Learn more: PMC Disclaimer | PMC Copyright Notice
WNK1, a kinase mutated in inherited hypertension with hyperkalemia, localizes to diverse Cl−-transporting epithelia
Abstract
Mutations in WNK1 and WNK4, genes encoding members of a novel family of serine–threonine kinases, have recently been shown to cause pseudohypoaldosteronism type II (PHAII), an autosomal dominant disorder featuring hypertension, hyperkalemia, and renal tubular acidosis. The localization of these kinases in the distal nephron and the Cl− dependence of these phenotypes suggest that these mutations increase renal Cl− reabsorption. Although WNK4 expression is limited to the kidney, WNK1 is expressed in many tissues. We have examined the distribution of WNK1 in these extrarenal tissues. Immunostaining using WNK1-specific antibodies demonstrated that WNK1 is not present in all cell types; rather, it is predominantly localized in polarized epithelia, including those lining the lumen of the hepatic biliary ducts, pancreatic ducts, epididymis, sweat ducts, colonic crypts, and gallbladder. WNK1 is also found in the basal layers of epidermis and throughout the esophageal epithelium. The subcellular localization of WNK1 varies among these epithelia. WNK1 is cytoplasmic in kidney, colon, gallbladder, sweat duct, skin, and esophagus; in contrast, it localizes to the lateral membrane in bile ducts, pancreatic ducts, and epididymis. These epithelia are all notable for their prominent role in Cl− flux. Moreover, these sites largely coincide with those involved in the pathology of cystic fibrosis, a disease characterized by deranged epithelial Cl− flux. Together with the known pathophysiology of PHAII, these findings suggest that WNK1 plays a general role in the regulation of epithelial Cl− flux, a finding that suggests the potential of new approaches to the selective modulation of these processes.
Mendelian forms of high and low blood pressure have revealed the central role of renal salt handling in blood pressure variation (1). Recently, mutations in two members of the WNK (with no K = lysine) family of serine–threonine kinases have been shown to cause pseudohypoaldosteronism type II (PHAII; OMIM no. 145260) (2). PHAII is an autosomal dominant disorder characterized by hypertension, with hyperkalemia (despite normal glomerular filtration) and renal tubular acidosis caused by impaired renal K+ and H+ excretion (3). These features are all chloride-dependent (4, 5).
WNK1 and WNK4 are both expressed in the kidney and are exclusively found in the distal convoluted tubule and collecting duct, sites involved in the determination of net salt reabsorption as well as net K+ and H+ secretion (2). WNK1 is cytoplasmic, whereas WNK4 predominantly localizes to the tight junction complex (2). These findings have established a role for the WNK kinases in a previously unrecognized signaling pathway involved in electrolyte homeostasis and blood pressure control. The physiologic abnormalities resulting from WNK mutations can be explained as the result of a primary increase in Cl− reabsorption in the distal nephron, which would be anticipated to raise blood pressure and impair H+ and K+ secretion (2).
Whereas WNK4 expression is limited to the kidney (2), WNK1 transcripts are present in many tissues in both human and rat (2, 6). This broad expression of WNK1 mRNA suggests a physiologic role for extrarenal WNK1, although the cell types and subcellular localization within these tissues are unknown. A more complete understanding of the biological function of WNK1 will require a detailed analysis of its tissue distribution and localization. We now report the extrarenal distribution of WNK1, with the notable result that WNK1 is predominantly localized to epithelia known to be involved in Cl− flux.
Methods
Northern Blot Analysis.
A segment of mouse WNK1 orthologous to exons 12–14 of human WNK1 (GenBank accession no. NM_018979) was amplified from mouse kidney cDNA by using specific primers and its identity verified by DNA sequencing; the segment was radiolabeled by random priming in the presence of 32P-labeled dCTP, and hybridized to poly(A)+ RNA of a mouse multiple tissue Northern blot according to the manufacturer's instructions (CLONTECH). After hybridization, blots were washed and exposed to x-ray film.
Preparation and Characterization of Antibodies.
Affinity-purified antibodies specific for WNK1 by both immunohistochemistry and Western blotting were prepared and characterized as described (2). The immunizing peptide comprising amino acids 1017–1033 of WNK1 is encoded in exon 12. Other primary antibodies used included a rat monoclonal anti-ZO-1 antibody (gift of James Anderson), a goat polyclonal anti-aquaporin-2 antibody (Santa Cruz Biotechnology), and a goat polyclonal anti-CFTR antibody (cystic fibrosis transmembrane conductance regulator; Santa Cruz Biotechnology). Affinity purified donkey anti-rat or goat IgG secondary antibodies were conjugated to the CY2, CY3, AMCA, or CY5 fluors (Jackson ImmunoResearch).
Tissue Preparation.
A survey of mouse tissue was used to study the immunolocalization of WNK1. Male mice consuming a normal chow diet were killed by cervical dislocation at age 10–12 weeks. Excised tissue was embedded in OCT mounting medium and snap frozen by immersion in isopentane at −140°C. Prepared blocks were then stored at −80°C until sectioning. In addition, frozen normal human skin and colon blocks were obtained from the Research Histology division of the Yale University Department of Pathology. Sections (5 μm) were cut and used for immunohistochemistry. These studies were approved by the Yale Human Investigation Committee and the Yale Animal Care and Use Committee.
Immunohistochemistry.
Tissue sections were processed and incubated with primary (rabbit anti-WNK1 (1:300), and rat anti-ZO-1 (1:100), goat anti-CFTR (1:400), or goat anti-AQP-2 (1:300)] and secondary antibodies as described (2). Stained sections were then visualized by means of immunofluorescence light and/or confocal microscopy.
All immunostaining with anti-WNK1 antibody was competed with a 3-fold excess of the immunizing peptide. In all tissues examined, WNK1 staining was eliminated by peptide competition, and staining with secondary antibody alone revealed no signal.
Results
Expression of WNK1 mRNA in Mouse.
The tissue distribution of mouse WNK1 mRNA was examined by hybridization of exons 12–14 of mouse WNK1 cDNA to a mouse multiple tissue Northern blot (Fig. (Fig.1).1). WNK1 transcripts are present in all tissues surveyed, with strongest signal in kidney, heart, and testis. Two major transcripts of ≈8 and 10 kb are observed. The 8-kb isoform is expressed most highly in kidney, whereas the 10 kb isoform is the predominant transcript in other tissues. This pattern is similar to that seen in human and rat (2, 6).
Northern analysis of mouse WNK1 transcripts. A mouse WNK1 cDNA probe orthologous to exons 12–14 of human WNK1 (GenBank accession no. NM_018979) was hybridized to RNA from a mouse multiple tissue Northern blot (CLONTECH). WNK1 transcripts are most strongly expressed in kidney, heart, and testis. Two transcripts of ≈8 and 10 kb are observed. The 8-kb isoform is predominant in kidney, whereas the 10-kb isoform is the major transcript in other tissues. The positions of the RNA size standards in kilobases are indicated.
Immunolocalization of WNK1.
The localization of WNK1 within different organs was examined by immunofluorescence microscopy using specific anti-WNK1 antibodies. As reported (2), WNK1 staining in the kidney is confined to the distal nephron, with staining in both the distal convoluted tubule and the collecting duct. In both nephron segments, WNK1 is cytoplasmic (data not shown).
As in kidney, WNK1 is not expressed in all cell types in other tissues, but is confined to polarized epithelia. For example, in liver, WNK1 immunolocalization reveals expression that is confined to cholangiocytes, the cuboidal cells that line the intrahepatic and extrahepatic bile ducts, which mediate the flow of bile from hepatic parenchyma (Fig. (Fig.2).2). Interestingly, in contrast to its cytoplasmic localization in kidney, WNK1 in bile ducts is predominantly localized to the lateral membrane (Fig. (Fig.2).2).
WNK1 immunolocalization in hepatic bile duct epithelium. Frozen sections of mouse liver were stained with anti-WNK1 (red) and anti-ZO-1 (green) antibody as described in Methods and analyzed by fluorescence microscopy. WNK1 is localized to cholangiocytes, cuboidal epithelial cells that line the biliary tract lumen. (A) A large intrahepatic bile duct. (Top) Composite (COMP) staining for WNK1 and ZO-1. (Middle and Bottom) Staining with each antibody alone. (B) Demonstration of WNK1 localization in a smaller bile ductule. As in A, WNK1 is localized to the lateral membrane (demonstrated by the arrow), whereas ZO-1 is localized to the tight junction. L, bile duct lumen.
In the pancreas, exocrine secretions flow from serous cells, to intercalated ducts, then intralobular and interlobular ducts, and finally the main pancreatic duct. WNK1 expression in pancreas is confined to the cuboidal epithelial cells of the interlobular and main pancreatic ducts (Fig. (Fig.3).3). As in liver, WNK1 localization is predominantly along the lateral membrane of these ducts. WNK1 was not found in intralobular ducts, acinar cells, or the islets of Langerhans.
Expression of WNK1 in exocrine pancreatic duct epithelium. Frozen sections of mouse pancreas were stained with anti-WNK1 (red) and anti-ZO-1 antibody (green) and analyzed by fluorescence microscopy as in Fig. Fig.2.2. A longitudinal section of pancreatic ducts is shown. WNK1 is present in the lateral membrane of cuboidal epithelial cells of large main exocrine pancreatic ducts (arrows). WNK1 is not expressed in smaller intralobular ducts (arrowheads).
In epididymis, WNK1 is expressed throughout the pseudostratified columnar epithelium, and its localization varies with cell type (Fig. (Fig.44 A–C). WNK1 becomes more tightly membrane associated with lateral accentuation with progression from basal cell layers to superficial layers. Costaining with the tight-junction protein ZO-1 demonstrates overlap of WNK1 and ZO-1 in the apical junctional complex in superficial layers, with WNK1 extending along the lateral membrane (Fig. (Fig.44 D–F). No staining of testis is observed.
WNK1 immunolocalization in epididymal epithelium. Frozen sections of mouse epididymis were stained with anti-WNK1 (red) and anti-ZO-1 antibody (green) and analyzed by fluorescence microscopy (A–C) and confocal microscopy (D–F). (A) WNK1 expression is seen in the columnar epithelial cells of the epididymis. WNK1 becomes more tightly membrane associated with lateral accentuation with progression from basal cell layers (Ba) to superficial layers (Su) of this pseudostratified epithelium. (B) Same view as in A, with ZO-1 staining alone. (C) Same view as in A, with WNK1 staining alone. (D) Confocal composite of WNK1 and ZO-1 demonstrating the lateral localization of WNK1 in epididymal epithelium. (E) Same view as in D, showing ZO-1 staining alone. (F) Same view as in D, showing WNK1 staining alone. Arrow demonstrates the lateral expression of WNK1, beginning at the tight junction.
WNK1 is also present in the gallbladder, where it is found in the single layer of tall columnar cells of the polarized epithelium lining the gallbladder lumen (Fig. (Fig.55A). Like its localization in kidney, WNK1 is cytoplasmic in gallbladder epithelium (Fig. (Fig.55B).
Localization of WNK1 in gallbladder epithelium. Frozen sections of mouse gallbladder were stained with anti-WNK1 antibody (red) and analyzed by fluorescence microscopy. (A) A low-power view demonstrates WNK1 staining of the simple columnar surface epithelium lining the gallbladder lumen (L). (B) A higher magnification demonstrates that WNK1 is cytoplasmic in gallbladder epithelium.
WNK1 localization was investigated in human skin and mouse esophagus. In the epidermis, WNK1 expression is prominent in the cytoplasm of keratinocytes in the basal layer (Fig. (Fig.66A). WNK1 is not seen in other layers of the epidermis. Similarly, WNK1 is also expressed throughout the stratified squamous epithelium lining the esophageal lumen (Fig. (Fig.66 B and C); the localization is cytoplasmic in basal layers but becomes more associated with the membrane in superficial layers.
WNK1 immunolocalization in epidermal basal keratinocytes and esophageal epithelium. Frozen sections of human skin and mouse esophagus were stained with anti-WNK1 antibody (red). (A) In the epidermis of human skin, WNK1 is expressed in the cytoplasm of basal keratinocytes, as indicated by arrow (SCo, stratum corneum epidermal layer; SBa, stratum basal epidermal layer). (B) A low-power view demonstrates WNK1 staining throughout the stratified squamous epithelium lining the esophageal lumen (L). (C) A higher-power view of B demonstrates that WNK1 is cytoplasmic in basal layers (Ba) but becomes more associated with the membrane in more superficial layers (Su).
These localizations of WNK1 are striking in that they strongly overlap with epithelia known to transport Cl−, which are involved in the pathogenesis of cystic fibrosis (see Discussion). This led us to examine expression of WNK1 in two additional epithelia involved in Cl− transport and cystic fibrosis: eccrine sweat ducts and colonic crypts. WNK1 shows striking expression in the cuboidal epithelial cells lining the lumen of eccrine sweat ducts in human skin (Fig. (Fig.7).7). In these cells WNK1 is localized to the cytoplasm. Moreover, WNK1 is also expressed in human colon, where it is localized in the epithelia of colonic crypts (Fig. (Fig.8).8). Here WNK1 is cytoplasmic, in contrast to the apical localization of CFTR.
Expression of WNK1 in eccrine sweat duct epithelium. (A) Sweat ducts of human skin were stained with anti-WNK1 antibody (red) and 4′,6-diamidino-2-phenylindole (purple) to identify nuclei in epithelial and surrounding cells. WNK1 is expressed in the cuboidal epithelial cells of eccrine sweat ducts. In these cells, WNK1 is cytoplasmic. (B) Same view as in A, showing WNK1 staining alone. Arrow, a sweat duct lumen in cross section. Arrowhead, a sweat duct in longitudinal section.
Expression of WNK1 in colonic crypt epithelium. Sections of human colon were stained with anti-WNK1 (red) and anti-CFTR (green) antibody, as well as 4′,6-diamidino-2-phenylindole (purple). WNK1 is expressed in the cytoplasm of colonic crypt epithelia, in contrast to the apical localization of CFTR in the same crypts. (A) Transverse section showing staining of the epithelium of colonic crypts. (B) Cross section of a colonic crypt (L = crypt lumen).
Discussion
The present study demonstrates that, despite its expression in many organs, extrarenal WNK1 is not ubiquitous but is predominantly confined to polarized epithelia. In addition, the cellular localization of WNK1 varies among tissues; WNK1 is cytoplasmic in some tissues but is associated with the lateral membrane in others. Both the significance and the mechanism of this varied localization is unclear. It is known that there is alternative splicing of WNK1 (2, 6, 7), and it will be of interest to determine whether these alternative splice forms account for its varied localization. Despite high levels of WNK1 mRNA, convincing staining of heart with this anti-WNK1 antibody was not seen (data not shown), raising the question of whether this organ contains a different WNK1 isoform. These issues will require further investigation.
The localization of both WNKs to the epithelium of the distal nephron and the phenotype resulting from their mutation is consistent with their role in the regulation of Cl− reabsorption in these nephron segments. The pattern of WNK1 localization in extrarenal tissues strongly suggests a similar role in these tissues, as WNK1 is largely confined to polarized epithelia; these epithelia are known to play an important role in Cl− flux (8–21).
For example, WNK1 is present in the ductal epithelium of eccrine sweat glands. Here, cells of the secretory coil produce an isotonic fluid from a blood ultrafiltrate. The ductal epithelium of the sweat gland modifies the composition of this fluid by apical Na+ absorption (8); this establishes an electrochemical gradient that drives transcellular Cl− reabsorption via CFTR-regulated Cl− channels, and also by a paracellular pathway across tight junctions (8).
Similarly, the gallbladder epithelium is responsible for concentrating biliary secretions more than 10-fold by electroneutral Na+-coupled Cl− transport accompanied by passive water flux (9–11). Evidence suggests that most of this Cl− transport occurs via the paracellular route (11). Even in epidermis and esophagus, epithelia not classically thought to transport electrolytes, there is recent evidence of significant Cl− flux (12–14).
In colonic epithelium, absorption of NaCl is achieved by two processes: electroneutral Na+/H+ and Cl−/HCO3− exchange, and electrogenic Na+ transport via ENaC, which drives both transcellular and paracellular absorption of Cl− (15). CFTR mediates both absorption and cAMP-dependent secretion of Cl−, and is predominantly expressed in crypts (15).
Finally, hepatic bile ducts (16–18), exocrine pancreatic ducts (19), and epididymis (8, 20) share a common mechanism of Cl− handling. In all three, net HCO3− secretion is achieved by coupling Cl−/HCO3− exchange to CFTR-regulated Cl− secretion. It is striking that in each of these epithelia, WNK1 localizes to the lateral membrane rather than the cytoplasm. Determining whether this correlation reflects a shared mechanism will require further study.
The presence of WNK1 in a wide variety of epithelia that have prominent Cl− transport, coupled with the demonstrated role of WNK1 in the regulation of renal Cl− transport, suggests that WNK1 may play a general role in the determination of Cl− flux, with similar or related targets in different tissues. The downstream targets of WNK1 in extrarenal tissues are unknown; however, the thiazide-sensitive Na–Cl cotransporter (NCCT) lies downstream of WNK4 in the kidney (21). This observation raises the possibility that one extrarenal target of WNK1 signaling is the related Na-K-2Cl cotransporter (NKCC1). NKCC1 is expressed in diverse epithelia, and its activity is regulated by intracellular Cl− via a kinase-dependent mechanism (22). The observed modulation of WNK1 kinase activity by salt (6) is consistent with a role in regulation of NKCC1 activity.
Moreover, the distribution of WNK1 expression is notable for its overlap with tissues that show pathology or physiologic defects in cystic fibrosis because of loss of the CFTR gene product (23–28). These abnormalities in cystic fibrosis include defective Cl− absorption in sweat ducts (24), impaired pancreatic function caused by inadequate ductal HCO3− secretion with sludging of secretions in exocrine ducts (25), male sterility caused by obstruction and atrophy in the distal epididymis (26), hepatic fibrosis/cirrhosis caused by inspissated biliary secretions (27), meconium ileus with distal intestinal impaction caused by defective Cl− transport and abnormal electrolyte absorption in the colon (28), and occasional cases of cholelithiasis caused by deranged gallbladder salt homeostasis (29). This striking overlap raises the possibility that WNK1 and CFTR may be involved in shared or parallel physiologic pathways that regulate Cl− flux in these epithelia. These correlations afford the speculation that pharmacologic modulation of WNK1 function could have an effect on Cl− flux in patients with cystic fibrosis. It is nonetheless clear that substantial work will be necessary to assess the relevance of these speculations. Determining the upstream regulators and downstream targets of the WNK1 signaling pathway in these tissues, as well as the physiologic consequence of selective knockout of WNK1 function in different epithelia, are approaches that have the potential to address this issue.
Acknowledgments
We thank Jim Anderson for his gift of the monoclonal anti-ZO-1 antibody, and Gerardo Gamba, Ali Gharavi, Ali Hariri, David Geller, Christoph Rahner, and Isabel Beerman for helpful discussions. This work was supported in part by the National Institutes of Health Specialized Center of Research in Hypertension. K.T.K. is a Howard Hughes Medical Student Research Fellow. K.A.C. is a trainee of the Medical Scientist Training Program. R.P.L. is an investigator of the Howard Hughes Medical Institute.
Abbreviations
| PHAII | pseudohypoaldosteronism type II |
| WNK | with no lysine = K kinase |
| CFTR | cystic fibrosis transmembrane regulator |