c-Met and NF-κB–Dependent Overexpression of Wnt7a and -7b and Pax2 Promotes Cystogenesis in Polycystic Kidney Disease

Abstract

The mechanisms of cystogenesis in autosomal dominant polycystic kidney disease (ADPKD) are not fully understood. Hyperactivation of the tyrosine kinase c-Met contributes to cyst formation, but we do not know the downstream mediators. Here, we found that hyperactivated c-Met led to increased NF-κB signaling, which in turn, drove de novo expression of Wnt7a and overexpression of Wnt7b in Pkd1^−/− mouse kidneys. Hyperactivated Wnt signaling increased expression of the transcription factor Pax2 in the cells lining cysts. Furthermore, blocking Wnt signaling with DKK1 decreased cyst formation in an organ culture model of ADPKD. In summary, these results suggest that the c-Met/NF-κB/Wnt/Pax2 signaling transduction axis may provide pharmacological targets for the treatment of ADPKD.

Autosomal dominant polycystic kidney disease (PKD) is a commonly inherited disorder caused by mutations in either the PKD1 gene, which encodes polcystin-1, or the PKD2 gene, which encodes polycystin-2.¹ Polycystins are localized to the primary cilia (in addition to other structures) along with several additional proteins encoded by genes associated with other cystic diseases in humans.² Despite enormous study of the molecular basis for cystogenesis over the past decade, there remain large gaps in our understanding of how mutations in polycystin genes or ciliary dysfunction result in cystogenesis.

Wnt pathways are among the most important regulatory signals that drive development from insects to vertebrates.³ Wnt signaling pathways are classified as canonical or noncanonical⁴; the former is involved in stabilizing β-catenin and allowing it to translocate to the nucleus, where it associates with tcf/lef transcription factors that regulate the expression of genes affecting proliferation.⁵ There are several noncanonical Wnt signaling pathways that affect cell behavior, including those pathways that regulate planar cell polarity (PCP) and affect the levels of intracellular calcium.^6,7 In the context of tubular epithelial development and maintenance, it is hypothesized that PCP determines mitotic spindle orientation and therefore, the orientation of cell division in relation to the proximal–distal axis of a developing tubule.^8,9 Therefore, it is apparent that abnormal Wnt signaling could affect cyst formation by stimulating excessive proliferation through the canonical pathway and deregulation of PCP.^9–11

In this report, we show abnormally high expression of Wnt7b, a gene normally expressed in the developing kidney, and de novo expression of Wnt7a (also recently reported in the work by Chen et al.¹²), a gene not normally expressed in the kidney tubules, in Pkd1 knockout (Pkd1^−/−) mouse. Elevated β-catenin transcriptional activity in PKD kidneys is shown using transgenic reporter mice. We recently reported that Wnt7b gene expression in the developing kidney is under control of the c-Met receptor tyrosine kinase.¹³ We have also recently reported that c-Met levels are elevated in Pkd1^−/− kidneys and kidneys from humans with PKD.¹⁴ Here, we show that Wnt7a and -7b expression in PKD is also under control of c-Met, acting through NF-κB. We find that Pax2 expression is greatly increased in Pkd1^−/− kidneys, and its expression is driven by Wnt7a and -7b. Finally, using an embryonic kidney organ culture model for PKD, it is shown that Dickkopf-related protein 1 (DKK1) an inhibitor of Wnt signaling,¹⁵ can block cyst formation. Together, these results suggest that the Wnt signaling pathway and Pax2-mediated regulation of gene expression may present additional targets to prevent cyst formation in PKD.

Results

Overexpression of Wnt7a and -7b and Increased Canonical Wnt Signaling in Pkd1^−/− Kidneys

To determine whether misexpression of Wnt genes may be involved in the pathogenesis of PKD, we performed a survey of Wnt gene expression in Pkd1 WT and Pkd1^−/− embryonic day 17.5 (E17.5) kidneys. Most notably, Wnt7a, which is not normally expressed in the developing kidney, was ectopically expressed in Pkd1^−/− kidneys (also recently shown in the work by Chen et al.¹²), and expression of Wnt7b, a Wnt gene crucial for kidney development,^13,16 seemed to be increased in Pkd1^−/− kidneys (Figure 1A). In situ hybridization showed that, in Pkd1 WT embryos, Wnt7b is mainly expressed in collecting duct tubules, and Wnt7a was not detected. In contrast, Wnt7a and increased Wnt7b expression were detected in the cyst-lining cells of Pkd1^−/− kidneys (Figure 1B). Increased Wnt7a and -7b expression in Pkd1^−/− kidneys was also observed at E13.5 before the onset of any observable cyst formation (Figure 1C), suggesting that increased Wnt gene expression may have a causal role in cyst formation. Canonical Wnt signaling prevents serine/threonine phosphorylation of β-catenin, and therefore, immunodetection of nonphosphorylated β-catenin can be used to indicate activation of the canonical pathway. Increased total β-catenin (∼2.4-fold) and nonphosphorylated β-catenin (∼3.7-fold) were detected in Pkd1^−/− kidneys (Figure 2A). β-catenin was mainly localized at cell–cell junctions in wild-type (WT) kidneys; however, nuclear staining of total β-catenin and nonphosphorylated β-catenin was observed in Pkd1^−/− kidneys, most prominently in cyst-lining epithelia (Figure 2C). Canonical Wnt signaling is also typically shown in vivo through the use of transgenic mice that express β-galactosidase (LacZ) from a β-catenin/tcf responsive promoter.^17,18 Increased staining for β-galactosidase was observed on a Pkd1^−/− genetic background but mainly in the cyst-lining cells (Figure 2D). We note that this last result differs from the recently published findings of Miller et al.,¹⁹ which failed to observe increased β-galactosidase activity in cysts from a T cell factor (TCF)/lymphoid enhancer factor-responsive LacZ transgene in Pkd1^−/− mice. It is known that different lines of TCF transgenic mice may give different results in different tissues, and it is possible that this finding may explain our disparate observations.

Figure 1.

Expression of Wnt genes in kidneys and immortalized cells. (A) Wnt gene expression in E17.5 Pkd1 WT (WT) or Pkd1^−/− mouse kidneys measured by RT-qPCR and normalized to 18S RNA from the same sample, with the Pkd1 WT value defined as one. (B) In situ hybridization of Wnt7a and Wnt7b in Pkd1 WT and E17.5 Pkd1^−/− kidneys. (C) Wnt7a/Wnt7b and Pax2 expression detected by RT-qPCR in E13.5 Pkd1 WT and Pkd1^−/− mouse kidneys. (D) RT-qPCR detection of Wnt7a/Wnt7b expression in immortalized cell lines.¹⁴

Figure 2.

Increased canonical Wnt signaling in PKD. (A and B) Western blot of lysates from E17.5 embryonic Pkd1 WT and Pkd1^−/− (A) kidneys or (B) cells for (left) total and (right) nonphosphorylated (active) β-catenin. (C) Staining of β-catenin and nonphosphorylated β-catenin in E17.5 Pkd WT and Pkd1^−/− paraffin-embedded kidney sections. Arrows indicate β-catenin at cell–cell junctions in Pkd1 WT kidneys, and arrowheads indicate nuclear β-catenin in Pkd1^−/− kidneys. (D) β-galactosidase staining in kidneys from TCF/lymphoid enhancer factor β-catenin responsive reporter/Pkd1^−/− E17.5 embryos showed increased LacZ activity in Pkd1^−/− kidneys, especially in cyst-lining cells.

Additionally, elevated β-catenin (∼2.9-fold) and nonphosphorylated β-catenin (∼4.3-fold) were also detected in immortalized epithelial collecting duct cell lines derived from E15.5 Pkd1^−/− kidneys compared with those cell lines derived from the papilla of Pkd1 WT kidneys (Figure 2B). Using these immortalized epithelial cell lines,¹⁴ it was also possible to detect increased expression of Wnt7b and ectopic expression of Wnt7a in Pkd1^−/− cells (Figures 1D and 3, A and B). Ectopic expression of Wnt7a in Pkd1^−/− cells was also shown by Western blot (Figure 3D).

Figure 3.

Met kinase inhibitor (SU11274) and HGF neutralizing antibody. Met kinase inhibitor (SU11274) and HGF neutralizing antibody can decrease (A) Wnt7a, (B) Wnt7b, and (C) Pax2 expression in Pkd1^−/− cells and kidney explants. (Left) Pkd1 WT and Pkd1^−/− cells and (right) ex vivo kidney explants, genotypes, and treatments as noted at the bottom of the panel were treated with 5 μM SU11274 (Met inhibitor), vehicle only (DMSO), HGF neutralizing antibody, or IgG control for 48 hours. mRNAs were measured by RT-qPCR and normalized to 18S rRNA (average of three sets of biologic repeats; all figures are the average of three sets of biologic repeats). *P<0.001. (D, left) Wnt7a and (right) Pax2 protein detected by Western blot of Pkd1 WT and Pkd1^−/− cells treated with SU11274 for 48 hours.

Increased Expression of Wnt7a and -7b in PKD Is Dependent on c-Met

Our recent work showed that Wnt7b expression in collecting duct epithelial cells is normally dependent on signaling from c-Met¹³ and that c-MET signaling is hyperactivated in PKD.¹⁴ Using SU11274, a pharmacological inhibitor of c-Met,¹⁴ it was shown that increased Wnt7a and -7b mRNA expression and Wnt7a protein were also dependent on c-Met in Pkd1^−/− cells (Figure 3, A and B, left panels). SU11274 also caused a decrease in Wnt7a and -7b mRNA in cultured Pkd1^−/− kidney explants (Figure 3, A and B, right panels). Similar to the action of SU11274, a hepatocyte growth factor (HGF) neutralizing antibody also caused decreased Wnt7a and -7b expression in Pkd1^−/− cells and ex vivo kidney explants (Figure 3, A and B), suggesting an autocrine mechanism for HGF stimulation of Wnt gene expression and suggesting that c-Met–dependent regulation of Wnt gene expression seems to have been pathologically co-opted in PKD.

Pax2 Is Highly Overexpressed and Dependent on c-MET Signaling in Pkd1^−/− Cells and Kidneys

Genetic experiments have suggested a role for Pax2 in cystic disease.^20,21 We found greatly elevated levels of Pax2 mRNA in immortalized Pkd1^−/− cells (Figure 4A) and kidneys at E13.5 and E17.5 (Figures 1C and 4B) by quantitative RT-PCR (RT-qPCR) and Western blot for Pax2 in Pkd1^−/− immortalized cells (Figure 4C) and kidneys (Figure 4D). This result was validated in vivo by in situ hybridization with Pax2 probe, showing increased Pax2 expression in cyst-lining epithelial cells of Pkd1^−/− kidneys (Figure 4E), and immunostaining (Figure 4F).

Figure 4.

Increased Pax2 expression in PKD. (A and B) RT-qPCR of Pax2 expression in (A) Pkd1^−/− cells or (B) Pkd1^−/− E17.5 kidneys normalized to 18S rRNA (average of three sets of biologic repeats). **P<0.005. (C and D) Pax2 protein in (C) Pkd1^−/− cells or (D) E17.5 Pkd1^−/− kidneys compared with Pkd1 WT. (E and F) In situ detection of (E) Pax2 mRNA and (F) protein in Pkd1 WT and Pkd1^−/− kidneys.

Upregulated Pax2 Expression Is Regulated by Wnt Signaling

In Pkd1^−/− cells and ex vivo Pkd1^−/− kidney explants, Wnt agonists and an antagonist were used to evaluate whether Wnt signaling can affect Pax2 expression. Recombinant DKK1 used as a Wnt antagonist¹⁵ decreased Pax2 expression in both Pkd1^−/− cells and Pkd1^−/− kidney explants (Figure 5). Conversely, Wnt7a protein could boost Pax2 expression to even higher levels in immortalized Pkd1^−/− cells and Pkd1^−/− kidney explants (Figure 5). If elevated Wnt signaling in the absence of polycystins is a primary driving force of cyst formation in PKD, it would be predicted that stimulation of Wnt signaling would elevate Pax2 expression in Pkd1 WT cells. Indeed, treatment of Pkd1 WT cells or explants with recombinant Wnt7a protein did dramatically increase Pax2 expression 10-fold (Supplemental Figure 1), although this finding was well below the level observed in Pkd1^−/− cells or kidneys. Pax2 expression at the mRNA and protein level was decreased in the presence of SU11274 or an HGF neutralizing antibody (Figure 3, C and D), confirming the role of c-Met as an upstream component of the Wnt signals that regulate Pax2 expression in PKD.

Figure 5.

Effect of Wnt signals on Pax2 expression. (A) Pkd1 WT and Pkd1^−/− cells or (B) kidney explants were treated with recombinant 0.3 μg/ml DKK1, 5 μg/ml Wnt7a, or control solution (0.1% bovine serum albumin in PBS). All comparisons marked by the brackets were significant. **P<0.005; ***P<0.01. Supplemental Figure 1 shows same data with Pkd1 WT and Pkd1^−/− cells and kidney explants at different scales to visualize differences in Pkd1 WT cells and explants.

Elevated Wnt and Pax2 Expression May Not Be Downstream of Mammalian Target of Rapamycin

A major area of PKD research currently focuses on the role of mTOR,²² and it has been shown that mammalian target of rapamycin (mTOR) can be activated downstream of c-Met.^14,23 However, treatment with rapamycin did not affect Wnt7a, Wnt7b, or Pax2 mRNA levels in Pkd1^−/− cells (Supplemental Figure 2). As a positive control, rapamycin did cause a decrease in S6-Kinase phosphorylation in Pkd1^−/− cells (Supplemental Figure 2D). Thus, the c-Met regulation of Wnt gene expression seems to be independent of mTOR, showing that multiple signal transduction pathways, both mTOR-dependent and -independent, may be operative in cystic disease.

NF-κB Pathway Is Hyperactivated by c-Met in Pkd1^−/− Cells and Regulates Wnt Gene Expression

Akt is hyperactivated in PKD downstream of c-Met and phosphoinositide 3-kinase.^14,24,25 In turn, the NF-κB family of transcription factors is a known mediator acting downstream of Akt.^26–28 Luciferase reporter activity expressed from an NF-κB–responsive promoter was increased over fivefold in Pkd1^−/− cells compared with Pkd WT cells and was sensitive to SU11274 (Figure 6A). Activation of the NF-κB pathway results in phosphorylation and nuclear translocation of the p65 subunit.²⁹ Increased basal phosphorylation of p65 (∼4.6-fold) and its nuclear localization were observed in Pkd1^−/− cells and augmented on stimulation with HGF, the ligand of c-Met (Figure 6, B and C). Nuclear localization of p65 was also increased in cyst-lining cells in Pkd1^−/− kidneys, which was detected by immunohistochemistry (Figure 6D).

Figure 6.

Hyperactivation of NF-κB in PKD. (A) Luciferase reporter assay showed increased NF-κB activity in Pkd1^−/− cells, which can be inhibited by 5 μM SU11274. (B) NF-κB p65 phosphorylation in Pkd1 WT and Pkd1^−/− cells detected by Western blot before and after treatment with HGF at 50 ng/ml for 1 hour. (C) Immunofluorescent staining of p65 in Pkd1 WT and Pkd1^−/− cells before and after treatment with HGF. (D) Immunohistochemistry of p65 in E17.5 kidney paraffin-embedded Pkd1 WT and Pkd1^−/− kidneys at (upper) lower and (lower) higher magnification. (E and F) RT-qPCR for Wnt7a and -7b and Pax2 in (E) Pkd1 WT and Pkd1^−/− cells or (F) kidney explants (E13.5 kidneys cultured for 3 days with media changed daily) treated with IKKγ NEMO binding domain inhibitor peptide at 100 μM for 48 hours. (G) Hematoxylin/eosin sections of E13.5 Pkd1 WT and Pkd1^−/− mice kidney explants cultured with 100 μM 8-Br-cAMP for 1 day, after which either IKKγ NEMO binding domain inhibitor peptide or control peptide was added for an additional 2 days with media changed daily. Quantification of cyst area is shown in right. (H) Higher magnification of G showing decreased cyst formation in the presence of the NF-κB inhibitory peptide but little effect on nephrogenesis, which was evidenced by the normal-appearing S-shaped tubule (representative of four biologic repeats). ***P<0.01.

To examine whether NF-κB regulates Wnt7a and -7b and Pax2 expression, we treated Pkd1 WT and Pkd1^−/− cells with IKKγ NEMO binding domain inhibitor peptide (Figure 6E) or NF-κB p65 (Ser276) inhibitory peptide (not shown); both are inhibitors of NF-κB pathway, and they decreased the expression of Wnt7a and -7b and Pax2. Dependency of Wnt7a and -7b and Pax2 expression on NF-κB was also shown in the ex vivo kidney explants, where both the IKKγ NEMO binding domain inhibitor peptide (Figure 6F) and NF-κB p65 inhibitory peptide (not shown) reduced Wnt7a and -7b and Pax2 expression and decreased cyst formation in Pkd1^−/− kidney explants (Figure 6G). These results show that NF-κB is an important component of the c-Met–Wnt –Pax2 axis that is dysregulated in PKD.

Wnt Blockade Limits Cystogenesis in Organ Culture

To show the close association between hyperactivation of Wnt signaling and cystogenesis in PKD, we analyzed the effect of a Wnt antagonist and agonist on cystogenesis of ex vivo embryonic kidney explants treated with 8-bromo-cyclic adenosine monophosphate (8-Br-cAMP) to induce cyst formation.^14,30 Treatment of cultured kidney explants with recombinant DKK1 to block canonical Wnt signaling¹⁵ significantly limited cyst formation in Pkd1^−/− kidney explants (Figure 7). Cyst area decreased from 29%±6% of the total organ culture area in control organ cultures kidney to 12%±1% in DKK1-treated kidney organ cultures (Figure 7B; P=0.001). Conversely, when recombinant Wnt7a was added to the organ culture media, larger cysts were formed in both Pkd1 WT and Pkd1^−/− kidney explants, increasing the cyst area in Pkd1^−/− kidney explants to 58%±7% (Figure 7; P=0.001) and the cyst area in Pkd1 WT kidneys from 7%±0.5% to 18%±3% (Figure 7B; P=0.005). Decreased cyst formation after treatment with DKK1 was not caused by increased cell death (Supplemental Figure 3).

Figure 7.

Effect of a Wnt inhibitor or agonist on cyst formation. (A) Hematoxylin/eosin-stained sections of E13.5 kidney explants cultured for 4 days in the presence of 100 μM 8-Br-cAMP and control, 0.3 μg/ml recombinant DKK1, or 5 μg/ml recombinant Wnt7a. (B) Quantitative analysis of three sets of biologic replicates as in A. DKK1 can decrease the cyst area in Pkd1^−/− kidneys from 29%±6% to 12%±1%, whereas Wnt7a can increase cyst area to 58%±7%. **P=0.005; *P=0.001.

Expression of Wnt7a and -7b and Pax2 Is Stimulated by 8-Br-cAMP

Elevated cAMP signaling is thought to be a major factor in cyst formation in autosomal dominant PKD,³⁰ and treatment with 8-Br-cAMP is required to obtain cysts in explants of embryonic kidneys.^14,30 Therefore, we examined whether increased Wnt and Pax2 gene expression in Pkd1^−/− kidney explants was affected by cAMP. Increased Wnt7a and -7b expression in Pkd1^−/− explants was observed in the absence of 8-Br-cAMP, but addition of 8-Br-cAMP did result in increased Wnt gene expression (Figure 8). Increased Wnt gene expression in untreated Pkd1^−/− explants may reflect increased levels of endogenous cAMP (which are insufficient to induce obvious cysts). Alternatively, these results may indicate that Wnt gene expression reflects an integration of cAMP-dependent and -nondependent pathways. However, 8-Br-cAMP did not boost Wnt gene expression in Pkd1 WT kidneys, indicating that the effect of cAMP is specific to Pkd1^−/− explants (Figure 8).

Figure 8.

cAMP effect on Wnt7a and -7b and Pax2 expression. (A) Pkd1 WT and Pkd1^−/− cells or (B) E13.5 embryonic kidney explants were treated with 100 μM 8-Br-cAMP or vehicle for 72 hours followed by examination of Wnt7a and -7b and Pax2 expression by RT-qPCR; genotypes and treatments are as noted at the bottom of the panels.

Discussion

PKD is a complex disease, and there is evidence for several signal transduction pathways being involved in cyst formation, including mean arterial pressure kinase, mTOR, calcium-dependent, cAMP, and Wnt signals. A role for Wnt signaling in the induction of the nephron during embryonic kidney development is well established.^31–33 Because cyst formation is sometimes regarded as an abnormal developmental process, a role for Wnt signaling in PKD has been the subject of many investigations.^4,34 Our results show increased Wnt7b expression and de novo Wnt7a expression in E13.5 embryonic Pkd1^−/− kidneys before the onset of cyst formation. Thus, Wnt signals may be involved in the initiation of cystogenesis, as well as the mediation of the progression of cystic disease, which is consistent with recently published results showing expression of known β-catenin target genes in Pkd1^−/− kidneys.³⁵ We also showed that inhibition of canonical Wnt signaling by DKK1 may decrease cyst formation and that the addition of Wnt7a to Pkd1 WT explants exacerbated cyst formation, lending additional support to previously published studies suggesting a role for canonical Wnt signaling in PKD.^36,37 For example, transgenic expression of a constitutively active β-catenin gene led to cyst formation in the kidney³⁸ as did conditional mutation of the adenomatous polyposis coli gene,³⁹ which would also result in higher levels of β-catenin. More recently, it was shown that Inversin, a protein encoded by a gene mutated in another cystic disease (nephronophthisis type II), may be involved in a switch between canonical and noncanonical Wnt signaling.⁴⁰ There is increasing evidence that cyst formation may involve abnormal manifestations of PCP pathways, resulting in randomized mitotic spindle orientation.^8,9,41 Bardet–Biedl syndrome involves cyst formation in the kidney and results from mutation of several genes, all of which encodes proteins that localize to the cilia and basal body.⁴² Mice deficient in Bardet–Biedl syndrome-causing genes display phenotypes consistent with abnormal regulation of PCP,⁴³ and the encoded proteins associate with proteins such as Vangl2, which are components of known PCP pathways in Drosophila.⁴⁴ Most recently, mutation of a protocadherin, Fat4, led to cysts within the kidney, which also seemed to result from abnormalities in PCP-related signaling pathways.¹¹

The relationship between cAMP signaling and Wnt signaling in PKD seems to be complex. Cyst formation in explants of embryonic kidneys is dependent on the addition of 8-Br-cAMP.³⁰ 8-Br-cAMP can also drive cyst formation in Pkd1 WT cultures, although the effect is much more pronounced in Pkd1^−/− explants.³⁰ 8-Br-cAMP boosts expression of Wnt7a and -7b but only in Pkd1^−/− explants. This finding is consistent with other published studies suggesting that the response to cAMP is very different in Pkd1^−/− cells and Pkd1 WT cells.^45,46 This observation also raised the question of whether the c-Met/NF-κB/Wnt/Pax2 axis is directly downstream of cAMP-dependent signals. Although we have not measured endogenous cAMP levels in kidney explants, our observations indicate that increased Wnt7a and -7b expression in Pkd1^−/− explants, although boosted by 8-Br-cAMP, was not dependent on 8-Br-cAMP, suggesting that increased Wnt expression may be primarily downstream of a non-cAMP–related pathway but that signaling through this pathway is augmented by cAMP-dependent signals. Even if increased Wnt and Pax2 gene expression is, to some extent, downstream of cAMP signals, the inability of Wnt7a to induce cyst formation in the absence of 8-Br-cAMP, even in Pkd1^−/− explants,^14,30 makes it clear that cAMP must regulate other pathways that are also involved in cyst formation. These pathways may relate to the effect of cAMP on fluid secretion or cell proliferation that may not be regulated by Wnt signaling. However, although Wnt7a cannot induce cysts in the absence of 8-Br-cAMP, 8-Br-cAMP–induced cyst formation and Pax2 expression could be almost completely blocked by the addition of DKK1, an inhibitor of canonical Wnt signaling. Thus, these observations suggest that increased canonical Wnt signaling is necessary but not sufficient for cyst formation in PKD.

Most cystic kidney diseases are considered to be ciliopathies,^47,48 and recent studies have shown involvement of primary cilia in Wnt signaling.^49,50 Increased Wnt7a and -7b expression was also observed in mice carrying a mutation in IFT20,⁵¹ which encodes a protein involved in assembly of cilia.^52,53 Our previous work showed that increased c-Met expression was caused by sequestration of the c-Cbl E3 ubiquitin ligase in the Golgi.¹⁴ The fact that abnormal Wnt7a and -7b expression may result from both abnormal protein trafficking and ciliary dysfunction suggests that either events in the cilia are affecting protein trafficking in the cilia or conversely, abnormal trafficking in the Golgi affects the cilia. Because IFT20 is also found in the Golgi, our studies and others^52,54–56 suggest a functional link between the Golgi and primary cilia that may be broadly involved in cyst formation.

The Pax2 gene is also known to have an important role in PKD. Pax2 is expressed in developing collecting ducts.⁵⁷ On a Pax2 heterozygous genetic background, cyst formation is reduced in kidneys of homozygous Pkd1 mutant or cpk mutant mice.^21,58 Of the two studies that observed this effect of Pax2 haploinsufficiency, one found that it was associated with decreased proliferation, and the other observed increased apoptosis in the cystic epithelium. Conversely, transgenic overexpression of Pax2 results in cyst formation in kidneys.^20,21,58 Pax2 has recently been shown to be part of a H3K4 methyltransferase complex,⁵⁹ which is known to activate gene expression.⁵⁹ The finding that Pax2 is involved in epigenetic modification of gene expression may provide additional insight as to why PKD is a chronic disease as well as help explain why cyst formation may be relatively silent for years⁶⁰ and then be activated and become quite dramatic. This finding may reflect the activation of an epigenetic switch that changes a broad pattern of gene expression, which results in the development of cysts. The identification of Pax2 regulatory targets may bring us closer to an understanding of the basic causes of cyst formation.

Concise Methods

Reagents

Antibodies were monoclonal anti-mouse β-catenin (610153; BD Transduction Laboratories), anti-mouse active β-catenin (clone 8E7; Millipore), rabbit anti-mouse NF-κB p65 (sc-109; Santa Cruz), rabbit anti-mouse phospho–NF-κB p65 (sc-101748; Santa Cruz), rabbit anti-mouse Pax2 (G. Dressler, University of Michigan, Ann Arbor, MI), mouse monoclonal anti-mouse α-tubulin (Sigma-Aldrich), goat anti-Wnt7a (sc-26361; Santa Cruz), rabbit monoclonal anti-S6K (2708; Cell Signaling), rabbit monoclonal antiphospho-S6K (9205; Cell Signaling). DKK1 (5897-DK-010; R & D Systems), Wnt7a (3008-WN-025; R & D Systems), HGF neutralizing antibody (AF-294-NA; R & D Systems), ApopTag Plus Fluorescein In Situ Apoptosis Detection Kit (S7111; Millipore), And Dual-Glo Luciferase Assay System (Promega). Unless otherwise stated, all other chemicals were purchased from Sigma-Aldrich. Met inhibitor (SU11274), Calbiochem (448101), IKKγ NEMO Binding Domain Inhibitor Peptide (IMG-2000; IMGENEX), and NF-κB p65 (Ser276) Inhibitory Peptide (IMG-2001; IMGENEX).

Cells and Mice

All mice experiments were approved by the Institutional Animal Care and Use Committee in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Pkd1^−/− mice were previously described.⁶¹ Mice carrying a Tcf/β-catenin–responsive reporter were obtained from Benjamin Alman⁶² (University of Toronto, Toronto, ON, Canada). Pkd1 WT and Pkd1^−/− cell lines (used at passages 9–14) were previously described.^63,64

Ex Vivo Kidney Explants Organ Culture

Timed pregnancies were generated by intercrossing mice heterozygous for the Pkd1^+/- allele. E13.5 mice kidneys were dissected out and cultured as described.³⁰ Kidneys from the same embryo were treated with specific additive or the vehicle solution as controls in the following days of kidney culture. Embryonic kidneys were cultured for 2–5 days, with media changed daily.

RT-qPCR

Real-time RT-qPCR was carried out on Smart Cycler II. SyBR Green was used for fluorescence detection. PCR parameters were 95°C for 2 minutes and 95°C for 15 seconds, 56°C for 30 seconds, and 72°C for 30 seconds for 40 cycles; melting temperature was measured between 60°C and 95°C. The specific mRNA amount was normalized by the 18S rRNA amount from the same cDNA sample. The primers used are listed in Supplemental Data.

Immunohistochemistry Staining

Paraffin-embedded kidney sections were deparaffinized through xylene, xylene/ethanol, and ethanol, hydrolyzed with water, and followed by antigen retrieval through boiling in a water bath in a container for 30 min with the antigen retrieval solution (H-3300; Vector Laboratories Inc). Block with Tris-buffered saline/10% goat serum/1% bovine serum albumin for 1 h was followed by primary antibody (anti–β-catenin or active β-catenin; 1:100) in 1% bovine serum albumin/Tris-buffered saline. After incubation with 1:100 horseradish peroxidase-conjugated secondary antibody, diaminobenzidine (Sigma) was added to the slides for color development.

In Situ Hybridization

Tissue in situ hybridization was performed as described previously.⁶⁵ Riboprobes were generated in the following region: Wnt7a from the 3′-untranslated region and Wnt7b from the 3′-untranslated region of the mouse genome, which is conserved in both transcripts of Wnt7b. The primer sequence for generation of these probes is listed in Supplemental Data. After the PCR product was subcloned into pCRII-TOPO vector (Invitrogen), sense and antisense probes were synthesized and labeled with digoxigenin-uridine 5'-triphosphate (Roche, Mannheim, Germany).

β-Galactosidase Staining

For β-galactosidase staining, 8- to 10-μM cryosection slides from embryos E17.5 kidneys were fixed in 0.2% glutaraldehyde/2% formaldehyde/2 mM MgCl₂ for 5–10 minutes, washed three times in PBS, and followed by incubation in 1 mg/ml X-gal/5 mM potassium ferricyanide/5 mM potassium ferrocyanide at room temperature.

Luciferase Reporter Assay

Luciferase reporter assay was performed using the manufacturer’s protocol. Briefly, Pkd1 WT and Pkd1^−/− cells were transfected with NF-κB firefly luciferase reporter plasmid and Renilla luciferase plasmid at 50%–70% confluence. Cells were collected 36–48 hours after transfection and lysed in passive lysis buffer provided by the kit manufacturer. NF-κB firefly luciferase activity was measured and normalized by Renilla luciferase activity.

Statistical Analyses

All data are presented as mean ± SEM. Two-tailed t test for unpaired groups was used to compare the mean of different groups. The difference between two means was significant when P was less than 0.01.

Disclosures

None.