I. INTRODUCTION TO ONTOGENESIS OF THE METANEPHRIC KIDNEY

Top Previous Next References

Most parenchymal epithelial organs follow a fairly simple scheme of embryonic organogenesis. An epithelial sheet or tube that is derived from one of the primordia enters a process of sequential branching (Fig. 1) to generate an arborizing or treelike structure. In the kidney, the epithelial tube is to become the arborizing nephric duct-derived collecting duct system.

View larger version (100K): [in this window] [in a new window]   Fig. 1. Histology of early metanephrogenic organization. Section through human kidney (~20.5 mm embryo) showing structures derived from Wolffian duct and metanephrogenic blastema in outermost zone of cortex. Peripheral branch of ureteric tree extends distally into an ampulla (see sect. III). Metanephric blastema has been induced (Fig. 2) to enter nephrogenic pathway (see sect. IV), and nephron anlage has completed mesenchyme-to-epithelium transition (see sect. V). [Modified from Horster (112).]

The particularly complex situation in renal embryonic development is, however, that it has not one but two distinct embryological origins. Most of the nephron, i.e., the structures beginning with the glomerulus and ending at the junction of connecting tubule and collecting tubule, are derived from another primordial tissue, the mesenchymal blastema. These two entirely different tissues interact so that the ureter-derived collecting duct is induced to branch while the mesenchymal blastema is induced to enter the critical process of mesenchyme-to-epithelium conversion or transition (MET). This earliest mesenchyme-derived epithelium follows a structurally well-defined morphogenetic pathway to generate most of the nephron.

Historically, the basic experimental model for work in nephrogenesis had been set up by Grobstein (89-91) some 45 years ago. These pioneering studies, at the National Institutes of Health, demonstrated in an organ system in vitro that 1) kidney rudiments when removed at embryonic day 11 (E11) (mouse) follow an almost normal developmental program in culture, 2) the isolated ureteric bud cannot develop without contact to the metanephric mesenchyme, and 3) the isolated metanephrogenic mesenchyme can be induced to go through the MET by a number of tissues, including the embryonic spinal cord and the ureteric bud. These and later tissue recombination experiments (164, 236, 242, 280) have set the stage for the application of today's molecular biology techniques to nephrogenesis (121, 148, 290).

The two developmental pathways for the two different tissues, the nephrogenic (mesenchymal) and the ductogenic (ureteric), are regulated by transcription factors and protooncogenes, polypeptide growth factors acting as signaling molecules, and their receptors. They are modulated by cell adhesion molecule (CAM) complexes and their associations with the cytoskeleton, by extracellular matrix (ECM) glycoproteins and ECM receptor molecules such as the integrin family, and by ECM degrading proteases. Protooncogenes regulate growth particularly in embryonic organogenesis, and they have the potential to gain tumorigenesis after gene mutations. Some of the protooncogenes that encode for receptor tyrosine kinases are involved in mesenchymal (nephrogenic)-epithelial (ductogenic) interactions, in which the protooncogene encoded tyrosine or serine/threonine kinase is the ureteric receptor (see sect. III) for signaling molecules secreted by the other primordial tissue, the metanephrogenic mesenchyme (see sect. IV).

The entire process is governed by changing gene expression patterns of transcription factors (see sect. VIII) such as Pax-2, of secreted signaling factors such as wnt-4, and of protooncogenes such as c-ret. Through the transfer of techniques from Drosophila to mouse and to human, some of the genes critical in mammalian development have been identified. These genes are part of complex programs that induce and control sequential morphogenetic and differentiation events, i.e., the stages of kidney ontogenesis.

To analyze the programs and their downstream effects during embryonic nephrogenesis, a wide spectrum of techniques is increasingly applied at the individual cell level. Single determinants of renal morphogenesis and of epithelial differentiation, uncovered through these techniques, have been presented in a number of excellent in-depth reviews, specifically on MET (16), conversion of mesenchyme to epithelium (66), growth factors (93), renal stem cells (100), gene targeting (147), signaling molecules (201), the transcription factor WT-1 (215), and basement membrane molecules (270).

This review intends to integrate data from these different areas of research. It centers on the progress accomplished over the past 10 years of research in embryonic nephrogenesis, and it intends to provide a comprehensive view of the many diverse aspects of embryonic renal epithelia. Among these, three processes in kidney organogenesis are emphasized, namely, 1) acquisition of functional properties in the collecting duct system, 2) MET, and 3) epithelial cell differentiation from an apolar to an apicobasal polarized phenotype.

	II. PRINCIPLES IN NEPHROGENESIS

Top Previous Next References

The development of the metanephric (permanent) mammalian kidney begins at gestational week 4-5 in humans and at E11 in mouse. Organogenesis and its governing principles have been studied mostly in the mouse. Metanephros formation, i.e., organogenesis of the permanent kidney (242), is initiated by the ureteric bud, which sprouts out of the posterior end of the Wolffian duct and invades the metanephrogenic mesenchyme (Fig. 2). The subsequent interaction between the two primordia induces the ureteric bud to branch dichotomously, thus initiating the morphogenesis of the collecting duct system (242). Induced metanephric mesenchyme condenses at the tips of the ureteric buds (Fig. 1), and mesenchymal cells form aggregates (Fig. 3), thus beginning the MET. Each aggregate epithelializes (156) and proceeds in stages to the vesicle stage, comma stage, and S-stage, from where each S-shaped body, after fusion with the ureteric bud-derived collecting duct (Fig. 3F), differentiates into one of the (2 × 10⁶) nephrons of the human kidneys. The architectural pattern, therefore, as a result of the sequential ureteric bud arborization, is designed to proceed from the deep cortex to the periphery in a repeat series of induction, morphogenesis, and differentiation (Fig. 2).

View larger version (36K): [in this window] [in a new window]   Fig. 2. Overview of principal events in early nephrogenesis. Ureteric bud, an offspring of Wolffian duct, invades mesenchymal blastema (left) and initiates reciprocal signaling (middle) between epithelial (ductal) and mesenchymal (metanephrogenic) cell types. Receptor tyrosine kinases are expressed almost exclusively in ureteric bud cell, whereas ligands are secreted by adjacent mesenchymal cells. Ligand for c-ros encoded receptor is not yet known. Ligand-receptor signaling activates successive stages shown in Fig. 3, AF. Mesenchymal-epithelial interactions, in addition to signaling growth factors, involve extracellular matrix (ECM), cell adhesion molecules (CAM), transcription factors, and protooncogene-encoded receptor tyrosine kinases (RTK). Mesenchymal blastema expresses stem cells of several cell lineages; stromogenic and nephrogenic ones are shown. Epithelial S-shaped body connects to ureteric bud-derived collecting duct. HGF, hepatocyte growth factor; GDNF, glial cell-derived neurotrophic factor. [Modified from Horster et al. (113).]

View larger version (14K): [in this window] [in a new window]   Fig. 3. Morphogenic stages in early metanephrogenesis. Nephrogenic pathway is initiated by inductive signaling between Wolffian duct-derived ureteric bud and adjacent mesenchymal blastema (Fig. 2). A: induced condensing mesenchymal cells adhere and begin cell remodeling to epithelial phenotype. B: globular aggregates close to tip of ureteric bud express basement membrane proteins and cell adhesion molecules. C: renal vesicle. Lumen of sphere suggests secreted fluid and solutes. D: comma-shaped body representing reorganized sphere and (E) S-shaped body suggest expression of an as yet unknown patterning program. F: junction of upper domain of mesenchymal blastema-derived S-shape body with most peripheral branch of Wolffian duct-derived ureteric bud. Scheme reduces some of morphological features, e.g., not shown is strictly dichotomous branching of ureteric tree, as visible in Fig. 1. [Modified from Horster et al. (113).]

The epithelial segments of the nephron, unlike the ureteric bud-derived collecting duct system, are created from mesenchymal cells by an intricate cascade of events. The early events (Fig. 2) result in the acquisition of an essentially epithelial character by the future nephron cells while these polarized cells form a sphere or vesicle (Fig. 3). The process of modeling the subsequent stages of comma and S-shape (Fig. 3) is not understood, although plenty of morphoregulatory molecules (see sect. VIIC) and transcription factors (see sect. VIIIA) are sequentially and differentially expressed. These stages of morphogenesis are the onset of nephron differentiation, i.e., epithelial segments begin to express their specific properties (112).

The mechanisms directing the segmentation of the nephron have not been identified. Some of the molecules involved in segmental morphogenesis are characteristically regulated in distinct segments, e.g., some members of the integrin family are expressed in the late S-stage (₂ distal; ₃ proximal), whereas others are upregulated only in the blastema (₁) or in the vesicle stage (₆) (143).

These stages of nephrogenesis have an ancestry that begins at the blastula stage, which determines the mesoderm; it follows the induction of the pronephros and the directed migration of the pronephric duct to proceed through the stage of the Wolffian duct and to induce the metanephric mesenchyme, which in turn directs branching of the ureteric tree. Cells of the metanephrogenic mesenchyme are induced by ureteric bud cells to become stem cells after rescue from apoptosis (see sect. IVB); they go on to condense and, guided by regulatory circuits of gene expression and repression (see sect. VIIIA), to enter the MET, and to polarize to apicobasal expression patterns (see sect. VI).

The metanephric kidney is derived from two different early embyronic tissue primordia: the nephric duct and the nephrogenic cord (Fig. 4). The nephric duct becomes the mesonephric duct and continues through the Wolffian duct stage to the ureteric bud. The nephrogenic cord, after inductive signaling with the pronephric duct-derived cells, develops into the nephroi of mesonephros and metanephros. The rudimentary pronephros, the transitory mesonephros, and the permanent metanephros form in sequence during mammalian renal ontogeny (Fig. 4), recapitulating the phylogeny of the excretory system.

View larger version (18K): [in this window] [in a new window]   Fig. 4. Embryonic precursors of metanephros. Rudimentary pronephros, transiently functioning mesonephros, and permanent metanephros are sequentially induced and formed, thus recapitulating phylogeny of excretory system. This embryonic continuity also pertains to some transcription factors and signal molecules. [Modified from Horster (112).]

Pronephric tubules are formed around E8 (mouse) from the intermediate mesoderm. In contrast, the primary nephric duct appears to arise from a distinctly different cell population within the intermediate mesoderm (198). The caudal nephric duct (frog) extends along the cranial-caudal axis not by mitotic apposition but through migration of cells from the foremost tip toward their later caudal organ destination where they reassemble into the duct (34). Pronephric (zebrafish) and metanephric (mouse) development, however, are believed to be governed by almost entirely different genetic programs, although the zebrafish Pax-2 and WT-1 homologs appear to be expressed during pronephros differentiation in a pattern similar to that in mouse metanephros (54).

The nephron of the mesonephros is functioning (173, 266, 298); it is composed of glomerulus and two segments (proximal/distal) merging in a collecting tubule that empties into the mesonephric or Wolffian duct. Wolffian duct cells appear to induce mesonephric progenitor cells to become differentiation competent (100); this interaction, in the murine mesonephros, occurs around E10.

As the Wolffian duct grows close to a mesenchymal cell population, it is induced to sprout (E10E11) and to form the ureteric bud, whereupon mesenchymal cells at the bud tip gather and appear to form a cap (Fig. 4). Next, the invading bud is directed to branch dichotomously to form a T-shape (Fig. 1), and this branching mode is maintained sequentially to express the arborizing structure of the collecting duct in the cortex, which establishes the general structure of the kidney.

C.  Cell Types of the Metanephros Are Derived From Different Lineages

1.  Nephric duct-derived ureteric bud

The embryonic kidney has served for almost 50 years (89) as a model system to study inductive interactions between mesenchyme and epithelium. The fact that the kidney is derived from two distinctly different primordia that can be grown to develop in culture has aided the experimental access to signal (growth factor) molecules and their receptors. The very first step in kidney development is the sprouting of the ureteric bud out off the Wolffian duct, followed by signaling from the ureteric bud to the mesenchyme to induce the mesenchyme to become epithelium.

The ductal system after initial sprouting develops by sequential dichotomous branching, and the induction of this repeat process of ureteric bud bifurcation in the kidney appeared to be through interaction with the nephrogenic mesenchyme (90). The nature of some of the inductive signals received by the ureteric bud cell from the mesenchyme has been disclosed, although the cues for the sequence of events remain elusive.

Branching morphogenesis of the collecting duct system, however, requires not only interactions between the embedding mesenchyme and the epithelial duct, but also between basement membrane (ECM) components and the epithelial cells.

The mechanisms of ductal growth and of ductal branching (10, 23, 231, 235) as well as differentiation by expression of plasma membrane transport proteins in the ureteric bud cell (113, 119) have finally come to be investigated.

For the metanephric mesenchymal blastema to produce the ~15 epithelial cell types of the metanephric kidney, it must be induced to undergo a conversion to the epithelial phenotype and subsequently differentiate into the highly specialized cell types of the nephron. Hypothetically, this pathway could start from two different points. One starting point would be a homogeneous mesenchymal population consisting of one multipotent cell type from which all nephron epithelial cell types are derived. Alternatively, the primary inductive event is not the conversion to the epithelial phenotype but a commitment of the mesenchymal cell type to different developmental pathways, and the secondary inductive event of phenotypic conversion then destines already committed cells to be recruited for the early nephron (101, 144, 213).

Studies on the temporospatial expression of two transcription factors, BF-2 (97) and Pax-2 (49, 50), have shed some light on this situation. It seems now justified to favor the hypothesis that all peripheral mesenchymal blastema cell types are induced to become stem cells through the first signal interactions. This initial step (see sect. IVA) rescues most of the nephrogenic stem cells now expressing Pax-2 from apoptosis (6, 145), whereas the uninduced mesenchymal cells enter programmed cell death (see sect. IVB).

Induction is a two-step event (6) that had been postulated already from earlier tissue recombination work (243), where it was found that a short-time (hours) exposure of uninduced mesenchyme to the ureteric inductor led to the stem cell phenotype but no further. Nevertheless, this first step to the stem cell phenotype rescues most of the mesenchyme from apoptosis. The second step, however, very likely differs in molecular nature from the first one (6). Two hypotheses, at present, are similarly supported by data, although not yet by complete lineage analysis. In one, the primary inductive interactions between mesenchyme and bud are believed to determine the distinct and final developmental pathways of both stromal and nephrogenic lineage (66). In the other, a bivalent stem cell progenitor population that gathers next to the outermost ureteric bud cells (Fig. 1) is available throughout nephrogenesis, and it may either take the nephrogenic (Pax-2) or the stromogenic (BF-2) pathway (8, 97, 242). It is interesting to note that the endothelial progenitor cell, the angioblast, may derive also from a bipotential (mesodermal) stem cell precursor (218). The further fate of the nephrogenic lineage is also determined by members of the superfamily of signaling peptides, as discussed in section VIIA.

Cell lineage analysis based on classic embryologic work (92, 100) clearly indicates that the definitive kidney is derived from two independent tissue compartments of the intermediate mesoderm, namely, the metanephrogenic mesenchyme and the Wolffian duct. This traditional view has been broadened by a set of data derived from embryonic kidney organ culture (213); when uninduced mesenchyme was isolated and tagged so that cells could be followed to their final destination, and then cocultured with isolated ureteric bud, mesenchymal cells were found to be inserted into the collecting duct, although the majority of collecting duct cells were derived from the ureteric bud.

Organogenesis of the kidney has long become a model system that represents principles in morphogenesis and cell differentiation. The continuous process of morphogenesis is guided by cascades of interactions between two different cell populations (Fig. 2). Regulation involves diverse families of genes and their products, including protooncogene-encoded receptors (see sect. VIIIB) and their polypeptide ligands (see sect. VIIA), transcription factors and their target genes (see sect. VIIIA), and regulating ECM proteins and CAM-mediated signals. All of these diverse systems interact to initiate and guide embryonic renal morphogenesis and cell differentiation. Unraveling the complexity of these interactions, which is described in the following sections, is a major challenge for many outstanding laboratories.

	III. URETERIC BUD AND BRANCHING MORPHOGENESIS

Top Previous Next References

Grobstein, in a classic series of experiments (89-91), had addressed the question of whether or not mesenchyme and epithelium derived from different embryonic organs can induce and maintain ductal morphogenesis and concluded that one common mechanism was not likely. However, there may be families of morphogenic molecules, either signaling via the ECM, as transforming growth factor- (TGF-) (230), or mesenchyme derived, as epimorphin (106), or ligands of protooncogene-encoded receptor tyrosine kinases, such as hepatocyte grwoth factor (HGF) (234) that participate in a general pattern governing branching morphogenesis. Several lines of evidence suggest that branching morphogenesis may be viewed within an integrated model system in which cell-matrix receptors (e.g., integrins) and basement membrane components (e.g., laminin-1) on one side and locally secreted signaling molecules [e.g., members of the bone morphogenetic protein (Bmp) and Wnt families] on the other organize budding and branching in different epithelial organs (112, 130, 232). Although specific modes of interactions between ECM and epithelial cells are expressed during organ morphogenesis, the general pattern of signal-directed remodeling of the matrix may apply to branching tubes so diverse as in kidney (232), lung (111), and salivary glands (130).

B.  Genes That Control the Ureteric Bud

1.  WT-1

Several genes are presently believed to be regulated by WT-1. Among those, in addition to Pax-2 (225), is insulin-like growth factor (IGF)-II and its receptor (53, 296) and TGF- (47), which are involved in branching morphogenesis, and platelet-derived growth factor (PDGF)-A (292). In the mesonephros, WT-1 is clearly expressed in the mesenchyme, the vesicle, and glomerular structures. In the metanephros, WT-1 is expressed in the uninduced mesenchyme and increasingly in the induced mesenchyme. Importantly, WT-1 is highly expressed in the proximal limb of the S-shaped body (121a), specifically in the future podocytes up to the mature glomerulus (41, 205).

As a member of the family of growth factor receptors characterized by an extraordinarily large extracellular binding domain, c-ret is a protooncogene required for ureteric bud branching and proliferation. The Ret receptor is first expressed in the Wolffian duct (E8E11.5), and as ureteric arborization proceeds (E13.5E17.5), c-ret is expressed only in ureteric bud (Fig. 5) tip cells (200). It is thus not surprising that in homozygous mutant mice (RETk) the ureteric bud does not outgrow from the Wolffian duct (247) while the mesenchyme from RETk mice maintains branching and growth of a wild-type ureter in vitro. Moreover, mesenchymal differentiation appears normal when induced with spinal cord, whereas the ureteric bud from RETk mice did not interact with wild-type mesenchyme (247). This impressive work indicated that 1) the Ret receptor response to an inductive signal is independent of the cell type, and 2) the Ret ligand glial cell-derived neurotrophic factor (GDNF) activates ureteric cell proliferation and branching very early (Fig. 5), beginning with the first visible Wolffian duct outgrowth.

View larger version (35K): [in this window] [in a new window]   Fig. 5. Temporospatial expression of signaling systems in nephrogenic and ureteric bud morphogenic pathways. Receptor tyrosine kinases (Ret, Met, Ros) are encoded by protooncogenes. Growth factor signaling molecules are expressed and secreted as indicated. Relative abundance of expression is specified by bold or normal type. Induced precondensing mesenchyme is shown on top; other stages of metanephric nephrogenesis correspond to those depicted highly schematically in Fig. 3, A, C, and E. PDGF, platelet-derived growth factor.

Limb deformity (ld) encodes a group of phosphoproteins, the formins, generated by alternative splicing of mRNA that are all expressed in the kidney. The ld mutation in the mouse results in defects of limb formation and urogenital development that, interestingly, are expressed with differing degrees of anomaly (167). The five known alleles of ld and the formin isoforms are expressed in the ureteric bud (Fig. 5) and in the mesenchyme. The primary defect in ld / mutant is a failure of the ureteric bud to induce the mesenchyme because outgrowth of the ureteric bud is either arrested or delayed or incomplete (167). Because formins are believed to operate in intracellular protein-protein interactions (27, 279), they might be part of a signaling pathway downstream of the ligand-receptor interaction in the ureteric bud tip cell. The epithelial cell at the tip of the invading ureteric bud (Fig. 1) faces the interspace between the two signaling tissues. When this cell (Fig. 11) was analyzed by the combined techniques of electronmicroscopy, electrophysiology, and molecular biology (119, 121), it was found that it differs functionally from the cells in deeper parts of the branching ureteric tree, and it undergoes the epithelial polarization and differentiation process in situ. An alternative hypothesis (100), however, postulates that the bud tip cell exhibits a mesenchymal phenotype that is capable to delaminate from the ureteric bud and to be incorporated into all segments of the mesenchyme-derived nephron. Ureteric bud cell differentiation, i.e., how this cell acquires the polar organization of apical and basolateral plasma membrane characteristic for the collecting duct cell, is discussed in section VI.

C.  Branching Morphogenesis

1.  Multiple controls regulate the arborizing duct system

The embryonic development of several organs, such as lung, mammary gland, pancreas, tooth, salivary glands, and kidney, depends decisively on branching morphogenesis. In all of these organs, a small epithelial rudiment is initially surrounded by mesenchymal cells. In the metanephric kidney (Fig. 2), a sequence of different events is initiated after reciprocal induction that leads to the formation of dichotomous duct branches. Two processes result in the arborizing collecting duct system starting from the first bifurcation of the ureteric bud, namely, the longitudinal growth of duct epithelia and the branching process. There is now accumulating evidence that both of these are regulated separately (40). In fact, a recent model for renal branching morphogenesis (229-231) proposes that a local ratio, at the branching point, of branch-promoting to branch-inhibiting factors might account for the architecture of the collecting duct system, and it was demonstrated for the first time that growth (ductogenesis) and branching are regulated by separate mechanism. Whereas HFG and TGF- are branching morphogens, TGF- is inhibitory to branching but not to ductogenesis. Nevertheless, HGF/scatter factor (SF) and its c-met receptor are the primary signaling system for branching and for ductal growth in nephrogenesis. Moreover, this inducing action on ductal morphogenesis is paralleled by HGF effects on its mesenchyme-based Met receptor, and antibodies to HGF/SF perturb branching morphogenesis and the early phase of MET (301).

Ureteric bud cells express high-affinity receptors for growth factors (see sect. VIIIB) and among them are the receptor tyrosine kinases Met, Ret, and Ros (Fig. 5). The polypeptide ligand of Ret, GDNF, is a branch-promoting growth factor, since GDNF / mice suffer from delayed or absent ureteric branching (186, 208, 232). Expression of GDNF (99) is high in wild-type condensing mesenchymal cells (E11.5) and downregulated after MET (Fig. 5) to maintain the arborizing morphology. Importantly, expression is maintained high in the outer cortical mesenchymal cell population, which might enable GDNF to organize the treelike patterning possibly by radial concentration gradients of GDNF along extracellular matrix components. Glial cell-derived neurotrophic factor binds to the receptor GDNFR- (127, 273), and the receptor-ligand complex binds to Ret (276, 285). Indeed, GDNF and Ret are a specific ligand-receptor entity, since GDNF when added to wild-type kidney cultures increased the number of (aberrant) ureteric branches (285). Ureteric growth and branching are also dependent on the intact Ret receptor, which is part of the GDNF signaling system (200, 247). Among the regulatory genes (Fig. 5) expressed in the ureteric bud cell, in addition to Pax-2, is ld, and its mutation inhibits ureteric growth (275). The intracellular signal systems, however, that link the ligand-dependent RTK activation to the ureteric bud effector systems for growth and branching remain undefined.

Extracellular matrix molecules are secreted and localized in complex temporospatial patterns (66, 141), and they have signaling roles particularly in early nephrogenesis, followed by regulation of basement membrane components and their receptors during epithelial polarization. When the peptide growth factors IGF-I or IFG-II (220) or nerve growth factor (NGF) (237) are blocked selectively by molecular or immunotechniques, mesenchyme and ureteric bud development is perturbed. On the other hand, renal morphogenesis is overall normal in mice with homozygous null mutations of the IGF-II (42) and the NGF (155) receptors. These apparently contradictory observations are not unusual in polypeptide signaling and suggest functional redundancy.

Laminin-1, a basement membrane constitutent, acts during cell differentiation in a dual way, namely, as a signaling molecule and as a structural component. The extracellular glycoprotein is a member of a large family present in basement membranes, and it has several biologically active domains including the proteolytic fragments E3 and E8 (72); importantly, the laminin ₁-chain is expressed at the onset of epithelial polarization (65). Antisera against the E8 and E3 domains inhibited the MET in embryonic renal organ culture (141). The role of laminin-1 in early branching morphogenesis was further demonstrated by monoclonal antibodies against the E3 fragment in organ culture of embryonic salivary gland, which inhibited branching morphogenesis, suggesting a role for the laminin E3 fragment in structuring basement membranes (130). Dystroglycan is the high-affinity receptor for laminin-1 and laminin-2, and -dystroglycan, the receptor for the E3 fragment, probably links the basement membrane to the cytoskeleton (70), and an antibody against -dystroglycan inhibits the embryonic epithelial polarization process (60). In addition to -dystroglycan, a second receptor specific for the E8 fragment of laminin-1, the ₆₁-integrin, is expressed on nephrogenic cells during MET (255). The ₆-integrin receptor, therefore, might participate in signal transmission during cell differentiation, and the ₆-integrin subunit associates with the ₁-subunit during the early nephrogenic process (65). The structural role of the laminins is within the two major complex networks of laminin and collagen IV that are probably linked by nidogen that binds to type IV collagen and to a domain in the laminin ₁-chain (174). The crucial role of nidogen as a link protein has been proven by perturbation experiments (67), demonstrating that interfering with the link between nidogen and laminin-1 inhibits ureteric bud branching morphogenesis. It is of interest in this context that an unexpected heterogeneity at the histochemical level (142) was demonstrated for the embryonic collecting duct, and a monoclonal antibody of the IgG-1 subclass was shown elegantly to react with an epitope at the basolateral side of the ampullar collecting duct epithelium (259). The roles of antibody and epitope in nephrogenesis remain to be elucidated.

Local proteolysis remodels the extracellular matrix during branching. Matrix proteolysis can be attributed to secreted matrix metalloproteinases (157) and to plasma membrane-bound proteases. Matrix metalloproteases and their regulation thus have a complementary role to the branch-promoting (157) and branch-inhibiting growth factors. This function implies that inhibitors of proteases block branching morphogenesis. Proteases can be secreted by epithelial or by adjacent stromal cells. Although almost all details of the design for collecting duct architecture are unknown, it might be permitted to speculate that stromally secreted proteases, after binding to epithelial receptors, could guide directed proteolysis along tissue gradients of proteases.

	IV. METANEPHRIC MESENCHYME AND NEPHROGENIC PATHWAY

Top Previous Next References

The classic studies of kidney organogenesis had established (89-91) that a signal, although its nature and mechanism could not be identified, must initiate outgrowth and branching of the ureteric bud, and another signal originating from the ureteric bud must induce MET. Today, some of the signals essential for interactive early nephrogenesis have been identified (and many more have been shown to exist), but the meaning of their temporospatial expression patterns and their precise downstream roles still remain poorly defined.

The discovery of BF-2 as an embryonic stromal gene (97) in the mesenchyme, i.e., in the intermediate mesoderm, has rekindled the debate of whether or not two cell populations are present in the uninduced mesenchyme whose final destination is either stromal or epithelial. This view requires that the uninduced mesenchyme had already received a signal, presumably from the early invading outgrowth of the Wolffian duct which is essential for the continuity of the nephrogenic process, since uninduced mesenchyme goes apoptotic (145). Alternatively, the decision would be made through molecules of the inductive interactions within a homogeneous mesenchymal cell type. A case for the latter was made by the observation (213) that induced ureteric bud cells can delaminate and incorporate themselves into the mesenchyme to become part of the epithelia-generating cell population.

The family of Pax genes encodes transcription factors expressed in various embryonic tissues, including the kidney (50). Two Pax genes are expressed in renal organogenesis, Pax-2 and Pax-8. Pax-2 is central for renal development, since it appears to specify all epithelial phenotypes derived from the intermediate mesoderm. It is first expressed in the nephric duct and in mesonephric tubules, extending caudally to the ureteric bud (49, 50). Pax-2 is expressed only in mesenchyme that has been induced by signaling from the ureteric bud, but not in the uninduced mesenchyme (207), and Pax-2 becomes the first indicator of the nephrogenic cell lineage at this early stage (Fig. 5). Pax-2 continues to be expressed throughout condensation and polarized vesicle formation (Fig. 15), and expression is downregulated in the proximal loop of the S-shaped body that is the site of the podocyte precursor cells. This repression of Pax-2 transcription is related to the expression of WT-1 (215) and its interaction with the first exon of the Pax-2 gene (225).

The regulatory loop between WT-1 and Pax-2 (Fig. 15) may provide clues for further analysis of those renal diseases in which epithelial cell proliferation continues. The Pax-2 gene is a member of a family that also includes genes involved in severe abnormalities as Waardenburg syndrome (264) and human aniridia (271). Pax-2 is essential for the MET (223), since interference with gene function inhibits mesenchymal condensation and the subsequent steps in nephrogenesis. The complete loss of Pax-2 function, by homologous recombination (272), resulted in abolished formation of nearly all of the epithelial components derived from the intermediate mesoderm. The central role of Pax-2 was further illustrated in a transgenic mouse model (52) where the deregulated gene produced severe abnormalities such as cystic kidney and undifferentiated glomerular epithelia, reminiscent of features in congenital nephrotic syndrome. In conclusion, the nephrogenic stem cell is characterized by a new pattern of gene expression, such as Pax-2, both at the level of transcription and of signaling molecules (Fig. 5).

Apoptosis or programmed cell death is a crucial regulated event in renal morphogenesis, since a large number of (blastemal) cells are produced and only a few are guided to follow a developmental program because they have been rescued from programmed cell death by induction. Cell death occurs by two distinct mechanisms. In one, necrosis, cellular ATP concentration declines immediately (as in renal ischemia) followed by a sequence of events, such as cellular swelling, leading to cell lysis. In the other, apoptosis, the primary event is a regulated breakdown of DNA into small, 200-bp fragments by the activation of a calcium-sensitive endonuclease. The cell constituents condense and ultimately break into fragments that are taken up mostly by macrophages. Mesenchymal cells are destined for apoptosis, as shown in transfilter culture of rat E13 isolated renal tissue (145, 295), unless they are induced to survive by signaling interactions with the ureteric bud. Apoptosis is a normal event in development, and apoptotic cells are frequently observed next to condensing aggregates and to vesicles (32, 116). Apoptosis is more prominent in mesenchymal cells of the nephrogenic outer cortex where cell death may serve to remove those mesenchymal cells that have not been chosen for MET. This implies that inductive signaling may be a two-step event (6) in which rescue from apoptosis is followed by conversion to the epithelial phenotype. Although the rescue or survival signal remains to be disclosed, a hint to a putative pathway may have come from the observation that LiCl (15 mM), added to the medium of isolated mouse (E11) mesenchyme in the absence of an inducing signal, was able to rescue these cells from entering programmed cell death; morphogenesis after the LiCl rescue was initiated up to the expression of the cell adhesion molecule N-CAM in the comma stage, but no further (39). In addition, some growth factors, such as epidermal growth factor (EGF) (295) or basic fibroblast growth factor (FGF) (206) are able to rescue cultured mesenchyme from apoptosis.

The tumor suppressor gene p53 is believed to not only act as controller in the cell cycle but also in apoptosis (211). However, wild-type p53 expression in comma- and S-shaped body does not provide a clue to its function (245), and a p53 loss-of-function mutation does not dramatically alter organogenesis (48), whereas a gain-of-function mutation induces ureteric alterations and small kidneys with a reduced number of nephrons and, notably, accelerated apoptosis (82). Of interest in this context are the findings that WT-1 binds the p53 protein, thus inhibiting p53-mediated apoptosis (169, 170), and p53 expression can be repressed by Pax-2 (260).

A protooncogene with death repressor activity, bcl-2, is expressed in nephrogenesis (197, 253). The bcl-2-encoded protein is on the outer mitochondrial membrane, in the endoplasmic reticulum and the nuclear envelope, and it is involved probably in an antioxidant pathway (286) to protect cells from oxidative damage (110). The bcl-2 gene is highly expressed in condensing mesenchyme and in the ureteric bud, and it is downregulated, as in most mature cells, in terminal epithelia and in glomerular cells (154, 253); bcl-2-deficient mice develop hypoplastic polycystic kidneys (286). In addition, they have abnormal postinductive nephrogenesis either after E13 (193) or weeks after birth (253, 286), associated with fulminant mesenchymal apoptosis (194).

WT-1 binds to p53, and WT-1-deficient mice show increased mesenchymal apoptosis (148). Also, when c-myc is constitutively expressed in cell lines, the ensuing apoptosis can be inhibited by coexpression of bcl-2 (17). Clearly, bcl-2-regulated apoptosis is a basic mechanism in nephrogenesis, and bcl-2 is an intrinsic negative regulator of the cell death pathway even beyond the inductive phase of morphogenesis.

Mesenchymal-to-epithelial transition of the metanephrogenic mesenchyme is the center event in early nephrogenesis. Mesenchymal cells are nonpolarized, loosely associated, and apolar cells embedded in ECM with a fibroblast-like shape and high mobility. Epithelial cells, in contrast, are asymmetric or polarized, form continuous sheets, express a basement membrane, have a cuboidal or "cobblestone" shape, and are generally little mobile. The phenotypic conversion of metanephrogenic mesenchymal to nephron epithelial cells (Fig. 6) mirrors changes in the expression of different gene families (121a), encoding transmembrane receptors, cell adhesion molecules, growth factors, ECM components, and specific basement membrane constituents. As discussed, one or more genes of the Pax family are expressed in each of the tissues that undergoes MET, and pro-, meso-, and metanephros are all formed through these interactions. The nature of the inductive signals that were postulated in the classic embryologic recombination experiments (86, 89) is now being uncovered. The inductive process is multifactorial and multiphasic. Although much effort has been invested in studying single factors (e.g., Refs. 37, 39, 102, 206, 295), these single factors appear to address and express only parts of the developmental program, i.e., their effects either have not induced tubulogenesis (39, 206, 295), or the inducing factor has activated another messenger system that in fact induced tubulogenesis (102, 133).

View larger version (25K): [in this window] [in a new window]   Fig. 6. Functional complexes in transition of mesenchymal to epithelial cells (MET). Nephrogenic mesenchymal cells (top) are induced to enter MET whereby several systems with signaling functions are activated. CAM-mediated signals and ECM-mediated signals interact with secreted growth factors to express epithelial phenotype (bottom).

The early stages of kidney organogenesis are governed by four different control genes (Fig. 5). These are WT-1 (148), Pax-2 (49, 51), Wnt-4 (258), and Bmp7 (55, 165). Their primary roles, as deduced from knockout experiments, suggest that they are local mediators in signaling interactions between the ureteric bud and the nephrogenic blastema.

WT-1 is expressed in the metanephrogenic mesenchyme but not in the ureteric bud cells (4), and the expression of WT-1 (191) is the earliest sign of commitment in the intermediate mesoderm-derived metanephric blastema that contains the nephron lineage(s). WT-1 is essential for mesenchymal competence to later respond to other inductive signals (148); it directs the genesis of the first ureteric bud off the Wolffian duct, and it interacts with Pax-2 in the nephrogenic stem cells at the transcriptional level (Fig. 15).

Another of the control genes in early nephrogenesis is Bmp-7, which encodes the bone morphogenetic protein-7 (55, 165). Bmp proteins belong to the TGF- family of secreted signaling molecules generally involved in morphogenesis (261). Bmp-7 loss-of-function kidneys differentiate up to the comma- and S-shaped stage, but further epithelial development is defective (56), and the mesenchymal cells undergo rapid apoptosis (165). Consequently, genes of early induction and transition stages are expressed in Bmp-7-deficient mutants, albeit mostly in aberrant patterns, such as Pax-2, Wnt-4, WT-1, Ret, and even Pax-8. Bmp-7 transcripts are seen in the mesonephric kidney, in the metanephric condensates, in comma and S-stage, and in the collecting duct (166). In line with the expression pattern of Bmp-7 and the mutant phenotype is the finding that Bmp-7 induces cultured metanephric mesenchyme to differentiate, and antibodies or antisense oligonucleotides inhibit Bmp-7 and tubulogenesis (288).

The homeobox gene lim1 is expressed in the intermediate mesoderm, mesonephric tubules, Wolffian duct, and induced mesenchymal cells (9, 75). Lim1 / mice have a complete defect of kidneys and gonads (249) probably due to premetanephric defects, suggesting that lim1 has a primary role in intermediate mesoderm specification. The gene that is expressed in the progeny of the Wolffian duct appears to have an important role after the initial signaling events (183). Metanephric mesenchyme and ureteric bud of Emx2 / mutants correctly expressed transcribed marker molecules of the the first signaling stage; the ureteric bud invaded the metanephric mesenchyme and induced Pax-2 expression. After this initial phase, however, signaling was discontinued, and the expression of Lim-1, c-ret, and Pax-2 in ureteric bud, as well as that of GDNF in the mesenchyme, was greatly reduced; the ureteric tip did not dilate and branch, Wnt-4 was not expressed, and MET was not initiated. The control experiment showed that wild-type ureteric bud or spinal cord was capable to induce the mesenchyme of the mutants, suggesting that Emx2 is required in early ureteric bud cells subsequent to the induction of mesenchymal Pax-2 expression.

Condensed mesenchymal cells, in addition to their shift from mesenchymal to epithelial surface proteins (Fig. 7), express a protein encoded by Wnt-4, a member of a family of genes that encode signaling molecules regulating early embryonic tissues (258). The Wnt-4 protein is secreted by induced metanephrogenic mesenchyme (138), and the Wnt-4 gene is expressed in condensates very soon after induction by tip ureteric bud cells, and expression persists in the vesicle, comma, and S-stages (Fig. 5). In mutant Wnt-4 / mice, mesenchymal cells condense next to the ureteric buds, Pax-2 is normally expressed (258), but Pax-8 that is normally expressed immediately after the condensation stage (Fig. 15) was not expressed. The failure of nephrogenesis to proceed beyond the condensation stage suggests that the cascade of signaling molecules directing the induction process was interrupted at the Wnt-4 plateau. The Wnt family (77) indeed has some characteristics of the putative mesenchymal inducer molecule(s), e.g., they are postulated to associate with secreted matrix or with basement membrane domains. Another Wnt family member, Wnt-1, was demonstrated to have properties comparable to Wnt-4 (102). In an elegant and important study, Wnt-1-secreting NIH 3T3 fibroblasts (transfected with Wnt-1 cDNA) were cocultured with isolated nephrogenic mesenchyme and shown to induce tubule formation after an initially increased proliferation of induced mesenchymal cells. It is of interest that Wnt-4 has a temporal expression pattern very similar to Bmp-7 (166, 288) and that sonic hedgehog (Shh) genes are coexpressed with Bmp genes (18), but nothing is known about interactions between these families of morphogens.

View larger version (46K): [in this window] [in a new window]   Fig. 7. MET-associated shift in gene expression. Relative abundance (symbol size) of gene expression or protein production of some CAM and ECM proteins, and their shift in process of mesenchymal (top left) to epithelial (top right) cell transition (MET), is shown. Reference numbers are given in parentheses.

To summarize, Wnt-4, in contrast to Wnt-11, which is expressed exclusively at the very tip of the ureteric bud (Fig. 5), appears to be essential, possibly as an autoactivator, for the transitional step from the induced mesenchyme to epithelia and most likely in subsequent early stages of tubulogenesis, but it is not required for the initial step(s) leading to the induced nephrogenic mesenchymal cells. This emerging role of Wnt as a regulator of E-cadherin-mediated cell-cell adhesion (105) denotes a further very important point along the multistep induction process.

The inductive process between nephrogenic mesenchyme and ureteric bud ultimately results in aggregates of induced cells that now express adhesion molecules (Fig. 6). Wnt-4 is expressed in these aggregates, and it is repressed upon completion of the nephron. In Wnt-4 / mice, condensations of induced mesenchymal cells appear, but few only express markers of epithelialization or Pax-8 (258), indicating that Wnt-4 is an essential regulator of MET, probably by controlling the expression of cell adhesion molecules (105, 281, 287). Control mechanisms in the subsequent stage of cell polarization are unknown, although some putative control genes, such as members of the Hox family and WT-1, are upregulated during the transition process. As the mesenchymal-to-epithelial program continues, tenascin is expressed in polarizing mesenchymal cells (5) together with SGP-2 (94).

The MET is completed when cells display the basic features of an epithelium, e.g., the asymmetry of apicobasal membrane proteins. The condensed and converted cells form a round body which, by as yet undefined mechanisms of fluid secretion, develops a small lumen and proceeds to the vesicle stage (Fig. 3C). The subsequent stage, some hours later, is reached by reorganization of the sphere into a longitudinally curved epithelial body, the "comma shape" (Fig. 3D). The mechanisms responsible for this rearrangement and for the intricate pattern formations that follow are unknown. The comma stage develops to the "S-shaped body" by a cleft involution at both curving ends (Fig. 3E). It is very likely by now that the S-stage expresses a proximal-distal profile of cell properties (121a) characteristic of the evolving proximal and distal nephron segments. Nevertheless, in situ hybridization and immunohistochemical work have already demonstrated that some transcription factors are repressed (e.g., WT-1, Pax-2, N-myc) while others are expressed (e.g., Pax-8; LFB-1) during this and the subsequent stages of nephron formation (4, 50, 223, 188, 153, 210). Their state of molecular differentiation is only now beginning to be explored (Fig. 12). The most proximal end of the S-shape enters angiogenesis (218), whereas adjacent tubular parts express a few apical membrane markers (7). In the most distal part of the S-shape, membrane fusion ultimately must connect the mesenchyme-derived nephron segments with those originating from the epithelial ureteric bud (Fig. 3F). Other cells of the distal tail of the S-body are to express the cell types of the distal convoluted tubule and of the macula densa. The importance of apoptosis at this stage, regulated by bcl-2 (110, 154), is clearly demonstrated by the fact that bcl-2 / mice, in addition to high apoptotic rates, have polycystic kidneys (286). Further development into nephron segments is accompanied by basement membrane scaffolding (68) and by signaling of a wide variety of growth factors (93). Collecting duct growth, in addition, is directed by IGF/R-mediated actions, since antisense oligonucleotides against the receptor inhibit growth of the collecting duct in organ culture (291). The repression of Pax-2 at the S-shape stage is a prerequisite for further cell differentiation; it is as essential here as is the expression of Pax-2 in previous stages for completion of MET (Fig. 15). Finally, the junction of the mesenchyme-derived segments of the nephron with those derived from the Wolffian duct has not been investigated, and it involves most likely mechanisms of plasma membrane reorganization similar to those supposed to take place in branching morphogenesis (229-231).

	V. MESENCHYME-TO-EPITHELIUM TRANSITION AND CELL ADHESION

Top Previous Next References

The first step of induction rescues part of the mesenchyme from apoptosis and leaves these mesenchymal stem cell populations committed for further induction to enter either the nephrogenic or the stromogenic pathway (see sect. IIC). Condensing cells transiently express the surface glycoprotein syndecan (281, 282), and they begin to express a new set of regulatory genes and of morphogenic molecules, whereas typical mesenchymal markers are repressed. Mesenchymal vimentin and N-CAM disappear, and epithelial E-cadherin, the -chain of laminin, and the ₆₁-integrins that act as transmembrane receptors for laminin A (66, 65, 139, 255) appear, to mention only some of them (Fig. 7). In an attempt to identify additional surface molecules expressed in the transitional process, monoclonal antibodies were raised against induced mesenchymal cells (83); their ability to inhibit tubule formation in culture and their expression patterns should promote further search for the corresponding antigens involved in epitheliogenesis. In conclusion, the conversion to the epithelial phenotype involves several levels of cellular, protein, and genetic changes (121a). Most of the putative interactions between these levels have not yet been defined.

Cell adhesion appears to be regulated through an interplay of several molecular species in which the cadherin family of adhesion molecules (171), the catenin family of intracellular adhesive junction proteins, and the product of the protooncogenes Wnt-1 and src participate (21, 105). Cadherins and catenins are fundamental molecules in embryonic epitheliogenesis, and changes in cadherin expression patterns designate events in morphogenesis, such as MET. The cadherins E-, N-, and P-, the classic cadherins, are able to form complexes with specific catenins, i.e., with -, -, and -catenin (plakoglobin) (21, 199). Cadherins are linked to the actin filament network by catenins, and cadherin-catenin complexes may interact with other cytoplasmic or transmembrane proteins (115, 214). Cadherins thus acquire a critical role in early epithelial polarization (152). The - and -catenins colocalize with E-cadherin at the zonula adherens (195), and interactions between the zonula adherens complex, as for -catenin/E-cadherin, and the actin filaments may participate in morphogenetic changes such as from the comma to S-shape. E-cadherin expression increases early at the contact site between two cells, ~1 h after contact has been made (176, 177). The next stage is characterized by interactions between E-cadherin and cytoskeletal proteins that result in the polar distribution of membrane proteins (176). Truncated N-cadherin lacking the extracellular domain and expressed in Xenopus embryos results in disruption of cell adhesion and abnormal development (137). Furthermore, epithelial cells deprived of -catenin are unable to adhere, although they express -catenin and E-cadherin; cells that lack cadherins but express - and -catenins can be induced to express an epithelial adhesive phenotype by the introduction of E-cadherin or N-cadherin, indicating that different cadherins can interact with one catenin subtype (107). E-cadherin is essential early in embryogenesis, since mice lacking this gene expression are unable to form a trophoectoderm epithelium and die from this preimplantation defect (217). The final phenotypic step in the polarization process, some 1.5 days after the first inductive step (mouse), is the gain of expression of cytokeratin components of the intermediate filament system (Fig. 7) and the loss of vimentin.

When mesenchymal cells aggregate next to the tip cells of ureteric buds to form packed condensates, this is considered the first visible morphological indication of transformation. Syndecan-1 (14) and N-CAM (196) are upregulated at the onset of this stage, as is Wnt-4, and they may therefore be involved in organizing the mesenchymal condensate. N-cell adhesion molecule coexpresses with the low-affinity receptor (p75NGFR) of NGF (297) and with NGF (237); this fact suggests an early role for N-CAM and NGF in the predetermined nephrogenic stem cell (140) even before epithelial polarization is expressed. N-cell adhesion molecule probably is a target gene for regulation by Pax and Hox gene products (64, 128); it may turn out that these homeobox and paired-box gene products are an important functional link between patterning genes and morphoregulatory CAM genes. A-cell adhesion molecule has a similar pattern except that its expression persists in the lower limb of the S-shaped body during later stages, whereas L-CAM is expressed in ureteric bud, collecting duct, and the limb of the S-shaped body that is to become the distal nephron where L-CAM continues to be expressed in the neonatal kidney (196). These segment-specific patterns for L-CAM and for A-CAM as early as in the S-stage may contribute to the expression of characteristic nephron segmental properties. In conclusion, upregulation of both syndecan (283) and N-CAM (140) denotes the onset of epithelial morphogenesis in the mesenchyme. However, it should be mentioned that neither the mechanism of aggregate formation within the condensations nor the regulation of cell polarization is currently understood. Answers to these open questions may come from studies on the "reverse" process in which epithelial cells revert to mesenchymal cells, a phenomenon also seen in carcinogenesis (16).

Syndecan-1 associates with the actin-containing cytoskeleton via its intracellular domain (14), and it is expressed highly in condensing (induced) mesenchymal cells only, i.e., it appears to be involved in the ligand-induced clustering of the committed mesenchymal cells (281). The proteoglycan is later reexpressed in bud-derived epithelia, while nephrogenic mesenchymal cells lose the expression of syndecan-1 with differentiation. Tenascin-C, a mesenchymal ECM glycoprotein, is transiently expressed close to the condensing (induced) stem cells and later next to the early epithelia (mouse E13), suggesting that its expression is upregulated by cortical epithelial signals, whereas in newborn kidney, tenascin C expression declines in the cortex and persists in medullary mesenchymal cells (5). Epimorphin, a 150-kDa protein, is expressed with syndecan (130) on the surface of condensing mesenchymal cells, but antibody interference with the protein had no effect on morphogenesis.

	VI. EPITHELIAL CELL POLARIZATION

Top Previous Next References

A.  Methods to Study Embryonic Renal Epithelia

1.  The classic system to study tissue interactions

The embryonic kidney, when taken in organ culture (mouse E11.5), continues with its tubulogenic developmental program (68, 86, 90, 92, 243, 280). When mesenchymal blastema and ureteric bud are separated before the inductive interactions, the ureteric bud can be induced to branching morphogenesis only by the renal mesenchyme, whereas the mesenchyme can be induced not only by the ureteric bud but also by other embryonic tissues (see sect. II). The fact that the induced metanephrogenic mesenchyme continues its differentiation in vitro to a certain stage of tubulogenesis provides an indispensable system to study early nephrogenesis because 1) each embryonic stage can be identified by its characteristic morphogenetic event, e.g., the comma shape (Fig. 13C), 2) immunolocalization and in situ hybridization (60) can be applied directly to the embryonic epithelia to resolve temporospatial expression patterns (227) of defined molecules, and 3) the in vitro situation provides direct access for mRNA analysis in single embryonic epithelial cells (121).

Functions of genes and their successive expression in the cascade of kidney development are being analyzed using either transgenic mice or mutants generated by loss of function/gain of function, i.e., by gene targeting (147). Mutation or modification of a gene at its chromosomal site ("knock-out gene") can be achieved by gene targeting in embryonic stem cells in culture; subsequently, these genetically manipulated cells are reimplantanted into the mouse, and the mutation can then be investigated in the whole (embryonic) mouse. In this way, mutations with a defective kidney phenotype have been created (Table 1), and their renal phenotype can be compared with the temporospatial expression pattern of the wild-type gene. Another approach is to identify gene families and their role in development by their homology of motifs with Drosophila genes, since many of the Drosophila melanogaster genes known as developmental control genes are conserved with evolution, e.g., in mice and humans. This pertains to the sequence elements of paired box (Pax) and homeobox (Hox) that are available to search for gene family members in other species. Ultimately, however, to understand the developmental (downstream) role of a gene, it is mandatory to define its function in the whole embryo under the conditions of constructed misregulation or of absent regulation of the gene under study. By the loss-of-function (targeted disruption) strategy, WT-1 was shown to prepare the mesenchyme for the inductive process (148), Pax-2 (272) to control multiple steps in nephrogenesis downstream of WT-1, Wnt-4 (258) to participate in the transformation from aggregate to epithelial tubule, Bmp-7 (55, 165) to be involved in stem cell conservation, c-ret (247) to regulate collecting duct branching and growth, and BF-2 (97) to advance the differentiation of aggregates.

3. Cell culture of ureteric bud and of induced metanephrogenic mesenchyme

A) MONOLAYER CELL CULTURE OF NEPHRIC DUCT-DERIVED EPITHELIA. The introduction of techniques for primary culture of single nephron segments (112) opened the in vitro access not only to most cell types of the nephron (114), but moreover to the apical plasma membrane (Fig. 8) for ion channel evaluation (119). In addition, this culture system (114) was modified to evaluate ion channel expression in the particular cell population covering the ureteric bud (Fig. 9) by applying the reverse transcription (RT)-PCR (Fig. 10) to monolayers and to the single cell (Fig. 11). These cells at the tip of the ureteric bud express specific properties, such as ClC-2 mRNA consistent with the notion that this channel is widely expressed in embryonic cells (see sect. VIC). In a newborn kidney culture system (259), significantly, a novel antigen, termed PCDAmp1, was found by immunohistochemical techniques on the basal aspect of ampullar collecting duct only in embryonic but not in differentiated collecting duct cells derived from cortical explants of newborn rabbit cortex (259). This important finding may point to a role of the antigen in nephrogenesis.

View larger version (146K): [in this window] [in a new window]   Fig. 8. Morphology of ureteric bud cells in culture. Scanning electron microscopy of apical cell side in monolayer (A) suggests that these cells express a principal cell-like phenotype, as indicated by central cilium (B).

View larger version (45K): [in this window] [in a new window]   Fig. 9. Scheme of single ureteric bud cell analysis. Ureteric duct with ureteric buds is microdissected and transferred on collagen matrix in culture dish. Patch-clamp recording is followed by mRNA harvesting into patch pipette. mRNA encoding housekeeper and ion channels are reverse transcribed and amplied by PCR, similar to protocol in Fig. 10.

View larger version (24K): [in this window] [in a new window]   Fig. 10. Scheme of quantitative RT-PCR in analysis of embryonic ion channel expression. Quantitative reverse transcriptase-PCR was applied to monolayer cultures of ureteric bud cells (Fig. 8) at different stages of embryonic and postembryonic cell differentiation (Fig. 14).

View larger version (60K): [in this window] [in a new window]   Fig. 11. Analysis of single ureteric bud cell. Single cells were patch-clamped, and cytoplasm for reverse transcriptase (RT)-PCR was harvested to measure ion channel conductance and mRNA expression. A: light microscopic view of ureteric bud. Inset: entire ureteric tree with mesenchyme; one bud had been dissected free of mesenchyme to provide access with patch pipette (arrow) to basal cell side. B: acrylamide gel showing products of ClC-2 chloride channel (bottom) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) housekeeper gene (top), amplified by single cell RT-PCR. Data are from 8 cytoplasmic samples and 6 controls. C: patch-clamp characterization of single cells at tip (A) of ureteric bud shows original whole cell traces and derived slowly activating inward current resembling ClC-2 chloride channel characteristics. [Modified from Huber et al. (121).]

View larger version (66K): [in this window] [in a new window]   Fig. 12. Microculture of metanephrogenic unit. To study MET, a nephrogenic unit as defined by a single ureteric bud with induced adherent mesenchyme is transferred in collagens. Schematic view illustrates in vitro MET and early nephrogenesis (Fig. 3, AF). These processes are documented by electron microscopy and by molecular analysis. Noninduced mesenchymal cells enter apoptosis, and ureteric bud cells proliferate and migrate to form a monolayer. [Data from Huber et al. (121a).]

Ureteric bud cells in culture were analyzed by electrophysiological (patch clamp) and by molecular biological techniques (RT-PCR) during embryonic and early postnatal development (116-121).

Embryonic ureteric bud cells (E17) express an apparently constitutively active, outwardly rectifying, whole cell chloride conductance (120), whereas perinatal and early postnatal ureteric bud/cortical collecting duct (CCD) cells acquire the expression of a mature type, hypotonic swelling-activated chloride conductance between E17 and postnatal day 1 (P1) before the onset of vectorial transport, i.e., before the fusion of the nephric duct-derived epithelia with the mesenchyme-derived nephron (Fig. 13). During this period, the embryonic-type chloride conductance, interestingly, is downregulated.

View larger version (39K): [in this window] [in a new window]   Fig. 13. Epithelial cell differentiation in branching morphogenesis: electrophysiological analysis of ion channel expression. Ureteric bud cell functional differentiation (E17 to P14) is characterized by patterns of conductance expression distinct for each channel type. Cell schemes (bottom) summarize channel types predominantly expressed or acquired in each of three stages of epithelial cell differentiation. CCD, cortical collecting duct.

Vectorial transport systems in nephrogenesis should be functional at the onset of glomerular filtration in any of the newly formed nephron generations to prevent loss of sodium, glucose, amino acids, and water into the urine. The ontogenic acquisition of one of the transport systems, therefore, was investigated by comparing the abundance of mRNA encoding the -subunit of the epithelial sodium channel (-ENaC) in embryonic and perinatal ureteric bud and in postnatal CCD cultures (117). This expression of -ENaC mRNA was compared with the appearance of low-conductance, sodium-selective channels in the apical plasma membrane in these ureteric bud and CCD cells (117). The -ENaC mRNA could be quantitated in embryonic ureteric bud (E15E17), and the abundance increased by a factor of two in postnatal ureteric bud (P1P6) and of five in postnatal CCD (P7P28), when compared with the embryonic stage (Fig. 14). This pattern indicates that the principal change in -ENaC mRNA expression occurs toward the end of morphogenesis. Expression of the functional channel protein, as evaluated by single-channel recording in apical patches of ureteric bud and CCD cells, showed the same single-channel characteristics in both stages, albeit at very different densities, suggesting the same channel type in ureteric bud and CCD. This was confirmed by the differing response of whole cell currents to amiloride (117). In conclusion, both -ENaC and the functional channel protein are expressed before glomerular filtrate for vectorial transport reaches the apical cell membrane.

View larger version (61K): [in this window] [in a new window]   Fig. 14. Epithelial cell differentiation in branching morphogenesis: molecular analysis of ion channel expression. Ureteric bud cell functional differentiation (E15 to P14) is characterized by distinct patterns of ion channel expression. Relative abundance of mRNA expression (bottom) in monolayer cultures (middle) derived from kidney at different stages of ureteric bud morphogenesis (top) is represented by width of lines. CFTR, cystic fibrosis transmembrane conductance regulator; ENaC, epithelial sodium channel.

C.  Membrane Transporters Acquire Their Apicobasal Patterns

1.  Ion channels and transporters

Ion channel expression in differentiating epithelia of the ureteric bud was studied by RT-PCR (Fig. 11) in cultured ureteric bud cell monolayers (Fig. 14) and by patch-clamp techniques to measure specific conductances of whole cell and apical plasma membrane as well as to identify apical ion channels (Fig. 13). Cortical collecting duct cells in culture express the mature principal cell phenotype, whereas ureteric bud cells maintain the embryonic phenotype, i.e., they conserve a nonpolar apicobasal distribution of ion conductances (119). Epithelial cell ontogeny, from ureteric bud to CCD, is characterized by a polar differentiation of plasma membrane ionic conductances. This implies the acquisition of amiloride-sensitive sodium channels, low-conductance potassium channels, and 9-pS nonselective cation channels in the apical plasma membrane (119).

The expression patterns of chloride channel type mRNA were evaluated during branching morphogenesis in cultured ureteric bud and CCD epithelia (118). The temporal pattern of ClC-2 expression suggested a specific embryonic function of this channel in epithelia undergoing branching morphogenesis, since it is downregulated after gestation, similar to its regulation in embryonic lung (192). ClC-2 and cystic fibrosis transmembrane conductance regulator (CFTR) mRNA have embryonic temporal expression patterns that change in the opposite direction with cell differentiation. While ClC-2 is downregulated, nucleotide-sensitive chloride current (CFTR) is expressed early, as demonstrated also in human fetal kidney (45), and it is increasingly expressed during the period of ClC-2 downregulation (Fig. 15). Similar to CFTR, the mRNA of nucleotide-sensitive chloride current (ICln) and of the CFTR truncated splice variant (TRN-CFTR) are increasingly expressed with embryonic cell differentiation (118, 116).

View larger version (45K): [in this window] [in a new window]   Fig. 15. Patterns of expression and repression of critical developmental genes. Genes that code for transcription factors and for signaling molecules interact positively (expression) or negatively (repression) or behave autoregulatory. Stages of metanephric morphogenesis (Fig. 3) require profound changes in gene expression, for cell condensation and adhesion, MET (Fig. 6), epithelial cell apicobasal polarization, nephron segmental pattern formation, and acquisition of membrane transport molecules (Fig. 13). Regulation of most expression events and downstream gene targets remains elusive. References to gene expression patterns are in text. IGF, insulin-like growth factor.

The developmental acquisition of vectorial transport-involved apical and basolateral potassium channels of members of the ATP-dependent Kir subfamily showed distinctly different patterns for potassium channel types (21a). ROMK1 and ROMK3 mRNA were not detected in E15 to P6 developmental stages. ROMK2 mRNA was apparent in postnatal ureteric bud cells and increased at P28. The basolateral channel KAB-2 mRNA, similarly, was not expressed in embryonic cells. Importantly, however, the expression of Kir6.1 and of SUR was high in embryonic cells and downregulated thereafter (21a). Significantly, later stages of ion channel development in collecting duct cells are characterized by an increase of sodium and potassium channel density in the apical plasma membrane at constant mature expression pattern (238, 239). Epithelial cell polarity in renal cells is acquired with cell differentiation (119). This has been demonstrated not only for apical membrane ion channel types (Fig. 13), but also for the isoforms NHE-3 and NHE-1 of the sodium/hydrogen antiporter in neonatal proximal tubule cells, where membrane location of these transporters did not show the mature distribution pattern (11), and for the Na⁺-K⁺-ATPase that is expressed in embryonic ureteric bud cells in basolateral and in apical plasma membrane (180), and lastly for an unusual appearance of immunoreactive surface markers (142). In contrast, the temporospatial expression of the Na⁺-K⁺-Cl symporter (NKCC2) in embryonic mouse kidney (E14.5) showed the terminal mature localization in the distal loop of Henle already at this stage (125).

Expression of most known members of the aquaporin (AQP) family was evaluated between E16 and P28 in rat kidney (303), and AQP1 and AQP2 were localized in human fetal kidney (46). Both AQP1 and AQP2 were expressed early in fetal kidneys, and AQP2, significantly, was localized to the apical membrane of fetal CCD principal cells from the onset of expression. As for sodium and potassium channels in later nephrogenesis (238, 239), water transport in later (perinatal) stages of CCD ontogeny appears to increase by a density change of water channel molecules in the plasma membrane (250).

Embryonic sodium-coupled glucose transport is characterized not only by a lower density of transporter molecules of the mature-type transporter in the apical proximal tubule membrane (12). Moreover, detailed analysis of sodium glucose transporter (SGLT)1 and SGLT2 mRNA revealed that expression of both molecular isoforms begins on E18 and E17, respectively, and a change in size of the SGLT2 mRNA suggested an embryonic splice variant of the transporter (306).

In conclusion, cell polarization requires a series of sequential expressions of basement membrane constituents and ECM molecules and the interactions of both with polypeptide signaling factors (304). It is not clear 1) how these processes direct epithelial cells to express a polar distribution of membrane transport proteins and 2) which developmental programs (54) direct the specific downstream patterns of membrane transporter mRNA expression.

	VII. GROWTH FACTORS AND EXTRACELLULAR MATRIX

Top Previous Next References

Growth factors, in addition to their mitotic (growth) action, not only mediate motogenic (migration) and morphogenic inductive signals, but also those for cell differentiation (polarization), proliferation, and apoptosis. The roles of the many growth factor families and their receptors, each of which consists of a multigene family, in regulating developmental programs through all of its stages appear to be overwhelmingly complex. The small polypeptides, synthesized and secreted by embryonic renal epithelia, activate in a paracrine/autocrine mode their renal target cells by way of basolateral plasma membrane recptors (84). They have been implicated in kidney organogenesis in vitro primarily because of their expression patterns in mice (93). The best-characterized growth factors pertinent to renal organogenesis are IGF-I and IGF-II, HGF, TGF-, and FGF. Responses of an epithelial renal embryonic cell can differ widely, from proliferation (123) to apoptosis to cell polarization as a consequence of the differing intracellular signal pathways activated by the growth factor ligand-receptor interactions. An additional problem has been posed by the observation that metanephric uninduced mesenchyme when cultured in defined medium selectively supplemented with the growth factors TGF-, EGF, TGF-, IGF-I, IGF-II, PDGF, FGF, or with retinoic acid failed to respond to any of them with tubulogenesis (295). Interactions are effective not only between individual soluble growth factors but also between them and ECM components. For example, when nephrogenic mesenchyme in culture was incubated with EGF (and pituitary extract) on different matrices (laminin, fibronectin, collagen type I or IV), the inductive process was initiated, but morphogenesis failed to be expressed (205). Although none of the growth factors per se may act as an inducer of nephrogenesis, the addition of antisense oligonucleotides or antibodies to individual growth factors in vitro often have revealed inhibitory effects on organogenesis, suggesting that growth factors in nephrogenesis may act synergistically. Growth factor families shown to be involved in morphogenesis and differentiation of the embryonic kidney are, among others, IGF-I/IGF-II, HGF/SF, TGF-, TGF-/EGF, PDGF-A/-B, Bmp, NGF, and neurotrophin-3. Some of these and their receptor families (219-222) have been localized to specific early structures, and their functional roles are discussed.

B.  Growth Factor Families Are Expressed in Temporospatial Patterns

1.  IGF-I and IGF-II

Early embryonic nephron epithelia express IGF-I, IGF-II, and IGF-binding proteins (109, 159, 160, 172, 220, 291). The IGF-I receptors are present in nephrogenic mesenchyme and in the ureteric bud cell plasma membranes (172) where immunoreactivity is particularly high. Antisense oligonucleotides against the IGF-I receptor of embryonic kidney, which mediate most of the actions of IGF-I and IGF-II, have shown that this signaling system is probably involved in branching morphogenesis of the collecting duct (291). Expression of IGF-II mRNA appears in nephrogenic blastema, is upregulated in the condensation stage (109), and decreases during further morphogenesis. Both IGF-I and IGF-II produced by the embryonic kidney act through extracellular IGF-binding proteins, and anti-IGF-I or anti-IGF-II antibodies, or anti-IGF-II receptor antibodies, completely inhibit the in vitro development of the embryonic kidney (220). Both IGF act differentially, since IGF-II, as shown by in situ hybridization (109), is localized in human fetal kidney to mesenchymal cells but not to epithelial cells of the cortex. Moreover, when antisense oligodeoxynucleotides to IGF-I receptor cDNA were added to mouse organ cultures, growth of induced condensations was inhibited, and ureteric bud branching was perturbed (162). Despite this clear evidence from in vitro studies in which antibodies against IGF-I, IGF-II, and the IGF-II receptor inhibited ureteric bud branching and mesenchymal differentiation (220), gene deletion experiments of IGF-I and the IGF receptor indicate that these genes are not essential for renal morphogenesis, but they may be important for overall organ size (160).

Hepatocyte growth factor is produced in mesenchymal cells (209) close to the epithelial ureteric bud target cells from day 11.5 onward (mouse), i.e., from the time of induction to the stage of metanephrogenic condensation (301). Met, its receptor (252), is located on the branching ureteric cells at the very tip (Fig. 5) where the other receptor tyrosine kinases, products of c-ros and c-ret, are also inserted. Signal transduction of HGF to its ureteric bud receptor may be paracrine in one of the mesenchymal cell populations and juxracrine in another (301). It is clear by now that HGF not only stimulates branching morphogenesis, as does TGF-, but it also is a potent inducer capable of morphogenic, motogenic (scatter), and mitogenic effects on nephrogenesis (16, 23, 151, 294). The morphogenic activity of HGF, first described in a cell culture assay system on a three-dimensional matrix (185), may also pertain to the MET (234, 277), since Met is coexpressed with HGF in at least a part of the mesenchymal tubulogenic cell population, where it may act either in the rescue from apoptosis or else have a part in the conversion process. When HGF is added to metanephrogenic mesenchymal cells in vitro, epithelial differentiation is stimulated (134), and anti-HGF serum in vitro inhibited ureteric bud development as well as nephrogenesis (234), consistent with the spatial receptor-ligand expression pattern of HGF. These apparently consistent observations, however, were not confirmed when the HGF gene was deleted by homologous recombination in mice (246, 278). The cultured mutant kidneys and the in vivo kidneys did not reveal any defect until E14, i.e., after tubulogenesis had begun. In addition, mice with a c-met / mutation do not reveal kidney defects (19). The scatter activity of HGF/SF puts this factor in the context of epithelium-to mesenchymal transition (EMT), since it dissociates cohesive epithelia that then may transform into spindle-shaped (mesenchymal-like) cells with increased migration/motility activity (95), suggesting that the "growth factor" induces or mediates those processes that MET and EMT have in common. Antibody perturbation of HGF in organ culture affects mesenchyme-to-epithelium conversion, inhibits branching of the ureteric tree, and induces excessive mesenchymal apoptosis (234, 301); however, as mentioned, kidneys of HGF / mice appear to develop normally (246, 278), suggesting redundancy of actions between HGF and other systems such as retinoic acid and formins might be involved (167, 178).

Transforming growth factor-1, TGF-2, and TGF-3 mRNA have distinct expression patterns (31) consistent with their differential functions in organogenesis (179). Specifically, TGF-1 mRNA is expressed in the nephrogenic mesenchyme, whereas later epithelial stages have low levels of TGF-2 mRNA and no TGF-3 expression (244). In branching morphogenesis of the ureteric bud/ collecting duct system (222, 230), the inhibitory effect of TGF- acts in a cross-talk with the activating role of HGF and TGF-. Indeed, the early expression pattern of TGF-1 and TGF-3 in embryonic (204) and of TGF-2 in neonatal basement membranes (269) suggests that this growth factor has a primary role in ECM assembly perhaps by differential regulation of collagen constituents.

Transforming growth factor- is synthesized and secreted in the embryonic kidney (221), where it binds to the EGF receptor (13), and it may also interact with laminin (83). The TGF- peptide and receptors for TGF- are present in the fetal human metanephros (85). Branching morphogenesis and subsequent tubulogenesis are inhibited by anti-TGF- antibodies added to the cultured embryonic organ (221), suggesting a morphogenetic role for TGF- particularly in early ureteric bud development (229). Epidermal growth factor/TGF- also appears to participate in the rescue of uninduced mesenchymal cells from apoptosis (32), and in vitro studies have shown that a mouse kidney cell line cultured on ECM forms aggregates and tubulelike structures when TGF- (or EGF) is supplemented (265). Although the role of TGF- (EGF) in mesenchymal differentiation remains largely undefined, its defined impact on branching morphogenesis has assigned this growth factor an important task in kidney organogenesis.

Platelet-derived growth factor-A and -B are produced, among other cell types of the kidney, in developing glomerular epithelial cells, in inner medullary collecting duct cells, and in mesangial cells. The two receptors for the isoforms of PDGF, the PDGFR- (73) and PDGFR-, are expressed in a variety of renal cells, such as in nephrogenic mesenchyme, interstitial cells, and vascular cells (2). Platelet-derived growth factor-B is expressed first in the metanephrogenic mesenchyme, but its later expression is limited to mesangial cells. Capillary vessels of the embryonic glomerulus are produced very early in the cleft of the comma-shaped body, but their cell lineage has not been resolved (218). Subsequent glomerular network organization depends on the expression of PDGF-B and PDGFR-, since mutant mice that have no mesangial cells (158, 254) fail to form an appropriate capillary tuft. Gene targeting of PDGF-B and of PDGFR- has further disclosed their critical role for mesangial cells, since PDGF knock-out embryos completely lack mesangial cells, but not endothelial cells; consequently, a glomerular capillary tuft is not formed, implying that the physiological function of mesangial cells (and of PDGF-B/R-) is the folding of the glomerular basement membrane and the formation of capillary branching into loops. Both processes are likely dependent on the mesangial production of a highly specialized ECM (1).

Of the many Bmp genes that are generally involved in embryonic patterning and mesoderm organization, Bmp-7 is expressed in the kidney (18, 55, 98, 165, 166, 219). Analysis of the homozygous mutant phenotype has indicated a requirement for Bmp-7 in nephrogenesis (135). The mouse mutant (56, 165) is characterized by very small kidneys, by polydactyly of hindlimbs or hind- and forelimbs, and by microphthalmie, indicating that Bmp-7 is required also for limb and eye development. In the homozygous Bmp-7 mutants, initial branching of the ureteric bud and formation of some early epithelia of the S-shape stage were observed, but further growth and differentiation in both cell systems were terminated on E11.5, followed by massive apoptosis (165). Although the regulation of Bmp expression is not understood, Bmp proteins are often expressed in embryonic cells close to or identical with those expressing hedgehog (18); moreover, Bmp-7 and Bmp-2 RNA colocalize in some organs (166), suggesting cooperation of both in cell signaling related to proliferation and apoptosis.

Nerve growth factor receptors have been identified by in situ hybridiziation (237) in the aggregates close the ureteric buds, and they are repressed in later stages. When these metanephric organ cultures were incubated with antisense NGF receptor oligonucleotide, ureteric bud branching and tubulogenesis were inibited, consequent to the suppressed nerve growth factor receptor expression. Although the low-affinity NGF receptor is expressed in uninduced metanephric mesenchyme (61), the high-affinity receptor is expressed in stromogenic cortical mesenchymal cells. This interesting, possibly important and nondiffusive signal system (132) awaits further molecular analysis.

In conclusion, the role of growth factors in nephrogenesis is particularly difficult to evaluate. Embryonic kidneys in organ culture synthesize and secrete a spectrum of growth factors; when these are added to the uninduced metanephric mesenchyme, however, growth and nephrogenesis do not occur, although some of them increase DNA synthesis. Furthermore, mouse gene knockout studies of growth factors and their receptors do not sustain the roles of single growth factors implicated from the in vitro expression studies. The other problem in evaluating the role of growth factors arises from the fact that addition of antisense DNA (or of an antibody) to only one of the many growth factors produced in the embryonic kidney can entirely block tubulogenesis, although all other factors are still effective.

Most ECM molecules contain multiple binding domains that are recognized to interact with integrins (72), their glycoproteins have signal transducing receptors (240), and they mediate events such as cell adhesion (300) in embryonic cells. While the composition of ECM depends on cell type and developmental stage (66), interactions between matrix molecules and growth factors are effective through differing mechanisms. For one, binding of a growth factor to ECM can alter its local activity (302). Furthermore, growth factors can alter ECM protein and receptor production importantly (161), and ECM molecules can modulate (69) the response of cells to growth factors.

With induction and the onset of MET, a profound change to the epithelial phenotype occurs (Fig. 6). Collagen I and II as well as N-CAM expressions are lost, whereas a multitude of epithelium-specific molecules begin to be expressed (Fig. 7), namely, the basement membrane proteoglycan syndecan-1, the laminin A chain, E-cadherin, ₆₁-integrin, and type IV collagen (66). Laminin B1 and B2 chains, but not the A chain, are expressed in the precondensation mesenchyme; induced mesenchymal cells during condensation synthesize high amounts of A chain for the expression of functionally active laminin. When laminin or ₆₁-integrin functions are disturbed during this transformation, cell differentiation to the polarized phenotype is inhibited (65, 141, 255). Thus a large number of ECM components are expressed by induced mesenchyme, and some of them are known to be required for epithelial morphogenesis.

The critical role of laminins is indicated by the fact that antilaminin antibodies inhibit mesenchymal-epithelial conversion and cell polarization (41, 255). As soon as the condensation stage has been reached (Fig. 3), laminin B expression is upregulated, indicating the onset of polarization. WT-1 transcription increases (Fig. 15) and initiates the downregulation of PDGF (76), IGF-II (53), and the IGF-I receptor (296), all of which are involved in mesenchymal cell proliferation. Laminin and receptors for laminin, such as dystroglycan and ₆₁-ntegrin, link ECM and its signals to the cytoskeleton. As the condensation stage proceeds toward polarization (Fig. 6), basement membrane polypeptides (59) such as the B chain of laminin and collagen IV, the A chain of laminin and its ₆-integrin receptor, and uvomorulin (287) are sequentially expressed. Laminin-1 probably is the principal component during the polarization process. Although laminin-₁ is expressed in embryonic epithelia only, laminin-₁ and -₁ are localized in both mesenchymal and epithelial cells. When mesenchymal cells are induced, laminin-₁ and -₁ mRNA increase immediately, but expression of laminin-₁ remains low until the mRNA expression is upregulated with epithelial cell polarization (66). In the same stage, expression of E-cadherin and ₆-integrin increase (Fig. 7). Specifically, the E3 and E8 fragments of laminin-1 are importantly involved in the polarization process. Antibodies against the E8 fragment of laminin-1 (141), as well as monoclonal antibodies against the integrin ₆-subunit (255), perturb tubulogenesis.

Laminin-1 appears to act via two receptors, the ₆₁-integrin and dystroglycan (60). The ₆-subunit is coexpressed with the laminin-₁ chain during mesenchyme-to-epithelium conversion, implying that it may act as receptor for laminin-1 (60). Importantly, polarization and further tubulogenesis are blocked at the vesicle stage by antibodies interfering with the binding of the A-chain laminin to the ₆-integrin (255). Dystroglycan mRNA is highly expressed in the basal cell membrane of renal embryonic epithelia, and it appears to be, in addition to ₆₁-integrin, a second important and independent cell receptor to attach the polarizing embryonic epithelial cell to the E3 fragment of laminin-1. With nephrogenesis, expression patterns of the many integrin subunits change differently in different nephron segments (196). Moreover, epithelial maintenance in the mature nephron also appears to be controlled by cell-matrix interactions as shown by the fact that basement membrane prevents apoptosis, suggesting a major role for mature cell-ECM interactions (74). These studies on the role of dystroglycan (60) have opened a new road of research in epithelial cell polarization.

The operational scheme common to all integrins (190) is the interaction with an ECM-derived ligand at its large extracellular domain (289), and transduction of the generated signal through a short cytoplasmic domain to components of the cytoskeleton and other intracellular signaling systems, as is expected from a conventional receptor. Synergistic actions between growth factor and integrin signal pathways (129) may be particularly important in cell adhesion, i.e., in MET. Expression patterns of integrins first change with induction (190). The uninduced mesenchyme expresses the ₁₁- and the ₄₁-integrins. The ₁₁-integrin receptor for fibronectin, interestingly, is coexpressed with cellular fibronectin in the uninduced mesenchyme, and they are both lost after induction (143). As soon as the ureteric bud, which then expresses ₂- and ₆-immunoreactivity, invades the mesenchyme, the now-induced mesenchyme expresses cell-cell contacts, loses the expression of the ₁- and ₄₁-integrins, and organizes into polarized epithelia that now express the ₆₁-integrin receptor for laminin in the cell basal membrane (143). This expression pattern announces the onset of interactions between the polarizing cell and basement membrane components. With progressing tubulogenesis (Fig. 3), the ₃₁-integrin appears expressed in the epithelia of Bowman's capsule and in podocytes, whereas ₁₁-integrin is expressed in the mesangium and ₂₁-intregrin in endothelial cells. Thus the terminal pattern of integrin subunit expression is established at the S-stage. In conclusion, these intricate developmental expression patterns of the integrin subunits might represent changing interactions between basement membrane components and integrins, and they might mirror events in the cell polarization process that are poorly understood.

	VIII. GENES THAT CONTROL RENAL ORGANOGENESIS

Top Previous Next References

A.  Transcriptional Regulation

1.  WT-1

The Wilms tumor suppressor gene WT-1 is a highly conserved (24, 136) transcription factor, essential for nephrogenesis (148, 215), and member of the early growth response family of the zinc-finger type (15, 26). Although WT-1 is faintly expressed in uninduced stem cells (Fig. 15), its expression is upregulated after induction (4) in condensing cells (212), and it continues to be expressed through the vesicle-, comma-, and S-shaped stages (121a), whereas high levels in the terminal nephron persist in podocytes only (Fig. 5). From this location in time and space, WT-1 is likely to control genes that code for developmental events (33, 168, 224, 225, 292), such as epithelial cell polarization and expression of differentiated properties. The roles of WT-1 in nephrogenesis were established from the analysis of WT-1 null mutants (148). In homozygous mice, the ureteric bud does not form and, consequently, the uninduced mesenchyme that is not capable to enter transition undergoes programmed cell death resulting in renal agenesis, whereas mesonephric duct and tubule development appear to be largely normal. Thus it appears by now that one of the major roles of WT-1, apart from the gene's roles in other organs (4), is to mediate the effects of the inductive signals during the MET, i.e., to render the stem cells competent for induction. WT-1 not only regulates the expression of several growth factors and growth factor receptors, but it also targets Pax-2 (225) and itself, the WT-1 (224), and transcriptional repression of both genes is mediated through WT-1 binding sites in the promoter regions. Therefore, regulation of the temporospatial expression pattern of Pax-2 (50) can now be linked to that of WT-1, since a rise of WT-1 protein in the proximal S-shape concurs with a fall of Pax-2 expression in the same cells (225). Pax-2 expression starts in the condensing mesenchyme, and Pax-8 expression follows in the vesicle stage (Fig. 15). It is only when both paired-box proteins are maximally expressed that WT-1 levels rise and thereby repress Pax-2 levels (225). This pattern (Fig. 15), taken together with the high abundance of Pax-2 in Wilms tumors (51), and with the positive modulation of WT-1 expression through transactivation by Pax-2 (43) and Pax-8 (44) proteins, indicates important regulatory interactions (175) between these early expressed and critical genes. A further transcriptional regulation by WT-1 is the repression of bcl-2 and of c-myc (28, 103), which emphasizes its role in controlling protooncogenes involved in apoptosis (bcl-2) and proliferation (c-myc) of uninduced mesenchyme. Thus WT-1 acts as a critical regulatory protein in metanephric blastema where it addresses also the growth factor-receptor pathway of IGF-I/IGF-II/IGFR-I. The finding that the IGF-II gene is repressed by WT-1 (Fig. 15) denotes a target gene that is critically important for normal growth and that, when deregulated, could advance tumor growth. In fact, IGF-II, the potent fetal mitogen, is overexpressed in most Wilms tumors, which emphasizes the important physiological downregulation of IGF-II by WT-1 to suppress mesenchymal proliferation (53, 168).

To summarize, WT-1 is a center transcription factor in early nephrogenesis (216), where it appears to organize much if not all of the intricate MET program. Finally, WT-1 is a striking example of how an alteration in transcription factor function (182) can result in oncogenesis (see sect. IX).

Pax-2 is one of the ancient renal genes, since it is expressed in the pronephros and in the Wolffian duct. Before renal organogenesis begins, i.e., before E11.5, Pax-2 is expressed (50) in the intermediate mesoderm from where the metanephros is derived. Within the paired-box gene family, Pax-2 and Pax-8 are expressed in overlapping domains (Fig. 5) in the embryonic mouse kidney (49-51, 210). Although Pax-2 proteins have properties of transcription factors such as sequence-specific DNA-binding and, particularly, transcriptional activation (51), downstream target genes of Pax-2 are not yet known. None of these two paired-box genes is expressed in the uninduced mesenchyme, which renders Pax-2 the herald of successful inductive interaction. In fact, the very first stage of cell polarization, i.e., the initial step toward the epithelial phenotype during mesenchymal-to-epithelial conversion, appears to be a major area of control by Pax-2 (223). This implies, of course, that the early expression requires induction by the ureteric bud. Also, it is readily understandable that Pax-2 is not expressed in the mesenchyme if induction has not occurred, as is the case in Danforth's short-tail mice with renal agenesis (207). Pax-2 expression is downregulated at the beginning of the S-stage. Persistent expression of Pax-2 in transgenic mice (52) inhibits terminal differentiation of the nephron and leads to abnormal glomerular and proximal tubular development, including absent foot processes and microcystic tubular dilatation. Thus the failure to repress Pax-2 in tubular and glomerular cells of transgenic mice illustrates the importance of WT-1 in the downregulation of Pax-2 (Fig. 15) for terminal nephrogenesis. In conclusion, the Pax-2 gene appears to be a pivotal mediator for signals transferred from the ureteric bud; other regulating genes, namely, Wnt-4 (258), Bmp-7 (55, 165), and WT-1 (148), are fundamentally involved in the Pax-2 stage of nephrogenesis.

Distinctive differences between Pax-2 and Pax-8 expression are apparent in nephrogenesis (63, 210). Although Pax-2 is expressed earlier than Pax-8 in the condensed murine mesenchyme, Pax-2 in the S-stage is downregulated before Pax-8 (Fig. 15). Pax-8 is not expressed in the Wolffian duct and its descendent epithelia (210), and potential target genes under the control of Pax-8 have been identified in the thyroid, encoding thyroglobulin and thyroperoxidase, but none has yet been found in the embryonic kidney. Importantly, Wnt-4 appears to be required for the sequential appearance of Pax-2 and Pax-8, since mesenchymal cells of Wnt-4 / mutants express Pax-2 after a regular induction, but they do not proceed to the expression of Pax-8 (258).

BF-2 is first expressed in mesenchymal cells adjacent to those induced cells that express Pax-2 (97). Thus Pax-2 and BF-2 are nephrogenic and stromogenic stem cell markers, respectively. In BF-2 mutants, stromal/interstitial cells persisted throughout embryonic cortical and medullary morphogenesis (97), whereas normally, stromal cells make up most of the embryonic medulla only and eventually disappear from there by apoptosis (228). Some days after the early induction, stromal cells carrying BF-2 are present in the medulla, and in a peripheral cortical (stem) cell population (97) that is different from the other (stem) cell population expressing Pax-2. In the mesenchymal cell population, BF-2 / kidneys formed only few but large condensates that expressed Wnt-4 and c-ret but did not progress to differentiation. Branching morphogenesis was greatly reduced, and the pattern of c-ret expression, normally limited to the very tips of the ureteric buds, now extended throughout the entire cortical system and into the medullary segments of the collecting duct. The important aspect of this latter finding is that BF-2 in wild-type kidneys is not expressed in the ureteric bud or the early collecting tubule. The conclusions from these most revealing data are as follows: 1) BF-2 expression follows the early bifurcating lineage into nephrogenic (Pax-2) and stromal (BF-2) cells, and 2) the action of BF-2 as a transcription factor is to regulate the synthesis and secretion of signaling molecules that guide morphogenesis and differentiation of both primordial tissues, the nephrogenic and the ductal. Clearly, the highly organized embryonic expressions of BF-2 and its broad range of developmental responsibilities will open new perspectives in nephrogenesis.

The genes contain a homeobox, a 183-bp sequence encoding a 61-amino acid sequence that has a DNA-binding motif for transcriptional regulation of sofar nondefined genes. Some of the homologs of Hox genes that exist in Drosophila are expressed in the meso- and metanephros, such as Hoxd-3 (263). Although Hox genes are believed to be involved primarily in regional patterning during embryogenesis, their specific roles in kidney organogenesis remain to be discovered, since very few of the targeted mutations so far have resulted in renal defects (149), suggesting redundant gene functions in the cluster or nested expression domains. Hoxb-9 can activate the promoter for N-CAM (128), a molecule that is expressed, as mentioned, during the early MET.

Hepatocyte nuclear factor (HNF)-1 and HNF1- are highly expressed in distinctly different temporospatial patterns in nephrogenesis (153). Although HNF-1 is expressed first in the S-stage only, HNF-1 is already expressed in the mesonephros, in the ureteric bud, in mesenchymal condensations, and in all subsequent stages up to the mature nephron, suggesting clearly different roles for the two HNF. The functional role of other HNF families in nephrogenesis (57) is poorly defined, although a gene homologous to HNF-3, the MFH-1, is highly expressed in uninduced mesenchymal cells (181).

N-myc is generally expressed in epithelia that are involved in inductive signaling (108), and it regulates, as demonstrated by homologous recombination in embryonic stem cells, early (precursor) proliferation (29), branching morphogenesis, and differentiation. The c-myc transcripts are found in the uninduced mesenchymal cells and in early differentiating stages, but, importantly, not in mature epithelia (189), where renal tubular hyperplasia and cysts occur if the gene is constitutively expressed in transgenic mice (274). The N-myc protooncogene appears to have an earlier frame of effects in the embryonic kidney (188), since as revealed by gene targeting of N-myc (256), the mesonephros in these transgenic mice had a lower number and smaller size of tubules and mesenchymal hypoplasia. The wild-type gene is expressed in a restricted pattern contrasting that of c-myc, since N-myc begins to be expressed during MET in the condensation stage, i.e., before Wnt-4 expression (258), and it persists through the S-stage (189); the mutant embryos die at E11.5 (256), indicating that not only renal but also generalized effects result from the N-myc loss of function (241).

Many protooncogenes are assumed or established to encode receptor proteins that participate in processes that are activated by growth factor signaling molecules, and they are expressed during embryonic organogenesis in distinct temporospatial patterns (248).

The c-met protooncogene product Met is the receptor for HGF, and it is expressed (Fig. 5) in ureteric bud cells and in early postinduced mesenchyme (3, 20, 252, 294, 301). The mesenchymal expression of c-met transcripts begins in the condensates close to the ureteric bud tip cells but not in the uninduced mesenchymal cells (206). The role of this protooncogene and its encoded receptor tyrosine kinase is unique in that it not only participates in branching morphogenesis but also in the induction of epithelial motility (5, 226). Notably, the Met receptor tyrosine kinase is capable of transducing signals for the diverse pathways and biologic functions of HGF/SF, such as proliferation (mitogen), motility ("scatter"), and branching growth (morphogen); transduction may be modified by the differential binding characteristics of intracellular signaling molecules to the phosphotyrosine residues of the activated cytoplasmic receptor domain (294).

The c-ret protooncogene encodes Ret, which is developmentally expressed in ureteric bud tip cells (Fig. 5) but not in the mesenchyme (200). The tip cell expresses a particular phenotype (121), and it is involved in multiple signaling loops, since it expresses also the protooncogene encoded receptor tyrosine kinases c-met and c-ros. The phenotype of homozygous c-ret mutant mice (247) and the embryonic expression pattern of c-ret (200) both indicate basic roles for the protooncogene in kidney (126) and in the intestinal nervous system. In c-ret / mice, the ureteric bud does not branch because c-ret directed signals from the mesenchyme cannot be received; however, mesenchymal cells enter the first stage of nephrogenesis which suggests that, in c-ret /, some early ureteric bud signaling has occurred.

The kidneys were either dysplastic rudiments characterized by reduced or absent ureteric branching and extended regions of uninduced mesenchmye, or they were absent (248). The primary defect of c-ret / is intrinsic to ureteric bud, since the metanephrogenic mesenchyme from these mutant embryos developed into tubules when cocultured with embryonic spinal cord (247). The Ret receptor and its ligand GNDF (58, 276), which is probably mesenchyme derived, constitute the second system known to signal between the two primordial tissues, and both share the same signaling pathway (58, 186, 200, 247, 273, 285). Remarkably, GDNF does not bind directly to the Ret receptor (233), but it binds first to its GDNFR-, and this complex interacts with Ret (273). This signaling pathway induces branching morphogenesis of the ureteric bud, in concert with the HGF-Met signal system. In conclusion, c-ret is another example of the dual role of protooncogenes in embryonic development and in cancer. Although the role of c-ret in embryonic signaling is fairly obvious from both wild-type localization (200) and renal and intestinal defects in c-ret recombinant mice (247), the neoplastic lesions in naturally occurring human mutations are less understood. These are associated with the familial medullary thyroid carcinoma and the multiple endocrine neoplasia type 2A (273), and the mutations appear to alter the extracellular domain of Ret such that a constitutive increase of the receptor kinase activity may be considered.

The murine gene c-ros (267), the homolog of Drosophila sevenless (30, 146), is development-dependently expressed in intestine, lung, and kidney. The c-ros transcripts Ros were the first protooncogene-encoded receptor tyrosine kinases demonstrated to express a temporospatial pattern in which they are highly expressed in ureteric bud tip cells in the outer cortex of embryonic kidney (251) and decline in the neonatal kidney (131). The ligand for Ros, in contrast to those for Ret and Met, is not yet known (Fig. 2). Protooncogenes and ECM proteins are functionally linked. Both c-ros and c-ret antisense oligonucleotides in the culture medium (163) inhibit c-ret expression more than c-ros and, remarkably, the expression of ECM proteoglycans in the interspace between mesenchymal and ureteric bud cells with resulting dysmorphogenesis (131). In conclusion, ECM glycoproteins modulate morphogenesis, and c-ros transcripts colocalize with these modulators that may imply that ECM-mediated signaling is part of the early inductive process. Finally, the signal for activation of Ros could involve direct cell-cell interactions across the interspace, between the ureteric bud receptor and the putative mesenchyme-derived ligand molecule (131, 163).

	IX. GENETIC ERRORS IN NEPHROGENESIS

Top Previous Next References

Inherited diseases, notwithstanding their deleterious effects in the patient, provide insight in developmental events. If renal malformations can be linked to mutations of specific genes, the role of the latter in normal embryogenesis of an organ can be deduced. In many examples, these errors in kidney development appear to be caused by erroneous transcriptional regulation. A few are discussed in their phenotypical context.

Polycystic kidney disease (PKD) is characterized by intrarenal cysts with an inside single cell layer of epithelial cells. The two major human genetic forms of PKD are the autosomal dominant polycystic kidney disease (ADPKD) and the autosomal recessive polycystic kidney disease (ARPKD). The cyst formation of ADPKD begins in the embryonic kidney and is maintained by fluid secretion into the cyst lumen (22, 87), which is mediated by CFTR (36). The cyst expansion process most likely is slow, since the disease is detected only in the third decade of life. In contrast, the recessive or juvenile form of PKD begins with a rapid embryonic cyst expansion, first in the proximal and next in all of the collecting tubules, leading to an early failure of the kidney (299) during the first year of life. In both forms, cell-matrix interactions are altered (35). The increased expression of the ₁-subunit in collecting duct cells and an irregular distribution of ₂-, ₃-, and ₆-subunits may contribute to the phenotype of the cystic disease.

Autosomal dominant PKD is caused by the PKD1 gene that encodes the PKD1 460-kDa protein, the polycystin (79, 122, 268, 293), which is a transmembrane protein (124) mediating interactions between ECM and the cell. Importantly, a second gene (PKD2) encodes a protein (PKD2 110 kDa) that possibly interacts with polycystin (184), and both may be part of a cyst-producing pathway. Of particular interest is the early embryonic expression of polycystin (293), which was not seen in uninduced metanephric mesenchyme and weakly expressed in the ureteric bud, in the comma- and S-shaped stages, and in fetal podocytes, whereas expression in cyst cells appeared to be prominent. Although polycystin is unequivocally expressed in embryonic renal epithelia (78, 88, 284), its role in nephrogenesis is not defined. It may be suggested, assuming normal copies of polycystin to act as ECM receptors, that this protein could participate in the epithelial polarization process following MET. Several basic features distinguish ADPKD from ARPKD. In the latter, the protooncogene c-myc is continued to be expressed, together with SGP-2, and a link to the cause of ARPKD may be provided by the fact that transgenic mice that express c-myc constitutively develop cysts much like those in ARPKD (274). Another phenotype of PKC results from a targeted disruption of bcl-2, where the homozygous mutants (194, 286) perinatally showed cystic dilatations in proximal tubule segments and, in addition, cysts in postnatal collecting duct (187); the gene generating this cystic phenotype remains to be defined.

Wilms tumor is the result of faulty MET. It shows as an embryonic cancer (nephroblastoma) that originates in early nephrogenesis from mesenchymal stem cells undergoing excessive proliferation, while their normal fate is to either differentiate into epithelial nephron cells or to enter apoptosis (202, 305). Wilms tumors frequently express Pax-2 in their epithelial components (62) and contain cells reminiscent of epithelia, stroma, and blastema. The genetic errors leading to nephroblastoma are understood only incompletely (25). The WT-1 protein binds to multiple sites in the fetal promoter of the IGF-II gene and represses IGF-II transcription (53). In this way, WT-1 can inhibit IGF-II synthesis, and it could be a negative regulator of renal embryonic mesenchymal growth. Furthermore, the PDGF-A promoter is a target structure for transcriptional repression by a splice variant of WT-1 (76). Both data suggest that WT-1 acts through the downregulation of growth factor production; WT-1 expression is upregulated at the onset of condensation (Fig. 15), and there it suppresses the transcription specifically of IGF-II and of PDGF-A interrupting normal proliferation of the mesenchymal cell population. Mutant loss (in some 15% of all patients with Wilms tumors and in patients with Denys-Drash syndrome) of this suppressing activity of WT-1 might be the decisive factor for the massive and erroneous blastemal growth. In conclusion, WT-1 is a striking example of the general notion that genes central to pattern formation in embryonic organogenesis or for cell lineage differentiation are important also in carcinogenesis. A type of WT-1 mutation is associated with the Denys-Dash syndrome, a rare congenital disease with renal failure, urogenital dysgenesis, and Wilms tumor. WT-1 mutations in Denys-Drash syndrome patients appear to have a single nucleotide exchange only (202), which may have altered the binding characteristics.

Renal cell carcinoma is a frequent adult malignancy that is assumed to be generated from proximal tubule epithelium, and a high percentage of primary tumors and renal cell carcinoma cell lines (73%) expressed Pax-2 (81). Significantly, Pax-2 protein synthesis in the cell lines could be inhibited by the addition of antisense oligonucleotides directed against Pax-2 mRNA, which resulted in growth inhibition of the cancer cells, whereas cancer cell lines that did not express Pax-2 did not change growth when the same antisense oligonucleotides were used in culture. These significant data suggest that the reexpression of Pax-2 in mature epithelium may be a predisposition for oncogenesis. In conclusion, developmental control genes regulating cell proliferation and differentiation in the embryo can be reexpressed in mature cells and in this unknown way contribute to the initiation and expansion of tumors.

Two human syndromes may be related phenotypically to the human homologs of Pax genes. Waardenburg syndrome is a mutation of the human homolog of Pax-3, which is a heritable autosomal dominant combination of lateral displacement of the inner canthii of the eye, pigment alterations, occasional (one-third) deafness, and mental retardation (264). The AN (aniridia) gene, the human homolog of Pax-6, is completely or partially deleted in patients with aniridia, a complete or partial absence of the iris and disturbed development of lens, cornea, and retina (104, 271). Aniridia is also expressed in the WAGR syndrome, which is characterized by several abnormalities such as Wilms tumor (W), aniridia (A), genitourinary tract defects (G), and mental retardation (R) (203). Lastly, a syndrome that alters early ureteric bud and mesenchymal development in a dominant mouse mutant, the Danforth's short tail (Sd), has not been characterized genetically (80).

	X. CONCLUDING REMARKS AND PERSPECTIVES

Top Previous Next References

Nephrogenesis today, through the convergence of Drosophila, mouse, and human genetics, represents a model system by which to study the basic regulation of ductal morphogenesis, inductive mesenchyme-epithelium tissue interaction, tissue remodeling, cell polarization, expression of epithelial cell transport properties, and the developmental programs behind. The route presently followed by many groups is to probe the temporospatial expression pattern of a particular gene and to relate it not only to a morphogenetic stage but also to simultaneously expressed patterns of other genes. In this way, patterns of expression on a time scale, corresponding to a morphogenic scale (Fig. 15), may suggest routes for future experiments. Furthermore, interference with transcriptional or posttranscriptional expression of single genes has contributed to establish its causal relation to a phenotype. Nevertheless, although some "upstream" regulations of gene expression or repression are known, there is little information about most of the "downstream" gene target molecules. Although many of the impressive crowd of registered kidney genes (38) are likely to be involved in organogenesis, it has been most revealing to work with only a few of them to establish some activation and repression patterns as morphogenesis and differentiation proceed from one stage to the next (Fig. 3). Thus one important area of progress that emerged in the past decade is the identification of some of the genes that initiate and direct the stages of nephrogenesis. In this context, the critical question of signaling molecules, their receptors, and their signal pathways has only just begun to be investigated, and the discovery of the ligand for the c-ros-encoded receptor tyrosine kinase might be one of the next landmarks in this field. Among the currently most interesting aspects are provided by work on biochemical interactions among gene products, as has been shown impressively for the regulatory circuit between Pax-2 and WT-1. Although these studies will undoubtedly lead to an understanding of causal relations in morphogenesis, cell polarization, and differentiation, they will ultimately serve to answer the question of how signaling molecules, receptors, and transcription factors contribute to those renal diseases that are believed to express an aberrant or disrupted developmental program.