The metanephric blastema differentiates into collecting system and nephron epithelia in vitro

Development 121, 3207-3214 (1995) Printed in Great Britain The Company of Biologists Limited 1995 3207 The metanephric blastema differentiates into collecting system and nephron epithelia in vitro J. Qiao, D. Cohen and D. Herzlinger* Department of Physiology and Biophysics, Cornell University Medical College, New York, NY 10021, USA *Author for correspondence (email: daherzli@mail.med.cornell.edu) SUMMARY The kidney forms from two tissue populations derived from intermediate mesoderm, the ureteric bud and metanephric mesenchyme. It is currently accepted that metanephric mesenchyme is committed to differentiating into nephrons while the ureteric bud is restricted to forming the renal collecting system. To test this hypothesis, we transferred lacz into pure metanephric mesenchyme isolated from gestation day 13 rat embryos. The fate of tagged mesenchymal cells and their progeny was characterized after co-culture with isolated ureteric buds. When induced to differentiate by the native inducer of kidney morphogenesis, lineage-tagged mesenchymal cells exhibit the potential to differentiate into collecting system epithelia, in addition to nephrons. The fate of cells deriving from isolated ureteric buds was also examined and results of these lacz gene transfer experiments indicate that the majority of ureteric bud cells differentiate into the renal collecting system. These cell fate studies combined with in situ morphological observations raise the possibility that collecting system morphogenesis in vivo occurs by growth of the ureteric bud and recruitment of mesenchymal cells from the metanephric blastema. Thus, metanephric mesenchyme may be a pluripotent renal stem population. Key words: kidney, mesoderm, renal stem cell, metanephric blastema, nephron epithelium, rat INTRODUCTION Renal function is dependent on the organization of diverse epithelial cell types into functional units called uriniferous tubules (Bulger, 1987). Each uriniferous tubule has a nephron and collecting tubule. According to anatomical conventions, the nephron is composed of a filtering glomerulus and a segmented epithelial tubule. Each nephron is contiguous with a collecting tubule which empties into the ureter via the renal calyces and pelvis. During embryonic development, kidney morphogenesis is initiated when the ureteric bud, an epithelial diverticulum that forms at the caudal aspect of the Wolffian Duct, induces surrounding intermediate mesoderm to differentiate (Saxen, 1987). This population of intermediate mesoderm, the metanephric mesenchyme or blastema, has been isolated and cultured in vitro (Grobstein, 1953, 1955, 1957; Grobstein and Dalton, 1956). Isolated metanephric blastemata differentiate into glomerular, proximal and distal convoluted nephron segments when cultured in the presence of embryonic spinal cord, a potent heterologous inducer of renal epithelial morphogenesis in vitro (Ekblom et al., 1981). Renal collecting system morphogenesis occurs only when isolated metanephric blastemata are co-cultured with the native inducer of kidney formation, the ureteric bud (Grobstein, 1953, 1955). It is on these observations that the current model describing the lineage of renal epithelia is based (Saxen, 1987). This model predicts that the ureteric bud induces mesenchyme of the metanephric blastema to differentiate into glomerular through distal convoluted tubule nephron segments, exclusively. The collecting tubule and intrarenal collecting system are believed to derive solely from the ureteric bud, after it is induced to branch and grow by metanephric mesenchyme. Since this model has yet to be tested by current lineage analysis techniques, we performed retroviral mediated gene transfer lineage analyses to examine the differentiated fate of cells deriving from both the ureteric bud and metanephric blastema (Price et al., 1987; Koseki et al., 1991; Herzlinger et al.,1992). Results of these experiments indicate that the majority of ureteric bud cells differentiate into the renal collecting system. Metanephric mesenchymal cells, however, are not restricted to differentiating into nephron epithelia. When co-cultured with the ureteric bud, the native inducer of kidney morphogenesis, metanephric mesenchyme differentiates into collecting system epithelia in addition to nephrons. These results combined with ultrastructural studies of the gestation day 13 rat kidney rudiment suggest that during embryogenesis, renal collecting system morphogenesis can occur by growth of the ureteric bud and recruitment of cells from the metanephric blastema. MATERIALS AND METHODS Isolation and characterization of embryonic tissues Timed pregnant Sprague Dawley rats were obtained from Hilltop Laboratories (PA). The day of appearance of the vaginal plug was designated as gestation day 0. Rats were killed by CO 2 asphyxiation on

3208 J. Qiao, D. Cohen and D. Herzlinger Fig. 1. Morphology of the gestation day 13 rat kidney rudiment. Frozen section (4 µm) of rat kidney rudiment isolated from a gestation day 13 embryo processed for peroxidase Dolichos- Biflorus histochemistry and poststained with toluidine blue. At this stage of development, the ureteric bud (u) is a single tubule with a terminal dilated ampulla. Mesenchymal cells of the metanephric blastema (m) are spindle shaped and dispersed. All ureteric bud cells are labeled with peroxidase-db, while metanephric mesenchymal cells do not bind this lectin. Peroxidase reaction product is not observed when sections are incubated in an excess of the competing sugar for DB, N-acetyl-D-galactosamine (data not shown). Bar, 25 µm. Fig. 2. Isolated metanephric blastemata and ureteric bud preparations are homogeneous. (A) Frozen section of gestation day 13 rat kidney rudiment processed for FITC-DB staining. Only the ureteric bud (u) is labeled by this lectin; the surrounding cells of the metanephric blastema (m) are not stained. Bar, 10 µm. (B) Kidney rudiments isolated from gestation day 13 rat embryos were dissociated into single cell suspensions and incubated with 50 µg/ml FITC-DB in the presence of 1 M N-acetyl-D-galactosamine, the competitive sugar for DB binding. Cells were analyzed by fluorescenceactivated cell sorting. Cells of total kidney rudiments run as a single peak that exhibits background fluorescence intensity. 50,000 cells were analyzed. (C) Kidney rudiments were processed as described above and incubated in FITC-DB in the absence of N-acetyl-Dgalactosamine. Two peaks can be easily resolved. The major peak, metanephric mesenchymal cells, exhibits background fluorescence staining. The minor peak, ureteric bud cells, exhibits fluorescence staining approximately 1 log unit greater than background values. 50,000 cells were analyzed. (D) Metanephric blastemata were isolated from gestation day 13 rat kidney rudiments by microdissection after collagenase digestion. Isolated metanephric blastemata preparations are composed of a homogeneous population of dispersed mesenchymal cells as determined by electron microscopic examination. Bar, 5 µm. (E) Isolated metanephric blastemata were dissociated into single cell suspensions and incubated with FITC-DB. Isolated blastemata preparations run as a single peak of cells exhibiting background fluorescence. 50,000 cells were analyzed and 2% exhibited fluorescent staining intensities above background levels as determined by profile from B. (F) Ureteric buds were isolated from gestation day 13 rat kidney rudiments by microdissection after collagenase digestion. The adluminal surface of isolated buds appeared free of adherent metanephric mesenchymal cells as determined by electron microscopic examination. Bar, 5 µm. (G) Isolated ureteric buds were dissociated into single cell suspensions, incubated with FITC-DB and analyzed as described. Isolated ureteric buds run as a single homogeneous population of DB-positive cells and exhibit fluorescent staining intensities approximately 1 log unit greater than background values. 25,000 cells were analyzed and 95.6% of this cell population exhibited fluorescent staining intensities above background levels as determined from the profile shown in B.

Cell commitment in kidney development 3209 Fig. 3. Specificity of histochemical procedures utilized to identify lineagetagged cells and cells of the renal collecting system. (A) Isolated ureteric buds were incubated with 20 µg/ml polybrene for 2 hours, washed extensively and then recombined with isolated metanephric blastemata. Recombination organ culture was grown for 7 days, fixed and incubated in x-gal to visualize β-galactosidase enzyme activity. This representative mock infected organ cultures lacks cells exhibiting the blue, β-gal reaction product. Bar, 15 µm. (B) Isolated ureteric buds were incubated in concentrated BAG retrovirus containing 20 µg/ml polybrene for 2 hours and washed extensively prior to culture with uninfected metanephric blastemata. Culture was grown for 7 days, fixed, processed for β-gal activity, followed by incubation in peroxidase-db in the presence of 1 M N-acetyl-D-galactosamine. β- gal-positive cells are easily visualized and the brown DB-peroxidase reaction product is not present. The presence of β-gal-positive cells is dependent on BAG infection. DB binding is inhibited by the competitive sugar, N-acetyl galactosamine. Bar, 15 µm. (C) Recombination organ culture established and grown as described in B. This culture was processed for β-gal activity and then incubated in peroxidase-db in the absence of the competitive sugar, N-acetyl-D-galactosamine. β-gal + (blue) and DB + (brown) cells are present. Bar, 15 µm. gestation day 13 and metanephric kidney rudiments dissected from embryos in Liebowitz s L-15 medium (L-15; Gibco). Metanephric kidney rudiments were incubated with Dulbecco s Modified Eagle s Medium (Gibco), 10% Fetal Calf Serum (Hyclone Laboratories)(DMEM-FCS) containing 0.2% collagenase (Gibco, prepared from histolyticum, approximately 200 units/mg) and 50 u/ml DNAase (Boehringer Mannheim) for 15 minutes at 37 C. Rudiments were then washed extensively in ice-cold DMEM-FCS. Ureteric buds were dissected from collagenase digested rudiments utilizing fine tipped minutia pins (Fine Science Tools). Isolated buds and blastemata were kept in DMEM-FCS on ice until use. For quantitative analysis, isolated ureteric bud and metanephric blastemata fractions obtained from collagenase-digested rudiments were trypsinized for 10 minutes (0.05% in 0.5 mm EDTA; Gibco). The resultant single cell suspensions were washed in DMEM-FCS and then phosphate-buffered saline (PBS) by low speed centrifugation and incubated for 30 minutes with fluorescein-conjugated Dolichos Biflorus (FITC-DB, vector; 50 µg/ml), a lectin that binds specifically to the ureteric bud. In addition, single cell suspensions were incubated with FITC-DB in the presence of 1 M N-acetyl galactosamine, the sugar moiety to which DB binds (Holthöfer, 1988). Cell preparations were washed and then fixed in 2% paraformaldehyde for 1 hour prior to cell sorting. In addition, whole metanephric kidney rudiments were subjected to collagenase followed by trypsin digestion, and the total cell suspensions assayed for FITC-DB binding as described above. Cells (a minimum of 25,000/sample) were run on a Becton Dickinson Fluorescent Activated Cell Sorter and analyzed with LYSIS II Ver 1.1 software. Gene transfer and organ culture The lacz gene was transferred into cells of isolated ureteric buds or metanephric blastemata by replicationdefective retroviral infection utilizing the BAG retrovirus (Price et al., 1987). Virus secreted into the culture medium from BAG-producing cell lines (provided by Fig. 4. In vitro cells of the metanephric blastema differentiate into nephron and collecting system epithelia. Hematoxylin and eosin-stained methacrylate sections (8 µm) of recombination organ cultures established with BAG-infected metanephric blastemata and uninfected ureteric buds. Cultures were fixed and processed for β-gal expression and DB-histochemistry after 7 days of culture. Bar, 15 µm. (A,B) Cells deriving from the metanephric blastema exhibit the morphology of condensed mesenchymal nephron progenitors (m) and glomerular epithelia (g). (C) In addition, what appear to be colonies of β-gal-tagged cells span from mesenchyme (m) condensing around the ureteric bud into the terminal tips of the forming renal collecting system (c). (D) A large percentage of metanephric mesenchymal cells differentiate into the segments of the renal collecting system (c) which can be identified by DB binding and a characteristic branched morphology.

3210 J. Qiao, D. Cohen and D. Herzlinger C. Cepko, Harvard University) exhibited an infective titre of 3 10 5 infectious particles/ml, as assayed by NIH 3T3 cell infection. BAGconditioned medium was concentrated 50 by ultracentrifugation and the concentrated viral stocks were stored at 80 C until use. All concentrated viral stocks were shown to be replication defective as described (Price et al., 1987; Mann et al., 1983). Isolated buds or blastemata were incubated with concentrated replication defective virus, in the presence of 20 µg/ml polybrene (Sigma) for 2 hours at 37 C. After retroviral infection, isolated tissues were washed extensively in DMEM-FCS and placed on the upper surface of 0.45 µmpore transwell culture chambers. Uninfected metanephric blastemata or ureteric buds were co-cultured with the appropriate BAG-infected primordia to facilitate renal differentiation. Each sample contained 8 buds and 8 blastemata. The bottom compartment of the transwell chambers were filled with DMEM/10% FCS and recombination organ cultures grown at the air/medium interface for indicated times. Culture medium was changed every 3 days. Histology and histochemistry All tissues were prepared for electron microscopy by conventional techniques as described (Herzlinger and Ojakian, 1984). Samples were viewed with a JEOL 100CX electron microscope. For frozen sections, tissue was fixed in 2% paraformaldehyde for 2 hours and then immersed in 30% sucrose for 12 hours. Samples were frozen in Tissue-Tek (Fisher) in isopentane cooled over liquid nitrogen, and sections cut at 20 C with a Hacker cryomicrotome. Sections were collected on Vectabond (Vector Laboratories)-coated glass slides. For Dolichos Biflorus staining (DB), sections were incubated with 50 µg/ml FITC or peroxidase-conjugated DB (Sigma) in the presence or absence of 1 M N-acetyl-D-galactosamine. After lectin incubation, sections were washed, fixed and viewed (FITC-DB) or peroxidase-db activity developed by incubation with 3,3 diaminobenzidine (DAB, Vector Laboratories). After development of the peroxidase reaction product samples were post-fixed in 2% paraformaldehyde. β-galactosidase activity was visualized by 5-bromo-4-chloro-3- indolyl-β-d-galactopyranoside (X-gal, New Jersey Glove, Inc.) hydrolysis (Koseki et al., 1991). This protocol is modified from that of Pearson and utilizing this protocol β-galactosidase (β-gal) activity endogenous to renal organ cultures is not observed (Koseki et al., 1991; Pearson et al., 1963). Dolichos Biflorus (DB) histochemistry was performed on whole mount samples after β-galactosidase histochemistry by incubation in 50 µg/ml peroxidase-labeled DB (Sigma)/PBS containing 0.075% saponin (Sigma). Specificity of DB binding was determined by the incubation of samples in DB and an excess of the competitive sugar, N-acetyl-D-galactosamine. The peroxidase reaction product was developed by 3,3-diaminobenzidine (Vector Labs). Organ cultures were prepared for light microscopic examination by embedding in methacrylate (Historesin, LKB). Serial, 8 um sections were cut with a MT-2 microtome (Sorvall). Samples were viewed with a Zeiss Axioskop. RESULTS On gestation day 13, the rat metanephric kidney rudiment is composed of the two primordia that form the kidney, the ureteric bud and the surrounding metanephric mesenchyme. By light microscopy, ureteric bud cells are organized into a single tubule with a dilated ampulla. All ureteric bud cells bind the lectin Dolichos Bifloris (DB). Cells of the metanephric blastema exhibit a dispersed, spindle-shaped morphology and are not labeled by DB (Fig. 1). The specificity of DB binding to the ureteric bud/forming collecting system is maintained throughout development and in the mature rodent kidney (Lahetonen et al., 1987; Holthöfer, 1988). Ureteric buds and metanephric blastemata were isolated from gestation day 13 rat kidney rudiments by microdissection after collagenase digestion. Serial sections (0.5 µm) of isolated ureteric bud and metanephric blastemata preparations were examined by light microscopy (data not shown). In addition, random sections were examined by electron microscopy (Fig. 2D,F). By morphological criteria, isolated ureteric buds appeared free of adherent mesenchymal cells and isolated blastemata were free of ureteric bud epithelia. The purity of isolated renal primordia was also assessed by quantitative florescence activated cell sorting using the ureteric bud-specific marker, FITC-DB (Fig. 2A-C,E,G). Results of these quantitative assays demonstrate that 96% of the cells in isolated ureteric bud preparations bind the ureteric bud-specific marker, DB. A maximum of 2% of the cells in isolated metanephric blastemata preparations exhibited DB binding above background levels. Finally, the purity of both isolated ureteric bud and metanephric blastemata preparations was tested by functional criteria. It is well established that neither the metanephric blastema nor ureteric bud grows or differentiates when cultured alone (Grobstein, 1953, 1955). Isolated ureteric bud and metanephric blastemata preparations grew and differentiated only when recombined with the appropriate complementary renal primordium (data not shown). Isolated renal primordia were incubated with BAG, a replication-defective retrovirus encoding lacz (Soriano and Jaenisch, 1986; Sanes et al., 1986; Price et al., 1987; Price, 1989; Koseki et al., 1991; Herzlinger et al., 1992). Integration of lacz into the genome of embryonic renal cells is cell cycle dependent, and the efficiency of gene transfer into embryonic renal cells was approximately 5%. All BAG retrovirus stocks utilized for these studies were shown to be replication defective, thus lacz transferred into embryonic renal cells can only be passed in a heritable manner (Mann et al., 1983). After BAG infection, isolated tissues were washed extensively and co-cultured with uninfected complimentary primordia to facilitate renal differentiation in vitro. Recombination organ cultures were grown for 3 or 7 days, fixed, and processed for β-galactosidase (β-gal) activity to identify lineage-tagged cells, and peroxidase-db histochemistry (DB) to identify the ureteric bud-forming renal collecting system. Serial sections of cultures were examined and the number and phenotype of β-galpositive cells assessed. Mock-infected, control organ cultures were processed in an identical manner. These control cultures lacked cells exhibiting the blue, β-gal reaction product (Fig. 3A). Only cultures established with BAG-infected renal primordia had β-gal-positive cells (Fig. 3B,C). Viral infected cultures were incubated with peroxidase-db in the presence and absence of the competitive sugar for DB binding (Fig. 3B,C). The renal collecting system is specifically labeled by DB-peroxidase. All lineage-marked cells deriving from isolated metanephric blastemata exhibited a mesenchymal, or aggregated mesenchymal morphology after 3 days of culture (Table 1). By 7 days of culture, 43% of cells derived from this primordium exhibited an aggregated mesenchymal morphology characteristic of differentiating renal epithelia. 15% of the β-gal-positive cells were located within morphologically identifiable glomerular and tubular nephron segments. Surprisingly, 41% of the tagged cells derived from isolated metanephric blas-

Cell commitment in kidney development 3211 temata were present in the forming renal collecting system (Table 1, Fig. 4). This portion of the developing kidney was identified by its unique branched architecture and DB staining. Tagged cells, derived from the metanephric blastemata, were located at the distal growing tips of the ureteric bud/renal collecting system and, in many samples, colonies of lineagetagged cells span from aggregated mesenchymal condensates into the forming renal collecting system. These data demonstrate that the metanephric blastemata co-cultured with the native inducer of kidney morphogenesis, the ureteric bud, differentiate into the renal collecting system in addition to nephrons. In recombined organ cultures established with BAGinfected ureteric buds, 97% of the β-gal-tagged cells were labeled with DB after 3 days of culture (Table 2). By 7 days of culture, the number of β-gal-positive cells rose approximately 1.5-fold. At this time, 73% of the tagged cells exhibited a ureteric bud phenotype. Thus, the majority of ureteric bud cells differentiate into the renal collecting system. At this time, however, 27% of the ureteric bud-derived cells no longer bound DB. The majority of these DB-positive ureteric budderived cells exhibited a morphology consistent with cells previously thought to be nephrogenic mesenchyme. In addition, ureteric bud-derived cells were observed in DB-negative tubular structures and in morphologically identifiable glomerular nephron segments (Table 2, Fig. 5). Thus, after isolation and organ culture, a significant percentage of ureteric bud cells exhibit a phenotype inconsistent with the forming renal collecting system. Collectively, these in vitro lineage studies imply that during embryogenesis in vivo, the renal collecting system forms by growth of the ureteric bud and recruitment of metanephric mesenchymal cells. We examined the ultrastructure of cells in the early rat kidney rudiment to determine if metanephric mesenchyme integrates into the forming collecting system. Electron microscopic examination of gestation day 13 rat kidney rudiments demonstrates that cells located in the tubular portion of the ureteric bud, closest to the Wolffian Duct, exhibit an epithelial morphology. Cells of this portion of the bud are columnar and have well-formed junctional complexes. The columnar ureteric bud epithelium here, is well displaced from the surrounding metanephric mesenchymal cells, which are fusiform and lack intercellular junctions (Fig. 6A). In contrast to the majority of ureteric bud Table 1. Differentiated fate of metanephric mesenchymal cells in vitro Ureteric bud coll. sys. (DB + ) Glom. Tubular Mesench. 3 days 0 0 0 100 7 days 41.6 (1.15) 9.6 (1.5) 4.5 (0.7) 43.6 (6.6) Isolated metanephric blastemata were infected with BAG and recombined with uninfected ureteric buds. Recombination organ cultures were fixed and processed for β-gal and DB peroxidase histochemistry after 3 or 7 days of culture. A minimum of 3 recombination organ cultures from separate experiments were analyzed for each time and a minimum of 120 and 547 β- gal-tagged cells were counted at 3 and 7 day cultures, respectively. Data is expressed as percentage of total β-gal-positive cell population (s.e.m.). (Glom., glomerular; mesench., mesenchymal; coll. sys., collecting system) cells, cells of the bud located at the growing ampullary tip, are rounded, stratified and lack intercellular junctions (Fig. 6B). Rounded, stratified, ureteric bud tip cells extend from the luminal to adluminal surface of the branched, lateral ureteric bud tips, as determined by examination of 0.5 µm thick serial sections (data not shown). Fusiform mesenchyme of the metanephric blastema appears to intermix with rounded, adluminal ureteric bud tip cells (Fig. 6C). These morphological observations are consistent with the hypothesis that the renal collecting system grows, in part, by recruitment of cells from the metanephric blastema. DISCUSSION Currently, it is believed that the metanephric blastema is committed to differentiating into nephron epithelia, exclusively. In this study, we demonstrate that metanephric mesenchyme differentiates into portions of the renal collecting system, in addition to nephron epithelia. These results are consistent with previous retroviral-mediated gene transfer lineage analyses. lacz-tagged metanephric mesenchyme gives rise to branched tubular structures in addition to nephron epithelia when cultured as explants on the cortex of neonatal rat kidneys (Koseki et al., 1991; Herzlinger et al., 1992). When we examined the differentiated fate of lineage-tagged cells deriving from isolated ureteric buds co-cultured with metanephric mesenchyme, we observed that the majority of ureteric bud cells formed the renal collecting system. After 3 days of culture, 97% of the lineage-tagged cells deriving from isolated ureteric buds bound the ureteric bud-specific marker, DB. This percentage is compatible with the initial purity of isolated ureteric bud preparations. By 7 days of culture, the majority of ureteric bud-derived cells were located in the renal collecting system, supporting the hypothesis that the ureteric bud gives rise to this segment of the kidney. However, 27% of the lineage-tagged cells deriving from isolated ureteric buds, exhibited a morphology consistent with differentiating and differentiated nephron epithelia. We cannot exclude the possibility that some of these lineage-tagged nephron progenitors derive from contaminating mesenchymal cells present in the initial isolated ureteric bud preparations. If this were the case though, lineage-tagged metanephric mesenchymal cells must have proliferated at a 10-fold greater rate than tagged ureteric Table 2. Differentiated fate of ureteric bud cells in vitro Ureteric bud coll. sys. (DB + ) Glom. Tubular Mesench. 3 days 97.7 (1.05) 0 0 2.2 (1.05) 7 days 72.6 (7.7) 3.3 (1.5) 1 (0) 23.3 (5.8) Isolated ureteric buds were infected with BAG and recombined with uninfected blastemata. Recombination organ cultures were fixed and processed for β-gal and DB peroxidase histochemistry after indicated times. A minimum of 3 recombination organ cultures from separate experiments were analyzed for each time and a minimum of 368 and 572 β-gal-tagged cells were counted at 3 and 7 day cultures, respectively. Data is expressed as percentage of total β-gal-positive cell population (s.e.m.). (Glom., glomerular; mesench., aggregated or condensed mesenchyme; coll. sys., collecting system)

3212 J. Qiao, D. Cohen and D. Herzlinger bud cells. The relative rates of division of metanephric mesenchymal versus ureteric bud cells can be estimated by comparing the total number of β-gal-tagged cells deriving from isolated renal tissues at 3 versus 7 days of culture. On the average, lineage-tagged cells deriving from isolated blastemata exhibited a three times greater rate of proliferation than tagged ureteric bud-derived cells. Thus, it is highly unlikely that contaminating mesenchyme present in isolated ureteric bud populations can account for the number of nephron progenitors observed to differentiate from isolated ureteric buds. This developmental plasticity of isolated ureteric buds in vitro may be related to our data demonstrating that portions of the renal collecting system can derive from metanephric mesenchyme. In short, the most recently formed ureteric bud/collecting system segments may not be stably differentiated collecting tubule epithelia. Electron microscopic observations support of this contention. Terminal tip ureteric bud cells are rounded, stratified and lack intercellular junctions; they do not exhibit an epithelial morphology. In addition, rounded terminal ureteric bud tip cells express both epithelial and mesenchymal intermediate filament proteins indicating that they are in the process of undergoing mesenchymal-toepithelial conversion (Bachmann et al., 1983). Thus, it is likely that this newly formed distal portion of the renal collecting system reverts back into metanephric mesenchyme after tissue isolation and in vitro culture. Alternatively, we cannot exclude the possibility that metanephric mesenchyme differentiates sequentially into the terminal ampullary tip of the ureteric bud, and then nephron epithelia. Collectively, these in vitro cell fate studies raise the possibility that the renal collecting system forms by ureteric bud growth and recruitment of cells from the metanephric blastema. Although we have previously reported that isolated metanephric mesenchyme does not give rise to the renal collecting system, this study was performed with the lineage marker, 1,1 1 -dioctadecyl-3,3,3 1,3 1 -tetramethylindocarbocyanine percholorate (DiI) (Honig and Hume, 1986; Herzlinger et al., 1994; Herzlinger, 1994). This carbocyanine dye labels cell membranes, and is diluted and lost if marked cells undergo more than three rounds of division (Honig and Hume, 1986; Mann et al., 1983). In this study, we transferred the lacz gene, a genomic marker into metanephric mesenchyme. Using this genomic lineage marker, which is not diluted by cell division, we demonstrate that metanephric mesenchyme is competent to differentiate into most, if not all, renal epithelial cell types (Price et al., 1987). Thus, it appears that metanephric mesenchyme must undergo multiple rounds of cell division prior to differentiating into the segments of the collecting system. Fig. 5. In vitro the majority of ureteric bud cells form the renal collecting system while some cells deriving from this primordium are competent to give rise to differentiating and differentiated nephron segments. Hematoxylin and eosin-stained methacrylate sections (8 µm) of recombination organ cultures established with BAG-infected ureteric buds and uninfected metanephric blastemata. Cultures were fixed and processed for β-gal expression and DB histochemistry after 7 days of culture. Bar, 15 µm. (A,B) The majority of cells deriving from isolated ureteric buds differentiate into cells of the renal collecting system (c) which can be identified by DB-peroxidase staining and a characteristic branched morphology. (C) Lineage-tagged ureteric bud-derived cells in this section are present in the collecting system (c) and an S-shaped body (s) which is an immature nephron (Saxen, 1987). The domain of the S-shaped body exhibiting tagged cells is thought to give rise to tubular nephron segments (Saxen, 1987). In addition, cells exhibiting a mesenchymal morphology (m) are marked. (D) A small percentage of cells deriving from isolated ureteric buds differentiated into morphologically identifiable glomerular nephron segments (g).

Cell commitment in kidney development 3213 In situ morphological observations indicate that metanephric mesenchyme can integrate into the forming renal collecting system within the embryo. Previous ultrastructural studies demonstrate that the basement membrane separating the distal ureteric bud tips and surrounding metanephric mesenchyme is incomplete (Lehetonen, 1975). Our electron microscopic studies demonstrate that metanephric mesenchymal and ureteric bud cells intermix in this area. Finally, unlike metanephric mesenchyme located at the periphery of the kidney rudiment, mesenchyme condensed around the ureteric bud tips expresses several proteins in common with the ureteric bud. (Dressler et al., 1990; Maas et al., 1994; Partenan and Thesleff, 1987; Vainio et al., 1992). Collectively, these data raise the possibility that, in vivo, the renal collecting system forms, in part, by recruitment of cells from the metanephric blastema. Although the ureteric bud/renal collecting system is absolutely required for kidney morphogenesis, little is known concerning the formation or growth of this embryonic tissue. It is believed to grow by mechanisms similar to those that guide Wolffian duct or nephric duct elongation. Recent lineage analyses performed in Xenopus indicate that the nephric duct extends along the anterior-to-posterior axis of the embryo by growth of the pre-existing nephric duct and induced recruitment of cells from intermediate mesoderm (Cornish and Etkin, 1993). Thus, our in vitro lineage analyses of the rodent collecting system are in agreement with the mechanisms reported for Xenopus nephric duct morphogenesis in vivo. When co-cultured with embryonic spinal cord, a potent heterologous inducer of nephron formation in vitro, the metanephric blastema differentiates into glomerular, proximal and distal tubule epithelia exclusively (Ekblom et al., 1981). We co-cultured isolated metanephric blastemata with embryonic spinal cord and DB + branched tubular structures were not observed (data not shown). In vitro, only the ureteric bud induces the metanephric blastema to differentiate into nephron and collecting system epithelia. This observation is compatible with those reported by Gruenwald (Gruenwald, 1942, 1952). In his classic experiments, Gruenwald obstructed the growth of the chick nephric duct inhibiting ureteric bud formation. He then transplanted nervous tissue into metanephric mesenchyme of embryos lacking a ureteric bud and showed that renal tubular-like structures form. The tubules that form in response to heterologous nervous tissue induction in ovo, however, are truncated and exhibit an architecture consistent with mesonephric tubules, the functional units of the vestigial, intermediary kidney formed during embryogenesis. The difference between the uriniferous tubules of the mesonephric and the final, or metanephric, kidney is that the latter possess collecting tubule segments (Saxen, 1987). Thus, our results are consistent with Gruenwald s; only the ureteric bud is competent to induce all the differentiative events leading to metanephric uriniferous tubule morphogenesis. Another difference noted between the inducing activities of the ureteric bud and spinal cord is that only the native inducer supports continued proliferation of metanephric mesenchymal cells. When isolated metanephric blastemata are co-cultured with the ureteric bud, lineage-tagged metanephric mesenchymal cells proliferate approximately 3.5-fold from culture day 3 to 7. When cultured with the heterologous inducer, embryonic spinal cord, metanephric mesenchymal cells Fig. 6. Cells at the terminal, growing ureteric bud tips are not organized as an epithelium and appear to intermix with mesenchymal cells of the metanephric blastema. Gestation day 13 rat kidney rudiments were processed for electron microscopy by conventional techniques. Bar, 1 µm. (A) Cells in the tubular portion of the ureteric bud (u) exhibit an epithelial morphology and are easily discriminated from mesenchymal cells of the metanephric blastema (m). (B) Cells at the ampullary ureteric bud tip (u) are rounded and stratified. (C) Cells at the adluminal surface of the distal ureteric bud tip (u) exhibit a fusiform morphology and are difficult to discriminate from surrounding cells of the metanephric blastema (m). At this location, the border separating the two primordia that form the kidney is not clear. undergo a transient proliferative burst within the first 24 hours of co-culture, and then pull out of the cell cycle (Saxén et al., 1983). In addition, embryonic spinal cord is a more potent inducer of terminal differentiation than the ureteric bud. When co-cultured with spinal cord metanephric mesenchyme differentiates into definitive stroma and nephrons. In contrast, the ureteric bud supports continued mesenchymal cell prolifera-

3214 J. Qiao, D. Cohen and D. Herzlinger tion, collecting system morphogenesis, and only limited terminal nephron differentiation. Interestingly, this latter pattern of differentiation more closely resembles the early stages of kidney morphogenesis in vivo. For example, in the developing human, although the ureteric bud forms by gestation week 4, nephrogenesis does not occur until gestation week 18. During gestation week 4-18, the ureteric bud grows and branches to form the ureter, renal pelvis and calyces while the population of mesenchymal nephron progenitors expands (Oliver, 1968). Similarly, in the rat, the ureteric bud is surrounded by metanephric mesenchyme on gestation day 12.5 but S-shaped bodies, immature nephron structures, do not appear until gestation day 15 (Gilbert et al., 1994). From gestation day 13-15, the initial segments of the renal collecting system are established (renal pelvis and large ducts of Bellini), while the mesenchymal nephron progenitors proliferate with little evidence of terminal nephron differentiation. In conclusion, although metanephric mesenchyme isolated from embryos at the earliest stages of metanephric kidney formation is competent to form nephrons when co-cultured with embryonic spinal cord, in vivo cells of this primordium may differentiate into portions of the renal collecting system in addition to nephrons. Thus, the metanephric blastema may be a pluripotent renal epithelial stem cell population. This work was supported by a New York Heart investigatorship and NIH grant R01 DK45218 awarded to D. H. We thank Ms Diane Mangraviti for the use of the fluorescence-activated cell sorter, Ms Lee Cohen Gould for assistance in preparing tissues for electron microscopy and Mr E. Lau for photographic processing. In addition, we thank Drs T. Mikawa, D. Bader, Q. Al-Awaqati, J. Barasch and D. Fischman for critically reviewing this manuscript, Dr C. Cepko for the gift of the BAG-producing cell line, and Dr D. Fischman for the use of the morphological facilities provided by the Department of Cell Biology and Anatomy, Cornell University Medical College. REFERENCES Bachmann, S., Kriz, W., Kuhn, C. and Franke, W. W. (1983). Differentiation of cell types in the mammalian kidney by immunofluorescence microscopy using antibodies to intermediate filament proteins and desmoplakins. Histochemistry 77, 365-394. Bulger, R. E. (1987). The urinary system. In Histology. (ed. L. Weiss, and R. O. Greep). pp. 831-888. New York: McGraw-Hill. Cornish, J. A. and Etkin, L. D. (1993). The formation of the pronephric duct in Xenopus involves recruitment of posterior cells by migrating pronephric duct cells. Dev. Biol. 159, 338-345. Dressler, G. R., Deutsch, U., Chowdhury, K., Nornes, H. O. and Gruss, P. (1990). Pax-2, a new murine paired-box-containing gene and its expression in the developing excretory system. Development 109, 787-795. Ekblom, P., Miettinen, A., Virtanen, I., Wahlstrom, T., Dawnay, A. and Saxen, L. (1981). In vitro segregation of the metanephric mesenchyme. Dev. Biol. 84, 88-96. Gilbert, T., Gaonach, S., Moreau, E. and Merlet-Binichou, C. (1994) Defect of nephrogenesis induced by gentamicin in rat metanephric organ culture. Laboratory Investigation 70, 656-666. Grobstein, C. (1953). Inductive epithelio-mesenchymal interaction in cultured organ rudiments of the mouse. Science 118, 52-55. Grobstein, C. (1955). Inductive interaction in the development of the mouse metanephros. J. Exp. Zool. 130, 319-340. Grobstein, C. (1957). Trans-filter induction of tubules mouse metanephrogenic mesenchyme. Exp. Cell Res. 10, 424-440. Grobstein, C. and Dalton, A. J. (1956). Kidney tubule induction in mouse metanephrogenic mesenchyme without cytoplasmic contact. J. Exp. Zool. 135, 57-73. Gruenwald, P. (1942). Experiments on distribution and activation of the nephrogenic potency in the embryonic mesenchyme. Phys. Zool. 15, 396-409. Gruenwald, P. (1952). Development of the excretory system. NY Acad. Sci. USA 55, 142-146. Herzlinger, D., Abramson, R. and Cohen, D. (1994) Phenotypic conversions in renal development. J. Cell Science 17, 61-64. Herzlinger, D. (1994). Renal stem cells and the lineage of the nephron. Annu. Rev. Physiol. 56, 671-689. Herzlinger, D., Koseki, C., Mikawa, T. and Al-Awqati, Q. (1992). Metanephric mesenchyme contains multipotent stem cells whose fate is restricted after induction. Development 114, 565-572. Herzlinger, D. and Ojakian, GK. (1984). Studies on the development and maintenance of epithelial cell surface polarity with monoclonal antibodies. J. Cell Biol. 98, 1777-1787. Holthöfer, H. (1988). Cell type-specific glycoconjugates of collecting ducts cells during maturation of the rat kidney. Cell Tissue Res. 253, 305-309. Honig, M. G. and Hume, R. I. (1986). Fluorescent carbocyanine dyes allow living neurons of identified origin to be studied in long-term cultures. J. Cell Biol. 103, 171-187. Koseki, C., Herzlinger, D. and Al-Awqati, Q. (1991). Integration of embryonic nephrogenic cells carrying a reporter gene into functioning nephrons. Am. J. Physiol. 261, C550-C554. Lahetonen, E., Virtanen, L. and Saxén, L. (1987). Changes in the glycosylation pattern during embryonic development of mouse kidneys as revealed with lectin conjugates. J. Histochem. Cytochem. 35, 55-65. Lehetonen, E. (1975). Epithelio-mesenchymal interface during mouse kidney tubule induction in vivo. J. Embryol. Exp. Morph. 34, 695-705. Maas, R., Elfering, S., Glaser, T. and Jepal, L. (1994). Deficient outgrowth of the ureteric bud underlies the renal agenesis phenotype in mice manifesting the limb deformity (ld) mutation. Dev. Dynamics 199, 214-228. Mann, R., Mulligan, R. C. and Baltimore, D. (1983). Construction of a retrovirus packaging mutant and its use to produce helper-free defective retrovirus. Cell 33, 153-159. Oliver, J. (1968) Nephrons and Kidneys New York: Harper and Row. Partenen, G. and Thesleff, E. (1987). Localization and quantitation of 125 I-epidermal growth factor binding in mouse embryonic tooth and other embryonic tissues at different developmental stages. Dev. Biol. 120, 186-197. Pearson, B., Wolf, P. L. and Vaquez, J. (1963). A comparative study of a series of new indolyl compounds to localize beta-galactosidase in tissues. Lab. Investigation 12, 1240-1259. Price, J., Turner, D. and Cepko, C. (1987). Lineage analysis in the vertebrate nervous system by retrovirus-mediated gene transfer. Proc. Natl. Acad. Sci. USA 84, 156-160. Price, J. (1989). Cell lineage and lineage markers. Curr. opin. Cell Biol. 1, 1071-1074. Sanes, J. R., Rubenstein, J. L. R. and Nicolas, J. (1986). Use of a recombinant retrovirus to study post-implantation cell lineage in mouse embryos. EMBO J. 5, 3133-3142. Saxen, L. (1987). Organogenesis of the Kidney. Cambridge: Cambridge University. Saxén, L., Salonen, J., Ekblom, P. and Nordling, S. (1983). DNA synthesis and cell generation during determination and differentiation of the metanephric mesenchyme. Dev. Biol. 98, 130-138. Soriano, P. and Jaenisch, R. (1986). Retrovirus as probes for mammalian development: allocation of cells to the somatic and germ cell lineages. Cell 46, 19-29. Vainio, S., Jalkanen, M., Bernfield, M. and Saxén, L. (1992). Transient expression of the syndecan in mesenchymal cell aggregates of the embryonic kidney. Dev. Biol. 152, 221-232. (Accepted 11 July 1995)