Kidney development begins with interactions between the ureteric bud, an epithelial outgrowth of the Wolffian duct, and the surrounding metanephric mesenchyme. These mutually inductive interactions results in formation of the branched collecting system from the ureteric bud and most of the tubular nephron as well as the epithelial portion of the glomerulus from the metanephric mesenchyme.As a result of a great deal of in vivo knockout and in vitro work, as well as expression profiling, key molecules regulating various aspects of kidney development have been identified.A systems biology perspective on renal organogenesis is gradually emerging.
kidney development; Wolffian duct; ureteric bud; metanephric mesenchyme; nephrogenesis; tubulogenesis
In the course of its development, the mammalian kidney goes through three distinct forms: the pronephros; the mesonephros; and the metanephros, ultimately leading to the formation of the mature kidney ( Figure 25.1 ). At day 22 of gestation in humans or at day 8 in mice, an epithelial streak called the pronephric duct arises in the cervical region of the developing embryo from intermediate mesodermal cells induced to undergo the transition to epithelial cells in response to signals arising from the somite and surface ectoderm. The pronephric duct then extends caudally to form the nephric duct or Wolffian duct. The most primitive kidney, the pronephros, is formed as the pronephric duct induces surrounding mesenchyme to form the pronephric tubules. Glomerulus-like structures (glomera) are also seen, but are not physically connected to the tubules forming a non-integrated nephron. The pronephros is functional only in fish and amphibians; it is thought to be rudimentary and non-functional in higher vertebrates.
Next, a more complex “protokidney,” the mesonephros, arises just caudal to the pronephros at day 24 in humans or day 9.5 in mice. As with the pronephros, mesonephric development starts with induction of the surrounding mesenchyme by the Wolffian duct. Unlike the pronephros, however, the mesonephros glomeruli are linked to the Wolffian duct via mesonephric tubules. In humans, about 30 nephrons are observed in the mesonephros; their function is unclear. The mesonephric duct and some tubules persist, and are ultimately integrated into the male genital system forming, in part, the vas deferens and tubules of the epididymis. In females, the mesonephros degenerates and disappears.
The permanent kidney of amniotes, the metanephros, starts to form at day 28 in humans or day 11 in mice. Unlike the pronephros and mesonephros, which are induced by the Wolffian duct, metanephric tubules are induced by an epithelial structure derived from the Wolffian duct: the ureteric bud. The ureteric bud is induced to evaginate from the Wolffian duct in response to signals arising from the metanephric mesenchyme, a loose aggregation of intermediate mesodermal cells. The emergence of this epithelial progenitor tissue of the metanephric kidney is a key initiating event, and depends upon differentiation of the metanephric mesenchyme from the intermediate mesodermal cells. As the ureteric bud invades the surrounding mesenchyme, it induces the mesenchymal cells to form epithelial metanephric tubules that eventually differentiate into the proximal through distal portions of the nephron. The ureteric bud, reciprocally induced by the metanephric mesenchyme, undergoes branching morphogenesis, eventually giving rise to the collecting system. Morphologically, nephron formation is completed by birth in humans, although only after birth does the nephron become fully functional.
Development of the Metanephros
Here, the development of the metanephros is described in more detail ( Figure 25.2 ), since it becomes the final kidney in mammals. As discussed in the previous section, the ureteric bud, the inducer of metanephric development, is formed from the Wolffian duct. This structure invades the metanephric mesenchyme, whereupon mesenchymal cells condense at the tip of the ureteric bud. The condensed mesenchymal cells then differentiate into epithelial cells: the so-called mesenchymal-to-epithelial transformation (MET). The newly formed epithelial cells gradually develop into distinct structures called “comma-shaped bodies,” which subsequently become “S-shaped bodies.” The S-shaped bodies, which begin to exhibit tubular morphology, continue to elongate; the end closest to the ureteric bud connects to it, while the opposite end forms podocytes and Bowman’s capsule. The middle part ultimately differentiates into the proximal through distal tubules of the nephron. At the same time, the tips of the ureteric bud, induced by the metanephric mesenchyme, continue to branch to ultimately form the collecting ducts, renal pelvis, calyces, and papillae.
The process of collecting system development has been studied in detail by microdissection of the human kidney. Initial ureteric bud branching is dichotomous and symmetric; the ureteric bud takes on a T-shape in this early stage of metanephrogenesis. Subsequently, the ureteric bud elongates and bifurcates at the tips, and eventually branches again dichotomously. At later branching events, the angle between branches lessens and branching may not be completely symmetrical, so that somehow the ureteric tree structure “fits” into the final shape of the kidney. The initial branches become dilated to form the renal pelvis, while terminal branches become collecting ducts. As the ureteric bud undergoes branching morphogenesis, its tips continue to induce more nephrons from the metanephric mesenchyme.
Vascular development occurs along with nephron development. Dissection of the developing kidney reveals that large vessels branch off from the dorsal aorta invading into the kidneys. As will be discussed later, the extent to which the microvasculature of the kidney is derived from cells of the metanephric mesenchyme versus cells from outside the kidney is still unclear (reviewed in ). The origin of the mesangial cells, which closely associate with the endothelial cells in the glomerulus, also remains uncertain.
Thus, the reciprocal mutual induction and feedback between the ureteric bud and the metanephric mesenchyme represent key events in metanephric epithelial tissue development leading to a functional kidney. These reciprocal interactions induce branching morphogenesis of the ureteric bud, together with epithelialization and tubulogenesis of the metanephric mesenchyme ( Figure 25.3 ). To better understand the mechanisms underlying induction, it is necessary to identify and analyze the key molecules that mediate signals between the metanephric mesenchyme and the ureteric bud.
Experimental Approaches to Kidney Development
Over the years, a variety of experimental approaches have been used for evaluation of the mechanistic details of the induction process between the ureteric bud and metanephric mesenchyme. For example, the developing kidney has been found to be amenable to extensive in vitro analysis. In addition to this “whole embryonic kidney organ culture,” it has been demonstrated that progenitor tissues (i.e., Wolffian duct, ureteric bud, and metanephric mesenchyme) can be isolated and cultured individually. Advances in genetic manipulation have allowed the analysis of kidney development in vivo in genetically-engineered mice. As in vitro culture techniques and in vivo genetic manipulation become increasingly sophisticated, it is becoming clear that these approaches should be viewed as complementary.
Whole Embryonic Kidney Organ Culture
The transfilter culture system, used by Grobstein and co-workers in the 1950s, has been the mainstay of in vitro organ culture work in the developing kidney. In this system, microdissected kidneys, from as early as the beginning of metanephrogenesis (gestational day 11.5 in mice or day 13.5 in rats), are cultured on top of filters for several days. In the presence of appropriately defined serum-free medium, kidney rudiments grow and differentiate. It is possible to observe branching morphogenesis of the ureteric bud, induction of the metanephric mesenchyme, and formation of nephrons by microscopy as the cultured embryonic kidneys develop ( Figure 25.4a ). Only vascular development does not occur to an appreciable extent. Thus, not only does the whole embryonic kidney culture resemble in vivo developmental processes, but it also appears to retain the inherent spatiotemporal complexity.
In this whole embryonic kidney culture, it is possible to manipulate humoral factors that play a role in nephrogenesis. The effects of growth factors or their inhibitors on kidney development can be evaluated in vitro by analysis of total kidney size, ureteric bud branching events, and metanephric mesenchyme tubulogenesis. For example, ureteric bud branching can be assayed by staining with a fluorescently-labeled lectin from Dolichos biflorus , which has been shown to bind specifically to the cells derived from the ureteric bud ( Figure 25.4b ). However, the organ culture method is not without its limitations. For example, when antibodies and antisense oligonucleotides are used to perturb in vitro nephrogenesis, care must be taken to ensure that the agent is delivered to the sites of interest, since antisense oligonucleotides do not seem to penetrate the ureteric bud as well as the metanephric mesenchyme. In addition, oligonucleotide toxicity may, in some instances, nonspecifically inhibit kidney growth. However, recent development of RNAi technology to perturb specific gene function may prove useful in this setting.
Isolated Wolffian Duct Culture
It is possible to culture the entire mesonephros–metanephros area on top of transfilters. Addition of humoral factors can induce outgrowth of ureteric bud-like structures from the Wolffian duct. For example, the addition of glial cell-derived neurotrophic factor (GDNF) induces numerous budding events at multiple foci along the length of the cultured Wolffian duct ( Figure 25.5 ). The epithelial Wolffian ducts can also be dissected away from most of the non-epithelial mesoderm and cultured in the presence of soluble growth factors to induce the outgrowth of ureteric bud-like structures. However, in this culture system GDNF alone is insufficient to induce the outgrowth of ureteric bud-like structures, and supplementation with other growth factors is required ( Figure 25.6 ). The Wolffian duct can also be cultured as a “naked” epithelial tube cleared of all surrounding mesodermal cells, although these isolated ducts must be cultured within a three-dimensional extracellular matrix. The ureteric bud-like structures can be excised from the Wolffian duct and induced to branch in culture ( Figure 25.7 ), indicating that the in vitro ureteric bud possesses the ability to branch and grow in a manner similar to that seen with the in vivo ureteric bud (see below). These ex vivo culture systems have proven useful in the elucidation of the mechanism of ureteric bud budding, and have allowed for the identification of multiple modulators/regulators (e.g., growth factors, signal transduction pathways, etc.) of this process.
Isolated Ureteric Bud Culture
Since the 1950s, in vitro culture of the two individual components of metanephros, the ureteric bud and the metanephric mesenchyme, has been attempted. Of the two progenitor tissues, in vitro growth of the isolated ureteric bud proved to be more difficult, and it was argued that cell–cell contact between the ureteric bud and metanephric mesenchyme was an important component in the process of ureteric bud branching. However, the isolated ureteric bud has since been shown to grow and branch extensively in the presence of appropriate extracellular matrix and soluble factors in the absence of direct contact with the metanephric mesenchyme ( Figure 25.8 ). The epithelial cells of the ureteric bud in this culture system appear to retain similar morphological characteristics to those of the ureteric bud in the whole embryonic kidney culture. This system has allowed for the isolation and identification of numerous molecules, including soluble factors that modulate ureteric bud branching morphogenesis.
Isolated Metanephric Mesenchyme Culture: Recombination with Heterologous Inductive Tissues
Another setting in which organ culture has been used focuses on metanephric mesenchyme induction and transformation to an epithelial phenotype. In this system, the isolated mesenchyme is cultured on one side of a filter while, on the other side of the filter, heterologous inducing tissues are placed. Various stages of metanephric mesenchyme induction (i.e., condensation, epithelialization, and tubulogenesis) can be observed, depending on the inductive capacity of the tissue. Using this method, it has been shown that embryonic spinal cord, salivary gland, and other tissues can induce the metanephric mesenchyme. As is the case for isolated ureteric bud branching described previously, a key question here is the relative contribution of humoral factor(s) or cell–cell contact in this process. Electron microscopic examination revealed that the inducing tissue can contact the metanephric mesenchyme via cellular processes extending through the filter. In fact, filters with pore sizes greater than 0.1 µm are unable to block cell-to-cell contact completely, indicating the importance of cell–cell contact. However, complete mesenchymal induction has been demonstrated in the presence of soluble factors without cell–cell contact with inductive tissue.
Isolated Metanephric Mesenchyme Culture: Recombination with Isolated Ureteric Bud
It has been shown that when co-cultured with freshly isolated metanephric mesenchyme, the isolated ureteric bud in culture (as described previously) is capable of inducing nephron tubules from the metanephric mesenchyme. At the same time, the pattern of ureteric bud growth is altered by the presence of the metanephric mesenchyme; it is only through contact with metanephric mesenchyme that the ureteric bud undergoes vectorial branching with elongation and tapering of newly induced branches, similar to those seen in cultured whole embryonic kidneys ( Figure 25.9 ). Although this patterning effect on the ureteric bud appears to be modulated, in part, by soluble factors, cell contact with the metanephric mesenchyme and/or short-acting factors produced by the interaction of these two progenitor tissues plays a key role in determining the arborization pattern. Another potential application of this “recombination” system is to pinpoint the defective tissue in knockout mice with a kidney phenotype. Through recombination of wild-type and gene knockout tissues (e.g., wild-type ureteric bud with knockout metanephric mesenchyme), it may be possible to determine the source of the kidney defect. For example, kidneys lacking heparan sulfate 2-O sulfotransferase (Hs2st −/− ) display renal agenesis, presumably due to defects in the inductive responsiveness of the ureteric bud to undergo branching morphogenesis. However, examination of cultures of recombined wild-type and Hs2st −/− ureteric buds and metanephric mesenchyme ( Figure 25.10 ) provided evidence that the key defect in the knockout kidney is the inability of the metanephric mesenchyme to undergo induction.
The aforementioned tissue culture systems allow one to observe certain key phenomenon in metanephric kidney development in vitro/ex vivo. The analysis of genetically-engineered kidneys (and/or their component tissues) in these in vitro systems, in combination with an advanced method of gene perturbation (e.g., RNAi), will undoubtedly provide a more mechanistic picture of kidney development.
Genetically-engineered mice allow one to manipulate the process of kidney development in vivo . Introducing null mutations of the gene of interest into mouse embryonic stem cells can be used to generate gene knockout mice. Generally speaking, knockout mice grow from conception without normal expression of the gene product. If the mice develop beyond the stage of kidney development, the effect of the gene disruption on nephrogenesis can be observed in vivo by direct examination of the tissue histology. In such cases, the abnormal kidney phenotype can be directly or indirectly ascribed to disruption of the gene. In fact, many important molecules involved in kidney development have been identified by gene knockout technology. In particular, the contribution of many transcription factors, molecules acting in the nucleus to regulate gene expression in the cell, have been demonstrated by this technique. However, knockout technology has its limitations. For example, although gene knockout mice can demonstrate the indispensability of a particular gene, how the gene product acts in the complex process of kidney development often remains unclear, owing to the spatiotemporal complexity of the developing organ. Thus, if one observes defective ureteric bud branching in knockout mice, the deleted/disrupted gene product could be affecting the ureteric bud directly or it could be affecting the metanephric mesenchyme, resulting in incompetent induction of the ureteric bud. To resolve this problem, some have used organ culture type approaches (“recombination”). Another limitation of gene knockouts is that the mice may not have an apparent phenotype due to redundancy. In other words, the expression of other molecules with features or functions that overlap with the targeted gene could compensate for the defect from the gene knockout.
Perhaps the major drawback of conventional knockouts for studying genes involved in development of the kidney is the possibility that the targeted gene is critical for early embryonic survival, rendering the analysis of kidney development impossible as the embryo dies before organogenesis. In recent years, this issue (as well as some of the others listed above) has been overcome by the use of tissue-specific knockout technologies which provides the means for gene disruption in a time- and organ-specific manner. This gene targeting system utilizes site-specific recombinases to excise out genes or portions of genes from the genomic DNA, resulting in the inactivation of the gene of interest. Cre-loxP is the most commonly employed site-specific recombinase for these “conditional” deletions, although other site-specifc recombinases are available, including Flp-FRT and Dre-Rox. In the case of the Cre-loxP system (although the general principles are shared), conditional deletion of a gene of interest is performed by crossing two transgenic mouse lines: (1) “floxed” mice that carry the gene of interest with flanking loxP sites which can be cleaved by the enzyme; and (2) Cre-recombinase transgenic mice under the control of a tissue-specific promoter. For example, deletion of β1 integrin from the epithelial cells destined to become the ureteric bud using Cre recombinase under the control of the HoxB7 promoter disrupts ureteric bud branching morphogenesis, and variably retards kidney growth, leading in a few instances to renal agenesis. While such approaches have proven useful, the main drawback of these conditional deletions in the study of the kidney is the availability of cell-specific promoters. The use of tet-operon and tamoxifen to induce recombinase activity and the modulation of gene activity have also proven to be useful. In these systems, animals are exposed to tetracycline or tamoxifen which activates the site-specific recombinase under its control, resulting in the modulation of gene activity. These systems have the advantage of being under the control of the investigator; however, they are not without problems, including the toxicity of the inducing agent. Nevertheless, these spatiotemporal conditional deletions have rapidly established themselves as an invaluable tool for investigating the development of the kidney.
Cell culture models have the advantage of simplicity. Since they use homogenous cell populations grown under controlled conditions, it is possible to perform biochemical analysis in great detail. Moreover, gene introduction by plasmid transfection and gene knockdown by RNAi is simpler in comparison to the organ culture system. Here, the most relevant system for branching morphogenesis of the ureteric bud, the three-dimensional cell culture system, is discussed.
When certain kidney-derived epithelial cells are suspended in an extracellular matrix gel (type I collagen or a collagen–Matrigel mixture) in the presence of morphogenetic humoral factors, they form tubules and undergo branching morphogenesis in vitro . The tubules in the three-dimensional culture have lumens and retain apical–basolateral polarity. The effect of humoral factors in epithelial tube and/or branch formation can be studied here. In addition to the humoral factors, the effect of extracellular matrix composition on morphogenesis and the cellular details of morphogenesis can be examined in this system. MDCK cells and murine inner medullary collecting duct (mIMCD3) cells have been used in the past, but these have the limitation of being derived from mature renal epithelial cells. To address this issue, an in vitro cell culture system using ureteric bud (UB) cells directly derived from embryonic day (E) 11.5 mouse ureteric bud has also been established. Although there is some difference in the responsiveness of the different cell lines to growth factors, all three cell lines respond to soluble factors produced by the metanephric mesenchyme by forming branching tubules.
A detailed mechanism of hepatocyte growth factor (HGF)-induced MDCK cell tubulogenesis model has been described. There appear to be two steps involved in this process of invasion: epithelial–mesenchymal transition; and the re-establishment of epithelial intercellular junctions. However, it has also been shown that ureteric buds in in vitro culture undergo branching morphogenesis through budding, a process in which epithelial cells never lose their junctions. It remains to be seen where and when these two morphogenetic processes, branching through invasion (“invadopodia”) and branching through budding, are utilized in vivo .
Molecular Approaches to Kidney Development
The development of high-density DNA microarray technology and global gene profiling has made it possible to analyze patterns of gene expression throughout embryonic and postnatal development and into adulthood in the whole developing rodent kidney. It has also been possible to analyze gene expression changes in in vitro culture systems such as the isolated UB and MM. For example, initial microarray analysis of a global time series of gene expression in the developing rat kidney revealed five discrete patterns or groups of gene expression ( Figure 25.11 ). mRNA encoding transcription factors and growth factors were found to be upregulated early in organogenesis (group 1). Among the genes whose expression level peaked in the middle of kidney organogenesis (group 2), many extracellular matrix related genes were found. Further representing the global time series gene expression data as self-organizing maps (SOMs) made it possible to define roughly six stages of gene expression during pre- and postnatal kidney development in the rat. Computational analysis suggested points of stability and transition based solely on gene expression and correlations where classically described anatomical changes were not intuitively obvious ( Figure 25.12 ). The most profound changes appear to occur at birth, when there is a sudden burst in the expression of many genes involved in redox metabolism and transport, including multispecific drug transporters such as Oat1 and Oct1.
There has also been a massive effort to create an atlas of gene expression in the developing kidney, the genitourinary developmental molecular anatomy project or GUDMAP. This multi-group international project is still continuing and has not only provided an atlas of localization information (e.g., ureteric bud, comma-shaped bodies, S-shaped bodies, renal vesicle), but has also yielded specific gene signatures for developing structures like branching ureteric bud tips. Although the function of many of these genes remains unknown, at a minimum they represent useful markers.
One of the key tasks in the future will be to place this localization information in the context of global gene expression time series data, to obtain a more accurate picture of the dynamics of kidney organogenesis and suggest new points of regulation. Growth factor-selective heparan sulfation interactions have been proposed as important regulators of the switching between stages. Moreover, based on current knockout and in vitro data, it has been suggested that key “hubs” in the network of genes regulating kidney development include the process of GDNF-dependent budding early on, and late tubulogenesis involving cilia-associated genes such as those implicated in various types of cystic kidney disease. It is becoming increasingly clear that an abundance of knockouts reported to have “renal phenotypes” cluster around these processes. In vivo branching morphogenesis, especially in the middle phases, appears largely protected from disruption in many knockouts of gene products known to be involved in in vitro branching. When branching phenotypes are reported, it is usually in the form of a small reduction in nephron number. One interpretation, supported by a wealth of in vitro data, is that there are many growth factor–heparan sulfate-dependent pathways regulating branching, and deletion of any single one is likely to be compensated by another. Double knockouts, for example of the EGF receptor and the HGF receptor (c-met), are beginning to provide support for this view.
In addition to these high-throughput gene profiling approaches, epigenetic transcriptional controls are becoming rapidly appreciated. These dynamic cell-inheritable processes alter transcriptional activity without affecting DNA sequence, and include covalent modifications of DNA and histones, DNA packaging, chromatin folding, and regulatory noncoding microRNAs (miRNAs). For example, conditional deletion of Dicer, the RNase involved in the production of miRNAs (which control gene expression at the post-transcriptional level), from cells of the nephron lineage lead to elevated apoptosis and premature termination of nephrogenesis. In addition, deletion of Dicer from ureteric bud epithelium disrupts branching morphogenesis, and leads to the development of renal cysts. Together with other studies on Dicer and miRNAs, the data clearly indicate a role for Dicer and Dicer-dependent miRNA activity in the development of the kidney, as well as in the development and progression of kidney disease.
Ultimately, high-throughput gene profiling, together with a thorough epigenetic analysis, may provide mechanistic insight into the very complex system of gene expression regulating kidney development. This may enable the development of a systems perspective on nephrogenesis. Attempts at creation of “coarse grained” models of kidney development have clearly begun. In the following section, a number of molecules which have been shown to be involved in kidney development are discussed. Over the past two decades, a large number of developing kidney phenotypes has been reported in gene knockout studies. Together with in vitro studies, they provide a great deal of functional information. We will highlight some of the results below. Several recent reviews describe them in much more detail. Moreover, we do not discuss in great detail the impressive amount of work that has been done in relation to the formation of the glomerular filtration barrier (reviewed in ), polycystic kidney disease (reviewed in ) or late nephron differentiation and acquisition of mature transport function (reviewed in ). We focus largely on the WD, UB, and early MM-derived structures.
Transcription Factors in Metanephrogenesis
Transcription factors bind to DNA and regulate the expression of other genes that are involved in, among other things, morphogenesis and differentiation. As a result of many gene disruption studies, several important transcription factors in kidney development have been demonstrated. With careful molecular marker analysis, it will soon be possible to draw a whole network of these molecules in this process.
Transcription Factors Regulating Glial Cell Line-Derived Neurotrophic Growth Factor
As will be described later, a key molecule in the process of the initial stage of metanephros development (i.e., ureteric bud formation and outgrowth from the Wolffian duct) is glial cell line-derived neurotrophic growth factor (GDNF). Many transcription factors regulating expression of this growth factor affect ureteric bud development, and thus kidney development.
Hox genes, mammalian homologs of Drosophilia homeotic genes, have been shown to be critically important for early nephrogenesis. While null mutants for Hoxa11 or Hoxd11 mice do not have a kidney phenotype, double knockouts of Hoxa11 and Hoxd11 show kidney agenesis or hypogenesis. Moreover, complete elimination of Hox11 paralogs ( Hoxa11 , Hoxc11 , and Hoxd11 ) result in a lack of ureteric bud outgrowth from the Wolffian duct. In this mutant, expression of another transcription factor, Six2 as well as Gdnf is lacking, suggesting that Hox11 paralogs regulate Gdnf expression.
Additional members of the homeotic gene family, the Pax genes, have been implicated in nephrogenesis. Compared with Hox genes, Pax genes appear to be restricted to certain tissues or organs. Pax2 and Pax8 have been shown to be expressed in the kidney. These Pax genes can be considered as early nephric lineage specification genes, as they are first expressed in the pronephric duct, and their simultaneous disruption causes failure in the formation of the epithelial pronephric duct from the intermediate mesoderm. After the pronephric duct, Pax2 expression is sequentially found in its extension, the Wolffian duct, the ureteric bud, as well as the condensed metanephric mesenchyme and the newly formed nephron tubules. As the kidney tubules mature, Pax2 expression decreases. The expression pattern suggests a role for Pax2 in mesenchymal–epithelial transformation. Homozygous null mutant mice lacking Pax2 show only a partially developed Wolffian duct, leading to kidney agenesis. It has also been shown that the mutant Wolffian duct does not respond to GDNF to form the ureteric bud. Furthermore, the mutant metanephric mesenchyme not only lacks Gdnf expression, it is not competent to form nephron tubules in response to wild-type spinal cord. Heterozygous Pax2 mutant mice have hypoplastic kidneys. In fact, there are Pax2-binding sites in the promoter region of Gdnf , and Pax2 can promote Gdnf expression in vitro . Pax2 has also been shown to promote the assembly of an H3K4 methyltransferase complex, which is involved in epigenetic transcriptional regulation.
Pax8 has a similar tissue expression pattern to Pax2 ; however, Pax8 expression peaks later than Pax2 . Although kidney development is apparently normal in Pax8 knockout mice, double knockouts of Pax2 and Pax8 show a complete absence of a urogenital system, due to failure in the formation of the pronephric duct from the intermediate mesoderm, suggesting some overlap in the roles of these two Pax genes in pronephric duct induction. Analysis of kidney development in mice heterozygous for Pax2 and for Pax8 , which form kidneys (albeit hypodysplastic with fewer ureteric bud tips and a reduced nephron number), indicates a dramatic reduction in the expression levels of Lim1 . Although normal levels of Ret and Gdnf were seen, Wnt11 (an important downstream target of Gdnf signaling, see below) was reduced. Thus, it has been postulated that Pax2 and Pax8 play a key cooperative role in nephron differentiation and branching of the ureteric bud.
Eya1, Six1, and Six2 Genes
These genes have been implicated in Drosophila eye development together with Pax6 , and are expressed in the metanephric mesenchyme in the kidney. Homozygous null mutants for Eya1 show kidney agenesis with loss of Gdnf and Six expression, suggesting that EYA1 acts upstream of SIX, and together they regulate Gdnf expression. In fact, EYA1 is shown to act as a co-activator of the genes regulated by SIX. In Six1 knockouts, which show various kidney phenotypes ranging from hypogenesis to agenesis, Gdnf expression is reduced, but Eya1 expression is preserved. Interestingly, metanephric mesenchyme derived from Six1 knockout mice is not competent in nephron tubule formation when it is cultured with spinal cord, a potent inducer of nephron tubulogenesis, suggesting that these factors not only control Gdnf expression, but also have a role in maintaining certain characteristics of metanephric mesenchyme. As described previously, another member of the Six family, Six2 appears to regulate Gdnf expression downstream of Hox11 paralogs. Although Six1 and Six2 show overlapping areas of expression, the fact that Six2 expression is reduced in Six1 knockouts suggests a close relationship between these two molecules. Six2 expression in vivo has been found to be directly activated by a novel protein complex composed of the Hox11 paralogous proteins, Pax2 and Eya1, which clearly demonstrates that Six2 and Gdnf are downstream targets of the Hox11 paralogs. Moreover, Six2 defines nephron progenitor populations in the metanephric mesenchyme and it cell-autonomously maintains progenitor populations.
Sall1 is a transcription-related protein expressed in the metanephric mesenchyme of the developing kidney. It is also expressed in extrarenal tissues such as the limb buds and central nervous systems. Knockout of this gene results in failure of the ureteric bud to undergo branching after invading the metanephric mesenchyme. Although Gdnf expression just before ureteric bud formation is reported to be normal, its expression in the metanephric mesenchyme subsequently decreases. It is unclear whether this reduction of Gdnf expression is due to a direct effect of the Sall1 mutation or if it is secondary to a loss of a ureteric bud-derived signal. Interestingly, exogenous Gdnf was unable to rescue branching in cultures of Sall1 knockout kidneys despite expression of Ret , suggesting that the ureteric buds are unable to respond to Gdnf. In situ hybridization demonstrates the expression of the stalk-specific marker, Wnt9b , as well as the β-catenin target gene Axin2 , in the ureteric bud tips of Sall1 knockout kidneys. Since reduction of β-catenin levels in Sall1 mutants rescued ureteric bud branching and overexpression of Wnt9b -inhibited branching in normal ureteric buds, the data indicate that Sall1 -dependent signals regulate the initiation of ureteric bud branching by modulating the expression of ureteric bud tip-specific genes. In addition, among metanephric mesenchyme cells, only those cells which express high levels of Sall1 are capable of nephron formation, suggesting the possibility that activation of this gene is key for nephron-forming capacity in the metanephric mesenchyme. Given the fact that Six2 maintains nephron progenitor cells, Sall1 may act together with Six2 to ensure multipotency of nephron progenitors.
Although some of the aforementioned genes affect a number of other genes, they all normally stimulate Gdnf expression. However, it is also important to restrict the area of Gdnf expression, to avoid multiple kidneys arising from a single Wolffian duct. In this regard, a member of the forkhead transcription factor superfamily, Foxc1 , appears to restrict Gdnf expression to the intermediate mesoderm around the Wolffian duct (i.e., metanephric mesenchyme). Foxc1 homozygous null mutants display ectopic ureteric bud outgrowth resulting in duplex kidneys. In this mutant, the restrictive expression pattern of Gdnf as well as Eya1 is perturbed, and is abnormally extended along the Wolffian duct.
Transcription Factors Regulating Ureteric Bud Formation (or Early Kidney Development)
Lim1 is a homeotic gene expressed in both the central nervous system and kidneys. By whole-mount in situ hybridization, its transcript is detected from the pronephric stage to the metanephros. In the metanephros, its expression is detected in the renal vesicles, S-shaped bodies, and ureteric bud branches. Knockout of Lim1 leads to kidney agenesis, suggesting its distinct role in early nephrogenesis. Since Pax2 expression in the mesonephros is detected in Lim1 knockouts, and the ectopic expression of Pax2 was found to induce Lim1 in the intermediate mesoderm, it is likely that Lim1 acts downstream of Pax2 in pronephros development. However, the exact role of Lim1 in metanephrogenesis remains to be determined.
One of the Wilms tumor suppressor genes, Wt1 , a zinc-finger transcription factor, is required for kidney development. Wt1 generally acts as a transcriptional repressor, and has been shown to repress Igf2 , Igf1 receptor, Pax2 , Myc , and Bcl2 expression. Most of these genes are related to cell proliferation, supporting the notion that loss of Wt1 -mediated repression could lead to disregulated proliferation (i.e., cancer). It seems paradoxical that the Wt1 knockout suffers from kidney agenesis, not tumors. In the homozygous deletion mutant of Wt1 , the ureteric bud fails to form, despite relatively normal development of the mesonephros and in the presence of Gdnf expression, suggesting that factor(s) other than GDNF might be required for ureteric bud initiation. In normal embryonic kidneys, Wt1 is expressed in uninduced mesenchyme, renal vesicles, and glomerular podocytes. Wt1 mutant mesenchyme is not responsive to the inductive signal from wild-type spinal cord, while the mutant Wolffian duct can induce wild-type metanephric mesenchyme, suggesting that the primary defect is in the mesenchyme. Interestingly, the mechanism(s) of WT1 remain to be fully elucidated. Genome-wide expression profiling analysis in cells expressing inducible WT1 identified some direct WT1-target genes, including EGF receptor ligands, chemokines, and transcription factors.
Limb Deformity Gene/ Fmn1
Kidney agenesis is observed in mice homozygous for the limb deformity ( ld ) gene, Fmn1 , mutation. The initial outgrowth of the ureteric bud is not observed in mutant mice. The metanephric mesenchyme from Fmn1 mutant mice is induced by embryonic spinal cord, suggesting that the defect is in the ureteric bud. The Fmn gene encodes formin, a gene product, which is present in both the ureteric bud and the metanephric mesenchyme. Although knockouts of certain formin isoforms display limb deformity and kidney defects, specific elimination of isoform 4 results in a pure kidney phenotype.
The Myc family members were first recognized as proto-oncogenes that function as transcription factors. While Myc is expressed in the uninduced metanephric mesenchyme and newly formed epithelium, Nmyc1 is transiently expressed in the area of mesenchymal–epithelial transformation. Another Myc family member, Lmyc1 , is expressed in ureteric bud-derived structures, and its expression increases as these structures mature. Both Myc and Nmyc1 gene knockouts are lethal at E9.5–10.5 and E10.5–12.5, respectively. Myc mutants apparently have no specific kidney phenotype. In Nmyc1 mutants, mesonephric development is affected. In both cases, mice die before metanephric development, which therefore cannot be assessed. The distinct pattern of expression of the various Myc genes makes them useful as markers: mesenchymal stromal cells are negative for Myc ; Nmyc1 is a marker for induced mesenchyme or early mesenchymal epithelialization; Lmyc1 is a marker for the collecting duct.
Transcription Factors Regulating Ureteric Bud Survival/Branching
Disruption of a homeotic gene, Emx2 results in urogenital defects in mice. In mutant mice, the initial formation of the Wolffian duct and ureteric bud is normal, as is the initial induction of the metanephric mesenchyme. Normal Pax2 expression is observed in these structures. However, the ureteric bud starts to degenerate around E12.5. At the same time, the expression of Pax2 and Ret , a receptor tyrosine kinase normally expressed at the tip of growing ureteric bud, is greatly reduced. Recombinant organ culture between wild-type and mutant ureteric bud and metanephric mesenchyme indicates a defect in the ureteric bud. Emx2 expression is observed at a later stage of epithelialization than Pax2 in normal mice. The expression pattern suggests that Emx2 regulates maturation and/or survival of epithelial cells, rather than formation of epithelial cells.
By differential gene expression screening in the epithelial cell three-dimensional culture system for branching tubulogenesis, the mammalian ortholog of Timeless gene was identified as a candidate for regulation of epithelial branching morphogenesis. Its expression is detected in the active region of ureteric bud branching in the developing kidney. Selective inhibition of this gene in various in vitro culture models resulted in inhibition of ureteric bud branching. Deletion of this gene is embryonic lethal prior to the onset of kidney development. The kidney (or ureteric bud)-specific knockout data are needed to provide a definitive role of this molecule in kidney development.
ETS Transcription Factor Genes
Etv4 and Etv5 , two members of the Pea3 family of E-twenty six (ETS) domain transcription factors, which are believed to function as transcriptional activator proteins, were identified in an analysis of gene expression in ureteric buds cultured in the absence or presence of Gdnf. Etv4 and Etv5 were found to have overlapping expression in ureteric bud tips which was positively enhanced by Gdnf. Gene dosage reductions and/or deletions of these transcription factors indicated a role in the formation of the ureteric bud tip domain.
Sox genes are developmental regulators containing a DNA-binding domain with high homology with the HMG box of the sex-determining gene Sry. Mice with double deletions of Sox9 and Sox8 have kidney defects, including renal agenesis. In situ hybridization demonstrates reduced expression of a number of downstream targets of Gdnf-Ret signaling, including Etv4 , Etv5 , Met , and Spry1 in the ureteric bud tips. Together, the data indicate that Sox8 and Sox9 have key roles in Gdnf-Ret signaling regulating/modulating ureteric bud branching morphogenesis.
Transcription Factors Regulating Stroma Development
Study of one of the forkhead box transcription factor superfamily members, Foxd1 ( Bf2 ), shed light on the role of the third cellular component of the developing kidney, the stroma. In developing kidneys, Foxd1 is expressed by the cortical stromal mesenchyme. Homozygous null mutants for Foxd1 die soon after birth due to renal failure. Mutant kidneys are small, fused, and located in the pelvis. Both ureteric bud branching morphogenesis and kidney tubulogenesis in the metanephric mesenchyme-derived segments are perturbed. Further analysis of this knockout mouse reveals that the mutant kidney capsule abnormally contains cells expressing bone morphogenetic protein (BMP) 4 or PECAM (endothelial marker). Abnormal signals from these cells are likely to cause disruption of normal ureteric bud branching and nephrogenesis. One of the target genes for this transcription factor is placental growth factor, a family member of vascular endothelial growth factor.
The Pod1 gene, a transcription factor expressed in mesenchymal cells surrounding the ureteric bud and visceral glomerular podocytes in the developing kidney, also plays a role in regulating ureteric bud branching and nephron formation. Null mutants for this gene exhibit a phenotype similar to that seen in Foxd1 knockouts; initial ureteric budding and mesenchymal condensation occurs, but the process appears to slow down beyond this point. By chimeric mouse analysis, Pod1 expression was shown to be critical for the medullary stroma.
Pbx1 gene encodes a homeodomain containing transcription factor expressed in metanephric mesenchyme, and both cortical and medullary stroma in the developing kidney. The kidney phenotype of Pbx1 knockouts is similar to that of Pod1/Foxd1 knockouts. It appears that expression of these genes is not dependent upon the others, and all three genes are required for the functioning stroma to be capable of supporting ureteric bud branching and nephron differentiation. Identifying the molecular nature of this “stromal effect” will provide considerable insight into kidney development.
Retinoic Acid Receptor Genes
One possible mechanism for control of ureteric bud branching morphogenesis by the stromal cells is through the vitamin A/retinoic acid pathway. Vitamin A deficiency has been known to result in small kidneys. Dietary vitamin A is converted to its active form, retinoic acid, and its signal is mediated through the retinoic acid receptor, which acts as a transcription factor. Retinoic acid synthesizing enzyme localizes to the cortical stromal cells, and double knockout of retinoic acid receptor Rara and Rarb2 results in Ret downregulation and impaired ureteric branching.
Transcription Factors Regulating Functional Maturation of Nephron Tubules
Hepatocyte nuclear factor ( Hnf )- 1 is a homeotic gene mainly expressed in liver and kidney. Hnf1 knockout mice have an enlarged liver and Fanconi syndrome, resulting in urinary wasting of sugars, amino acids, and electrolytes that normally are reabsorbed in the renal proximal tubules, suggesting an important role for Hnf1 in regulating the expression of proximal tubule transporters.
Maturation of Henle’s loop and distal tubule is controlled by Brn1 , a POU transcription factor. Brn1 is expressed only in part of the mesenchymal condensate, then the prospective Henle’s loop and distal tubule in the maturing kidney. There is no expression in the glomerulus, proximal tubule or collecting duct. Knockout of this transcription factor results in an elongation and maturation defect of the Henle’s loop, macula densa, and distal tubule.
Ureteric Bud Outgrowth from the Wolffian Duct
Outgrowth of the ureteric bud from the Wolffian duct in response to signals arising from the metanephric mesenchyme is the initiating event in the development of the mammalian kidney. The major growth factor involved in this process is Glial-derived neurotrophic factor (Gdnf) (see below).There are several levels of regulation surrounding the GDNF pathway.
Restriction of GDNF Expressed Region by Slit-Robo
The secreted protein Slit2 and its receptor Robo2 , previously reported as a chemo-repellant factor for axon guidance, also functions to restrict Gdnf expression. Null mutations of either Slit2 or Robo2 result in supernumerary ureteric buds, caused by an abnormally extended Gdnf expression area.
Regulators for GDNF Signaling Pathway
Sprouty (Spry) is an intercellular protein that acts as a negative feedback regulator for FGF and other receptor tyrosine kinase-mediated signaling. Knockouts of Spry1 display multiple ureters and multiplex kidneys. It appears that mutant Wolffian ducts are abnormally more sensitive to GDNF, as reduction of Gdnf expression rescues the phenotype. Double knockouts of Spry and Gdnf ( Spry −/− ; Gdnf −/− ) rescue normal ureteric bud outgrowth and kidney development. When Fgf10 was also deleted from these mice ureteric bud outgrowth failed to occur, suggesting that Fgf10 is likely to function as the receptor tyrosine kinase responsible for ureteric bud outgrowth in the absence of Spry and Gdnf .
Activin and FGF
One of the puzzles in early nephrogenesis has been that a substantial fraction of Ret knockouts develop very rudimentary kidneys, suggesting the existence of a “bypass” pathway for ureteric bud formation. A GDNF-independent pathway for in vitro budding has been described that appears to involve a FGF, but it is possible that inhibition of activin signaling also enables a “bypass” of the GDNF-ret pathway. The FGF-dependent bypass pathway has been recently supported by in vivo evidence.
NPY and BMP through PKA
Utilizing ex vivo cultured Wolffian ducts, microarray analysis of ducts maintained in the absence or presence of Gdnf identified neuropeptide Y ( Npy ) as a novel modulator of ureteric bud outgrowth. Npy also rescues budding in Bmp4-treated Wolffian ducts, suggesting that this neuropeptide is reciprocally regulated by Gdnf and Bmps. Comparison of budded and non-budded portions of Gdnf-induced cultured Wolffian ducts also reveals an almost 15-fold increase in protein kinase A (PKA) activity in non-budded Wolffian ducts. Microarray analysis reveals a marked decrease in the expression of Ret following activation of the PKA pathway in cultured Wolffian duct. Bmp2 expression is also increased in unbudded Wolffian ducts, and exogenous Bmp2 inhibits ex vivo budding from the Wolffian duct with a three-fold increase in PKA activity. Taken together, the data suggest a role for PKA in regulating the site of ureteric bud outgrowth, potentially via a Bmp-dependent downregulation of Ret/Gfrα1 co-receptor expression.
Another member of this family, growth/differentiation factor ( Gdf )11 is expressed in the Wolffian duct and the metanephric mesenchyme. Knockouts of this gene result in kidney hypoplasia to agenesis, thought to be caused by downregulation of Gdnf in the mesenchyme. In these mice, molecules known to regulate Gdnf expression such as Eya1 , Six2 , Pax2 , and Wt1 are expressed in the metanephric mesenchyme region, and the metanephric mesenchyme from this mutant undergoes nephron tubule formation when it is cultured with embryonic spinal cord. Molecular markers for the Wolffian duct such as Ret , Pax2 , Emx2 , and Lim1 were expressed in the right place, and mutant Wolffian duct responds to exogenous GDNF. Thus, GDF11 is likely to be indispensable for Gdnf expression.
Outside the kidney, this mutant mouse shows deranged anterior/posterior patterning, with alteration of Hox gene expression pattern. Given the fact that Hox11 paralogs control Gdnf expression, downregulation of Gdnf in Gdf11 mutants may be mediated through Hox11 expression. GDF11 acts through the activin receptor (ACVR) IIA and IIB, and knockouts of Acvr2b result in a similar though milder phenotype than that of Gdf11 knockouts.
Unidentified Signal(s) from Metanephric Mesenchyme
Another pathway involved in the regulation of ureteric bud outgrowth is revealed by single deletions of Fgfr2 (but not Fgfr1 , which appeared to be normal) from the metanephric mesenchyme, which leads to the outgrowth of duplicated ureteric buds from the Wolffian duct. Although it is not clear which factors are secreted from metanephric mesenchyme, this result suggests that certain FGF signaling plays a role in the metanephric mesenchyme to suppress ectopic ureteric bud formation.
Ureteric Bud Branching Morphogenesis
As discussed previously, ureteric bud branching morphogenesis is induced by signals from the metanephric mesenchyme. Soluble growth factors, direct cell-to-cell contact, and cell–matrix contact play key roles in the process. The molecules likely to be involved in ureteric bud branching morphogenesis are summarized in Table 25.1 .
|Process||Soluble Factors||Transcription Factors||ECM/Protease/Integrin|
|Slit/Robo||Six1 , 2|
|Sprouty(intracellular molecule)||Hox11 paralogs|
|Wnt11||Sox9||Integrin α 3 β 1|
|Maintenance/maturation of collecting system||EGFR ligands||Emx2|
The embryonic kidney or isolated metanephric mesenchyme can induce branching morphogenesis of MDCK, mIMCD3 or UB cells grown in type I collagen gels in the absence of apparent cell contact. Moreover, isolated ureteric buds from E13 rats have been shown to undergo branching morphogenesis in the presence of soluble factors. This suggests that the metanephric mesenchyme elaborates soluble growth factors capable of inducing branching morphogenesis in the ureteric bud. A number of key growth factors have been identified.
Glial Cell Line-Derived Neurotrophic Factor
The importance of GDNF and its receptors GFRa1 and RET in kidney development is strongly supported by gene knockout data. Kidney agenesis or severe kidney hypogenesis is found in Gdnf , Gfra1 or Ret knockout mice. GDNF is expressed in the prospective metanephric mesenchyme area, while GFRa1 and RET are expressed in the Wolffian duct at the time of ureteric bud induction. As discussed previously, a wide variety of genetic manipulations affecting the transcription of the GDNF/GFRα1/Ret axis lead to disruption of kidney development. Moreover, in vitro application of GDNF-soaked beads to the whole genitourinary tract culture induces ectopic budding of the Wolffian duct. GDNF is also important in subsequent branching morphogenesis of the ureteric bud, as inhibition of this factor perturbs further branching of the isolated ureteric bud in vitro . Gene dosage of Gdnf is important, as heterozygous null mutants of this gene have smaller kidneys. However, unlike ureteric budding from the Wolffian duct, where GDNF appears necessary and sufficient, GDNF is necessary but not sufficient to support ureteric bud branching morphogenesis, at least in the isolated ureteric bud culture system.
Fibroblast Growth Factors and Receptors
Many fibroblast growth factors (FGFs) and their receptors are expressed in the developing kidney. Initial demonstration of the importance of this signaling pathway came from an analysis of transgenic mice that overexpress a soluble chimera of FGFR 2IIIb and human IgG Fc. In these animals, where FGF signaling is broadly inhibited, kidney agenesis or severe hypogenesis ensues. In addition, minor kidney defects are reported in Fgf7 , Fgf10 , and Fgfr2IIIb knockout mice. In the isolated ureteric bud culture system, FGF2 and FGF7 induce a less branched globular ureteric bud growth, while FGF1 and FGF10 support branching growth, suggesting that FGFs may play a role in ureteric bud morphogenesis. Consistent with this, a ureteric bud-specific knockout of Fgfr2 , (but not Fgfr1 ) was found to result in abnormal ureteric bud growth, as well as abnormally thickened cortical stroma. Three-dimensional reconstructive imaging reveals decreases in ureteric bud tip number and increases in the length of the ureteric bud segments.
This heparin-binding growth factor has been implicated in neurite outgrowth and mesenchymal–epithelial interaction during organogenesis. In the context of kidney development, pleiotrophin (isolated from the conditioned medium made by metanephric mesenchyme-derived cells) was found to induce ureteric bud branching morphogenesis in the presence of GDNF in the in vitro culture system. It is present in the developing kidney at the basement membrane of the ureteric bud. Although pleiotrophin may act as a mitogen for the ureteric bud after its budding through GDNF action, knockout of this gene does not appear to have a major affect in kidney development.
Wnts and Related Molecules (Frizzled-Related Proteins)
A member of the WNT family of secreted glycoprotein, WNT2b is expressed in the stroma, and its presence promotes ureteric bud branching in vitro . Null mutants of the gene encoding another member of WNT family, Wnt11 , result in decreased ureteric bud branching with reduced Gdnf expression in the mesenchyme. Ret knockout mice also show reduced expression of Wnt11 . In the trunk or stalk region of the ureteric bud, Wnt9b is expressed, and genetic deletion of this molecule results in cystic kidneys, presumably due to the disruption of the planar cell polarity (non-canonical beta catenin-independent) pathway of Wnt signaling. Later in collecting duct development, collecting duct-specific inactivation of Wnt7b displays a similar phenotype.
Secreted frizzled-related proteins (sFRPs) are secreted proteins that function as WNT modulators. sFRP1 is expressed in the stroma and periureter area; sFRP2 is expressed in early mesenchymal condensates and also in the periureter area in the developing kidney. Exogenous administration of sFRP1 to embryonic kidney organ culture leads to decreased ureteric bud branching and nephron induction. While exogenous sFRP2 alone does not have a major effect, its administration to the organ culture treated with sFRP1 partially reverses the inhibitory effect of sFRP1.
Transforming Growth Factor β Superfamily
Most of the soluble factors discussed thus far facilitate ureteric bud branching and growth. However, unopposed proliferation and branching is not desirable for normal kidney development. Potential candidates for these “negative regulators” include members of the transforming growth factor (TGF) β superfamily. Generally, TGFβ inhibits epithelial cell growth. In organ culture, exogenous TGFβ1 or another member of the TGFβ superfamily, activin, inhibits ureteric bud development and/or disrupts the branching pattern, suggesting their role not only in regulating proliferation, but also in correct patterning. In this regard, it is interesting that TGFβ selectively inhibits HGF-induced mIMCD3 branching events with little effect on tubule formation ; HGF plus TGFβ induces long, straight tubules, whereas HGF alone induces branching tubules. Furthermore, detailed image analysis reveals alteration in the ureteric bud branching pattern in embryonic kidneys treated with TGFβ superfamily members. In isolated ureteric bud culture, administration of TGFβ superfamily members causes growth inhibition, as well as morphological changes similar (although not so striking) to that observed in mIMCD cell culture.
Heterozygous mutation of another family member, bone morphogenetic protein ( Bmp ) 4 reveals loss of ureteric bud elongation, together with ectopic budding from the Wolffian duct. Its expression is detected at the intermediate mesoderm surrounding the Wolffian duct and metanephric mesenchyme surrounding the stalk of the ureteric bud. Taken together, these data suggest that TGFβ superfamily members inhibit branching events, but have somewhat less effect on ureteric bud elongation (and may even facilitate it in the presence of stimulatory growth factors), and play a role in regulating the pattern of ureteric bud branching.
Gremlin is a secreted BMP antagonist expressed initially in the Wolffian duct, followed by induced mesenchyme in metanephros development. Knockout of this gene results in kidney agenesis. The ureteric bud forms but fails to invade the metanephric mesenchyme. Subsequently, the metanephric mesenchyme undergoes apoptosis. Although the kidney phenotype of Gremlin1 knockouts resembles that of Sall1 knockouts, Sall1 expression is unchanged in Gremlin1 knockouts. However, inactivation of one copy of the Bmp4 gene or the complete absence of Bmp7 in gremlin knockout animals rescues ureteric bud invasion and branching. Thus, it possible that gremlin antagonization of BMP activity at ureteric bud tips acts to restrict and guide ureteric bud outgrowth and branching, although the mechanism remains unclear. Furthermore, treatment of embryonic kidneys from Six1 knockout animals with gremlin restores ureteric bud branching morphogenesis, while heterozygous inactivation of Bmp4 ( Bmp4 +/− ) also rescues ureteric bud branching morphogenesis and kidney development in Six1 knockout animals. Taken together, the data indicate that interplay between Bmps, gremlin, and Six1 likely plays a key role in the regulation/modulation of ureteric bud growth and branching.
Hepatocyte Growth Factor
Embryonic kidneys express HGF and its receptor Met. Neutralizing anti-HGF antibodies inhibit the growth of the embryonic kidney and disrupt ureteric bud branching morphogenesis in serum-free organ culture. These results support the notion that HGF is an important morphogen for the ureteric bud that is secreted from metanephric mesenchyme. However, kidney development appears to be unaffected up to embryonic day 14 in Hgf or Met knockout mice, which die around this time from liver failure. Although HGF may act at later stages of ureteric bud branching morphogenesis, it is probably not critically important in the initial stages.
Epidermal Growth Factor Receptor Ligands
Epidermal growth factor (EGF), transforming growth factor α (TGFα), heparin-binding epidermal growth factor-like growth factor (HBEGF), amphiregulin, and betacellulin all bind and activate the EGF receptor. TGFα is present in the embryonic kidney, and disruption of TGFα signaling by neutralizing antibodies results in a small, less well-developed kidney in organ culture. As mentioned above, when mIMCD3 cells grown in collagen gels are co-cultured with embryonic kidney, the cells undergo branching morphogenesis. Unlike MDCK cells, which only undergo branching morphogenesis in the presence of HGF, mIMCD3 cells respond to EGF receptor ligands as well. Similar results are obtained with the UB cell three-dimensional culture system. Moreover, Met (HGF receptor) knockout kidney epithelial cells grown in three-dimensional extracellular matrix (ECM) gels undergo in vitro tubulogenesis in response to EGF or TGFα. A conditional deletion of Met from the ueteric bud results in kidneys with a reduced number of nephrons and increased epidermal growth factor (EGF) receptor expression. Mice which lack both normal Egfr and Me t signaling have decreased ureteric bud branching and small kidneys with a reduced number of glomeruli, suggesting that Met and Egfr can act cooperatively to regulate ureteric bud branching. Although Tgfα knockout mice do not have a kidney phenotype, knockout of the EGF receptor in mice with a certain genetic background leads to dilated collecting ducts and renal dysfunction. These results suggest an important role for EGF receptor ligands in later collecting duct development.
Insulin-Like Growth Factors
In serum-free organ culture, the embryonic kidney produces insulin-like growth factors (IGFs) 1 and 2. When neutralizing antibodies against IGFs are added to the culture, kidney growth is suppressed. The ureteric bud expresses IGF1 receptor. Addition of antisense oligonucleotides against IGF1 receptor to the embryonic kidney in organ culture leads to a small kidney with disrupted ureteric bud branching morphogenesis. However, knockout mice for either Igf1 or Igf2 do not display a kidney phenotype. Molecular redundancy may be part of the explanation; however, as with HGF, the apparent discrepancies between the in vitro and in vivo data need to be addressed experimentally.
Soluble growth factors are not the only molecules involved in ureteric bud branching morphogenesis. Cells of the ureteric bud are surrounded by ECM proteins, and to form branching tubules the cells must digest the ECM. Cells have receptors for ECM proteins, such as integrins, as well as other specific receptors. Integrins can transmit signals to cytosolic and intranuclear proteins in a fashion similar to growth factor receptors. The cell modifies its behavior in response to the combined signals from growth factors and ECM proteins.
The importance of the specific composition of the ECM in kidney epithelial cell branching morphogenesis has been shown using the three-dimensional cell culture model. When MDCK cells are cultured in type I collagen gels in the presence of HGF, the cells undergo branching morphogenesis. When MDCK cells are suspended in growth factor-reduced Matrigel, a basement membrane protein mixture secreted by EHS sarcoma cells, HGF-induced tubulogenesis is inhibited. By mixing individual Matrigel component proteins into type I collagen gels, it was found that collagen I, laminin, fibronectin, and entactin facilitate MDCK cell tubulogenesis, whereas collagen IV, vitronectin, and heparan sulfate proteoglycan inhibit it. However, a mixture of type I collagen and Matrigel, not pure type I collagen, is the optimum ECM for UB cell (a cell line derived from embryonic kidney tissue) tubulogenesis. Interestingly, when these cells are cultured in growth factor-reduced Matrigel alone, UB cells develop into cystic structures ( Figure 25.13 ). Together with the fact that isolated ureteric buds can be cultured in an ECM containing Matrigel but not in pure type 1 collagen gels, this indicates that ECM composition modulates tubulogenesis and branching morphogenesis.