| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |

*INSERM U489 and The University Pierre and Marie Curie, Paris, France; and
Pasteur Institute, Paris, France.
Correspondence to Dr. Jerome A. Rossert, INSERM U489 and Department of Nephrology, Tenon Hospital, 4 rue de la Chine, 75020 Paris, France; Phone: +33-1-56-01-60-29; Fax: +33-1-56-01-69-99; E-mail: jerome.rossert{at}tnn.ap-hop-paris.fr
| Introduction |
|---|
|
|
|---|
As for many other structures, the combinatorial action of different cell-specific transcription factors is very likely to play a critical role in the development of the ureteric bud and of the metanephric mesenchyme. In this review, we focus on transcription factors that have been shown to play a role in vivo in the differentiation of metanephric mesenchymal cells into epithelial or stromal cells and in the differentiation of ureteric bud cells into epithelial cells of the excretory system.
| Early Differentiation of Metanephric Mesenchymal Cells |
|---|
|
|
|---|
|
superfamily that is produced by the metanephric mesenchyme and that is necessary for ureteric budding (8). A possible role for Pax2 thus would be the induction of mesenchymal competence and of Gdnf expression, which is in agreement with the ability of Pax2 to activate the transcription of Gdnf in vitro. However, analysis of Eya1-/- mice and of Hox11 null mutant mice has shown that Pax2 is not sufficient to activate Gdnf expression (cf. infra) and that induction of the metanephric blastema probably requires a combinatorial action of different transcription factors. The Eya genes encode proteins that possess transactivation activity but do not seem to bind DNA and probably act as coactivators. They contain a so-called Eya domain that is indispensable for coactivation activity. In Human, haploinsufficiency for Eya1 results in the branchio-oto-renal and branchio-otic syndromes that associate craniofacial abnormalities, hearing loss, and, for the branchio-oto-renal syndrome, kidney defects (9). During kidney development, Eya1 is expressed in the metanephric mesenchyme but not in the ureteric bud or its derivatives (10). At birth, Eya1 knockout mice lack kidneys and ureters because of a failure of the ureteric bud to form, and analysis of kidney development in Eya1-/- embryos has shown that Pax2 is expressed but that expression of Gdnf and Six2 cannot be detected (11). Thus, Eya1 seems to act downstream of Pax2 and to be indispensable for acquisition of mesenchymal competence and for expression of Gdnf by metanephric mesenchymal cells.
The Hoxa11, Hoxc11, and Hoxd11 genes are expressed in the metanephric mesenchyme but not in the Wolffian duct or the ureteric bud or its derivatives (1214). At birth, mice harboring null mutant alleles for all three of these genes have no kidney, and analysis of kidney development has shown that the ureteric bud never forms (15). In situ hybridization experiments have shown that Wt1, Pax2, and Eya1 are normally expressed, whereas Six2 and Gdnf are absent. Furthermore, analysis of embryos with only five mutant alleles support the hypothesis that the levels of expression of Six2 are dependent on the numbers of functional Hox11 alleles, because these mutants, which have one functional allele, show weak expression of Six2 (15). These data are in agreement with the hypothesis that Pax2, Eya1, and Six2 form a functional network and that induction of Six2 by Eya1 and Hox11 gene products is necessary for Gdnf induction (Figure 1).
Six2 is a transcription factor that is characterized by the presence of a Six-type homeodomain and of a Six domain located N terminal to the homeodomain. Six and Eya have been shown to interact directly and to induce synergistic activation of promoters (16). Unfortunately, targeted disruption of Six2 has not yet been reported, and the role of this gene in metanephric blastema differentiation and induction of Gdnf expression is purely hypothetical.
Foxc1/Mf1/Fkh1 is a transcription factor that belongs to the forkhead/winged-helix family and that seems to interact with the Pax2-Eya1-Six2 pathway. The forkhead family contains transcription factors that share an evolutionarily conserved DNA-binding domain named forkhead domain. Studies of knockout mice have shown that members of this family play essential roles during embryonic development and in particular that they regulate cell proliferation, cell fate determination, and differentiation. During embryonic development, Foxc1 is expressed in various tissues, including the metanephric mesenchyme, but not in the Wolffian duct or the ureteric bud (17). A proportion of mice that lack functional Foxc1 genes display duplex kidneys associated with double ureters (17). The additional ureter arises from the Wolffian duct more anteriorly to the normal ureter and is not connected to the bladder. It probably results from anterior persistence of Gdnf expression domain, which suggests that Foxc1 may negatively regulate the expression of Gdnf. Because the expression of Eya1 also extends anteriorly in mice that do not have functional Foxc1 genes (17), Foxc1 may inhibit the expression of Eya1, which, in turn, inhibits the expression of Gdnf (Figure 1). The Pax2-Eya1-Six2-Hox11-Foxc1 pathway is not the only one to be involved in early induction of metanephric mesenchyme, and other genes that do not seem to interact directly with this pathway, such as Wt1 or Sall1, are also indispensable for early kidney development.
The Wilms tumor-suppressor gene Wt1 encodes a DNA-binding protein that contains four zinc fingers of the C2H2 type. It exists as different isoforms that are generated by alternative splicing and by usage of alternative translation start sites and that have roles as transcription factors but also in RNA processing (18). In Human, mutations in WT1 are responsible for Wilms tumors but also for the WAGR (Wilms tumor- Aniridia-genitourinary abnormalitiesmental retardation), Denys-Drash, and Frasier syndromes, which all are characterized by the presence of kidney abnormalities (18). During kidney development, Wt1 is expressed at relatively low levels in the metanephric mesenchyme before induction by the ureteric bud and at higher levels in condensing mesenchymal cells, in vesicles, in comma-shaped bodies, and at the proximal part of S-shaped bodies in cells that will become podocytes (19,20). In Wt1-null mice, the metanephric blastema forms, but the ureteric bud does not sprout from the Wolffian duct, and thus the metanephric mesenchyme undergoes apoptosis and kidney does not form (21). Ex vivo coculture experiments using metanephric blastema from Wt1-/- mice have shown that it is unable to condense and that Wt1 is indispensable for induction of the metanephric mesenchyme by normal ureteric buds (21). Wt1-null mice complemented with a YAC transgene spanning the Wt1 locus survive until birth, which allows a more precise analysis of the role of Wt1 during kidney development (22). Analysis of these mice has shown that Wt1 is required not only for early induction of the metanephric mesenchyme but also at later stages of kidney development. Although the majority of mice show a total absence of ureteric bud development, similar to what is seen in uncomplemented null mice, some embryos have hypodysplastic kidneys where metanephric mesenchyme condenses around the tips of the ureteric bud branches but forms no or very few epithelial structures. Furthermore, whereas Pax2 expression is present but reduced in kidney mesenchyme of Wt1-/- mice, it is normal in complemented mice. This suggests that Wt1 modulates the early differentiation of metanephric mesenchyme but does not directly upregulate Pax2. Thus, Wt1 would act in combination with Pax2, Eya1, Six2, and Hox genes to induce early metanephric mesenchyme differentiation (Figure 1).
Sall1 (Sal-like 1) is the mammalian homolog of the Drosophila region-specific homeotic gene Spalt. It is a homeotic gene that contains 10 zinc finger motifs and that is expressed mostly in the metanephric mesenchyme surrounding the ureteric bud during metanephros development (23). Heterozygous mutations of SALL1 in Human lead to the Townes-Brocks syndrome, which is characterized by dysplastic ears, polydactyly, imperforate anus, and kidney and heart abnormalities (24). Sall1-null mice have either complete kidney agenesis or severe renal dysplasia (23). Analysis of metanephros development has shown that the ureteric bud forms but fails either to invade the metanephric mesenchyme or to grow and branch within the metanephric mesenchyme, which leads to its apoptosis. Coculture experiments have shown that Sall1 is required for the metanephric mesenchyme to attract the ureteric bud, and analysis of genes expression has shown that it does not have direct effects on Wt1, Pax2, Eya1, or Gdnf expression (23). Thus, still another pathway seems to be needed for proper development of the metanephric mesenchyme, besides the Pax2-Eya1-Six2 pathway and besides Wt1 (Figure 1). It is of note that analysis of knockout mice has shown that Sall2, which is closely related to Sall1 and which is also expressed in the metanephric mesenchyme surrounding the ureteric bud, is dispensable for kidney development (25).
| Differentiation of Tubular Epithelial Cells and of Podocytes |
|---|
|
|
|---|
|
HNF1
/LF-B1 is a dimeric homeodomain transcription factor that is expressed in different organs, including liver and kidney. It can bind DNA as homo- or heterodimers, and HNF1
/vHNF1/LF-B3 and DcoH (dimerization cofactor for HNF1) have been shown to heterodimerize with HNF1
. During kidney development, HNF1
is expressed selectively in proximal tubular cells, and HNF1
-/- mice display defects in terminal differentiation of proximal tubular cells responsible for Fanconi syndrome (27,28). In particular, they have glucosuria, aminoaciduria, and phosphaturia. In Human, heterozygous mutations of HNF1
also lead to impaired renal glucose reabsorption (29).
As described above, Pax2-/- mice lack ureteric bud (6). Thus, the metanephric mesenchyme undergoes apoptosis, and the role of Pax2 during late stages of metanephrogenesis cannot be analyzed. However, in organotypic cultures, inhibition of Pax2 expression prevents the condensation and epithelial conversion of metanephric mesenchymal cells (30). Furthermore, overexpression of Pax2 in transgenic mice is responsible for abnormal differentiation of tubular cells (31). Thus, Pax2 may play a role in epithelial conversion of mesenchymal cells and in differentiation of tubular cells.
Lim1 is a transcription factor that contains a LIM class homeodomain and two LIM domains. During kidney development, Lim1 is expressed in ureteric bud branches and in collecting ducts but also in vesicles, comma-shaped bodies, S-shaped bodies, and developing tubules (32). Analysis of the role of Lim1 during kidney development in Mammals is difficult because null mutant mice lack head structures and die around day 10 pc, before development of the metanephros. However, the four mice that survived until birth completely lacked kidneys and gonads (33). Study of the role of Xlim1 during development of Xenopus embryos has shown that this gene synergizes with Pax8 to induce pronephric tubule differentiation and growth (34). In Mammals, Lim1 thus could also be involved in inducing epithelial cell differentiation.
Although podocytes are highly differentiated cells that express many specific markers, only a few transcription factors that are involved in podocyte differentiation have been identified by using knockout mice (Figure 2). Because early kidney development is grossly abnormal in mice that completely or partially lack functional Wt1 alleles, the role of Wt1 in podocyte differentiation cannot be assessed using these mice (21,22). However, the presence of glomerular defects in subjects with Denys-Drash syndrome or Frasier syndrome suggested that this gene was necessary for normal differentiation of podocytes (18). This hypothesis has been recently confirmed by generation of mice specifically lacking the +KTS or the -KTS isoform of Wt1 (35). These two isoforms result from the existence of two alternative splice donor sites at the end of exon 9, and they differ by the presence (+KTS) or absence (-KTS) of three amino acids between zinc fingers 3 and 4. In mice that lack the +KTS isoform of Wt1, podocytes remain cuboidal and formation of foot processes is severely impaired (35). Mice that lack the -KTS isoform have very small glomeruli and synaptopodin, and
3 integrin are almost completely absent, which suggests that early steps of podocyte differentiation are impaired (35). Furthermore, in vitro experiments suggest that Wt1 directly activates the transcription of podocalyxin, a gene that encodes a podocyte-specific protein (36).
As stated above, Pod1 is expressed in podocytes at all stages of differentiation, and Pod1 null mice show a clear defect in podocyte differentiation (26). In particular, they remain cuboidal with few foot processes. Furthermore, the density of the capillary network is markedly reduced, suggesting that podocyte differentiation is necessary for normal development of the capillary tuft.
Lmx1b is a DNA-binding protein that contains two LIM domains separated by one homeodomain. In Human, heterozygous mutations of LMX1B are responsible for the nail-patella syndrome (37). During embryonic development, Lmx1b is expressed in limbs but also in kidney, where its expression is restricted to podocytes (38). Lmx1b-/- mice display abnormal glomeruli that are not fully differentiated, contain dysplastic podocytes, and lack a capillary network (3841). The podocytes remain cuboidal, and no foot processes or slit diaphragms can be detected. Furthermore, they do not express podocin, and they synthesize decreased amounts of pro-
3(IV) and pro-
4(IV) collagen chains (3941).
| Differentiation of Stromal Cells |
|---|
|
|
|---|
Foxd1 is a member of the forkhead/winged helix family of transcription factors, which is expressed in different organs, including kidney, during embryonic development. In kidney, it is selectively expressed in stromal cells that surround condensed mesenchymal cells and differentiating nephrons and at lower levels in stromal cells located in the medulla around ureteric bud branches (42). Foxd1-/- mice die at birth, while they have small kidneys that are abnormally positioned. Analysis of kidney development in these mice shows a dramatic reduction in the number of S-shaped and comma-shaped bodies, and the presence of large amounts of condensed mesenchymal cells, which suggests first that stromal cells do not differentiate normally and second that they are necessary for epithelial differentiation of condensed mesenchymal cells (42). Furthermore, because expression of Ret (a receptor that is expressed by ureteric bud cells and that binds Gdnf) is abnormal in ureteric bud cells of Foxd1-/- mice, stromal cells may also be necessary for normal differentiation of ureteric bud cells (42). This hypothesis is in agreement with results obtained by analyzing null-mutant mice for RAR (cf. infra).
RAR are transcription factors that belong to the nuclear receptor superfamily and transduce the retinoid signal. In the presence of retinoic acid, they bind to enhancer elements and activate transcription. Analysis of double mutant mice that lack Rara (retinoic acid receptor
) and Rarb2 (retinoic acid receptor
2) shows that retinoic acid is indispensable for normal development of stromal cells. During kidney development, Rara is expressed at low levels throughout the embryonic kidney, whereas Rarb2 is expressed selectively in stromal cells (43). Unlike Foxd1, Rarb2 is expressed at similar levels in all stromal cells (43). Rara-/- Rarb2-/- double mutant mice have severely dysplastic kidneys with reduced branching of the ureteric bud and accumulation of stromal cells beneath the renal capsule (43). This inhibition of ureteric bud branching is associated with a reduced expression of Ret in ureteric bud cells, and overexpression of Ret in these cells can almost completely rescue renal development (44). Thus, binding of retinoic acid to its receptor is indispensable for normal development of stromal cells, and interactions between stromal cells and ureteric bud cells are necessary for normal development of the excretory system. However, these interactions may be either direct or indirect, mediated by interactions between stromal cells and epithelium-forming mesenchymal cells.
| Differentiation of Ureteric Bud Cells |
|---|
|
|
|---|
A role of Pax2 in differentiation of ureteric bud cells is mostly suggested by analysis of nephric duct formation. Pax8 is another member of the Pax family of transcription factors that is expressed during pro-, meso-, and metanephros development, but kidney development is normal in Pax8-/- mice (47,48). Whereas initial development of the nephric duct is normal in Pax2-/- mice, in Pax2-/- Pax8-/- double-mutant embryos, the intermediate mesenchyme does not undergo mesenchymal-epithelial transition and the nephric duct does not form (4749). In zebrafish, lack of functional Pax2.1 gene leads to abnormal epithelial differentiation of the pronephric duct (50). Conversely, in chick embryos, ectopic expression of Pax2 in the intermediate mesoderm and genital ridge can induce the formation of ectopic nephric ducts (50). Similarly, a role for Lim1 in ureteric development is suggested by the fact that ectopic expression of XPax8 and Xlim1 can induce the formation of additional pronephric tubules in Xenopus embryos (34).
In conclusion, although different transcription factors that are necessary for early differentiation of metanephric mesenchymal cells have been identified, little is known about the genes that regulate terminal differentiation of tubular epithelial cells or of ureteric bud cells. However, understanding the development of the renal excretory system is of critical importance because, in children, congenital abnormalities of kidney and urinary tract are the main cause of end-stage kidney disease. One approach could be to identify the enhancers that drive gene expression in a temporally and spatially controlled manner during differentiation of ureteric bud cells.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
E. de Laplanche, K. Gouget, G. Cleris, F. Dragounoff, J. Demont, A. Morales, L. Bezin, C. Godinot, G. Perriere, D. Mouchiroud, et al. Physiological oxygenation status is required for fully differentiated phenotype in kidney cortex proximal tubules Am J Physiol Renal Physiol, October 1, 2006; 291(4): F750 - F760. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
HOME
CURRENT ISSUE
ARCHIVES
JASN Express
ONLINE SUBMISSION
AUTHOR INFO
EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP |