Inhibition of Hedgehog signaling does not mitigate polycystic kidney disease severity in a Pkd1 mutant mouse model

ABSTRACT

Autosomal dominant polycystic kidney disease (ADPKD) is a monogenic disorder caused by mutations in PKD1 or PKD2, encoding polycystin-1 and polycystin-2, respectively. These polycystins form a cilia-localized complex that, when mutated, fails to inhibit an uncharacterized cilia-dependent cyst activation (CDCA) signal. This leads to progressive bilateral cyst growth and ultimately compromised renal function. Previous in vitro and in vivo studies from our group have demonstrated that Hedgehog (Hh) signaling inhibition reduces renal cystic severity in PKD models. To further investigate, we inactivated several Hh pathway components (Gli1, Gli2, Gli3 and Smo) in a Pkd1 hypomorphic mouse model through conditional deletion by tamoxifen-induced Cre-Lox recombination. We assessed cystic severity using kidney weight assessment and a microcomputed tomography (micro-CT)-based 3D imaging assay. Contrary to expectations, inactivation of Gli1 and Smo significantly increased cystogenesis. These findings suggest that Hh signaling does not mediate the CDCA signal.

Summary: Inactivation of several Hedgehog pathway components in a Pkd1 hypomorphic mouse model does not mitigate cystic disease severity.

INTRODUCTION

Polycystic kidney disease (PKD) refers to a group of monogenic disorders characterized by the formation of bilateral renal cysts. Autosomal dominant polycystic kidney disease (ADPKD), the most common form of PKD, is caused by mutations in the genes PKD1 (78% of cases) and PKD2 (15% of cases) (Chebib et al., 2025). The disease is marked by the progressive growth of fluid-filled cysts derived from renal tubular epithelial cells. Over time, cyst expansion displaces and compresses the surrounding kidney parenchyma, impairing renal function and leading to end-stage renal disease at a median age of 58 for PKD1-related mutations and 79 for PKD2-related mutations (Cornec-Le Gall et al., 2013).

Although the precise pathogenesis of PKD remains unclear, primary cilia are strongly implicated in the disease mechanism (Boletta and Caplan, 2025). Primary cilia are non-motile, microtubule-based organelles that protrude from the apical surface of renal epithelial cells into the lumen, functioning as sensory structures that detect and respond to extracellular signals. Proper ciliary function relies on intraflagellar transport (IFT), a bidirectional protein transport system that moves components along the microtubule core (Ishikawa and Marshall, 2017). Disruptions in IFT-related genes or other cilia related genes can lead to a spectrum of disorders collectively known as ciliopathies, one of which is PKD. Early studies have shown that mutations in the IFT gene Tg737 (Ift88) result in ciliary malformations and cyst formation in renal tubular cells (Yoder et al., 1995; Pazour et al., 2000). Moreover, evidence shows that NPHP proteins, encoded from genes involved in nephronophthisis, a type of recessive cystic kidney disease, localize and are involved with proper ciliary function (Otto et al., 2003; Wolf and Hildebrandt, 2011). For example, defects in the NIMA-related (Nek) kinase Nek8, which is found in the inversin compartment of cilia, cause the development of juvenile cystic kidney disease in the jck mouse, further cementing the role of cilia in cystogenesis (Liu et al., 2002; Manning et al., 2013).

Among the known primary cilia-associated proteins are polycystin-1 (PC1) and polycystin-2 (PC2), which are encoded by PKD1 and PKD2, respectively. PC1 and PC2 colocalize to the membrane of primary cilia in the epithelium of the kidney and act in tandem to form a receptor–ion channel complex that regulates cellular Ca²⁺ ion levels (Nauli et al., 2003). Additionally, the polycystin complex has also been hypothesized to negatively regulate an uncharacterized cilia-dependent cyst activation (CDCA) signaling pathway, distinct from MAPK/ERK, mTOR or cAMP pathways (Ma et al., 2013).

Given the established connection between cilia and PKD pathogenesis and the evidence of a novel cyst activation pathway that require functional primary cilia, we sought to investigate the role of Hedgehog (Hh) signaling, a pathway mediated by primary cilia, in ADPKD. Specifically, we hypothesized that aberrant Hh signaling could underlie the CDCA pathway involvement in cyst growth activation. Previously, using a N-ethyl-N-nitrosourea-derived Ttc21b-mutant mouse model, which carries a mutation in an IFT gene and develops cystic kidneys (Tran et al., 2008), our group has demonstrated that suppressing overactive Hh signaling both genetically and pharmacologically reduces the cystogenic potential of mutant kidney explants (Tran et al., 2014). This was also observed in the jck cystic kidney disease model, which carries a mutation in the kinase Nek8 (Liu et al., 2002). Most importantly, with respect to this study, inhibition of Hh signaling in Pkd1-null mutant kidney explant cultures reduces cyst formation (Tran et al., 2014).

Although none of these models represented adult-onset polycystin-mediated cystic disease, the results suggested Hh signaling be considered as a candidate for the CDCA pathway. To explore this, we used the hypomorphic Pkd1 p.R3277C (Pkd1^RC) mouse model (Hopp et al., 2012), which mimics the cystogenic process observed in human ADPKD, and provides an attractive choice for our study. Mice that carry the typical human disease genotype Pkd1^+/− or Pkd2^+/− only develop a few kidney cysts, making them poor models to study PKD. Biallelic Pkd1 inactivation models are embryonically lethal at mid gestation whereas inducible Pkd1 deletion models can present with a spectrum in the intensity and progression of the PKD phenotype depending on timing of inactivation or recombination efficiency, which could differ between mice (Piontek et al., 2007). Although all of the polycystin mutant models have strengths and limitations, the Pkd1^RC allele offers a consistent model of progressive cyst growth, which facilitates quantitative assessment of disease severity and amelioration. Over 20 studies using the Pkd1^RC mouse line are listed in the Mouse Genome Informatics database (https://www.informatics.jax.org/), testing the effects of genetic interactions, diet, microRNA, inflammation and drug treatment.

Using this model, we aimed to assess the effect of Hh pathway inactivation on cyst progression and further elucidate the mechanistic role of Hh signaling in ADPKD pathogenesis. We leveraged the Cre-LoxP system to conditionally delete components of the Hh signaling pathway (Smo, Gli2 and Gli3) at postnatal day 13.5 (P13.5) and assessed the cystic state of the kidneys at 90 days using gross morphology and microcomputed tomography (micro-CT) analysis. Gli1-null homozygous cystic kidneys were also analyzed. In contrast to our previous results, we found that repression of the Hh pathway increased the progression of polycystic disease in a hypomorphic Pkd1 model, and we conclude that it does not contribute to PKD pathogenesis as a CDCA signal.

RESULTS

Micro-CT analysis of Pkd1^RC/RC mice illustrates progressive renal cyst development 90 days postnatally

To investigate the role of Hh signaling in PKD pathogenesis, we utilized a mouse model carrying the knock-in Pkd1 p.R3277C mutation (Pkd1tm1.1Pcha, called here Pkd1^RC). This mutation mimics the human PKD1 p.R3277C allele and leads to an ∼60% reduction in mature PC1 due to inefficient cleavage and folding (Hopp et al., 2012). Homozygous Pkd1^RC/RC mice exhibit early-onset cystogenesis, with cystic expansion progressing throughout their lifespan.

To evaluate the cystic profile of the diseased kidneys, we utilized micro-CT imaging, a method that enables a comprehensive and quantitative assessment. This technique allowed for efficient and high-throughput analysis of kidney samples. To validate the reliability of micro-CT for assessing renal cystic burden and to establish the time-course of postnatal cyst development of Pkd1^RC/RC kidneys, we compared the cystic volume (CV) index derived from micro-CT imaging to the conventional kidney weight-to-body weight (KW/BW) ratio metric. Kidneys were collected and analyzed at four time points: 21, 44, 67 and 90 days postnatally (Fig. 1A–C).

Fig. 1.

Pkd1^RC/RC kidneys progressively develop cysts 90 days postnatally. (A) Micro-CT cross sections from representative Pkd1^RC/RC kidneys from each time point (P21, P44, P67, P90) in the time series analysis with a P90 Pkd1^+/+ control. (B,C) Mean cystic volume percentage (CV %) (B) and kidney weight-to-body weight (KW/BW) ratio of cystic kidneys at each time point with s.d. (C). (D) CV % and KW/BW % plotted for each mouse to assess correlation between the two metrics. P21 mice were excluded due to weak correlation, likely reflecting early disease stage and minimal cystic burden. n=11, P21; n=20, P44; n=12, P67; n=13, P90.

We observed a correlation between CV index and KW/BW ratio, with an R² value of 0.65 (P<0.0001), confirming the validity of micro-CT as an independent measure in the evaluation of cystic kidneys (Fig. 1D). These findings validate the utility of micro-CT for cystic analysis and establish this model as a robust platform for studying progressive PKD pathogenesis and therapeutic interventions.

Gli1 absence increases cystic severity in Pkd1 mutant mice

To evaluate the role of Hh signaling in PKD pathogenesis, we first investigated the effects of inactivating Gli1, a transcriptional activator that mediates Hh signaling expression. Unlike other Hh-associated transcription factors, Gli1 is not essential for mouse embryonic development or post-natal survival (Park et al., 2000), allowing us to generate viable Gli1-null homozygous mice. Using the Gli1^lacZ-null mutant knock-in mouse (Bai et al., 2002), we crossed Gli1-null homozygous mice to Pkd1^RC/RC mice to produce double mutants.

We compared the cystic profiles of Gli1; Pkd1 double mutant kidneys and Pkd1 single mutant kidneys at 90 days postnatally. Analysis revealed a significantly larger cystic burden in the double mutant kidneys compared to the single mutants, with KW/BW ratio indicating a 1.45 average fold increase (Fig. 2A) and CV index indicating a 1.78 average fold increase (Fig. 2B). These findings demonstrate that attenuating Hh pathway activity via Gli1 inactivation exacerbates renal cyst development and expansion in the context of Pkd1 mutations.

Fig. 2.

Attenuated Hh signaling through Gli1 inactivation increases cystic severity in Pkd1^RC/RC mice. (A) Kidney weight/body weight (KW/BW) ratio and (B) Cyst volume percentage (CV %) of 90-day Pkd1^RC/RC mouse kidneys with and without Gli1 expression. Comparison of the datasets were done using the Mann–Whitney test and presented as mean± s.e.m. ***P=0.0001; ****P<0.0001 (Mann–Whitney test).

Deletion of Gli2 and Gli3 does not affect the cystic phenotype in Pkd1 mutant mice

Given that Gli1 is not a primary driver of Hh signaling and functions as a downstream target of Gli2 and Gli3 to amplify their activity (Hui and Angers, 2011), we next investigated whether the deletion of Gli2 and Gli3, the primary modulators of Hh signaling, influences the cystic phenotype in Pkd1 mutant mice. To test this, we utilized R26CreER (Ventura et al., 2007), a tamoxifen-inducible Cre-mediated recombination system that is ubiquitously expressed, to inactivate Gli2 and Gli3 floxed genes (Corrales et al., 2006; Blaess et al., 2008). Progeny from crosses between Gli2^flox/flox; Gli3^flox/flox; Pkd1^RC/RC and ROSA26Cre^ERT/+ carriers were used to generate two groups: experimental mice (Gli2; Gli3; Pkd1 triple mutants with Cre expression) and control mice (Gli2; Gli3; Pkd1 triple mutants without Cre expression) (Fig. S2).

At P13.5, we administered tamoxifen to induce Cre-mediated recombination for Gli2 and Gli3 deletion. At 90 days, we assessed the cystic profiles of the kidneys. Both groups exhibited comparable cystic phenotypes, with the average KW/BW ratio being 2.4% in the experimental group and 2.5% in the control group (Fig. 3A), whereas the CV indices were 80.7% and 80.9%, respectively (Fig. 3B). Given that both kidneys from the experimental mice were fixed and stained for imaging, we tested a subsequent litter to assess recombination efficiency of our protocol. This revealed an average recombination rate of 21% for Gli2 and 74% for Gli3 (Table 1). These results indicate that reduction of Gli2 and Gli3 gene expression does not ameliorate cyst formation or progression in a mutant Pkd1 mouse model. However, the inefficient recombination reduces the utility of this analysis.

Fig. 3.

Attenuated Hh signaling through Gli2 and Gli3 inactivation has no effect on cystic severity in Pkd1^RC/RC mice. (A) Kidney weight/body weight (KW/BW) ratio and (B) Cyst volume percentage (CV %) of mice with and without Cre recombinase expression. Gli2^flox/flox; Gli3^flox/flox; Pkd1^RC/RC mice with or without R26CreER were administered tamoxifen at P13.5 to induce the deletion of Gli2 and Gli3. Kidneys were analyzed at 90 days. Comparison of the datasets were done using the Mann–Whitney test, presented as mean±s.e.m., and demonstrated no significant differences (Mann–Whitney test).

Table 1.

qPCR analysis of tamoxifen-mediated gene deletion in Gli2^flox/flox; Gli3^flox/flox; Pkd1^RC/RC; ROSA26Cre^ERT/+ mice

Sample	Gli2 Recombination %	Gli3 Recombination %
1	6.0	21.2
2	13.8	67.4
3	50.6	95.4
4	20.2	117.5
Average	22.6	75.4
Standard deviation	16.9	36.0

Deletion of Smo increases the cystic burden on Pkd1 mutant kidneys

To definitively characterize the role of Hh signaling in cyst formation in Pkd1 cystic kidneys, we examined the effects of inactivating Hh signaling through the deletion of Smoothened (Smo). Smo serves as a primary activator of Hh signaling through mediating canonical ligand-dependent pathway activation by accumulating at the distal tip of the primary cilium to facilitate the processing of Gli transcription factors into their active forms (Corbit et al., 2005; Rohatgi et al., 2007).

One potential confounder for this analysis is the evidence that the severity of the Pkd1^RC mutation is sensitive to strain background (Arroyo et al., 2021). We therefore introgressed a well-characterized floxed Smo allele (Long et al., 2001) into the C57BL/6J (B6) background for more than ten generations (Houghtaling et al., 2025) and mated these with Pkd1^RC/RC and ROSA26Cre^+/+ mice to generate progeny with the genotypes Smo^flox/flox; Pkd1^RC/RC; ROSA26Cre^+/+ (control) and Smo^flox/flox; Pkd1^RC/RC; ROSA26Cre^ERT/+ (experimental) (Fig. S3). Pups were administered tamoxifen at P13.5 and euthanized at 90 days to assess their phenotypes.

Quantitative (q)PCR analysis confirmed partial deletion of Smo, with an average deletion efficiency of 40% across the experimental cohort (Table 2). Despite this incomplete deletion, the mice exhibited a more-severe cystic phenotype. Compared to controls, the experimental group demonstrated a 1.24-fold increase in KW/BW ratio (Fig. 4A) and a 1.61-fold increase in CV percentage (Fig. 4B). These findings suggest that inactivation of Hh signaling through Smo deletion exacerbates, rather than mitigates, the cystic phenotype associated with the Pkd1 mutation, even when deletion is incomplete.

Table 2.

qPCR analysis of tamoxifen-mediated gene deletion in Smo^flox/flox; Pkd1^RC/RC; ROSA26Cre^ERT/+ mice

Sample	Smo Recombination %
505	67.8
552	38.0
553	20.5
557	27.2
558	23.7
565	74.1
569	46.6
Average	40.0
Standard deviation	18.8

Fig. 4.

Attenuated Hh signaling through Smo inactivation increased cystic severity in Pkd1^RC/RC mice. (A) Kidney weight/body weight (KW/BW) ratio and (B) Cyst volume percentage (CV %) of mice with and without Cre recombinase expression. Smof^lox/flox; Pkd1^RC/RC mice with or without R26CreER were administered tamoxifen at P13.5 to induce the deletion of Smo. Kidneys were analyzed at 90 days. Comparison of the datasets were done using the Mann–Whitney test and presented as mean±s.e.m. **P=0.0093; ***P=0.0001 (Mann–Whitney test).

DISCUSSION

A significant advance in the characterization of the biology related to PKD was the recognition that a number of proteins implicated in cystic kidney disease in humans and mouse models appear to have a role in cilial function (Taulman et al., 2001; Hou et al., 2002; Yoder et al., 2002). These include genes causal for disorders such as Bardet–Biedl syndrome (BBS) and nephronophthisis (NPHP) (Hildebrandt et al., 2011). In most cases, cystic disease is associated with loss-of-function mutations.

Given this, an elegant study in which the disruption of cilia was found to suppress cyst growth in Pkd1 and Pkd2 conditional mutants was a surprise (Ma et al., 2013). Those authors conclude from their experiments that there is a “polycystin-inhibited, cilial dependent proliferative cyst-promoting pathway”. This is of note, as inhibition of this presumptive pathway would be a potential therapeutic avenue for PKD.

Importantly, we had good evidence that Hh signaling should be evaluated as a candidate for this cilial-dependent, proliferative cyst-promoting pathway. This hypothesis is based on our studies of PKD in a very different context; specifically, as a consequence of the perinatal deletion of mouse Ttc21b, the mammalian ortholog of IFT139. We originally discovered Ttc21b in a screen for N-ethyl-N-nitrosourea (ENU)-induced mutations that affect organogenesis. Our studies revealed that a mutation in this cilial gene resulted in defective retrograde intraflagellar transport and overactive Hh signaling (Tran et al., 2008). The constitutive mutation had severe developmental defects, consistent with the known importance of Hh signaling in early development, which resulted in embryonic lethality.

Surprisingly, when Ttc21b is deleted in the perinatal period, the most robust phenotypic consequence was rapidly progressive PKD (Tran et al., 2014). We further showed that genetic and pharmacologic inhibition of Hh signaling reduced cystic kidney disease in this model and in kidney explants carrying mutations of Nek8 or Pkd1. Our inference that Hh signaling mediates cystic disease was supported by examination of renal epithelium from human individuals with ADPKD (Silva et al., 2018). A role for Hh signaling in exacerbating cystic disease was shown in additional studies, although these did not utilize polycystin mutant models (Hsieh et al., 2022; Sato et al., 2018).

A prior study by Ma et al. tested Hh pathway inactivation in a Pkd1 conditional mouse model and reported no significant evidence of Hh signaling mediating renal cystogenesis (Ma et al., 2019). Our approach differed by utilizing a Pkd1 hypomorphic model with constitutive disease, a larger sample size for all tested genotypes and a micro-CT-based method to comprehensively assess cystic profiles. As noted previously, there is no ideal mouse model for polycystin-mediated PKD, and it seems reasonable (and arguably necessary) to test the CDCA hypothesis in different mutants.

We established that micro-CT imaging is a reliable tool for characterizing and analyzing cystic kidneys by comparing it to the conventional kidney weight-to-body weight (KW/BW) ratio when defining the 90-day developmental period of Pkd1^RC/RC mice. Using this validated approach, we found that Gli1 inactivation led to a significant increase in cystic severity as measured by both KW/BW ratio and micro-CT analysis. Conversely, conditional deletion of Gli2 and Gli3 resulted in no significant change. qPCR assessment revealed a 3-fold higher recombination rate of Gli3 compared to Gli2, the prominent activator the Hh pathway. Additionally, deletion occurs stochastically, with variability spatially within the kidney and cellularly for both genes, which could confound the interpretation of the results.

To more conclusively assess whether the Hh pathway mediates the hypothesized CDCA signal, we tested the consequences of deleting Smo, a key mediator of canonical Hh signaling, on disease severity in Pkd1^RC/RC mice. Even incomplete deletion of Smo did not ameliorate the cystic phenotype but instead exacerbated cystogenesis. Individually, there was not a strong correlation between the percentage recombination and cystic severity; however, this might be due to the natural individual phenotypic variability that occurs in cystic kidney disease manifestation that can be obscuring the relation.

The finding that genome-wide attenuation of Hh signaling increases PKD severity could be explained by a combination of mechanisms. Increased cell proliferation and a pro-inflammatory microenvironment are well-documented drivers of cystogenesis in PKD (Torres and Harris, 2009). Moreover, studies have demonstrated increased T cell infiltration into cystic lesions in Pkd1^RC/RC mice, with disease severity correlating with the extent of inflammation (Kleczko et al., 2018). Furthermore, genetic and drug inhibition of the inflammatory response in Pkd1^RC/RC mice ameliorated disease (Swenson-Fields et al., 2022). Hh plays a role in regulating mesenchymal cell quiescence; specifically, epithelial-specific loss of Sonic Hedgehog (Shh) and mesenchymal-specific deletion of Smo in the adult lung can lead to mesenchymal cell proliferation and expansion around the airway epithelium and in alveolar interstitial regions (Peng et al., 2015). Impaired Hh signaling in CD4+ T cells has been shown to enhance inflammatory cytokine production, exacerbating neuroinflammatory conditions like multiple sclerosis and experimental autoimmune encephalomyelitis (Benallegue et al., 2021). This disruption in immune response modulation by Hh inactivation might lead to unchecked activity that confers increased inflammation and tissue damage, worsening cystic kidney disease progression. However, assessing this would be challenging, given the myriad tests of immune activity and the potential difficulty of differentiating it as a cause of or a response to cystic disease.

In summary, our findings support previous results that suggest that Hh signaling is unlikely to contribute to the hypothesized CDCA signal (Ma et al., 2019). However, the results implicate its potential role in maintaining renal epithelial and parenchyma integrity and highlight its involvement in PKD pathogenesis. Our study is the first to demonstrate a harmful outcome from Hh inactivation in cystic kidney disease. Previous reports have reported a protective effect from Hh attenuation in various cystic kidney models (Tran et al., 2014; Silva et al., 2018; Hsieh et al., 2022; Sato et al., 2018), whereas one study found no effect (Ma et al., 2019). As the role of Hh signaling in cystogenesis remains ambiguous, further study is warranted to elucidate its contribution to renal cystic disease.

MATERIALS AND METHODS

Mouse husbandry

The Pkd1tm1.1Pcha mutant line, called here Pkd1^RC was kindly provided Dr Peter Harris (Mayo Clinic, Rochester, MN, USA) (Hopp et al., 2012). The Hh pathway mutant alleles we used included Gli1^tm2Alj/J (Bai et al., 2002), Gli2^tm6Alj/J (Corrales et al., 2006), Gli3^tm1Alj/J (Blaess et al., 2008) and Smo^tm2Amc/J (Long et al., 2001; Houghtaling et al., 2025). The Gli2 and Gli3 alleles were carried in a double mutant line generously shared by Dr Samantha Brugmann (Cincinnati Children's, Cincinnati, OH, USA). Tamoxifen-induced recombination was mediated using the strain B6.129-Gt(ROSA)26Sor^{tm1(cre/ERT2)Tyj}/J (Ventura et al., 2007). Breeding strategies to generate Hh-deficient mutants and controls are shown in Figs S1–S3. This study was approved by the Seattle Children's Research Institute IACUC Committee, and we adhere to the NIH Guide for the Care and Use of Laboratory Animals.

Tamoxifen administration

Tamoxifen solution (Sigma, T5648) was made up in corn oil at a concentration of 10 mg/ml. 50 µl was injected intraperitoneally at P13.5.

Kidney processing

Ninety-day-old mice were euthanized via carbon dioxide inhalation, with gas flow maintained for 1–2 min after the cessation of respiration. Once confirmed unresponsive, a midline abdominal incision was made to expose the kidneys, which were excised and weighed. In preparation for micro-CT scanning, kidneys were fixed in 4% paraformaldehyde solution overnight at 4°C. To provide contrast for micro-CT analysis, kidneys were stained with phosphotungstic acid hematoxylin [PTAH, 2% phosphotungstic acid (Sigma, P4006), 0.1% Harris hematoxylin (Epredia, 72711)]. Kidneys were washed several times with PBS to remove residual PFA before immersion in PTAH solution until fully stained. Stained kidney samples were stored in 70% ethanol at 4°C.

For preservation for genomic DNA acquisition, kidneys were flash frozen in liquid nitrogen and stored at −80°C.

Micro-CT image acquisition and analysis

Micro-CT imaging was performed on the Bruker SkyScan 1272 CMOS Edition, with scanning parameter of 55 kV and 179 µA at a voxel size of 17.63 µm with a 0.5 mm aluminum filter. Data reconstruction was done using the NRecon software. To calculate the cystic volume percentages of the kidneys, the image reconstructions were segmented using the threshold tool in the 3D slicer program with the Slicermorph plugin (Rolfe et al., 2021). Two separate segments were created, one for the tissue volume, which excludes cystic space, and one for the total kidney volume. To create the tissue segmentation, a threshold value was chosen automatically for each sample using the auto-threshold option following the IsoData algorithm. To create the segmentation for the total kidney, the previous segment was cloned, dilated and shrunk, resulting in a fully filled segmentation with identical boundaries. The segment statistics module was then used to calculate the cystic volume ratio.

Quantitative PCR

Genomic DNA was isolated from the kidney of tamoxifen-injected mice using a PureLink Genomic DNA Mini Kit. Nucleic acid concentration was quantified using the NanoDrop ND-1000 Spectrophotometer and diluted to 10 ng/µl. qPCR was performed to detect the recombined Smo, Gli2 or Gli3 allele on a CFX96 Real-Time PCR Detection System (Bio-Rad) using SYBR Green PCR Master Mix (Applied Biosystems) with the following primers: DB898: 5′-GCGGTCTGGCAGTAAAAACTATC-3′ (ROSA f); DB899: 5′-GTGAAACAGCATTGCTGTCACTT-3′ (ROSA r); DB1073: 5′-TCCGGGGAGCTCGTTTAAAC-3′ (Smo rec f); DB1074: 5′-CCATCACGTCGAACTCCTGG-3′ (Smo rec r); DB1081: 5′-TGAAAAGCCGAATTCTGCAGA-3′ (Gli2 rec F); DB1082: 5′-TCGTGGAAACTATGCATCATGT-3′ (Gli2 rec R); DB1083: 5′-CGTGGGATCCTCTAGAGTCG-3′ (Gli3 rec F); and DB1084: 5′-AGTCCACACACATAGACCATCA (Gli3 rec R).

To calculate the percentage recombination in tamoxifen-injected mice, the recombined Smo, Gli2, and Gli3 PCR product was normalized to the ROSA control product using the ΔΔCt method. Importantly, mice homozygous for the recombined Smo allele are not viable, so genomic DNA from a Smo null/+ (heterozygous) sample was used as the control, which is equivalent to 50% total recombination. Using this as a reference, the quantification of the recombined Smo alleles in the tamoxifen-injected mice were corrected by a factor of two. Individual samples were performed in triplicate. Cycling conditions were 95°C for 3 min followed by 40 cycles of a two-step PCR of 95°C for 10 s and 60°C for 30 s.

Statistical testing

Statistical significance was determined using the Mann–Whitney test. All data is presented as mean±s.e.m. A significance threshold of 0.05 was used. Males and females were tested for all three of the experimental groups and no significant statistical differences were found, using both KW/BW and CV metrics.