Activation of podocyte Notch mediates early Wt 1 glomerulopathy

Rowan I. Asfahani, Mona M. Tahoun, Eve V. Miller-Hodges, Jack Bellerby, Alex K. Virasami, Robert D. Sampson, Dale Moulding, Neil J. Sebire, Peter Hohenstein, Peter J. Scambler and Aoife M. Waters Programme of Developmental Biology of Birth Defects, Great Ormond Street Institute of Child Health, University College of London, London, UK; Clinical and Chemical Pathology Department, Faculty of Medicine, Alexandria University, Alexandria, Egypt; MRC Institute of Genetics and Molecular Medicine, University of Edinburgh, Edinburgh, Scotland; Great Ormond Street Hospital NHS Foundation Trust, London, UK; Institute of Ophthalmology, University College of London, London, UK; and The Roslin Institute, Edinburgh, Scotland

The Wilms' tumor suppressor gene, WT1, encodes a zinc finger protein that regulates podocyte development and is highly expressed in mature podocytes. Mutations in the WT1 gene are associated with the development of renal failure due to the formation of scar tissue within glomeruli, the mechanisms of which are poorly understood. Here, we used a tamoxifen-based CRE-LoxP system to induce deletion of Wt1 in adult mice to investigate the mechanisms underlying evolution of glomerulosclerosis. Podocyte apoptosis was evident as early as the fourth day post-induction and increased during disease progression, supporting a role for Wt1 in mature podocyte survival. Podocyte Notch activation was evident at disease onset with upregulation of Notch1 and its transcriptional targets, including Nrarp. There was repression of podocyte FoxC2 and upregulation of Hey2 supporting a role for a Wt1/ FoxC2/Notch transcriptional network in mature podocyte injury. The expression of cleaved Notch1 and HES1 proteins in podocytes of mutant mice was confirmed in early disease. Furthermore, induction of podocyte HES1 expression was associated with upregulation of genes implicated in epithelial mesenchymal transition, thereby suggesting that HES1 mediates podocyte EMT. Lastly, early pharmacological inhibition of Notch signaling ameliorated glomerular scarring and albuminuria. Thus, loss of Wt1 in mature podocytes modulates podocyte Notch activation, which could mediate early events in WT1-related glomerulosclerosis.
G lomerulosclerosis accounts for 5% to 10% of pediatric and adult end-stage kidney disease and recurs in 15% to 30% of patients following kidney transplantation. 1 Mutations in genes encoding transcription factors which regulate podocyte differentiation have been implicated in human glomerulosclerosis and include Wilms' tumor 1 (WT1). 2-7 WT1 encodes a nuclear protein containing 4 zinc fingers that bind DNA and RNA and is highly expressed in mature podocytes. [8][9][10][11] Mutations in the regions involved in DNA binding or zinc finger formation have been reported in patients with Denys-Drash syndrome, a disorder associated with infantile diffuse mesangial sclerosis, gonadal dysgenesis, and Wilms' tumor. [12][13][14] Mutations disrupting an alternative splice donor site in intron 9, results in Frasier syndrome, a disorder associated with focal segmental glomerulosclerosis (FSGS), predisposition to gonadoblastoma, and male pseudohermaphroditism. 15,16 Furthermore, mutations in WT1 have been reported in isolated primary steroid-resistant nephrotic syndrome, supporting a role for aberrant Wt1 function in the pathogenesis of glomerulosclerosis. 5,7 Mechanisms underlying the development of WT1-related glomerulosclerosis are poorly understood with no effective treatments available. Recent studies employing chromatin immunoprecipitation sequencing, RNA sequencing, exon array, and bioinformatic analyses have shown a specific role for WT1 in regulating the podocyte-specific transcriptome. WT1 can bind to both promoters and enhancers of 18 known podocytopathy genes. 17,18 Mutations in WT1 could lead to altered transcriptional regulation of these genes, which is necessary for podocyte differentiation and function, and contribute to the pathogenesis of glomerulosclerosis. 17,18 Genes regulating podocyte differentiation have also been implicated in the pathogenesis of WT1-related disease. De novo expression of PAX2 protein and mRNA, a paired box transcription factor repressed by WT1 during nephrogenesis, has been observed in podocytes of Denys-Drash syndrome patient biopsies. 19 Furthermore, re-expression of Pax2 in podocytes, cell cycle re-entry, and reduced expression of podocyte proteins such as nephrin and a-actinin-4 have been reported in heterozygous Wt1 tmT396/þ mice with glomerulosclerosis by 8 months of age. 20 These findings suggested that Wt1 tmT396/þ podocytes dedifferentiate and revert to an immature phenotype during disease progression. 20 During glomerulogenesis, podocyte fate induction is regulated by Notch, a highly conserved and ubiquitous pathway that transduces short-range signals between neighboring cells. [21][22][23][24] Ligand binding initiates regulated intramembrane proteolysis with subsequent nuclear translocation of the Notch intracellular domain (NICD) where it associates with RBPJ-k, a DNAbinding protein, and promotes transcription of target genes (e.g., hairy enhancer of split [Hes]), which regulates tissuespecific differentiation. Notch1, Notch2, and downstream transcriptional targets Hes1 and Hey1 are expressed in podocyte precursors in the S-shaped body and are down-regulated during terminal differentiation. 25,26 Ectopic podocyte Notch activation in differentiating and mature podocytes is associated with both diffuse mesangial sclerosis and FSGS, respectively, the latter being mediated by p53-induced podocyte apoptosis and the former associated with de novo Pax2 expression. 27,28 Given the interplay between WT1 and Notch during glomerulogenesis and the fact that both diffuse mesangial sclerosis and FSGS phenotypes occur with mutations in WT1 as well as ectopic podocyte Notch activation, we hypothesized a role for podocyte Notch activation in the development of glomerulosclerosis related to loss of Wt1 function.
In a model where Wt1 was deleted in mature podocytes, using a tamoxifen-based CRE-LoxP system, we observed increased podocyte loss during the development of glomerulosclerosis. At disease onset, we found upregulation of Notch pathway transcripts in mutant podocytes. At the same time point, we observed repression of FoxC2 and upregulation of Hey2 transcripts in primary mutant podocytes. Cleaved Notch1 and HES1 proteins were evident in mutant podocytes at disease manifestation. Induction of podocyte HES1 expression in vitro was associated with increased expression of Slug and Snail transcripts, genes implicated in epithelial to mesenchyme transition. Pharmacological inhibition of Notch with gamma-secretase inhibitors at disease onset rescued the severity of glomerulosclerosis and albuminuria. These data support a role for early Notch activation in the manifestation of Wt1 glomerulopathy, which may be mediated via repression of podocyte FoxC2.

RESULTS
Early glomerulosclerosis is evident at 5 days post tamoxifen induction in CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice Wt1 deletion in mature podocytes results in glomerulosclerosis with compromised renal function by day 7 post tamoxifen induction in adult CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice. 29 To investigate events leading to the induction of disease in these mice, we first determined the earliest point at which we could detect glomerulosclerosis after Wt1 deletion. Tamoxifen was administered for 3 consecutive days by i.p. injection to 5-week-old CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice and mice were nephrectomized at 4, 5, 6, and 12 days following injection. Successful Wt1 deletion was demonstrated by recombination polymerase chain reaction (PCR) and the reduction of Wt1 expression in glomeruli (Supplementary Figure S1). Following light microscopy analysis of periodic acid-Schiff (PAS)-stained kidney sections, we determined the severity of glomerulosclerosis by a semiquantitative analysis at each time point in CAGG-CreER TMþ/À ;Wt1 f/f mutants, CAGG-CreER TMÀ/À ;Wt1 f/f controls, and heterozygous CAGG-CreER TMþ/À ;Wt1 f/þ mice (Figure 1a-d).
Heterozygous mice did not develop glomerulosclerosis by day (D) 12 postinduction (PI) (Supplementary Figure S2). At D4 PI, mutants exhibited early segmental glomerulosclerosis with focal foot process effacement and a trend toward higher levels of albuminuria (P ¼ 0.21) (Figure 1a and a 0 ; Supplementary Figures S3A and S4). By D5 PI, glomerular scarring was more extensive ($ score 2) in mutants compared with control and heterozygous mice (*P < 0.05) (Figure 1b Figure S5). We conclude that podocyte function appears compromised within 6 days of Wt1 deletion in mature podocytes. Therefore, investigations into the mechanisms underlying manifestation of disease should be undertaken within this time frame.
Podocyte apoptosis has previously been implicated in the pathogenesis of glomerulosclerosis. 30 Therefore, we next investigated whether apoptosis is evident during disease induction in CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice. Expression of Cleaved Caspase-3 protein was observed within mutant glomeruli as early as D4 PI but quantitatively, was not statistically significantly different compared with control glomeruli (Figure 2a) (P ¼ 0.06, NS). By D5 PI, at the onset of early glomerulosclerosis, we observed an increased number of Cleaved Caspase-3/DAPI-positive cells in mutant glomeruli (Figure 2c, c 0 , and d) (***P < 0.0001), consistent with a temporal increase in apoptosis. Terminal deoxynucleotidyltransferase-mediated 2 0 -deoxyuridine 5 0triphosphate nick end-labeling (TUNEL)-positive mutant glomerular cells were evident at D5 PI but were absent in controls (Supplementary Figure S6). TUNEL staining of D6 PI primary mutant podocytes was also higher than for controls (Figure 2e Figure S9). Together, these studies suggest that loss of Wt1 in mature podocytes is associated with podocyte apoptosis and development of glomerulosclerosis.
Podocyte Notch activation precedes glomerulosclerosis in CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice Constitutive Notch activation in terminally differentiated podocytes results in podocyte apoptosis and glomerulosclerosis, 27,28 supporting a role for Notch in podocyte injury. Following our observation of increased podocyte apoptosis in CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice, we hypothesized that podocyte Notch activation plays a role in early Wt1 glomerulopathy. At D4 PI before manifestation of disease, Notch pathway transcripts in primary podocytes of CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice were not statistically different to those of controls (Supplementary Figure S10). By D6 PI, concomitant with early glomerulosclerosis and albuminuria, we observed increases in canonical Notch pathway transcripts, Notch1, Nrarp, and Hey2. Increased  during signal transduction. 31,32 We also found upregulation of Pofut1 protein in primary mutant podocytes at D6 PI (Figure 3e). Pofut1 is an O-fucosyltransferase 1 enzyme that mediates fucosylation of the extracellular domain of the Notch. 33 Together, these data suggest that post-translational modification of Notch components are active in early Wt1 glomerulopathy.
We next sought to determine whether the Notch ligand Jagged1 is expressed in CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice owing to previous studies implicating podocyte Jagged1 in glomerulosclerosis. 28 We observed a striking upregulation of Jagged1 protein in the parietal epithelium of mutants compared with controls (Figure 4a and b). Within segments of mutant glomeruli, foci of Podocin-positive cells expressing Jagged1 were also observed (Figure 4b 0 , high-power images). Immunoblotting of primary mutant and control podocyte lysates revealed induction of Jagged1 expression at D6 PI (Figure 4c). Together, these data suggest that Notch signaling is activated when glomerulosclerosis first manifests following deletion of Wt1 in mature podocytes.
Given previous studies that showed that Wt1 and Foxc1a can inhibit NICD activation, 34 we explored FoxC2 expression in this model. Using semiquantitative PCR analysis, we found decreased podocyte FoxC2 transcript in CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice compared with their control littermates (*P ¼ 0.04) (Figure 3a). This decrease was observed at the same time point when Notch components were upregulated and concordant with disease manifestation. These findings support the hypothesis that Notch activation in mature podocytes is normally repressed by FoxC2 and WT1 and loss of these proteins is associated with increased podocyte Notch activation.
Podocyte HES1 expression is evident at onset of glomerulosclerosis and is associated with podocyte epithelial to mesenchyme gene expression Following detection of podocyte Hes/Hey mRNA expression in primary mutant podocytes, we sought to validate the expression of HES1, a bHLH transcription factor in podocytes by triple immunofluorescence labeling. We observed clusters of HES1-expressing cells that were positive for the podocyte marker, Synaptopodin, at onset of glomerulosclerosis in mutants compared with controls ( Figure 5). HES1-positive glomerular epithelial cells were observed in regions distinct from HES1-positive Lotus tetragonolobus lectin (LTL)-positive tubules (Figure 5b and b 0 ). Clusters of Synaptopodin-positive HES1-podocytes were distinct from platelet-endothelial cell adhesion molecule-positive glomerular endothelial cells (Figure 5c, d, and d 0 ). As HES1 has been previously implicated in epithelial to mesenchymal transition (EMT), we next determined that Snail and Slug transcript was upregulated in primary D6 PI mutant podocytes compared     (Figure 5e). To further understand the role of HES1 in podocytes, we transfected cultured primary Nphs2;rtTA transgenic murine podocytes with constructs expressing either TetOHes1 or green fluorescent protein (GFP) alone. HES1 was not expressed in untreated TetOHes1 nor doxycycline-treated GFP transfected Nphs2;rtTA podocytes. Treatment with doxycycline led to a dose-dependent increase in podocyte Hes1 mRNA and protein expression (Figure 5f, Supplementary Figure S11). Induction of Hes1 expression led to a 3-fold upregulation of Snail and Slug transcript compared with untreated TetOHes1 transfected and doxycycline-treated GFP transfected Nphs2;rtTA podocytes (Figure 5f). These results suggest that podocyte HES1 induction could mediate manifestation of glomerulosclerosis through regulation of EMT genes in podocytes.
Following our observation of Notch activation in murine Wt1 glomerulopathy, we next tested biopsy samples from a human subject with FSGS associated with the WT1 c.1390G>T mutation. We found JAGGED1 expression in cells with focal nephrin staining but also found that expression was most marked in the parietal epithelium ( Figure 6). The control biopsy from a nondiseased time 0 renal allograft biopsy revealed JAGGED1 expression in regions distal to nephrin staining suggesting endothelial JAGGED1 expression. JAGGED1 expression in the control parietal epithelium was much less than in mutant WT1 patient tissue. Expression of the Notch bHLH transcription factor, HES1, was also observed in nuclei of mutant WT1 glomeruli compared with in control biopsy tissue ( Figure 6). We conclude that these data support a role for Notch activation in human WT1-mediated glomerular disease.
Gamma secretase inhibition of Notch ameliorates disease severity in Wt1 glomerulopathy As podocyte Notch activation is a feature of early disease, we hypothesized that pharmacological Notch inhibition at onset of glomerulosclerosis could influence severity of disease. CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice were treated by i.p. injection with the gamma secretase inhibitor GSI-IX, N-[N-(3, 5-Difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester (DAPT), or dimethylsulfoxide (i.e., vehicle only) late D4 PI and early D5 PI (Figure 7a-d). In vehicle-treated mutants, we observed hyaline-filled tubules representing proteinaceous material and sclerotic glomeruli with mesangial proliferation (Figure 7b). These features were not evident in GSI-IX-treated mutants (Figure 7c). The proportion of vehicle-treated mutant glomeruli exhibiting extensive glomerulosclerosis was significantly higher than the proportion of GSI-IX-treated mutants (**P ¼ 0.008) (Figure 7d). We validated inhibition of canonical Notch pathway transcripts by semiquantitative PCR (Figure 7e). Analysis of Wt1 transcript and protein expression did not reveal a significant difference between vehicle-treated or GSI-IX-treated mutant mice (Supplementary Figure S12). GSI-IX-treated mutants, compared with vehicle-treated mutants, showed an improvement in urine albumin-creatinine ratio at D5 PI (Figure 7f). Western blot analysis of urine albumin confirmed absence of albuminuria in GSI-IX-treated mice compared with in vehicle-treated mutants at D5 PI (Figure 7g). We conclude that the observed efficacy of GSI inhibition in amelioration of glomerulosclerosis supports the hypothesis that Notch is activated during disease development.
We also tested the role of starting treatment in established disease by administering GSI-IX to CAGG-CreER TMþ/À ; Wt1 f/f mutant mice on D7 PI. Following 2 doses of GSI-IX, mice were nephrectomized late on D8 PI, and kidney histology was analyzed for 2 mice per treatment. We did not observe a difference in degree of glomerulosclerosis, albuminuria, nor tubulointerstitial expression of vascular smooth muscle actin immunofluorescence (Supplementary Figure S13). We hypothesize that podocyte Notch activation plays a role in early Wt1 glomerulopathy.
Mfng and Rbpsuh short hairpin RNA knockdown does not rescue podocyte EMT gene expression Given the upregulation of MFng and Rbpsuh transcripts at D6 PI, we sought to determine whether knockdown of Mfng or Rbpsuh would influence podocyte EMT gene or apoptosis gene expression. While we did observe evidence of Mfng and Rbpsuh down-regulation with repression of Hes1 and Hes3 transcripts, we did not observe significant repression of podocyte Snail and Slug transcript nor significant upregulation of podocyte-specific transcripts such as Nphs1 nor Nphs2 (Supplementary Figure S14). Reduced primary mutant podocyte viability precluded optimal transfection owing to slow growth and increased apoptosis. We conclude that genetic rescue of Wt1 glomerulopathy by Notch inhibition would be best validated with in vivo strategies.

DISCUSSION
Using an inducible model of Wt1 deletion, we demonstrate a role for Notch activation in the pathogenesis of Wt1 glomerulopathy. The data provided in this study establish the utility of temporal semiquantitative analysis of glomerular scarring as a platform for the study of early pathological events in an inducible model of mature podocyte injury. We show that podocyte apoptosis is evident as early as the fourth day following tamoxifen administration to adult CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice before glomerulosclerosis  Quantitative graphs of a proportion of glomeruli per genotype revealing the extent of intraglomerulosclerosis (score 0: <25% glomerulus sclerosed; score 1: 25%-50% sclerosis; score 2: 50%-75% sclerosis; score 3: >75% sclerosis). Bars represent the mean percentage of each score per genotype. Error bars represent the SEMs. At day 5 postinduction, a higher proportion of GSI-IX-treated CAGG-CreER þ/À ;Wt1 f/f transgenic mice (n ¼ 52 glomeruli, n ¼ 5 mice) exhibited normal glomerular morphology compared with vehicle-treated CAGG-CreER þ/À ;Wt1 f/f (mutants) (n ¼ 52 glomeruli, n ¼ 4 mice): vehicle-treated CAGG-CreER þ/À ;Wt1 f/f versus GSI-IX-treated CAGG-CreER þ/À ;Wt1 f/f mutant transgenic mice: score 0: 18 AE 0.9% versus 59 AE 1.0%, *P < 0.02, Student t-test; score 2: 39 AE 5% versus 8 AE 3%, **P < 0.008, Student t-test. podocyte apoptosis, dedifferentiation, and both diffuse mesangial sclerosis and FSGS phenotypes, 27,28 which are also associated with mutations in WT1. 5,7 Following Wt1 deletion in this model, we show upregulation of several Notch components, including Notch1 and its transcriptional target gene, Nrarp, 35 as well as Notch bHLH transcription factors. The finding of increased Jagged1 and Pofut1 protein in primary mutant podocytes at disease induction suggested a liganddependent mechanism of podocyte Notch activation at disease manifestation. Furthermore, induction of HES1 expression in transgenic Nphs2;rtTA primary podocytes led to increased Snail and Slug expression, suggesting that Notch is activating genes promoting EMT in podocytes. Notch inhibition using gamma secretase inhibitors, on the D4 and D5 following tamoxifen administration, led to a reduction in the severity of glomerulosclerosis and albuminuria. We propose a model in which loss of Wt1 in mature podocytes induces podocyte apoptosis and EMT, which could be mediated via activation of Notch. These findings are consistent with previous reports demonstrating podocyte Notch activation in chemically induced models of glomerulosclerosis and in human biopsies of glomerular disease. 28,[36][37][38] Increased podocyte Hes1, Hes3, Hes5 and Hey1, Hey2, and HeyL have been found in streptozotocin-and puromycin aminonucleosideinduced glomerulosclerosis. 28 Podocyte apoptosis has been shown to play an instigating role in the pathogenesis of FSGS, and Notch activation in mature podocytes induces apoptosis. 28,30 Conditional activation of the intracellular domain of Notch1 in mature podocytes is associated with positive TUNEL staining and upregulation of podocyte Trp53 and Apaf1. Pifithrin-a inhibits apoptosis in podocytes transduced with the Notch1 intracellular domain, thereby suggesting that Notch1 induces podocyte apoptosis via the p53 pathway. 28 Conditional deletion of podocyte Rbpj in mice with diabetic nephropathy is associated with reduced podocyte apoptosis providing further support for a role for Notch in podocyte apoptosis. 28 In the current study, we speculate a role for p53-mediated apoptosis in the pathogenesis of Wt1 glomerulopathy in our model. It would be interesting to explore whether apoptosis is inhibited with conditional inactivation of Notch1 in podocytes of adult CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice. Indeed, conditional deletion of Notch1, but not Notch2, in podocytes of mice with diabetic nephropathy abrogates glomerulosclerosis. 37 Furthermore, reduced expression of primary podocyte Snail1 protein and mRNA has been observed in these mice, 37 suggesting that podocyte Notch1 activation is associated with both apoptosis and epithelial mesenchymal transition. Our study adds to previous studies demonstrating that ectopic Notch activation in mature podocytes is associated with the development of glomerular scarring. Conditional inactivation of Notch1 and its transcriptional targets in both early and late stages of Wt1 glomerulopathy would further define the role and window for Notch activation in disease pathogenesis.
Pharmacological block of Notch signaling has previously defined a narrow window for Notch in proximal nephron identity. [21][22][23] However, a recent study suggests that Notch is required for the formation of all nephron segments and primes nephron progenitors for differentiation. 24 Notch1 and Notch2 are expressed in overlapping patterns in the proximal domain of the S-shaped body where podocyte precursors reside. 23 During terminal podocyte differentiation, Hes/Hey genes are progressively down-regulated. 25,26 Regulation of vertebrate podocyte differentiation has been proposed to involve a multimeric transcriptional network involving Wt1, FoxC1/C2, and Rbpj. 34,39 FOX transcription factor binding motifs have been found in a large proportion of WT1-bound regions supporting coordinated action of these transcription factors in regulating podocyte-specific genes. 17,18 Double knockdown of either wt1a/rbpj or wt1a/foxc1a in zebrafish caused a reduction in podocyte number in contrast to a single knockdown of any of the 3 genes, supporting a genetic interaction between these transcription factors in regulation of podocyte specification. 34 Co-immunoprecipitation studies revealed putative interactions among Wt1, Rbpj, and FoxC2 proteins, and together, combinations of Wt1, FoxC1/2, and NICD can synergistically induce Hey1 expression. 34 In Xenopus, knockdown of xWT1 decreased the early glomus-specific expression of XHRT1, the Hey1 orthologue, but did not perturb its late expression in the pronephros anlagen, suggesting that xWT1 mediates expression of XHRT1 early in glomerulogenesis. 40 HeyL expression in pretubular aggregates is also regulated by Wt1 during murine metanephric development and studies of Wt1 target genes in embryonic mouse kidney tissue revealed that Wt1 can bind to the HeyL promoter. 41 These studies support a role for modulation of Notch transcriptional targets by Wt1 and FoxC1/2 during nephrogenesis.
After the proximal nephron forms, podocytes function normally in the absence of Notch. 25,26 Following Wt1 deletion in mature podocytes, we find upregulation of Notch pathway components, Notch1, Nrarp, Hey1, Hey2, HeyL, Hes1, Hes3, and Hes5. These findings are consistent with previous studies where Wt1 and Foxc1a can inhibit the ability of NICD1 to activate a synthetic Notch reporter driven by Rbpj sites 34 and suggest a model where Wt1 and FoxC1/2 have antagonistic effects on Notch signaling in the mature podocyte. Our study also demonstrated that loss of Wt1 in mature podocytes is associated with a reduction of FoxC2 expression and upregulation of Notch pathway components coincident with onset of glomerulosclerosis and albuminuria. We speculate that FoxC2 could repress expression of Hey2 and other Notch bHLH genes in mature podocytes, based on our observation of increased Hey2 transcript in mutant podocytes at a point when FoxC2 is down-regulated. This would be consistent with a previous report showing that Hey2 is a transcriptional target of FoxC2 in endothelial cells. 42 Restoration of FoxC2 levels in adult CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice may be sufficient to restore podocyte-specific gene expression following injury, perhaps via repression of Notch bHLH gene expression.
An alternative mechanism of podocyte Notch activation in Wt1-mediated injury could also be mediated via activation of Hippo signaling. Kann et al. 18 identified podocyte-specific enrichment for TEAD transcription factor motifs (effectors of Hippo signaling) in the vicinity of WT1 chromatin immunoprecipitation sequencing peaks. Notch ligands are transcriptional targets of Hippo signaling, 43 and it is possible that loss of Wt1 expression in mature podocytes mediates Notch activation via regulation of Hippo components.
Our observation of increased Pofut1 in mutant podocytes supports activation of Notch in CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice. Notch activation relies on O-linked glycosylation of the extracellular domain of Notch receptors and ligands that influence receptor sensitivity to ligand stimulation. 24 Pofut1 is an O-fucosyltransferase 1 enzyme that mediates fucosylation of the extracellular domain of the Notch protein and has recently been shown to play an important role in the regulation of cell surface expression of Notch1. 33 We also found increased Mfng, a b3-N-acetylglucosaminyltransferase that mediates glycosylation of the Notch extracellular domain in mutant podocytes. 31,32 In most contexts, Fringe-mediated glycosylation of Notch1 renders it more sensitive to Dll1-mediated activation. 32 We have found evidence for Jagged1 expression in CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice and in human biopsies of WT1-mutated glomerulosclerosis. It will be of interest to study the consequences of Mfng deletion in CAGG-CreER TMþ/À ;Wt1 f/f mice on disease manifestation.
In the early stages of disease, we observed Jagged1 expression in podocytes and parietal epithelium of mutant glomeruli. This observation is consistent with previous reports for a role for Jagged1 in glomerulosclerosis. 28,36 In contrast, we did not observe any quantitative difference in Delta1 protein expression in primary mutant podocytes compared with controls in the early stages of disease. Both Delta1 and Jagged1 share overlapping expression patterns within the middle segment of the S-shaped body during kidney development. 23 Combined loss of both ligands in Six2-Cre;Dll1 f/f ;Jag1 f/f transgenic mice results in a severe reduction in numbers of proximal tubules and glomeruli, thereby supporting a role for ligand-mediated Notch activation in defining proximal nephron identity. 23 Replacement of one Jag1 allele in the Dll1-null background can rescue some WT1positive podocytes, suggesting that an important role for Jag1 in podocyte fate induction. 23 In the context of podocyte injury, Niranjan et al. 28 reported that TGF-b1 mediated treatment of podocytes resulted in Jagged1 upregulation. Future studies will be directed at examining the effects of conditional deletion of Jagged1 in adult CAGG-CreER TMþ/À ;Wt1 f/f mice on disease manifestation.
In summary, we identify podocyte apoptosis as an early event in the pathogenesis of glomerulosclerosis mediated by loss of Wt1 function in mature podocytes. At disease onset, we find induction of podocyte EMT gene expression and upregulation of several Notch pathway components. Induction of podocyte HES1 expression is associated with increased Snail and Slug expression, suggesting that HES1 regulates podocyte EMT. Early pharmacological blockade of Notch signaling leads to a reduction in the severity of glomerulosclerosis and albuminuria. We speculate that activation of Notch is mediated by repression of FoxC2. Given the recent advances in our understanding of the complex biological roles of WT1, transgenic mice carrying point mutations relevant to human disease will provide invaluable tools to investigate the transcriptional networks and posttranscriptional mechanisms underlying WT1-related glomerulosclerosis.

MATERIALS AND METHODS Mouse strains
Wt1 deletion in adult mice was achieved following generation of bitransgenic mice by crossing CAGG promoter-driven CreER TM mice with homozygous Wt1 conditional mice, where the first exon of Wt1 is flanked by LoxP sites. 29 Site-specific recombination between the LoxP sites of the Wt1 gene results in a ubiquitous Wt1 null allele. Successful Wt1 deletion was demonstrated by recombination PCR and the depletion of Wt1 expression in podocytes (Supplementary Figure S1). Mice were maintained on a mixed genetic background consisting of C57BL/6J and CD1. However, for comparison of phenotypes such as proteinuria, littermates were used.
Cre recombinase was induced by i.p. administration of tamoxifen (1 mg/40 g body weight for 3 days; Sigma Aldrich) to 5-week-old mice. All animal work was carried out under the permission of license. Mice were housed and bred in animal facilities at the Western Labs, UCL Great Ormond Street Institute of Child Health. Mice were killed at D4, D5, D6, D8, and D12 PI of tamoxifen. Bilateral nephrectomies were performed under sterile conditions. Nphs2;rtTA transgenic mice expressing the tetracycline transactivator specifically in podocytes were used for primary podocyte cultures and Hes1 overexpression experiments. 44 Phenotype analysis Urinary albumin and creatinine were determined using the mouse albumin enzyme-linked immunosorbent quantitation set (E90-134; Bethyl Laboratories) and creatinine assay (KGE005; R&D Systems) kits, respectively.
Western blot detection of albuminuria was determined following loading 4 microliters of urine samples (albumin antibody [E90-134; Bethyl Laboratories] 1:1000 dilution). Kidneys were fixed in in 4% formaldehyde in phosphate-buffered saline (PBS) and paraffinembedded kidney sections were stained with PAS.
Primary glomerular harvest and podocyte culture Fresh kidney cortices were dissected into ice-cold Hanks balance salt solution (24020-133; Gibco/Invitrogen, Thermo Fisher Scientific), decapsulated, and minced. After rinsing thoroughly with fresh Hanks balance salt solution, the pieces were pushed through a 100-mm cell strainer (Falcon 352360; BD Biosciences) into a chilled beaker. Hanks balance salt solution was added to a total of 6 ml, and this was divided between 2 chilled 15-ml Falcon tubes. Then 2.2 ml of Percoll (17-0891-02; Amersham) was added to each tube. The samples were spun at 400 rpm for 10 minutes at 4 o C to separate glomeruli across the Percoll gradient. 45 The presence of glomeruli in the top of the gradient was verified microscopically. The top 1 to 2 ml were removed and passed again through a 100-mm cell strainer to catch any large tubule fragments. The filtrate was then passed through a 40-mm cell strainer (Falcon 352340, BD Biosciences) to trap the glomeruli. For podocyte culture, harvested glomeruli were placed onto culture dishes coated with 0.1 mg/ml rat tail collagen type I as previously described. Culture medium from D1 to D3 was RPMI 1640 medium containing 15% fetal bovine serum (FBS) (Atlanta Biologicals), penicillin streptomycin, and amphotericin B. On D3 of culture, unattached glomeruli were washed away and medium was changed to 10% FBS. Podocytes were examined on D6 of culture after harvest.

Immunofluorescence studies
Tissues were fixed and embedded in paraffin or O.C.T. Compound (Sakura Finetek) as previously described. 27 For double and triple immunofluorescence labeling, formalin-fixed sections were deparafinized according to previously published protocols. 27 Microwave antigen retrieval was carried out in citrate buffer in 4 5-minute cycles at medium-high setting (NN-S758WC, 950W maximum output; Panasonic) following by a 20-minute cooling period at room temperature (RT). Blocking was performed in Universal Blocking Reagent (DAKO). O.C.T. Compound (Sakura) embedded cryosections were permeabilized in 0.5% Triton X-100 for 5 minutes, followed by 2 5-minute PBS washes. They were then incubated in blocking buffer (10% goat serum, 1% BSA, 0.1% Triton X-100) for 1 hour at RT. Sections were probed with antibodies diluted in blocking reagent and incubated at 4 C overnight. The following day, sections were washed in PBS and incubated with Alexa Fluor (Molecular Probes, Invitrogen, Thermo Fisher Scientific)-conjugated secondary antibodies for 1 hour at RT. Slides were mounted using VECTASHIELD (Vector Laboratories) and nuclei were stained with DAPI. Confocal imaging was performed using a Zeiss LSM-710 system with an upright DM6000 compound microscope (Leica Microsystems) and images were processed with Zen software suite (Zen Software). Z stacks were acquired at 0.5-mm intervals and converted to single planes by maximum projection with FiJi software.
Apoptosis studies TUNEL assay. Cells were seeded on Matrigel (356234; BD Biosciences, San Jose, CA) coated 8-well chamber slides. The following day, the cells were fixed in 4% paraformaldehyde for 15 minutes and washed in PBS twice for 5 minutes each time at RT. Cells were then permeabilized with 0.2% Triton X-100 in PBS for 5 minutes, then washed in PBS for 5 minutes at RT. Then 100 ml of Equilibration buffer from the DeadEnd Fluorometric TUNEL kit (CAT.G3250; Promega) was added to the cells for 5 to 10 minutes at RT. Following Equilibration buffer, 50 ml rTdT incubation buffer (45 ml Equilibration buffer þ 5 ml Nucleotide mix þ 1 ml rTdT enzyme) was added to the cells; to ensure even distribution of the buffer, the slides were covered with plastic cover slips and incubated at 37 C for 1 hour. Plastic coverslips were removed by immersing the slides 2x SSC for 15 minutes at RT. Slides were then washed 3 times for 5 minutes each with PBS at RT. Sections were mounted with VECTASHIELD plus DAPI mounting medium (H-1200; Vector Laboratories). Slides were imaged on a Zeiss fluorescent microscope.
Cleaved Caspase-3/7 staining. Collagen-coated gas-permeable bottom plates (Corning, Corning, NY) were used to culture glomeruli for all apoptosis assays. Glomeruli were harvested from mice 6 days following tamoxifen induction. Following 6 days of glomerular harvest, unattached glomeruli were washed away with Dulbecco PBS and the medium was replaced with Roswell Park Memorial Institute with 0.2% FBS. After 18 hours, CellEvent Caspase-3/7 Green (Life Technologies, Thermo Fisher Scientific) to measure apoptosis were added to each plate per the manufacturers' instructions. Sections were counterstained with DAPI.
Annexin V staining. Following tissue dissociation, the samples were spun down at 320g for 5 minutes and washed with the Annexin V Binding Buffer (88-8103-74; Thermo Fisher Scientific). The cells were then stained with Annexin V-PE-Cy7 for 30 minutes on ice, in the dark. After the incubation period, the samples were again washed and resuspended in the Annexin V Binding Buffer. Just prior to sample acquisition, each sample was stained with Sytox Blue at a final concentration of 0.3 mmol/l (S34857; Thermo Fisher Scientific). The samples were acquired using a 5-laser BD LSRFortessa X-20 Analyser, equipped with 355 nm (ultraviolet), 405 nm (violet), 488 nm (blue), 561 nm (yellow), and 640 nm (red) lasers. Prior to cell acquisition, the cells were filtered through a 35-mm cell strainer to prevent cellular aggregation during sample acquisition.

Antibodies and reagents
The following antibodies were used: cleaved Notch1 (Val1744, rab- and administered by i.p. injection (100 mg GSI-IX/40 g body weight) on the evening of day D4 PI of tamoxifen to CAGG-CreER TM þ/À ;Wt1 f/f transgenic mice. Vehicle treatment with dimethyl sulfoxide (W387520; Sigma Aldrich) administered by i.p. injection (100 mg GSI-IX/40 g body weight) on the evening of D4 PI of tamoxifen to CAGG-CreER TM þ/À ;Wt1 f/f transgenic mice were used as controls. For late treatments, CAGG-CreER TMþ/À ;Wt1 f/f mutants were treated by i.p.
injection with the gamma secretase inhibitor GSI-IX DAPT (100 mg/40 g mouse body weight) on the evening of D7 PI of tamoxifen and treated again with DAPT the next morning (16 hours later). Urine was collected at least 8 hours following the second DAPT treatment and mice were then sacrificed. Following nephrectomy, light microscopic examination of PAS-stained specimens were scored for severity of glomerulosclerosis. Comparison was made between GSI-DAPT and vehicletreated CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice.
Transfection experiments Primary transgenic murine Nphs2;rtTA podocytes were transfected with TetOHes1 plasmid (6154; Addgene) 46 or control-GFP-only plasmid with Lipofectamine 3000 kit (L3000001; Thermo Fischer Scientific). Cells were transfected at 70% confluency. Following 24 hours, both TetOHes1 and control-plasmid transfected cells were treated with doxycycline (2 and 4 mg/ml) for 72 and 96 hours' duration. RNA extraction was undertaken according to manufacturer's instructions using the Qiagen microRNA extraction kit. Protein was extracted as previously described. 47 shRNA experiments For MFng and Rbpsuh gene knockdown, the Thermo Scientific Open Biosystems pGIPZ MFng (V3LMM_20363), and Rbpsuh (V3LMM_437682) and nonsilencing control vectors (RHS4346) were used. Primary podocytes from D6 PI CAGG-CreER TMþ/À ;Wt1 f/f transgenic mice were cultured in Roswell Park Memorial Institute medium in 10% FBS and 1% insulin, transferrin, and selenium. Next, 1 Â 10 5 cells were seeded per well in 6-well dishes (plastic bottom) in duplicates, 24 hours prior to being transfected with 4 mg of MFng and Rbpsuh short hairpin RNA (shRNA) and 4 mg nonsilencing control vector. One plate was harvested to determine knockdown efficiency by quantitative real-time (qRT)-PCR 48 hours following transfection.

RNA extraction and qRT-PCR
Following primary podocyte derivation, total RNA was isolated using the RNeasy Micro Kit (Qiagen). cDNA was synthesized by reverse transcription using a high-capacity RNA-to-cDNA kit (4387406; Life Technologies, Thermo Fisher Scientific), qRT-PCR was performed with 250 ng cDNA on a Bio-Rad qPCR machine (CFX96 touch realtime PCR detection system; Hercules, CA) using SYBR Green PCR Master Mix (Applied Biosystems) and 0.45 mg of the oligonucleotides (Sigma Aldrich) outlined in Supplementary Tables S1 and S2. For each gene, the reaction was run in duplicate for between 6 and 10 samples, and for each primer pair, a no-template control was included. The data were normalized to Gapdh gene levels within each sample and analyzed using the DD Ct method. 48 Western blotting Protein was isolated using radioimmunoprecipitation assay buffer supplemented with phosphatase and protease inhibitors. Lysis was completed by shearing through a 26-gauge syringe needle. Samples were denatured with 10% b-mercaptoethanol (Sigma Aldrich) in 4X Laemmli sample buffer at 95 C for 10 minutes. Primary podocytes were run on 4% to 15% sodium dodecylsulfate-polyacrylamide gel electrophoresis gradient gels (Bio-Rad Laboratories). Immortalized podocytes were run on 10% sodium dodecylsulfate-polyacrylamide gel electrophoresis gels. All gels were transferred onto polyvinylidene fluoride and blocked in 5% nonfat milk in PBS before being probed with primary antibodies and secondary antibodies. Blots were developed with Pierce ECL Western Blotting Substrate (Thermo Fisher Scientific).