|Home | About | Journals | Submit | Contact Us | Français|
Kidney development has been studied over the past sixty years as a model of embryonic induction during organogenesis. Wilms’ tumor-1 (WT1), that encodes a transcription factor and RNA-binding protein, was one of the first tumor suppressor genes identified, and was soon thereafter shown to be associated with syndromic forms of childhood kidney disease and gonadal dysgenesis. Kidney agenesis, resulting from a null mutation in the WT1 gene, was one of the first examples of organ agenesis resulting from a gene targeting experiment. Thus, the study of the WT1 gene and its encoded proteins has been at the forefront of developmental biology, tumor biology and the molecular basis for disease. WT1 is now known to have an important role in kidney progenitor cells during development. This review will discuss recent advances in our understanding of kidney progenitor cells, and the recent identification of WT1 target genes in these cells.
This review will focus on recent studies of progenitor cells in the developing kidney, and on the role of the Wilms’ tumor-1 gene (WT1) in these progenitor cells. We have previously demonstrated WT1 to be essential for kidney development.1 Our most recent work has demonstrated that WT1 is likely to be a master control gene, that regulates the expression of a large number of genes that have a critical role in kidney development.2 Thus, by increasing our understanding of the function of the WT1 gene, we will fundamentally increase our understanding of how progenitor and stem cells are regulated in the developing kidney, and bring the field of kidney development closer to the time when we can either use stem cell lines from the developing kidney to regenerate kidney tissue in children or adults with ESRD, or alternatively, transform ES or IPS cells into kidney stem cells for the same purpose.
The initial metanephric mesenchyme contains the cells that will give rise to a nephron progenitor population in the developing kidney. This progenitor population is also known as the “cap mesenchyme.”3 Figure 1 provides a summary of nephron induction. After the initial condensation of induced progenitor cells around the bud, a portion of the condensed mesenchyme begins to form pre-tubular aggregates (PTAs) adjacent and inferior to the tips of the branching ureteric bud (Fig. 1B). Figure 2 summarizes a selected number of knockout phenotypes that have greatly informed our understanding of the early inductive stages of kidney development.
The metanephric kidney is the product of a series of inductive interactions between the metanephric mesenchyme and the epithelial ureteric bud.4 A transcription factor complex that includes Eya1 and Six1 appears to be required for the initial specification of the metanephric mesenchyme at the caudal end of the urogenital ridge.5–7 Ureteric bud outgrowth from the Wolffian duct is attracted by the neurotrophin GDNF, that is secreted by the metanephric mesenchyme under control of the transcription factor, Pax2.8–12 An additional transcription factor, WT1, is required for bud outgrowth.1 At least one target of WT1 appears to be VEGF-A,13 that appears to act on an angioblast population that may produce an as yet unidentified factor that is involved in stimulating Pax2 expression. In the absence of either WT1, Pax2, Eya1, Six1 or GDNF, there is usually complete renal agenesis.
Simultaneous with bud outgrowth, the metanephric mesenchyme becomes histologically distinct from the surrounding mesenchyme (blue in Fig. 1).3 After the initial contact of the ureteric bud with the metanephric mesenchyme, the condensed mesenchyme is also hypothesized to generate a population of stem and progenitor cells, which are found at the periphery of the nephrogenic zone in the developing kidney.14 Six2 is a transcription factor that appears to be responsible for maintaining cells in a progenitor state; in its absence there is premature differentiation of progenitors and loss of the progenitor population.15 For the purposes of this review, I will define a “kidney stem cell” as the population of cells that both has the capacity to self renew as a stem cell, and give rise to a more differentiated cell type. The histology of kidney development mandates the existence of a stem cell population. Otherwise, it would be impossible for the progenitor cells present at the beginning of kidney development to give rise to all the pre-tubular aggregates required for form the full complement of nephrons. What is presently unknown is whether the stem cell population is a subset of what we refer to as the “progenitor” population, that is functionally defined as the Six2 or Osr1-expressing cells, recognizing that Osr1 has a slightly more limited expression domain.16 Alternatively, the entire Six2 or Osr1 expressing population could be equivalent to the stem cell population, and there might not be any distinction between Six-2 expressing progenitors and stem cells in the developing kidney.
Why is it important to distinguish these two possibilities? First, if there is a true stem cell population, then these would be the most important cells to identify and manipulate for therapeutic use. It is likely that they would be able to maintain the ability to give rise to all the cell type in the nephron, and would have a better chance of being maintained indefinitely in culture, once the correct conditions are identified. In contrast, a non-stem progenitor would be more likely to be limited in its ability to be passaged indefinitely. Secondly, if the stem cell population is indeed a subset of the progenitors, then by definition, the non-stem cell progenitors are more differentiated, and may perhaps be regarded as a transient amplifying population, similar to the multipotent progenitor cells that are hypothesized to be present in the hematopoietic system. This leads to another important question—if the progenitors are more differentiated than the stem cell population, have they also become heterogeneous?
The nephrogenic zone of the developing kidney generates several rounds of nephrons before disappearing by around P6/7 in the developing mouse. Although a histologically distinct zone is apparent even until P5 or 6, there are actually few, if any, new nephrons being induced at that point, and that induction of new nephrons is probably complete by P3 or 4.17 Thus, the window for induction of the great majority of nephrons seems to be between E16.5 and P2. These observations are consistent with the quantitative results of Cebrian et al. who observed that there are only 700 nephrons present by E16.5, but approximately 8,000 nephrons by P0.18
Although the histology of the nephrogenic zone strongly suggested the existence of a stem-like population within the progenitor cell population, rigorous proof for stem-like cells has only been published in the past few years. In three studies, an inducible lineage specific Cre recombinase was transiently expressed in the cap mesenchyme, under the control of the either the Cited1, Osr1 or Six2 regulatory regions.19–21 In all three cases, cell lineages were permanently marked using a reporter transgene in which the expression of the LacZ gene is turned on after Cre recombinase removes a poly-A addition sites flanked by LoxP sites. Once Cre-mediated recombination removes the poly-A addition sites, LacZ is transcribed in all descendants of the cell in which the recombination took place, regardless of whether Cre remains expressed in those cells. These studies showed that if Cre was induced early in kidney development, and kidneys were examined for LacZ expression several days later, that LacZ was expressed both in descendent nephrons and in the cap mesenchyme, demonstrating that the cap mesenchyme contained a stem-like population that could both self renew and give rise to differentiated structures. In particular, Osr1 appears to be expressed in a somewhat more restricted population than Six2, and may mark a subset of the cap mesenchyme that contains stem-like cells. Importantly, these experiments do not distinguish whether the stem-like population is a small subset of the Six2, Cited1 or Osr1-expressing cells, or all of these cells. Moreover, as previously discussed, the interpretation that these experiments demonstrate the presence of a self-renewing population is based on the assumption that at the time of transient activation of Cre recombinase, there were insufficient progenitor cells present to give rise to all the pre-tubular aggregates that would be required to provide the full complement of nephrons in the adult kidney.
BMP7 shows three patterns of expression during nephrogenesis. First, it is expressed by the progenitor cells themselves, early in kidney development.22–24 After this mesenchymal expression, it becomes more highly expressed by the ureteric bud than by the mesencyhme. Additionally, as nephrons are induced, a high level of BMP7 expression is observed in the nascent podocytes.24,25 Bmp7 null embryos do not exhibit renal agenesis, but have poorly developed kidneys, that appears to be due to premature loss of the progenitor population.22,23
In the Bmp7 null mutant kidney, the first round of nephrons is induced, after which there is loss of the progenitor population.22,23 To better understand the role of BMP7 in progenitor cells and in nephron differentiation, we targeted conditional mutations to the Bmp7 gene in podocytes.25 Conditional knockout of Bmp7 in podocytes, did not yield as severe a phenotype as observed in Bmp7 null embryos, but nevertheless led to dysgenic nephron development and premature loss of the progenitor population.25
A role for FGF signaling in kidney development was first suggested by the work of Alan Perantoni, who showed that FGF2 had the ability to maintain survival of rat metanephric mesenchymes,26 removed from the ureteric bud, and not placed next to embryonic neural tubes. Additionally, Dudley et al. placed isolated murine metanephric mesenchymes in the presence of FGF2 and BMP7 for 48 hours prior to being exposed to embryonic neural tube and showed that together these factors were able to maintain the survival and inductive competence of the mesenchyme, whereas neither factor by itself, nor medium alone was effective.27 (The difference between rat and murine organ culture experiments is not fully resolved, though it may relate to differences in maturity of the rudiments when they are commonly removed from the respective embryos.) These experiments provided an early indication that BMP and FGF signaling were likely to interact in kidney development. Subsequently, the laboratory of Dr. Carlton Bates provided genetic evidence for a crucial role for FGF signaling in early kidney development, by showing that conditional knockout of Fgfr1 and Ffgr2 in the metanephric mesenchyme led to a renal agenesis phenotype, and conditional knockout of Fgfr2 in the ureteric bud also greatly decreased branching.28 In the Fgfr1/2 conditional mesenchymal knockout, early expression of Eya1 and Six1 are not affected, whereas WT1, Pax2 and Sall1 are dramatically decreased. This is consistent with other published reports suggesting that the Eya1/Six1 pathway is independent of the WT1-related pathway. 6,7 Additionally, chromatin immunoprecipitation data and organ culture knockdown results (Fig. 3) now demonstrate that Pax2 and Sall1 are targets of WT1.2
Despite the early work with FGF2, the more recent results discussed above suggest that FGF8 may be the more important FGF with regard to a role affecting nephron induction and maintenance of the progenitor population.29,30 PTAs express Fgf8 and Wnt4.29,30 Wnt4 has the ability to induce a mesenchymal to epithelial transformation of the aggregate to become the renal vesicle,31,32 a simple tubule that undergoes extensive growth, segmentation and differentiation to give rise to the mature nephron, that contains many different cell types divided between multiple segments that have a distinct physiological functions. As the pre-tubular aggregates are forming, the ureteric bud continues to branch (at least for several additional round of branching), and these dichotomous branching events lead to the induction of additional rounds of nephrons. In the absence of FGF8, Wnt4 is not expressed, but a small number of mesenchymal to epithelial transformations proceed as far as the renal vesicle, though no further,29,30 raising the question about whether Wnt4 is indeed primarily responsible for the mesenchymal to epithelial transformation, though this may also be related to incomplete knockout of FGF8 in this model.
The Wilms’ tumor-1 (WT1) gene encodes a zinc finger protein that acts both as a transcription factor and an RNA binding protein.33,34 We have focused on the transcription factor functions. Mouse embryos developing without WT1 display renal and gonadal agenesis.1 Over the past fifteen years, many genes have been published to be WT1 target genes.35 However, nearly all these studies depended on the use of immortalized cell lines, and few of the identified targets offered further insights as to why there was renal agenesis in the absence of WT1, though it is more likely that these studies may be relevant to the role of WT1 in neoplasia. Therefore, we sought to identify bona fide WT1 target genes by using a genome wide chromatin immunoprecipitation approach, using chromatin obtained from embryonic kidneys.2 Furthermore, to avoid biased PCR-based amplification of chromatin, we obtained chromatin from over 1,000 embryonic mouse kidneys. Many of our identified targets have been verified by direct ChIP assays, also using chromatin from embryonic kidneys. Amongst those tested thus far by direct ChIP, 26/28 genes with an enrichment score of >8 on the microarray verified as being bound by WT1, whereas only 4/15 genes with enrichment scores <8 verified by direct ChIP. Therefore, while some bona fide WT1 targets may have lower enrichment scores on the microarray, a score of >8 suggests a high likelihood of being a true WT1 target to the extent these can be identified by WT1 direct ChIP. Additionally, we used a novel membrane permeable-morpholino-based approach to knockdown WT1 expression in embryonic kidney organ culture, to validate these targets in a biological assay (Fig. 3). This study yielded a set of target genes that included genes already known to have an important role in kidney development, such as Pax2, Bmp7 and Sall1 (reviewed in refs. 36 and 37) (Fig. 3), as well as many novel target genes. In contrast, Six2, one of the most prominent markers of the progenitor population,20 which had an enrichment score of 2.0, did not validate either by direct ChIP or in the morpholino experiment, suggesting that it is not a target of WT1. This was a particularly useful result, as the retention of Six2 expressing cells after morpholino treatment (Fig. 3) served to demonstrate that intermediate doses of the WT1-specfic morpholino did not simply cause the apoptosis of the progenitor population. Rather, these Six2-expressing cells were still present, allowing us to specifically demonstrate loss of expression of WT1 target genes.
This set of target genes makes it very clear why there is renal agenesis in the absence of WT1. Several of the more interesting novel target genes are as follows:
Finally, we should note some histological and gene expression relationships that would also support a role for factors expressed by the PTA and renal vesicle in the maintenance of the progenitor population. Figure 4A shows an antibody stain for WT1 in a section through a 24-week human fetal kidney. Figure 4B is an in situ hybridization for WT1 mRNA expression through an E16 mouse embryonic kidney. In Figure 4A, (a) marks WT1 expression in the progenitor population; (b) WT1 expression in stroma, as the cells appear looser and in between the trunks of the ureteric bud (U); (c) marks WT1 expression in a forming PTA inferior to the ureteric bud; (d–g) are segments of a S-shaped tubules, (d) shows the cup-like arrangement of podocyte precursors, (e) is the proximal tubule segment that is heading out of the plane of section, (f) is the distal tubule segment that comes back into the plane of section, and (g) is the connection between the S-shaped tubule and the ureteric bud (U). (i and h) is an area of WT1 expression that is also probably part of an immature glomerulus from an irregular section of the domain marked by (d). Figure 4B is an in situ hybridization showing WT1 expression in an E16.5 mouse kidney, the letter markings are the same as in panel A.
The importance of showing these expression patterns and histology is to draw attention to several major aspects of the WT1 expression domains and histological relationships. (1) The PTA is immediately adjacent to the progenitor population such that FGF8 expressed by PTA would not have to travel far to find target cells in the progenitor pool; (2) there is contiguous WT1 expression from the progenitor domain to the PTA (i.e., from (a) to (c), suggesting that despite the restriction of Six2 to the progenitor domain and its exclusion from the PTA, that there may also be common molecular processes occurring in both the progenitor domain and PTA, such that the progenitor domain and PTA could be viewed as a continuum, rather than as two entirely separate compartments. (3) Given the close proximity of the PTA and nascent podocytes to the progenitor domain, that the high level of WT1 expression in the PTA and podocytes could represent a large source of BMP7 expression that could also affect adjacent PTAs and progenitor cells. (4) The PTAs, marked as (c), are at the same locations relative to the ureteric bud, as S-shaped tubules (d–f). Therefore it appears that more mature S-shaped tubules are probably adjacent to less mature PTAs. This three-dimensional relationship is often lost in schematic drawings that depict the S-shaped tubules only as a successive structure of the PTA and not as structures that are adjacent to other PTAs and that have the potential to affect their growth and differentiation. Nevertheless, the close proximity between PTAs and progenitor populations supports the hypothesis that maintenance of progenitor cells could be affected by gene expression in the PTAs, renal vesicles and S-shaped tubules that were previously induced, and explain why the failure to form PTAs or induce nephrons could affect the progenitor population.
Much has been learned about WT1 and kidney progenitor cells from the past 20 years of intense study. The approaches of gene targeting, genetic lineage tracing, and chromatin immunoprecipitation have helped focus these studies and have provided a strong in vivo component to a wealth of information obtained from studies performed in vitro. Future studies will most likely involve further use of the organ culture morpholino system, more sophisticated approaches to gene targeting and lineage tracing, and more complex studies that combine the two. Although not commented on in this review, it is likely that we will learn as much from the study of WT1 as an RNA-binding protein in the next ten years, as we have learned about its function as a transcription factor in the previous decade.
Dr. Rafi Kopan, Professor of Medicine and Developmental Biology Washington University School of Medicine: I would like to ask you a question about BMP7. In the flox BMP7 you showed this deterioration along the nephrogenic zone even though you produced a deletion in the glomerulus. This is fascinating, because in our NOTCH 2 knockouts,46 there is no glomerulus and there are no podocyte precursors but the nephrogenic zone is intact and it continues. Is there any chance that you got some dilution of the progenitors themselves and that is part of the expression or do you think that there is some compensating mechanisms we should begin to look at in our system.
Dr. Jordan Kreidberg: That is a fascinating question. When I was talking with some of your fellows, we discussed this a little bit as well. I think that there are a few things going on. First of all, I think that we only make one round of probably good nephrons here. Whereas, I think in your system you are continuing to make them, even though they are all missing proximal segments. I am not sure we can reconcile the two sets of results at present, with regard to their respective effects on the progenitor population. In both systems, it will be important to examine FGF8 expression in the pretubular aggregate, as that is known to be important for maintaining progenitor cells.
Dr. Kopan: We are clearly downstream of where you are, it is just fascinating to me that you have BMP expression segregating to the podocyte precursor and that is the cell type we totally loose. But something still compensates.
Dr. Kreidberg: Yes.
Dr. Sun-Young Ahn, Instructor in Pediatrics, Washington University School of Medicine: You mentioned that Six2 plays a role in maintaining the nephron progenitor pool and that Six2-null mice present with small kidneys and premature differentiation. The podocyte-specific BMP7 knockout mice you presented share many similar features with the Six2-null mice. Did you also observe the premature differentiation seen in Six2-null mice?
Dr. Kreidberg: This is a very interesting point. We are very interested in whether BMP7 has a role in maintaining self-renewal of the progenitor population, or is mainly stimulating proliferation.
Dr. Jeffrey Miner, Professor of Medicine and Cell Biology and Physiology Washington University School of Medicine: You seem to have gone into the ChIP on chip study with a bias that WT1 is a positive activator rather than being a negative regulator in some cases. Have you found any instances where it seems to be negatively regulating any targets?
Dr. Kreidberg: I don’t think we had that bias. It is more how we have ended up. You are correct in that we have not yet found any instances where it seems to be negatively regulating targets. That may be either because we just haven’t looked at enough examples or it may be because WT1, at this early stage, mainly has a positive stimulatory effect on transcription. Again, most of those studies on the repression were from immortalized cell lines and involved what was probably huge overexpression of WT1. On the other hand, in regarding the TGFbeta/BMP family as potential targets of WT1, certainly as you flush that out it will probably work out better if some of those turn out to be negatively regulated. For example, two of the targets are SMAD6 and SMAD7 that are inhibitory SMADS. So if on the one hand, WT1 is stimulating BMP7, why would it also stimulate inhibitory SMADs? If I had to pick my two leading candidates to be negatively regulated by WT1 it might be those inhibitory SMADS, but we won't know until we actually test them.
Dr. Helen Liapis, Professor of Pathology and Immunology and Medicine, Washington University School of Medicine: My questions is somewhat related to the one you just answered. I would like to hear from you the implications of your work to human disease and particularly to patients with WT1 mutations, in that they appear to have two distinct glomeruli phenotypes. As you mentioned one phenotype is the diffuse mesangial sclerosis (DMS) and the other is FSGS which we think of as a podocyte disease. Can you summarize what you think may underline the phenotypes in humans when it comes to WT1?
Dr. Kreidberg: Well let me start with the Wilms’ tumor and then I will come back to the glomerular disease. One other tumor suppressor gene for Wilms’ tumor is WTX, identified by Dan Haber’s lab.47 Soon thereafter, Randy Moon’s lab showed it to be a negative regulator of WNT signaling.48 So now we have 2 cases in Wilms’ tumor where negative regulators of WNT signaling are tumor suppressor genes, WT1, thru CXXC5 or WTX. It may be that hyper-expression of beta catenin is very important in Wilms’ tumor and that that is normally negatively modulated by WT1. This is certainly the present thinking with Wilms’ tumors.49 With regard to glomerular disease, I think we know less and it will be very intriguing to see, when we finally get done with the ChiP/seq for WT1, how much overlap there is in the targets identified in progenitor cells. BMP7 would be a prime candidate to be regulated by WT1 in podocytes since they are coexpressed there and the role of podocyteexpressed BMP7 in kidney disease is a totally unexplored area. There is also an evolving story of beta-catenin and Wnt signaling in glomerular disease.50,51 WT1 could be having an influence in minimizing WNT signaling in the podocyte as a way of maintaining normal podocytes. Beyond that, alpha-actinin-1 was one of our top targets. Alpha-actinin-4 is the main one so far associated with FSGS, but alpha-actinin-1 is probably expressed in podocytes as well. It may be that through the regulation of cytoskeletal associated proteins, WT1 may have an important role in maintaining normal podocytes.
Dr. Fanxin Long Associate Professor of Medicine, Washington University School of Medicine: I was wondering if you would comment on what is the percentage of the target genes you identified by ChIP chip that are actually changed in expression when you do expression profiles comparing wild type with mutant embryos. Do you have any information?
Dr. Kreidberg: You are asking at the microarray level?
Dr. Long: Right. If you compare the two lists and all the targets you identified by ChIP chip are validated by microarray, what is the percentage.
Dr. Kreidberg: We haven’t been able to do that at a global level yet. Recall that the WT1 knockout has essentially renal agenesis. For that reason, we can’t get enough valid RNA from those mutant rudiments at E11.5. We may be able to use the transgenic knockdown that I showed you to do microarrays. However, the knockdown is really brand new and so we haven’t gotten as far as doing microarray studies on gene expression in this model.
Dr. Scott Boyle, Postdoctoral Fellow in Nephrology Washington University School of Medicine: I wonder if you could tell us anything about WT1 binding partners, both as a way to link WT1 into core developmental pathways and to account for the location specific targets of WT1, presumably as different targets in mesenchyme than have been in the podocyte.
Dr. Kreidberg: We tried using informatic studies to see if we could find any other known transcription factor sites that were commonly adjacent to this binding site that I showed you using standard data bases for transcription factors. That was unsuccessful in showing that any known transcription factor site was often adjacent to putative WT1 binding sites. I think it is going to mainly take more proteomic studies to find out what other transcription factors or histone modifying enzymes WT1 may be associated with to regulate gene transcription.
Dr. Hila Barak, Postdoctoral Fellow in Developmental Biology, Washington University School of Medicine: I know that you mentioned in WT1 morphant kidney the same number of Six2 expression cells were detected. However, in the picture that you showed, the structure looks different and the cells around the ureteric bud are not intact. Do you think that part of the effect that you have found might occur as a result of a morphology change?
Dr. Kreidberg: I think what you are noticing is that that the Six2 population does look less compact. The normal compaction of the progenitor population around the ureteric bud may not be occurring. Interestingly, syndecan1 was identified some years ago as a potential WT1 target.52 Syndecan1 did not turn up as a target in our study. However, it is certainly possible that proteoglycans involved in that compaction may not be expressed properly, whether they are direct targets or indirect targets.
Dr. Kopan: You made the case that perhaps one role for WT1 is to create a permissive environment by regulating a cohort of chromatin modulators. Did you try to take your morpholino-treated tissue into culture and determine whether a combination of sodium butyrate and maybe 5-Aza-Cytidine could undo the effect of loss of WT1?
Dr. Kreidberg: No, we haven’t done anything to rescue the morpholino with those sorts of studies.
Previously published online: www.landesbioscience.com/journals/organogenesis/article/11928
Edited transcripts of research conferences sponsored by Organogenesis and the Washington University George M. O’Brien Center for Kidney Disease Research (P30 DK079333) are published in Organogenesis. These conferences cover organogenesis in all multicellular organisms including research into tissue engineering, artificial organs and organ substitutes and are participated in by faculty at Washington University School of Medicine, St. Louis, Missouri USA.