INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

IN THE MAMMALIAN EMBRYO, kidney development depends on reciprocal interactions between two mesodermal derivatives: the ureteric bud, an epithelial outgrowth of the wolffian duct, and the metanephric mesenchyme. On induction by the ureteric bud extremities, the metanephric mesenchyme undergoes a series of morphogenetic events leading to nephron formation via epithelial differentiation. In turn, the mesenchymal tissue dictates growth and branching of the ureteric bud (16, 33). Regulation of kidney organogenesis is known to require proper activation of a set of genes (4, 23, 43). Advances in technique such as gene targeting in mice have allowed identification of growth factors and receptors as candidates to regulate the nephrogenic signals between the mesenchyme and the ureteric bud (26, 29, 31, 35, 36). At present, the glial cell line-derived neurotrophic factor (GDNF)-GDNF receptor- (GDNFR-)-Ret signaling complex is among the best candidates to mediate the early inductive events of renal organogenesis. GDNF is a mesenchyme-derived secreted signal, considered to be the inducer of ureteric bud formation and required for growth and branching of the ureteric bud (30, 43). Ureteric epithelial cells mediate their response to GDNF via activation of the receptor tyrosine kinase Ret, which is expressed in the ureter tips (10, 42). However, GDNF needs first to bind a cell-surface-associated protein, GDNFR-, to interact with Ret (32, 41). The role of exogenous GDNF on in vitro nephron formation on wild-type metanephros remains to be clarified because contradictory data on ureteric bud morphogenesis have been reported (29, 30, 44).

During embryonic development, retinoids play a fundamental role in early differentiation, morphogenesis, and pattern formation in vertebrates (9, 24). Evidence for a developmental retinoid requirement during urogenital tract development originally came first from analysis of the offspring of severely vitamin A-deficient rats (46). Severe congenital malformations occur in such rats, frequently including renal abnormalities (48). These can be reversed by vitamin A treatment at various times during gestation (47). These morphogenetic abnormalities were recapitulated with murine embryos having null mutations for two retinoid nuclear receptor isoforms and exhibiting renal hypoplasia or agenesis (25). However, these experiments did not specify the underlying mechanisms of RA environment on nephron ontogeny. Recently, using a model of in vitro renal development, we showed that RA is a modulating factor controlling renal organogenesis (45). We also demonstrated that the branching pattern of the ureteric bud was significantly more developed on retinoid exposure (45).

Up to now, the effect of all-trans-retinoic acid (RA) on gene expression during renal development has received little attention. In this study, we identify some of the molecular targets of the retinoids in the developing rat kidney. Using paired metanephroi grown in serum-free medium, either RA-free or supplemented with RA, we focus our analysis on the expression of the GDNF-GDNFR--Ret complex components. We also study the effects of exogenous GDNF on nephron formation. Our results demonstrate that c-ret mRNA expression is regulated by RA in a dose-dependent manner, highlighting the control that vitamin A derivatives exert on the nephron mass.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Metanephros organ culture. Metanephros organ culture was performed as previously described (1, 14). Fetuses from Sprague-Dawley female rats of known mating date (day 0 of pregnancy was the day after overnight mating) were taken at embryonic day 14 (E14), and whole metanephroi were collected and freed of exogenous tissue. Kidney rudiments were placed onto a 0.8-µm Millipore AA filter (Millipore, Saint-Quentin-en-Yvelines, France), floating on a defined serum-free medium and incubated for 2, 4, or 6 days in 35-mm petri dishes at 37 ± 0.5°C in a humidified incubator (5% CO₂). The medium originally described by Avner and colleagues (1, 2) was used. Culture medium was changed daily. No antibiotic was present throughout the experiment, because we previously showed that they alter in vitro nephrogenesis via impaired ureteric bud branching morphogenesis (13).

Special care was taken to prepare stock solutions of RA. Ten-millimolar RA solution was prepared in absolute ethanol, stored at 20°C in the dark, and used within 1 wk. RA was used at a concentration ranging from 1 nM to 10 µM, extemporaneously prepared. Experiments were performed under paired conditions; i.e., one metanephros was grown as a control, and the opposite kidney from the same fetus was grown in the RA-supplemented medium. In some experiments, RA was added after 48 h of culture. Human recombinant GDNF (R & D Systems) was used at 15 or 150 ng/ml. All tissue culture reagents were from Sigma France.

In vitro differentiation assessed by lectin histochemistry. The effect of GDNF on in vitro renal development was studied either on pairs of metanephroi grown in absence of RA or on pairs of metanephroi grown with 100 nM RA in the culture medium for 6 days. Briefly, metanephroi were fixed with 2% paraformaldehyde in PBS, permeabilized with saponin, treated with neuraminidase, and labeled with rhodamine-coupled Arachis hypogaea agglutinin that binds podocyte membranes, as already described (14). The total number of nephrons present within the cultured metanephroi was then accurately determined.

RT-PCR. Due to the small size of the samples and thus to the very low amounts of RNA recoverable from kidney rudiments in vitro, we analyzed gene expression by RT-PCR. Sixty metanephroi were collected on E14, and 30 or 15 were pooled after 2 or 4 days of culture, respectively. Messenger RNAs were extracted by using Dynabeads mRNA purification kit (Dynal, Oslo, Norway), as recommended by the supplier, and were quantified at A₂₆₀ by spectrophotometry.

Single-stranded cDNAs were synthesized as follows: 0.1 µg of mRNA in 10 µl, containing 5 µM of random hexanucleotides, was heated at 70°C for 5 min. After cooling, 10 µl of an RT mixture was added to final concentrations of 50 mM Tris · HCl (pH 8.3), 75 mM KCl, 3 mM MgCl₂, 10 mM dithiothreitol, 0.6 mM of each dNTP (Stratagene), together with 20 U RNase inhibitor (Promega) and 200 U Moloney murine leukemia virus-RT (GIBCO-BRL). After incubation at 37°C for 1 h, the reaction was stopped at 90°C for 5 min to inactivate the transcriptase enzymes and cooled at 4°C. Reaction tubes without RT were prepared to control for the absence of genomic DNA contamination.

The transcripts analyzed in this study were the tyrosine kinase receptor c-ret, its proposed ligand GDNF, the accessory protein GDNFR-, and -actin as control. They were studied by amplification of the obtained cDNA using different primer pairs, chemically synthesized (Genosys, Cambridge, UK), that are listed in Table 1. C-ret-specific primers were chosen on the basis of the nucleotide sequence of the mouse c-ret (GenBank accession no. X67812) (19). Oligonucleotide primers were devised to amplify the region 1,578-2,018 encoding the transmembrane domain and most of the cysteine-rich domain. No cross-reaction of these primers with known sequences was found using BLAST/NCBI software. The GDNF primers we used were based on the rat nucleotide sequence (GenBank accession no. L15305) and located at positions 141-160 and 584-605 (34). GDNFR- primers were selected according to the rat nucleotide sequence (GenBank accession no. U59486) and located at positions 1,177-1,201 and 1,598-1,622, as suggested by P. Towers (personal communication). The primers of the internal control were sequences 1,509-1,528 and 2,457-2,476 of the rat -actin sequence (GenBank accession no. J00691) (27).

One-half of the RT reaction volume was added to 90 µl of the following PCR Mastermix prepared immediately before use: 75 mM Tris · HCl (pH 9), 20 mM (NH₄)₂SO₄, 0.01% (wt/vol) Tween 20, 200 µM of each dNTP, and 2 U Taq DNA polymerase (Pro-HA, Eurogentec). Primers (2 µl) were added to a final concentration of 0.5 µM. The final MgCl₂ concentration was 4.5, 1.5, and 1 mM for c-ret, GDNF, and GDNFR-, respectively, and 1.5 mM for -actin. The reaction mixture was overlaid with 100 µl mineral oil and then placed in a thermal cycler (Perkin-Elmer) programmed as follows: first a 3-min step at 95°C, then n cycles [95°C (1 min), annealing temperature (1 min), 72°C (1 min)] (see Table 1). PCR amplification was followed by a 7-min step at 72°C. For each amplification, the number of cycles was chosen such that the reaction products do not reach the plateau phase. (As a general rule, a plateau is reached when 5 additional cycles do not yield at least twice as much product.) The RT products solution was diluted 10 times before amplification of GDNF and GDNFR-, routinely performed with 30 cycles. Undiluted RT products solution was used for 35 cycles of c-ret amplification. Plateau is reached after 40 cycles for GDNF and GDNFR-, but not for c-ret. Ten microliters of amplified products were run on a 2% agarose gel containing ethidium bromide and molecular markers. The size of PCR products was analyzed and their identity confirmed by restriction enzyme digestion. The PCR products were visualized by ultraviolet transillumination and photographed using 667 Polaroid films.

Semiquantitative analysis of c-ret expression by PCR. To study the effect of a wide range of RA concentrations on c-ret gene expression, we performed simultaneous amplification of c-ret and -actin cDNA. The latter was used as an internal control, and addition of its primers was delayed, as proposed by Kinoshita et al. (20). After RT, c-ret and -actin cDNAs were amplified as follows: primers for c-ret sequence were added to the PCR reaction mixture as described above. After 10 cycles of amplification, primers for -actin were added for 22 additional cycles. The PCR products were visualized as just described. Coamplification of c-ret and -actin proceeds exponentially, and semiquantitative analysis was performed in the logarithmic phase of amplification (5). Intensity of the bands was analyzed following scanner densitometry using image analysis software (NIH Image).

Western blotting analysis. Twenty-five metanephroi were collected on E14, and 10 or 5 were pooled after 2 or 4 days of culture with or without RA. After homogenization in 50 µl of lysis buffer (Tris · HCl 50 mM, pH 8.6 containing 5 mM EDTA and protease inhibitors), samples were boiled in Laemmli buffer. Twenty micrograms of each homogenate were loaded on top of a 6-20% gradient gel, electrophoretically separated under reducing conditions, and blotted onto nitrocellulose membrane (Hybond-C extra, Amersham). Membranes were washed in Tris-buffered saline (TBS; 50 mM Tris · HCl, 150 mM NaCl, pH 8) and incubated in 5% skimmed milk in TBS overnight. Then they were incubated in 0.2 µg/ml primary antibody rabbit anti-Ret (Santa-Cruz Biotechnology) for 1 h in TBS + Tween 20 at 0.05%. After washing, membranes were incubated in secondary antibody donkey anti-rabbit and probed with rabbit peroxidase anti-peroxidase soluble immune complexes (Jackson Immunoresearch) used at a 1:5,000. The signal was visualized by enhanced chemiluminescence (Amersham kit) using the recommended protocol.

Data analysis. Data are reported as means and SE. Comparisons between control and GDNF-exposed metanephroi were performed using the Wilcoxon's paired test. Significance was determined by P < 0.05.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Analysis of the transcripts coding for Ret revealed products of roughly the expected size (439 bp). However, based on the mouse sequence, Msp 1 enzymatic treatment led to irrelevantly sized products, and Pvu II was ineffective. The PCR product was sequenced (Eurogentec), and the nucleotide sequence of this fragment is illustrated in Fig. 1. It displays 89% similarity with the mouse homologous c-ret cDNA and had a three-nucleotide insert (TGT) in position 1,928, giving a length of 442 bp. Most of the differences between the mouse and rat sequences are found within the transmembrane domain (~68% homology). The sequence of the cysteine-rich domain is highly conserved, with 94% homology with the mouse sequence (19). For the corresponding protein, amino acid sequence comparison reveals 88% homology between mouse and rat Ret proteins (Fig. 2). Note that the additional amino acid we detected was a valine, a residue that is present at this position in both human and chicken Ret proteins (37, 40). All cysteine residues are conserved, and none of the amino acids known to be mutated in multiple endocrine neoplasia syndromes or Hirschsprung's disease are modified in the amplified rat c-ret sequence (12). Thus the amplified fragment corresponds to rat c-ret transcripts.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Comparison of mouse c-ret cDNA and amplified rat fragment. Match between nucleotides is marked by a dot. A high degree of homology between mouse and rat sequences was found. In rat, 3 nucleotides (TGT) were inserted in position corresponding to base 1,928 of mouse cDNA. Msp 1 and Pvu II restriction enzymes sites are indicated. No Pvu II site is present in this partial rat sequence.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Comparison of mouse and rat protein partial sequences. Transmembrane domain is represented by shaded area, and cysteine residues located in extracellular domain are enclosed by boxes. Identical amino acids are identified by a single dot. Asterisks above some amino acids represent mutation sites in multiple endocrine neoplasia syndromes or Hirschsprung's disease. Insertion of nucleotides results in addition of a valine (V) residue in transmembrane domain at position 344.

The profile of c-ret expression in E14 kidney before and during organ culture is depicted in Fig. 3. After 2-4 days of culture of metanephros in a RA-free medium, c-ret expression in the metanephros decreases significantly. By contrast, in the presence of 100 nM of RA in the culture medium, c-ret is strongly expressed throughout the culture period, even reaching a level of expression higher than the level at the time of explantation. To determine whether RA is able to upregulate c-ret expression, metanephroi were grown for 2 days without RA so as to lower c-ret expression, and then RA was added to the culture medium for the third and fourth day of in vitro development (Fig. 3). C-ret expression is markedly stimulated under these conditions. Western blot analyses directed against Ret proteins yield the same results (Fig. 3).

View larger version (46K): [in this window] [in a new window]   Fig. 3.   Product analysis of RT-PCR for c-ret mRNA and Western blotting. Pairs of metanephroi were grown in absence () or presence (+) of 100 nM RA for 2 or 4 days (d). In some experiments, all-trans-retinoic acid (RA) was added during the 3rd and 4th days (3-4d). Two percent agarose gel stained with ethidium bromide revealed c-ret (442 bp) and -actin control (504 bp) bands. Western blot of Ret reveals one band at ~170 kDa, as expected. RA stimulates both c-ret mRNA and protein amounts. E14, embryonic day 14.

Because we previously demonstrated that RA was able to promote nephron formation in organ culture in a dose-dependent manner via enhanced branching morphogenesis (45), we subjected metanephroi cultures to a wide range of RA concentrations over a period of 4 days to analyze c-ret expression levels. As shown in Fig. 4, the presence of an RA concentration as low as 1 nM was able to significantly sustain a high level of c-ret expression, compared with paired controls. The c-ret mRNAs augmented in a dose-dependent manner, linearly from 10⁹ to 10⁶ M RA (r = 0.82, P < 0.001, n = 12). The highest expression of c-ret was found in metanephroi cultured with 1 µM of RA, leading to an eightfold increase compared with metanephroi grown in the absence of RA. A concentration of 10⁵ M RA caused c-ret mRNA levels to drop.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Semiquantitative RT-PCR of c-ret expression. Pairs of metanephroi were grown for 4 days in absence or presence of RA concentration ranging from 109 to 105 M. A: after simultaneous amplification, detection of both -actin (top) and c-ret (bottom) bands were visualized in 2% agarose gel stained with ethidium bromide. Quantification of the relative amount of c-ret to -actin is presented in B. A linear increase of c-ret expression is observed in response to increasing RA concentration in the medium (log scale). Ratio was normalized to value obtained with 1 nM of RA. Experiments were done in triplicate.

Concerning GDNF and GDNFR- expression, their mRNAs were present at very high levels in E14 kidneys compared with c-ret. In vitro development did not modify their expressions, irrespective of the presence of RA in the culture medium (Fig. 5). The amount of GDNF peptides remained unaffected by the duration of the culture and by the presence or absence of RA (data not shown).

View larger version (47K): [in this window] [in a new window]   Fig. 5.   RT-PCR analysis of glial cell line-derived neurotrophic factor (GDNF) and GDNF receptor- (GDNFR-) mRNAs. Pairs of metanephroi were grown in absence () or presence (+) of 100 nM RA for 2 or 4 days. Two percent agarose gel stained with ethidium bromide revealed GDNF (464 bp) and GDNFR- (445 bp) bands.

We studied the effect of exogenous GDNF on rat metanephros development in vitro in cultures grown with or without RA. The total number of nephrons present within the explanted metanephroi was quantified after 6 days of culture. Addition of 15 ng/ml of GDNF to the standard culture medium had no effect on in vitro nephron formation. However, the presence of a 10-fold higher concentration of GDNF induced a 35% stimulation of in vitro nephrogenesis (65 ± 4 vs. 90 ± 6, n = 15 pairs, P < 0.01) (Fig. 6, A and B). In the presence of 100 nM of RA, a concentration known to induce per se a 200% increase in nephron formation in vitro (45), the concomitant presence of 150 ng/ml of GDNF allowed the formation of 33% supernumerary nephrons (179 ± 9 vs. 231 ± 9, n = 8 pairs, P < 0.02) (Fig. 6, C and D).

View larger version (111K): [in this window] [in a new window]   Fig. 6.   Microphotographs showing metanephroi differentiation in vitro assessed by lectin histochemistry. Pairs of metanephroi were grown under control (A and B) or RA-stimulated conditions (C and D), in absence (A and C) or presence (B and D) of 150 ng/ml GDNF in the medium. A moderate increase in nephron number had occurred with GDNF, far from the stimulation observed with RA alone. Bar represents 150 µm.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Whole metanephros organ culture has emerged as an attractive system for studying the underlying mechanisms of renal organogenesis because it allows optimal metanephros differentiation. Here we demonstrate that the c-ret tyrosine kinase receptor expression is highly dependent on the presence of RA. Moreover, c-ret mRNA level is regulated by RA concentrations in a dose-dependent manner. The increased abundance of Ret proteins by RA is consistent with the stimulated branching morphogenesis we observed leading to enhanced nephron formation (45), i.e., the greater the number of ureter tips, the greater the number of sites to induce differentiation of the metanephric mesenchyme into nephrons. Therefore, not only is c-ret involved in the initial outgrowth of the ureter from the wolffian duct (35), but also its level of expression correlates with the number of nephrons formed in vitro (45).

At the time we explanted the embryonic kidney, the ureteric bud had few extremities. Whereas c-ret is expressed all over the T-shaped bud at an earlier stage, c-ret is only expressed at the tips of the ureter at the moment of explantation (28). This contributes to c-ret's relatively low level of expression compared with GDNF or GDNFR-, which, in turn, are highly expressed in the metanephrogenic blastema (17, 38, 41). Analysis of Ret proteins clearly indicates that more tyrosine kinase receptors are present on the ureteric bud, which allows a more effective response to the high level of endogenous GDNF. Because the dichotomous branching pattern of the ureteric bud was maintained in RA-exposed metanephroi (45), we speculate that Ret proteins are not ectopically expressed, but further investigations are needed to clarify this point. It remains an open question whether RA acts primarily on the bud itself or whether an enhanced production of mesenchymal signals by RA is the leading event that stimulates the branching pattern. There may be a direct effect of RA on c-ret expression: the nucleotide sequence of the c-ret promoter region thus far identified (453 bp) does not show any RA-responsive elements (18), but this does not rule out the possibility that c-ret expression is directly controlled by RA because RA-responsive elements can be located further upstream from the transcription start sites. It is significant that the dose-dependent response we measured clearly indicates that lowering or increasing the RA concentration below or above some value may impair renal organogenesis via inadequate Ret expression. Upregulation of c-ret expression by RA exposure has also been reported in neuroblastoma cells before neuronal differentiation, but only for very high RA concentrations (7, 39). It is very unlikely that c-ret is the only target of RA in the embryonic kidney because numerous potentially RA-responsive genes are known to be expressed in the developing kidney (4). But, so far, c-ret is the first identified gene present in the ureteric bud that may play a role in controlling the number of nephrons to be induced.

Given the strong stimulation of in vitro nephron formation on RA exposure we observed (Fig. 5C; Ref. 45) and the RA-dependent Ret expression we have reported here (Fig. 4), we would have expected a much more pronounced stimulation of in vitro nephrogenesis on simultaneous exposure to RA and GDNF. The most likely explanation of the observed moderate response to large concentrations of exogenous GDNF, and as also reported for branching morphogenesis of the ureteric bud by others (44), is that the amount of endogenous GDNF was not limiting, as suggested by the presence of abundant GDNF transcripts and proteins. In this regard, it should be noted that although GDNF is able to promote primary ureteric buds from the wolffian duct (29), Sainio et al. (30) have shown recently that exogenous GDNF is not critical for ureteric bud branching of late T-shaped bud. It is clear that there is an absolute requirement for a sufficient amount of GDNF produced by the mesenchymal cells to allow the ureter to bud out from the wolffian duct leading to metanephros formation. The occurrence of unilateral renal agenesis in GDNF heterozygous deficient embryos supports this view (26, 31). It seems, however, that later in metanephros development, the levels of endogenous GDNF remain high, as they do in vitro, suggesting that it may act as a ureteric bud survival factor. Despite the unchanged amount of GDNF in the metanephros throughout the culture period, c-ret expression decreased sharply, indicating that GDNF does not modulate c-ret expression. In addition, the weak GDNF responsiveness cannot rely on a limiting amount of GDNFR- because GDNFR- expression remained unchanged in cultured metanephroi regardless of whether RA was present. The same results were obtained using the rat instead of human recombinant GDNF (data not shown).

The question of whether GDNF is the more potent ligand to induce c-ret signaling in the embryonic kidney remains open. The fact that superior cervical ganglion neurons do not survive in Ret / mice whereas they are almost unaffected in GDNF / mice supports this hypothesis (11, 31, 35). This suggests that the Ret receptor can be activated by other signaling molecules. Interestingly, a new neurotrophic factor structurally related to GDNF and named neurturin (NTN) has been identified (22). Its responsiveness, like GDNF, requires a glycosyl-phosphatidylinositol-linked receptor (NTNR-) to induce Ret activation (8, 21). Its role during renal development is not yet established but is likely, given the large amounts of NTNR- detected in the metanephros (3, 6). RA may thus control the effects of a variety of differentiation factors interacting first with a specific glycosyl-phosphatidylinositol membrane-anchored protein before transducing their signals via a shared transmembrane protein kinase receptor such as c-ret.

In conclusion, we demonstrate here that RA stimulates the expression of the proto-oncogene c-ret in serum-free metanephros organ culture, suggesting that the vitamin A environment may play a crucial role in nephron formation. In the cascade of events resulting in the initial interaction between the metanephric blastema and the ureteric bud, the local RA concentrations could be the triggering event leading to c-ret expression, to respond to GDNF or to another as yet uncharacterized signaling molecule. Thus the RA-dependent c-ret expression within the differentiating metanephros may control the nephron mass.