How Does the Ureteric Bud Branch?

CREATION OF TIPS AND STALKS

In vitro and genetic approaches have helped to identify many promoters and inhibitors of UB branching (reviewed in^1,2), but the fundamental mechanism of how a straight epithelial tube gives rise to a branched tree remains obscure. Before the advent of systems specifically designed to study the UB, branching morphogenesis of the UB was described through the analysis of branching in renal cell lines such as Madin-Darby canine kidney (MDCK) and UB cells^4–6; however, recent data suggest that UB branching proceeds through a fundamentally different mechanism of outpouching of wedge-shaped cells that are created through an apical cytoskeletal “purse-string” mechanism (Figure 2A).⁷ The “branching through budding” model implies that the remodeling of the contiguous epithelial tube occurs through the creation of a secretory epithelium via differential localization of growth factor receptors and matrix-degrading enzymes to UB tips relative to stalks that initiate new branch formation, a concept supported by cell lineage studies,⁸ and microarray analysis of differential gene expression in UB tip versus stalk cells.⁹ Thus, this challenges the traditional notion that the epithelial tube is composed of homogeneous cells and suggests that microenvironments within the UB, possibly in the form of gradients, are key to vectorial branching morphogenesis and the generation of new tips.^10–12

KIDNEY DEVELOPMENT IS A STAGED PROCESS

In addition to the creation of tips and stalks, branching morphogenesis underpins the basic architecture of the kidney. Global patterns of gene expression during kidney organogenesis, together with in vitro and genetic data and morphologic analyses, suggest that branching morphogenesis of the UB is an iterative yet simultaneously vectorial process that can be broadly conceptualized in terms of developmental stages: (1) Outgrowth of the UB from the WD; (2) rapid, iterative branching of the UB; (3) deceleration of UB branching accompanied by differentiation of the metanephric mesenchyme; and (4) termination of branching and completion of mesenchymal differentiation (Figure 1A).^11,13,14 These stages are separable into in vitro modules that have been used to reconstitute “engineered” kidney tissue that is capable of early vascularization and rudimentary tubular function.¹⁵ Each stage is typified by various sets of heparin-binding growth factors, receptor tyrosine kinases, signaling pathways involving intracellular kinases, and effectors that mediate cell adhesion and basement membrane remodeling (reviewed in^2,16–18). Although disruption of certain pathways produces catastrophic effects (i.e., renal agenesis), numerous instances in which mutation of key molecules has minimal apparent phenotypic consequences exist, suggesting that each of these stages is characterized by a distinct network structure of gene and protein interactions that confer varying resilience to mutation.^19,24,25

UB OUTGROWTH

The initiating step in metanephric development is emergence of the UB from the WD; failure of this critical step leads to renal agenesis, whereas incorrect positioning of the UB leads to a variety of urinary tract anomalies ranging from mega-ureter to vesicoureteral reflux. A multitude of positive and negative regulatory factors, converging on glial cell line–derived neurotrophic factor (GDNF) signaling, play key roles in this stage (reviewed in¹⁶). Genetic deletion of GDNF or its receptor, Ret, most often results in renal agenesis, although rudimentary kidneys form in up to 50% of these mice,^20–22 suggesting that GDNF-dependent budding of the WD may be bypassed through activation of other signaling pathways. This concept has been validated through in vitro studies in which the combined effect of stimulatory (fibroblast growth factor 7 [FGF7]) and blockade of inhibitory (activin) molecules is able to induce bud formation in isolated WDs.²³ The existence of such a bypass pathway may be sufficient to explain the relative infrequency of renal agenesis despite the seeming dependence of UB outgrowth on GDNF signaling; therefore, it may be that a “second hit” is necessary to manifest the phenotype.

EARLY UB BRANCHING

The next conceptualized stage of collecting system development, rapid iterative UB branching, depends on reciprocal epithelial–mesenchymal interactions. A common theme emerging from the study of diverse branching systems is the existence of both positive and negative feedback loops in the complex interplay among epithelium and mesenchyme.²⁴ Like UB outgrowth (and subsequent stages), a combination of stimulatory (e.g., FGFs, pleiotrophin, heregulin) and inhibitory (e.g., TGF-β1, bone morphogenic protein 4, activin) molecules have been shown to modulate the extent of ureteric branching, but, in general, positive feedback mechanisms seem to predominate.¹⁴

The iterative, or “feed-forward,” nature of the branching tree suggests that the expression of a particular set of proteins becomes stable and self-maintaining such that an autocatalytic network, or “tip-stalk generator,” is established (Figure 1A)²⁵; therefore, only a minimal alteration in the expression of a small set of genes may be required for the impressive morphologic changes that occur between the stages of UB branching. Such a “consensus set” of conserved signaling molecules can also be found among a number of organs, and it is the relative “weight” of certain pathways (along with those guiding final differentiation) that may be organ specific.²⁶ Thus, it may be the structure of the branching network through the capture of nodes and the “tightness” of the links between nodes that determines epithelial cell fate.²⁵ Such self-organization has been proposed in gene regulatory networks in origin of life scenarios.²⁷

The question remains, then, how is the complex patterning of the nephron achieved? To explain patterning during embryogenesis, the concept of gradient morphogens was proposed more than a century ago.²⁸ Secreted proteins of the WNT; hedgehog; and members of the EGF, FGF, and TGF-β families, all of which have been found to be important in kidney morphogenesis, have been recognized as candidate substances to provide positional information (reviewed in^12,29,30). That TGF-β may be an arbiter of tubule spacing has been explored in other branching systems in which TGF-β repulses epithelial cell processes to space out the branches of the epithelial tree.^12,31,32 It is interesting to note that most of the secreted proteins involved in UB branching are heparin binding. Heparan sulfate proteoglycans are known to modulate cell surface localization of ligands and thus are ideal candidates as positive and negative regulators of signaling by morphogen gradients (Figure 1B).³³ In this context, heparan sulfate proteoglycan diversity has been proposed to be a key (although largely unexplored) driving mechanism of stage-specific regulation of UB morphogenesis.³⁴

LATE UB BRANCHING

As organogenesis comes to completion, branching slows down, presumably owing to negative feedback. In the kidney and other branching systems, it seems that members of the TGF-β superfamily are the primary molecules involved in branching inhibition, but there is mounting evidence that negative feedback signals may also arise from mesenchyme cell surface molecules.^5,14 This is anatomically suggested in the kidney as fusion between a lateral ureteric branch, and metanephric tubule effectively removes the ureteric branch from further divisions; however, signals to slow branching may be present even earlier during MM-derived tubule formation, where it has been noted that branch-inhibitory factors are expressed in comma and S-shaped bodies and thus may regulate the extent and pattern of branching (reviewed in²⁶).

Mutations in the pathways that regulate early and late UB branching usually do not result in significant branching defects.¹¹ The phenotypes that are manifest are generally quantitative leading to a decrease in nephron number. There is evidence that low nephron number in humans predisposes to hypertension and chronic kidney disease^35,36; although it is unlikely that essential hypertension is due to a loss-of-function mutation in a single gene, it is conceivable that variant alleles, identified by single-nucleotide polymorphisms, that interact differently with modifiers, suppressors, and enhancers are sufficient to cause subtle changes in nephron number that eventually lead to disease.¹¹

BRANCHING CESSATION

The events that signal the termination of branching remain enigmatic. Miscues in the cessation of kidney development can lead to a variety of renal disorders ranging from reduced nephron number to cystic kidney disease, which can be considered a disorder of tubule maintenance.¹¹ In vitro studies suggest soluble factors produced by the differentiating MM modulate the expression of specific subsets of matrix proteases that can modify the extracellular matrix and influence branch termination.¹² For example, release of endostatin at UB tips modulates UB branching and specific sizes and concentrations of the basement membrane component hyaluronic acid seems to independently regulate UB branching and promote tubular maturation.^37,38 Thus, extracellular matrix components may act as a potential switch for ending branching morphogenesis, as well as initiating nephron differentiation.

Stop/maturation signals also seem to be correlated with the differentiation of the MM.³⁹ UB-derived soluble factors that promote mesenchymal epithelialization, such as leukemia inhibitory factor, inhibits branching of the isolated UB.^14,40 In mice, loss of nephrogenic mesenchyme (as a result of conversion into nephrons) at the time of birth leads to the disruption of dichotomous branching.⁴¹ Such environmental cues seem to play a large role in the determination of the extent of UB branching, especially in the case of maternal nutrient restriction, which seems to affect both nephron number and differentiation.^42,43 Recombination of the MM with the UB cultured in isolation suggests that it is able to regulate branch tapering and luminal caliber, raising the possibility that a “second hit” within the MM plays a key modulating role in the range of cystic or renal disease phenotypes.

It becomes clear from the concepts discussed here that UB branching is a very complex process, yet from the systems, or “nephrome,” perspective, UB branching can be viewed as elegantly efficient: Using a basic set of molecular building blocks to fashion the complex three-dimensional structure of the ureteric tree out of a single bud from the WD. These dramatic changes can potentially occur through small alterations in the signaling network, resulting in a branching module. It seems likely that this branching module is conserved among multiple branching organs,²⁶ and although it seems plausible to hypothesize a unifying theory of epithelial branching through the establishment of an autocatalytic network, there are undoubtedly pathways that are stage and organ specific. It will be through elucidation of those pathways that we will understand what makes a kidney a kidney and why perturbation of specific pathways results in renal disease.

DISCLOSURES

None.