Metanephric mesenchyme contains embryonic renal stem cells

Juan A. Oliver, Jonathan Barasch, Jun Yang, Doris Herzlinger, Qais Al-Awqati

Abstract

Renal epithelial cells derive from either cells of the metanephric mesenchyme or ureteric bud cells, but the origin of other renal cells is unclear. To test whether metanephric mesenchymal cells generate cells other than epithelial, we examined the developmental potential of a metanephric mesenchymal cell line (7.1.1 cells) and of primary cultures of metanephric mesenchymal cells. 7.1.1 Cells express both mesenchymal and epithelial markers and, on confluence, form well-defined monolayers expressing epithelial junctional proteins. However, 7.1.1 cells as well as primary cultures of metanephric mesenchymal cells also generate spindle-shaped cells that are positive for α-smooth muscle actin, indicating that they are myofibroblasts and/or smooth muscle; this differentiation pathway is inhibited by collagen IV and enhanced by fetal calf serum or transforming growth factor-β1. Transforming growth factor-β1also induces expression of smooth muscle proteins, indicating that the cells differentiate into smooth muscle. 7.1.1 Cells as well as primary cultures of metanephric mesenchymal cells also express vascular endothelial growth factor receptor 2 and Tie-2, suggesting that the metanephric mesenchymal cells that generate epithelia may also differentiate into endothelial cells. The pluripotency of the 7.1.1 cells is self-renewing. The data suggest that the metanephric mesenchyme contains embryonic renal stem cells.

  • nephrogenesis
  • cell differentiation
  • myofibroblast
  • transforming growth factor-β
  • collagen IV

cell division is infrequent in the adult kidney, but this organ has the capacity to regenerate and there is abundant cellular proliferation during recovery from diseases, such as acute tubular necrosis (35). The origin of newly generated renal cells is largely undefined, but by analogy to other organs, organ-specific pluripotent cells (i.e., renal stem cells) are likely precursors of new cells. Although several strategies may be attempted to identify adult renal stem cells (1), we reasoned that such cells would share molecular and functional characteristics with embryonic renal stem cells. However, unlike blood cells, all of which derive from a single bone marrow stem cell, the origin of the different lineages in solid organs, such as the kidney, is obscure, and it is unknown whether the embryonic kidney contains organ-specific pluripotent cells that generate the many cell types present in the adult kidney.

Renal embryogenesis starts when a branch of the Wolffian duct called the ureteric bud invades an aggregate of mesenchymal cells called the metanephric mesenchyme. Reciprocal interactions between the metanephric mesenchyme and the ureteric bud result in growth and branching of the duct while the mesenchyme grows and converts to epithelia. The ureteric bud cells give rise to the epithelia of the collecting duct, and the metanephric mesenchymal cells give rise to the epithelia of the rest of the nephron.

Epithelia are the most prominent cells in the adult kidney, but this organ contains many other cell types, including mesangial, vascular smooth muscle, endothelial, and renal medullary interstitial cells, as well as myofibroblasts, fibroblasts, macrophages, and neurons. Of all these nonepithelial cells, and likely because nephron capillaries have such a critical role in nephron function, the potential origin of renal endothelial cells has long attracted attention. Embryonic kidneys grown in vitro do not develop capillaries, and when the developing kidneys of mice were grown on quail allantoic membranes, glomeruli were vascularized by quail endothelia, suggesting that renal endothelial cells may be exogenous to the kidney (33). However, recent studies by Robert et al. (30, 31) demonstrated that when embryonic kidneys are first grown in vitro and then grafted into the anterior eye chamber of a normal animal, the kidney develops endothelial cells (including glomerular capillaries) that are of graft origin. These studies suggest that embryonic kidneys contain endothelial cell precursors (angioblasts). However, it is not clear whether these angioblasts invade the embryonic kidney from outside, are a mesenchymal population of cells distinct from the metanephrogenic mesenchymal cells that will became renal epithelia, or derive from some distinct progenitor cell. Even less is known about the origin of the renal vascular smooth muscle and mesangial cells, and during kidney development, cells with smooth muscle characteristics are generated much later than nephrons and capillaries (5).

Except for epithelia and endothelia, the developmental origin of other renal cells remains little explored, and as an initial approach, three possibilities can be postulated: 1) there is migration of fully differentiated cell types into the developing kidney;2) the embryonic kidney contains progenitor cells, each giving rise to an adult organ-specific cell type; and 3) the embryonic kidney contains progenitor cells with the ability to generate many cell types (i.e., multipotent progenitors or embryonic renal stem cells).

To test whether the embryonic kidney possesses organ-specific progenitor cells capable of differentiating into several cell types, we examined the differentiation potential of cells from rat metanephric mesenchymes. The mesenchymes were isolated at day 13 of embryonic age (E13), the day when, in the rat, the ureteric bud invades the metanephric mesenchyme, inducing its conversion to epithelia and starting renal development. We used an immortalized cell line of metanephric mesenchymal cells and primary cultures of these cells. The results indicate that metanephric mesenchyme contains cells that are pluripotent and can differentiate, in addition to epithelia, into myofibroblasts and smooth muscle cells and likely also into endothelial cells.

METHODS

Cell lines.

The metanephric mesenchymal cell line (7.1.1 cells) has been previously described (17, 18). Cells were grown in MEM containing 10% fetal calf serum (FCS), 2 mM glutamine, and penicillin plus streptomycin and incubated at 37°C in a 5% CO2atmosphere. When cultures were passed, cells were dissociated with 0.5% trypsin in calcium- and magnesium-free HBSS with 0.02% EDTA. The cells were also cultured in the absence of FCS (serum-free media), in DMEM-Ham's F-12 supplemented with 5 μg/ml insulin and transferrin, 5 ng/ml sodium selenite, and 20 ng/ml of dexamethasone andl-thyroxine (all from Sigma) plus glutamine and antibiotics as described above. For passage, cells cultured in serum-free media were dissociated with calcium- and magnesium-free HBSS containing 0.02% EDTA without trypsin.

Primary cultures of metanephric mesenchymal cells were obtained from rat E13 mesenchymes dissociated as described (3) and grown in either of the two culture media described for the 7.1.1 cells but always supplemented with 50 ng/ml basic FGF and 10 ng/ml transforming growth factor (TGF)-α.

The cell line derived from mouse ureteric bud cells and the rabbit intercalated cell line (Clone C) have been previously described (2, 9). The mouse endothelial pancreatic islet cell line (MS-1) was obtained from American Type Culture Collection.

RT-PCR.

RNA from cells was isolated and transcribed as previously described (26), and PCR was performed with the primers shown in Table 1.

View this table:
Table 1.

Primers used for RT-PCR

Antibodies and reagents.

The following antibodies were used: monoclonal to α-smooth muscle actin (α-SMA), clone 1A4 (Sigma); monoclonal to calponin, clone M3556 (DAKO); monoclonal to E-cadherin, clone 36 (Transduction Laboratories); polyclonal to cytokeratin (Zymed Laboratories); polyclonal to neural cell adhesion molecule (NCAM; Chemicon); monoclonal to myosin light chain kinase (MLCK), clone K36 (Sigma); monoclonal to rat endothelial cell antibody 1, clone HIS52 (Serotec); Tie-2 polyclonals (RDI and Alpha Diagnostic, no. 1), and monoclonal Ab33 (the gift of Dr. C. Kontos); monoclonal to smooth muscle myosin heavy chain, clone M3558 (DAKO); VEGF receptor 2 [VEGFR-2; R&D Systems and Cruz (SC505 and SC315)]; polyclonal to vimentin (Chemicon); polyclonal to WT-1 (Santa Cruz); and polyclonal to ZO-1 (Zymed Laboratories).

Mouse collagen IV and Matrigel were from Becton Dickinson; r-human TGFβ1, r-human basic FGF, and r-human TGF-α were from R&D Systems; Dil-Ac-LDL was from Biomedical Technologies; and Sytox green was from Molecular Probes.

Immunodetection.

Immunofluorescence was done in methanol-fixed cells or tissue sections as described (26, 27). Immunohistochemistry was done in methanol-fixed cells, serially blocked with 3% H2O2, 10% FCS/2% donkey serum (both dialyzed), and avidin/biotin (Vector Laboratories). After blocking, sections were incubated with the corresponding primary antibodies, and after washing, sections were incubated with a biotin-linked donkey secondary antibody. Avidin-linked horseradish peroxidase was next added with the ABC kit (Vector Laboratories), and peroxidase activity was localized with diaminobenzidine (DAB Substrate Pack, Biogenex) and enhanced by diaminobenzidine-enhancing solution (Vector Laboratories). Slides were finally counterstained with hematoxylin.

Immunobloting was done in protein homogenates as previously described (26).

Cloning experiments.

7.1.1 Cells were dissociated with trypsin, diluted, and plated at a density of ∼30 cells/35-mm tissue culture dish. The cells were grown in MEM containing 10% FCS and supplemented with 2 mM glutamine and antibiotics. Five days later, individual clones were stained with α-SMA and ZO-1 antibodies. In addition, individual clones were isolated with clonal rings and again cloned. These secondary clones were expanded and analyzed for their ability to differentiate into epithelia and into myofibroblast/smooth muscle.

RESULTS

Metanephric mesenchymal cells.

To examine the differentiation potential of metanephric mesenchymal cells, two preparations of cells were used. Most experiments were performed with a clonal cell line (7.1.1 cells) of metanephric mesenchymal cells. This cell line was generated as previously described (17, 18) and cloned by serial dilution. Briefly, metanephric mesenchymes from embryos at E13 were dissected from the ureteric bud and trypsin dissociated, and the cells were immortalized by the introduction of SV40 large T antigen using ts58 retrovirus. Transfected cells were selected with G4198. A clone was established and named 7.1.1 cells. Southern blots for the large T antigen in three subclones showed identical insertion sites, confirming that the line derives from a single clone (not shown).

Experiments were also performed with primary cultures of rat metanephric mesenchymal cells isolated from E13 embryos, the same age in which the 7.1.1 cells were isolated.

Expression of mesenchymal and epithelial markers in 7.1.1 cells.

RT-PCR of 7.1.1 cells revealed expression of the genes for the mesenchymal proteins vimentin, NCAM, and WT-1 (Fig.1 A); the latter two are essential for kidney development (7, 21). In addition, the cells also expressed the genes for E-cadherin, Pax-2, and Wnt-4, all epithelial proteins with important roles during nephrogenesis (21, 29). Surprisingly, despite the fact that the cell line derives from metanephric mesenchymes obtained the day that rat renal development starts, RT-PCR was also positive for mRNA of several genes characteristic of fully differentiated renal epithelia, such as megalin, aminopeptidase A, the ClC-5 chloride channel, and the thiazide-sensitive Na-Cl cotransporter.

Fig. 1.

Expression of mesenchymal and epithelial markers in 7.1.1 cells. A: RT-PCR with RNA from 7.1.1 cells showed the expression of mesenchymal and epithelial genes, including genes characteristic of well-differentiated renal epithelia.B: RT-PCR with RNA from 7.1.1 cells did not detect the message for the transcription factor BF-2, whereas it was easily detected in RNA from day 17 of embryonic age (E17) rat kidneys. C: immunofluorescent microscopy of 7.1.1 cells showed that these cells synthesize mesenchymal (a–c) and epithelial proteins (d–f). Bar = 21 μm. D: electron micrograph of monolayers of metanephrogenic mesenchymal cell lines. A cell line not expressing epithelial phenotype (a) contrasts with the 7.1.1 cells (b). Note that only the 7.1.1 cells have a distinct apical tight junction as well as well-defined apical (AP) and basolateral (B) domains, the former with many microvilli. Arrowheads, intercellular space between two cells. NCAM, neural cell adhesion molecule.

In addition to mesenchymal cells that are epithelial precursors, metanephric mesenchyme contains a distinct population of cells called stroma that express the transcription factor BF-2 (15). Although RT-PCR detected BF-2 transcripts in RNA from E17kidneys, transcripts were not detected in 7.1.1 cells (Fig.1 B), indicating that the cells are not stroma cells.

Imnunofluorescent microscopy of the 7.1.1 cells (Fig. 1 C) revealed that in addition to mesenchymal proteins (vimentin, WT-1, and NCAM), the cells expressed markers of well-differentiated epithelia, such as cytokeratin, and formed monolayers with tight junctions mediated by ZO-1 and adherent junctions formed by E-cadherin. In contrast to another clonal cell line also derived from metanephric mesenchyme but without epithelial characteristics (Fig. 1 D, a), confluent 7.1.1 cells (Fig. 1 D, b) demonstrated, in addition to well-developed tight junctions, distinct basolateral and apical domains, the former with many microvilli and the latter with extracellular matrix. These results indicate that the 7.1.1 cells are metanephric mesenchymal cells that differentiate into epithelia, likely because of ureteric bud induction (34). Hence, the 7.1.1 cells are metanephric mesenchymal epithelial precursors.

Expression of smooth muscle and endothelial genes in 7.1.1 cells.

To examine the possibility that 7.1.1 cells may generate progeny other than epithelial cells, we examined them for expression of genes characteristic of vascular cells (Fig.2 A). mRNA for several proteins characteristic of smooth muscle, including α-SMA, calponin, and tropoelastin, were detected. The figure further shows that mRNA for several proteins characteristic of endothelial or endothelial cell precursors [VEGFR-2, Tie-2, CD31, and von Willebrand factor (vWF)] was also detected. Because the expression of endothelial genes was somewhat unexpected, to rule out the possibility that it was due to the transformation of the cells, two other immortalized renal cell lines were examined for the presence of VEGFR-2 and Tie-2. Although RT-PCR in 7.1.1 cells easily detected the mRNA for these two receptors, a mouse ureteric bud cell line (2) and a rabbit collecting duct cell line (9) transformed with the same virus used for 7.1.1 cells were both negative (Fig. 2 B).

Fig. 2.

RT-PCR for smooth muscle and endothelial genes in 7.1.1 cells. A: RT-PCR with RNA from 7.1.1 cells showed expression of genes characteristic of smooth muscle and endothelial cells.B: although RT-PCR easily detected the endothelial-specific receptor tyrosine kinases VEGF receptor 2 (VEGFR-2) and Tie-2 in 7.1.1 cells, neither of these genes was detectable in two other cell lines of renal origin, ureteric bud (UB) cells and intercalated cells. α-SMA, α-smooth muscle actin.

Differentiation of metanephric mesenchymal cells into myofibroblasts and/or smooth muscle.

The presence of mRNA for several smooth muscle genes in 7.1.1 cells raised the possibility that these cells, in addition to giving rise to epithelia, may also be capable of differentiating into myofibroblast and/or smooth muscle. To explore this, we first examined for the presence of α-SMA by immunofluorescent microscopy. We found that when the cells were grown at low density, individual cells expressed α-SMA as well as the epithelial marker ZO-1 (Fig.3 A, a) . However, in confluent cultures, the α-SMA-positive cells became spindle shaped, no longer expressed ZO-1 (Fig. 3 A, b), and, by changing the depth of field, could be seen to grow on top of the monolayer of cells expressing ZO-1. In contrast to the 7.1.1 cells, cells from the ureteric bud cell line did not express α-SMA. These results indicate that although most 7.1.1 cells differentiate into epithelia, they can also differentiate into myofibroblasts and/or smooth muscle.

Fig. 3.

Expression of α-SMA in 7.1.1 and metanephric mesenchymal cells.A: at low density (a), individual 7.1.1 cells may express both α-SMA (rhodamine) and ZO-1 (FITC; bar = 21 μm), whereas, at confluence (b), cells expressing α-SMA are negative for ZO-1 and locate on top of cells that form a monolayer, are ZO-1 positive, and α-SMA negative (bar = 63 μm). B: E13 rat renal anlage (arrows; a) does not express α-SMA, whereas this protein is easily detectable in a fragment of aorta concomitantly isolated (double arrow). Many metanephric mesenchymal cells isolated fromE13 rat kidney and grown overnight express α-SMA (b). Bars = 63 μm.

7.1.1 Cells derive from E13 metanephric mesenchyme, and at this age, α-SMA is not detectable in rat developing kidney (Fig.3 B, a). Indeed, α-SMA-positive cells are only detected in rat embryonic kidney 3 days later (5), indicating that the appearance of myofibroblasts and/or smooth muscle cells does not start until nephrogenesis is well advanced. Hence, to determine whether the ability of 7.1.1 cells to generate both epithelia and myofibroblast/smooth muscle cells is an intrinsic capacity of metanephric mesenchymal cells rather than the result of the transformation of the 7.1.1 cell line, we isolated primary cultures ofE13 rat metanephric mesenchymes. Confirming previous results (39), on isolation and overnight culture, many of the metanephric mesenchymal cells were positive for α-SMA (Fig.3 B, b), indicating that in vitro these cells differentiate into myofibroblasts and/or smooth muscle cells. Of note, this occurred whether the cells were grown in the presence of FCS or in serum-free media.

Regulation of metanephric mesenchymal cell differentiation into myofibroblasts and/or smooth muscle.

Because E13 kidneys do not contain myofibroblasts and/or smooth muscle cells (Fig. 3 B, a), the expression of α-SMA by 7.1.1 cells and primary cultures of metanephric mesenchymal cells in vitro could be due to one or the combination of two factors. First, there may be factors in the embryonic kidney that inhibit differentiation of metanephric mesenchymal cells into a myofibroblast and/or smooth muscle fate (i.e., α-SMA expression). Second, there may be factors in vitro that facilitate differentiation of mesenchymal cells into this fate. As the metanephric mesenchyme initiates differentiation into epithelia, there is loss of collagens I and III and appearance of collagen IV (10) so that mesenchymal-to-epithelial conversion and collagen IV expression are closely linked (10, 22). Hence, we postulated that collagen IV may facilitate mesenchymal conversion to epithelia, thereby preventing differentiation of induced metanephric mesenchymal cells into myofibroblasts and/or smooth muscle cells. Indeed, immunoblots demonstrated that when 7.1.1 cells were grown on collagen IV for 7–10 days, α-SMA became undetectable (Fig.4 A). Similarly, immunofluorescence of 7.1.1 cells (Fig. 4 A, a) and of primary cultures of metanephric mesenchymal cells (Fig. 4 A, b) grown on collagen IV revealed that all cells expressed ZO-1 and that cells expressing α-SMA disappeared from the cultures. These results indicate that collagen IV prevents differentiation of induced metanephric mesenchymal cells into myofibroblasts and/or smooth muscle cells.

Fig. 4.

Regulation of 7.1.1 and metanephric mesenchymal cell differentiation. A: immunoblots detected α-SMA in 7.1.1 cells grown on cell culture dishes in MEM with 10% fetal calf serum (FCS), but the protein became undetectable in 7.1.1 cells grown on dishes coated with 10 μg/cm2 of collagen IV (30 μg of protein/lane). Monolayers of 7.1.1 cells (a; bar = 63 μm) and primary cultures of metanephrogenic mesenchymal cells (b; bar = 21 μm) were seeded on collagen IV and stained with ZO-1 (FITC) and α-SMA (rhodamine). α-SMA-positive cells became exceedingly rare.B: immunoblots easily detected α-SMA in 7.1.1 cells grown on cell culture dishes in MEM with 10% FCS, but the protein became undetectable in 7.1.1 cells grown in serum-free media (30 μg protein/lane). 7.1.1 Cells were dissociated without trypsin and grown overnight in serum-free media without (a) and with 4 ng/ml of transforming growth factor (TGF)-β1 (b) and stained for α-SMA (rhodamine). Nuclei were stained with 10 nM Sytox green. Bar = 127 μm. C: immunoblots of 7.1.1 cells grown in serum-free media in the absence or presence of 4 ng/ml TGF-β1 (20 μg protein/lane). D: immunofluorescence of confluent cultures of 7.1.1 cells grown in the presence of 4 ng/ml of TGF-β1. Bars = 21 μm.E: electron micrograph of 7.1.1 cells treated with TGF-β1. Note bundles of thin and intermediated filaments (arrows). MLCK, myosin light chain kinase; SM-MHC, smooth muscle-myosin heavy chain.

To test whether factors present in vitro may facilitate differentiation of the metanephric mesenchymal cells into myofibroblasts and/or smooth muscle cells, we grew 7.1.1 cells in defined serum-free media (seemethods). Although α-SMA was easily detectable in immunoblots from 7.1.1 cells grown in 10% FCS, the protein was undetectable when cells were grown for >1 wk in serum-free media (Fig.4 B). This result suggests that a factor(s) in serum facilitates differentiation of 7.1.1 cells into a myofibroblast and/or smooth muscle phenotype. Of the several factors present in serum capable of regulating α-SMA gene expression, TGF-β1 is the most prominent. This cytokine is an important regulator of vascular smooth muscle differentiation and has been shown to induce 10T1/2 cells to became smooth muscle cells (20). When 7.1.1 cells were grown overnight in serum-free media, very few of the cells were positive for α-SMA (Fig. 4 B, a); the addition of TGF-β1 to these cultures markedly increased the number of cells expressing α-SMA (Fig. 4 B, b), indicating that this cytokine facilitates differentiation of 7.1.1 cells into myofibroblast and/or smooth muscle.

Immunoblots revealed that TGF-β1 induces 7.1.1 cells to differentiate into smooth muscle; TGF-β1 increased the expression of α-SMA and MLCK and induced the synthesis of smooth muscle-specific myosin (Fig. 4 C). Although these results indicate that smooth muscle differentiation is a potential developmental fate of 7.1.1 cells, immunofluorescent microscopy revealed that only occasional cells were positive for smooth muscle markers; in confluent cell cultures, only a few cells were positive for MLCK, smooth muscle myosin, or calponin (Fig. 4 D). In addition, electron microscopy of 7.1.1 cells treated with TGF-β1 showed (Fig. 4 E) occasional cells containing bundles of thin and intermediate filaments similar to those found in smooth muscle cells grown in vitro (32). Thus it appears that at least in vitro, TGF-β1 is capable of inducing differentiation of a small population of 7.1.1 cells into smooth muscle.

Three independent subclones were established, and experiments showed that the ability of 7.1.1 cells to express α-SMA and markers of smooth muscle was self-renewing (not shown).

Endothelial phenotype of metanephric mesenchymal cells.

The presence of mRNA for VEGFR-2 and for Tie-2 in 7.1.1 cells suggested that metanephric mesenchymal cells could, in addition to generating epithelia, myofibroblasts, and smooth muscle cells, also be precursors for endothelial cells. Accordingly, we examined the endothelial phenotype of the 7.1.1 cells. Immunoblots with three antibodies to distinct protein domains of VEGFR-2 demonstrated that similar to the endothelial cell line MS-1, the 7.1.1 cells synthesize VEGFR-2 (Fig.5 A). The blot with antibody SC-315 shows that the peptide used to raise the antibody fully blocked the signal.

Fig. 5.

VEGFR-2 and Tie-2 in 7.1.1 cells and primary cultures of metanephric mesenchymal cells. A: three antibodies to VEGFR-2 directed to three different domains of the mouse protein (amino acids 1–762 of the extracellular domain for R&D, 931–997 for SC505, and the COOH terminal for SC315) all detected the protein in the endothelial cell line MS-1 and 7.1.1 cells. The peptide used to rise antibody SC315 blocked the signal.B: three antibodies to Tie-2 raised to different domains of the human protein (amino acids 25–200 of the extracellular domain for RDI, 1–745 for Ab33, and the COOH terminal for AD no. 1) detected the protein in the endothelial cell line and 7.1.1 cells. The peptide used to rise antibody AD no. 1 fully blocked the signal (30 μg protein/lane). C: immunomicroscopy detection of VEGFR-2 (a) and Tie-2 (b) in the endothelial cell line MS-1. Similarly, VEGFR-2 was detected in all 7.1.1 cells with antibody SC505 (c) and antibody SC315 (e). All 7.1.1 cells were positive for Tie-2 with antibody RDI (d) and antibody AD no. 1 (f). Primary cultures of metanephric mesenchymal (MM) cells were also positive for VEGFR-2 (g) and Tie-2 (h). Immunogenic peptides for VEGFR-2 antibody SC315 and for Tie-2 antibody AD no. 1 eliminated their respective signals (not shown). Bars = 21 μm.

Similarly, immunoblots of 7.1.1 cells with three antibodies raised to different domains of Tie-2 detected a robust signal for this receptor, comparable to the signal in the endothelial cells (Fig. 5 B). Furthermore, the blot with antibody AD1 shows that the immunogen peptide completely inhibited the signal. These results, together with the results with RT-PCR, indicate that the 7.1.1 cells synthesize the two receptor tyrosine kinases that are the hallmark of the endothelial phenotype.

To examine whether VEGFR-2 and Tie-2 were synthesized by all 7.1.1 cells or, similar to the case of α-SMA and smooth muscle markers, these receptors were synthesized only by a distinct population of cells, we examined cell monolayers by immunomicroscopy with several antibodies. As expected, cultures of the MS-1 endothelial cells stained with antibodies to VEGFR-2 (Fig. 5 C, a) as well as Tie-2 (Fig. 5 C, b). Similarly, two antibodies to VEGFR-2 (Fig.5 C, c and e) and two to Tie-2 (Fig. 5 C, d and f) showed that these receptors are present in all 7.1.1 cells. Staining for VEGFR-2 and for Tie-2 was changed by neither the confluence of the cultures nor the absence of FCS. To determine whether the presence of these tyrosine kinase receptors in 7.1.1 cells was due to the metanephric mesenchymal origin of the cells, primary cultures of cells derived from E13 metanephric mesenchymes were analyzed; VEGFR-2 (Fig. 5 C, g) and Tie-2 (Fig. 5 C, h) were also detected in primary cultures of E13 metanephric mesenchymal cells grown in vitro.

To determine whether the presence of VEGFR-2 and Tie-2 in the 7.1.1 cells and in primary cultures of E13 rat metanephric mesenchymal cells was due to in vitro culture conditions, we performed two experiments. First, we examined the presence of these two receptors in acutely dispersed cells of E13 rat metanephric mesenchymes cytospun into slides. These cells, similar to 7.1.1 cells and the primary cultures, were also positive for VEGFR-2 and Tie-2 (not shown). Second, we examined for the presence of these two receptors by immunofluorescent microscopy in rat E13 kidneys. Remarkably, VEGFR-2 was detectable in all (Fig. 6, a–c) and Tie-2 in many (Fig. 6, d–f) cells of E13metanephric mesenchymes, including endothelial cells (identified with rat endothelial cell antibody 1) (8). The widespread distribution of Tie-2 was only apparent at the very beginning of renal development (E13 in the rat), and in embryonic kidneys of more advanced age, this receptor was quickly restricted to endothelial cells (not shown).

Fig. 6.

Immunodetection of VEGFR-2 and Tie-2 in E13 kidneys. All cells of the metanephrogenic mesenchyme, including the endothelial cells (identified with rat endothelial cell antibody 1, rhodamine) stained for VEGFR-2 (a–c). Many cells of the mesenchyme, including endothelial cells, also stained for Tie-2 (d–f). Respective immunogenic peptides blocked the staining. Bar = 21 μm.

The presence of endothelial markers in the 7.1.1 cells and in primary cultures of metanephric mesenchymal cells suggested that the cells of renal embryonic mesenchyme that generate epithelia may also be endothelial cell precursors and that under appropriate conditions, they may express markers, such as VE-cadherin and vWF, that define a fully differentiated endothelial phenotype. A variety of maneuvers were tried but had no effect in this regard. However, 7.1.1 cells formed capillaries when grown on Matrigel and internalized acetylated LDL (not shown), two characteristics of endothelial cells.

DISCUSSION

When the ureteric bud invades the metanephric mesenchyme and development of the kidney starts, the renal anlage contains ureteric bud cells (progenitors of the collecting duct epithelia) and metanephric mesenchymal cells. Of the latter, two distinct populations of cells have been identified: cells that are the progenitors of the nephron epithelia (3), excluding the collecting duct, and metanephric mesenchymal stroma cells (15), which presumably are the progenitors of some interstitial cells. Cells expressing VEGFR-2 have also been identified in the metanephric mesenchyme (30, 31) and are believed to be endothelial precursors (i.e., angioblasts). It is unclear, however, whether these angioblasts are distinct from the mesenchymal epithelial precursors or indeed whether they are exogenous to the embryonic kidney.

The origin of the nonepithelial cell types in adult kidney is unknown and, except for endothelial cells, little explored. To determine whether a renal embryonic cell could generate several of the cell types present in adult kidney, we examined the differentiation potential of metanephric mesenchymal cells isolated on the first day of kidney development (E13 in the rat). We used two types of metanephric mesenchymal cells. With an immortalized clonal cell line (7.1.1 cells), we found that the cells express mesenchymal as well as epithelial proteins, suggesting that the mesenchyme from which the cells were derived had initiated, after ureteric bud induction, its transformation into epithelia (3, 34). When cultured in vitro, the cells develop monolayers, express epithelial junctional proteins, and have well-defined apical and basolateral domains, indicating that the cells differentiate into epithelia. In addition, RT-PCR showed that the cells express a variety of renal genes characteristic of fully differentiated nephron epithelia, suggesting that these cells are kidney-specific mesenchymal cells. We also found that the cells possess the capacity to differentiate into cell types other than epithelia and, with the appropriate culture conditions, differentiate into cells that are spindle shaped, do not express ZO-1, do not make a monolayer, and express α-SMA, indicating that they are myofibroblasts and/or smooth muscle cells. Although expression of α-SMA is not indicative of a smooth muscle phenotype, the appearance of cells with several specific smooth muscle proteins (Figs.2 A and 4 C) suggests that 7.1.1 cells can indeed fully differentiate into smooth muscle cells.

We identified two exogenous cues that regulate differentiation of the 7.1.1 cells toward myofibroblast and/or smooth muscle. Seeding the cells on collagen IV decreased, and FCS and TGF-β1enhanced, the number of spindle-shaped cells expressing α-SMA. Whether these effects result from selection by collagen IV and TGF-β1 of cells with an already determined phenotype (resulting from asymmetrical cell division) or these factors act by inducing expression of a genetic program that causes a specific cell fate is now unclear. Because TGF-β1 was found to transdifferentiate mammary epithelia into mesenchyme (24), it would appear that at least in the case of this cytokine, the second possibility is more likely. Regardless, it is worth noting that the appearance of myofibroblasts in the renal parenchyma is a prominent feature of many renal diseases, and these cells are believed to participate in the mechanism of disease (reviewed in Ref.13). Although it has long been known that TGF-β1 increases production of renal extracellular matrix (4) and may increase the number of renal myofibroblasts (11), it may be worth exploring the possibility that instability of collagen IV may also be a mechanism contributing to myofibroblast appearance and fibrosis in renal diseases. Of note, humans with genetic mutations of different collagen IV chains (36) or animals with deletion of these genes (6) develop interstitial fibrosis with renal failure.

At the embryonic age at which the primary cultures of metanephric mesenchymal cells were isolated (E13), α-SMA is not detectable in the cells of the metanephric mesenchyme in vivo, but isolation and overnight culture causes most of these cells to become spindle shaped and express α-SMA (Fig. 3 B). Similar to the 7.1.1 cells, this differentiation was markedly inhibited by collagen IV because this matrix protein restricted the cells toward an epithelial fate. The similar behavior of the 7.1.1 cell line and the primary cultures of metanephric mesenchymal cells indicate that the plasticity of the line is not due to its transformation but rather reflects the embryonic origin of the cells. In agreement with our results, previous studies have shown that isolation of neuronal stem cells with propagating genes does not alter the ability of the cells to enter into multiple differentiation pathways (12, 14).

In a previous study using a lineage marker (19), it was found that metanephric mesenchymal cells could generate all the different types of nephron epithelia (except collecting duct), indicating that these cells are renal epithelial stem cells. Although in that study no vascular cells were detected, the present results indicate that metanephric mesenchymal cells have a broad differentiation potential and, in addition to epithelia, can generate myofibroblasts and smooth muscle cells. In addition, because clonal experiments demonstrated that the ability of the cells to differentiate into either epithelia or myofibroblast/smooth muscle is self-renewing, it is apparent that some metanephric mesenchymal epithelial precursors are embryonic renal stem cells.

Because fetal and adult organ-specific stem cells are likely to have similar molecular characteristics, characterization of renal embryonic renal stem cells may provide clues to identify adult renal stem cells.

We also found that the 7.1.1 cells express two receptor tyrosine kinases, VEGFR-2 and Tie-2, that are the hallmark of an endothelial or endothelial precursor phenotype. The two receptors were also detected in primary cultures of metanephric mesenchymal cells and in vivo metanephric mesenchymes. Although somewhat surprising, these results are consistent with previous reports detecting VEGFR-2 (31) or its mRNA (25, 30) in metanephric mesenchyme and the mitogenic effect of VEGF in many cells, including epithelia, of isolated renal mesenchymes (37). Although the presence of Tie-2 in the embryonic kidney has been suspected to be restricted to well-differentiated endothelial cells (40), there has been no previous analysis of early metanephric mesenchyme. Furthermore, in E13 rat embryos, Tie-2 has been found in tongue and intestinal mesenchymes as well as heart epicardium (23). Of note is that as renal development advances and shortly after E13, Tie-2 quickly becomes restricted to endothelial cells (not shown), which is in agreement with findings from in situ hybridization (40).

The finding of VEGFR-2 and of Tie-2 in 7.1.1 and primary cultures of metanephric mesenchymal cells suggests that metanephric mesenchymal cells that are epithelial precursors and the angioblasts previously detected in the embryonic kidney (30, 31) are the same cells; i.e., metanephric mesenchymal cells that generate epithelia are also progenitors of endothelial cells. Recent studies have shown that most early mesoderms contain cells with characteristics of endothelial precursors (reviewed in Ref. 16), but our studies suggest that at least in the kidney, the endothelial precursors may be the same cells that under appropriate developmental cues became nephron epithelia, myofibroblasts, or smooth muscle. Interestingly, a vascular progenitor capable of giving rise to both smooth muscle and endothelial cells was recently derived from embryonic stem cells (38); the fate of this vascular progenitor is determined by the specific growth factors to which it is exposed.

Confirmation that our cell line and metanephrogenic mesenchymal epithelial precursors are also endothelial progenitors is presently not available. Despite synthesis of VEGFR-2 and Tie-2 and detection by RT-PCR of genes that are characteristic of fully differentiated endothelial cells (e.g., vWF, CD 31), proteins that characterize mature endothelia, including VE-cadherin, vWF, and CD31, could not be detected in 7.1.1 cell immunoblots (not shown). In addition, multiple attempts to induce the cells to increase the synthesis of these proteins and become fully differentiated endothelial cells were unsuccessful. Needless to say, it is possible that the appropriate differentiation signals were not provided or that our conditions contained inhibitory cues for endothelial differentiation, much like collagen IV is for smooth muscle fate. Alternatively, the presence of Tie-2 and VEGFR-2 in metanephric mesenchymal cells may have an as yet undefined function.

Regardless of whether the metanephric mesenchymal cells that generate epithelia also are endothelial progenitors, the finding that they are embryonic renal stem cells makes their phenotype (i.e., simultaneous expression of mesenchymal and epithelial proteins, ability to differentiate into epithelia or myofibroblast/smooth muscle and VEGFR-2+, and Tie-2+) potentially useful in the identification of adult renal stem cells. Work in the central nervous system suggests that embryonic and adult stem cells share molecular characteristics and developmental potential (28).

Acknowledgments

We thank David Williams for help with electron microscopy and Irene Diaz for secretarial assistance.

Footnotes

  • This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-46934.

  • Address for reprint requests and other correspondence: J. Oliver, Dept. of Medicine, Columbia Univ., 630 West 168 St., New York, NY 10032 (E-mail: jao7{at}columbia.edu).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • May 29, 2002;10.1152/ajprenal.00375.2001

REFERENCES

View Abstract