Background. Wilms' tumor suppressor gene (WT1) is essential for normal podocyte function, and transforming growth factor (TGF)-beta contributes to focal segmental glomerulosclerosis (FSGS). We aimed to address whether TGF-beta affects WT1 expression in podocytes.
Methods. A human podocyte cell line treated with TGF-beta1 and kidneys in Alb/TGF-beta1-transgenic mice were analyzed for WT1 expression.
Results. In cultured podocytes, TGF-beta1 reduced WT1 protein expression determined by western blotting beginning at 8 h and decreased WT1 messenger RNA (mRNA) expression measured by quantitative reverse transcription–polymerase chain reaction beginning at 3 h. Knockdown of Smad4 by small hairpin (sh) RNA partially rescued the TGF-beta1-induced reduction of both WT1 protein and mRNA expressions in the cultured podocytes. TGF-beta1 did not alter luciferase activity of the reporter construct for a human WT1 promoter but reduced that for a human WT1 5′ enhancer construct, suggesting that TGF-beta1 may regulate WT1 expression by altering the 5′ enhancer activity. In the transgenic mice, WT1 protein expression in podocytes was decreased at 1 and 3 weeks of age, while glomeruloclerosis developed after 3 weeks.
Conclusion. TGF-beta1 reduces WT1 expression in cultured human podocytes and podocytes in mice before overt glomerulosclerosis begins. The effects are at least partially Smad4 dependent. Our findings identify a novel pathway linking TGF-beta1 to podocyte injury and FSGS. The WT1 reduction may be a useful marker for early podocyte injury.
Focal segmental glomerulosclerosis (FSGS) and collapsing glomerulopathy are important causes of proteinuria and chronic kidney disease. Podocyte injury has been implicated in the initiation and progression of FSGS [1, 2].
Wilms' tumor suppressor gene (WT1) encodes a zinc finger transcription factor that regulates expression of target genes by binding to both DNA and messenger RNA (mRNA). WT1 was identified in 11p13 by positional cloning in an effort to determine a responsible gene for Wilms tumor, a malignant tumor of the kidney [3, 4]. WT1 is expressed in all stages of kidney development, while the expression becomes restricted to podocytes in the mature kidney.
Mutations in various regions of WT1 have distinct phenotypes. Mutations in exon 8 or 9, which encode the zinc finger domain of WT1, can lead to Denys–Drash syndrome; this syndrome manifests an early-onset nephrotic syndrome with diffuse mesangial sclerosis and, less frequently, FSGS . Mutations in the donor splice site in intron 9 of WT1 result in a reduction of KTS+ isoform and are responsible for the Frasier syndrome, characterized by pseudohermaphroditim and FSGS [5, 6]. Deletion of several neighboring genes in 11p13 including WT1 and PAX6 can cause WAGR syndrome (Wilms' tumor, aniridia, genitourinary anomalies and mental retardation), which is also associated with FSGS [5, 7]. In addition, podocyte WT1 expression is decreased in idiopathic FSGS and HIV-associated collapsing glomerulopathy [2, 8]. WT1-null mice fail to develop kidneys and urogenital tract , and reduction of WT1 expression in mice results in crescentic glomerulonephritis and mesangial sclerosis . Substantial evidence suggests an essential role for WT1 in mature podocytes as well as developing kidney.
Transforming growth factor (TGF)-beta expression is increased in podocytes in idiopathic FSGS and HIV-associated collapsing glomerulopathy [8, 11, 12]. TGF-beta induces podocyte injury through various mechanisms, including apoptosis , proliferation , epithelial–mesenchymal transition (EMT)  and increased production of extracellular matrix proteins including collagen .
In the present study, we have addressed the hypothesis that TGF-beta1 reduces WT1 expression in podocytes. We have used a conditionally immortalized human podocyte cell line and Alb/TGF-beta1 transgenic mice, in which elevated circulating levels of active TGF-beta1 are associated with progressive glomerulosclerosis . We found that WT1 expression was significantly decreased in both TGF-beta1-treated cultured podocytes and podocytes in transgenic mice kidneys. We also demonstrated that Smad4 mediated the suppression of WT1 by TGF-beta1 and that TGF-beta1 reduced a recently identified 5′ enhancer activity, which is conserved among vertebrates .
A conditionally immortalized human podocyte cell line AB8/13 (provided by Dr Moin Saleem)  was grown on type-I collagen-coated culture ware (BD Biosciences, San Jose, CA) in RPMI supplemented with 10% fetal bovine serum, 100 U/mL of penicillin G sodium, 100 μg/mL of streptomycin sulfate and insulin–transferrin–selenium G supplement (Invitrogen, Carlsbad, CA). For each experiment, cells were differentiated by culture for 5 days under growth-restricted conditions at 37°C unless specially noted. Recombinant human TGF-beta1 was purchased from PeproTech (Rocky Hill, NJ).
Smad4 knockdown with small hairpin RNA
pLKO.1-puro lentiviral plasmids encoding small hairpin (sh) RNA for human Smad4 or nontarget control (NTC), together with a puromycin-resistant gene, were purchased from Sigma–Aldrich (St Louis, MO). HIV-1-based lentiviral particles were generated in HEK293FT cells (American Type Culture Collection, Manassas, VA) by cotransfection of shRNA or NTC plasmid with pCMV R8.91 and pCMV-VSV-G using FuGENE6 reagent (Roche, Indianapolis, IN). Cultured podocytes were infected with the lentiviral particles in the presence of 8 μg/mL polybrene for 5 h at 33°C. Effective viral titer was confirmed by observation of cellular survival in the presence of 1 μg/mL puromycin in a preliminary experiment. Four shRNA clones were tested for the ability to knockdown Smad4 mRNA, and the best-performing clone was selected.
Ectopic expression of SNAIL
pBabe-puro-hSNAIL.ER.NoTag, a retroviral expression vector for human SNAIL tagged with mouse estrogen receptor (ER-SNAIL) , and pBabe-puro, a control empty vector, was provided from each Dr Eric R. Fearon and Dr Bob Weinberg through Addgene (Cambridge, MA). Retroviral particles were generated in Phoenix-Ampho cells by transfecting each vector using FuGENE6. Cultured podocytes were infected with retroviral particles in the presence of 8 μg/mL polybrene at 33°C overnight. At 72 h after infection, the cells were incubated with 1 μg/mL puromycin, a selection drug for each vector, for 5 days to eliminate uninfected cells. The cells stably expressing ER-SNAIL or control cells were treated with or without 4-hydroxytamoxifen (4-OHT) (Sigma–Aldrich) at indicated concentration for 8 h to induce nuclear translocation of ER-SNAIL .
Luciferase reporter assay
p3TP-lux, a TGF-beta-responsive luciferase reporter construct containing a part of human PAI-1 promoter and three TPA responsive elements was provided from Dr Jeff Wrana and Dr Joan Massague through Addgene . p3TP-lux harbors Smad3/Smad4-binding sites and is widely used to monitor a DNA-binding activity of the Smad complex. Reporter constructs for a human WT1 full-length promoter, hW10/luc (−443 to +182), were gifts from Dr Youqi Han and Dr Mark Minden . A reporter construct for human WT1 5′ enhancer, WTA, was supplied by Dr Frank Bollig and Dr Christoph Englert . Empty control vectors for these constructs (pGL2 basic for p3TP-lux, pGL2 enhancer for hW10/luc and pGL3 promoter for WTA) were purchased from Promega (Madison, WI). The reporter vectors or empty vectors were cotransfected into the cultured podocytes with a control reporter construct, pCMV-beta-gal, using FuGENE6 at 33°C. After 24 h, the culture medium including plasmids and Fugene6 was replaced with fresh medium with or without 5 ng/mL TGF-beta1. After another 24-h incubation, the cells were lysed in luciferase lysis buffer (Promega) and luciferase activity was determined with a luciferase reporter assay system (Promega). The data were normalized to beta-galactosidase activity derived from pCMV-beta-gal measured with Galactosto-Star System (Applied Biosystems, Foster City, CA).
TGF-beta1 transgenic mice
Albumin/TGF-beta1 transgenic mice, in which the albumin promoter drives expression of an active TGF-beta1 molecule, have elevated circulating levels of TGF-beta1 at the earliest time point studied (2 weeks of age) and develop progressive glomerulosclerosis . Male transgenic and wild-type C57B6 × CBA mice were sacrificed at 1, 3 and 5 weeks of age (N = 5 or 6 of each genotype at each time point), and kidneys were collected. The experiments were approved in advance by the NIDDK Animal Care and Use Committee and performed under the NIH Guide for the Care and Use of Laboratory Animals.
Isolation of glomeruli from transgenic or wild-type mice kidneys at 3 weeks of age was carried out as previously reported . Briefly, the mice were anesthetized and perfused through the heart with Dynabeads M-450 tosylactivated (Invitrogen) prepared in Hank’s buffered salt solution with calcium and magnesium. Kidneys were minced with a razor blade, digested with collagenase and DNAse for 30 min and passed through 100-μm cell strainer. Gomeruli were collected with a magnet and washed with phosphate-buffered saline (PBS), and RNA was extracted. Equal amounts of glomerular RNA were obtained from mice in each group (wild-type, 2.05 ± 0.45 μg per a mouse; transgenic, 1.98 ± 0.51 μg).
Cultured podocytes were lysed or mouse kidney tissues were homogenized in RIPA lysis buffer (50 mM Tris–HCl, pH 7.4; 150 mM NaCl; 2.5% deoxycholic acid; 1% NP-40; 1 mM ethylenediaminetetraacetic acid) with protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN). Total protein lysates equal to 20 μg of protein were separated on NuPAGE 4–12% bis–tris gels (Invitrogen) under reducing condition and transferred to nitrocellulose membranes (Invitrogen). Membrane were probed with the following antibodies: mouse monoclonal anti-WT1 antibody, Clone 6F-H2 (Millipore, Billerica, MA) at a dilution of 1:500; mouse monoclonal anti-Smad4 antibody, clone SMD46 (Abcam, Cambridge, MA) at 1:250; mouse monoclonal anti-beta-actin antibody (Sigma–Aldrich) at 1:5000; rabbit polyclonal anti-Snail antibody (Abcam) at 1:500; rabbit polyclonal anti-phospho-Smad2 (Ser245/250/255) (Cell Signaling Technology, Danvers, MA) at 1:1000; rabbit monoclonal anti-phospho-Smad 3 antibody, clone EP823Y (Epitomics, Burlingame, CA) at 1:1000 and goat polyclonal anti-Smad2/3 (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:200 as a primary antibody. After incubation with horseradish peroxidase-conjugated secondary antibodies, signals were detected by chemiluminescence. The intensity of bands was quantified by densitometory.
Cells cultured on type I collagen-coated cover slips (BD Biosciences) or mouse-frozen kidney sections were fixed with 4% paraformaldehyde in PBS for 30 min at room temperature, permeabilized with 0.2 % Triton X-100 for 10 min and incubated with mouse monoclonal anti-WT1 antibody (Millipore) at a dilution of 1:100 or rabbit anti-podocin antibody at 1:200 as a primary antibody (kind gift from Dr Peter Mundel). The first antibody was detected with Alexa 588-conjugated anti-mouse antibody or Alexa 594-conjugated anti-rabbit antibody (Invitrogen).
Quantitative reverse transcription–polymerase chain reaction
Total RNAs were purified from cultured cells or mouse kidney tissues with miRNAeasy mini-kit (Qiagen, Valencia, CA). Sufficient purity of RNA was confirmed by observing 260/280 absorbance ratio of 1.8–2.2, measured by a NanoDrop spectrophotometer (Thermo Scientific, Waltham, MA). Complementary DNA was synthesized with High Capacity RNA-to-cDNA Kit (Applied Biosystems) according to the manufacturers instruction. Quantitative reverse transcription–polymerase chain reaction (qRT–PCR) was carried out in ABI 7900 sequence detection system (Applied Biosystems) with Power SYBR green PCR master mix (Applied Biosystems). Corresponding primers were as follows: human WT1 forward (F) primer ATAACCACACAACGCCCATC, reverse (R) primer TCAGATGCCGACCGTACAAG; human Snail (F) GCACATCCGAAGCCACAC, (R) GGGAGAAGGTCCGAGCACAC; human Slug (F) GCCCGCCGCGATGCTGTAGG, (R) CTTGCCAGCGGGTCTGGCGG; human Zeb1 (F) GCTCTAACCCGCCTTCATCC, (R) TGTCCCAGTGTTTCAGTTTCTCTG; human Zeb2 (F) AGGCGCAAACAAGCCAATC, (R) GCTGGACTCGTCTCCTGGTC; human beta-actin (F) AGCACAGAGCCTCGCCTTTG, (R) AGCGCGGCGATATCATCATC; mouse WT1 (F) CCAGCCTACCATCCGCAAC, (R) GGGTCCTCGTGTTTGAAGG and mouse beta-actin (F) GTCCACACCCGCCACCAG, (R) TGACCCATTCCCACCATCAC.
Data are presented as mean ± SD. Statistical significance was evaluated by one-way analysis of variance, using Bonferroni testing for intergroup comparisons. A P value of <0.05 was taken as significant.
We addressed whether TGF-beta1 affected WT1 expression in a human podocyte cell line AB8/13 (Figure 1). TGF-beta1 (5 ng/mL) reduced WT1 protein expression in the podocytes compared with the control at 8, 24 and 48 h, although the reduction is modest at 8 h (Figure 1a). In a dose–response study, TGF-beta1 at every concentration spanning 0.2–20 ng/mL decreased WT1 protein expression (Figure 1b). In immunolocalization studies, WT1 was localized to the nuclei, and the staining was substantially diminished by exposure to TGF-beta1 (5 ng/mL) for 24 h (Figure 1c). TGF-beta1 (5 ng/mL) exposure for 3, 8 and 24 h reduced WT1 mRNA expression, assessed by qRT–PCR (Figure 1d).
TGF-beta signaling often involves the Smad family, although other pathways also contribute to signal transduction [24, 25]. The TGF-beta type I receptor phosphorylates receptor-regulated (R)-Smads (Smad2 and Smad3), and each phosphorylated R-Smad binds the common mediator Smad (Smad4). The R-Smad/Smad4 complex translocates to the nucleus and regulates gene transcription. We confirmed by western blotting that incubations with TGF-beta1 (5 ng/mL) for 30 and 90 min induced phosphorylation of Smad2 and Smad3 in the cultured podocytes (Figure 2a). To examine whether the Smad signaling mediates WT1 gene regulation by TGF-beta1, we carried out a knockdown assay for Smad4 with shRNA. Comparing podocytes infected with the lentiviral ShRNA vectors targeting Smad4 (Sh-Smad4) with the cells with the vectors encoding NTC, we confirmed an effective knockdown of Smad4 by observing significant reduction of both Smad4 mRNA measured by qRT–PCR (Figure 2b) and Smad4 protein assessed by western blotting (Figure 2c). We performed luciferase reporter assay using p3TP-lux, which is activated by TGF-beta via Smad signaling . Following treatment with TGF-beta1, the luciferase activity was abrogated in podocytes expressing Sh-Smad4 compared with the negative control (NTC), suggesting that Sh-Smad4 effectively inhibits the Smad signal transduction (Figure 2d). We addressed the effect of Sh-Smad4 on WT1 protein and mRNA expressions in the podocytes. Smad4 knockdown substantially rescued the TGF-beta1-induced suppression of WT1 protein after 24 h (Figure 2e). Furthermore, Sh-Smad4 exposure for 8 h partially blocked TGF-beta1 suppression of WT1 mRNA (Figure 2f). Taken together, these data suggest that TGF-beta1 suppression of WT1 mRNA and protein are mediated at least in part by the Smad pathway.
TGF-beta control of EMT typically involves induction of one or more zinc finger transcription factors, such as Snail, Slug, Zeb1 and Zeb2 , and therefore, we examined a possible role for these proteins in the observed effects on WT1 expression. We cultured the podocytes in the presence or absence of TGF-beta1 (5 ng/mL) for 30 min and measured mRNA expression of the four genes by qRT–PCR. Among these, TGF-beta1 increased Snail mRNA by nearly 10-fold, increased Slug mRNA to a minor degree and unaffected Zeb1 and Zeb2 mRNA expression (Figure 3a). At 90 min, only Snail mRNA was upregulated (data not shown). After podocytes were exposed to TGF-beta1 (5 ng/mL), Snail protein expression was upregulated after 90 min and 6 h (Figure 3b). The addition of Smad4 shRNA was able to partially decrease Snail mRNA induction by TGF-beta1 at 8 h (Figure 3c), suggesting that Smad signaling contributed to the Snail upregulation. We next sought to elucidate a functional role for Snail in WT1 expression; we used human Snail tagged with mouse estrogen receptor (ER-Snail), a chimeric molecule which makes Snail nuclear translocation dependent on 4-hydroxytamoxifen (4-OHT) . Expression of ER-Snail was confirmed by western blotting (Figure 3d) and by qRT–PCR (data not shown). Unexpectedly, administration of 4-OHT did not alter WT1 mRNA expression (Figure 3e), suggesting that Snail may not be involved in WT1 regulation.
We analyzed the effect of TGF-beta1 on human WT1 promoter activity using a reporter construct hW10/luc . TGF-beta1 (5 ng/mL) did not alter luciferase activity of the construct (Figure 4a). We tested the effect of TGF-beta1 on a human WT1 enhancer located ∼4.3 kb upstream of the transcription start site using a reporter vector WTA . As shown in Figure 4b, TGF-beta1 reduced luciferase activity of WTA by ∼40%. While this effect was modest, it would be consistent with a direct suppressive effect of TGF-beta1 on WT1 transcription.
To address whether these observations in cultured podocytes might have relevance to in vivo, we studied Alb/TGF-beta1 transgenic mice, which develop severe progressive global glomerulosclerosis . We carried out histological evaluation for glomerulosclerosis using Periodic acid-Schiff-stained kidney sections of the transgenic and wild-type mice at 1, 3 and 5 weeks of age (Figure 5a, 1-week-old mice kidneys are not shown). Glomeruli in 1 and 3-week-old transgenic mice were histologically intact, while glomerulosclerosis was evident in the transgenic mice at 5 weeks of age, with diffuse mesangial matrix expansion and glomerular capillary obliteration. In the previous study, we reported that the transgenic mice manifested glomerulosclerosis at 3 weeks of age, earlier than the present study . This discrepancy represents phenotypic drift, as the selection of breeder mice at 6–8 weeks of age likely selects for a milder renal phenotype. To evaluate whether WT1 expression might be an early marker of podocyte injury, and because podocyte numbers were likely reduced at 5 weeks of age, we used 1- and 3-week-old transgenic and wild-type mice kidneys. Using whole kidney lysates for western blotting, WT1 protein expression, normalized to beta-actin expression, was significantly lower in transgenic mice kidneys at both 1 week of age (75.8 ± 0.10%) and 3 weeks of age (36.1 ± 20.2%) when compared with wild-type mice (Figure 5b and c). We carried out double-immunofluorescent staining for WT1 (green) and podocin (red) (Figure 5d). In wild-type mice kidneys, WT1 localized to the nuclei of podocytes, as identified by cytoplasmic podocin expression. In transgenic mice kidneys, WT1 expression was greatly diminished, while podocin was unaffected. We measured WT1 mRNA expression by qRT–PCR in isolated glomeruli from 3-week-old mice and found no significant difference between transgenic and wild-type mice (data not shown).
In the present study, we have shown that TGF-beta1 reduced WT1 mRNA and protein expression in a human podocyte cell line and in podocytes in mouse kidneys, and this suppressive effect is mediated at least in part via Smad signal transduction. It may be of relevance that TGF-beta1 partially suppressed activity of a WT1 enhancer recently identified ∼4.3 kb upstream of the transcriptional start site, although the effect was modest and the functional importance of this effect remains to be shown.
WT1 is recognized as an important transcriptional regulator of podocyte phenotype [10, 27], but the regulation of WT1 in podocytes is not well understood. TGF-beta1 increases both WT1 protein and mRNA expressions in cultured hepatocytes, which is divergent from our findings and suggests that tissue-specific regulatory networks may be operating . Kashikar et al.  showed that serine threonine receptor-associated protein (STRAP) reduced WT1 and E-cadherin expression in mouse embryonic fibroblasts and that STRAP is involved in TGF-beta signaling as a negative regulator , although the relationship between TGF-beta and WT1 was not defined. To our knowledge, the current study is the first to show that TGF-beta1 can suppress WT1 expression at the protein and mRNA levels.
TGF-beta is a well-recognized cause of podocyte injury and glomerulosclerosis , and reduced expression of WT1 also causes glomerulosclerosis . The data presented here are compatible with a model in which TGF-beta1 acts via multiple pathways to induce podocyte injury, one of which is to reduce WT1 expression and thereby reduce expression of certain critical WT1-dependent genes. We note that the reduction in podocyte WT1 expression was observed in transgenic mice prior to the onset of histologic injury, suggesting that it may be an early injury marker.
One discrepancy in the data presented here is that TGF-beta1 reduced expression of both WT1 mRNA and WT1 protein in a human podocyte cell line, while WT protein but not WT1 mRNA was reduced in Alb/TGF-beta1 transgenic mouse glomeruli. These results suggest that additional factors are operative in vivo.
The transcriptional regulation of WT1 in podocytes is not well understood. In human embryonic kidney (HEK) cells, a yeast artificial chromosome (YAC) containing ∼620 kb including 5′ and 3′ sequence of mouse WT1 gene can induce WT1 expression to a considerable level, and deletion of 15 kb upstream site from transcription start site in the HEK cells with the YAC causes complete loss of the gene expression . To date, retinoic acid , PAX2 , PAX8 , Sp-1 , NF-κB , GATA-1 , GATA-2  and WT1 itself [22, 36] are known to interact with a promoter or an enhancer region of the WT1 gene and modulate WT1 gene expression. In this study, we showed, using a reporter assay, that TGF-beta1 reduced activity of a 5′ enhancer of the WT1 gene, while it did not alter the promoter activity. The 5′ enhancer is located in conserved region at ∼4.3 kb upstream of the human WT1 transcription start site, responds to retinoic acid and contributes to WT1a expression in embryonic zebra fish in the intermediate mesoderm at an early developmental stage and in glomeruli at a later stage . The present study lacks data addressing whether Smads or other regulatory proteins, stimulated by TGF-beta1, interact with the WT1 5′ enhancer sequence. We acknowledge this as an important limitation of our work that will require further effort to clarify.
In conclusion, we have demonstrated that TGF-beta1 reduced WT1 expression in a human podocyte cell line and in podocytes in Alb/TGF-beta1 transgenic mice, prior to the onset of overt glomerular injury and that the effects on WT1 gene expression are at least partially Smad4 dependent. Our findings suggest a novel mechanism by which TGF-beta1 may contribute to podocyte injury and glomerulosclerosis and further suggest that reduced WT1 expression might prove to be a useful marker of early podocyte injury.
We are grateful to Hideko Takahashi and Huiyan Lu for technical assistance, Youqi Han, Mark Minden, Frank Bolig, Christoph Englert, Jeff Wrana, Joan Massague, Eric Fearon and Bob Weinberg for providing plasmids, Moin Saleem for providing a human podocyte cell line, Peter Mundel for providing anti-podocin antibody and Jurgen Schnermann for a critical review.
Funding. This study was supported by Intramural Research Program, NIDDK, NIH.
Conflict of interest statement. None declared.