Hepatocyte nuclear factor 4 alpha (Hnf4α), a member of the transcriptional nuclear receptor superfamily expressed in the liver, kidney, intestine, and pancreatic β-cells is required for their homeostasis and metabolism. Hnf4α is considered a powerful dominator, because its protein binds to approximately 40% of the promoter binding region of the genes expressed in hepatocytes and pancreatic islets (Odom et al.,2004). Hnf4α has at least nine isoforms (Hnf4α1 through α9), resulting in alternative splicing and alternative promoter usage (Furuta et al.,1997; Thomas et al.,2001). Hnf4α1 through α6 contain a domain translated starting from exon 1A by the P1 promoter, whereas Hnf4α7 through α9 contain a domain translated from exon 1D by the P2 promoter. The important difference is which type of Hnf4α is translated predominantly in each tissue. In the liver and intestine, both types of Hnf4α are present, but in the pancreas, only type P2 is mainly present. In a previous study, we found that type P1 was predominantly present in mouse kidney and that type P2 appeared only in the embryonic period (Kanazawa et al.,2009).
Expression of the Hnf4α gene is mainly up-regulated by transcription factors Hnf1α (also known as Tcf1/Hnf1/LFB1) and Hnf1β (also known as Tcf2/vHnf/LFB3). Hnf1α binds to the Hnf4α promoter region combining with Hnf6, and Hnf1β binds to the same region with Gata6 (GATA binding protein 6) in an alternative manner (Hatzis and Talianidis,2001). Hnf4α also occupies Hnf1α and Hnf1β promoter regions and up-regulates expression of these genes as positive feedback, but Hnf1α inhibits its own gene transcription by protein–protein interaction with Hnf4α as negative feedback. The Hnf family members are also known as genes related to mature onset diabetes of the young (MODY), because mutations in the Hnf4α, Hnf1α, and Hnf1β genes cause the MODY1, MODY3, and MODY5 forms of non–insulin-dependent diabetes mellitus (NIDDM; Yamagata et al.,1996a,b; Horikawa et al.,1997; Furuta et al.,1997).
Hnf4α is known to play an essential role in mammalian embryogenesis, because disruption of the Hnf4α gene results in embryonic lethality owing to cell death in the embryonic endoderm and failure of gastrulation (Chen et al.,1994). Furthermore, use of the cre-recombinase–mediated tissue-specific gene knockout system revealed that Hnf4α was related to mesenchymal-to-epithelial transition (MET) in the liver, formation of crypts and maturation of goblet cells in the colon, and secretion of insulin in pancreatic β-cells (Parviz et al.,2003; Garrison et al.,2006; Miura et al.,2006). By using in vitro assays, its functions in cellular adhesion, differentiation, and metabolism were identified gradually (Spath and Weiss,1998; Hayhurst et al.,2001; Olsen et al.,2005; Grewal et al.,2005; Chiba et al.,2006). However, the role of Hnf4α in the developing kidney is almost unknown.
Kidney development is started by interaction between the ureteric bud (UB) arising from the Wollfian duct and the metanephric mesenchyme (MM). Around embryonic day (E) 10.5, the UB invades the MM and begins to branch, resulting from Ret proto-oncogene/glial cell derived neurotrophic factor (Ret/Gdnf) signaling regulated by LIM homeobox protein 1 (Lim1), paired box 2 (Pax2), Wilms tumor 1 (Wt1), and sine oculis-related homeobox/eyes absent/dachshund (Six/Eya/Dach; Kreidberg et al.,1993; Sainio et al.,1997; Tang et al.,1998; Moore et al.,1999; Li et al.,2003; Brodbeck et al.,2004; Davies et al.,2004; Kobayashi et al.,2005). Invasion of the UB drives the loose MM to concentrate around the UB tips and forms the condensed mesenchyme (CM) resulting from MET, in which wingless-related MMTV integration site 9b/4 (Wnt9b/Wnt4) is found as an inducer (Stark et al.,1994; Carroll et al., 2003). In parallel with the branching and extension of the UB, the protubular cells in the CM undergo an epithelialization that consecutively leads to renal vesicles, the comma-shaped body (CSB) and S-shaped body (SSB). Then endothelial cells invade the distal end of S-shaped body to form glomerulus by means of vascular endothelial growth factor (Vegf) signaling and the dorsal end connects to the UB with the result that functional nephrons are produced (Eremina et al.,2003). Most of these molecules enumerated here (and more) also act in multistage nephrogenesis and form a complex network mutually (Dressler,2006).
In the present study, to identify the function of Hnf4α in nephrogenesis, we performed gene silencing during development of the kidney by a combination of organ culture and the RNA interference (RNAi) method. We had hypothesized that a decrease of Hnf4α gene expression would result in failure of transformation to the CSB/SSB, because Hnf4α protein was, first, detected in a part of the CSB during nephrogenesis. However, contrary to our hypothesis, Hnf4α gene silencing resulted in the induction of apoptosis in the edge of the CM where immunostaining of Hnf4α was negative. Additionally, we tried to clarify the microenvironmental expression manner of the Hnf gene family during kidney development using laser microdissection (LMD).
Localization of Hnf4α Protein During Kidney Development
First, we investigated the localization of Hnf4α protein in adult and embryonic kidneys by immunohistochemistry. In the adult mouse kidney, Hnf4α was located in the nuclei of epithelial cells constituting proximal tubules which are also stained by AQP1 antibody and walls of the glomerular capsule, but not in the other components (Fig. 1A–D). In the embryonic day (E) 12.5 metanephric kidney constructed by the UB and CM before formation of prenephric structure such as the CSB or SSB, immunoreactions for Hnf4α were observed only in remitted number of the cells constituting the CM. They were detected in cytoplasm, but not in nuclei (Fig. 1E,F). In the E15.5 kidney built with the CSB, SSB, and immature nephric structures, Hnf4α was detected at immature tubules and parts of the CSB and SSB, whereas no immunoreaction was detected in the CM (Fig. 1G–J, region encircled by blue dashed line). The analysis for the localization of Hnf4α protein was carried out using two different antibodies, and the same results were obtained. Parts of these results corresponded to those of our previous study (Kanazawa et al.,2009).
Hnf4α Gene Silencing Induced Apoptosis in CM
To determine the function of Hnf4α during kidney development, we produced a gene silencing system combining organ culture and the RNAi method. At first, to confirm whether organ culture was performed correctly, cultured kidneys were analyzed by submicroscopy, histology, and molecular biology (Fig. 2). The development of the cultured kidney was judged to be normal, since the following evidence was obtained. First, the size of the whole kidney and the number of UB tips increased day by day. Second, several stages of developing nephrons were observed without cell death in E13+3 and +5 kidneys showing canonical localization of Hnf4α.
Next, we checked the effects of several small interfering RNA (siRNAs; #1, #2, and #3) targeting different sequences of Hnf4α mRNA in NIH3T3Hnf4α−/− cells. In these, the highest reduction was observed in Hnf4α siRNA #3-transfected cells, showing a 70% decrease of the mRNA level (three samples showed 80–90% decreases, but one sample showed only a 40% decrease) compared with negative control siRNA-transfected cells (Fig. 3I).
To investigate the effect of gene silencing in the developing kidney, we transfected Hnf4α siRNA into embryonic kidneys. It was found that the first transformation of the CM into the CSB/SSB expressing Hnf4a gene occurs within 1 day when E12.5 kidney was cultured. As we expected the effects of siRNA in this period, E12.5 kidneys were assayed at 24 or 48 hr posttransfection. In the Hnf4α siRNA #3-transfected E12.5 kidney, the Hnf4α mRNA level decreased 30% compared with the negative control (P < 0.05; Fig. 3J). Because the fluorescence was observed in almost all areas of FAM-labeled siRNA-transfected kidney, it was confirmed that siRNA penetrate into the center of kidney by electropolation (Fig. 3K). The siRNA-transfected E12.5 kidneys decreased 8% in size, whereas the number of lobules was almost the same in the Hnf4α siRNA-transfected and negative control kidneys (Fig. 3L,M). The Hnf4α siRNA-transfected E11.5+3 kidney decreased the size in 24 and 48 hr posttransfection (Fig. 4). Notably, cellular organization fell into disorder and apoptotic cells appeared in the edge of the CM (Fig. 5B,D). By anti-ssDNA staining, it was revealed that no or only a few apoptotic cells appeared on the surface of the negative control kidney, but a large number of apoptotic cells appeared in the Hnf4α siRNA-transfected kidney (Fig. 5E,F). These results suggested that Hnf4α was essential for differentiation in the CM and its down-regulation induced cell death in the CM during kidney development. No effect was observed in the medulla, including the CSB and SSB (Fig. 5D).
Hnf4α Begins to Activate in the CM Region
Although Hnf4α had the potential to induce differentiation of the CM as a result of gene silencing in the developing kidney, its protein was not detected in the CM immunohistochemically (Fig. 1G,H). To resolve this paradox, we measured the mRNA of Hnf4α in the CM using laser microdissection-reverse transcriptase-polymerase chain reaction (LMD-RT-PCR). First, to confirm whether LMD was carried out correctly, we checked the Pax2 and Wt1 mRNAs for control in each part of developing nephron (Fig. 6C–E). By immunohistochemistry, Pax2 protein was localized at the UB, CM, and CSB (Fig. 6C), whereas Wt1 protein was found in the CM, a part of the CSB and immature podocytes (Fig. 6D). In LMD-RT-PCR, both mRNAs were detected in the CM and CSB but not in the MM, corresponding with the protein localization (Fig. 6E). These results suggest that LMD is a useful tool for detecting microenvironmental gene expression.
As shown in Figure 6F,G, we checked pan-Hnf4α mRNA in the MM, CM, and CSB by RT-PCR using primer pairs based on the nucleotide sequences including all variants of Hnf4α. Hnf4α gene expression was observed in the CM and CSB but not in the MM. The Hnf4α mRNA level of the CM region was lower than that in the CSB.
To identify which type of Hnf4α was expressed in the CM and CSB, nested PCR was carried out by using primer pairs for types P1 and P2 (Fig. 7A,B). In the first PCR, type P1 was confirmed only in the CSB, but it was detected in the CM and CSB by the second amplification. On the other hand, although type P2 was negative in all regions at the first PCR, it became newly positive in the CM with the second amplification. These results suggested that Hnf4α gene expression began in the CM stage and that the Hnf4αP2 isoform was expressed tentatively only in the CM stage and then disappeared in the CSB/SSB stage.
To investigate how Hnf4α gene expression was regulated during kidney development, we checked the expression of Hnf4α upstream genes Hnf1α, Hnf1β, Hnf6, and Gata6. As a result, Hnf1α mRNA was detected in the CSB/SSB, whereas Hnf1β mRNA was found in the CM and CSB/SSB (Fig. 7C).
In this study, we performed siRNA-based Hnf4α gene silencing and LMD-RT-PCR for Hnfs gene expression during early development of the kidney. Our results suggested the following: (1) Expression of the Hnf4α gene begins at the CM stage in nephrogenesis and is required for cell survival in the CM. (2) In the CM stage, Hnf4αP1/P2 are expressed with Hnf1β, whereas later in the CSB/SSB stage, only Hnf4αP1 is expressed with Hnf1α/Hnf1β.
In the present study, cellular organization fell into disorder and many apoptotic cells were observed at the edge of the CM in the kidney treated with Hnf4α siRNA. These results suggested that Hnf4α was related to the suppression of apoptosis in renal stem/progenitor cells differentiating actively to pretubular aggregates (Herzlinger et al.,1992; Sariola,2002; Nishinakamura and Osafune,2006; Boyle et al.,2008; Couillard and Trudel, 2008). Because in the CM immunoreactions were not obtained by immunohistochemistry (Fig. 1) but expression of mRNA was detected by LMD-RT-PCR (Figs. 6, 7), we suppose that the cells of the CM express Hnf4α protein at low level which could not be detected by immunohistochemistry and it acts for cell survival in the CM. In microarray analysis of Hnf4α-expressing Hek293 cells, it was noted that several apoptosis/proliferation-related genes were regulated by Hnf4α (Lucas et al.,2005).
Although our Hnf4α gene silencing model revealed the new possibility that Hnf4α was associated with cell survival in the CM, it was still unknown what role this gene plays in the CSB/SSB because the CSB/SSB seemed to be normal formation in the Hnf4α siRNA-transfected kidney. It was hypothesized that Hnf4α in the CSB/SSB is related to cell–cell adhesion, homeostasis, and the microvillus formation as shown in some previous studies (Hayhurst et al.,2001; Olsen et al.,2005; Grewal et al.,2005; Chiba et al.,2006). To verify these hypotheses, biological and immunohistochemical analysis targeting the cell junction proteins, the metabolic enzymes and the cytoskeleton proteins will be needed.
We investigated Hnf4α gene expression in several areas of kidney during nephrogenesis, using LMD-RT-PCR, and found expression of Hnf4αP1/P2 and P1 in the CM and CSB, respectively. In a previous study, we showed by RT-PCR that Hnf4αP1 dominantly presented in the adult kidney but that Hnf4αP2 appeared only in the embryonic period (Kanazawa et al.,2009). It was considered from the present results, that Hnf4αP2 temporally detected in the embryonic kidney originated from the CM. Basically, Hnf4α protein contains two transactivation domains (AF1 and AF2) and a negative regulatory domain (Hadzopoulou-Cladaras et al.,1997), whereas Hnf4αP2 lacks the AF1 domain and transactivation by Hnf4αP2 is weaker than that by Hnf4αP1 (Torres-Padilla et al.,2002; Eeckhoute et al.,2003; Ihara et al.,2005). No differences in the amino acid sequence were noted between the DNA binding domains of Hnf4αP1 and Hnf4αP2. In addition, it was reported that Hnf4α protein bound to the promoter region of its own gene (Bailly et al.,2001). Taken together, it might be considered that Hnf4αP2 acts as an accelerator to express Hnf4αP1 stably in the initial period of nephrogenesis.
Additionally, in LMD-RT-PCR for Hnf4α upstream gene expression, we detected Hnf1β in the CM, and later Hnf1α and Hnf1β both were detected in the CSB/SSB. These results corresponded with the data on in situ hybridization for Hnf1α and Hnf1β (Lazzaro et al.,1992). In the adult kidney, Hnf1β is expressed in all segments of the nephron, but Hnf1α and Hnf4α are expressed only in proximal tubules (Coffinier et al.,1999a; Jiang et al.,2003; Kanazawa et al.,2009). In addition, while Hnf1α−/− mice showed a normal level of Hnf4α and up-regulated Hnf1β expression in the liver, Hnf1β−/− mice lack expression of Hnf1α and Hnf4α, resulting in the failure to develop visceral endoderm (Pontoglio et al.,1996; Barbacci et al.,1999; Coffinier et al.,1999b). These studies provided evidence that Hnf1β acts as a more fundamental factor than Hnf1α during embryogenesis. In a previous study, it was proposed that Hnf4α promoter activity depended on either Hnf1α or Hnf1β in an alternative manner in hepatocytes (Hatzis and Talianidis,2001). According to this regulation model, our results suggest the possibility that Hnf4α plays two different roles in each nephrogenic stage: Hnf4αP1 and P2 are induced by Hnf1β in the CM stage, and Hnf4αP1 is induced by Hnf1α in the CSB/SSB stage. Further investigations are necessary to confirm whether Hnf1α/1β protein binds to Hnf4α promoter in the developing kidney.
In conclusion, we suspect that the expression of Hnf4α gene is required for cell survival in the CM during early kidney development. Because other functions of this gene, such as inducing MET, microvillus formation, and maintenance of homeostasis, are expected to characterize the epithelial cells of proximal tubule in several stages of kidney development, further studies are needed.
Embryos were obtained from timed mating of C57BL/6 mice (Japan SLC, Hamamatsu, Japan). The time at E0.5 was considered to be noon of the day on which the vaginal plug was detected. Embryos at E12.5–E17.5 and newborn mice were obtained by cervical dislocation, and analyzed by the following histological and molecular biological procedures. Mice were maintained and killed according to the Guide for the Care and Use of Laboratory Animals, Hokkaido University, Graduate School of Veterinary Medicine.
Whole kidneys isolated from embryos at E11.5 and E12.5 were cultured on Millicell culture plate inserts (Millipore, Billerica, MA) supported at the surface of the medium. The medium prepared constituted of 9.8 mg/ml RITC80-7 in 0.01 M HEPES (pH 7.2) supplemented with 1 mg/ml transferrin (Roche, Mannheim, Germany), 4 μl/ml insulin (Novo Nordisk, Tokyo, Japan), 0.01 ml/ml penicillin streptomycin (Gibco/Invitrogene, Carlsbad, CA), and 5% fetal bovine serum. Organs were cultured at 37°C with 5% CO2. The term “E12.5+X” means that the kidney was isolated at E12.5 and then cultured for X days.
Cell Line and Plasmid
NIH3T3 cells (mouse fibroblasts, EC93061524, Dainippon Sumitomo Pharma, Osaka, Japan) were cultured in Dulbecco's modified Eagle medium with 10% calf serum. The pCImHnf4α plasmid was generated to express mouse Hnf4α gene in mammalian cell, in which mouse Hnf4α CDS sequence was cloned as an EcoRI fragment from TA cloning pGEM-T Easy vector (Promega, Madison, WI) into pCI-neo mammalian expression vector (Promega). Mouse Hnf4αCDS cDNA was obtained by PCR using mouse liver cDNA and a primer pair (forward: 5′-CTTGGTCATGGTCAGTGTGA-3′ reverse: 5′-AGCTTGCTAGATGGCTTCTTG-3′). NIH3T3Hnf4α−/− cell line was generated by pCImHnf4α transfection into NIH3T3 cells with Fugene6 (Roche) and selection with G418. Expression of the Hnf4α gene in NIH3T3Hnf4α−/− cells was confirmed by RT-PCR.
NIH3T3Hnf4α−/− cells were seeded onto six-well plates with 2.5 ml of growth medium. siRNA transfection to the cells was performed according to the siPORT Neo FX (Ambion/Applied Biosystems, Foster City, CA) protocol. SilencerR predesigned siRNAs (AM16704 #1; siRNA ID#62230, #2; ID#158083, #3; ID#62318) and SilencerR negative control siRNA (AM4611) were purchased from Ambion. The siRNA (25 pmol, finally 10 nM) and 5.0 μl of Neo FX reagent were mixed in 200 μl of opti-MEM (Gibco). After 10-min incubation, this mixed reagent was applied to 70% confluent NIH3T3Hnf4α−/− cells in growth medium. After 24-hr culture, transfected cells were analyzed.
Transfection to the embryonic kidney isolated was performed according to the siPORT siRNA Electroporation Buffer (Ambion) protocol. The siRNA (200 pmol) was applied to the embryonic kidney in 75 μl of electroporation buffer with 150 V, 100 μs single electropulse by using ECMR 830 (BTX, San Diego, CA). After a 10-min incubation in electropolation buffer, kidneys were transferred onto normal growth medium and organ-cultured for 24 or 48 hr until analysis. To visualize the efficiency of penetration into embryonic kidney, SilencerR FAM-labeled negative control siRNA (AM4620) was used.
Five-micrometer-thick cryosections obtained from fresh embryos at E13.5–E15.5 were prepared for mounting on glass slides precoated with LMD film (Meiwafosis, Osaka, Japan) and fixed with absolute methanol for 3 min at 4°C. After staining with 1% toluidine blue for 10 sec, LMD for the MM, CM, and CSB was performed by using Ls-Pro300, according to the manufacturer's protocol (Meiwafosis). All procedures were done in the RNase-free conditions.
Total RNAs were extracted from tissues using Trizol reagent (Invitrogen) and digested with DNase I (DAKO) to eliminate genomic DNA. RNAs from LMD samples were purified with an RNAqueous kit (Ambion) and Turbo DNase (Ambion). RT reactions were carried out using 1.2 μg of RNA with random hexamers (Invitrogen) and ReverTraAce (Toyobo, Osaka, Japan). PCR was carried out using 1.0 μg/μl cDNA with Ex Taq (Takara Bio, Tokyo, Japan) and the appropriate primer pairs shown in Table 1. Nested PCR reactions were performed using 1/20 volume of the first PCR products with the primer pairs designed at the inside of the sequence between the first primer pairs. The amplified samples were electrophoresed with 1.5% agarose gel containing ethidium bromide, and finally photographed under an ultraviolet lamp. For measurement of band density, ImageJ (http://rsb.info.nih.gov/ij) was used. Real-time PCR was performed with Brilliant II SYBRR Green QPCR master mix (Stratagene, La Jolla, CA) on Mx 3000P (Stratagene) according to the manufacturer's instructions. The expression levels of genes were normalized by that of Actb as a housekeeping gene.
Details for Hnf4α (1), (2), (3), and (4) are shown in Figure 6. The polymerase chain reactions (PCRs) were carried out with 35 cycles (for tissue sample) or 50 cycles (for laser microdissection sample). The nested PCRs were carried out with 35-35 cycles.
Histology and Immunohistochemistry
The samples for histology were fixed with 4% paraformaldehyde at 4°C overnight. Three- to 5-micrometer-thick paraffin sections were routinely prepared, stained with hematoxylin and eosin, and immunostained with the following procedure. For antigen retrieval, sections were incubated in citrate buffer (pH 6.0) for 10 min at 105°C. After cooling, slides were soaked in methanol containing 1% H2O2 to remove internal peroxidase. After washing, sections were blocked in 10% normal goat or rabbit serum in phosphate buffered saline (PBS) for 30 min at room temperature and incubated with the primary antibody at 4°C overnight. After washing in PBS three times, sections were incubated with secondary antibodies for 1 hr at room temperature, washed again and incubated with Vectastain elite ABC reagent (Vector Laboratories, Burlingame, CA) for 30 min at room temperature. After washing, sections were incubated with 3,3′-diaminobenzidine tetrahydrochloride-H2O2 solution until the stain developed (< 5 min).
For primary antibodies, goat polyclonal anti-Hnf4α (sc-6556, Santa Cruz, Santa Cruz, CA) diluted 1:600, rabbit polyclonal anti-Hnf4α (#3113, Cell Signaling, Danvers, MA) diluted 1:200, rabbit polyclonal anti-Aqp1 (AB3065, Chemicon/Millipore) diluted 1:200, rabbit polyclonal anti-Pax2 (71-6000, Zymed/Invitrogen) diluted 1:100, rabbit polyclonal anti-Wt1 (CA1026, Calbiochem/EMD, Darmstradt, Germany) diluted 1:600, and anti-ssDNA (A4506, DAKO, Glostrup, Denmark) diluted 1:300 were used. For secondary antibodies, biotin-conjugated rabbit anti-goat (Chemicon/Millipore) 1:100 and biotin-conjugated goat anti-rabbit (Funakoshi, Tokyo, Japan) were used.