Epithelial polarity and differentiation in polycystic kidney disease

Renal cysts are the central pathological feature of a number of human congenital and acquired diseases. The two most common forms of genetically determined renal cystic disease, autosomal dominant polycystic kidney disease (ADPKD) and autosomal recessive polycystic kidney disease (ARPKD) affect over 500,000 individuals in the United States (Welling and Grantham, 1991). Thus, polycystic kidney disease is more common than the combination of trisomy 21, sickle cell anemia, cystic fibrosis, hemophilia and Duchenne’s Muscular Dystrophy, and is responsible for 8 to 10% of all patients receiving dialysis or transplantation for end stage renal disease. The cost of these therapies alone, apart from the significant morbidity and mortality caused by renal cystic diseases, was approximately $300,000,000 in 1992. In both ADPKD and ARPKD progressive accumulation of fluid in a subpopulation of renal tubules leads to distortion and destruction of normal adjacent renal tissue and, ultimately, to renal failure. Despite intensive investigation, the genes responsible for ADPKD and ARPKD have not been identified (Germino and Somlo, 1992). Our efforts have predominantly focused on the pathophysiology of renal cyst formation and progressive enlargement (Avner, 1988, 1993; Avner et al., 1990). Data generated from such studies may focus molecular genetic approaches in identifying candidate genes, and, even in the absence of gene identification, may lead to specific therapeutic strategies directed at reducing the formation and enlargement of cystic lesions. This review will focus on our studies in the C57 BL/6J cpk/cpk (CPK) mouse, a murine model of ARPKD (Avner et al., 1987a,b, 1988), and specifically address the role of altered epithelial polarity and differentiation in PKD.

In 1977, a spontaneous mutation in the C57 BL/6J murine strain produced animals with autosomal recessive polycystic kidney disease (CPK) (Russell and McFarland, 1977). The cystic CPK strain has subsequently been maintained through controlled breeding of obligatory heterozygotes. CPK animals that appeared normal at birth developed progressive lethargy, abdominal protuberance and wasting, and died in renal failure at 3-4 weeks of postnatal age with massively enlarged kidneys (Preminger et al., 1982). We have studied the morphology and ontogeny of tubular cyst formation in the CPK mouse by light and transmission electron microscopy, and intact nephron microdissection (Avner et al., 1987a). The earliest morphological alterations consisted of tubular dilatation and cyst formation in proximal tubular segments of 17-day affected fetuses. Cysts occurred as outpouchings of proximal tubular segments and remained in continuity with other nephron segments. Transmission electron microscopy revealed widening of the tubular intercellular spaces without evidence of cell injury. Nephron microdissection revealed a shift in site of nephron involvement from proximal tubules to cortical and outer medullary collecting tubules as the disease progressed. The distinct shift in the site of nephron involvement at specific stages of renal organogenesis suggests stage-specific alterations in cystic gene expression or a developmentally regulated pattern of tubular responsiveness to cyst-promoting processes.

Studies in a variety of in vivo and in vitro experimental models as well as mathematical analysis of cyst growth kinetics demonstrate that two basic criteria necessary for cyst formation are increased epithelial cell proliferation and altered transtubular fluid transport (Avner et al., 1990; Welling, 1990; Wilson and Sherwood, 1991). Altered fluid transport in renal cystic tubular epithelium leading to net tubular secretion and intratubular fluid accumulation could result from alterations in ion pump activity or more global abnormalities of polarized cell structure and function. In normal renal tubular epithelial cells, sodium-potassium adenosine triphosphatase (Na⁺,K⁺-ATPase) is restricted to the basal-lateral membrane. Through coupled sodiumpotassium countertransport, Na⁺,K⁺-ATPase is the major driving force for renal tubular sodium reabsorption and the secondary active transport of a number of other solutes (Skou and Esmann, 1992). We, therefore, sought to determine whether alterations in Na⁺,K⁺-ATPase activity or polarized distribution were correlated with tubular cyst formation and enlargement. In CPK kidneys, the earliest phases of proximal tubular cyst formation were paralleled by increases in whole kidney Na⁺,K⁺-ATPase activity (Avner et al., 1988). Increased renal Na⁺,K⁺-ATPase activity occurred before significant epithelial hyperplasia and was not associated with abnormalities in tubular basal laminae glycoprotein expression. This suggested that increased Na⁺,K⁺-ATPase activity with consequent tubular epithelial hyperplasia were early markers of CPK gene expression and cyst formation. Subsequent studies suggested that proximal tubular cyst formation in this murine model, like cyst formation in organ culture models of developing renal tissue, might involve increased organic anion secretion driven by a normally localized basolateral Na⁺,K⁺-ATPase (Avner et al., 1985, 1987b, 1988, 1989, 1990). The intratubular sequestration of osmotically active organic anions would then obligate net intratubular fluid accumulation, resulting in cyst formation and progressive enlargement (Avner, 1988; Avner et al., 1990).

Altered Na⁺,K⁺-ATPase activity might also mediate intratubular fluid accumulation and cyst formation if the ion pump was mislocated to apical membranes of cystic tubular cells. In this location, it could stimulate net basal to apical vectorial transport of sodium and fluid, resulting in tubular fluid secretion and cyst formation. Such mislocation has recently been reported in cultured ADPKD epithelium (Wilson and Sherwood, 1991; Wilson et al. 1991). In collaboration with Dr W. J. Nelson, we, therefore, sought to determine whether abnormal Na pump distribution was associated with cystic tubular maldevelopment in the CPK model (Avner et al., 1992). Such studies were of particular interest given the developmental stage specific formation of cysts in distinct nephron populations in this model, and the fact that Na⁺,K⁺-ATPase was reported to be transiently expressed in the apical membrane of collecting tubule epithelium during normal renal development (Holthofer, 1987; Minuth et al., 1987). In addition, the Na⁺,K⁺-ATPase is sorted to both apical and basal-lateral domains during establishment of polarity in MDCK cells, a canine renal epithelial cell line (Hammerton et al., 1991).

In these studies, kidney tissue was obtained from control C57BL/6J and cystic CPK mice at 7 postnatal stages: days 0 (newborn), 3, 5, 8, 10, 12 and 21. In CPK mice, day 0 through 5 reflect a proximal and early collecting tubule stage of cystic development, days 8-12 represent predominant cystic collecting tubule development, and day 21 represents the terminal phase of cystic disease (Avner et al., 1987a). Sequential serial sections cut parallel to the long axis of the kidney were stained with antibodies specific for either the a or b subunit of the Na⁺,K⁺-ATPase, biotinylated lectins and a series of antibodies to apical and basal lateral membrane marker proteins. Apical membrane markers included y-glutamyltranspeptidase, GP-135, and the lectins Lotus tetragonologbus agglutinin (LTA) and Dolichos biflorus agglutinin (DBA). Basolateral domain markers included ZO1, type 4 collagen, laminin, entactin, band 3 anion exchanger, and carbonic anhydrase II. Staining patterns were evaluated separately in outer cortical, inner cortical and medullary zones, which permitted direct comparison of nephron segments in control and cystic tissue at similar developmental stages. Distributions of Na⁺,K⁺-ATPase subunits were compared in proximal tubules, proximal tubular cysts, collecting tubules and collecting tubule cysts through systematic evaluation of lectin profiled serial sections. By this method, lectin staining patterns allow clear discrimination of proximal tubules and proximal tubule cysts (LTA positive, DBA negative) from collecting tubule and collecting tubule cysts (LTA negative, DBA positive) in developing murine renal tissue (Avner and Sweeney, 1990; Sweeney and Avner, 1991).

Analysis of control proximal tubules demonstrated restriction of both the αl and β1 subunits of Na⁺,K⁺-ATPase to the basal-lateral membrane domain at all developmental stages (Fig. 1A). Similarly, in cystic CPK proximal tubules, Na⁺,K⁺-ATPase subunits were expressed exclusively in basal-lateral membranes (Fig. IB). Apical membrane expression of Na⁺,K⁺-ATPase subunits was not detected in control proximal tubules or unaffected or cystic CPK proximal tubules at any developmental stage. Analysis of other marker proteins of the apical or basal-lateral membrane domain revealed normal distributions in proximal tubule cells.

In the majority of control collecting tubules both al and pl subunits of Na⁺,K⁺-ATPase were localized to the basallateral membrane of cells in all nephrogenic zones. However, a subpopulation of control collecting tubules in the outer and inner cortical zones of active tubulogenesis demonstrated apical-lateral as well as low intensity basal membrane staining of both Na⁺,K⁺-ATPase subunits (Figs 1C and 2A,B). Apical Na⁺,K⁺-ATPase expression was observed in a maximum of 16% of outer cortical, 6% of inner cortical and 2% of medullary collecting tubules at days 0 through 5 (Fig. 2). In all zones, apical membrane Na⁺,K⁺-ATPase staining declined progressively at subsequent developmental stages as the normal basal-lateral membrane staining pattern became prominent. These data demonstrate that Na⁺,K⁺-ATPase is transiently localized to both apical and lateral membranes of epithelial cells as a normal feature of renal collecting tubule, but not proximal tubule, development. Apical membrane Na⁺,K⁺-ATPase expression is a normal, transient phenotypic feature of early collecting tubule differentiation that is lost with subsequent cell maturation.

In cystic CPK collecting tubules, apical and lateral membrane staining of Na⁺,K⁺-ATPase al and pl subunits was significantly increased when compared with control tubules (Figs ID and 2). A maximum of 63% of medullary cystic tubules, 46% of inner cortical, and 47% of outer cortical collecting tubule cysts expressed apical membrane Na⁺,K⁺-ATPase. Significant differences were present between Na⁺,K⁺-ATPase distributions in cystic and control collecting tubules at progressive developmental stages (Fig. 2). In all nephrogenic zones, the percentage of cystic collecting tubules with apical membrane Na⁺,K⁺-ATPase distribution decreased over time paralleling the normal developmental pattern seen in control collecting tubules. No abnormalities in immunolocalization of other apical or basal lateral cell surface marker proteins were detected at any stage of development in control or cystic collecting tubules. Like immature control collecting tubules, cystic collecting tubules demonstrated apical and lateral membrane Na⁺,K⁺-ATPase expression. In parallel with control collecting tubules, the percentage of cystic collecting tubules with apical and lateral membrane Na⁺,K⁺-ATPase expression decreased in relation to the total cystic nephron population at progressive developmental stages. Thus, a large proportion of cystic collecting tubules exhibit a relatively undifferentiated phenotype in terms of the membrane distribution of Na⁺,K⁺-ATPase. This conclusion is consistent with recent reports that cystic CPK collecting tubule epithelium exhibits elevated steady state mRNA levels of the SGP2 gene and a variety of protooncogenes that are normally expressed during an early stage of collecting tubule differentiation (Cowley et al., 1991; Harding et al., 1991). Further, ultrastructural analysis of cystic CPK collecting tubules reveals characteristic anatomical features of a relatively undifferentiated epithelium (Fig. 3).

Distributions of other apical or basolateral membrane marker proteins were unaffected in immature control and cystic collecting tubule epithelial cells. Apical membrane distribution is, thus, not a result of general deregulation of mechanisms involved in establishing and maintaining cell surface polarity. Further, the fact that Na⁺,K⁺-ATPase distribution is restricted to the basal lateral membrane of both normal and cystic proximal tubular epithelium indicates that CPK cells are capable of normal Na⁺,K⁺-ATPase sorting. On the basis of these data we have speculated that the sorting pathway of Na⁺,K⁺-ATPase is the same in normal and cystic epithelial cells, but that Na⁺,K⁺-ATPase retention in the apical membrane is transient in normal development, but persists in cystic epithelium. At present, it is not clear whether retention of Na⁺,K⁺-ATPase on the apical membrane during normal collecting tubule development or in cystic tubule epithelial cells is due to alterations the assembly of the membrane cytoskeleton at that membrane. Similarly, it is not known whether the distribution of glycosphingolipids is altered in the apical membrane of immature control or cystic tubule cells, or whether retained apical membrane Na⁺,K⁺-ATPase functions as an active Na pump. Studies are currently underway to characterize the distribution of membrane cytoskeletal proteins and glycosphingolipids in control and cystic epithelial cells, and determine whether Na⁺,K⁺-ATPase activity is affected by its apical membrane microenvironment.

On the basis of our studies to date, we conclude that increased basal-lateral membrane Na⁺,K⁺-ATPase activity in cystic kidney proximal tubule epithelia may drive organic anion secretion with consequent intratubular fluid accumulation. Further, increased apical membrane Na⁺,K⁺-ATPase expression in cystic kidney collecting tubule epithelia may be a marker of the relatively undifferentiated phenotype of cyst lining cells. If such apically expressed enzyme is active, it may have pathogenic import in collecting tubule cyst formation and progressive enlargement by stimulating net basal to apical vectorial solute and fluid transport.

The work described in this review was supported by National Institutes of Health grants DK34891 and DK 44875. The author acknowledges the excellent secretarial assistance of Ms Debbie McNamara in manuscript preparation.