Disease processes in epithelia: the role of the actin cytoskeleton and altered surface membrane polarity

Biochimica et Biophy sica Acta, 1225 (1993) 1-13

1

© 1993 Elsevier Science Publishers B.V. All rights reserved 0925-4439/93/$06.00

Minireview

BBADIS 61316

Disease processes in epithelia: the role of the actin cytoskeleton and altered surface membrane polarity Jeffrey Leiser c a n d B r u c e A. Molitoris a,b,d Department of Medicine, University of Colorado School of Medicine, The Veteran Affairs Medical Center, Dem,er, CO (USA), b Department of Cellular and Structural Biology, Uni~:ersity of Colorado School of Medicine, Dent~er, CO (USA), c Department of Pediatrics, University of Colorado School of Medicine, The Children's Hospital, Dent:er, CO (USA) and a Department of Medicine, Indiana Unit,ersity Medical Center, Indianapolis, IN (USA) (Received 14 April 1993)

Key words: Epithelial polarity; Cytoskeleton; Ischemia: Proximal tubule; Microvillous inclusion disease; Polycystic kidney disease

Contents I.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

II.

Characteristics of polarized epithelial cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Distinct surface m e m b r a n e domains .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Cytoskeletal and junctional organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Establishment and maintenance of epithelial cell polarity . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 2 3 5

III. Altered epithelial cell polarity in pathologic states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Ischemia in kidney proximal tubule cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Microvillus inclusion disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Polycystic kidney disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8 8 10 10

IV. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11

I. Introduction

Polarized epithelial cells play an essential role in the vectorial movement of ions, water and macromolecules

Correspondence to: B.A. Molitoris, D e p a r t m e n t of Medicine, Indiana University School of Medicine, 1120 South Drive, Indianapolis, IN 46202, USA. Fax: + 1 (317) 274-8575. Abbreviations: PKD, polycystic kidney disease; A D P K D , autosomal dominant PKD; A R P K D , autosomal recessive PKD; BLM, basolateral membrane; BBMV, brush-border m e m b r a n e vesicles; C / P L , cholesterol/phospholipid; EGF, epidermal growth factor; MID, microvillous inclusion disease; PC, phosphatidylcholine; PI, phosphatidylinositol; SPH, sphingomyelin; TGN, trans-Golgi network.

between biological compartments. These vectorial processes include absorption (e.g., enterocytes and renal proximal tubule cells), secretion (e.g., hepatocytes, endocrine, and exocrine cells) and exchange (e.g., alveolar and capillary endothelial cells). In order to conduct these processes, the surface membranes of polarized epithelial cells are organized into distinct apical and basolateral domains. These domains are distinguishable biochemically, structurally and physiologically, each containing specific ion channels, transport proteins, enzymes and lipids. The establishment and maintenance of this specialized organization is a multistage process involving the formation of cell-cell and cell-substratum contacts, and the targeted delivery of plasma membrane and cyto-

plasmic components to the appropriate domains. The actin cytoskeleton plays an increasingly understood role in these processes, specifically interacting with the surface membrane at both junctional and non-junctional sites, and being involved in the transport of surface membrane-destined vesicles. Aberrations in the actin cytoskeleton affect these processes, interfering with the specialized organization of polarized epithelial cells. This, in turn, results in cell- and tissue-specific abnormalities, and ultimately in the development of a pathologic state. This review will first describe recent findings regarding basic cellular mechanisms involved in the organization of polarized epithelial cells, with an emphasis on the actin cytoskeleton. These fundamental concepts will then by used as a foundation for discussing the known and potential roles of alterations in the actin cytoskeleton in a variety of disease processes. II. Characteristics of polarized epithelial cells

H-A. Distinct surface membrane domains Polarized epithelial cells have a characteristic cellular organization that includes organization of the surface membrane into biochemically distinct apical and basolateral domains (reviewed in Refs. 1, 2), with extensive differences in protein composition (Table I). The apical domain faces the 'external' compartment

(e.g., the urinary lumen in renal tubular cells, the intestinal lumen in enterocytes), and possesses a complement of specialized intrinsic and extrinsic proteins that primarily function in cell-specific, often vectorial, processes (e.g., absorption). The basolateral domain faces the 'internal' compartment, and possesses a complement of proteins similar to those found over the entire surface membrane of non-polarized cells. They primarily function in general cell processes, such as maintaining the normal physiologic state, and also in signal recognition and transduction. The existence of these specific apical and basolateral membrane protein compositions is essential for the efficient functioning of the polarized epithelial cell. For example, sodium reabsorption by proximal renal tubular cells is dependent on the polarized localization of specific carrier proteins such as the N a + / H + antiporter and the Na+-depen dent glucose, amino acid and phosphate cotransporters to the apical membrane, combined with localization of the Na+/K+-ATPase to the basolateral membrane (BLM). Extensive differences in membrane external leaflet lipid composition also exist between the apical and basolateral domains, and result in large physicochemical differences (Table I) [3-5]. The apical membrane cholesterol/phospholipid ( C / P L ) and sphingomyelin/phosphatidyicholine ( S P H / P C ) ratios are high, resulting in high anisotropy and insulating capacity. In contrast, the BLM C / P L and S P H / P C ratios are low,

TABLE I

Biochemical asymmetry of the surface membrane of polarized epithelial cells

Proteins Enzymes

Apical membrane

Basolateral membrane

Leucine aminopeptidase Maltase GPl-linked proteins (e.g. alkaline phosphatase)

Adenylate cyclase

Insulin

Receptors

Parathyroid hormone

ATPases Carriers

H +-ATPase Mg2+-ATPase Amiloride-sensitive Na + channel Na +-dependent cotransporters Na+/K+/2CI - cotransporter Na+/H + antiporter

Epidermal growth factor Laminin Na +/K +-ATPase Ca2+-ATPase CI /HCO 3 exchanger Na +-independent glucose carrier

Lipids Sphingomyelin Phosphatidylcholine Phosphatidylinositol

High High Low LOw

Low Low High High

Physical properties Electrical resistance Membrane fluidity

High Low

High

Cholesterol

Low

teins mediating such interactions are the cadherins and the integrins. Cadherins are a superfamily of transmembrane glycoproteins that bind with each other in a homophilic manner to mediate Ca2+-dependent cell-cell adhesion (reviewed in Refs. 10, 11). The family of 'classical' cadherins includes at least 12 members, including uvomorulin (E- or epithelial cadherin). However, a variety of cadherin-like molecules with similar amino-acid sequences, but different overall structures, have also been identified, e.g., desmoglein I and the desmocollins. The extracellular self-recognition sequence, at the N-terminus, is unique to each cadherin. Specific intracellular domains, at the C-terminus, associate with

with the phosphatidylcholine (PC) and phosphatidylinositol (PI) contents being relatively high. This results in a membrane that is more fluid (low anisotropy) and across which diffusion can occur more rapidly. Membrane lipid composition and anisotropy also influence the function of intrinsic membrane proteins, e.g., Na+/K+-ATPase [6-8] and the Na÷-dependent glucose cotransporter [9].

II-B. Cytoskeletal and junctional organization Establishment and maintenance of cell surface membrane polarity is dependent on cell-cell and cellsubstratum contacts. Two important families of pro-

A

APICAL " ~.~("

\L~,]"

~

\'

.,,~'~,~"

.]lIIUUII-U

MyoslnI

-,=

Vlllln, Flmbrin

o

iJ I

MEMBRANE

B

.++++..z+,,++,.c,. mlcrofilaments

II I

'::'."~'~'::~-] ZO

TW

.~ CATEI~INS

C

,.-.,#/

1'

ii:

II

BASAL

Y

M

F-ACTIN]

CAPINNG.g PROTE'N~

~

"~

(X-ACTININ

~I~VINCUC'N

rl,+-ECM

Fig. 1. Cytoskeletal and junctional components in polarized cells. F-actin microfilaments, bundled by villin and fimbrin and bound to the surface membrane via a myosin I, form the microvillus (MV) structural core and extend into the terminal web (TW). Microtubules (MT) extend basally from the terminal web. Vesicles (V) more apically along microtubules and microfilaments using dynein and myosin I motors, respectively, and basolaterally along microtubules using a kinesin motor. The zonula occludens (ZO) encircles the apex of the cell and interacts with microvillar microfilaments. The zonula adherens (ZA) lies below this, encircling the cell and interacting with the terminal web (TW). Desmosomes (D) lie in a loose band below this, and are connected to other desmosomes and to basolateral hemidesmosomes (HD) via intermediate filaments (IF). Gap junctions (G J) interconnect adjacent cells. A cortical web (inset B), comprised of actin cytoskeletal elements, parallels the basolateral membrane surface and forms complexes with integral membrane proteins (e.g., N a + / K + - A T P a s e and cadherins) via ankyrin, protein 4.1, and possibly other proteins. Focal adhesions (FA) (inset C), containing talin, vinculin, and other proteins, link cytoskeletal F-actin bundles to the extracellular matrix (ECM) via a variety of integrins. The nucleus (N) is also labelled.

cytoplasmic proteins termed catenins. Interactions with catenins appear important both for cell-cell binding and for interactions with the actin cytoskeleton [12]. Cadherins are also known to be phosphorylated, and tyrosine kinases may be involved in their regulation. lntegrins are a family of heterodimeric transmembrane glycoproteins that mediate Ca2+-dependent cell-substratum and heterophilic cell-cell binding (reviewed in Refs. 10, 13). There are at least 20 members, generated from permutations of pairings between at least 14 a- and 8 /3-subunit isoforms. The pattern of binding to the various components of the extracellular matrix, such as fibronectin, laminin and various collagens, varies with each integrin family member, but generally involves the Arg-Gly-Asp (R-G-D) tripeptide sequence. Intcgrins have also been implicated in heterophilic cell-cell adhesion [14,15]. Integrin function appears regulated by both extracellular and intracellular signals. Binding of an extracellular ligand apparently induces conformational changes in the intracellular portion of the /3-subunit, which in turn regulate interactions with the actin cytoskeleton [16]. Integrins are also known to be phosphorylated at serine and tyrosine sites. The actin cytoskeleton interacts with the cell surface membrane at several levels, including the apical microvilli, the terminal web, cell-cell and cell-substratum junctional complexes, other (non-junctional) ceil-cell complexes, and complexes with other extrinsic and cytoplasmic membrane proteins (Fig. 1A). GTP-binding proteins may play a significant role in regulating these interactions [17,18]. Apical membrane amplification occurs via formation of individual microvilli (reviewed in Ref. 19). The structural backbone of a microvillus consists of 20-30 vertically-oriented, polarized (barbed/'plus' ends at the microvillus tip) actin microfilaments that extend along its length and into the terminal web area [20-23]. Specific actin 'bundling proteins' such as fibrin and villin serve to cross-link the individual actin microfilaments. These actin-associated proteins may also serve other regulated functions. For example, depending on Ca 2+ concentration, villin may induce actin microfilament nucleation, capping, bundling or severing [24]. Attachment of actin microfilaments to the microvillus tip involves an as yet poorly characterized protein complex, while attachment to the lateral portions of the microvillus membrane involves a brushborder myosin-I protein/calmodulin complex [21,25]. Attachment to and movement of vesicles apically along the actin microfilaments, perhaps mediated by myosin1, may also occur [26]. The microfilaments terminate in association with both the terminal web and the zonula occludens. The terminal web (reviewed in Ref. 19) is composed primarily of a dense meshwork of fine, non-actin fila-

mentous material that is oriented perpendicular to and cross-links adjacent actin microfilament rootlets. Both myosin and (non-erythroid) spectrin isoforms have been identified, and are thought to function in this cross-linkage [23]. Tropomyosin is also present, and is felt to stabilize the actin microfilaments. Junctional complexes occur along the basolateral domain of polarized epithelial cells and serve a variety of specialized functions. The zonula occludens (tight junction) encircles the apex of the cell, fusing adjacent cells together and delimiting the border between the apical and basolateral domains. It forms a barrier to intradomain movement of intrinsic membrane proteins and outer leaflet phospholipids (the 'fence' function), and to paracellular movement of solutes between biologic compartments (the 'gate' function) [27-32]. Actin microfilaments, including those from microvilli, attach directly to the zonula occludens, perhaps interacting with either of the two known protein components, ZO-1 or cingulin [33,34]. Transcellular resistance may be affected by intracellular signals (e.g., Ca 2+, cAMP or protein kinase C) through interactions between the actin cytoskeleton and the zonula occludens [27-30] or by other processes [32]. The zonula adherens (intermediate junction) consists of a circumferential band of actin microfilaments of mixed polarity encircling the apex of the cell just basal to the zonula occludens. Myosin, a-actinin and tropomyosin have also been identified, and are thought to both cross-link and provide contractile potential for the actin microfilaments [21]. This complex is attached to the lateral membrane by an adhesion plaque containing c~-actinin, zyxin, vinculin and radixin [10]. These, in turn, link to c a t e n i n / c a d h e r i n complexes, which mediate cell-cell adhesion [10]. Desmosomes (maculae adherens) lie in a loose band below the zonula adherens and are also distributed along the lateral membrane, forming cell-cell adhesions mediated by cadherins. Hemidesmosomes are distributed along the basal membrane, and form cellsubstratum adhesions mediated by integrins. Desmosomes and hemidesmosomes display common ultrastructural features, such as trilayered cytoplasmic plaques and linkage via cytokeratin intermediate filaments, but they are analogous (rather than homologous) structures with biochemically distinct compositions (reviewed in Refs. 10, 35). Gap junctions are distributed along the lateral membrane. The basic structural unit is a transmembrane hexamer that forms an aqueous channel. Clusters of these hexamers, aligned with similar clusters on an adjacent cell, enable electrical and chemical coupling between cells. Focal adhesions are distributed along the basal membrane, and anchor cells to the extraceIlular matrix (Fig. 1C; reviewed in Refs. 10, 36). Actin micro-

filaments, cross-linked by a-actinin to form stress fibers, are linked to the membrane by a protein complex that includes talin, vinculin and actin microfilament capping proteins. These, in turn, link to integrin heterodimers, which mediate cell-substratum adhesion. Focal adhesions and stress fibers may be regulated by intracellular signals. Conversely, extracellular matrix interactions via integrins may generate intracellular signals both via interactions with the cytoskeleton and via signal transduction pathways similar to those associated with hormone receptors [37,38]. Actin is also distributed along the BLM as part of a cortical web (Fig. 1B; reviewed in Ref. 10). Short actin microfilaments and associated proteins are linked via direct or indirect means to spectrin tetramers, hexamers and octamers to form a two-dimensional network. This network is then linked via ankyrin and protein 4.1 to a variety of basolateral cell-surface proteins such as uvomorulin [39], Na+/K+-ATPase [40-43] and the C I - / H C O - 3 exchanger [44]. Of note is that actin microfilaments also associate with and may modulate the activities of the basolaterally-located N a + / K + / 2CI- cotransporter [45] and epidermal growth factor (EGF) receptor [46], and the apically-located Na + channel [47-49]. Other cytoskeletal elements include microtubules and intermediate filaments. Polarized cells have two distinct sets of microtubules [50]. One set consist of short filaments extending laterally as a dense mat underneath the terminal web, oriented randomly. The other set consists of longer filaments originating diffusely from just below the terminal web and extending longitudinally in bundles, oriented with their 'plus' ends directed basally [50,51]. Vesicles and other membrane-bound organelles can travel vectorially along this system of fibers powered by kinesin- and dynein-like mechanoproteins. Intermediate filaments form an array about the cell nucleus, extending to the periphery. They are most prominent in cells subject to mechanical stress, and are thought to play a tension-bearing role.

II-C. Establishment and maintenance of epithelial cell polarity Establishment of a polarized epithelium begins with cell-cell recognition and contact mediated by cadherins [9,34,52-55]. This allows for establishment of local domains that serve as foci for assembly of junctional and non-junctional cytoskeletal structures, leading to delineation of the apical and basolateral domains. For example, in non-confluent MDCK cells soluble spectrin/ankyrin/uvomorulin complexes can be identified [39]. Cell-cell contact induces aggregation of these complexes at the contact sites, forming a Triton-X100-insoluble, basolaterally-located cortical cytoskeletal web [56]. Other membrane proteins, such as

APICAL

RER~~'R~

go BASAL

2'

Fig. 2. Left: Synthesis, sorting, and targeting pathways for membrane components. Proteins are synthesized in the rough endoplasmic reticulum (RER), migrate sequentially through the transitional endoplasmic reticulum (TER) and Golgi apparatus, and arrive at the trans-Golgi network (TGN). In some cell types, apically- and basolaterally-targeted components then move via vesicles directly to their corresponding surface membrane domains (pathways A and C, respectively). In other cell types, apically-targeted proteins are first delivered to the basolateral domain, then endocytosed and re-delivered to the apical domain (pathway B). The nucleus (N), microvilli (MV) and extracellular matrix (ECM) are also labeled. Right: Endocytic pathways. Apical receptor-mediated endocytosis begins with the budding off of a clatherin-coated pit (la) to form a clatherin-coated vesicle. Disassembly of the clatherin coat (2a) yields a primary or early endosome. From these endosomes, membrane components may recycle back to the apical domain (3a), transfer to lysosomes (L) via multivesicular bodies (MVB) (4a), or transcytose to the basolateral domain (5a). Similar, though less extensive, endocytic pathways exist for the basolateral domain (lb-5b).

Na+/K+-ATPase, can subsequently be incorporated and constrained within this cortical web [57-59]. Further genesis and maintenance of a polarized surface membrane depends upon the sorting and targeting of newly-synthesized proteins (reviewed in Refs. 60-63). A variety of pathways appear to be involved (Fig. 2). In MDCK cells, newly-synthesized proteins generally are sorted at the level of the trans-Golgi network (TGN) and directly targeted to apical or basolateral domains (pathways A and C, respectively). For example, glycosyl phosphatidylinositol (GPI)-anchored proteins target directly to the apical domain, perhaps via coclustering with hydrogen-bonded glycosphingolipid patches that bud from the TGN. Other proteins, such as the low density lipoprotein receptor, utilize a short (approx. 14 residue) amino-acid signal sequence to target directly to the BLM. Proteins related to those involved in cell-surface endocytosis may recognize and mediate the budding of these proteins from the TGN.

In hepatocytes, on the other hand, newly-synthesized proteins generally are all initially transported to BLM, with apically-destined proteins then undergoing transcytosis via endocytic pathways (Fig. 2, pathway C). Sorting may occur at either of two sites, the BLM or the endosome. At the BLM level, there may be selective internalization or selective retention of apical and basolateral proteins, respectively, with the apical proteins undergoing transcytosis (Fig. 2, pathway lb ~ 2b -~ 5b). Alternatively, BLM could be internalized randomly (Fig. 2, pathway lb-~ 2b), and then undergo protein sorting in the endosome to either recycle to the basolateral domain (pathway 3b) or undergo transcytosis to the apical domain (pathway 5b). Establishment of surface membrane lipid polarity is less well-understood, in part because of the lack of immunologic techniques and domain-specific markers. Sphingolipids are preferentially transferred from the TGN to the apical domain, perhaps via vesicles containing coclustered GPI-anchored proteins [64]. Possible mechanisms for the sorting of other lipid species are less clear, though lipid transfer proteins may play a role [65,66]. Once established in the appropriate surface membrane domain, apical and basolateral proteins and lipids are maintained by the zonula occludens 'fence' function. In addition, lateral diffusion of proteins within the membrane domain may be constrained locally by interactions with the cortical actin cytoskeleton (described earlier). However, in most polarized epithelial cells both membrane domains, but particularly the apical domain, are in a constant state of flux due to endocytosis (Fig. 2, right panel). To maintain polarity, endocytosed membrane proteins and lipids must be recycled to their membrane of origin, with sorting occurring at the endosomal level (pathways 3a and 3b). Selective retention in the membrane with exclusion from endocytosis may also play an important role. For example, in at least one strain of MDCK cells Na+/K+-ATPase is initially delivered equally to the apical and basolateral membranes. The assembly, with cell-cell contact, of a basolateral actin cortical cytoskeleton (described earlier) allows for selective retention and stabilization of Na+/K+-ATPase in the basolateral membrane. No such retention or stabilization occurs for the apical membrane, and apical N a + / K +ATPase continues to be catabolized. The end result is a polarized distribution, with higher levels in the basolateral domain [57]. Vesicle traffic through the cytoplasm occurs mainly via vectorial translocation along microtubules and actin microfilaments mediated by vesicle-bound mechanoproteins, as noted previously. In polarized epithelial cells, transport along microtubules predominates [51,62]. However, translocation along actin microfilaments may play an important role in transport across

the terminal web for apically-targeted newly-synthesized proteins and in apical endocytosis [26,67]. Vesicle recognition and fusion with the surface membrane is also an actively controlled process involving GTP-binding proteins [61]. In addition, second messenger systems (e.g., Ca 2+) may play a role, regulated by cell-cell and cell-substratum contacts, with signal transduction occurring through cadherins and integrins [61]. III. Altered epithelial cell polarity in pathologic states The establishment and maintenance of epithelial cells polarity is a multicomponent, multistage process. As with any complex system, there is great potential for malfunction. The following sections will examine several conditions, both acquired and congenital, in which abnormal function leading to clinical disease appears in large part to be related to dysfunction of the actin cytoskeleton.

IliA. lschemia in kidney proximal tubule cells Tissue ischemia occurs when a reduction in regional blood flow decreases delivery of oxygen and substrates to levels inadequate to maintain cellular energy status. ATP depletion ensues, inducing a series of structural, biochemical and functional derangements. The kidney proximal tubule is particularly susceptible to this process, and serves as a useful model both in vivo and in vitro (reviewed in Ref. 68). Of central importance appears to be a rapidly-occurring, duration-dependent disruption and dissociation of the actin cytoskeleton and associated structures (Fig. 3A,B). In vivo, ischemia induced disruption of microvillar actin cores and the apical circumferential actin microfilament network, with redistribution of stainable actin from the apical pole to throughout the cytoplasm [6973]. In vitro, using the proximal-tubule-derived LLCPK~ cell line, antimycin-A-generated ATP depletion induced disruption of the cortical cytoskeleton and redistribution of actin into large cytoplasmic aggregates primarily located in the perinuclear region [74]. Within the first 5 rain of ATP depletion, there was a marked (thermodynamically-favored) conversion of monomeric (G) actin to polymeric (F) actin [74], presumably involving effects on G-actin sequestering proteins and F-actin capping and nucleation-inhibiting proteins [75,76]. Concurrent with this disruption of the actin microfilament network was a dissociation of the cortical cytoskeleton. In vivo, Na+/K+-ATPase, spectrin, and uvomorulin all became Triton X-100 soluble (i.e., cytoskeleton-dissociated), with the Na+/K+-ATPase being free of actin, spectrin and uvomorulin [77]. In vitro, ankyrin and spectrin distributed throughout the cytoplasm and were not associated with each other [78].

With or following this disruption and dissociation of the cortical actin cytoskeleton, the surface membrane undergoes extensive changes (outlined in Fig. 4). These

include alterations in microvillous morphology, disruption of cellular junctions and loss of surface membrane polarity. Apical microvilli are lost by coalescence in-

Lumen APICAL

- -

:

.....

.....

Apical proteins Myosin I Villin, Fimbrln ZO-1, Cingulin F-Actin mlcrofilsments Fodrin Ankyrln Cadherin Integrin NS+, K+-ATPase

°

BASAL Blood Lumen APICAL

C

BASAL Blood Fig. 3. (A,B), Effect of ischemia on the actin cytoskeleton. Symbols correspond to those of Fig. 1. With ischemic injury, the actin cytoskeleton dissociates. Microvilli fuse and fragments are shed into the lumen. The zonula occludens, zonula adherens, terminal web, cortical web and focal adhesions fragment. F-actin aggregates distribute throughout the cytoplasm and accumulate perinuclearly. Other actin cytoskeletal elements (e.g., spectrin, ankyrin) distribute throughout the cytoplasm. Apical and basolateral integral membrane proteins (e.g., apical leucine aminopeptidase; basolateral Na+/K+-ATPase and integrins) distribute randomly over the surface membrane of the cell. (C,D), Effect of ischemia on Na +/K+-ATPase polarity and sodium transport. In the normally-functioning cell, sodium enters across the apical membrane via apical channels and cotransporters, travelling down its electrochemical gradient and providing the driving force for H ÷ secretion and glucose, amino acid and phosphate reabsorption. Sodium is then transported across the BLM via basolateral Na +/K+-ATPase, moving up its electrochemical gradient in an energy-requiring process. With ischemic injury, a portion of BLM Na +/K+-ATPase redistributes to the apical domain. Apically redistributed Na +/K+-ATPase maintains its functional activity and is capable of competing for intracellular sodium. The resultant Na +/K+-ATPase-mediated apical sodium secretion sets up a futile cycle that may uncouple vectorial sodium transport and ATP utilization. Apical sodium secretion also markedly diminishes the efficiency of sodium reabsorption and induces a high urinary sodium excretion. Intercellular movement (dotted line) may also contribute to abnormal fluxes.

ISCHEMIA

]

ATP D E P L E T I O N [

LOSS OF MICROVILLOUS INTEGRITY

D I S R U P T I O N OF J U N C T I O N A L COMPLEXES AND FOCAL ADHESIONS

[ I

\

APICAL

MEMBRANE

FUSION, B L E B B I N G , AND INTERNALIZATION

U N T E T H E R I N G OF A C T I N A S S O C I A T E D M E M B R A N E PROTEINS (Na÷,K÷-ATPase, INTEGRINS)

M O V E M E N T OF LIPIDS A N D P R O T E I N S TO ALTERNATE DOMAIN

LOSS OF N O R M A L C E L L CELL AND CELLSUBSTRATUM ATTACHMENT

/ REDUCED PLASMA MEMBRANE SURFACE AREA

LOSS OF S U R F A C E M E M B R A N E LIPID A N D PROTEIN POLARITY

DISRUPTION EPITHELIAL CONTINUITY

OF CELL

CELLULAR DYSFUNCTION ]

Fig. 4. Proposed role of ischemia-induced actin cytoskeletalalterations in the pathophysiologyof kidney proximal tubule cell injury. volving membrane fusion and by fragmentation involving shedding into the lumen and internalization into the cytoplasm [79-82]. The zonula occludens becomes disrupted. Intercellular permeability progressively increases (loss of 'gate' function) [32,70,83], and membrane lipids and proteins are able to redistribute across domains (loss of 'fence' function). Apical domain SPH and cholesterol contents decrease while PC, PI and total phospholipid contents increase, with apical membrane fluidity markedly increasing secondary to the reduction in S P H / P C and C / P L ratios [4,71,72,84]. Na+/K+-ATPase, untethered from the cortical cytoskeleton, redistributes into the apical domain, while the apical marker protein leucine aminopeptidase redistributes into the basolateral domain [71,72,77,84]. Similarly, investigations using BS-C-1 cell line and H 2 0 2 oxidative stress have revealed disruption of focal adhesions with loss of talin from the basal cell surface [85]. The disruption untethers the associated integrins, allowing them to redistribute into the apical domain

[85].

The functional ramifications of these changes in cytoskeletal structure and surface membrane polarity are substantial. Loss of microvilli results in a markedly decreased apical membrane surface area. Increases in intercellular permeability allow for increased 'backleak' [15], though this is of uncertain significance in the kidney proximal tubule. Redistribution of membrane lipids changes membrane physicochemical properties, which in turn may affect integral membrane protein function. For example, reduced glucose reabsorption after ischemic injury appears related in part to decreases in the apical S P H / P C ratio. In brush-border membrane vesicles (BBMV) prepared after 15 rain of ischemia, the V,..... for Na+-dependent glucose transport and the number of phlorizin-binding sites decreased dramatically compared to controls [86]. Though this could represent a redistribution of Na+-dependent glucose 'carriers' to the alternate domain, carrier-mediated Na+-dependent alanine transport was not similarly affected [86]. Moreover, glucose transport in BBMV was highly correlated

with the S P H / P C ratio and inversely correlated with membrane fluidity [86], and with recovery from ischemia glucose transport increased to control levels concurrently with the normalization of the apical S P H / P C ratio [84]. Finally, redistribution of membrane proteins to the alternate domain may result in substantial changes in domain-specific functions. Perhaps the best-studied example in this category involves Na+/K+-ATPase. Under physiologic circumstances, sodium enters the cell via a variety of apical transport proteins, providing the energy for uptake of a variety of solutes while moving down its electrochemical gradient. Sodium is then transported out of the cell up its electrochemical gradient via basolateral Na+/K+-ATPase (Fig. 3C). Under these circumstances vectorial transport of sodium and other solutes is coupled to ATP utilization. With ischemia-induced redistribution of Na+/K+-ATPase to the apical membrane, sodium that has entered the cell may be transported out of the cell across either the apical or the basolateral membrane (Fig. 3D). Apical transport via Na+/K+-ATPase results in a futile cycle, with transport of sodium now uncoupled from ATP utilization. In vivo evaluation of sodium reabsorption by micropuncture [87] and by lithium and sodium reabsorption [84] confirmed decreased proximal tubule sodium absorption after ischemia, correlated with apical redistribution of Na+/K+-ATPase, but independent of apical sodium permeability, cellular ATP content and cellular morphology. Though as yet undocumented, basolateral redistribution of apical transport proteins could similarly affect cellular transport. The apical redistribution of integrins may also have important functional ramifications, as adherence of unattached cells to cell monolayers is increased in cells previously exposed to oxidant stress [15]. Such cell-cell interactions within a renal tubule lumen may lead to cell clumping and lumenal obstruction [85]. Though as yet undocumented, redistributed cadherins could function similarly. Partial confirmation of a central role for actin cytoskeletal changes in surface membrane alterations comes from studies using cytochalasin D, which disrupts actin microfilaments by binding to their 'plus' end. Perfusion of kidneys with cytochalasin D, under conditions in which disruption was selective for the actin microfilaments, resulted in loss of microvilli with membrane fusion and fragmentation, and well-correlated time-dependent decreases in sodium and lithium reabsorption [69,73]. Morphologically and physiologically, these changes were similar to changes seen during ischemia. Cytochalasin D has also been shown to disrupt the zonula occludens 'gate' function in small intestine enterocytes [88]. An intriguing and potentially important problem is the increased sensitivity to ischemia of proximal corn-

pared to distal tubule cells with respect to a variety of morphologic, biochemical, and functional derangements [68]. For example, in vivo renal ischemia for up to 50 min induces extensive apical actin microfilament disruption in proximal tubule cells, but essentially no alterations in distal cells [73]. Similarly, ischemia for 15 min induces extensive disruption of tight junctions and redistribution of Na+/K+-ATPase to the apical domain in proximal cells, but no such changes in distal cells (Refs. 70, 77 and Molitoris, B.A. and Dahl, R., unpublished data). Finally, ischemia-induced impairment of sodium reabsorption is also largely due to proximal tubule dysfunction, with distal tubule function little-affected, as evidenced by good correlations between the fractional excretions of sodium and lithium [84]. The observation that cytochalasin D infusion also affects proximal but not distal tubular sodium reabsorption [69] raises the possibility that intrinsic differences in actin cytoskeletal susceptibilities may play an important role in this differential sensitivity. Reestablishment of apical and basolateral membrane polarity occurs during the recovery phase of ischemic acute renal failure and has been demonstrated for leucine aminopeptidase, Na+/K+-ATPase and apical and BLM lipids [71,84,86,87]. The mechanisms responsible for restitution of surface membrane polarity have not been determined, but involve actual remodeling of surface membrane domains Ln previously damaged cells rather than cellular proliferation [79,80]. The rate at which repolarization occurs is dependent on the severity of the injury. 15 min of ischemia (mild injury) required only 24-48 h of reperfusion, while 50 min of ischemia (moderate to severe injury) required several days for reestablishment of apical and basolateral membrane polarity [84]. In both cases, reestablishment of surface membrane polarity occured hours to days after cellular morphology had normalized. The study of factors involved in and responsible for cellular recovery from ischemia is now being pursued with vigor, but remains in its infancy (reviewed in Refs. 89, 90). Thyroid hormone [91], EGF [92] and transforming growth factor-a [93] have all been shown to accelerate recovery, c-fos, Egr-1 and TRPM-2 gene activation has been observed [94,95], but interpretation of these data is problematic [96]. Heat shock proteins have been shown to accumulate [97-103]. However, the underlying cellular mechanisms remain to be determined. For instance, such basic questions as whether the disrupted and dissociated actin cytoskeletal components are reutilized or if synthesis of new proteins is required remain unanswered. In summary, ischemia in kidney proximal tubule cells induces a rapidly-occurring, duration-dependent disruption of the actin cytoskeleton. This, in turn, leads to disruption of associated surface membrane structures and untethering of integral membrane proteins.

10 Lipids and proteins move to alternate domains, and cell-cell and ceIl-substrate contacts are lost. The cell is no longer able to perform vectorial transport, and cellular and organ-level dysfunction ensues. This proposed scheme is summarized in Fig. 4.

III-B. Microl;illous inclusion disease Microvillous inclusion disease (MID, Davidson's disease) is an autosomal recessive disorder that presents clinically as protracted diarrhea from birth and is generally fatal [104,105]. In all patients there is decreased absorption of water, electrolytes and nutrients, and in some there is net secretion of water and sodium. Morphologically, in the small intestine there is loss of villous height without crypt hyperplasia. Enterocytes along these short villi are strikingly abnormal, with scanty, disorganized, short microvilli and apical cytoplasmic vesicles possessing a microvillous membrane surrounded by a terminal web (microvillous inclusions). In contrast, nearby non-enterocytic ceils (Goblet cells, Paneth's cells and enteroendocrine cells) are ultrastructurally normal. Deeper in the intestinal crypts, the enterocytes have a more normal microvillous structure, but there is a marked increase in apical electron-dense vesicles thought to be secretory vesicles containing glycocalyx and microvillous material [106]. Colonocytes, of the large intestine, are similarly affected, containing microvillous inclusions and increased electron-dense vesicles. MID is probably a heterogeneous disorder. The specific underlying defects are unknown, but likely involve abnormalities in delivery of Golgi-derived vesicles to the apical membrane [9,106]. As noted previously, both microtubule- and actin microfilament-associated processes appear to be involved in delivery of apically-destined vesicles, and models involving disruption of each have been investigated. In vivo treatment of intestinal cells with microtubule-depolymerizing drugs resulted in abnormal delivery of apical proteins and glycocalyx to and formation of microvilli along the basolateral membrane [51,107]. With clearance of these drugs, basolateral membrane was endocytosed to form microvillus inclusions throughout the cytoplasm; these subsequently fused with the apical membrane to restore normal cell morphology [51]. Although basolateral microvilli are not typical of MID, they have been described in one patient with intestinal microvillous inclusions [106]. SDS-PAGE of brush-border proteins from this patient revealed absence of a 130-kDa band, thought possibly to be vinculin [106]. On the other hand, treatment of fetal intestinal epithelium cultures with microfilament-disrupting drugs was reported (in preliminary form) to result in collapse of the apical microvillous domain and formation of

apical microvillous inclusions [106]. Basolateral microvilli were seen only when microtubule-depolymerizing drugs were used instead. In a different model, MDCK epithelial cells with incomplete cell-cell contacts (grown in low-calcium medium at a low density) displayed decreased numbers of microvilli and decreased levels of an apical marker protein, but were still highly polarized; in addition, they possessed inclusions, termed vacuolar apical compartments, that possessed microvilli and apical (but not basolateral) surface membrane proteins [1(/8]. Return to normal calcium medium resulted in fusion of the vacuolar apical compartments with the apical surface membrane and normalization of morphology [109]. Both these models seem reasonably similar to typical MID. In addition, SDS-PAGE of brush-border proteins from two infants with MID revealed a striking diminution of a 200-kDa band that ran parallel to myosin and was bound by anti-myosin sera [106,110]. Finally, it should be noted that abnormal delivery a n d / o r regulation of N a ~ / K + - A T P a s e or other surface membrane ion transport proteins, perhaps associated with abnormalities in the actin cytoskeleton, may also be involved in the pathophysiology of MID. Measurement of electrolyte transport across excised jejunum from an affected patient revealed substantially reduced unidirectional absorptive and secretory fluxes of sodium and chloride, with net secretion of both, the latter finding not explicable simply on the basis of reduced microvillous surface area [111]. Intestinal cells treated with microtubule-depolymerizing agents had a normal basolateral distribution of Na+/K+-ATPase [51], but direct data on transport protein surface membrane distribution in MID or with microfilament-disrupting agents are not yet available.

III-C. Polyqvstic kidney disease Polycystic kidney disease (PKD) generally refers to a group of inherited disorders in which there are multiple cysts in both kidneys (reviewed in Refs. 112, 113). The cysts are functioning portions of individual nephrons; over time they close off from their nephron of origin and progressively enlarge, mechanically obstructing adjacent nephrons and eventually inducing renal failure. There are two clinically distinguishable forms of inherited PKD. Autosomal dominant PKD (ADPKD) is fairly common (1 in 1000 live births) and has a relatively indolent course. Most cases are due to a gene linked to a highly polymorphic marker on the short arm of chromosome 16 [114], but approx. 10% of A D P K D cases appear due to a different gene. Autosoreal recessive PKD (ARPKD) is rarer (1 in 100000 live births) but generally has a much more severe course, with many patients dying in the first year of life.

11 Morphologically, the renal cysts appear as fluid-filled sacs lined by a continuous single layer of epithelium. On the basis of a variety of studies (reviewed in Ref. 112), it is believed that a combination of epithelial cell proliferation and altered fluid transport are responsible for the pathophysiology of cyst formation. Specifically, maintenance of a continuous epithelial cell lining within an expanding cyst, in the absence of marked changes in cell size, requires cell proliferation. Moreover, fluid accumulation within a cyst closed off from its nephron of origin requires secretion of fluid across the epithelial cell layer. Underlying each of these processes may be abnormalities in epithelial polarity. Cell proliferation is controlled by a complex variety of intracellular and extracellular factors. Recent data have implicated one such factor, EGF, in epithelial proliferation in PKD. To begin, EGF has been found in elevated levels in cyst fluid from human ADPKD patients and from the C57BL/6J cpk/cpk mouse, a model of ARPKD [112,115]. In turn, EGF applied to renal tubule cultures induced more marked mitogenesis in ADPKD tissue than in normal tissue [112,116]. Finally, as with many receptors, the EGF receptor is associated with the actin cytoskeleton (see Ref. 46 and references therein) and is normally located basolaterally. However, in both ADPKD and ARPKD tissues mislocation of the EGF receptor to the apical domain has been found [112,117]. Thus, cyst fluid EGF, acting in an autocrine manner via abnormally localized EGF receptors, may play a significant role in the abnormal epithelial proliferation seen in PKD. As noted earlier, normally in proximal tubule cells vectorial apical to basal sodium transport occurs via coordinated function of apical transport proteins and basolateral Na+/K+-ATPase (Fig. 3C). Water follows sodium, yielding net fluid absorption. However, in ADPKD functionally competent Na+/K+-ATPase is mislocated to the apical membrane [112,118,119]. This is associated with ouabain-inhibitable net basal to apical (reverse) transport of sodium (Fig. 3D) [119], and hence fluid secretion. The mislocation is specific. The cortical cytoskeletal Na+/K+-ATPase-associated proteins spectrin and ankyrin are also mislocated apically. However, other proteins such as the apical N a + / H + antiporter and the basolateral C 1 - / H C O - 3 exchanger remain normally distributed [112,119]. Of note is that in ARPKD and in a new mouse model of ADPKD, apical mislocation of Na+/K+-ATPase has not been found [117,120], and other mechanisms for net fluid secretion are presumably operative. Finally, a gene from the ADPKD region of chromosome 16 that has recently been cloned contains ankyrin repeats and a region homologous to several tyrosine protein kinases [121]. As noted by the authors, ankyrin repeats are common to proteins involved in cell-cycle control and differentiation, and tyrosine kinase recep-

tors are involved in cell proliferation, raising the possible involvement of the gene product in abnormal cell proliferation. Alteratively, the gene product may be a cytoskeletal protein, perhaps involved in the misapplication of Na+/K+-ATPase to the apical membrane.

IV. Summary The establishment and maintenance of cell polarity is essential for normal epithelial function. Disruption of the underlying processes, either as a primary inborn defect or as a secondary result of other pathologic processes, can lead to loss of epithelial polarity and further cellular and organ-level dysfunction. Continued elucidation of the processes involved may prove fruitful both in the understanding of basic cell biology and in the understanding and treatment of a variety of disease states.

Acknowledgements We thank Mary. Lee for her secretarial assistance. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK41126 (to B.A.M.) and grants from the Veterans Affairs Research Service and the American Heart Association. B.A.M. is an AHA Established Investigator.

References 1 Simons, K. and Fuller, S.D. (1985) Annu. Rev. Cell Biol. 1, 243-288. 2 Rodriguez-Boulan, E. and Nelson, W.J. (1989) Science 245, 718-725. 3 Carmel, G.F., Rodriquez, F., Carriere, S. and LeGrimellec, C. (1985) Biochim. Biophys. Acta 818, 149-157. 4 Molitoris, B.A. and Hoilien, C. (1987) J. Membr. Biol. 99, 165-172. 5 Molitoris, B.A. and Simon, F.R. (1985) J. Membr. Biol. 83, 207 -215. 6 Chong, P.L.G., Fortes, P.A.G. and Jameson, D.M. (1982) J. Biol. Chem. 29, 43-49. 7 Molitoris, B.A. (1987) Semin. Nephrol. 7, 61-67. 8 Sutherland, E., Dixon, B.S., Leffert, H.L., Skally, H., Zaccaro, L. and Simon, F.R. (1988) Proc. Natl. Acad. Sci. USA 85, 8673-8677. 9 Molitoris, B.A. and Nelson, W.J. (1990) J. Clin. Invest. 85, 3-9. 10 Luna, E.J. and Hin, A.L. (1992) Science 258, 955-964. 11 Takeichi, M. (1991)Science 251, 1451-1455. 12 Ozawa, M., Ringwald, M. and Kemler, R. (1990) Proc. Natl. Acad. Sci. USA 87, 4246-4250. 13 Cheresh, D.A. (1992) Lab. Immunol. II 12, 217-236. 14 Lampugnani, M.G., Resnati, M., Dejana, E. and Marchisio, P.C. (1991) J. Cell Biol. 112, 479-490. 15 Goligorsky, M.S., Lieberthal, W., Racusen, L. and Simon, E.E. (1993) Am. Physiol. Soc. 264, F1-F8. 16 Reszka, A.A., Hayashi, Y. and Horwitz, A.F. (1992) J. Cell Biol. 117, 1321-1330. 17 Hall, A. (1992) Mol. Biol. Cell 3, 475-479. 18 Lokeshwar, V.B. and Bourguignon, L.Y.W. (1992) J. Biol. Chem. 267, 22073-22078.

12 19 Heintzelman, M.B. and Mooseker, M.S. (1992) Curr. Opin. Cell Biol. 26, 93-122. 20 Coudrier, E., Kerjaschki, D. and Louvard, D. (1988) Kidney Int. 34, 309-320. 21 Drenckhahn, D. and Derietzel, R. (1988) J. Cell Biol. 107, 1037-1048. 22 Mooseker, M.S. (1985) Annu. Rev. Cell Biol. 1, 2(19-241. 23 Rodman, J.S., Mooseker, M. and Farquhar, M.G. (1986) Eur. J. Cell Biol. 42, 319-327. 24 Janmey, P.A. and Matsudaira, P.T. (19881 J. Biol. Chem. 263, 16738-16743. 25 Conzelman, K.A. and Mooseker, M. (1987) J. Cell Biol. 105, 313-324. 26 Fath, K.R. and Burgess, D.R. (1993) J. Cell Biol. 120, 117-127. 27 Gumbiner, B. (1987) Am. J. Physiol. 253, C749-C758. 28 Madara, J.L. (1989) J. Clin. Invest. 83, 1089-1094. 29 Madara, J.L., Barenberg, D. and Carlson, S. (1986) J. Cell Biol. 102, 2125-2136. 30 Madara, J.L. and Dharmsathaphorn, K. (1985) J. Cell Biol. 101, 2124-2133. 31 Van Meer, G. and Simons, K. (1986) EMBO J. 5, 1455-1464. 32 Mandel. L.J., Bacallao, R. and Zampighi, G. (1993) Nature 361, 552-555. 33 Citi, S.H., Sabanay, H., Jakes, R., Geiger, B. and Kendrick-Jones, J. (19881 Nature 333, 272-276. 34 Gruenberg, J. and Howell, K.E. (19891 Annu. Rev. Cell Biol. 5, 453-481. 35 Garrod, D.R. (1993) Curr. Opin. Cell Biol. 5, 30-40. 36 Burridge, K., Fath, K., Kelly, T., Nuckolls, G. and Turner, C. (1988) Annu. Rev. Cell Biol. 4, 487-525. 37 Juliano, R.L. and Haskill, S. (1993) J. Cell Biol. 120, 577-585. 38 Ben-Ze'ev, A. (1991) BioEssays 13, 207-212. 39 Nelson, W.J., Shore, E.M., Wang, A.Z. and Hammerton, R.W. (1990) J. Cell Biol. 110, 349-357. 40 Koob, R., Kraemer, D., Trippe, G., Aebi, U. and Drenckhahn, D. (1990) Eur. J. Cell Biol. 53, 93-100. 41 Morrow, J.S., Cianci, C.D., Ardito, T., Mann, A.S. and Kashigarian, M. (19891 J. Cell Biol. 108, 455-465. 42 Nelson, W.J. and Veshnock, P.J. (1987) Nature 328, 533-536. 43 Bertorello, A.M. and Cantiello, H.F. (1992)J. Am. Soc. Nephrol. 3, 8(/3. (Abstract) 44 Drenckhahn, D., Schluter, K., Allen, D.P. and Bennett, V. (1985) Science 230, 1287-1289. 45 Matthews, J.B., Awtrey, C.S. and Madara, J.L. (1192) J. Clin. Invest. 90, 1608-1613. 46 Den Hartigh, J.C., Van Bergen en Henegouwen, P.M.P., Verkleij, A.J. and Boonstra, J. (1992) J. Cell Biol. 119, 349-355. 47 Cantiello, H.F., Stow, J.L., Prat, A.G. and Ausiello, D.A. (1991) Am. Physiol. Soc. 261, C882-C888. 48 Smith, P.R., Saccomani, G., Joe, E.H., Angelides, K.J. and Benos, D.J. (1991) Proc. Natl. Acad. Sci. USA 88, 6971-6975. 49 Prat, A.G., Ausiello, D.A. and Cantiello, H.F. (1992) J. Am. Soc. Nephrol. 3, 817. (Abstract) 50 Bacallao, R., Antony, C., Dotti, C., Karsenti, E., Stelzer, E.H.K. and Simons, K. (1989) J. Cell Biol. 109, 2817-2832. 51 Achier, C., Filmer, D., Merte, C. and Drankahahn, D. (1989) J. Cell Biol. 109, 179-189. 52 Simmons, C. (1992) Semin. Perinatol. 16, 78-89. 53 Nelson, W.J., Hammerton, R.W., Wang, A.Z. and Shore, E.M. (1990) Cell Biol. 1,359-371. 54 Wiley, L.M., Kidder, G.M. and Watson, A.J. (1990) BioEssays 12, 67-73. 55 McNeil, H., Ozawa, M., Kemler, R. and Nelson, W.J. (1990) Cell 62, 309-316. 56 Nelson, W.J. and Veshnock, P.J. (1987) J. Cell Biol. 104, 15271537. .

57 Hammerton, R.W., Krezeminski, K.A., Mays, R.W., Ryan, T.A., Wollner, D.A. and Nelson, W.J. (1991) Science 254, 847-849. 58 Nelson, W.J. and Hammerton, R.W. (1989) J. Cell Biol. 108, 893-902. 59 Salas, P.J.I., Vega-Salas, E., Hochman, J., Rodriguez-Boulan, E. and Edidin, M. (1988) J. Cell Biol. 107, 2363-2376. 60 Mostov, K., Apodaca, G., Aroeti, B. and Okamoto, C. (1992) J. Cell Biol. 116, 577-583. 61 Nelson, W.J. (1992) Science 258, 948-955. 62 Nelson, W.J. (1991) Cell Biol. 2, 375-385. 63 Wollner, D.A. and Nelson, W.J. (1992) J. Cell Sci. 102, 185-190. 64 VanMeer, G. and Simons, K. (1988) J. Cell Biochem. 36, 51-58. 65 Koval, M. and Pagano, R.E. (1989) J. Cell Biol. 108, 2169-2181. 66 Nichols, J.W. (1987) J. Biol. Chem. 262, 14172-14177. 67 Bornens, M. (1991) New Biol. 3, 627-636. 68 Weinberg, J.M. (1991) Kidney Int. 39, 476-500. 69 Kellerman, P.S., Clark, R.A.F., Hoilien, C.A., Linas, S.L. and Molitoris, B.A. (1990) Am. J. Physiol. 259, F279-F285. 70 Molitoris, B.A., Falk, S.A. and Dahl, R.H. (1989) J. Clin. Invest. 84, 1334-1339. 71 Molitoris, B.A., Hoilien, C.A., Dahl, R.H., Ahnen, D.J., Wilson, P.D. and Kim, J. (19881 J. Membr. Biol. 1(/6, 233-242. 72 Molitoris, B.A., Wilson, P.D., Schrier, R.W. and Simon. F.R. (19851 J. Clin. Invest. 76, 2097-2105. 73 Kellerman, P.S. and Bogusky, R.T. (1992) Kidney Int. 42, 896902. 74 Molitoris, B.A., Geerdes, A.E. and McIntosh, J.R. (19911 J. Clin. Invest. 88, 462-469. 75 Sander, M.C. and Wang, Y.L. (1990)J. Cell Biol. i10, 359-365. 76 Southwick, F.S. and Young, C.L (19901 J. Cell Biol. 110, 19651973. 77 Molitoris, B.A., Dahl, R. and Geerdes, A.E. (1992) Am. J. Physiol. 263, F488-F495. 78 Molitoris, B.A. and Geerdes, A.E. (19911 J. Cell Biol. 115, 350. (Abstract) 79 Glaumann, B., Glaumann, H., Berezesky, I.K. and Trump, B.F. (1977) Virchows Arch. B Cell Pathol. 24, 1-18. 80 Glaumann, B., Glaumann, H. and Trump, B.F. (1977) Virchows Arch. B Cell Pathol. 25, 281-308. 81 Venkatachalam, M.A., Bernard, D.B., Donohoe, J.F. and Levinsky, N.G. (1978) Kidney Int. 14, 31-49. 82 Venkatachalam, M.A., Jones, D.B., Rennke, H.G., Sandstrom, D. and Patel, Y. (1981) Lab. Invest. 45, 355-365. 83 Canfield, P.E., Geerdes, A.E. and Molitoris, B.A. (1991) Am. J. Physiol. 261, F1038-FI045. 84 Spiegel, D.M., Wilson, P.D. and Molitoris, B.A. (1989) Am. J. Physiol. 256, F430 F436. 85 Gailit, J., Colflesh, D., Rabiner, I., Simone, J. and Goligorsky, M.S. (1993) Am. J. Physiol. 264, FI49-F157. 86 Molitoris, B.A. and Kinne, R. (1987) J. Clin. Invest. 80, 647-654. 87 Molitoris, B.A., Chan, L.K., Shapiro, J.I., Conger, J.D. and Falk, S.A. (1989) J. Membr. Biol. 107, 119-127. 88 Madara, J. (1987) Am. J. Physiol. 253, C171-Cl75. 89 Bacallao, R. and Fine, L.G. (1989) Am. J. Physiol. 257, Fg13F924. 90 Toback, F.G. (19921 Kidney Int. 41, 226-246. 91 Sutter, P.M., Thulin, G., Stromski, M., Ardito, T., Gaudio, K.M., Kasgarian, M. and Siegel, N.J. (1988) Pediatr. Nephrol. 20, 1-7. 92 Humes, H.D., Cieslinski, D.A., Coimbra, T.M., Messana, J.M. and Galvao, C. (1989) J. Clin. Invest. 84, 1757-1761. 93 Reiss, R., Funke, A.J., Cielinski, D.A. and Humes, H.D. (1990) Kidney Int. 37, 492. (Abstract) 94 Rosenberg, M.E. and Paller, M.S. (19911 Kidney Int. 39, 1156 1161.

13 95 Ouellette, A.J., Malt, R.A., Sukhatme, V.P. and Bonventre, J.V. (1990) J. Clin. Invest. 85, 766-771. 96 Molitoris, B.A. (1992) in Acute Renal Failure (Lazarus, J.M. and Brenner, BM., eds.), pp. 1-32, Churchill Livingstone, Edinburgh. 97 Lindquist, S. (1986) Annu. Rev. Biochem. 55, 1151-1191. 98 Enami, A., Schwartz, J.H. and Borkan, S.C. (1991) Am. Physiol. Soc. 260, F479-F485. 99 Subjeck, J.R. and Shyy, T.T. (1986) Am. J. Physiol. 250, C1-C17. 100 Currie, R.W. (1987)J. Mol. Cell. Cardiol. 19, 795-808. 101 Van Why, S., Ardito, T., Thulin, G., Siegel, N. and Kashgarian, M. (19911 J. Am. Soc. Nephrol. 2, 656. (Abstract) 102 Cairo, G., Bardella, L., Schiaffonati, L. and Bernelli-Zazzera, A. (1985) Hepatology 5, 357 361. 103 Van Why, S., Hildebrandt, F., Igarashi, P., Kashgarian, M. and Siegel, N. (19911 Pediatr. Res. 29, 353A. (Abstract) 104 Davidson, G.P., Cutz, E., Hamilton, J.R. and Gall, D.G. (1978) Gastroenterology 75, 783-790. 105 Cutz, E., Rhoads, J.M., Drumm, B., Sherman, P.M., Durie, P.R. and Forstner, G.G. (1989) N. Engl. J. Med. 320, 646-651. 106 Carruthers, L., Dourmashkin, R. and Phillips, A. (1986) Clin. Gastroenterol. 15, 105-120. 107 Pavelka, M., Ellinger, A. and Gangl, A. (1983) J. Ultrastruc. Res. 85, 249-259. 108 Vega-Salas, D.E., Salas, P.J.I. and Rodriguez-Boulan. E. (1987) J. Cell Biol. 104, 1249-1259. 109 Vega-Salas, E.E., Salas, P.J.I. and Rodriguez-Boulan, E. (1988) J. Cell Biol. 107, 1717-1728. 110 Carruthers, L, Phillips, A.D., Dourmashkin, R. and WalkerSmith, J.A. (1985) J. Pediatr. Gastroenterol. Nutr. 4, 904-907.

111 Rhoads, J.M., Vogler, R.C., Lacey, S.R., Reddick, R.L., Keku, E.O., Azizkhan, R.G. and Berschneider, H.M. (1991) Gastroenterology 100, 811-817. 112 Wilson, P.D. and Sherwood, A.C. (1991) Kidney Int. 39, 450-463. 113 Resnick, J. and Vernier, R.L. (1987) in Pediatric Nephro[ogy (Holliday, M.A., Barratt, T.M. and Vernier, R.L., eds.), pp. 371-383, Willams and Wilkins, Baltimore. 114 Reeders, S.T., Breuning, M.H., Davies, K.E., Nicholls, R.D., Jarman, A.P., Higgs, D.R., Pearson. P.L. and Weatherall, D.J. (1985) Nature 317, 542-544. 115 Moskowitz, D.W., Bonar, S.L., Marcus, M.D., Clayman, R.V. and Avner, E.D. (1989) Clin. Res. 37, 497A. (Abstract) 116 Das, K., Gabow, P.A. and Wilson, P.D. (19891 Kidney Int. 35, 310. (Abstract) 117 Avner, E.D. and Sweeney, W.E. (1992)J. Am. Soc. Nephrol. 3, 292. (Abstract) 118 Carone, F., Nakumara, S., Caputo, M., Jin, H., Nelson, J. and Kanwar, K. (1992) J. Am. Soc. Nephrol. 3, 293. (Abstract) 119 Wilson, P.D, Sherwood, A.C., Palla, K., Du, J., Watson, R. and Norman, J.T. (1991) Am. Physiol. Soc. 26(I, F420-F430. (Abstract) 120 Schafer, K., Kriz, W., Gretz, N., Luft, F.C. and Bacbmann, S. (19921 Clin. Nephrol. 301. (Abstract) 121 Gerimno, G.G., Somlo, S., Weinstat-Saslow, D., Schneider, M., Barton, N., Waterbury, G., Stevenson, S., Pohlschmidt, M., Snarey, A., Kraus, B., Frischauf, A.M. and Reeders, S.T. (1992) J. Am. Soc. Nephrol. 3, 295. (Abstract)