role for increased Hh signaling in renal cystogenesis, which has been largely unexplored. Reduc- ing Gli2 dosage or culturing with Hh inhibitors reduced renal cystogenic potential in Thm1aln/aln kidneys and deleting Gli2 at- tenuated cystic kidney disease of Thm1 cko mice, implicating a causal role for increased Hh signaling in Thm1 renal cystogenesis. In mice, loss of GLI2 does not result in a renal phenotype, but increased Figure 5. Small molecule Hh inhibitors act independently of PKA and Wnt inhibitors to reduce cystogenic potential of kidney explants. (A) E13.5 CD1 kidneys were cultured in expression of GLI3R in a mouse model of the presence of 100 mM 8-bromo-cAMP, in combination with DMSO, 10 mM Gant61, 5 Pallister-Hall syndrome causes nonob- mMSant2,or10mM of PKA inhibitor H89 for 4 days. Untreated contralateral kidneys structive hydronephrosis.12 These findings were used as control. (B) Quantitative assessment of kidney images following 4-day suggest redundancy among the GLI activa- culture. Bars represent mean6SEM; n=6 kidneys for each small molecule. (C) Western tors in the kidney and highlight the es- blot for P-CREB, CREB, GLI1, GLI2, and GLI3. Cultured kidneys were pooled and sential function of GLI3R in maintaining homogenized to form protein lysates. H89-treated kidneys show a dramatic decrease appropriate Hh activity levels. Because in P-CREB. In contrast, Gant61- and Sant2-treated kidneys do not show this decrease. GLI2 normally does not show a functional Sant2-treated kidneys show a reduction of GLI1, slight reduction of GLI2, and in- role in kidney development, attenuation of creased GLI3R. Gant61-treated kidneys do not show alteration of GLI protein levels. Thm1 renal cystogenesis by loss of GLI2 sug- H89-treated kidneys show slight reduction of GLI1 and GLI3A. (D) MEFs derived gests that the Thm1 ciliary defect enhances from E14.5 wt and Thm1aln/aln mice harboring the BAT-gal allele were treated with L- (control) or Wnt3a-conditioned media overnight and assayed for b-galactosidase GLI2 function in the postnatal kidney. Sim- activity. Bars represent mean6SEM of 3 experiments. aln MEFs show elevated re- ilarly, enhanced GLI2 activity is evident and sponse to Wnt3a ligand. (E) Treatment of cultured Thm1aln/aln kidneys with 40 mM causative of the neural tube defects in the aln/aln IWR-1 does not reduce cystogenic potential. (F) Quantification of kidney images fol- Thm1 developmental mutant. En- lowing 4-day culture. Bars represent mean6SEM; n=8 control and n=3 Thm1aln/aln hanced GLI3A activity also contributes to from 2 experiments. *P,0.05; **P,0.005; ***P,0.0005. the Thm1aln/aln neural tube defects, and Western blot analyses of Thm1aln/aln anterior

J Am Soc Nephrol 25: 2201–2212, 2014 Hedgehog Signaling in Renal Cysts 2207 BASIC RESEARCH www.jasn.org

regulating CFTR and fluid secretion, H89 may also act by partially inhibiting the Hh pathway. Although Gant61 has been shown to inhibit GLI1 and GLI2 from binding to their target DNA sites32 and prevents GLI2 from localizing to the ciliary distal tip in cellular studies (data not shown), which is required for GLI2 activation,43 Gant61-treated kidneys did not alter GLI protein levels. Aside from the Hh components that are also targets of the pathway, ChIP assays have revealed Pax2 and Cyclin D1 as transcriptional targets of Hh signaling (N. Rosenblum, personal communication). Examination of such targets would help in assessing Gant61 action. The molecular mechanism by which Hh signaling may contribute to renal cystogenesis remains undefined. It is possible that Hh signaling contributes directly to increased proliferation of cyst-lining epi- Figure 6. Expression of Hh signaling genes is increased in Thm1 cko, jck,andPkd1 thelial cells. Alternatively, it has been sug- cko cystic kidneys. (A) Quantitative RT-PCR analysis using RNA lysates of 5 wt and 5 gested that low levels of Hh activity in Thm1 cko kidneys, of (B) 5 wt and 5 jck kidneys, or of (C) 5 wt and 4 Pkd1 cko kidneys. mature kidneys maintain renal epithelial 2+ Bars represent mean6SEM of Ptch and Gli fold expression, normalized to house- cells in a differentiated state.44 Altered Ca keeping gene Oaz1, proposed as one of the more stable, reliable control genes for homeostasis is another hallmark of PKD, quantitative PCR.44 *P,0.05; ***P,0.0005; ****P,0.00005. and several studies suggest that Hh signaling modulates Ca2+ levels. In a lung cancer cell line, an SMO inhibitor, GDC-0449, limb buds revealed that loss of THM1 increases GLI3A was shown to increase steady-state levels of Ca2+.45 Con- levels, without altering GLI3R levels.17 Such differences in versely, SMO may act as a G protein–coupled receptor, and GLI protein levels were not detectable in whole kidney ex- Ca2+ imaging of Xenopus embryonic spinal cells showed Sonic tracts. It is possible that assaying whole kidney extracts HH ligand acutely increased Ca2+ spike activity through dilutes expression differences that might occur in a cell type– SMO activation.46 Although these studies show different effects specific manner. Finally, Hh inhibitors also prevented cysts of Hh activity on Ca2+, the possibility of an interplay between in jck and Pkd1m1Bei cultured kidneys, suggesting that in- Hh and Ca2+ signaling in kidney cells may merit investigation. creased Hh signaling may have a general role in renal tubular If increased Hh signaling plays a role in Thm1 renal cysto- dilation and cyst formation. genesis, then molecular mechanisms must be different between Like IFT complex B mutant cells, Thm1-null cells showed Thm1 and the complex B mutants, which cannot fully activate the an elevated response to Wnt3a ligand, reflecting upregulated Hh pathway (reviewed by Eggenschwiler and Anderson10). Even canonical Wnt signaling. However, small molecule Wnt inhib- among IFT complex B mouse mutants, there is variation in which itors did not reduce cysts in the kidney explant assay. Two Wnt branch—canonical or noncanonical—is perturbed before reports suggest that overactive canonical Wnt signaling does cysts arise and once cysts have formed.20,21,33,47–49 Thus, not contribute to renal cystogenesis in Pkd1, Pkd2, and inver- despite a unifying primary cilia hypothesis in renal cystogenesis, sin mouse mutants.41,42 Thus, our data may reflect the possi- differences appear between mouse models in alteration of signal- bility that canonical Wnt signaling does not play a role in ing pathways, which likely reflect heterogeneity of cystic kidney cAMP-mediated tubular dilation. Alternatively, our results disease pathogenesis. This highlights the importance of better may suggest that examining the role of Wnt signaling in renal understanding the molecular mechanisms by which cilia/IFT cystogenesis in vivo is more appropriate. Regardless, lack of mediate Hh and Wnt signaling. prevention of renal cysts in this cAMP culture assay suggests Cystic kidneys of the complex A IFT140 cko mouse also that the preventive effect of the Hh inhibitors is not occurring showed increased Gli expression.23 Thus, data from five mouse through the canonical Wnt pathway. models—IFT140,23 Glis2,15 Thm1, jck,andPkd1—suggest el- The Hh inhibitors also did not exert their beneficial effects by evated Hh signaling in cystic kidneys. Additionally, a global inhibiting PKA. Conversely, the PKA inhibitor H89 decreased gene expression analysis of kidneys from patients with levels of GLI1 and GLI3A, suggesting that in addition to ADPKD revealed increased expression of Hh signaling

2208 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2201–2212, 2014 www.jasn.org BASIC RESEARCH

Figure 7. Small molecule Hh inhibitors reduce cystogenic potential of cultured jck/jck and Pkd1m1Bei/m1Bei kidneys. (A) E13.5 jck/jck kidneys and (B) E14.5 Pkd1m1Bei/Pkd1m1Bei were cultured in the presence of 100 mM 8-bromo-cAMP, in combination with 10 mM Gant61, 5 mMor10mM Sant2 (for jck/jck or Pkd1m1Bei/Pkd1m1Bei kidneys, respectively), or control DMSO for 4 days. Graphs represent quantiﬁcation of kidney images following 4-day culture. Bars represent mean6SEM; n=3 wt, n=9 jck/+,andn=5 jck/jck kidneys from 3 experiments for Gant61; n=5 wt, n=9 jck/+,andn=8 jck/jck kidneys from 4 experiments for Sant2; n=6 wt, n=12 Pkd1m1Bei/+,andn=9 Pkd1m1Bei/Pkd1m1Bei from 4 experiments for Gant61; n=7 wt, n=10 Pkd1m1Bei/+ and n=6 Pkd1m1Bei/Pkd1m1Bei from 4 experiments for Sant2. *P,0.05; **P,0.005; ***P,0.0005; ****P,0.00005.

components, including GLI2, Ptch1,andGAS1.50 While the CONCISE METHODS Thm1 ciliary defect, and likely that of IFT140, directly up- regulates the Hh pathway and Glis2 represses Hh signaling Generation of Thm1 and Pkd1 conditional knock-out in the kidney, further investigation will be required to deter- mice mine the mechanism by which Hh signaling is upregulated in AThm1-lacZ knockout mouse (Ttc21btm2a(KOMP)Wtsi) was purchased jck and Pkd1 mutant kidneys. from KOMP (Knockout Mouse Project). The major components of In summary, our data reveal a role for THM1 loss in renal the targeting vector consisted of a LacZ gene flanked by Frt sites in in- fl fl cystogenesis and a protective role for downregulating Hh tron 3 of Thm1 and of loxP sites flanking exon 4. To create Thm1 ox/ ox signaling in the Thm1 cko mouse and in cultured kidneys of mice with conditional deletion potential, Thm1-lacZ mice were mated three independent genetic mouse models of cystic kidney dis- to FLPeR mice (gift from Susan Dymecki, Harvard Medical School), ease. This compels analysis of whether Hh inhibitors reduce which express FLPe recombinase. Resulting progeny expressing renal cystogenesis in these mouse models in vivo.IftheseHh FLPe recombinase showed excision of the Frt-flanked lacZ gene inhibitors are effective in vivo, several Hh inhibitors are being and continued Thm1 transcription from exon 3 to exon 4. These fl fl fl tested in clinical trials for cancer (reviewed by Tran et al.51), Thm1 ox/+ mice were intercrossed to generate Thm1 ox/ ox mice. The which will facilitate translation of these experiments to ther- tamoxifen-inducible ROSA26CreERT mouse was purchased from apeutic application. The Jackson Laboratory (004847) and mated to Thm1aln/+ mice

J Am Soc Nephrol 25: 2201–2212, 2014 Hedgehog Signaling in Renal Cysts 2209 BASIC RESEARCH www.jasn.org to create Thm1aln/+; ROSA26CreERT+ mice. Male Thm1aln/+; Immunofluorescence fl fl ROSA26CreERT+ mice were time-mated with Thm1 ox/ ox females. After removal of renal capsules, dissected kidneys were bisected and E17.5 pregnant females were injected intraperitoneally with tamox- fixed in 4% paraformaldehyde overnight at 4°C. Tissue was dehydra- ifen (Sigma) at a dose of 9 mg/kg mouse body weight to generate ted through a series of ethanol, xylene, and paraffin, and then 2 Thm1aln/ ; ROSA26CreERT/+ (or Thm1 cko) mice. Alternatively, 5- embedded in paraffin. Sections of 7 mm were obtained using a mi- 2 week-old wt and Thm1aln/ ; ROSA26CreERT/+ mice were injected crotome. Antigen retrieval using sodium citrate buffer (pH, 6) was intraperitoneally with tamoxifen to examine the effects of deleting performed. Tissue was blocked with 1% BSA in PBS for 1 hour, then Thm1 in fully developed kidneys. incubated with lectins, LTL, and DBA (1:50; Vector Laboratories) or fl fl Pkd1 ox/ ox mice were obtained from Jackson Laboratories (010671). primary antibodies against a6F (1:1000; Developmental Studies Hy- fl fl fl Pkd1 ox/ ox females were mated to Pkd1 ox/+; ROSA26-CreERT/+ males bridoma Bank), Tamms–Horsfall Protein (1:500; Santa Cruz Bio- fl fl fl fl to create Pkd1 ox/ ox; ROSA26-CreERT/+ mice. At P2, Pkd1 ox/ ox nursing technology), PCNA (1:2000; Cell Signaling Technology), acetylated mothers were injected with 10 mg tamoxifen/kg mouse body weight to a-tubulin (1:4000; Sigma-Aldrich). Sections were washed three times induce Pkd1 deletion in pups. with PBS, and then incubated with appropriate secondary antibody conjugated to AlexaFluor 488 or AlexaFluor 594. After three washes 9 Mouse Genotyping of PBS, sections were mounted with Vectashield with 4 ,6-diamidino- Tail biopsies (1–2 mm) or yolk sacs were used to extract DNA by alkaline 2-phenylindole (DAPI) (Vector Laboratories). Staining was visual- lysis. Tails were boiled in 50–200 ml 50 mM NaOH for 10 minutes and ized using a Zeiss Axiophot fluorescence microscope and imaged briefly vortexed. One tenth the volume of 1 M Tris-HCl (pH, 8.0) with a Leica DFC 350 camera or a Nikon 80i fluorescence microscope was added, followed by centrifugation at 14,000 rpm for 6 minutes. equipped with a Nikon DS-Fi1 camera. One microliter of supernatant was used for subsequent genotyping. Thm1aln/aln mice were genotyped as described elsewhere,17 using primers Quantification of Proliferation Rates + alndiag-F and alndiag-R, followed by amplicon digestion with AvaII. jck PCNA cells and DAPI-stained nuclei were manually counted using mice were genotyped using primers jck-F and jck-R and subsequent the Cell Counter plug-in of ImageJ software (National Institutes of amplicon digestion with BseYI. Pkd1m1Bei mice were genotyped using a Health, Bethesda, MD). fl Taqman assay as described previously.25 Pkd1 ox allele was genotyped using primers, Pkd1f-F and Pkd1f-R, as described by The Jackson Lab- Kidney Explant Cultures oratory. Primer and probe sequences are listed in Supplemental Table 1. Kidneys were dissected from E13.5 Thm1aln/aln or jck/jck embryos or from E14.5 Pkd1m1Bei/m1Bei embryos and cultured on 1-mmporein- Histology serts in a six-well plate with DMEM/F12 media (Life Technologies) Kidneys were dissected and renal capsules were removed. Kidneys were containing penicillin and streptomycin. Kidneys were cultured in the then bisected longitudinally and fixed in Bouin solution for several days. presence of 100 mM 8-bromo-cAMP (Sigma-Aldrich), with or with- Tissue was dehydrated through a series of ethanol, xylene (two changes), out Sant1 (Sigma-Aldrich), Sant2 (Alexis Biochemicals), Gant61 paraffin(twochanges), thenembeddedinparaffin. Sections of 7 mmwere (Alexis Biochemicals), IWR-1 (Sigma-Aldrich), IWP (Sigma- obtained using a microtome and stained with hematoxylin and eosin. Aldrich), or H89 (Sigma-Aldrich) for 4 days. Media and small mol- ecules were refreshed daily. Kidneys were imaged using a Leica BUN and cAMP Measurements MZ12.5 stereoscope and a DFC500 camera. Images were quantified Trunk blood was collected and serum was obtained using Microvette using ImageJ software. For each kidney, the cystic areas were calcu- CB 300 Blood Collection System tubes (Kent Scientific). Serum BUN lated and summed, then divided by the total area of the kidney in was measured using the QuantiChrom Urea Assay Kit (BioAssay question. Systems) according to manufacturer’s protocol. Dissected kidneys were halved or quartered and snap-frozen. Kidney Quantitative RT-PCR pieces were homogenized in 0.1 M HCl using a Bullet Blender Storm RNA was extracted using Trizol (Life Technologies) according to (MidSci).LevelsofcAMPwereobtainedfromkidneyhomogenatesusing manufacturer’s protocol. One microgram of RNA was converted to the cAMP Enzyme Immunoassay Kit, Direct (Sigma-Aldrich) according cDNA using Quanta Biosciences qScript cDNA Supermix (VWR In- to the manufacturer’s instructions. Protein concentrations of kidney ternational). Quantitative PCR analysis was performed using Quanta homogenates were obtained using a BCA Protein Assay (Pierce). Biosciences Perfecta qPCR Supermix (VWR International) and a Bio- Rad CFX Connect Real-Time PCR Detection System. Primers used Scanning Electron Microscopy for detection of Ptch1, Gli1, Gli2, Gli3, and housekeeping gene Oaz1 Six-week-old kidneys were dissected in PBS and fixed in 2% (proposed as one of the more stable, reliable control genes for quan- glutaraldehydein0.1Msodiumcacodylatebufferat4°C.Kidneys titative PCR)52 were designed using the Roche Applied Science RT- were dehydrated in an ethanol series, dried in an EMS Critical Point qPCR Assay Design Center (http://www.roche-applied-science.com/ Dryer, mounted onto metal stubs, and coated in a Pelco SC-6 Sputter shop/CategoryDisplay?catalogId=10001&tab=&identifier=Universal Coater. Renal primary cilia were visualized using a Hitachi S-2700 +Probe+Library&langId=-1&storeId=15006). All amplicons span an Scanning Electron Microscope equipped with a Quartz PCI digital intron and were annealed at 60°C. Primer sequences are listed in Sup- camera. plemental Table 1.

2210 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2201–2212, 2014 www.jasn.org BASIC RESEARCH

Western Blot mechanosensation in the primary cilium of kidney cells. Nat Genet 33: Protein extracts were obtained by pooling four cultured kidneys and 129–137, 2003 homogenizing them using Passive Lysis Buffer (Promega) containing 5. Wang Q, Pan J, Snell WJ: Intraflagellar transport particles participate proteinase inhibitor cocktail (Pierce). Western blots were performed directly in cilium-generated signaling in Chlamydomonas. Cell 125: 549–562, 2006 17 as described, using primary antibodies against CREB, P-CREB, 6. Ocbina PJ, Eggenschwiler JT, Moskowitz I, Anderson KV: Complex GLI1 (Cell Signaling Technology), GLI2 (generous gifts from Drs. J. interactions between genes controlling trafficking in primary cilia. Nat Eggenschwiler and B. Wang), GLI3 (R&D Systems), and a-tubulin Genet 43: 547–553, 2011 (DM1A from Sigma-Aldrich). Signal was detected using SuperSignal 7. Ingham PW, Nakano Y, Seger C: Mechanisms and functions of Hedgehog – West Femto Chemiluminescent Substrate (Pierce). signalling across the metazoa. Nat Rev Genet 12: 393 406, 2011 8. Corbit KC, Aanstad P, Singla V, Norman AR, Stainier DY, Reiter JF: Vertebrate Smoothened functions at the primary cilium. Nature 437: BAT-gal Assays 1018–1021, 2005 aln/+ Thm1 heterozygous mice were mated to mice harboring a 9. Rohatgi R, Milenkovic L, Scott MP: Patched1 regulates hedgehog sig- b-galactosidase transgene (B6.Cg-Tg[BAT-lacZ]3Picc/J; The Jackson naling at the primary cilium. Science 317: 372–376, 2007 Laboratory; Stock 005317). MEFs were then isolated from E14.5 wt; 10. Eggenschwiler JT, Anderson KV: Cilia and developmental signaling. – BAT-lacZ and Thm1aln/aln; BAT-lacZ embryos, then treated with either Annu Rev Cell Dev Biol 23: 345 373, 2007 11. Wang B, Fallon JF, Beachy PA: Hedgehog-regulated processing of Gli3 34 L (control) or Wnt3a-conditioned media as described previously. produces an anterior/posterior repressor gradient in the developing BAT-gal activity was measured using the Galacto-Light Plus System vertebrate limb. Cell 100: 423–434, 2000 (Applied Biosystems). 12. Cain JE, Islam E, Haxho F, Blake J, Rosenblum ND: GLI3 repressor controls functional development of the mouse ureter. JClinInvest121: 1199–1206, 2011 Statistical Analyses 13. Christensen ST, Ott CM: Cell signaling. A ciliary signaling switch. Sci- Statistical significance was calculated using a t test (Excel; Microsoft, ence 317: 330–331, 2007 Redmond, WA). 14. Hu MC, Mo R, Bhella S, Wilson CW, Chuang PT, Hui CC, Rosenblum ND: GLI3-dependent transcriptional repression of Gli1, Gli2 and kidney patterning genes disrupts renal morphogenesis. Development 133: ACKNOWLEDGMENTS 569–578, 2006 15. Attanasio M, Uhlenhaut NH, Sousa VH, O’Toole JF, Otto E, Anlag K, Klugmann C, Treier AC, Helou J, Sayer JA, Seelow D, Nurnberg G, We thank Barbara Fegley of the University of Kansas Medical Center Becker C, Chudley AE, Nurnberg P, Hildebrandt F, Treier M: Loss of (KUMC) Electron Microscopy Core and Jing Huang of the KUMC GLIS2 causes nephronophthisis in humans and mice by increased ap- Histology Core for their technical assistance. We thank members of optosis and fibrosis. Nat Genet 39: 1018–1024, 2007 the Beier Lab, the Harvard Center for PKD Research, and the KUMC 16. Chan SK, Riley PR, Price KL, McElduff F, Winyard PJ, Welham SJ, Woolf Kidney Institute for helpful discussions. AS, Long DA: Corticosteroid-induced kidney dysmorphogenesis is associated with deregulated expression of known cystogenic mole- ThisworkwassupportedbyaPilotandFeasibilityProjectAwardfrom cules, as well as Indian hedgehog. Am J Physiol Renal Physiol 298: the Harvard Center for PKD Research (PI: Jing Zhou), R21DK088048, F346–F356, 2010 ASN Gottschalk Research Scholar Award, National Institutes of Health 17. Tran PV, Haycraft CJ, Besschetnova TY, Turbe-Doan A, Stottmann RW, (NIH) Center of Biomedical Research Excellence (P20-GM104936-06, Herron BJ, Chesebro AL, Qiu H, Scherz PJ, Shah JV, Yoder BK, Beier PI: Dale Abrahamson) and NIH K-INBRE (P20-GM103418, PI: Joan DR: THM1 negatively modulates mouse sonic hedgehog signal trans- duction and affects retrograde intraflagellar transport in cilia. Nat Hunt) to P.V.T., an NIH T32 Postdoctoral Fellowship (DK71496-6 A1, Genet 40: 403–410, 2008 PI: Jared Grantham) to D.T.J., and an R01-HD36404 to D.R.B. 18. Qin J, Lin Y, Norman RX, Ko HW, Eggenschwiler JT: Intraflagellar transport protein 122 antagonizes Sonic Hedgehog signaling and controls ciliary localization of pathway components. Proc Natl Acad Sci DISCLOSURES USA108: 1456–1461, 2011 None. 19. Yoder BK, Tousson A, Millican L, Wu JH, Bugg CE Jr, Schafer JA, Balkovetz DF: Polaris, a protein disrupted in orpk mutant mice, is required for assembly of renal cilium. Am J Physiol Renal Physiol 282: F541–F552, 2002 REFERENCES 20. Jonassen JA, San Agustin J, Follit JA, Pazour GJ: Deletion of IFT20 in the mouse kidney causes misorientation of the mitotic spindle and 1. Torres VE, Harris PC: Mechanisms of disease: Autosomal dominant and cystic kidney disease. JCellBiol183: 377–384, 2008 recessive polycystic kidney diseases. Nat Clin Pract Nephrol 2: 40–55, 21. Patel V, Li L, Cobo-Stark P, Shao X, Somlo S, Lin F, Igarashi P: Acute quiz 55, 2006 kidney injury and aberrant planar cell polarity induce cyst formation in 2. Quinlan RJ, Tobin JL, Beales PL: Modeling ciliopathies: Primary cilia in mice lacking renal cilia. Hum Mol Genet 17: 1578–1590, 2008 development and disease. Curr Top Dev Biol 84: 249–310, 2008 22. Lin F, Hiesberger T, Cordes K, Sinclair AM, Goldstein LS, Somlo S, 3. Nauli SM, Rossetti S, Kolb RJ, Alenghat FJ, Consugar MB, Harris PC, Igarashi P: Kidney-specific inactivation of the KIF3A subunit of kinesin-II Ingber DE, Loghman-Adham M, Zhou J: Loss of polycystin-1 in human inhibits renal ciliogenesis and produces polycystic kidney disease. Proc cyst-lining epithelia leads to ciliary dysfunction. JAmSocNephrol17: Natl Acad Sci U S A 100: 5286–5291, 2003 1015–1025, 2006 23. Jonassen JA, SanAgustin J, Baker SP, Pazour GJ: Disruption of IFT 4. Nauli SM, Alenghat FJ, Luo Y, Williams E, Vassilev P, Li X, Elia AE, Lu W, complex A causes cystic kidneys without mitotic spindle misorientation. Brown EM, Quinn SJ, Ingber DE, Zhou J: Polycystins 1 and 2 mediate JAmSocNephrol23: 641–651, 2012

J Am Soc Nephrol 25: 2201–2212, 2014 Hedgehog Signaling in Renal Cysts 2211 BASIC RESEARCH www.jasn.org

24. Davis EE, Zhang Q, Liu Q, Diplas BH, Davey LM, Hartley J, Stoetzel C, 38. McCooke JK, Appels R, Barrero RA, Ding A, Ozimek-Kulik JE, Bellgard Szymanska K, Ramaswami G, Logan CV, Muzny DM, Young AC, MI, Morahan G, Phillips JK: A novel mutation causing nephronophthisis Wheeler DA, Cruz P, Morgan M, Lewis LR, Cherukuri P, Maskeri B, in the Lewis polycystic kidney rat localises to a conserved RCC1 domain Hansen NF, Mullikin JC, Blakesley RW, Bouffard GG, Gyapay G, Rieger in Nek8. BMC Genomics 13: 393, 2012 S, Tonshoff B, Kern I, Soliman NA, Neuhaus TJ, Swoboda KJ, Kayserili 39. Herron BJ, Lu W, Rao C, Liu S, Peters H, Bronson RT, Justice MJ, H, Gallagher TE, Lewis RA, Bergmann C, Otto EA, Saunier S, Scambler McDonald JD, Beier DR: Efficient generation and mapping of recessive PJ, Beales PL, Gleeson JG, Maher ER, Attie-Bitach T, Dollfus H, developmental mutations using ENU mutagenesis. Nat Genet 30: 185– Johnson CA, Green ED, Gibbs RA, Hildebrandt F, Pierce EA, Katsanis 189, 2002 N: TTC21B contributes both causal and modifying alleles across the 40. Parelkar SV, Kapadnis SP, Sanghvi BV, Joshi PB, Mundada D, Oak SN: ciliopathy spectrum. Nat Genet 43: 189–196, 2011 Meckel-Gruber syndrome: A rare and lethal anomaly with review of 25. Magenheimer BS, St John PL, Isom KS, Abrahamson DR, De Lisle RC, literature. J Pediatr Neurosci 8: 154–157, 2013 Wallace DP, Maser RL, Grantham JJ, Calvet JP: Early embryonic renal 41. Miller MM, Iglesias DM, Zhang Z, Corsini R, Chu L, Murawski I, Gupta I, tubules of wild-type and polycystic kidney disease kidneys respond to Somlo S, Germino GG, Goodyer PR: T-cell factor/beta-catenin activity cAMP stimulation with cystic fibrosis transmembrane conductance is suppressed in two different models of autosomal dominant polycystic regulator/Na(+),K(+),2Cl(-) Co-transporter-dependent cystic dilation. J kidney disease. Kidney Int 80: 146–153, 2011 Am Soc Nephrol 17: 3424–3437, 2006 42. Sugiyama N, Tsukiyama T, Yamaguchi TP, Yokoyama T: The canonical 26. Natoli TA, Gareski TC, Dackowski WR, Smith L, Bukanov NO, Russo RJ, Wnt signaling pathway is not involved in renal cyst development in the Husson H, Matthews D, Piepenhagen P, Ibraghimov-Beskrovnaya O: Pkd1 kidneys of inv mutant mice. Kidney Int 79: 957–965, 2011 and Nek8 mutations affect cell-cell adhesion and cilia in cysts formed in 43. Haycraft CJ, Banizs B, Aydin-Son Y, Zhang Q, Michaud EJ, Yoder BK: kidney organ cultures. Am J Physiol Renal Physiol 294: F73–F83, 2008 Gli2 and Gli3 localize to cilia and require the intraflagellar transport 27. Piontek K, Menezes LF, Garcia-Gonzalez MA, Huso DL, Germino GG: A protein polaris for processing and function. PLoS Genet 1: e53, 2005 critical developmental switch defines the kinetics of kidney cyst for- 44.LiB,RauhauserAA,DaiJ,SakthivelR,IgarashiP,JettenAM,Attanasio mation after loss of Pkd1. Nat Med 13: 1490–1495, 2007 M: Increased hedgehog signaling in postnatal kidney results in aberrant 28. Davenport JR, Watts AJ, Roper VC, Croyle MJ, van Groen T, Wyss JM, activation of nephron developmental programs. Hum Mol Genet 20: Nagy TR, Kesterson RA, Yoder BK: Disruption of intraflagellar transport 4155–4166, 2011 in adult mice leads to obesity and slow-onset cystic kidney disease. 45. Tian F, Schrodl K, Kiefl R, Huber RM, Bergner A: The hedgehog path- Current biology: CB,17: 1586-1594, 2007. way inhibitor GDC-0449 alters intracellular Ca2+ homeostasis and in- 29. Smith LA, Bukanov NO, Husson H, Russo RJ, Barry TC, Taylor AL, Beier hibits cell growth in cisplatin-resistant lung cancer cells. Anticancer Res DR, Ibraghimov-Beskrovnaya O: Development of polycystic kidney 32: 89–94, 2012 disease in juvenile cystic kidney mice: Insights into pathogenesis, ciliary 46. Belgacem YH, Borodinsky LN: Sonic hedgehog signaling is decoded by abnormalities, and common features with human disease. JAmSoc calcium spike activity in the developing spinal cord. Proc Natl Acad Sci Nephrol 17: 2821–2831, 2006 USA108: 4482–4487, 2011 30. Hopp K, Ward CJ, Hommerding CJ, Nasr SH, Tuan HF, Gainullin VG, 47. Nishio S, Tian X, Gallagher AR, Yu Z, Patel V, Igarashi P, Somlo S: Loss of Rossetti S, Torres VE, Harris PC: Functional polycystin-1 dosage gov- oriented cell division does not initiate cyst formation. J Am Soc Nephrol erns autosomal dominant polycystic kidney disease severity. JClinIn- 21: 295–302, 2010 vest 122: 4257–4273, 2012 48. Simons M, Gloy J, Ganner A, Bullerkotte A, Bashkurov M, Kronig C, 31. Torres VE, Wang X, Qian Q, Somlo S, Harris PC, Gattone VH 2nd: Ef- Schermer B, Benzing T, Cabello OA, Jenny A, Mlodzik M, Polok B, fective treatment of an orthologous model of autosomal dominant Driever W, Obara T, Walz G: Inversin, the gene product mutated in polycystic kidney disease. Nat Med 10: 363–364, 2004 nephronophthisis type II, functions as a molecular switch between Wnt 32. Lauth M, Bergstrom A, Shimokawa T, Toftgard R: Inhibition of GLI- signaling pathways. Nat Genet 37: 537–543, 2005 mediated transcription and tumor cell growth by small-molecule an- 49. Saburi S, Hester I, Fischer E, Pontoglio M, Eremina V, Gessler M, tagonists. Proc Natl Acad Sci U S A 104: 8455–8460, 2007 Quaggin SE, Harrison R, Mount R, McNeill H: Loss of Fat4 disrupts PCP 33. Saadi-Kheddouci S, Berrebi D, Romagnolo B, Cluzeaud F, Peuchmaur signaling and oriented cell division and leads to cystic kidney disease. M, Kahn A, Vandewalle A, Perret C: Early development of polycystic Nat Genet 40: 1010–1015, 2008 kidney disease in transgenic mice expressing an activated mutant of the 50. Song X, Di Giovanni V, He N, Wang K, Ingram A, Rosenblum ND, Pei Y: beta-catenin gene. Oncogene 20: 5972–5981, 2001 Systems biology of autosomal dominant polycystic kidney disease 34. Corbit KC, Shyer AE, Dowdle WE, Gaulden J, Singla V, Chen MH, (ADPKD): Computational identification of gene expression pathways Chuang PT, Reiter JF: Kif3a constrains beta-catenin-dependent Wnt and integrated regulatory networks. Hum Mol Genet 18: 2328–2343, signalling through dual ciliary and non-ciliary mechanisms. Nat Cell Biol 2009 10: 70–76, 2008 51. Tran PV, Lachke SA, Stottmann RW: Toward a systems-level un- 35. ChenB,DodgeME,TangW,LuJ,MaZ,FanCW,WeiS,HaoW,Kilgore derstanding of the Hedgehog signaling pathway: Defining the complex, J, Williams NS, Roth MG, Amatruda JF, Chen C, Lum L: Small molecule- robust, and fragile. Wiley Interdiscip Rev Syst Biol Med 5: 83–100, 2013 mediated disruption of Wnt-dependent signaling in tissue re- 52. de Jonge HJ, Fehrmann RS, de Bont ES, Hofstra RM, Gerbens F, Kamps generation and cancer. Nat Chem Biol 5: 100–107, 2009 WA, de Vries EG, van der Zee AG, te Meerman GJ, ter Elst A: Evidence 36. Liu S, Lu W, Obara T, Kuida S, Lehoczky J, Dewar K, Drummond IA, Beier based selection of housekeeping genes. PLoS ONE 2: e898, 2007 DR: A defect in a novel Nek-family kinase causes cystic kidney disease in the mouse and in zebrafish. Development 129: 5839–5846, 2002 37. Otto EA, Trapp ML, Schultheiss UT, Helou J, Quarmby LM, Hildebrandt F: NEK8 mutations affect ciliary and centrosomal localization and may This article contains supplemental material online at http://jasn.asnjournals. cause nephronophthisis. JAmSocNephrol19: 587–592, 2008 org/lookup/suppl/doi:10.1681/ASN.2013070735/-/DCSupplemental.

2212 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2201–2212, 2014 Table S1. Oligonucleotide sequences

Figure S1. Thm1 cko cystic kidney disease initiates by P15. Histological analyses of

P10, P15 and P20 wt and Thm1 cko kidneys.

Figure S2. Proliferation persists in P20 Thm1 cko medulla. (A) Immunostaining of

P20 wt and Thm1 cko kidney sections for PCNA. In wt kidney, band of PCNA+ cells on left side of section is cortex. Note marked reduction in PCNA+ cells in medulla, right of cortex. Similar corticomedullary region is photographed for Thm1 cko kidney. Note absence of a reduction in PCNA+ cells in Thm1 cko medulla. All scale bars represent

500µm. (B) Proliferation rates in cortices and medullas of 3 wt and 3 Thm1 cko mice.

Figure S3. Deletion of Thm1 at 5 weeks of age does not cause renal cysts 3 months following deletion. Histological analyses of wt and Thm1 cko kidneys 3 months following injection of mice with tamoxifen at 5 weeks of age.

Figure S4. GLI protein levels are similar between wt and Thm1aln/aln whole kidney extract. Representative Western blot analyses of protein extracts from (A) E13.5 wt and

Thm1aln/aln kidneys cultured in cAMP for 4 days and (B) E18.5 wt and Thm1aln/aln kidneys.

Three cultured kidneys were pooled each for wt and Thm1aln/aln from 2 organ culture experiments. Kidneys of 2 E18.5 wt and 2 Thm1aln/aln mice were analyzed separately.

Figure S5. Small molecule Wnt inhibitors, IWR-1 and IWP-2, do not reduce cystogenic potential of CD1 kidney explants. E13.5 CD1 kidneys were cultured in the presence of 8-bromo-cAMP, in combination with 10 µM, 20 µM or 40 µM IWR-1 or 10

µM, 20 µM or 40 µM IWP-2 for 4 days. Contralateral kidneys were cultured in control

DMSO at concentrations present in IWR-1 and IWP-2 treatments. Graphs represent quantitative assessment of kidney images following 4-day culture. Bars represent mean ±

SEM. n=4 kidneys. *P<0.05

Figure S6. SMO inhibitor, Sant1, prevents tubular dilation in CD1 kidneys cultured with cAMP. E13.5 CD1 kidneys were cultured in the presence of 8-bromo-cAMP, in combination with either control DMSO or 20µM Sant1 for 4 days. Graphs represent quantitative assessment of kidney images following 4-day culture. Bars represent mean ±

SEM. n=10 kidneys from 2 experiments. *P<5.0 x 10-6.

Table S1. Oligonucleotide Sequences Figure S1. Thm1 cko renal tubular dilations begin by P15

wt Thm1 cko P10

P15

P20 Figure S2. Proliferation persists in P20 Thm1 cko kidneys

A wt Thm1 cko P20

B 40 35 30 25 wt 20 Thm1 cko 15 10 5 0 cortex medulla Figure S3. Deletion of Thm1 at 5 weeks of age does not cause renal cysts 3 months following deletion

wt Thm1 cko Figure S4. GLI protein levels are similar between wt and Thm1aln/aln whole kidney extracts

A B GLI1 GLI1

GLI2 GLI2

GLI3A GLI3A

GLI3R GLI3R

DM1A DM1A Figure S5. Small molecule Wnt inhibitors do not reduce cystogenic potential in CD-1 kidneys cultured with cAMP

0.001% 0.002% 0.004% A 30 25 * * 20 DMSO 15 10 10µM 20µM 40µM

% cystic area%cystic 5 0 10 20 40 IWR-1 DMSO IWR

B 0.001% 0.002% 0.004% 25 * * 20

DMSO 15

10µM 20µM 40µM 10 5 % cystic area%cystic

IWP-2 IWP-2 10 20 40

DMSO IWP Figure S6. SMO inhibitor, SANT1, prevents tubular dilation in CD1 kidneys cultured with cAMP

DMSO 20µM SANT1 12 10 8 6 4 * % cystic area%cystic 2 0 DMSO 20uMµ Sant1