The instructive role of metanephric mesenchyme in ureteric bud patterning, sculpting, and maturation and its potential ability to buffer ureteric bud branching defects

Mita M. Shah, James B. Tee, Tobias Meyer, Catherine Meyer-Schwesinger, Yohan Choi, Derina E. Sweeney, Thomas F. Gallegos, Kohei Johkura, Eran Rosines, Valentina Kouznetsova, David W. Rose, Kevin T. Bush, Hiroyuki Sakurai, Sanjay K. Nigam

This article has a correction. Please see:


Kidney organogenesis depends on reciprocal interactions between the ureteric bud (UB) and the metanephric mesenchyme (MM) to form the UB-derived collecting system and MM-derived nephron. With the advent of in vitro systems, it is clear that UB branching can occur independently of MM contact; however, little has been done to detail the role of MM cellular contact in this process. Here, a model system in which the cultured isolated UB is recombined with uninduced MM is used to isolate the effects of the MM progenitor tissue on the development and maturation of the collecting system. By morphometrics, we demonstrate that cellular contact with the MM is required for vectorial elongation of stalks and tapering of luminal caliber of UB-derived tubules. Expression analysis of developmentally significant genes indicates the cocultured tissue is most similar to an embryonic day 19 (E19) kidney. The likely major contributor to this is the functional maturation of the collecting duct and proximal nephron segments in the UB-induced MM, as measured by quantitative PCR, of the collecting duct-specific arginine vasopressin receptor and the nephron tubule segment-specific organic anion transporter OAT1, Na-Pi type 2 cotransporter, and Tamm-Horsfall protein gene expressions. However, expression of aquaporin-2 is upregulated similarly in isolated UB and cocultured tissue, suggesting that some aspects of functional maturation can occur independently of MM cellular contact. In addition to its sculpting effects, the MM normalized a “branchless” UB morphology induced by FGF7 or heregulin in isolated UB culture. The morphological changes induced by the MM were accompanied by a reassignment of GFRα1 (a receptor for GDNF) to tips. Such “quality control” by the MM of UB morphology may provide resiliency to the branching program. This may help to explain a number of knockout phenotypes in which branching and/or cystic defects are less impressive than expected. A second hit in the MM may thus be necessary to make these defects fully apparent.

  • kidney development
  • branching morphogenesis
  • mesenchymal-to-epithelial transformation
  • polycystic kidney disease
  • kidney tissue engineering

the mammalian kidney arises from reciprocal inductive interactions of two primordial tissues, the ureteric bud (UB) and the metanephric mesenchyme (MM). Development of the kidney is initiated by the outgrowth of the epithelial UB from the mesonephric or Wolffian duct, presumably in response to inductive signals from the MM (a loose aggregation of cells derived from the intermediate mesoderm). The UB then rapidly invades the MM where it is induced to undergo numerous iterations of branching morphogenesis, leading to the formation of the treelike collecting system of the kidney extending from the collecting tubules to the ureter's connection to the trigone of the bladder (29). Simultaneously, the branching tip of the UB induces epithelialization of the mesenchymal cells, resulting in the formation of the nephron from Bowman's capsule through the distal tubule (19).

Use of in vitro culture systems in which either of the two kidney progenitor tissues, UB or MM, can be grown and studied in isolation have provided valuable information about the independent growth and differentiation of both the UB and MM, implicating a number of soluble factors as important components of the morphogenetic events that lead to formation of the nephron (2). For example, soluble factors present in a MM cell-conditioned medium (BSN-CM), in combination with glial cell line-derived neurotrophic factor (GDNF), are capable of inducing branching morphogenesis of the isolated UB (iUB) (21, 26), while other growth factors, including members of the fibroblast growth factor (FGF) family (i.e., FGF1, FGF2, FGF7, and FGF10) were also found to facilitate growth and, to varying degrees, branching morphogenesis of the isolated UB (20). In addition, it has been demonstrated that growth/branch-promoting factors and growth/branch-inhibiting factors [such as transforming growth factor (TGF)-β, BMP2, BMP4, and activin] modulate UB morphogenesis and cooperate in the patterning of the arborized UB (5). These growth factors, as well as certain extracellular matrix molecules, also regulate branching in cultured renal epithelial cells (28). Although these models provide clear-cut evidence that UB branching, independently of the MM, occurs, because they rely on soluble factors present in the medium, they do not address the short-range, mutual, inductive events which lead to the development of the mammalian kidney. Such events would include, among other things, cell-cell contact between the UB and MM as well as soluble and matrix molecules acting at short range, important interactions which have thus far not been easily amenable to in vitro analysis.

It has been demonstrated previously that the epithelial UB can be reprogrammed (through repatterning of the expression of certain proteins and by branching parameters) by lung mesenchyme to resemble lung epithelium (11, 12), suggesting such signals emanate from the mesenchymal tissue or through mutual induction. In the present study, we sought to explore, among other things, the role of such short-range, mutual, inductive processes between the UB and MM in the patterning of the developing collecting duct tree by utilizing a previously described recombination model that exploits the ability of the cultured branched isolated UB to recombine with freshly isolated MM to form nephron units (21, 31). These kidney-like tissue recombinants not only provide insights into the reciprocal interactions of the UB and MM but have also been shown to be capable of undergoing early vascularization when implanted in vivo into a host rat, and transporting organic anion transporter (OAT) substrates in a manner similar to functional nephrons (23), suggesting that this recombinant has many properties analogous to the developing in vivo kidney. Analysis of UB growth and branching following recombination demonstrates that in addition to providing important growth factors for UB growth, the MM also provides essential cues for UB arborization patterns, establishment of UB stalk length and tubular caliber (including tapering), as well as nephron maturation. We quantitatively demonstrate that cultured UBs are instructed to undergo further and more complex branching and elongation when recombined with fresh MM while the cultured UB maintains the potential to induce MM mesenchymal-to-epithelial transition. We also show that the MM has the capability of inducing patterning and differentiation in the isolated UB so that the recombined tissue is quantitatively similar to whole embryonic kidney. The patterning effect of MM upon the isolated UB appears to be modulated, in part, by soluble factors, but it is very likely that cell contact or short-acting factors play a key role in determining the arborization pattern. Moreover, we demonstrate that inductive signals are able to affect some level of functional maturation in the proximal tubule and collecting duct, leading to a structure that is similar, at the level of gene expression, to a fairly advanced embryonic kidney. Finally, we show that the addition of MM to dysmorphic UBs leads to normalization of the UB branching phenotype, suggesting that MM may have the potential to correct errors in UB branching through a sculpting process and may be a source of resiliency in the kidney developmental program.



Cellgro cell culture media (DMEM/F12) was from Mediatech (Herndon, VA) and FBS was from BioWhittaker (Walkersville, MD). Growth factor-reduced Matrigel and rat type I collagen were obtained from Becton Dickinson (Franklin Lakes, NJ). Growth factor GDNF was from R&D Systems (Minneapolis, MN), while FGF1 was from EMD Biosciences (San Diego, CA). The primary antibodies against E-cadherin (mouse monoclonal, 1:100), Pax2, WT1, and collagen IV were from Transduction Laboratories (Los Angeles, CA). Secondary antibodies used were from Jackson ImmunoResearch Laboratories (West Grove, PA) and Invitrogen (Carlsbad, CA). Fluorescein or rhodamine-conjugated Dolichos bifloris lectin was from Vector Laboratories (Burlingame, CA). Except for the Transwells (clear, polyester, 0.4-μm pore; Costar, Cambridge, MA), all plasticware was from Falcon (Lincoln Park, IL). All other reagents and chemicals, unless otherwise indicated, were from Sigma (St. Louis, MO).

Animal care and research protocols were performed in accordance with University of California San Diego (UCSD) institutional guidelines, following approval by the Animal Subjects Program and the institutional animal care and use committee of UCSD.

Generation of BSN-conditioned medium.

MM-derived cells (BSN cells) were cultured in DMEM/F12 supplemented with 10% fetal calf serum at 37°C in an atmosphere of 5% CO2. Conditioned medium was collected after incubation with serum-free DMEM/F12 for 3–4 days. After collection, the conditioned media was concentrated approximately fivefold, and the buffer was changed to fresh DMEM/F12 media using an Ultrasette filtration device (5K MWCO; Pall; San Diego, CA).

Isolation and culture of the UB.

Isolated UB cultures were performed as previously described (21) with minor modifications. Briefly, embryonic day 13 (E13) rat kidneys were lightly digested with trypsin and the UBs were separated from the MM using fine-tipped needles. The UBs were suspended in Transwell filters within a matrix containing growth factor-reduced Matrigel diluted 1:1 with DMEM/F12 media. The isolated UBs were then cultured in BSN-CM supplemented with 10% FCS, 125 ng/ml GDNF, and 125–375 ng/ml FGF1 or FGF7 or heregulin as indicated in the text. UBs were allowed to grow for 6–7 days.

Recombination cultures of isolated UBs with MM.

Following 6–7 days of culture, the UBs were dissected away from the majority of the surrounding ECM gel, positioned on top of a new Transwell tissue culture insert and freshly isolated or induced MM from E13.5 kidneys were placed in close proximity to the cultured UB. The recombined tissues were then cultured in DMEM/F12 supplemented with 10% FCS at 37°C in an atmosphere of 5% CO2 and 100% humidity for up to 10 days as previously described (21), (31). Experiments were all performed in triplicate with an n = ≥3 unless otherwise indicated.


Specimens were fixed in 4% paraformaldehyde (EM Sciences, Fort Washington, PA) for 30 min at room temperature and then washed with PBS. UBs were dissected from the Transwell insert and excess extracellular matrix gel was removed. Immunohistochemistry was performed as previously described (13, 21). Specimens were examined by scanning laser confocal microscopy (Zeiss LSM-510, Nikon EZ-C1 confocal systems). Images were processed with Photoshop software (Adobe, San Jose, CA).


To visualize the lumens of the recombined tissues (n = 2), rhodamine-labeled dextran sulfate between 3 and 40 kDa was injected into the luminal space using an Eppendorf semiautomated microinjection apparatus mounted on a Nikon inverted microscope via pulled glass capillary needles (13) fluorescently labeled lumens were visualized using a Nikon dissecting microscope equipped with epifluorescence.


Fluorescent images of lumen-stained structures were acquired at ×4 magnification [D. bifloris in embryonic kidney and isolated (i) UB; rhodamine-labeled dextran sulfate microinjected in recombined tissues]. Images were automatically processed with Photoshop software (Adobe) via channel selection for the fluorescent lumen color and edge tracing with the software Sobel filter. For each measurement, three or more specimens were analyzed. A first-generation branch was defined as the primary trunk of both the embryonic kidney and iUB, and the first branches of the ureteric bud extending into the mesenchyme in the recombined tissue. Branch length, width, and paired-branch angles were manually measured with the Photoshop-supplied ruler and protractor.

Quantitative RT-PCR analysis.

The RNA from tissues (MM, iUB, UB-MM coculture) either freshly isolated or from tissue culture was isolated at the specified time points (n = 3) using an RNAqueous Micro kit from Ambion (Austin, TX). Quantitative (q) PCR analysis was performed in triplicate using Invitrogen's SuperScript III Platinum Two-Step qRT-PCR Kit with SYBR Green and the ABI 7500 Fast Real-time PCR System (Applied Biosystems, Foster City, CA). All primers were designed using the Primer3 website and synthesized by Allele Biotechnology (San Diego, CA). The following primer sets were used for qPCR: rat aquaporin-2 (sense and antisense: 5′-agagctcttcctgaccatgc-3′ and 5′-ccggtgaaatagatcccaag-3′, respectively); rat OAT-1 (sense and antisense: 5′-gaagagcaggagcagaggaa-3′ and 5′-accccactccctttagtgct-3′, respectively); rat Na-Pi type 2 cotransporter (NaPi-2; sense and antisense: 5′-gccaatgtcatccagaaggt-3′ and 5′-tgctctggacaacaaacgtc-3′, respectively); rat Tamm-Horsfall protein (THP; sense and antisense: 5′-atcacacgacaaggtgtcca-3′ and 5′-gccacaccaggttttctgtt-3′, respectively); rat GAPDH (sense and antisense: 5′-aaggtcatcccagagctgaa-3′ and 5′-cctgcttcaccaccttcttg-3′, respectively); and mouse arginine vasopressin receptor 2 (AVPR2; sense and antisense: 5′- cacacctacggaaaggcatc-3′ and 5′-ctgttgctgggagagctagg-3′, respectively). Samples were incubated at 50°C for 2 min followed by 95°C for 2 min followed by the 45 cycles of 15 s at 95°C, 30 s at 56°C, and 30 s at 72°C. Dissociation steps were run on all samples.

Microarray analysis.

The E13, E16, E19, 4-wk-old kidney, and recombined kidney-like tissues were dissolved in lysis buffer, and the RNA was isolated using the RNAqueous Micro kit from Ambion. RNA was submitted to the UCSD Genechip core facility for processing with the Affymetrix (Santa Clara, CA) rat 230 2.0 whole genome chip. All in vivo time points as well as the recombined tissue were analyzed in biological triplicates. Gene expression analysis was performed using Agilent (Santa Clara, CA) Genespring GX 7.3 software. Statistical comparisons of gene subsets were made with Welch's t-test (parametric, variances of biological sample triplicates not assumed to be equal) and multiple testing correction via Benjamini and Hochberg's false discovery rate procedure (3). Raw data for the whole embryonic kidney time points are available at the NCBI GEO website with accession number GSE9570.


For recombination experiments, isolated rat UBs cultured for 6–7 days were dissected from the original extracellular matrix gel and then placed in contact with freshly isolated E13.5 MM as previously described (Fig. 1A) (21, 31). These recombined tissues were further cultured in DMEM/F12 media supplemented with 10% FCS for an additional 7–10 days (day 0 being the day of recombination) without the addition of any exogenous soluble growth factors. During these 7–10 days of coculture, the MM underwent a mesenchymal-to-epithelial transition as evidenced by the evolution of cap condensates, pretubular aggregates, and comma and S-shaped bodies (Fig. 1, BE). The recombined UB itself underwent branching and vectorial elongation (Fig. 1F). The leading edge of the elongating UB retained tiplike properties during recombination as evidenced by immunolocalization of GFRα1 to the tips of newly forming branches (Fig. 1G). During later recombination, GFRα1 was also found in the condensing pretubular mesenchyme and early secretory nephrons (Fig. 1H) as has been described in the developing kidney (24). As previously noted, connecting segments between metanephric S-shaped bodies and the UB were also observed in the UB/MM coculture (Fig. 2, A and B) (31). These connecting segments appeared to effectively remove the UB tip from further branching events consistent with the branching that occurs in in vivo embryonic kidney development. The tubules were joined by a single lumen as evidenced by contiguous filling of both central and peripheral UB segments via microinjection of rhodamine-labeled dextran (Fig. 2, CF), demonstrating that the tubules are in fact similar to those that form in vivo.

Fig. 1.

Isolated ureteric bud (UB) induces mesenchymal-to-epithelial transformation. Phase-contrast of uninduced metanephric mesenchyme (MM) placed around the cultured UB (A) and cocultured for 7 days (B and C) is shown. DF: scanning-laser confocal images of UB and MM recombinations after 8 days of coculture where the UB is shown in green (Dolichos bilious; DB) and MM cells are in red (E-cadherin). The boxed area in D indicates the boundary of the isolated UB before coculture. Structures derived from mesenchymal-to-epithelial transformation, including cap-condensate (B, arrows) and coronas (indicated by asterisks) similar to those seen in the embryonic kidney, are observed. Connection of the UB with the MM effectively removes the UB tip from further branching events similar to branching in the developing embryonic kidney (arrow in E). F: magnified view of an area of UB branching during coculture. Boxed area indicates elongation of UB branches and vectorial growth toward the MM; arrow indicates an area of the UB that has not undergone recombination with the MM and maintains its original architecture. G: immunolocalization of GFRα1 (green) to the UB tip demonstrates that the invading UB branches retain tip properties during recombination while at a later stage of recombination (H), GFRα1 is also found in the condensing pretubular mesenchyme and early secretory nephrons (E-cadherin, red).

Fig. 2.

Continuous tubular connection between UB ampullae and MM tubuli during UB/MM coculture. A and B: scanning-laser confocal image of UB/MM coculture on day 8 indicating the presence of the mesenchymally derived renal epithelial cell marker peanut lectin agglutinin (PNA; red) and D. bifloris (DBA; green). Note the staining pattern of DBA in this transition zone (white arrows). CF: immunofluorescence microscopy (C and E) and merged images with phase-contrast microscopy (D and F) of the UB/MM coculture on day 8 after microinjection of labeled dextran into the UB lumen. Note the contiguous filling of dextran into the new UB branches and tubules derived from the MM induced by coculture (white arrows). Dashed line outlines new luminal area.

Soluble factors instruct patterning of the UB.

To better define the role of soluble factors vs. cell-cell contact with the MM in the patterning of the UB, we utilized the well-characterized iUB assay (21) to first determine whether soluble factors alone can induce structural changes in the iUB. It has previously been shown that a variety of soluble growth factors, which include members of the FGF and TGF-β superfamilies, as well as pleiotrophin and heregulin, can modulate both the stalk and tip phenotype of the UB (20). For example, it was shown that the exogenous addition of TGF-β1 altered the normal UB phenotype, resulting in UBs with thicker stalks and reduced branching (5). Interestingly, the combination of FGF7 (a growth factor which induces the formation of a globular, nonbranched UB) (20), with members of the TGF-β superfamily partially reversed this reduced-branching phenotype (5). This suggests that combinations of growth-facilitatory (such as FGF7) and branch-inhibitory growth factors (such as TGF-β1) are potentially important in the growth and branching of the UB (28). It is possible then that, upon growth of the UB into the MM, the UB comes into contact with or induces the MM to produce soluble factors which alter its phenotype.

To investigate this possibility, we cultured isolated UBs suspended in ECM in standard growth media of DMEM/F12 with 10% FCS and 125 ng/ml of GDNF and FGF7 for 5 days. Following these 5 days of culture, the suspended UB was then cultured in BSN-CM with 125 ng/ml GDNF with and without 125 ng/ml FGF1 (Fig. 3). These UBs were then allowed to grow for another 6 days. This experimental setup allows us to examine whether MM soluble factors contribute to UB patterning. We have previously found that iUBs cultured in the presence of FGF7 assume a predominantly “tip” phenotype, evidenced by the presence of less differentiated cells with abundant secretory granules throughout the bud, with essentially no stalk formation (20) (Fig. 3, A and D). When they are then cultured in the presence of BSN-CM, however, they assume a more “normal” phenotype, with distinct stalks (Fig. 3C, filled arrow) and ampullae (Fig. 3C, open arrow). The presence of FGF1 did not appear to significantly alter the phenotype (data not shown), and this effect was not seen if the media was changed to BSN-CM containing FGF7 (Fig. 3, DF). The morphological change in the UB appeared to be due to an effect of active remodeling by factors contained in BSN-CM as well as withdrawal of FGF7 since change in either one alone had no effect on UB morphology. This result suggests that soluble factors can modulate UB patterning and that iUBs cultured in a three-dimensional ECM possess structural plasticity. Interestingly, the patterning of the UB occurred in all UB segments; this is unlike isolated UBs recombined with MM where mostly distal branches and new growth showed tapering and elongation (Fig. 1F), suggesting that the secreted soluble factors only act over a short range. Thus both soluble factors present in BSN-CM and the MM itself can “repattern” the isolated UB in culture toward tips and stalks that structurally resemble those in the intact developing kidney. Key cell membrane proteins may be involved in MM-UB interactions related to patterning; we sought to further investigate this possibility by evaluating the effects of addition of MM to UB branching.

Fig. 3.

Soluble factors can pattern the isolated UB. A, B, D, and E: brightfield images of the isolated UB initially grown in DMEM/F12 with 10% FCS and 125 ng/ml FGF7 for 5 days and then cultured in BSN-CM with 10% FCS and 125 ng/ml FGF1 (AC) and isolated UB initially grown in DMEM/F12 with 10% FCS and 125 ng/ml FGF7 for 5 days and then cultured in BSN-CM with 10% FCS and 125 ng/ml FGF7 (DF). C and F: fluorescent images of isolated UBs 6 days after media change and stained for D. bifloris. Filled arrow in C points to an area of proximal stalk formation, and open arrow points to an ampulla that were not initially present in the isolated UB cultured in DMEM/F12 with 10% FCS and 125 ng/ml FGF7 and which do not form in the presence of BSN-CM with 10% FCS and 125 ng/ml FGF7 (F). Magnification ×10.

MM instructs the iUB to branch and elongate.

As shown above in Fig. 1, the recombinations revealed that the cultured iUB underwent impressive morphological changes in the presence of MM. The ampullae of preexisting branches in the iUB developed many smooth outpouches, which elongated to form new branches. Thus branching morphogenesis of the iUB after recombination with MM appeared to be different from that observed in the presence of conditioned media (21). Whereas the iUB developed short, thick branches with broad, rounded ampullae during in vitro culture (Fig. 1A), the recombined UB underwent branching and elongation, but with branches and ampullae that were thinner and longer (Fig. 1, D and F). Quantitative morphometric analysis of the isolated UB compared with the UB after recombination and UB in embryonic kidney culture demonstrated that the iUB was essentially an iterative structure with uniform branch length, branch caliber, and branch angle from one generation to the next (Fig. 4). Conversely, the UB in cultured embryonic kidney was a patterned structure with varying branch lengths, caliber, and angles. In this case, the UB stalks showed progressive shortening in later generations of branching in addition to tapering of luminal caliber (Fig. 4, A and B). Branch angles also became wider as branching progressed (Fig. 4C). When MM was recombined with the iUB, new UB branches that formed showed a pattern of growth similar to UB branches found in the embryonic kidney, with a strong correlation of between tapering length and width over successive branching generations (r2 = >0.7) and weaker correlation with widening angle (r2 = 0.4). These data strongly support the view that cell-cell contact between the MM and UB regulates UB architecture in addition to inducing growth and branching, although it is possible that alteration in the distribution and local concentration of soluble factors after the addition of MM to the UB also contributes to these processes.

Fig. 4.

Morphometric analysis of the UB. AC: UB branch length (A), branch width (B), and branch angles (C) in the isolated UB culture, UB/MM coculture, and in embryonic kidney cultured for 5 days. Isolated UB branch length, width, and angles are generally uniform over 5 generations of branching, while UB branching in the cultured embryonic kidney shows tapering length and width with increasing branch angle over successive branching generations. Addition of MM to the isolated UB under coculture conditions alters the morphometry of newly induced UB branches to resemble that of the UB in embryonic kidney, suggesting that MM has the ability to pattern the UB and ultimately influence final kidney architecture. The coefficient of determination of a linear regression comparison between the ordinal generation and morphometric measurement is indicated by r2. *P < 0.001 for the slope of the least squares line.

Coculture stimulates more maturation of MM than the UB.

In the aforementioned results, inductive signals between the UB and MM stimulate UB branching morphogenesis and the mesenchymal-to-epithelial transition, respectively. It remains unknown, however, whether these signals also induce maturation of these structures into the collecting duct and tubular nephron. We sought to answer these questions using the coculture assay by first comparing the global gene expression pattern of the cocultured tissue with embryonic kidney at various developmental intervals from the early E13 kidney to the postdevelopmental 4-wk-postnatal kidney. We focused on a set of 11,108 “developmentally significant genes,” defined as genes with at least a threefold expression change between any two of the in vivo whole kidney stage conditions (E13, E16, E19, 4-wk-postnatal) (33a). By comparing the expression of these genes in the recombined tissue with that in each of the four in vivo time points, we found that the cocultured tissue most closely resembled the E19 kidney. Scatter plots comparing absolute gene expression provided a visual representation of this greater similarity between the recombined tissue and the E19 kidney (Fig. 5A). An evaluation of endothelial-specific genes (Pecam, Tie1, Vegf, Et1, and Cdh5) demonstrated that these genes are expressed at a level corresponding to an E13E16 kidney (Fig. 5B) (7). Conversely, epithelial-specific genes (Emp1, Emp2, Emp3, and Cdh1) in the recombined tissue were expressed at a level corresponding to a more mature stage (Fig. 5C). Further statistical analysis revealed that 96.9% of developmentally significant genes in the recombined tissue were not differently expressed in the E19 kidney (P ≤ 0.05). This association between the recombined tissue and the in vivo E19 kidney time point persisted when further analyzed by narrowing the set of developmentally significant genes to a subset of genes exhibiting progressive upregulation or downregulation, revealing 92.6 and 92.8% gene expression homology, respectively (P ≤ 0.05) (Fig. 5D). Gene ontology analysis of the 341 genes that had statistically different expression between the E19 kidney and the recombined tissue revealed that, in addition to the aforementioned endothelial and epithelial genetic differences as would be expected in the avascular recombined tissue, one-fifth of these genes were associated with non-renal-specific organ development and cell differentiation.

Fig. 5.

cDNA microarray analysis. Global microarray results for whole kidney stages at embryonic day 13 (E13), E16, E19, and 4-wk old were narrowed to a set of “developmentally significant genes” (11,108 of 31,000), defined as genes with at least a 3-fold expression change between any 2 stages. A: scatter plot of normalized expression of developmentally significant genes for E13, E16, E19, and 4-wk-old whole kidneys compared with recombined tissue (diagonal lines represent ±1.5- and ±3-fold change, respectively). B and C: linear plot of normalized expression of vascular and nonvascular (epithelial) genes in E13 to 4-wk-old whole kidneys and recombined tissue, respectively. D: tabulated results of the subset of developmentally upregulated or downregulated genes with normalized expression in recombined tissue that exhibit statistically similar expression in various stages of the in vivo whole kidney (P ≤ 0.05).

We complemented the structural analysis with an examination for nephron differentiation by examining markers of glomerular differentiation. We found that the glomerular structures were positive for WT1 and collagen IV, suggesting that the epithelial glomerulus and glomerular basement membrane were present in the coculture and that nephronogenesis proceeded up to the point of vascularization (Fig. 6, AC). Examination of Pax2 localization within the recombined tissue demonstrated high Pax2 expression in collecting ducts and condensing mesenchyme but not in preglomerular and mature renal tubule (proximal tubule) structures, a pattern consistent with an older developing kidney (6) (Fig. 6D). We then investigated the extent of functional transformation by looking for specific markers of collecting duct and proximal tubule differentiation. The collecting duct-specific marker included the AVR. Using qPCR, we compared the expression levels of AVR in E13 MM, the iUB cultured for 1 wk, and the iUB cultured for 2 wk, to the recombined tissue cultured for 7 days (Fig. 7A). We found that while addition of MM to the iUB appeared to result in an increase in AVR expression, the low levels of expression impeded the power for significance. Because of this low qPCR expression level, we compared the expression of AVR at various in vivo embryonic kidney stages by cDNA microarray, which revealed that the normalized expression level of AVR best matched an E19 kidney (Fig. 7B). This result strongly suggested that the newly formed connecting tubular structures in the cultured recombined tissue are differentiated structures and not simply elongations of the iUB itself.

Fig. 6.

Nephron differentiation in in vitro tissue recombination between metanephric mesenchyme (MM) and cultured isolated ureteric bud (UB). A and B: immunolocalization of WT1 (green) and PNA (red) in the recombined tissue demonstrates formation of the epithelial glomerulus. C: immunolocalization of collagen IV (red) to the glomerular structures suggests maturation of the glomerular basement membrane. D: Pax2 immunolocalization (red) to the collecting ducts (asterisk) and condensing MM (black arrow) but not preglomerular (white arrow) or proximal tubular structures suggests maturation of these structures within the recombined tissue. DAPI, blue.

Fig. 7.

Expression levels of collecting duct- and proximal tubule-specific markers. A: quantitative expression levels relative to GAPDH normalized to maximal expression of the collecting duct-specific marker arginine vasopressin receptor (AVR), respectively, in E13 MM, isolated UB cultured for 1 wk (the time at which coculture is initiated in the recombined coculture model), UB/MM cocultured for 7 days, and isolated UB cultured for 2 wk (the duration of isolated UB+coculture). B: bar graph of normalized cDNA microarray expression level of AVR in in vivo whole kidneys at E13, E16, E19, and the recombined tissue culture demonstrating that the expression of AVR in the recombined tissue most closely approximates the level of AVR expression of the E19 kidney. C: analysis of the quantitative expression levels relative to GAPDH normalized to maximal expression of the proximal tubule-specific markers Tamm-Horsfall protein (THP), Na-Pi transporter-2 (NaPi-2), and the organic anion transporter-1 (OAT1). Error bars indicate SE.

The proximal tubular markers we investigated included OAT-1, an early proximal tubular marker, NaPi-2, a middle proximal tubular marker, and THP, an ascending loop of Henle-specific marker. We compared their expression levels between uninduced E13 MM, the isolated cultured UB and the UB/MM recombined tissue via qPCR. Consistent with the observation that robust mesenchymal-to-epithelial transition occurs in the cocultured tissue, we found that the recombination of UB and MM leads to a functional maturation of the MM through the upregulation of these proximal tubular markers (Fig. 7C).

In aggregate, these results demonstrate that although relatively differentiated, the iUB retained the ability to induce the MM and, in the process, created functional nephron units and at the same time could be stimulated by fresh MM to undergo further branching. That the MM was capable of stimulating new branch formation from structures that had some degree of terminal differentiation suggests a certain level of plasticity to the ureteric epithelial cells.

MM has the ability to correct dysmorphic UB growth and branching.

A number of factors identified as important to in vitro branching morphogenesis in a variety of embryonic tissues have failed to result in a significant phenotype when tested in in vivo studies. Many of these factors, which include FGF2 and FGF7, likely serve redundant functions along with other factors from the same family of molecules while others, such as GDNF, are clearly important in UB branching (8, 14, 17, 18). Nevertheless, the phenotype of renal agenesis in GDNF knockout mice is only partially penetrant, suggesting alternate signaling pathway activation (12a, 16). Given the large number of factors involved in UB branching, it remains surprising that there are only relatively few developmental errors of branching. Many sources of resiliency exist in kidney development although one that has been largely overlooked is the MM itself. We hypothesized that the MM could “correct” UB patterning errors, thus buffering against many branching mutations. As shown, iUBs cultured with FGF7 had no discernable stalks (Fig. 3, A and D). When the FGF7-treated UBs are recombined with MM, the resulting newly formed UB branches are slender with clear stalks and tips (Fig. 8A). Further evaluation revealed that addition of MM to the FGF7 iUB induced GFRα1 to become localized to the tips of the newly formed branches (Fig. 8, B and C). This is in distinct contrast to the nearly uniform presence of GFRα1 along the isolated FGF7 UB (Fig. 8, F and G). A similar correction of UB morphology was seen with iUBs cultured in the presence of heregulin, a factor thought to play a role in the growth and differentiation of the stalk portion of the epithelial UB (Fig. 9) (25). This is interesting in light of the finding that, in contrast to isolated UBs cultured with FGF7, GFRα1 expression is substantially downregulated in heregulin-treated iUBs. These data demonstrate that the MM does, in fact, induce the differentiation of both functional tips and stalks and suggest that there are compensatory mechanisms within the MM that can correct or override developmental errors in UB branching.

Fig. 8.

MM can normalize dysmorphic, tip-predominant UB growth. A and B: scanning confocal laser images of the isolated UB cultured in the presence of FGF7 then recombined with freshly isolated MM. Despite the lack of distinct stalk and tips in the cultured UB (see Fig. 3, A and D), new UB branches induced by MM are elongated and tapered and have morphometry similar to branches induced from isolated UBs cultured in the presence of FGF1 which have distinct stalks and tips (green, D. bifloris; red, PNA; blue, lotus lectin). BG: scanning confocal laser images of immunofluorescence localization of GFRα1 (red) and D. bifloris (green) in FGF7 UB/MM coculture (B and C), FGF1 isolated UB (D and E), and FGF7 isolated UB (F and G). Note the uniform distribution of GFRα1 along the FGF7 isolated UB compared with the localization of GFRα1 at the UB tips in FGF1 isolated UB. Coculture of the FGF7 isolated UB with MM normalizes GFRα1 localization. Scale bars = 100 μm.

Fig. 9.

MM can normalize stalk-predominant UB growth induced by heregulin. A: brightfield image of the isolated UB cultured in the presence of heregulin then cocultured with freshly isolated MM. B: fluorescent image of the heregulin UB-MM coculture after 5 days of culture. D. bifloris (green) staining demonstrates how, despite the lack of distinct stalk and tips in the cultured UB, new UB branches induced by MM (nephric tubules; NT) are elongated and tapered and have morphometry similar to branches induced from control isolated UBs. Scale bars = 100 μm.


A key morphogenetic event in kidney collecting system development is the mutual induction that takes place between the epithelial UB and the surrounding MM. The signals that are transmitted between these tissues ultimately determine the kidney's architecture, arborization of collecting ducts, and nephron number. Although in vitro systems have been used to study the role of soluble factors on UB and MM development separately and in whole organ culture, neither these existing systems, nor knockouts, have shed much light on how the MM regulates UB sculpting, patterning, and maturation. In this study, we used an assay in which cultured isolated UB is recombined with freshly isolated MM to describe the effect of the MM on UB patterning. This allowed us to compare UB branching patterns and differentiation in the absence and presence of the MM in a novel manner. We have previously shown that the iUB is capable of inducing the MM and that connections form between the UB and MM (21, 31). Here, we showed that the MM was also able to repattern the iUB from an iterative structure with uniform stalk length and thickness to one with elongated stalks and tapering tubule caliber. The quantitative morphometric transformation in iUB branching was consistent with tubule changes that occur in cultured embryonic kidney during later stages of branching. While it has previously been suggested that the MM is capable of patterning the UB (10–12, 21, 31, 32), this study demonstrates how the MM specifically influences UB tubular architecture.

Although the precise molecular mechanisms that govern UB patterning remain to be elucidated, our data suggest that a combination of soluble factors as well as cell-cell contact with the MM likely plays a role in this process. In our study, addition of MM-conditioned media along with withdrawal of FGF7, a growth factor that has previously been shown to induce a predominantly “tip phenotype” in the iUB (20), sculpted the iUB from an amorphous structure to one with distinct tips and stalks. This restructuring of the iUB was similar to that seen in the new branches that form in the UB/MM assay, with the important difference that new UB branches that form in the UB/MM coculture system grew vectorially toward the MM and also demonstrate tapering of UB caliber, an effect not seen in the isolated UB system without the presence of MM. We suggest that the vectoriality and tapering of the UB in vivo are an effect of both soluble factors and cell-cell contact and that the soluble factors responsible for this patterning must only be acting over a short range.

Current views suggest that epithelial cell-fate determination is controlled via the inductive signals emanating from mesenchyme and that epithelial-mesenchymal cross talk is required to complete morphogenesis and differentiation (4). Our gene expression data clearly demonstrate that inductive cross talk between the UB and MM has the potential to advance kidney development to a relatively mature state. However, our data also suggest that while cell-cell contact is required for some aspects of kidney differentiation (patterning), others, such as the acquisition of certain membrane proteins specific to the collecting system, are not dependent upon UB-MM cross talk.

Based on gene-inactivation experiments, branching in the developing kidney is relatively resistant to mutation of single genes (15, 27, 30). In the many gene deletions in mice that might be expected to result in renal phenotypes (i.e., members of the FGF and TGF-β superfamilies), few result in unambiguous branching defects. Mutations that cause clear-cut kidney phenotypes tend to affect either the early bud outgrowth event, leading to kidney agenesis, or very late events in development, leading to tubular ectasia or cyst formation (9, 15, 16a, 27, 34). Our finding that the MM was able to correct dysmorphic UB growth and branching is interesting in this regard. In this setting, the MM showed the ability to take a UB that predominantly had a tip or stalk phenotype and induced structures that had a mature architecture with both tips and stalks. This implies that the MM, apart from providing cues for vectorial guidance of UB-derived branching tubules as well as their elongation, tapering, and differentiation, appears to provide a level of “quality control” to limit defects in UB branching morphogenesis, presumably through the action of as yet unidentified MM cell surface proteins or soluble molecules. This resiliency may be partially responsible for the relative lack of “intermediate” branching phenotypes seen in UB branching morphogenesis (30).

Several clinical renal diseases have roots that can be traced to errors during ureteric bud morphogenesis. The phenotypic expression of tubular diseases exist as early as the developing fetus for forms of renal dysplasia as well as cystic diseases such as polycystic kidney disease (PKD) (33). Even when dominant genetic defects such as PKD1 in autosomal PKD are known to be present, a variable range of phenotypes can exist, resulting in differences in disease onset or severity. This has led to a search for other disease-modifying factors or “second hits” that may tip the balance between disease expression and phenotypic absence (1). One such research area has been growth factors, which have been implicated in modulating disease expression in the development of the distal nephron (22). Our finding that the MM was able to correct or override developmental errors in UB branching, such as the cystlike structures resulting from our use of the EGF-related peptide heregulin, suggests that the mesenchyme has the potential to modify and rescue dysmorphogenetic tubular disorders. These results also suggest that certain mesenchymal genes may play a decisive role in determining phenotypic expression of predisposed disorders of renal tubulogenesis, potentially acting as second hits to make the phenotype fully manifest, and thus serve as an important model for future studies into renal disease modification. Polymorphisms in this set of genes may be a key source of “genetic background” variation in development renal disease.


This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grants RO1-DK57286, RO1-DK65831 (to S. K. Nigam). M. M. Shah is supported by a Research Career Award from the NIDDK (K08-DK069324). J. B. Tee is supported by a Research Fellowship Award from the Canadian Institutes of Health Research. Y. Choi is supported by a training grant from the National Institutes of Health (T32-HL007261). H. Sakurai is also supported by an American Heart Association Scientist Development Award. K. T. Bush was also supported by a Normon S. Coplon Extramural Grant from Satellite Healthcare, Inc.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 12a.
  14. 13.
  15. 14.
  16. 15.
  17. 16.
  18. 16a.
  19. 17.
  20. 18.
  21. 19.
  22. 20.
  23. 21.
  24. 22.
  25. 23.
  26. 24.
  27. 25.
  28. 26.
  29. 27.
  30. 28.
  31. 29.
  32. 30.
  33. 31.
  34. 32.
  35. 33.
  36. 33a.
  37. 34.
View Abstract