J Am Soc Nephrol 12: 1188–1196, 2001 Activation of by the Wilms’ Tumor Product Mediates Apoptosis of Renal Cells

KAY-DIETRICH WAGNER,* NICOLE WAGNER,† VIKAS P. SUKHATME,‡ and HOLGER SCHOLZ* *Johannes-Mu¨ller-Institut fu¨r Physiologie and †Klinik fu¨r Innere Medizin I, Humboldt-Universita¨t, Charite´, Berlin, Germany; and ‡Beth Israel Deaconess Medical Center, Renal Division, Boston, Massachusetts.

Abstract. The Wilms’ tumor WT1 is re- sensus site, which bound to recombinant WT1 protein. Tem- quired for development, but little is known about WT1 poral changes of vdr and mRNA levels in developing rat downstream signaling in renal cells. This study reported an kidneys were correlated closely. The active metabolite 1,25- approximately fivefold upregulation of dihydroxyvitamin D3 (1,25-(OH)2D3) strongly inhibited the (VDR) mRNA and protein in human embryonic kidney (HEK) proliferation of wt1-transfected HEK 293 cells. Exposure to

293 cells that stably expressed WT1 at a level comparable to 1,25-(OH)2D3 caused apoptosis of cultured wt1 immunoposi- the developing kidney in vivo. Co-transfection of HEK 293 tive cells from mouse embryonic kidney cortex. These findings cells with expression plasmids encoding four different WT1 suggest that transcriptional activation of the VDR by WT1 can splicing variants stimulated mouse vdr promoter activity more mediate programmed death of renal embryonic cells in re- than fourfold. A 201-bp fragment was identified in the proxi- sponse to 1,25-(OH)2D3. The results provide the first evidence mal vdr promoter that was required for transactivation by for a role of the vitamin D endocrine system in renal cell WT1. This critical sequence contained a predicted WT1 con- growth and differentiation during development.

The Wilms’ tumor gene product, WT1, first was identified suggesting that not all of the putative targets are normally as a tumor suppressor by its mutational inactivation in a controlled by WT1. subset of Wilms’ tumors (reviewed in reference 1). Wilms’ Expression of the WT1 gene is developmentally regulated tumor (nephroblastoma) is a childhood malignancy of the in embryonic kidneys. WT1 mRNA and protein levels in the kidney caused by a failure of the metanephric mesenchyme metanephric mesenchyme increase upon induction by the to differentiate into glomeruli and tubules (2). WT1 gene ureteric bud (15). Levels remain high in the renal vesicles expression is critical for genitourinary development, and and the comma- and S-shaped bodies, where WT1 expres- homozygous disruption of wt1 in mice caused agenesis of sion is restricted to the podocyte layer of the differentiating the kidneys, likely as a result of a loss of metanephric glomeruli (15). Little is known about the physiologic func- blastemal cells (3). tion of WT1 during development. WT1 induced apoptosis in WT1 is related structurally and functionally to the early osteosarcoma cells by suppressing the synthesis of epider- growth response family of transcription factors mal growth factor receptor (16). Because programmed cell and originally was characterized as a transcriptional repres- death occurs at a large scale in embryonic kidneys (17) and sor (reviewed in reference 4). More recent studies indicate because epidermal growth factor rescued the cultured renal that WT1 also can stimulate gene transcription (5–7) and mesenchyme from apoptosis (18), WT1 has been proposed possibly is involved in RNA processing (8). Putative WT1 to act as a proapoptotic signal in the developing kidney as downstream targets include growth- and differentiation-pro- well (16). However, direct experimental evidence that WT1 moting , such as insulin-like growth factor 2 (9), Pax2 can indeed mediate programmed death of renal embryonic (10), amphiregulin (11), and E-cadherin (12), among others cells is still missing. (reviewed in reference 13). However, the endogenous tran- To characterize further the function of the WT1 transcription script levels of most candidate genes with WT1-sensitive factor, we aimed to identify novel WT1 downstream target promoters failed to correlate with WT1 expression (14), genes. Using the approach of educated guessing, we tested the hypothesis that vitamin D receptor (VDR) expression is regu- lated by WT1. VDR belongs to the steroid superfamily of Received July 5, 2000. Accepted October 25, 2000. nuclear receptors and mediates the genomic actions of the Correspondence to Dr. Holger Scholz, Johannes-Mu¨ller-Institut fu¨r Physiolo- active metabolite 1,25-dihydroxyvitamin D3 (1,25-(OH)2D3). gie, Humboldt-Universita¨t, Charite´, Tucholskystrasse 2, 10117 Berlin, Ger- Upon binding to its ligand, VDR heterodimerizes with the many. Phone: ϩ49-30-2802-6610; Fax: ϩ49-30-2802-6662; E-mail: [email protected] (RXR), and this complex physically inter- 1046-6673/1206-1188 acts with specific DNA recognition sites that are present in the Journal of the American Society of Nephrology regulatory regions of putative downstream target genes (re- Copyright © 2001 by the American Society of Nephrology viewed in reference 19). J Am Soc Nephrol 12: 1188–1196, 2001 WT1 Activates Vitamin D Receptor 1189

The following lines of evidence suggested to us that VDR agnostics Ltd., Hamburg, Germany). 1,25-(OH)2D3 concentrations in might be a potential target for WT1. First, the promoters of the the extracts were measured by RIA as described by the manufacturer VDR genes from human and mouse contain several predicted (Immunodiagnostics Ltd.). WT1 consensus-binding sites (20,21). Second, a number of studies have shown that 1,25-(OH)2D3 can act as a signal for RNA Preparation and Northern Blot Hybridization cell differentiation and also may induce apoptosis (reviewed in Isolation of total RNA and Northern blot hybridization were per- reference 22). Finally, VDR mRNA and protein have been formed as described elsewhere (9). Equal amounts of total RNA were detected in embryonic kidneys, thus pointing to a role for the pooled from stable clones grown separately. vitamin D system in renal development (23). Sodium Dodecyl Sulfate–Polyacrylamide Gel Materials and Methods Electrophoresis Cell Culture and Transfections Total cell lysates from subconfluent cultures were prepared by Human embryonic kidney (HEK) 293 cells (ATCC CRL-1573) heating the samples in TBS/1% sodium dodecyl sulfate (SDS) buffer and HeLa cells (ATCC CRL-7923) were obtained from the Amer- to 95°C for 3 min. Twenty ␮g of protein were loaded per lane and ican Type Culture Collection (ATCC). The cells were grown in transferred after separation on a 10% SDS–polyacrylamide gel elec- Dulbecco’s modified Eagle’s medium (DMEM; Life Technologies, trophoresis onto polyvinylidenedifluoride membranes (Amersham Eggenstein, Germany) supplemented with 10% fetal calf serum Pharmacia Biotech, Freiburg, Germany) with the use of a semidry (FCS; Biochrom KG, Berlin, Germany), 100 IU/ml penicillin (Life blotting apparatus (BioRad, Mu¨nchen, Germany). A polyclonal anti- ␮ Technologies), and 100 g/ml streptomycin (Life Technologies). WT1 antibody from rabbit (WT 180; Santa Cruz Biotechnology, ϫ 6 The cells were seeded at a density of 3 10 cells/100-mm dish Heidelberg, Germany) was used at a 1:1000 dilution for immunoblot- on the day before the transfection. Mouse wt1 expression con- ϩ ting. After incubation with a goat anti-rabbit secondary antibody structs with no genomic mutations (wt1 cDNA in pCB6 (24)) (1:5000), the reaction products were detected by enhanced chemolu- were transfected into HEK 293 cells at 70% confluence with the minescence system (Amersham Pharmacia Biotech). A rat monoclo- use of the phosphate precipitation technique (25). Stable nal anti-VDR antibody (1:500 dilution; Affinity Bioreagents, Ham- ␮ clones were selected with 300 g/ml G418 (Life Technologies) burg, Germany) and a biotinylated goat anti-rat IgG secondary and expanded separately. Transient co-transfections were per- antibody (1:1000 dilution; Dianova, Hamburg, Germany) were used ␮ formed with the indicated amounts of plasmids and 1 gofa for immunodetection of VDR protein with the biotin-streptavidin ␤ cytomegalovirus promoter-driven -galactosidase expression con- technique. struct as an internal control for transfection efficiency.

Luciferase and ␤-Galactosidase Assays WT1 Immunostaining of Cultured Mouse Renal The cells were lysed 48 h after the transfection, and luciferase Embryonic Cells activities were measured in a luminometer (Lumat LB 9501, After fixation with 3% paraformaldehyde in phosphate-buffered Bertholdt, Germany) with the use of beetle luciferin as a substrate saline, the cells were permeabilized with 0.1% Triton X-100 and (Promega, Mannheim, Germany). ␤-galactosidase activities were de- endogenous peroxidase activities were blocked for 5 min in a termined spectrophotometrically (Beckman DU 540 spectrophotom- solution of 3% H2O2 in methanol (1:4). After washing in TBS, the eter) with the use of a commercial kit according to the manufacturer’s cells were incubated for 16 h at 4°C with a rabbit polyclonal instructions (Promega). Results shown are averages of five transfec- anti-WT1 antibody (Santa Cruz Biotechnology) diluted 1:150 in tion experiments, each performed in duplicate. P Ͻ 0.05 was consid- TBS with 5% normal goat serum. This incubation was followed by ered significant (ANOVA). a 2-h treatment with biotinylated secondary antibody (goat anti- rabbit, 1:100 in TBS with 1% bovine serum albumin; Vector Determination of Cell Proliferation Laboratories Inc.) and the streptavidin-peroxidase complex (Sig- ma, Deisenhofen, Germany). WT1-positive cells were identified by Cell proliferation was estimated by counting aliquots of the their brown color after visualization with diaminobenzidine and trypsinized cells in a Neubauer chamber and by measuring 5-bromo- hydrogen peroxide (Sigma). 2'-deoxy-uridine incorporation into genomic DNA followed by an enzyme-linked immunosorbent assay detection (Roche Diagnostics, Mannheim, Germany). For this purpose, HEK 293 cells stably trans- Construction of Reporter Plasmids fected either with a wt1 expression construct or with the empty A 1451-bp fragment from the mouse vdr promoter (from 70 to ϩ 4 pCB6 vector were seeded into 96-well plates at a density of 10 cells 1521 bp of the published sequence (20)) was cloned by PCR with the 2 per cm with variable concentrations of 1,25-(OH)2D3 in the medium. use of genomic DNA from mouse liver as a template and the follow- The cultures were incubated for 4 d, and the medium containing ing primers: 5'-TGCCCTAAGGTGTTGGCT-3' (forward primer), 5'- Ϯ 1,25-(OH)2D3 was renewed daily. Data presented are means SEM TGGACACACAGCTCGGCG-3' (reverse primer). The PCR product of 5 experiments performed as duplicates. was ligated into the SmaI/BglII restriction sites of the pGL2basic reporter plasmid and confirmed by dideoxy sequencing of both

Measurement of 1,25-(OH)2D3 Concentrations strands. This construct was designated pVDR1451. 5'-deletion mutants Concentrations of 1,25-(OH)2D3 were measured in tissue culture were generated from pVDR1451 by PCR with the use of the following supernatants at different time points (4, 8, 12, 16, and 24 h) after forward primers: 5'-CGGATCATCACAGGCAGA-3', 5'-TT- incubation. The supernatants were delipidated by treatment with a GAGGGGCAGGGCGGTC-3', 5'-CCAGGTGCTGAGCAGTCT-3'. dextran sulfate/magnesium chloride reagent, followed by immunoex- The PCR products were cloned into the KpnI/BglII sites of pGL2basic

traction with a monoclonal anti–1,25-(OH)2D3 antibody (Immunodi- and designated pVDR306, pVDR105, and pVDR60, respectively. 1190 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 1188–1196, 2001

Electrophoretic Mobility Shift Assays zinc fingers 3 and 4 of the WT1 molecule (28). Renal cells Mouse recombinant WT1 protein was generated in the presence of were chosen because the fetal kidney is a physiologic site of [35S]methionine with the use of the TNT quick coupled transcription- WT1 expression and because low levels of the endogenous translation system (Promega). An aliquot of the translation product WT1 mRNA were transcribed in HEK 293 cells (29), suggest- was analyzed on a 10% SDS-polyacrylamide gel, which showed a ing that the transacting factors required for normal WT1 ex- single band of the size of WT1 protein (approximately 52 kD). DNA pression are operating in this cell line. No WT1 protein was binding reactions were performed for 30 min at room temperature detectable by immunoblotting in HEK 293 cells that were with 20 ng of the unlabeled recombinant WT1 protein (ϩ and Ϫ KTS ϩ transfected stably with the empty pCB6 expression vector isoforms) in 10 ␮lofa1ϫ reaction buffer (10 mM Tris-HCl [pH 7.5], (Figure 1A). WT1 protein levels in pooled samples (n ϭ 5 50 mM NaCl, 1 mM MgCl2, 0.5 mM ethylenediaminetetraacetate, 0.5 mM DTT, l4% glycerol, 0.05 mg/ml poly [dI-dC]). For the supershift different clones) of the wt1 transfectants were approximately experiments, 1 ␮g of a polyclonal rabbit anti-WT1 antibody (C19; fourfold higher (by densitometry) than in embryonic rat kidney Santa Cruz Biotechnology) was added to the reaction mixture. The on gestational day 18 (E18). Taking into account that wt1 is end-labeled 21-bp double-stranded oligonucleotide (5'-TGAACT- expressed only in a small percentage of cells in E18 kidneys TAGTGGGCGTGGTTG-3') contained the predicted WT1 element (15), the average amount of wt1 protein per cell therefore may from the proximal mouse vdr promoter. A 28-bp DNA fragment be approximately the same in our wt1-HEK 293 transfectants including the WT1 consensus binding site from the platelet-derived and in a restricted cell population in the developing kidney in growth factor A-chain promoter served as a specific competitor (5'- vivo. GGGGCGGGGGCGGGGGCGGGGGAGGGG-3') (26). VDR Expression in Transfected HEK 293 Cells Induction and Analysis of Apoptosis in Cultured Mouse We performed Northern blot hybridization using a full- Renal Embryonic Cells length human cDNA probe to compare VDR mRNA levels in Primary cultures of kidney cortical cells were prepared from mouse wt1Ϫ and pCB6ϩ-transfected clones. As shown in Figure 1B, embryos on gestational day 12 (E12) by an enzymatic digestion VDR mRNA content was significantly higher (fivefold by procedure as described previously (27) and grown for1dinDMEM- 10% FCS. The viable cells were split into 96-well plates at a density of 103 cells/cm2 and cultured for 72 h in DMEM-10% FCS supple- mented with variable concentrations (1 nM to 1 ␮M nominally) of

1,25-(OH)2D3. The active vitamin D compound (Leo Pharmaceutical, Ballerup, Denmark), which was obtained as a 4 ϫ 10Ϫ3 M stock solution, was diluted to 10Ϫ6 M in ethanol. The final ethanol content in the tissue culture medium was Յ0.1 vol %. Control experiments were performed by incubation of the cells for 72 h with the appro- priate amounts of ethanol (in DMEM-10% FCS). The following techniques were applied to identify apoptotic cells: (1) the character- istic signs of apoptosis, including membrane blebbing, cytoplasmic shrinkage, and nuclear condensation, were observed at 400ϫ magni- fication after hematoxylin and eosin staining of the cultures (Axioplan 2; Zeiss, Göttingen, Germany); (2) free 3'-OH termini in apoptotic cells were fluorescence-labeled with the terminal deoxynucleotidyl transferase (TdT) technique with the use of a commercial kit accord- ing to the manufacturer’s instructions (Roche Diagnostics); (3) apo- ptotic cells were stained with a FITC-conjugated anti-annexin V antibody (Roche Diagnostics) and visualized under a fluorescence microscope (400ϫ magnification, Axioplan, Zeiss). The annexin V Figure 1. Expression of Wilms’ tumor transcription factor WT1 and immunopositive cells were counted in 10 optical fields from two vitamin D receptor (VDR) in human embryonic kidney (HEK) 293 dishes each. Only those annexin V–positive cells that did not stain cells. (A) Immunoblot demonstrating WT1 protein in HEK 293 cells with propidium iodide were considered apoptotic. The fraction of cells that were transfected stably either with a mouse wt1 expression that exhibited annexin V immunoreactivity was calculated in five construct (WT1) or with the empty pCB6ϩ expression vector independent experiments and taken as an estimate for the proapoptotic (pCB6ϩ), respectively. WT1 protein levels (approximately 52 kD- action of 1,25-(OH)2D3. band) were only slightly higher in pooled samples (n ϭ 5) of wt1- transfected HEK 293 cells than in kidneys from rat embryos on gestational day 18 (E18). VDR mRNA (B) and protein (D) levels in Results ϭ ϩ Constitutive WT1 Expression in HEK 293 Cells pools (n 5) of wt1- and pCB6 -transfected HEK 293 cells. Note the different intensities (fivefold by densitometry) of the specific North- To identify novel putative downstream target genes of the ern blot hybridization signals between the wt1-expressing and the WT1 transcription factor, we established a human embryonic pCB6ϩ-transfected cells (B). Equal amounts of ␤-actin transcripts kidney cell line (HEK 293) with constitutive wt1 expression. were detected in the two cell populations (C). The increase of VDR The transfected cDNA corresponded to the most abundant wt1 mRNA in wt1-expressing HEK 293 cells was mirrored by increased isoform containing a 17 amino acid insertion from alternatively VDR protein levels (D). HeLa cells, which also expressed the approx- spliced exon 5 and three additional amino acids (KTS) between imately 50-kD VDR protein, served as positive controls (D). J Am Soc Nephrol 12: 1188–1196, 2001 WT1 Activates Vitamin D Receptor 1191

densitometry) in wt1-expressing than in pCB6ϩ-transfected cells. The increase of VDR mRNA in wt1-HEK 293 was associated with upregulation of VDR protein as indicated by immunoblot analysis with a rat monoclonal anti-VDR antibody (Figure 1D). ␤-Actin transcript levels were approximately the same in the wt1Ϫ and pCB6ϩ-transfected clones (Figure 1C). Because HEK 293 cells that were transfected stably either with wt1 cDNA (in pCB6ϩ plasmid) or with pCB6ϩ empty vector shared a similar genetic background, the observed differences in VDR likely were due to the different WT1 levels.

Activation of the VDR Gene Promoter by WT1 To test whether transcription from the vdr promoter could be activated by wt1, we transiently co-transfected HEK 293 cells with wt1 expression vectors (four different splicing variants) and a construct containing approximately 1.5 kb of mouse vdr promotor sequence (20) upstream of the firefly

luciferase reporter gene (pVDR1451). Transient co-transfec- tions were performed because the permanent lines were extremely difficult to transfect, yielding transfection effi- ciencies of less than 5%. Normalized luciferase activities that were taken as a measure for the transcriptional activity of the vdr promoter are indicated in Figure 2. Co-transfec- ␮ tion of pVDR1451 (2 g) along with the different wt1 ex- pression constructs (15 ␮g each) stimulated vdr promoter

activity more than fourfold. Luciferase activity of pVDR1451 increased in proportion with the amount (2 to 15 ␮g) of co-transfected wt1 expression plasmid (data not shown). In Figure 2. WT1 activates the vdr promoter in HEK 293 cells. (A) contrast, transcription from the vdr promoter was not Effect of transient co-transfection of expression constructs encoding changed significantly by co-transfection of a pCB6ϩ con- four different wt1 splicing variants (Ϯ17 amino acids/ϮKTS) on the struct that contained the full-length wt1 cDNA in antisense transcriptional activity of an approximately 1.5-kb mouse vdr pro- (WT1 rev.) orientation (Figure 2A). Serial 5'-deletion anal- moter construct (pVDR1451). Shown are the relative luciferase activ- ities normalized to ␤-galactosidase activity in each sample. Values are ysis detected a 201-bp fragment in the proximal vdr pro- means Ϯ SEM of five experiments in duplicate. *, statistical signifi- moter that was required for transcriptional activation by wt1 cances (P Ͻ 0.05) versus transfection of the vdr promoter construct (Figure 2B). This sequence contained a predicted WT1 ϩ (pVDR1451) alone; WT1 rev., a pCB6 construct containing the consensus site (-GNGGGNGNG-) 115 bp upstream of exon full-length wt1 cDNA in reverse orientation. (B) Serial 5'-deletion 1 (20). Electrophoretic mobility shift assays were performed analysis of the mouse vdr promoter. Promoter constructs of various to examine whether recombinant wt1 protein (ϩ and Ϫ KTS lengths were co-transfected transiently into HEK 293 cells either ϩ isoforms) could physically interact with the putative WT1 along with empty pCB6 expression plasmid (□) or with the wt1 ϩ ϩ Ⅵ element. Using the 21-bp oligonucleotide from the mouse cDNA ( KTS isoform) in pCB6 ( ). The relative lengths of the vdr promoter constructs and the locations of the putative WT1 consensus vdr promoter as a probe (5'-TGAACTTAGTGGGCGTG- binding sites are indicated in the schematic drawing. Normalized GTTG-3'), we obtained a single retardation band with both luciferase activities are expressed as relative light units (RLU). Values wt1 isoforms (Figure 3). A similar retardation signal was shown are averages Ϯ SEM of five experiments performed in dupli- seen with nuclear extracts prepared from the permanent cate. *, statistical significance (P Ͻ 0.05) between wt1- and pCB6ϩ-

wt1-expressing clones (data not shown). The shifted band transfected cells. Note that deletion of a 201-bp sequence (pVDR306 could be competed with excess amounts of a 28-bp unla- versus pVDR105) abolished the wt1 responsiveness of the vdr beled oligonucleotide including a WT1 binding site from the promoter. platelet-derived growth factor A-chain promoter (26). The retardation band was supershifted upon incubation of the

reaction mixture with a polyclonal rabbit anti-WT1 antibody 1,25-(OH)2D3 Inhibits the Proliferation of HEK 293 (Figure 3). No band shift was obtained with an oligonucle- Cells otide containing an Sp1 instead of a WT1 consensus binding To test whether renal cell growth might be controlled by a site (data not shown). vitamin D–dependent signaling mechanism, we studied the 1192 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 1188–1196, 2001

Figure 3. Electrophoretic mobility shift assay performed with 20 ng of recombinant wt1 protein (ϩ and Ϫ KTS isoforms) and a 21-bp double-stranded sequence containing the predicted WT1 consensus site from the proximal mouse vdr promoter. A 28-bp unlabeled oligonucleotide including a WT1 binding site in the platelet-derived growth factor A-chain promoter (27) was used at the indicated molar excess in competition experiments. Supershift assays were performed by incubating the binding reactions with a polyclonal anti-WT1 anti- body (pAb). Figure 4. (A) Proliferation of HEK 293 cells that were transfected stably either with a wt1 expression construct (WT1) or with the empty ϩ effect of 1,25-(OH) D on the proliferation of pCB6ϩ- and pCB6 expression plasmid. The cells were seeded into 96-well plates 2 3 at an initial density of 104 cells/cm2 and grown for3dinDulbecco’s wt1-transfected HEK 293 cells. As shown in Figure 4A, the modified Eagle’s medium (DMEM)-10% fetal calf serum (FCS). basal proliferation rates were significantly lower in the wt1- ϩ Cells from two wells were harvested each day and counted in a expressing lines as compared with the pCB6 -transfected Neubauer chamber to estimate cell proliferation. Values are means Ϯ cells. Incubation with the active metabolite 1,25-(OH)2D3 SEM of five experiments. *, statistically significant differences (P Ͻ dose-dependently inhibited 5-bromo-2'-deoxy-uridine incorpo- 0.05) between wt1- and pCB6ϩ-transfected cells. (B) 1,25-dihy- ration into chromosomal DNA. A comparison between the droxyvitamin D3 (1,25-(OH)2D3) inhibits the proliferation of HEK slopes of the growth curves in Figure 4B reveals that the 293 cells. The stable cell lines were grown for 4 d with the indicated wt1-expressing HEK 293 cells were by far more sensitive to concentrations of 1,25-(OH)2D3 nominally adjusted in the tissue cul- ϩ ture medium. Cell proliferation was assayed by measurement of the antiproliferative action of 1,25-(OH)2D3 than the pCB6 transfectants. Notably, the 1,25-(OH) D concentrations mea- 5-bromo-2'-deoxy-uridine incorporation with an enzyme-linked im- 2 3 Ϯ sured in the tissue culture supernatants after 24 h of incubation munosorbent assay technique. Values presented are means SEM of five experiments performed in duplicate. Symbols indicate statistical were approximately 15-fold lower than those nominally ad- significance (P Ͻ 0.05) within (*) and between (ϩ) the wt1- and justed in the medium. This finding is consistent with an esti- pCB6ϩ-transfected lines, respectively. Incubation of stable HEK 293 mated in vitro half-life of 1,25-(OH)2D3 of approximately 6 h cells with vehicle alone (0.1% ethanol in DMEM-10% FCS) had no (30). significant effect on proliferation rates (data not shown).

1,25-(OH)2D3 Induces Apoptosis of Wt1-Expressing Renal Embryonic Cells examined whether programmed cell death also was involved

Because the active vitamin D metabolite can induce ap- in the action of 1,25-(OH)2D3 on embryonic renal cells. As optosis in a variety of cells (reviewed in reference 22), we a more physiologic model than the permanent HEK 293 cell J Am Soc Nephrol 12: 1188–1196, 2001 WT1 Activates Vitamin D Receptor 1193

Figure 5. 1,25-(OH)2D3 induces apoptosis in primary cultures of mouse renal embryonic cells. Isolated cells were grown for3dinDMEM-10% FCS containing either 1 nM 1,25-(OH)2D3 or 0.1% ethanol as a vehicle (five experiments each performed in duplicate). Representative microscopy of cells from mouse embryonic kidney cortex after a 72-h incubation period either with 1,25-(OH)2D3 (B, D, F) or with 0.1% ethanol as a vehicle (A, C, E), respectively. WT1-positive cells (brown color) were detected with the use of a polyclonal anti-WT1 antibody followed by immunoperoxidase labeling with an indirect biotin-streptavidin technique. Note that the fraction of wt1-immunopositive cells decreased dramatically (B versus A) after 1,25-(OH)2D3 treatment. The characteristic signs of apoptosis including DNA condensation became visible after hematoxylin and eosin staining of the 1,25-(OH)2D3–treated cultures (arrows in D). Apoptosis was confirmed by end-labeling of free 3'-OH termini in cultures incubated with the active vitamin D3 metabolite (F). Significantly fewer apoptotic cells were seen in the controls (E). Magnifications: ϫ200 in A and B; ϫ400 in C and D; ϫ100 in E and F.

lines, we used renal cells that were freshly isolated from Whereas vdr mRNA levels remained high throughout adult- mouse embryonic kidney cortex. These primary cultures hood, wt1 transcripts were barely detectable in kidneys from consisted of more than 80% wt1-immunopositive cells, a adult rats (Figure 6). significant fraction of which exhibited the signs of apoptosis (e.g., membrane blebbing, cytoplasmic shrinkage, chroma- tin condensation) after a 72-h incubation with 1 nM 1,25-

(OH)2D3 (Figure 5D versus Figure 5C). Programmed cell death was confirmed by the demonstration of nuclear DNA fragmentation with the use of in situ DNA nick end labeling assay (Figure 5F versus Figure 5E) and by annexin V immunocytochemistry. The effect of the active vitamin D metabolite was dose-dependent, and the fraction of annexin V–positive cells was 45 Ϯ 6%, 62 Ϯ 5%, 73 Ϯ 6%, and 82 Ϯ 4% in the presence of 1 nM, 10 nM, 100 nM, and 1 ␮M

1,25-(OH)2D3, respectively. Of note, the active vitamin D metabolite caused apoptosis preferentially of the wt1-ex- pressing renal embryonic cells, and—by consequence—the fraction of wt1-immunonegative cells increased from less than 20% to more than 90% after a 72-h exposure to Figure 6. WT1 and vdr mRNA levels at different time points in 1,25-(OH) D (Figure 5B versus Figure 5A). developing and adult rat kidneys. Total kidney RNA was prepared 2 3 from 13 animals each, pooled at equal aliquots, and loaded (20 ␮g/lane) on a 1.2% agarose-formaldehyde gel. Note the striking VDR and WT1 Expression in the Developing Kidney correlation of wt1 and vdr expression between days 15 (E ) and 21 Time courses of wt1 and vdr expression were studied by 15 (E21) of embryonic development. The same membrane first was Northern blot analysis in the developing rat kidney in vivo. hybridized with a full-length wt1 cDNA probe and then used for Intrarenal wt1 and vdr transcript levels were closely related rehybridization with a 32P-labeled rat vdr cDNA. Ethidium bromide between days 15 and 21 of embryonic development (Figure 6). staining indicates roughly equal loading of the RNA gel (bottom). 1194 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 1188–1196, 2001

Discussion development. It seems likely, therefore, that VDR is a physio- The purpose of this study was to analyze the downstream logic target gene for WT1 not only in vitro but also in the signaling mechanisms of the Wilms’ tumor transcription factor developing kidney in vivo. WT1 in embryonic renal cells. We report the following novel The functional significance of our findings is supported by observations: the observation that the growth inhibitory effect of 1,25- (OH)2D3 was enhanced dramatically in wt1-expressing HEK 1. VDR mRNA and protein levels were upregulated in HEK 293 cells. Accordingly, less than nanomolar concentrations of 293 cells with constitutive WT1 expression. the active vitamin D metabolite 1,25-(OH)2D3 induced pro- 2. WT1 transactivated the mouse vdr promoter probably grammed death predominantly of the wt1-immunopositive through direct interaction with a predicted consensus bind- cells from mouse embryonic kidney cortex. These observations ing site. suggest that the sensitivity to the proapoptotic action of 1,25- 3. WT1 significantly increased the susceptibility of HEK 293 (OH)2D3 was enhanced in the wt1-expressing cells. Apoptosis cells to the antiproliferative action of 1,25-(OH)2D3. seems to be particularly important in renal development as the 4. The active vitamin D metabolite induced apoptosis in pri- kidneys were the most severely affected organ in Bcl2 knock- mary cultures of wt1-expressing renal embryonic cells. out mice (37), and approximately 3% of all cells within the 5. The temporal expression patterns of wt1 and vdr genes nephrogenic renal cortex are apoptotic at any given time during correlated closely in the developing kidney. These findings development (17). Our results suggest that the vitamin D suggest that transcriptional activation of the VDR by the endocrine system contributes to the high rate of apoptosis in WT1 zinc finger protein can mediate growth inhibition and the developing kidney. Because 1,25-(OH)2D3 may induce apoptosis of renal embryonic cells. programmed cell death through a VDR-dependent signaling mechanism (38), upregulation of VDR by the WT1 transcrip- Recent studies have shown that the active metabolite 1,25- tion factor therefore would increase the susceptibility of renal

(OH)2D3 acting through the intracellular VDR can promote embryonic cells to the proapoptotic action of vitamin D me- cell differentiation in addition to its effects on calcium and tabolite. Similar results have found recently with breast cancer phosphate metabolism (reviewed in references 19 and 22). cells, which exhibited increased rates of apoptosis in response Interestingly, some of the molecular targets of VDR, e.g., to chemotherapeutic agents after pretreatment with 1,25- epidermal growth factor receptor (31) and cyclin-dependent (OH)2D3 (39). kinase inhibitor p21 (32), also are regulated by the WT1 A role of the vitamin D endocrine system in renal develop- transcription factor (6,16), indicating that WT1 and VDR act ment is seemingly in conflict with the normal embryonic through common intracellular signal transduction pathways. development of vdrϪ/Ϫ mice (40,41). Surprisingly, the vdr null Functional synergism between WT1 and VDR is supported by mutant animals were phenotypically normal at birth and did not our observation that WT1 stimulated VDR expression in human exhibit the signs of until weaning. After embryonic kidney cells. Upregulation of VDR by WT1 most weaning, however, vitamin D–dependent type II and likely occurred at the transcriptional level, because co-trans- growth retardation developed in homozygous mice, leading to fection of WT1 expression constructs stimulated vdr promoter death within 15 wk after birth (40,41). The normal phenotype activity probably through direct interaction of WT1 protein of the vdr null mice does not exclude a role of the vitamin D with a predicted consensus sequence in the proximal vdr system in renal cell differentiation but rather suggests a redun- promoter. dancy in vitamin D–dependent signaling throughout embry- The transcriptional effect of WT1 may depend on the type of onic and early postnatal development. Similar findings were expression vector used for transfection (33). Because we trans- made recently with amphiregulin, a member of the epidermal fected cytomegalovirus promoter–based constructs that have growth factor family, that has been identified as a transcrip- been found to repress rather than activate gene transcription tional target of WT1 (11). Thus, amphiregulin stimulated the (33), it might even underestimate the stimulatory potency of branching morphogenesis of metanephric kidney explants (11), wt1 in our experimental setting. For comparison, a threefold but no gross developmental defects were observed in kidneys induction of the human forkhead gene promoter (34) and a from amphiregulin knockout mice (42). fivefold activation of the syndecan-1 promoter (5) by WT1 Taken together, our findings demonstrate that the WT1 gene have been reported. The exact mechanism for the dual regula- product transcriptionally activates VDR expression in human tory functions of WT1 that can either repress or enhance gene embryonic kidney cells. Upregulation of VDR by the WT1 transcription is still unclear but may involve specific protein– transcription factor may mediate apoptosis of renal embryonic protein interactions (35,36). cells in response to 1,25-(OH)2D3. These findings suggest a The metanephric blastema and the podocyte precursors of role for the vitamin D endocrine system in the regulation of the immature glomeruli are major sites of wt1 expression in the renal cell growth and differentiation during development. developing kidney (15). Vdr immunopositive cells also have been detected recently in the renal mesenchyme and the vis- Acknowledgments ceral and parietal glomerular epithelium of embryonic rat kid- The authors appreciate the expert technical assistance of A. Richter ney (23). Our findings demonstrate that the temporal expres- and I. Gra¨tsch. We thank Dr. F. Priem for the measurement of sion patterns of wt1 and vdr correlate closely during renal 1,25-(OH)2D3 and Dr. G. Walz for critical reading of the manuscript. J Am Soc Nephrol 12: 1188–1196, 2001 WT1 Activates Vitamin D Receptor 1195

The wt1 expression constructs were a gift from Dr. D. Haber, and 17. Coles HSR, Burne JF, Raff MC: Large scale normal cell death in ␣ 1 ,25(OH)2D3 was kindly provided by Dr. L. Binderup (Leo Phar- the developing rat kidney and its reduction by epidermal growth maceutical Products). This study was financially supported by a grant factor. Development 118: 777–784, 1993 from the Deutsche Forschungsgemeinschaft (Scho 634/3-1). 18. Koseki C, Herzlinger D, Al-Awqati Q: Apoptosis in metanephric development. J Cell Biol 119: 1327–1333, 1992 19. Haussler MR, Whitfield GK, Haussler CA, Hsieh JC, Thompson References PD, Selznick SH, Dominguez CE, Jurutka PW: The nuclear 1. Hastie ND: The genetics of Wilms’ tumor—A case of disrupted vitamin D receptor: Biological and molecular regulatory proper- development. Annu Rev Genet 28: 523–558, 1994 ties revealed. J Bone Miner Res 13: 325–349, 1998 2. Beckwith JB, Kiviat NB, Bonadio JF: Nephrogenic rests, nephro- 20. Jehan F, DeLuca HF: Cloning and characterization of the mouse blastomatosis, and the pathogenesis of Wilms’ tumor. Pediatr vitamin D receptor promoter. Proc Natl Acad Sci USA 94: Pathol 10: 1–36, 1990 10138–10143, 1997 3. Kreidberg JA, Sariola H, Loring JM, Maeda M, Pelletier J, 21. Miyamoto KI, Kesterson RA, Yamamoto H, Taketani Y, Nishi- waki E, Sawako T, Inoue Y, Morita K, Takeda E, Pike JW: Housman D, Jaenisch R: WT-1 is required for early kidney Structural organization of the human vitamin D receptor chro- development. Cell 74: 679–691, 1993 mosomal gene and its promoter. Mol Endocrinol 11: 1165–1179, 4. Rauscher FJ III: The WT1 Wilms tumor gene product: A devel- 1997 opmentally regulated transcription factor in the kidney that func- 22. Carlberg C, Polly P: Gene regulation by vitamin D . Crit Rev tions as a tumor suppressor. FASEB J 7: 896–903, 1993 3 Eukaryot Gene Expr 8: 19–42, 1998 5. Cook DM, Hinkes MT, Bernfield M, Rauscher FJ III: Transcrip- 23. Johnson JA, Grande JP, Roche PC, Sweeney WE Jr, Avner ED, tional activation of the syndecan-1 promoter by the Wilms’ Kumar R: 1␣,25-dihydroxyvitamin D : Receptor ontogenesis in tumor protein WT1. Oncogene 13: 1789–1799, 1996 3 fetal renal development. Am J Physiol 38: F419–F428, 1995 6. Englert C, Maheswaran S, Garvin AJ, Kreidberg J, Haber DA: 24. Buckler AJ, Pelletier J, Haber DA, Glaser T, Housman DE: The Induction of p21 by the Wilms’ tumor suppressor gene WT1. murine Wilms’ tumor gene (WT1): Isolation, characterization, Cancer Res 57: 1429–1434, 1997 and expression during kidney development. Mol Cell Biol 11: 7. Guan LS, Rauchman M, Wang ZY: Induction of Rb-associated 1707–1712, 1991 protein (RbAp46) by Wilms’ tumor suppressor WT1 mediates 25. Gorman C, Moffat LF, Howard BH: Recombinant genomes growth inhibition. J Biol Chem 273: 27047–27050, 1998 which express chloramphenicol acetyltransferase in mammalian 8. Larsson SH, Charlieu JP, Miyagawa K, Engelkamp D, Rassou- cells. Mol Cell Biol 2: 1044–105, 1982 lzadegan M, Ross A, Cuzin F, van Heyningen V, Hastie ND: 26. Wang ZY, Qiu QQ, Huang J, Gurrieri M, Deuel TF: WT1, the Subnuclear localization of WT1 in splicing or transcription factor Wilms’ tumor suppressor gene product, represses transcription domains is regulated by alternative splicing. Cell 81: 391–401, through an interactive nuclear protein. Oncogene 10: 415–422, 1995 1995 9. Drummond IA, Madden SL, Rohwer-Nutter P, Bell GI, 27. Drummond IA, Mukhopadhyay D, Sukhatme VP: Expression of Sukhatme VP, Rauscher FJ III: Repression of the insulin-like fetal kidney growth factors in a kidney tumor line: Role of FGF2 growth factor gene by the Wilms tumor suppressor WT1. Science in kidney development. Exp Nephrol 6: 522–533, 1998 257: 674–678, 1992 28. Haber DA, Sohn PL, Buckler AJ, Pelletier J, Call KM, Housman 10. Ryan G, Steele-Perkins V, Morris JF, Rauscher FJ III, Dressler DE: Alternative splicing and genomic structure of the Wilms G: Repression of Pax-2 by WT1 during normal kidney develop- tumor gene WT1. Proc Natl Acad Sci USA 88: 9618–9622, 1991 ment. Development 121: 867–875, 1995 29. Scholz H, Bossone SA, Cohen HT, Akella U, Strauss WM, 11. Lee SB, Huang K, Palmer R, Truong VB, Herzlinger D, Kolquist Sukhatme VP: A far upstream cis-element is required for Wilms’ KA, Wong J, Paulding C, Yoon SK, Gerald W, Oliner JD, Haber tumor-1 (WT1) gene expression in renal cell culture. J Biol Chem DA: The Wilms tumor suppressor WT1 encodes a transcriptional 272: 32836–32846, 1997 activator of amphiregulin. Cell 98: 663–673, 1999 30. Satchell DP, Norman AW: Metabolism of the cell differentiating 12. Hosono S, Gross I, English MA, Hajra KM, Fearon ER, Licht agent 1alpha,25(OH)2-16-ene-23-yne vitamin D3 by leukemic JD: E-cadherin is a WT1 target gene. J Biol Chem 275: 10943– cells. J Steroid Biochem Mol Biol 57: 117–124, 1996 10953, 2000 31. Petkovich PM, Wrana JL, Grigoriadis AE, Heersche JN, Sodek J: 13. Reddy JC, Licht JD: The WT1 Wilms’ tumor suppressor gene: 1,25-dihydroxyvitamin D3: increases epidermal growth factor How much do we really know? Biochim Biophys Acta 1287: receptors and transforming growth factor beta-like activity in a 1–28, 1996 bone-derived cell line. J Biol Chem 262: 13424–13428, 1987 14. Tha¨te C, Englert C, Gessler M: Analysis of WT1 target gene 32. Liu M, Lee MM, Cohen M, Bommakanti M, Freedman LP: expression in stably transfected cell lines. Oncogene 17: 1287– Transcriptional activation of the Cdk inhibitor p21 by vitamin D3 1294, 1998 leads to the induced differentiation of the myelomonocytic cell 15. Pritchard-Jones K, Fleming S, Davidson D, Bickmore W, Por- line U937. Genes Dev 10: 142–153, 1996 teous D, Gosden C, Bard J, Buckler A, Pelletier J, Housman D, 33. Reddy JC, Hosono S, Licht JD: The transcriptional effect of van Heyningen V, Hastie ND: The candidate Wilms’ tumour WT1 is modulated by choice of expression vector. J Biol Chem gene is involved in genitourinary development. Nature 346: 270: 29976–29982, 1995 194–197, 1990 34. Ernstsson S, Pierrou S, Hulander M, Cederberg A, Hellqvist M, 16. Englert C, Hou X, Maheswaran S, Bennett P, Ngwu C, Re GG, Carlsson P, Enerba¨ck S: Characterization of the human forkhead Garvin AJ, Rosner MR, Haber DA: WT1 suppresses synthesis of gene FREAC-4. J Biol Chem 271: 21094–21099, 1996 the epidermal growth factor receptor and induces apoptosis. 35. Johnstone RW, See RH, Sells SF, Wang J, Muthukkumar S, EMBO J 14: 4662–4675, 1995 Englert C, Haber DA, Licht JD, Sugrue SP, Roberts T, Rang- 1196 Journal of the American Society of Nephrology J Am Soc Nephrol 12: 1188–1196, 2001

nekar VM, Shi Y: A novel repressor, par-4, modulates transcrip- breast cancer cells to chemotherapy-induced cell death. Cancer tion and growth suppression functions of the Wilms’ tumor Res 60: 2040–2048, 2000 suppressor WT1. Mol Cell Biol 16: 6945–6956, 1996 40. Li YC, Pirro AE, Amling M, Delling G, Baron R, Bronsson R, 36. Maheswaran S, Park S, Bernard A, Morris JF, Rauscher FJ, Hill DE, Demay MB: Targeted ablation of the vitamin D receptor: An Haber DA: Physical and functional interaction between WT1 and animal model of vitamin D-dependent rickets type II with alo- proteins. Proc Natl Acad Sci USA 90: 5100–5104, 1993 pecia. Proc Natl Acad Sci USA 94: 9831–9835, 1997 37. Veis D, Sorenson C, Shutter J, Korsmeyer S: Bcl-2-deficient 41. Yoshizawa T, Handa Y, Uematsu Y, Takeda S, Sekine K, Yoshi- mice demonstrate fulminant lymphoid apoptosis, polycystic kid- hara Y, Kawakami T, Arioka K, Sato H, Uchiyama Y, Masushige S, neys, and hypopigmented hair. Cell 75:229–240, 1993 Fukamizu A, Matsumoto T, Kato S: Mice lacking the vitamin D 38. Davoust N, Wion D, Chevalier G, Garabedian M, Brachet P, receptor exhibit impaired bone formation, uterine hypoplasia and Couez D: Vitamin D receptor stable transfection restores the growth retardation after weaning. Nat Genet 16: 391–396, 1997

susceptibility to 1,25-dihydroxyvitamin D3 cytotoxicity in a rat 42. Luetteke NC, Qiu TH, Fenton SE, Troyer KL, Riedel RF, Chang A, glioma resistant clone. J Neurosci Res 52: 210–219, 1998 Lee DC: Targeted inactivation of the EGF and amphiregulin genes 39. Wang Q, Yang W, Uytingco MS, Christakos S, Wieder R: reveals distinct roles for EGF receptor ligands in mouse mammary

1,25-Dihydroxyvitamin D3: and all-trans-retinoic acid sensitize gland development. Development 126: 2739–2750, 1999