Advertisement
Commentary Free access | 10.1172/JCI22029
1Departments of Medicine and of Physiology and Cellular Biophysics, College of Physicians and Surgeons, Columbia University, New York, New York, USA. 2Departments of Pediatrics and of Medicine, University of Rochester, Rochester, New York, USA.
Address correspondence to: Qais Al-Awqati, Departments of Medicine and of Physiology and Cellular Biophysics, College of Physicians and Surgeons, Columbia University, 630 West 168th Street, New York, New York 10032, USA. Phone: (212) 305-3512; Fax: (212) 305-3475; E-mail: qa1@columbia.edu.
Find articles by Al-Awqati, Q. in: JCI | PubMed | Google Scholar
1Departments of Medicine and of Physiology and Cellular Biophysics, College of Physicians and Surgeons, Columbia University, New York, New York, USA. 2Departments of Pediatrics and of Medicine, University of Rochester, Rochester, New York, USA.
Address correspondence to: Qais Al-Awqati, Departments of Medicine and of Physiology and Cellular Biophysics, College of Physicians and Surgeons, Columbia University, 630 West 168th Street, New York, New York 10032, USA. Phone: (212) 305-3512; Fax: (212) 305-3475; E-mail: qa1@columbia.edu.
Find articles by Schwartz, G. in: JCI | PubMed | Google Scholar
Published June 1, 2004 - More info
The collecting ducts of the kidney are composed of intercalated cells (responsible for acid/base transport), principal cells (mediating salt and water absorption), and inner medullary cells, which mediate all three types of transport. Forkhead box (Fox) genes are a large family of transcription factors that are important in cell-type specification during organogenesis. In this issue, Blomqvist et al. find that mice lacking Foxi1 have no intercalated cells in the kidney . The collecting ducts of the null mice contained primitive cells that expressed both intercalated cell and principal cell proteins, yet the acid/base transport function of the kidney was disrupted and the mice exhibited distal renal tubular acidosis. These findings suggest that Foxi1 plays a critical role in determining cell identity during collecting duct development.
The mechanism of generation of the 400 or more cell types of mammals from a single cell remains the principal problem of developmental cell biology. Classical embryologists discovered that progenitor cells first undergo specification before they become determined to form a differentiated cell type. Under certain conditions, specified cells can differentiate in a reversible manner, whereas determined cells autonomously differentiate even if placed in a different location in the embryo. The molecular mechanism of these changes in cell fate depends on signals received from the niche or “ecosystem” in which the cell is located, but ultimately it must involve the activation of a network of transcription factors that is specific for each tissue and cell type.
Each nephron of the adult mammalian kidney is composed of 12 epithelial segments. The collecting duct has three segments, some of which have more than one cell type. These tubular segments begin to develop when the ureteric bud (an outgrowth of the wolffian duct) invades the metanephric mesenchyme. The tip of each ureteric bud induces the conversion of this mesenchyme into an epithelial structure that undergoes morphogenetic transformation to form the nephron from the glomerulus to the end of the distal tubule, while the ureteric bud forms the collecting duct. The adult collecting duct has several cell types, one of which — the intercalated cell — is specialized for the transport of H+ and HCO3–. The other type — the principal cell — mediates salt and water transport. A third cell type, present only in the inner medullary collecting duct, performs salt, water, and acid transport. All of these cells derive from the ureteric bud, but the pathway of their differentiation remains obscure. The paper by Blomqvist et al. in this issue of the JCI is the first to provide insight into the molecular mechanism of cell-identity determination of at least the intercalated cell (1).
The intercalated cell exists in two functionally distinct subtypes in the cortical collecting duct: one secretes HCO3–, while the other absorbs it. Reabsorption of HCO3– (i.e., secretion of protons) is accomplished in α-intercalated cells by an apical H+-ATPase and a basolateral Cl–/HCO3– exchanger. Secretion of HCO3– by β-intercalated cells is mediated by the same kinds of transporters but located on the opposite cell membranes. We found that when animals were fed an acid diet, the β form converted to the α phenotype (2). Others have shown, using a variety of methods, that there are more than two subtypes of intercalated cells, but we interpret these findings as showing a continuity of changes taking place between the two extremes of α and β. Recent studies based on single identified β-intercalated cells showed that these cells converted to an α phenotype in response to acidification of the basolateral media (3). Recently, we found in an intercalated cell line that this conversion represents an instance of terminal differentiation, with the β-intercalated cell representing a less differentiated (protoepithelial) type and the α phenotype being more differentiated (4).
Blomqvist et al. (1) find that mice lacking Foxi1, a forkhead transcription factor, have no intercalated cells as judged by electron microscopy, or by the absence of specific proteins of these cells. Foxi1 is expressed in the intercalated cell and in the inner ear. Could the transcription factor be a critical determinant of cell identity of these cells? The authors show a remarkable result: each cell of the collecting ducts of the mutant mice expresses both intercalated proteins (carbonic anhydrase II) and principal cell proteins (aquaporin 2). It would have been useful to find out whether the embryonic collecting duct also resembled the mutant ducts in their pattern of expression. Although studies using these excellent markers have not been performed, the entire collecting duct seems to be composed of a single cell type, staining weakly for aquaporin 2 (K. Madsen, personal communication).
Nephrogenesis proceeds in mice, rats, and rabbits for one week postnatally. Electron microscopic studies have shown that the cortical collecting duct of the newborn shows a single cell type, resembling a principal cell, with a single cilium, few organelles, smooth apical and basolateral surfaces, and large glycogen deposits (5). No intercalated cells can be demonstrated in the developing neonatal collecting duct, using physiological criteria (6). Minuth and colleagues found that the early collecting tubule contains a “scrambled” expression of principal and intercalated lectin-binding sites (7). But it remains to be shown definitively that development of the collecting tubule lineage progresses from a generic ureteric bud cell to a specified collecting tubule cell that finally terminally differentiates into principal, intercalated, and inner medullary collecting duct types. Incidentally, the latter cell contains both principal and intercalated cell proteins, which raises the question of whether these cells have been arrested in their developmental program at the specification stage.
There is an interesting connection between renal anomalies and deafness. Deletion of ATP6B, of pendrin, and of Kcc4, the K-Cl cotransporter present in intercalated cells, can cause deafness. One of the supporting cells of the sensory epithelium in the vestibular epithelium has many similarities to the intercalated cell, and it serves to control the acidic pH of the endolymph. Remarkably, deletion of Foxi1 depletes the epithelium of these cells, causing deafness (8).
Forkhead genes (termed Fox, for forkhead box) encode a subset of the helix-turn-helix class of transcription factors (9). They are divided into 17 classes (A to Q) and are involved in a wide range of cell-type specification during organogenesis in animals and humans. In Caenorhabditis elegans, a forkhead protein, pha-4 (an orthologue of FoxA), specifies the identity of the pharynx and causes the expression of all pharyngeal genes (10). Remarkably, the order of appearance of the pharyngeal genes depends on the affinity of the promoters to the forkhead protein, providing an elegant mechanism that controls the timing of appearance of these genes.
Foxi1 controls the expression of pendrin and the band 3 anion exchangers (AE1), and probably also the ATP6B subunit of the vacuolar H+-ATPase, all characteristic proteins of the intercalated cells. But the absence of this factor resulted not only in the absence of these genes but also in the lack of this cell type. The intercalated cells contain a very large number of mitochondria, and their shape is quite characteristic because of the presence of vigorous apical endocytosis and exuberant microvilli. It seems unlikely that all of these characteristics are secondary events produced by expression of a few transport proteins. One possibility is that absence of the transport proteins leads to cell death during development. But since the expression of carbonic anhydrase and aquaporin 2 becomes scrambled in the mutant mice, it is likely that Foxi1 activates a whole program of expression of genes that control cell shape, organellar biogenesis, apical structure (microvilli and microplicae), and apical functions (endocytosis and exocytosis). Using microarray analysis it should be possible to identify the developmental program that specifies this cell type. But what is the developmental pathway by which Foxi1 determines the identity of the intercalated cell? Blomqvist et al. favor the idea that it has a simple and direct effect of cell determination similar to the proposed mechanism of pha-4 (1). However, it appears that Foxi1 is expressed in the kidneys at embryonic day 16 in a uniform pattern in the ureteric bud and collecting duct (ref. 11; note that this is our interpretation of Figure 2A of ref. 11, which differs from the opinion of the authors). This raises the question of whether Foxi1 plays a more complex role in cell specification in the collecting duct. For instance, its disappearance could also mediate principal cell differentiation. Forkhead genes often collaborate with other transcription factors to control the timing of differentiation, exerting negative and positive influences on these processes. For instance, Foxe1 collaborates with pax8 and TTF1 to control the differentiation of the thyroid cell. Foxe1 mutations result in thyroid agenesis (12), yet the gene is expressed in the thyroid before the differentiation of thyrocytes and is suppressed when thyroglobulin begins to be expressed (13). What seems clear is that the identification of Foxi1’s role in intercalated cell differentiation has opened the door to an exciting new chapter in the development of the kidney.
See the related article beginning on page 1560.
Conflict of interest: The authors have declared that no conflict of interest exists.