Abstract
Podocytes are highly specialized cells, but their chromatin status and the precise molecular events leading to their differentiation remain poorly defined. Here we used ChIP-Seq analysis for H3K4me3, H3K4me1 and H3K27me3 to establish the histone methylation map in adult mouse podocytes. Our data demonstrate open chromatin across podocyte specific genes and reveals that genes expressed in the mesoderm lineage become actively repressed upon podocyte differentiation. To better understand the transcriptional control of podocyte differentiation, we studied the role of transcription factor MAFB. ChIP-Seq experiments and functional analysis in conditional knockout mice identified a set of direct MAFB targets including Nphs1, Nphs2, Vegfa and Tcf21. Loss of MafB led to the deposition of extracellular matrix, progressive foot process effacement, and kidney disease. ChIP experiments in knockout animals revealed that during development MAFB is essential for H3K4me3 methylation and the recruitment of WT1 to the promoters of the podocyte specific genes Nphs1 and Nphs2. Taken together our data reveal the crucial function of MAFB by permitting chromatin accessibility at podocyte-specific genes during development and maintaining terminal differentiation in adults.
Introduction
Cellular differentiation relies on the combined action of transcriptional regulators that activate specific genes to give a cell type its characteristic traits. Epigenetic modifications, such as histone methylations modulate chromatin structure and thus strongly impact the ability of transcription factors to bind to DNA and activate downstream target genes. Depending on the type of protein modifications, histone marks can be activating (H3K4me3), enhancing (H3K4me1, H3K27ac) or suppressing (H3K27me3) 1. Epigenetic marks are not static, but are shaped by developmental processes and cellular differentiation. Deposition of the repressive mark H3K27me3 involves the polycomb repressive complex 2 (PRC2) that has been shown to be essential for early embryonic development2. Removal of suppressive histone marks is achieved through specific enzymes that work together with so called pioneering factors that are able to access closed chromatin.
Podocytes are highly specialised cells that are crucial components of the glomerular filter. Differentiation of progenitors involves a number of transcription factors including members of the FOXC gene family 3–7, LMX1B 4,8–10, TCF21 11, the Wilms’ tumor homologue WT1 12–18 and MAFB 19,20. WT1 in particular has been shown to be essential for podocyte differentiation and as many as 50% of podocyte specific genes appear to be direct WT1 targets. Indeed, deletion of WT1 in adult mice leads to a rapid loss of podocyte specific gene expression and, as a consequence renal failure 14,22. While WT1 is essential for the expression of podocyte specific genes such as nephrin (Nphs1) 23, it is not sufficient for their activation. How WT1 can access enhancers and promoters and activate podocyte specific genes is presently poorly understood.
MAFB belongs to the family of large MAF proteins that are characterised by a basic domain involved in DNA binding, a leucine zipper structure that permits hetero and homo-dimerization and an acidic domain involved in transactivation of gene expression 24,25. MAFB has originally been identified as the gene underlying the Kreisler mutation in mice 26, a spontaneous mutant that results in defects of segmentation, brain patterning, pancreas development and macrophage differentiation (for review see 27). Further analyses have revealed an important function for podocyte development and function 20,28,19, and heterozygous point mutations in MAFB have been identified in patients suffering from FSGS and Duane retraction syndrome 21, as well as in cases of multicentric carpotarsal osteolysis29, a syndrome that is associated with a high risk of developing chronic kidney disease. How MAFB contributes to podocyte differentiation and maintenance, as well as its direct targets in this cell type remain poorly defined.
Here we set out to perform an unbiased study of the epigenetic state of the fully mature podocytes and identify the genome wide targets of MAFB. By employing constitutive and inducible podocyte specific knockout animals, we show that MAFB is essential for the activation of the podocyte specific gene program, by opening chromatin and permitting WT1 to bind to promoters of podocyte specific genes such as Nphs1 and Nphs2.
Materials and Methods
Mice
Mice were housed in a conventional animal facility with free access to food and water. Deletion of MafB in fully differentiated podocytes was performed using the inducible Wt1:CreERT2 line 30 in combination with the conditional MafbLoxP/LoxP 31 allele and the GFP reporter Rosa26RmTmG (Ref. 30). Mafb deletion was induced in adult male mice (8-10 weeks) with two consecutive days intraperitoneal injection of Tamoxifen (SIGMA) at 40mg/kg dissolved in corn oil. The developmental analysis was performed on E18.5 embryonic kidneys obtained by mating animals carrying the MafbGFP/+ constitutive knockin allele 32. All experiments were performed according to national guidelines and approved by the French Minister of Higher Education (APAFIS#22471-2019011513265550v8).
Chromatin Immune Precipitation
Adult podocytes were isolated as previously described33 with minor modifications. Briefly, glomeruli were isolated by paramagnetic beads perfusion and podocytes released by further digestion. Single cell suspension was then crosslinked with methanol-free Formaldehyde 1% (Thermofisher) for 10 minutes at RT before quenching, permeabilization and staining with anti-Podocin (SIGMA). Isolated podocytes were aliquoted in sonication buffer (50 mM HEPES pH 7.4, 1 mM EDTA, 150 mM NaCl, 0.4 % SDS, 1% Triton ×100, 1x protease inhibitor) and immediately sonicated using the sonicator Covaris S220 (peak incidence 150W, duty factor 10%, duration 20 minutes). For the epigenetic analysis during kidney development, whole kidneys were isolated from E18.5 embryos, crosslinked, mechanically dissociated in sonication buffer and sonicated with the previous parameters. Chromatin immuno-precipitation was performed using 200000 cells per condition (adult podocytes) or 3µg of chromatin per condition (immature podocytes) in combination with the following kit, reactives and antibodies from Diagenode, if not otherwise specified: True MicroChip-seq kit (ref. C01010132), anti-H3K4me3 (ref. C15410030), anti-H3K4me1 (ref. C15410037), anti-H3K27me3 (ref. C15410195), rabbit IgG (C15410206), anti WT1 (SantaCruz, ref. sc192), anti MAFB, (SIGMA, ref. HPA005653). Libraries were synthetized from the obtained chromatin using the MicroPlex Library Preparation Kit v2 (ref. C05010013) and sequenced.
Bionformatic analysis was performed on two independent biological replicates for each epigenetic mark or transcription factor using Galaxy based approaches (http://usegalaxy.org). Trimmed, groomed reads were aligned to the reference genome mm10 and the peaks were called using the MACS2 algorithm (FDR<0.05). WT1 and MAFB ChIP peaks were further refined using IDR<0.01 and IDR<0.05, respectively, as a threshold. The datasets obtained were further polished with the exclusion of blacklisted loci recently published. For histone marks no IDR analysis was performed but enriched regions consistent between both datasets were considered as peaks. MEME analysis 34 (http://meme-suite.org) was carried out to define de novo consensus binding regions and search for other TF binding sites using in silico databases. GREAT algorithm 35 (http://bejerano.stanford.edu/great/public/html/) was used to analyze the gene ontology terms correlated with the regions bound by the transcription factors or the modified histones. Heat maps and genomic tracks were generated using the Easeq software 36.
H3K4me3 enrichment in embryonic kidneys was assessed by RT-qPCR. Immunoprecipitated chromatin was purified and the enrichment for Nphs1, Nphs2 and Synpo promoters was compared against IgG ChIP experiments and regions distant from the discovered peaks. Primers sequences used for the RT-qPCR are listed in Suppl. Table 5.
Gene expression analysis
RNA-seq analysis was performed on FACS-sorted isolated podocytes. Kidneys were harvested from 8 weeks old mice, 2 days after tamoxifen induction. Podocytes were isolated using previously described protocols33 and taking advantage of the Rosa26R mTmG reporter. In brief, GFP+ (recombined) podocytes were sorted by FACS from dissociated kidneys and isolated cells immediately processed with RNeasy micro kit (Qiagen) to obtain purified total RNA. A total of 5 samples for control (MafbLoxP/+; WT1 Cre ERT2; Rosa26mTmG) and KO (MafbLoxP/LoxP; WT1 Cre ERT2; Rosa26mTmG) were selected. RNAs were quality assessed (Agilent 2100 Bioanalyzer, RIN>9) and RNA libraries were prepared using NEBNext® Ultra TM RNA Library Prep Kit for Illumina® (NEB, USA) following manufacturer’s protocols. The clustering of the index-coded samples was performed by Novogene on a cBot Cluster Generation System using PE Cluster Kit cBot-HS (Illumina). Paired-end clean reads were mapped to the reference genome using HISAT2 software and for samples with biological replicates, differential expression analysis of two conditions/groups was performed using the DESeq2 R package.
Single cell analysis in Fig. 4B was performed using the fetal nephron dataset from 37 in the publicly available database (https://cellxgene.cziscience.com/e/08073b32-d389-41f4-a4fd-616de76915ab.cxg/).
Tissue analysis, histology, EM analysis
Immune-fluorescence and histological analyses were performed on renal tissue, embedded in paraffine after fixation ON at 4°C with PFA 4%, and sectioned at 5um thickness. Primary antibodies used in the study included: anti-WT1 (Dako, ref M3561, clone 6F-H2), anti-MAFB (BETHYL Laboratories, ref IHC-00351), anti-NPHS1 (R&D Systems, ref AF3159), anti-NPHS2 (Sigma, ref P0372), anti-SYNPO (PROGEN Biotechnik, ref 65194).
Transmission electron microscopy (TEM) and scanning electron microscopy (SEM) analyses were performed essentially as described 12,38. For the quantification of foot process effacement and glomerular basal membrane expansion, 3 animals per genotype were analyzed, using 20 images per animal containing an average of 1.33 basal membrane transversal sections. The basal membrane thickness was measured every two micrometers along each section, whereas the diameters of the foot processes was assessed for each event present along each segment.
Urine analysis
Spot urines were collected at the same hour of the day (10 am) and centrifuged to eliminate aggregates and impurity. 5ul of the urines for each sample and time point was loaded in denaturing/reducing buffer (1x Laemmli) in 10% acrylamide/bis acrylamide gel and resolved by electric field. The gel was then stained by Blue Coomassie solution for 3 hours and destained until proper contrast was achieved.
Results
Determining the histone profile in adult mouse podocytes
Epigenetic marks are highly cell type specific and it is therefore essential to start with pure cell populations to map histone modifications. To achieve this goal we isolated glomeruli from perfused adult mice, dissociated them into single cells and FACS sorted them based on the expression of podocin 33 (Fig. 1A). ChIP-Seq experiments (GEO GSE233911) for H3K4me3 (active), H3K4me1 (primed) and H3K27me3 (repressed) were carried out to determine the genome wide chromatin status in podocytes under homeostatic conditions. Biological duplicates showed highly reproducible peaks (Suppl. Figure 1A&B) and all further analyses were therefore carried out with a merged data set. Heat maps across a set of podocyte-specific genes (Suppl. Table 1: GSE119531) 39 showed the expected histone methylation patterns with strong H3K4me3 at the transcription start site (TSS), enrichment of H3K4me1 in the gene body and depletion of H3K27me3 throughout expressed loci (Fig. 1B-C). H3K4me3 enriched regions mostly cluster around the TSS and relate to GO biological processes involved in ‘nucleosome assembly’ and ‘protein folding’ (Fig. 1D). Consistent with the podocyte specific signature, the GO-term ‘Absent podocyte slit diaphragm’ featured in the Mouse Phenotype category (Suppl. Table 1 worksheet 2; FDR q-value= 0.000119119.) Indeed, close inspection of the methylation pattern around podocyte-specific genes confirmed strong H3K4me3 peaks at TSS, enrichment of H3K4me1 throughout out the gene body and depletion of the H3K27me3 repressive mark as exemplified by Wt1, Nphs1, Nphs2 and Tdrd5 (Fig. 1F-G; Suppl. Figure 1A).
By contrast H3K27me3 peaks were broader, and GREAT analysis identified GO biological terms that relate to embryonic development and differentiation processes (Fig. 1E). Examples included the early mesoderm marker T (Brachyury) (Fig. 1H), the paraxial mesoderm markers paraxis (Suppl. Figure 1C), the lateral plate markers Gata4 (Suppl. Figure 1D), the gonadal marker Dmrt1 (Suppl. Figure 1E), and the kidney progenitor marker Six2 (Suppl. Figure 1F). This H3K27me3 methylation pattern is consistent with an active suppression of genes that orchestrate development into alternative developmental linages or genes that are expressed in early stages of kidney progenitors (such as Six2), but that are switched off during terminal differentiation. Interestingly, Hox gene clusters were split with the Hoxc10/11/12 paralogues being actively repressed, whereas downstream Hox genes (Hoxd9-Hoxd1) were devoid of H3K27me3 marks (Suppl. Figure 1G), which is in agreement with their expression in adult podocytes 40. Taken together this data set provides a blueprint of the methylation status of podocytes under homeostatic conditions.
Identification of direct MAFB targets in adult podocytes
To gain further insights into podocyte-specific gene regulation, we performed ChIP-Seq analysis for the transcription factors WT1 and MAFB on isolated glomeruli of adult mice (GEO GSE233910). Consistent with published data 13,41, WT1 bound a large number of genomic regions (17097 peaks; IDR <0.01) mapping not only close to genes, but also to intergenic regions, which is in agreement with the somewhat non-classical action of WT1 (Suppl. Figure 2B). Comparison of our data with previously published findings 13 confirmed the WT1-binding motif (Suppl. Figure 2C) and an overlap of 87% of previously identified peaks (14946 peaks out of 17098 total peaks, data summarized in Suppl. Fig. 2D). As expected, GREAT analysis of the 17097 peaks showed an enrichment of podocyte-related GO terms (Suppl. Figure 2E-F).
Bioinformatic analysis of MAFB ChIP-Seq data from adult glomeruli identified 347 consistent peaks (IDR<0.05). MEME analysis (http://meme-suite.org) confirmed the presence of the MAFB core binding sequence that has been previously identified in silico (jaspar.genereg.net) (Fig. 2A). The analysis revealed a high recurrence of palindromic MAFB sites, a finding that is consistent with binding of this transcriptional regulator as a homodimer 42. 57% of MAFB binding sites were found within 3kb of the transcriptions start sites (Fig. 2B). GREAT analysis of peak-associated genes showed GO term enrichment for cell matrix adhesion (q-value=0.004) and the cellular components of the podosome (q-value=0.01). In the mouse phenotype category, the most enriched GO terms related to abnormal podocyte physiology, including abnormal/expanded mesangial matrix morphology (q-value=0.0125 and q-value=0.0107), podocyte foot process effacement (q-value= 0.0161) and abnormal slit junction morphology (q-value=0.0164) (Fig. 2C and Suppl. Table 2, Worksheet2). Comparison of the list of genes associated with MAFB binding peaks with a list of 462 podocyte specific genes39 showed that MAFB binds to 42 genes (Suppl. Table 2; worksheet2).
Genome wide comparison between WT1 and MAFB IDR peaks revealed 256 chromosomal regions that were bound by both transcription factors (Fig. 2D-E and Suppl. Table 2 worksheet 5). Closer inspection of these showed no precise spacing between the TF binding sites, which may suggest that the two TFs bind to podocyte specific genes in an independent manner. Consistent with this hypothesis, pull-down experiment in co-transfection experiments (HEK293 cells) did not show direct interaction between WT1 and MAFB (data not shown).
MAFB is required to maintain adult podocytes fully differentiated
A recent study has shown that MAFB is not only important during development, but also for podocyte maintenance in adult animals19. To obtain further insights into the molecular mechanisms driven by MAFB, we employed a tamoxifen inducible Wt1-CreERT2 driver30 (Mafbflox/flox; Wt1-CreERT from now on called Mafbcko). Despite very efficient deletion of Mafb as early as 2 days after tamoxifen injection (Fig. 3A&B), loss of kidney function was slow, developing into overt proteinuria only 8 weeks after induction (Fig. 3C). Scanning and transmission electron microscopy showed a diffuse fusion of podocyte foot processes and thickening of the basement membrane (Fig. 3D I to IV). Histological analysis confirmed focal segmental glomerulosclerosis with excessive deposition of collagen matrix (Fig. 3D V and VI). Interestingly, qPCR analysis revealed only a mild reduction of podocyte specific genes even 30 days after deletion (e.g. THSD7a, Synpo, Podxl, Nphs2; Fig. 3F).
To obtain an unbiased view of genes that directly depend on MAFB in adult podocytes, we next performed RNA-Seq analysis (GEO GSE233912) on glomeruli isolated two days after tamoxifen induction, a time point when the expression of MafB was lost (Fig. 3B), Nphs2 (a known target of MAFB) was reduced (Suppl. Figure 3A), but no overt phenotypic abnormalities were present. Pearson and PCA analysis confirmed clustering of wildtype and knockout samples into two different groups (Suppl. Figure 3B&C). Differential expression analysis identified 1109 (693) and 1178 (724) genes to be decreased and increased in Mafbcko animals at an adjusted p-Value of padj <0.05 (padj <0.01), respectively (Fig. 4A and Suppl. Table 3). Enrichr analysis of downregulated genes identified GO-terms including ‘Genes controlling nephrogenesis’ (p=1.6×10-6), nephrotic syndrome (p=2.0×10-5) and ‘VEGFA-VEGFR2 signaling’ (p=2.0×10-5) (Fig. 4B). GO- terms for upregulated genes included ‘Tgf-β regulation of extracellular matrix’ (p=9×10-6), ‘nephrin interactions’ (p=1×10-4) and ‘Rac1/Rho1 motility signaling pathways’ (p=1×10-3) (Fig. 4C), which probably reflects the cellular response of the podocyte to the reduction of key components of the slit diaphragm such as Nphs1 and Nphs2. GSEA analysis with a list of podocyte-specific differentiation markers 40 showed a significant loss of podocyte specific genes (Fig. 4D; enrichment score = 0.65) and a return to a precursor phenotype (Fig. 4E; enrichment score=-0.4).
Comparison of significantly deregulated transcripts with genes that map close to MAFB ChIP-Seq peaks indicates 64 genes to be directly regulated by this transcription factor, 11 of which appeared to be podocyte specific (Fig. 4 F&G). The much higher number of significantly downregulated genes in absence of MAFB (p<0.01, n=693) compared to the relatively low number of direct targets detected by ChIP-Seq may suggest secondary effects already at this early stage after deletion. Interestingly, we found Pod1/Tcf21, a gene with important functions in glomerular development 8–11, to be reduced upon MafB deletion (padj=2.0×10-7; 1.8fold; Suppl. Table 3). Furthermore, GSEA analysis revealed ‘Tcf21 targets’ as a key GO-term enriched in control over knockout samples (enrichment score 0.6; Fig. 4H). Examination of the genomic locus revealed peaks for MAFB and WT1 48.5kb upstream of the Tcf21 transcription start site (Fig. 4I&J), a region that is evolutionary conserved in placental animals and has been identified as a candidate cis-regulatory region (Encode project). We conclude that MAFB achieves its podocyte-specific function by directly regulating a subset of podocyte specific genes including other transcription factors required for terminal differentiation, such as Tcf21.
MAFB is required for chromatin accessibility of podocyte specific promoters
To gain further insights into the molecular mechanisms underlying podocyte differentiation we turned our attention to development. Comparison of the expression pattern on mouse tissues revealed that MAFB expression occurs after WT1, but slightly before NPHS1 and NPHS2 (Suppl. Figure 4A). Analysis of a publicly available single cell database (https://cellxgene.cziscience.com/e/08073b32-d389-41f4-a4fd-616de76915ab.cxg/) 37 confirmed a similar expression profile in human nephron development (Suppl. Figure 4B). Mice homozygous for the MafbGFP knock-in that represents a loss-of-function allele completely lacked NPHS1 and NPHS2 despite the persistence of WT1 expression (Fig. 5A). We therefore hypothesized that MAFB may be involved in chromatin opening thus allowing other transcription factors such as WT1 to access promoter regions of podocyte specific genes. To test this model, we carried out ChIP experiments with antibodies against the active chromatin mark H3K4me3 on E17.5 wildtype and MafbGfp/Gfp kidneys. As expected, Nphs1, Nphs2, Synpo promoter regions were enriched for H3K4me3 modification implying an open chromatin structure in wildtype samples (Fig. 5B). Regions downstream of these genes were not enriched and served as negative controls. By contrast, in MafbGfp/Gfp samples chromatin at the Nphs1 and Nphs2 promoter failed to be precipitated with H3K4me3 antibodies. Of note, the Synpo promoter remained open as demonstrated by maintenance of the enrichment and the residual expression level observed in the IF. To test if changes in the methylation pattern would directly impact binding of other podocyte specific transcription factors, we evaluated the binding of WT1 to these regions. Strikingly, WT1 enrichment was completely lost at Nphs1/Nphs2 promoters, but only slightly reduced at the regulatory region of Synpo (Fig. 5B), a gene that is not bound by MAFB (Suppl. Table 2, Worksheet 1). We conclude that MAFB is crucial to permit chromatin accessibility at its targets, thus allowing transcription factors, such as WT1, to bind and activate key genes of the podocyte.
Discussion
The transcriptional profile of a cell is shaped by the epigenetic landscape that permits or interferes with binding of TFs and thus the activation of cell type specific genes. In this study we determined the map of histone marks in the adult podocyte, which provides important insights into specific gene regulation and will serve as a resource for future analysis. Not surprisingly, strong H3K4me3 peaks were found at transcription start sites of a large number of genes, with podocyte specific genes often showing very broad peaks. H3K4me3 peak breadth has been shown previously to be highly dynamic and directly associated with transcriptional activity 43. In contrast to H3K4me3 modifications, the repressive H3K27me3 mark was predominantly found covering developmental genes. In particular, genes encoding transcription factors that are active at an earlier time point (e.gi. T, Gata4, Six2) were covered with repressive marks. This is consistent with the known role of the PRC2 complex that actively represses genes to lock in developmental decisions 44. Hox genes are peculiar as they show collinearity of gene expression during early stages of development and progressive silencing of members within a cluster is accompanied by histone methylation via the PRC complex 45. In podocytes, H3K27me3 methylation followed this pattern with only the silent portion of each cluster being CpG-marked.
Active suppression by H3K27me3 is important as podocyte specific deletion of Ezh2, a crucial member of the PRC2 complex, sensitizes mice to glomerular disease46. On the molecular level this has been suggested to be due to an increase of Notch signalling and in particular a loss of Jag1 H3K27me3 methylation. In our data we did not find strong H3K27me3 methylation at the Jag1 locus, which may suggest that demethylation at this locus is not a primary cause of dysfunction in Ezh2 mutant podocytes. In the future it will be interesting to investigate in detail epigenetic changes on a genome wide scale in glomerular diseases and link them with the observed changes in gene expression.
We also describe the genome wide binding sites of MAFB in podocytes in vivo. Compared to WT1, which shows very promiscuous binding (17098 peaks), the number of MAFB peaks with an IDR <0.05 was relatively low (349 peaks), suggesting a more selective target site selection. This is consistent with a recent study that showed enrichment for WT1 and FOXC2, but not MAFB in glomeruli specific super-enhancers3. Previous studies in other organ systems have identified a larger number of MAFB targets, which may be due to a less stringent IDR analysis in their studies 47,48. The binding motif detected precisely fits the expected MAFB recognition sequence, with a large proportion of sites showing palindromic sites indicating binding as a homodimer. 256 of the identified MAFB regions (including Nphs1 and Nphs2 promoters) appear to be also bound by WT1, which may suggest that these two transcription factors cooperate in the activation of downstream targets. The absence of a precise spacing between the binding sites and the lack of co-precipitation after co-transfection into HEK293 cells suggests this cooperation not to involve physical interaction. Instead we propose that MAFB is required to allow WT1 factors to access chromatin and thus permit activation of a subset of podocyte specific genes (Fig. 5C).
While this study was in progress, a manuscript reported that deletion of MAFB with the Nphs2-CreERT2 driver in adult podocytes leads to glomerular defects 8 weeks after induction19. Our own data using a different Cre driver (Wt1-CreERT2) that induces >95% deletion of Mafb at protein level three days after induction confirms this relatively slow, but progressive kidney disease. In contrast to the previously published study that reported transcriptomic analysis at a rather late time point (8 weeks after induction), our RNA-Seq analysis was performed only two days after tamoxifen induction, which allowed us to look at primary events. Despite the rapid loss of MAFB protein after inactivation, we observed only a relatively mild reduction of a broad spectrum of genes, including podocyte marker proteins. The intersection of RNA- seq and ChIPseq data allowed us to elaborate a transcriptional cascade. Indeed, Lmx1b and TCF21/Pod1, are directly regulated by MAFB, whereas the key regulator WT1 remains normally expressed in absence of MAFB. MAFB can therefore be considered in an intermediate hierarchic position, downstream of WT1 (by which it is activated), but responsible for high TCF21/Pod1 and Lmx1b expression levels and, as a consequence, their downstream effector genes. Contrary to published in vitro experiments19, we did not find significant enrichment at the Tcf21 start site in our in vivo ChIP-Seq experiments. Instead we identified an evolutionary conserved element 49.5kb upstream of Tcf21 that is bound by both WT1 and MAFB and has been identified as a potential cis-regulatory element by the ENCODE project. Given the important function of TCF21 in podocyte differentiation, it will be interesting to test this region for mutations in patients with unexplained cases of glomerulopathies.
In contrast to the relatively mild kidney phenotype after tamoxifen induced deletion of Mafb in adults, constitutive deletion abolishes terminal podocyte differentiation. Our data point to a crucial role for MAFB as a pioneer transcription factor that allows chromatin to open and other factors to bind 49. Expression of MAFB has been shown in the epidermis to be both necessary and sufficient to exit cell cycle and promote progenitor differentiation 50. Once chromatin has been opened, MAFB may no longer be essential to permit continuous transcription of podocyte specific genes. This model is supported by our ChIP studies in developmental kidneys which demonstrate that loss of MAFB leads to a lack of open chromatin and strongly reduced binding of WT1 at the Nphs1 and Nphs2 loci.
Taken together our study describes the epigenetic landscape of mouse podocytes, identified genome wide binding sites for MAFB and discovered an important role of MAFB during differentiation in opening chromatin and other TFs expression and binding to promoters of podocyte specific genes.
Acknowledgements
We would like to thank the entire staff of the iBV animal facility for their dedication. We are grateful to William Pu for providing us with the Wt1CreERT2 strain. This study was supported by institutional grants from TU Dresden, the Institut National de la Santé et de la Recherche Médicale, Centre National de la Recherche Scientifique, University Cote d’Azur and Aix-Marseille University and grants to A.S. from the Fondation pour la Recherche Médicale (ING20160435020), ANR (ANR-12-BSV 1-0033-02; ANR-11-LABX-0028-01), the European Commission (EURenOmics Grant agreement 305608), Conseil Régional Sud Est PACA”, the “Conseil Départemental 06”, to M.H. Sieweke from Fondation pour la Recherche Médicale (DEQ. 20110421320), the ‘Agence Nationale de la Recherche’ (ANR-11-BSV3-026-01, ANR-17-CE15-0007-01 and ANR-18-CE12-0019-03), INCA (13-10/405/AB-LC-HS), Fondation ARC pour la Recherche sur le Cancer (PGA1 RF20170205515), an INSERM-Helmholtz cooperation and the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement number 695093 MacAge). M.H. Sieweke is an Alexander von Humboldt Professor at TU Dresden. The authors acknowledge the University’s CCMA, Electron Microscopy facility (Centre Commun de Microscopie Appliquée, Université Côte d’Azur) and MICA Imaging platform Côte d’Azur supported by Université Côte d’Azur, the “Conseil Régional Sud Est PACA”, the “Conseil Départemental 06 “ and Gis Ibisa, and Christelle Boscagli for her technical help.
Suppl. Table 1: Histone methylation peaks detected by ChIP-Seq analysis
Suppl. Table 2: ChIP-Seq peaks for MAFB and WT1
Suppl. Table 3: RNA-Seq analysis
Suppl. Table 4: List of Primers
References
- 1.: Roles and regulation of histone methylation in animal development [Internet]Nat. Rev. Mol. Cell Biol 20:625–641
- 2.: The Polycomb -Group Gene Ezh2 Is Required for Early Mouse DevelopmentMol. Cell. Biol. [Internet 21:4330–4336
- 3.: Super-Enhancer-Associated Transcription Factors Maintain Transcriptional Regulation in Mature PodocytesJ. Am. Soc. Nephrol. [Internet] ASN 2020081177
- 4.: Lmx1b and FoxC combinatorially regulate podocin expression in podocytesJ Am Soc Nephrol [Internet 25:2764–2777
- 5.: Wt1a, Foxc1a, and the Notch mediator Rbpj physically interact and regulate the formation of podocytes in zebrafishDev Biol 358:318–330
- 6.Notch signaling, wt1 and foxc2 are key regulators of the podocyte gene regulatory network in XenopusDevelopment 137:1863–1873
- 7.: Foxc1 and Foxc2 are necessary to maintain glomerular podocytesExp. Cell Res 352:265–272
- 8.: Regulation of glomerular basement membrane collagen expression by LMX1B contributes to renal disease in nail patella syndromeNat Genet [Internet 27:205–208
- 9.: The LIM-homeodomain transcription factor Lmx1b plays a crucial role in podocytesJ. Clin. Invest 109:1073–1082
- 10.: Transcriptional induction of slit diaphragm genes by Lmx1b is required in podocyte differentiationJ. Clin. Invest. [Internet 109:1065–1072
- 11.: Loss of the podocyte-expressed transcription factor Tcf21/Pod1 Results in podocyte differentiation defects and FSGSJ. Am. Soc. Nephrol 25:2459–2470
- 12.: Alternatively spliced isoforms of WT1 control podocyte-specific gene expressionKidney Int
- 13.: Genome-Wide Analysis of Wilms’ Tumor 1-Controlled Gene Expression in Podocytes Reveals Key Regulatory MechanismsJ Am Soc Nephrol [Internet
- 14.: Integration of Cistromic and Transcriptomic Analyses Identifies Nphs2, Mafb, and Magi2 as Wilms’ Tumor 1 Target Genes in Podocyte Differentiation and MaintenanceJ Am Soc Nephrol
- 15.: Donor splice-site mutations in WT1 are responsible for Frasier syndromeNat Genet 17:467–470
- 16.: Frasier syndrome is caused by defective alternative splicing of WT1 leading to an altered ratio of WT1 +/-KTS splice isoformsHum Mol Genet 7:709–714
- 17.: Germline WT1 mutations in Wilms’ tumor patients: preliminary resultsMed.Pediatr.Oncol 27:404–407
- 18.Germline mutations in the Wilms’ tumor suppressor gene are associated with abnormal urogenital development in Denys-Drash syndromeCell 67:437–447
- 19.: Transcription factor MafB in podocytes protects against the development of focal segmental glomerulosclerosisKidney Int. [Internet 98:391–403
- 20.: MafB is essential for renal development and F4/80 expression in macrophagesMol Cell Biol [Internet 26:5715–5727
- 21.: A mutation in transcription factor MAFB causes Focal Segmental Glomerulosclerosis with Duane Retraction SyndromeKidney Int. [Internet 94:396–407
- 22.Acute multiple organ failure in adult mice deleted for the developmental regulator wt1
- 23.Schedl A: The major podocyte protein nephrin is transcriptionally activated by the Wilms’ tumor suppressor WT1J. Am. Soc. Nephrol 15:3044–3051
- 24.The Maf transcription factors: Regulators of differentiation [Internet]Trends Biochem. Sci 22:437–441
- 25.The world according to MafNucleic Acids Res. [Internet 25:2953–2959
- 26.The mouse segmentation gene kr encodes a novel basic domain-leucine zipper transcription factorCell 79:1025–1034
- 27.Functional analysis of large MAF transcription factors and elucidation of their relationships with human diseasesExp. Anim 70:264–271
- 28.The mouse Kreisler (Krml1/MafB) segmentation gene is required for differentiation of glomerular visceral epithelial cellsDev. Biol. [Internet 249:16–29
- 29.Multicentric carpotarsal osteolysis is caused by mutations clustering in the amino-terminal transcriptional activation domain of MAFBAm. J. Hum. Genet. [Internet 90:494–501
- 30.Epicardial progenitors contribute to the cardiomyocyte lineage in the developing heartNature [Internet 454:109–113
- 31.Microglia development follows a stepwise program to regulate brain homeostasisScience [Internet 353
- 32.MafB deficiency causes defective respiratory rhythmogenesis and fatal central apnea at birthNat Neurosci [Internet 6:1091–1100
- 33.Molecular fingerprinting of the podocyte reveals novel gene and protein regulatory networksKidney Int. [Internet 83:1052–1064
- 34.The MEME Suite:W39–W49
- 35.Bejerano G: GREAT improves functional interpretation of cis-regulatory regionsNat Biotechnol [Internet 28:495–501
- 36.An interactive environment for agile analysis and visualization of ChIP-sequencing dataNat. Struct. Mol. Biol. [Internet 23:349–357
- 37.Clatworthy MR: Spatiotemporal immune zonation of the human kidneyScience [Internet 365:1461–1466
- 38.Collombat P: The inactivation of Arx in pancreatic alpha-cells triggers their neogenesis and conversion into functional beta-like cellsPLoS Genet [Internet 9
- 39.Humphreys BD: Cell profiling of mouse acute kidney injury reveals conserved cellular responses to injuryProc. Natl. Acad. Sci. U. S. A. [Internet 117:15874–15883
- 40.Defining the molecular character of the developing and adult kidney podocytePLoS One [Internet 6
- 41.WT1 controls antagonistic FGF and BMP-pSMAD pathways in early renal progenitorsNat. Commun 5
- 42.Design of a bZip transcription factor with homo/heterodimer-induced DNA-binding preferenceStructure [Internet 22:466–477
- 43.Distinct features of H3K4me3 and H3K27me3 chromatin domains in pre-implantation embryosNat. 2016 5377621 537:558–562
- 44.The Polycomb complex PRC2 and its mark in lifeNature [Internet 469:343–349
- 45.Epigenetic Temporal Control of Mouse Hox Genes in VivoScience (80-.) 324:1320–1323
- 46.Shifts in podocyte histone H3K27me3 regulate mouse and human glomerular diseaseJ. Clin. Invest. [Internet 128:483–499
- 47.Transcriptional regulation of endothelial cell behavior during sprouting angiogenesisNat. Commun. 2017 81 8:1–14
- 48.Pancreatic islet enhancer clusters enriched in type 2 diabetes risk–associated variantsNat. Genet. [Internet 46
- 49.Pioneer Transcription Factors Initiating Gene Network ChangesAnnu. Rev. Genet 54
- 50.A LncRNA-MAF:MAFB Transcription Factor Network Regulates Epidermal DifferentiationDev. Cell 32:693–706
Article and author information
Author information
Version history
- Preprint posted:
- Sent for peer review:
- Reviewed Preprint version 1:
Copyright
© 2023, Massa et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.
Metrics
- views
- 268
- downloads
- 9
- citations
- 0
Views, downloads and citations are aggregated across all versions of this paper published by eLife.