PHYSIOLOGICAL REVIEWS   Vol. 78 No. 4 October 1998, pp. 1109-1129
Copyright ©1998 by the American Physiological Society

Microtubule-Dependent Vesicle Transport: Modulation of Channel and Transporter Activity in Liver and Kidney

SARAH F. HAMM-ALVAREZ AND MICHAEL P. SHEETZ

Department of Pharmaceutical Sciences, University of Southern California, Los Angeles, California; and Department of Cell Biology, Duke University, Durham, North Carolina

I. INTRODUCTION
II. BACKGROUND
    A. Microtubule Organization
    B. Cytoplasmic Microtubule-Based Motors
    C. Microtubule-Based Vesicle Transport Facilitates Endocytosis and Transcytosis
III. MICROTUBULE-DEPENDENT VESICLE TRANSPORT AND NORMAL LIVER FUNCTION
    A. Transport and Secretion of Bile Salts
    B. Electrolyte/Fluid Secretion at the Canaliculus
    C. Transcytosis in Hepatocytes
    D. Cytoplasmic Dynein and Receptor-Mediated Endocytosis in Hepatocytes
    E. Autophagic, Peroxisomal, and Lysosomal Trafficking Pathways in Hepatocytes
    F. Future Challenges
IV. MICROTUBULE-DEPENDENT VESICLE TRANSPORT AND LIVER DISEASE: ETHANOL MAY IMPAIR MICROTUBULE-BASED VESICLE TRANSPORT
V. MICROTUBULE-DEPENDENT VESICLE TRANSPORT IN NORMAL KIDNEY FUNCTION
    A. Maintenance of Apical Water and Ion Channels in Renal Epithelia
    B. Apical Endocytosis
    C. Transepithelial Transport of Organic Anions
    D. Basolateral Localization of Transporters
    E. Future Questions
VI. MICROTUBULE-DEPENDENT VESICLE TRANSPORT AND KIDNEY DISEASE
    A. Polycystic Kidney Disease
    B. Ischemia and Reperfusion
VII. CONCLUSION
REFERENCES

	ABSTRACT

Top References

Hamm-Alvarez, Sarah F., and Michael P. Sheetz. Microtubule-Dependent Vesicle Transport: Modulation of Channel and Transporter Activity in Liver and Kidney. Physiol. Rev. 78: 1109-1129, 1998.  Microtubule-based vesicle transport driven by kinesin and cytoplasmic dynein motor proteins facilitates several membrane-trafficking steps including elements of endocytosis and exocytosis in many different cell types. Most early studies on the role of microtubule-dependent vesicle transport in membrane trafficking focused either on neurons or on simple cell lines. More recently, other work has considered the role of microtubule-based vesicle transport in other physiological systems, including kidney and liver. Investigation of the role of microtubule-based vesicle transport in membrane trafficking in cells of the kidney and liver suggests a major role for microtubule-based vesicle transport in the rapid and directed movement of ion channels and transporters to and from the apical plasma membranes, events essential for kidney and liver function and homeostasis. This review discusses the evidence supporting a role for microtubule-based vesicle transport and the motor proteins, kinesin and cytoplasmic dynein, in different aspects of membrane trafficking in cells of the kidney and liver, with emphasis on those functions such as maintenance of ion channel and transporter composition in apical membranes that are specialized functions of these organs. Evidence that defects in microtubule-based transport contribute to diseases of the kidney and liver is also discussed.

	I. INTRODUCTION

Top Next References

Rapid, directed movement of ion channels and transporters to and from the plasma membrane of cells within kidney and liver is critical for organ function and homeostasis. This channel and transporter modulation is now being seen as a product of microtubule (MT)-based vesicle transport. The known function of MT polymers as networks supporting the bidirectional movement of membrane vesicles driven by the MT-based motor proteins, kinesin and cytoplasmic dynein, is consistent with this proposed role in the liver and kidney. Different cargoes can be transported via MT-dependent vesicle transport including various types of endocytic and exocytic vesicles (reviewed in Refs. 31, 56). Many physiological processes utilize MT-based vesicle transport; examples include fast axonal transport (reviewed in Refs. 19, 75, 77), acinar secretion in lacrimal gland (25, 35, 129), lactation in mammary gland (120, 128), release of atrial natriuretic factor in myocytes (81, 96), and release of lytic granules from cytotoxic T cells (24, 88, 92, 107). Within this context, it is not surprising to find MT-dependent vesicle transport implicated in the essential physiology of the liver and kidney.

This review discusses the physiological role of MT-based vesicle transport in the liver and kidney. Our focus is on how MT-based vesicle transport facilitates specialized functions such as biliary secretion and glomerular filtration and reabsorption in these organs. A brief review of MT organization is followed by a short discussion of the properties and regulation of kinesin and cytoplasmic dynein. A summary of the individual steps in membrane trafficking facilitated by MT-dependent vesicle transport is also included. The focus then shifts to a detailed discussion of the role of MT-dependent vesicle transport in the liver, including bile acid and electrolyte secretion, transcytosis, receptor-mediated endocytosis, and autophagic and lysosomal trafficking. The potential contributions of defects in MT-based vesicle transport to alcoholic liver disease are also considered.

The second part of this review focuses on MT-dependent transport in kidney function, discussing the movement of channels and transporters to and from the apical membrane of renal epithelia. Current theories implicating defects in MT-based vesicle transport in polycystic kidney disease and in kidney damage caused by ischemia and reperfusion are also delineated.

	II. BACKGROUND

Top Previous Next References

The biochemically distinct ends of the MT can be organized into several different arrays that are used as cues to delineate cell polarity. Structurally, MT consist of 13 protofilaments organized around a central hollow core. These protofilaments are assemblies of tubulin organized in a head-to-tail array, providing the MT with its intrinsic biochemical polarity. The distinct ends of the MT are referred to as plus ends and minus ends, which describe the dynamics of MT assembly from each end in vitro (reviewed in Ref. 48). Microtubule minus ends are usually anchored at a MT organizing center (MTOC), which consists of a pair of centrioles surrounded by diffuse centriolar material. Through interactions with a third form of tubulin, -tubulin (84, 123, 148) and with other structural proteins of the MTOC, the polar MT are organized into an array characteristic of the cell type. Microtubule organization determines the positioning of the MT scaffold supporting intracellular MT-dependent vesicle movements and membrane organization.

Microtubule organization in epithelia is shown in Figure 1; this arrangement is characteristic of cells from intestine and kidney (2, 11, 57, 62, 108, 134). The MT are organized with their minus ends toward the apical surface and their plus ends extending through the cell body to the basolateral surface (11). This sequestration of MT plus and minus ends to the basolateral and apical surfaces distinguishes these two separate domains in epithelia. Centrosomal structures are present below the apical membranes of polar epithelia such as Caco-2 cells (108) and chick enterocytes (45), but these structures are not synonymous with MTOC. Although most -tubulin is insoluble in fibroblasts and other cell types exhibiting a radial MT organization, studies in Caco-2 cells have shown that almost half the -tubulin is soluble (108). Because the elements of the organizing centers are spread more diffusely within the apical region where the MT minus ends are anchored, it is unclear whether MT in polar epithelia such as renal epithelia are localized at distinct apical MTOC.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Microtubule arrays in different cell types. Two organizations are shown: parallel organization typical of intestinal and kidney epithelia (renal epithelia) and radial organization emanating from apical pole that is characteristic of hepatocytes (hepatocyte). Nuc, nucleus.

Recent investigations on MT organization within individual hepatocytes and hepatocyte cords (122) have revealed that hepatocyte MT radiate from an organizing center positioned at the pericanalicular region. However, additional curving MT intersected the straight MT emanating from this organizing center; it is hypothesized that this curving web defines a "centrospheric" region where the Golgi and trans-Golgi membranes are located. Some workers (152) have also shown that regrowth of MT in cultured rat hepatocytes is initiated at the bile canalicular region. These results are combined in the model shown in Figure 1, which reveals that most hepatocyte MT originate in the pericanalicular space and extend to the sinusoidal regions. A few of the curving MT described by Novikoff et al. (122) define a central sphere within the hepatocyte.

Two types of MT-based motor proteins are implicated in intracellular vesicle transport: kinesin and cytoplasmic dynein. Each of these proteins utilizes ATP hydrolysis to power membrane vesicle movements along the MT track. Both kinesin- and cytoplasmic dynein-driven vesicle motility requires additional accessory factors for vesicle movement (141, 142).

Conventional kinesin, the first kinesin to be implicated in vesicle transport, is a multimeric protein containing two heavy chains (KHC, 110-130 kDa) and two light chains (KLC, 65 kDa) (reviewed in Ref. 164). This kinesin first isolated from squid axoplasm is now known to be a member of a kinesin superfamily that includes a large and structurally diverse group of MT-based motor proteins that participate in numerous force-generating cellular activities including chromosomal segregation, protein and vesicle transport, and intracellular organization (see recent reviews in Refs. 76, 77, 116, 158). The identification of conventional kinesin resulted from a search for a "motor" protein in axoplasm that might be responsible for the rapid and directed movement of membranes within the axon seen by light microscopy (18, 159). Systematic molecular biological and genetic studies have since revealed a family of 90 related proteins that can be grouped into three different classes based on the location of the catalytic core at the NH₂ terminus of the KHC (KIN N motors, includes conventional kinesin), the COOH terminus of the KHC (KIN C motors), or the interior (KIN I motors) (158). The variation within this protein superfamily has produced both dimeric and monomeric motors capable of carrying different cargo at widely varying motility velocities and to either MT plus or minus ends. Conventional kinesin is an MT plus-end-directed motor, as are most but not all of the other cytoplasmic vesicle motors.

A comprehensive discussion on the kinesin superfamily is beyond the scope of the current review. However, some examples drawn from information on brain kinesins will illustrate general features of kinesin vesicle motors that may be echoed by future findings in liver and kidney. Molecular studies identifying heterogeneity within cytoplasmic brain kinesins (3) were rapidly followed by the demonstration that several of these kinesins had different functional properties (reviewed in Ref. 77). Some kinesins only transport specific cargo; for instance, kinesin superfamily protein (KIF) 1A transports synaptic vesicle precursors in axons, whereas KIF1B transports mitochondria. Although most brain KIF have so far demonstrated movements toward MT plus ends, recent studies suggest that KIFC2 may be a MT minus end-directed motor, consistent with its proposed role in dendritic transport (132). Another kinesin, KIF2, is involved in the transport of vesicles distinct from synaptic vesicle precursors, but only in juvenile axon. The velocities of each of these brain kinesins vary within a 10-fold range (from 0.2 to 1.5 µm/s). Other brain KIF have only been identified through molecular analysis and remain to be fully characterized. These studies in the neuron suggest that different kinesins may be utilized during development, to power the movement of different vesicles, and that these kinesins may exhibit large variations in the dynamics of vesicle movements.

Cytoplasmic dynein is a multimeric protein consisting of two heavy chains (DHC, 440-530 kDa), two to three intermediate chains (DIC, 70-74 kDa), and varying numbers of light chains (DLC, 50-55 kDa) (reviewed in Refs. 140, 164). Cytoplasmic dynein drives membrane vesicle movements to the MT minus ends. In addition to the motor complex, cytoplasmic dynein-driven vesicle motility requires another accessory factor, the dynactin complex, for vesicle motility (142). The dynactin complex is an assembly of several different proteins: dynactin (52, 80); Arp1, an actin-related protein which forms a filament constituting the major structural motif present in the dynactin complex (99, 138); p50/dynamitin (40, 126); actin-capping protein (138); and other uncharacterized polypeptides. Recent work on the organization of this complex macromolecular assembly has shown that DIC bind to dynactin (86, 162), whereas DLC are tightly associated with DHC (51).

A single cytoplasmic dynein gene was originally isolated from brain (110), and until recently, little information has been available regarding possible diversity within the cytoplasmic dynein family. Within the last few years, workers have shown that at least two cytoplasmic dynein transcripts are present in nervous tissue (153) and in sea urchin (49). Also, the localization of three distinct DHC to three different organelle populations in cultured cells has been demonstrated (157). Whether the heterogeneity of structure tolerated within the kinesin superfamily will be found in this newly emerging family of cytoplasmic dyneins has yet to be determined (discussed in Ref. 50). With the consideration of the large number of proteins involved in the formation of the massive cytoplasmic dynein/dynactin complex, it is conceivable that alternative splicing and assembly of different forms of these polypeptides can provide diversity equivalent to that provided by the many members of the kinesin superfamily (6).

Microtubule-dependent vesicle transport is an integral component of many of the membrane-trafficking events involved in endocytosis, secretion, transcytosis, and membrane organization and maintenance, as discussed in several recent reviews (31, 56). A detailed analysis of the individual trafficking events facilitated by MT-based transport is beyond the scope of this review. However, some findings on the role of MT and motor proteins in endocytosis and transcytosis that have particular relevance for the role of MT-based vesicle transport in liver and kidney are presented here to provide the appropriate framework for subsequent discussions.

	III. MICROTUBULE-DEPENDENT VESICLE TRANSPORT AND NORMAL LIVER FUNCTION

Top Previous Next References

In this section, the role of MT-dependent vesicle transport in the movement of channels and transporters involved in bile and electrolyte secretion from hepatocyte canalicular membranes is discussed in depth. We also consider the role of MT-dependent vesicle in liver transcytosis, receptor-mediated endocytosis, autophagy, and peroxisomal trafficking. When considering the many diverse functions in liver that are facilitated by MT-based vesicle transport, it is not surprising to find that the relative abundance of cytoplasmic dynein in liver (as measured by the ratio of cytoplasmic dynein to microtubule polymer) is 15 times higher than the ratio found in brain (135).

Many of the studies described here assess the impact of various MT-targeted drugs (colchicine, nocodazole, or taxol) on a particular process to infer the possible involvement of MT-based vesicle transport in that process. The concept behind this approach is that removal of MT with colchicine or nocodazole or stabilization of MT with taxol can both significantly reduce MT-based vesicle movements (for example, Ref. 66). However, colchicine, nocodazole, and taxol can elicit many other effects including changes in cellular architecture, disruption of plasma membrane subdomains, inhibition of protein synthesis, or even changes in cellular signal transduction pathways. Each of these effects may also impact the process under evaluation. The first indication that a MT-dependent vesicle transport step participates in membrane traffic is often evidence obtained from inhibitor studies. Demonstrations that a given cellular process utilizes MT-based vesicle transport can be quite convincing using MT-targeted drugs if additional information is available suggesting that vesicle movements participate in the process, that the activity of the protein itself is not directly affected, or if the changes can be confirmed by localization studies.

The ability of inhibitors of MT polymerization to reduce canalicular bile salt secretion by hepatocytes has been extensively investigated, revealing some intriguing effects. No major effect of colchicine on the basal excretion of bile salts was observed in rats (32). However, when an increased bile salt load of taurocholate was imposed on the liver, a delay in biliary taurocholate secretion was observed with colchicine (43). Also, colchicine had a differential inhibitory effect on the efflux of different bile salts into bile both under normal conditions and also after biliary drainage (depleted/reinfused). Under both conditions, the increased inhibition of bile salt secretion by colchicine was correlated with increased hydrophobicity of the bile salts (32, 59). Further investigations of the role of MT-dependent vesicle transport in secretion of bile salts of differing hydrophobicity revealed that increases in the hydrophobicity/ability to form micelles was correlated with an increased sensitivity of the secretion of that particular bile salt to colchicine treatment in rats (34). Colchicine also inhibits the biliary secretion of other organic anions including bilirubin and diethylmaleate conjugates (33, 38).

Much debate has centered on the mechanism by which protein-bound bile salts traverse the hepatocyte; both diffusion-based and vesicular mechanisms have been proposed (reviewed in Ref. 43). Bile salts enter hepatocytes at the sinusoidal membrane and bind to cytosolic bile acid-binding proteins. These complexes move to the canalicular surface, where the bile salts are extruded into bile. Because the process of biliary secretion is sensitive to colchicine, the data seemed consistent with the vesicular movement of bile salts. In fact, early support for the vesicular movement of bile salts to the canalicular region was provided by autoradiograpy, which localized internalized bile salts to the Golgi and, to a lesser extent, to the endoplasmic reticulum (43, 54, 150). Electron and fluorescence microscopy studies also revealed that bile salts detected either by antibody binding (95) or by direct labeling with FITC (90) were localized to the Golgi apparatus. Arguments against the vesicular transport model of bile salt movement across hepatocytes include a recent study that has examined the movement of fluorescent bile salts across hepatocytes using confocal microscopy in rat hepatocyte couplets (42); no evidence for sequestration or microtubule-dependent trafficking of these labeled bile salts was found. Also, the time course of bile acid movement across hepatocytes is measured in minutes, whereas a vesicular transport process might take much longer.

To explain these contradictory observations, several models have been considered (reviewed in Ref. 43). The most attractive model involves recruitment of transporters, similar to the studies discussed for renal epithelia. Imposition of a bile salt load could initiate the recruitment of additional bile salt transporters via a MT-dependent vesicle transport process governing the exocytic insertion and endocytic retrieval of membranes containing these transporters. Some additional evidence supports this hypothesis, including the observations that changes in bile acid load lead to corresponding changes in hepatocyte volume and in the flow of pericanalicular vesicles to the canalicular membrane (23, 64, 70, 71). Although the apparent localization of bile salts to the Golgi that has been reported (43, 54, 90, 95, 150) is somewhat puzzling, this phenomenon could be due to the presence of transporters in the Golgi and Golgi-derived membranes that are in transit to or from the canalicular membrane. Alternatively, this pool could represent a newly synthesized pool of bile salts that are synthesized to supplement dietarily derived bile from plasma.

Although biliary secretion at the canalicular region clearly has MT-dependent components, not all secretion occurs through MT-dependent pathways. The biliary secretion in isolated hepatocyte couplets exposed to dibutyryl cAMP and/or nocodazole was recently investigated (17). Although dibutyryl cAMP significantly elevated bile acid secretion by more than twofold in tandem with a major increase in canalicular circumference, nocodazole significantly reduced bile acid secretion by more than fourfold, concomitantly decreasing the canalicular circumference. Exposure of nocodazole-treated couplets to dibutyryl cAMP resulted in an elevation of biliary secretion and an increase in canalicular circumference relative to nocodazole-treated couplets, but both parameters were still reduced relative to controls. These studies support the presence of both MT-dependent (nocodazole-sensitive) and MT-independent (dibutyryl cAMP-sensitive) mechanisms for introducing bile acid transporters into the canalicular membrane. That biliary secretion might be differentially regulated in MT-dependent and MT-independent steps is not surprising, since investigations in recent years have identified, cloned, and characterized several different transporters responsible for the efflux of organic acids and bile salts from the hepatocyte (reviewed in Ref. 117).

As a way of sorting through the complexity inherent in the study of multiple bile salt transporters regulated by different processes, a recent study has utilized skate hepatocytes as a model system to investigate the effects of nocodazole on Na⁺-independent bile salt transport. Skates and other lower vertebrates do not express the Na⁺-coupled bile acid transporter thought to be responsible for much of the Na⁺-dependent bile salt secretion in mammalian systems. This study utilized fluorescently labeled taurocholate (NBD-taurocholate) and measured the extrusion of this fluorescent probe into the canalicular lumen (112). The Na⁺-independent transport of this probe to the lumen was reversibly inhibited by cold treatment or nocodazole washout, consistent with the model of MT-dependent insertion of Na⁺-independent transporters into the canalicular region after imposition of a bile salt load.

The MT-dependent movement of transporter-containing vesicles to the canalicular membrane is an attractive model for coping with an increased bile salt load, particularly since this general mechanism occurs in many other epithelia (reviewed in Ref. 4). However, the increased sensitivity to colchicine as bile salt hydrophobicity increases is puzzling. The cytoskeleton plays a major role in organization of the plasma membrane and maintenance of subdomains; changes in membrane fluidity caused by increasing hydrophobicity of bile salts might lead to an increased requirement for MT in maintaining transporters already in the plasma membrane, rendering hepatocytes exposed to more hydrophobic bile salts more sensitive to colchicine through a mechanism independent of MT-based vesicle transport. Microtubule and microfilament fragmentation is thought to govern access of secretory vesicles to the apical plasma membrane in some specialized secretory epithelia (for example, Ref. 85) in a model known as the cytoskeletal barrier model. However, if MT served an analogous barrier function in hepatocytes, their disassembly by colchicine should lead to insertion of canalicular vesicles containing bile salt transporters. However, it is difficult to reconcile the differential effects of colchicine on the secretion of bile salts of increasing hydrophobicity with the cytoskeletal barrier model.

Within the context of the working model discussed above, it is possible that transporters for the more hydrophobic bile salts are enriched in cytoplasmic vesicles, whereas the transporters for less hydrophobic bile salts are enriched at the canalicular surface. It is also possible that micelle formation by hydrophobic bile salts can directly signal to components of the membrane trafficking machinery. Along these lines, a recent study has revealed that liver kinesin MT gliding activity is inhibited by bile salts of increasing hydrophobicity (chenodeoxycholate conjugates), but less so by bile salts with less hydrophobicity (ursodeoxycholate or cholate conjugates) (106). This phenomenon suggests MT-based motor activity may be directly modulated by bile salts and that such effects may mediate the adaptive response in liver to a high bile salt load under physiological conditions.

Several recent studies have addressed the mechanism of activation of the Cl/HCO₃ exchanger, revealing a potential role for MT-dependent transport in the process of exchanger mobilization. Use of the Na⁺-independent Cl/ HCO₃ exchanger is one means by which hepatocytes compensate for increased alkalinization (109). This exchanger is somehow activated by imposition of an alkaline load. One study on activation of the Cl/HCO₃ exchanger utilized isolated perfused rat liver to demonstrate that alkalinization led to a 40% increase in bile flow that could be reduced by 80% in the presence of colchicine (23). A plausible explanation for this finding would be that the Cl/HCO₃ exchangers and one or more bile salt transporters were localized on the same pericanalicular membranes mobilized in response to alkalinization in hepatocytes. In this study, horseradish peroxidase (HRP) excretion into bile was also significantly increased with alkalinization, revealing that upregulation of the transcytotic pathway appeared to occur in tandem with mobilization of Cl/HCO₃ exchangers.

Subsequent studies revealed that the culture of hepatocytes under alkaline conditions normally promotes an increase in the maximum acid influx due to Cl/HCO₃ exchanger activity; however, this increase could be blocked by prior exposure of hepatocytes to colchicine (12). Also, elevation of cAMP could mimic the effects of alkaline medium, but this response to cAMP was also prevented by colchicine. These findings were consistent with the previous work in isolated perfused rat liver (23), supporting the existence of a pool of Cl/HCO₃ exchangers that could be mobilized by alkaline medium in a colchicine-sensitive manner; this work also suggested that hepatocytes rather than ductal epithelia were responsible for the increased bile flow in response to alkaline conditions reported previously.

Another explanation for the sensitivity of exchanger activity to colchicine in these studies (12, 23) might be that colchicine caused effects on Cl/HCO₃ exchanger activity in the membrane independently of MT-dependent vesicle transport (MT may interact directly with the exchanger in the plasma membrane or maintain optimal membrane fluidity for exchanger activity). This proposed model of vesicle movement to and from the apical membrane along MT was strengthened by observations that the imposition of an alkaline load on hepatocytes also caused movement of the peanut agglutinin receptor and the ecto-ATPase from their normal pericanalicular localization to the canalicular region, an effect which could again be blocked by colchicine treatment (12).

The mechanism involving movement and/or fusion of these pericanalicular membranes with the canalicular membrane is likely to involve MT-based vesicle transport. Another possible explanation involves abolition of the regulated response by colchicine due to premature fusion of vesicles containing channels/exchangers with the canalicular membrane, analogous to the cytoskeletal barrier model mentioned above. Microtubule stabilization (with taxol) would be expected to inhibit MT-based vesicle transport but to permit MT fragmentation, so taxol might be used to further discriminate between these working models. Further support for the role of MT-based vesicle transport would also be confirmation that the channels/exchanger composition of the apical membrane is increased after alkalinization but not after nocodazole or colchicine exposure.

Changes in hepatocyte volume are linked with changes in bile flow, which in turn involves MT-dependent vesicle transport. In particular, the role of MT-based vesicle transport in the biphasic increase in bile flow elicited by hypotonic cell swelling was probed in isolated perfused rat liver from control and colchicine-treated rats exposed to hypotonic stress before measurements of biliary and HRP excretion (22). Colchicine but not lumicolchicine blunted the normal increase in bile flow in response to hypotonic stress. Furthermore, the normal increase in HRP excretion in response to hypotonic stress was also reduced by colchicine. These researchers proposed that hypotonic stress elicited exocytosis of pericanalicular vesicles containing fluid and bile salt transporters via MT-dependent vesicle transport. These observations are similar to the phenomenon of coordinate alkalinization-induced increases in Cl/HCO₃ exchanger activity and bile flow reported by Bruck et al. (23).

The work summarized here demonstrates that diverse perturbations (increased bile salt load, alkalinization) result in the insertion of the appropriate transporters from pericanalicular stores into the canalicular membrane through an apparent MT-dependent vesicle transport mechanism to maintain homeostasis. No evidence for major changes in the MT array itself are found under these conditions, supporting the concept that sensor molecules are able to convey these diverse signals to the vesicle transport machinery. This working model for insertion of pericanalicular transporters will be strengthened as antibodies to the transporters of interest become available, so that colchicine-induced changes in transporter activity can in some cases be correlated with biochemical and microscopic localization of the transporters on the appropriate membranes before and after stimulation. A major area for future investigations should be the elucidation of the possible intracellular signaling mechanisms activated by these diverse treatments and the examination of the nature of the signal(s) derived from these signaling mechanisms that converge on the MT-based vesicle transport machinery.

Several studies support the involvement of MT-based vesicle transport in transcytosis. Hepatic transcytosis serves a major role in the excretion of ligands and fluid-phase proteins such as dimeric IgA, HRP, and low-density lipoproteins from the sinusoidal circulation into the bile (reviewed in Refs. 119, 149). Also, transcytosis facilitates the sorting of proteins that leads to the development of functionally and biochemically distinct canalicular and sinusoidal plasma membranes.

Studies by Durand-Schneider et al. (39) have shown that rat hepatocytes isolated and cultured in the presence of colchicine and nocodazole develop bile canaliculi that appear normal by electron microscopic examination. However, colchicine or nocodazole exposure inhibits the ability of these hepatocytes to transport and concentrate fluorescein diacetate into the bile canalicular region, a property exhibited by untreated hepatocytes. Moreover, the membrane protein B10 that normally exhibits an exclusively apical distribution in untreated hepatocytes is shifted to a punctate vesicular distribution in colchicine- or nocodazole-pretreated hepatocytes. No differences in the distribution of B1, a membrane protein that distributes to sinusoidal membranes, were seen in colchicine- or nocodazole-pretreated hepatocytes.

Recent investigations by Hemery et al. (74) have expanded on this earlier work. By exposing cells to fluorescently labeled anti-B10 monoclonal antibody, these workers followed the traffic of the canalicular marker B10 from the sinusoidal to the canalicular membrane. In parallel with fluorescently labeled internalized transferrin (Tf) or asialoorosomucoid (ASOR), B10 moved first to a perinuclear compartment and then to its final destination at the canalicular membrane. Nocodazole pretreatment prevented the movement of B10, Tf, and ASOR to this perinuclear compartment, causing the dispersal of these markers at the cell periphery. When the hepatocytes were given anti-B10 antibody under conditions that allowed the labeling of the perinuclear compartment before exposure to nocodazole, the movement of B10 from the perinuclear compartment to the canalicular region was also blocked. This study suggests the existence of at least two MT-dependent steps in the transit of a resident canalicular protein from the sinusoidal surface to the bile canaliculus.

Consistent with a role for MT-based vesicle transport in the transcytosis of internalized constituents followed by their release into bile, colchicine treatment decreases dimeric IgA secretion into bile (53, 118). One study has correlated this impaired traffic through the transcytotic pathway with an accumulation of dimeric IgA in the peripheral region of hepatocytes after colchicine treatment (53). Similarly, Saucan and Palade (137) have shown that exposure of rats to colchicine results in a marked delay in the trafficking of polymeric IgA receptor to the canalicular membrane. Isolated rat hepatocyte couplets exposed to HRP revealed an initial labeling of submembranous structures below the sinusoidal surface followed by movement of the label to tubulovesicular structures at the pericanalicular region after 10 min; the translocation to the pericanalicular region was blocked by treatment with colchicine (133).

Surprisingly, the work by Saucan and Palade (137) also revealed that the albumin secreted into bile was derived from newly synthesized stores in the hepatocytes and that albumin secretion into bile was increased by colchicine pretreatment. This particular finding is reminiscent of models proposing a cytoskeletal barrier function in regulated secretion, in contrast to the many other studies discussed previously that movement of transporters to the canaliculus depends on MT. The movement of newly synthesized secretory proteins to bile may therefore proceed via a different pathway than transcytosed material or resident pericanalicular transporters.

As mentioned previously, the activation of biliary secretion via imposition of a bile salt load also enhances the flow of material through the transcytotic pathway. Pulse loading of isolated perfused rat liver with HRP followed by stimulation of bile flow with taurocholate revealed a biphasic excretion pattern into bile, with a rapid early phase (paracellular diffusion) and a prolonged second phase (transcytotic pathway). This second phase of secretion into bile, representing the transcytotic pathway, was blocked by colchicine pretreatment, consistent with the movement of HRP through a MT-dependent vesicle transport pathway (72). This study also observed that stimulation with taurocholate resulted in accumulation of HRP in small pericanalicular vesicles after 2 min of pulse loading; the loading and density of these vesicles were not reduced by colchicine, suggesting that a rapid MT-independent pathway responsible for transport to the pericanalicular region may exist (72).

These findings support a role for MT-based vesicle transport in the movement of resident canalicular proteins, ligands, and soluble proteins through the transcytotic pathway in at least two places: from the peripheral regions of the hepatocyte to the perinuclear region and from the perinuclear region to the canalicular region. The upregulation of transcytosis observed in parallel with imposition of a bile salt load also seems to utilize MT-dependent vesicle transport, although insufficient information is available to determine whether two MT-dependent steps are also present in this activated pathway.

Perhaps the best-characterized form of MT-dependent vesicle transport in the liver is the cytoplasmic dynein-driven receptor-mediated endocytosis of asialoglycoproteins (ASG). The hepatocyte-specific endocytosis of ASG occurs when ASG binds to its receptor, the ASG receptor (ASGR), and the complex is internalized and processed. This process has been extensively analyzed using microscopic and biochemical methodologies. Density gradient centrifugation of membranes obtained from rat hepatocytes incubated with ASG for various times showed that ASG were first recovered in lower density fractions with properties of endosomes and prelysosomes; with increasing exposure, ASG began to accumulate in higher density fractions with properties of dense lysosomes (13). Exposure of hepatocytes to colchicine reduced the amount of ASG recovered in the dense lysosomal fraction, suggesting that trafficking from endosomal/prelysosomal to lysosomal compartments was inhibited after colchicine treatment. Consistent with this study, density gradient centrifugation of membranes from rat hepatocytes showed that endocytosed ASG accumulated in early endosomes after vinblastine treatment (78). Another study used electron microscopy to confirm that the uptake and intracellular movement of endocytosed ASG in hepatocyte couplets was significantly inhibited by colchicine (69).

The nature of the MT-dependent step in endocytosis of ASG-ASGR was explored in biochemical studies that probed the affinity of vesicles containing ASG-ASGR for MT (55). In these studies, the investigators utilized transmission electron microscopy to show that endocytosed ASG was recovered in endosomal vesicles that were closely associated with MT. They devised a procedure for purification of membrane vesicles by cosedimentation with MT and asked whether ASG and ASGR were recovered in membranes purified by this MT-affinity technique. Their results revealed that both ASG and ASGR were pelleted with MT but that only ASG was released from membranes after addition of ATP; cytoplasmic dynein recovery from MT paralleled that of ASG but not ASGR. These findings suggested that cytoplasmic dynein might facilitate the sorting of ASG from ASGR in the early endosome and power the movement of ASG-containing vesicles to lysosomes.

To further resolve the role of cytoplasmic dynein in trafficking of the ASG-ASGR complex, an additional series of biochemical experiments was performed (124). The parameters of kinesin, cytoplasmic dynein, and vesicular ASG and ASGR binding and release from MT were determined. The results clearly showed that cytoplasmic dynein and vesicular ASG binding and release from MT were highly correlated, supporting the hypothesis that cytoplasmic dynein mediated ASG traffic to the lysosome. Kinesin-MT binding was not correlated with either ASG- or ASGR-MT binding. In confirmation of the proposed role for cytoplasmic dynein in the trafficking of ASG, immunoprecipitation of cytoplasmic dynein from liver yielded recovery of 15% of the internalized ASG, whereas immunoprecipitation of kinesin yielded <1% of internalized ASG. This evidence is consistent with concurrent work in cultured cells implicating cytoplasmic dynein as the motor powering MT-based vesicle movement to the cell interior from the endosomes (7).

Although the investigations on internalization of ASG-ASGR reveal a role for cytoplasmic dynein-driven vesicle transport in movement from endosomes to lysosomes, recent studies on internalization of Tf via Tf receptor (TfR)-mediated endocytosis in hepatocytes have also indicated a MT-dependent step at an earlier point within the endocytic pathway (130). Nocodazole treatment caused a significant reduction in initial (15 min) as well as steady-state uptake Tf uptake; this effect could be mimicked by inhibition of protein phosphatase 2A with microcystin or okadaic acid, which are known to reduce MT-dependent vesicle movements in hepatocytes (67). Because no significant changes in the number of TfR in the plasma membrane were observed at the beginning of the pulse with ¹²⁵I-labeled Tf under these conditions, the effects of nocodazole were thought to delineate a MT-dependent step in TfR internalization. Although MT dependent, this step may not necessarily involve MT-based vesicle transport but rather involve a role of MT in receptor clustering or sequestration in the plasma membrane. These studies on an early MT-dependent step in receptor-mediated endocytosis of TfR are consistent with previous reports in cultured cells (83, 154).

A recent study on the trafficking of autophagic vesicles in liver has implicated MT-based transport. Autophagy is the degradation of intracellular macromolecules after their sequestration from cytoplasm into autophagic vesicles formed by membranous endoplasmic reticulum-derived cisternae. As the autophagic vesicles mature, they acquire lysosomal properties and ultimately converge with the endocytic pathway at the lysosomes (143). Fengsrud et al. (46) demonstrated that vinblastine blocked the accumulation of BSA in the peripheral endosomes that was normally seen in hepatocytes. Also, vinblastine prevented incorporation of BSA into the autophagic vacuole while increasing the number of prelysosomal autophagic vacuoles present in hepatocytes. The conclusion of this study was that MT-dependent vesicle transport is required to mediate the fusion of autophagic vacuoles with the endocytic pathway at prelysosomes or lysosomes.

A role for MT in organizing the peroxisomal compartment in liver has recently been proposed. Peroxisomes are responsible for the oxidative catabolism of lactate, oxalate, fatty acids, polyamines, and other metabolically active low-molecular-weight compounds; the oxidases found in peroxisomes generate H₂O₂ as a by-product of oxidation, which is reduced by catalase also found in peroxisomes (97). The peroxisomal compartment (enriched in catalase, which constitutes 16% of total peroxisomal protein) was shown to have a heterogeneous distribution in Hep G2 cells (139); elongated tubular organelles of up to 5 µm in length were seen in the lower density cells, whereas a mixture of rod-shaped (0.5 µm) and spherical single organelles was seen in higher density cells. In untreated cells, the rod-shaped and spherical organelles were never seen in the same cell with the tubular organelles. Treatment with colcemid, nocodazole, or vinblastine promoted increases in tubular peroxisome formation, although taxol had no effects. Fluorescence microscopy showed that nocodazole-treated cells contained clusters of spherical peroxisomes and tubular peroxisomes. If the treatments were removed, normal morphology was recovered and confocal microscopy could verify that the peroxisomes were localized to the MT. In vitro MT binding was assessed by video-enhanced DIC microscopy with purified peroxisomes; binding to MT was avid but abolished in response to proteinase K, suggesting that a linker protein present on the peroxisome (motor or cytoplasmic linker protein) mediated this attachment. Recent studies have confirmed the association of MT and peroxisomes, using antibodies to a resident 70-kDa peroxisomal membrane protein (73).

Liver MT also seem to support membrane traffic from lysosomes to the canalicular membrane. The release of hepatic lysosomal contents into bile has been shown to have both an MT-dependent and an actin-dependent component (144). Also, this release of lysosomal enzymes to bile is stimulated by bile acids. The mechanism of bile acid-stimulated traffic from lysosomes to the bile canaliculus was recently examined in depth (100). To rule out the possibility that bile acids were able to stimulate traffic from the lysosomes via effects on membrane fluidity, hydrophobic, hydrophilic, and non-micelle-forming bile salts were evaluated for their ability to enhance lysosomal secretion and alter plasma membrane and lysosomal fluidity. Although all groups equally enhanced output, only the hydrophobic bile salts altered membrane fluidity, suggesting that changes in hydrophobicity did not explain the bile acid stimulation. In support of the earlier work by Sewell et al. (144), taurocholate-dependent increases in lysosomal enzyme secretion were completely blocked by colchicine or vinblastine. Fluorescence microscopy revealed that taurocholate caused the translocation of lysosomal membranes to the bile canalicular region in rat hepatocyte couplets with a time course consistent with the taurocholate-stimulated release of lysosomal enzymes to bile. These findings support a MT-dependent vesicle transport mechanism responsible for driving the movement of lysosomal membranes to the canalicular membrane. The current status of the findings on the role of MT-based vesicle transport in organization of the autophagic and peroxisomal pathways and the traffic to and from the lysosomes is shown in Figure 2.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Role of microtubule-based vesicle transport in movement to and from lysosomal compartment in hepatocytes. Role for microtubule-based vesicle transport and, specifically, cytoplasmic dynein in movement of endocytic carrier vesicles (ECV) from endosomal compartment (EC) to lysosomal compartment (LYS) is well established. Recent studies suggest that uptake to EC may also be facilitated by microtubules. Additionally, movement of autophagic vacuoles (AV) to lysosomal network and maintenance of peroxisomal (PX) network is facilitated by microtubules, although an individual motor has not yet been implicated in these events. Finally, movement of lysosomal constituents to bile is facilitated by microtubules.

Many of the studies to date on the role of MT-based vesicle transport in the liver have been plagued by the necessity of using difficult models for the study of partially characterized processes. Most of the evidence that MT and vesicle transport play a role in liver membrane trafficking has been obtained through inhibitor studies. The application of biochemical and molecular approaches would reinforce many of the conclusions drawn from the initial studies. For instance, where colchicine-induced changes in bile salt secretion into bile are thought to arise from the MT-dependent movement and fusion of vesicles containing bile salt transporters to the canalicular membrane, demonstration via localization studies that transporter contents are increased and decreased accordingly in the apical membrane would be important confirmation of this theory. Some biochemical approaches have been initiated such as the characterization of the role of cytoplasmic dynein in ASGR-mediated endocytosis (55, 124). As the transporters (in this case, bile salt transporters or organic anion transporters) become better characterized and more antibodies are generated to these proteins, information about their trafficking through MT-dependent pathways will also become more accessible. The cloning and characterization of several bile salt and organic anion transporters will facilitate this area tremendously (reviewed in Ref. 117). The availability of WIF-B cells and other simpler cell models that retain essential properties of hepatocytes such as formation of bile canaliculi will also enable the introduction of more molecular approaches.

	IV. MICROTUBULE-DEPENDENT VESICLE TRANSPORT AND LIVER DISEASE: ETHANOL MAY IMPAIR MICROTUBULE-BASED VESICLE TRANSPORT

Top Previous Next References

At the cellular level, exposure to ethanol results in decreases in receptor-mediated endocytosis (27, 28) and also decreased biliary secretion and accumulation of proteins within the Golgi apparatus (60). The reduction in endocytosis is known to encompass reductions in the rates of internalization within coated pits (26) and reductions in the movement of ligand from endosomes to lysosomes (29). As discussed in the previous section, many of these membrane trafficking changes exhibited by the alcoholic liver are normally facilitated by MT-dependent vesicle transport.

The inhibition by ethanol of these membrane trafficking events is not well understood, although impaired MT-dependent vesicle transport has been suggested as one cause. In support of this hypothesis, MT-dependent vesicle movement in hepatocytes treated by acute or chronic ethanol exposure was measured by video microscopy; both acute and chronic ethanol exposure reduced MT-dependent vesicle movements (151). When the mechanism underlying this observation was explored, no changes in kinesin and cytoplasmic dynein synthesis or membrane binding were observed in samples from the ethanol-exposed livers. However, kinesin and cytoplasmic dynein MT-activated ATPase activity were substantially decreased in vitro after exposure of motors and MT to acetaldehyde, the active by-product of ethanol metabolism. This same study also revealed changes in the association of dynamin, a protein involved in generation of endocytic vesicles, to membranes.

Acetaldehyde can form covalent adducts with the Lys residues on different proteins. Several studies have demonstrated that tubulin has a high affinity for acetaldehyde, specifically -tubulin (82, 156). The formation of acetaldehyde -tubulin inhibits tubulin assembly into MT in a number of systems (146, 147). It is possible that reduced MT formation in ethanol-exposed hepatocytes also leads to impairment of the MT-based transport events involved in endocytosis and secretion. Although the relationships between acetaldehyde-modified MT and inhibition of membrane trafficking have not been conclusively demonstrated, this model remains an attractive one to explain the major deficits in membrane trafficking in the alcoholic liver.

	V. MICROTUBULE-DEPENDENT VESICLE TRANSPORT IN NORMAL KIDNEY FUNCTION

Top Previous Next References

As in the liver, numerous studies show that the insertion and/or retrieval of channels and transporters at the apical surface of kidney epithelia is MT dependent. This constant and dynamic cycling of channels and transporters is responsible for ultrafiltration of plasma by the glomerulus, the reabsorption of water and solutes from the ultrafiltrate, and the secretion of solutes including organic anions and cations into the tubular fluid. The process of reabsorption and secretion enables the renal tubules to modulate the volume and composition of the urine. Because the bulk of the water, Na⁺, Cl⁺, and other solute absorption from the glomerular ultrafiltrate occurs in the proximal tubule, most studies on channel modulation have focused on these cells; however, a few studies on epithelia from the collecting duct and distal nephron reveal similar patterns of trafficking. Although renal epithelia maintain a vigorous apical endocytosis and recycling pathway involving clathrin-coated vesicles (20), recent and ongoing investigations suggest that MT-dependent vesicle transport is integral in the establishment of epithelial polarity and the maintenance and regulation of channel and transporter composition of the apical plasma membranes in the kidney.

Establishment of the appropriate distribution of proteins in the apical membrane during polarization of kidney epithelia is dependent on the MT network. These epithelia lining the proximal and distal tubules maintain a MT distribution similar to gut epithelia or the MDCK cell line, with apical organizing centers anchoring MT minus ends and plus ends extending to the basolateral region (10). Work examining the establishment of polarity in LLC-PK₁ cells, a proximal tubule-derived cell, revealed that segregation of the Na⁺-hexose carrier to the apical plasma membrane was inhibited by pretreatment with either nocodazole or colchicine (161). This same study revealed parallel effects of nocodazole and colchicine on two other apical membrane markers, -glutamyltransferase and alkaline phosphatase, strengthening the conclusions of these data derived from inhibitor studies. No effect of these MT-targeted agents on the sorting of Na⁺-K⁺-ATPase to the basolateral membranes was elicited by nocodazole or colchicine, suggesting that the role of MT in governing the traffic of plasma membrane proteins was specific to apically targeted proteins in this system.

The apical expression of the intrinsic factor-cobalamin receptor in renal proximal tubular epithelia is also dependent on the cellular MT array (127). This receptor is responsible for scavenging cobalamin bound to intrinsic factor at the apical surface and then transcytosing this complex from the apical to the basolateral surface. Treatment of cells with either nocodazole or colchicine resulted in loss of the intrinsic factor-cobalamin receptor from the apical plasma membrane; also, transcytosis of cobalamin was reduced in parallel with loss of apical receptor. Consistent with these findings, in proximal tubules isolated from colchicine-treated rats, the gp330 membrane protein was redistributed into cytoplasmic and basolateral vesicles, in contrast to its normally apical localization (61).

The distribution of water channels in proximal tubule and collecting duct is also determined by MT-dependent vesicle transport events. The factors controlling the distribution of the water channel aquaporin-2 (AQP2) were examined in rat collecting duct principal cells (131). Normally this channel is localized at the apical plasma membrane and in apical vesicles. Exposure of rats to colchicine led to redistribution of AQP2 into cytoplasmic vesicles, with little localization to the apical membrane. Elkjaer et al. (41) obtained similar results for aquaporin-1 (AQP1) in proximal tubule; the normally intense labeling of the apical membrane obtained by exposing proximal tubular epithelia with antibodies to AQP1 is shifted to a diffuse cytoplasmic labeling in colchicine-treated rats. Consistent with these findings, MT disruption significantly reduces vasopressin-induced transepithelial water flow and H⁺ secretion (reviewed in Ref. 20).

The regulation of channel and membrane protein distributions within the apical membrane of renal epithelia in response to diverse extracellular signals is also modulated by MT-dependent vesicle transport mechanisms, as evidenced by the sensitivity of this process to changes in the MT array. In epithelia from the distal nephron of Xenopus laevis, the role of MT in antidiuretic hormone (ADH)-induced protein secretion and amiloride-sensitive transepithelial conductance due to insertion of apical Na⁺ channels was investigated (163). Nocodazole blocked both the ADH-stimulated protein secretion and ADH-mediated increase in amiloride-sensitive transepithelial conductance (proportional to insertion of Na⁺ channels). Because these two events normally occur in parallel, the investigators proposed that the effects of nocodazole on both protein secretion and Na⁺ channel insertion were elicited through inhibition of a MT-dependent step in exocytosis.

Microtubule-dependent vesicle transport has been suggested to mediate the adaptive response to phosphate governed by the Na⁺-P_i cotransporter. Epithelia within the kidney are acutely sensitive to the concentration of extracellular phosphate (P_i). The Na⁺-P_i cotransporter modulates P_i balance; when extracellular P_i is low, the Na⁺-P_i cotransporter is inserted into the apical membrane causing an increase in P_i influx, and when extracellular P_i is high, the Na⁺-P_i cotransporter is internalized, causing a decrease in P_i influx. Internalization of Na⁺-P_i cotransporter in low P_i-adapted cells can also be stimulated by parathyroid hormone (PTH). When proximal tubule-derived opossum kidney (OK) cells were exposed to low P_i, the normal increased influx of P_i attributable to mobilization of Na⁺-P_i cotransporter was reduced by 50% after exposure to either colchicine or nocodazole (68). Interestingly, taxol had no effect on this adaptive response. Nocodazole did not block the PTH-induced reduction in P_i influx in adapted cells due to cotransporter loss. These results led to the authors' conclusion that MT are involved in the adaptive response modulating the number of Na⁺-P_i cotransporters in the apical membrane; moreover, because effects on retrieval of transporters by PTH were not seen, these investigators concluded that the nocodazole- and colchicine-sensitive step involved insertion and not retrieval of the cotransporter.

Similar data demonstrating a role for MT in the adaptive response to low P_i were obtained by Lotscher et al. (103). Their studies revealed that a low-P_i diet was associated with a threefold increase in the Na⁺-P_i cotransporter abundance within the apical brush-border membrane. This increase appeared to be due to redistribution of intracellular stores, since total Na⁺-P_i cotransporter content of the cells was unaltered. Colchicine-induced disruption of the MT network blocked the adaptive increase in cotransporter content of the apical membrane; however, colchicine did not block the downregulation of Na⁺-P_i cotransporter due to exposure of low-P_i adapted cells to high P_i. These results are similar to those of Hansch et al. (68), although the effects of taxol were not determined.

Another study has associated the low-P_i adaptive response with a transient (5 min) loss of MT and microfilaments that corresponds to an increase in P_i influx (125). However, taxol-induced MT stabilization did not block the increased P_i influx in this system. These findings led to the consideration of an alternative model to explain the role of MT in the adaptive response that may also account for the dissimilar effects of colchicine and taxol also reported by Hansch et al. (68). If MT-dependent vesicle transport drives the movement of Na⁺-P_i cotransporters to the apical membrane, colchicine and taxol would be expected to elicit comparable effects. If MT instead serve a barrier function, then their removal by nocodazole could allow premature/unregulated insertion of Na⁺-P_i cotransporters into the apical membrane, abolishing the subsequent adaptive response. Also, this role would still be consistent with the observed increase of Na⁺-P_i cotransporters in membranes after replenishment of P_i in low-P_i adapted cells (103). If fragmentation of the MT barrier is involved in mediation of this adaptive response, as indicated by the observations of Papakonstanti et al. (125), then taxol treatment might maintain the MT barrier but permit MT fragmentation during the adaptive response, explaining the discrepancy between the effects of taxol and nocodazole/colchicine. As precedent, MT and microfilaments serve a barrier function in regulated secretion in other systems including the secretory epithelia comprising the lacrimal gland and pancreas (for example, Ref. 85).

A role for MT-based vesicle transport in secretion of cGMP through insertion of an organic anion transporter in LLC-PK₁ cells has also been proposed (30). Exposure of renal tubules to nitric oxide (sodium nitroprusside) or atrial natriuretic peptide normally results in stimulation of soluble or membrane-associated guanylyl cyclase, production of cGMP, and mobilization of channels causing increased Na⁺ excretion (decreased short-circuit current). This response is enhanced by additional release of cGMP by the stimulated cells that has a downstream effect. Addition of nocodazole to LLC-PK₁ cells before stimulation with sodium nitroprusside blocked the decrease in short-circuit that would normally result; however, addition of nocodazole before stimulation directly with cGMP did not block the decrease in short-circuit current. These findings are consistent with inhibition of cGMP release by the nocodazole-treated cells. The authors (30) hypothesize that addition of nocodazole prevents extrusion of an organic anion transporter that is normally mobilized to the plasma membrane to allow release of cGMP to downstream cells, although it seems conceivable again that the absence of a MT barrier rather than the loss of MT-based vesicle transport may also have eliminated the response to stimulation with nitroprusside.

The vigorous apical endocytosis characteristic of renal epithelia involves MT-based vesicle transport. A detailed study of the role of MT in apical endocytosis was recently performed in renal proximal tubule cells from untreated versus colchicine-treated rats (41). In untreated rats, the apical surface of the proximal tubular epithelia contained numerous endocytic invaginations, small vacuoles, and dense apical tubules (which contain contents destined for insertion into apical plasma membrane). In colchicine-treated rats, the proximal tubular epithelia revealed major changes at the apical membrane including the absence of endocytic invaginations, an accumulation of large endocytic vacuoles, and a fivefold reduction in the dense apical tubules. Colchicine also inhibited the apical uptake of HRP and increased the numbers of small vesicles found in the basolateral regions of proximal tubular epithelia.

These authors suggest that the primary disruption of apical traffic by MT-targeted drugs that is manifested both by altered channel and membrane protein composition (previous section) and by reduced endocytosis is due primarily to disruption of endocytosis rather than exocytosis. They propose that dense apical tubules normally originate from the fusion of small and large endocytic vesicles at the apical plasma membrane; however, the colchicine-induced inhibition of endocytosis reduces the amount of incoming membrane available for dense apical tubule formation. Disruption in dense apical tubule function then leads to reduction in the outgoing traffic from the dense apical tubules, explaining the reduced insertion of apical channels such as AQP1 and gp330 into the apical membrane. They attribute the increase in vesiculation at the basolateral surface to an accumulation of incoming vesicles that cannot fuse into dense apical tubules, because of an additional block in MT-dependent traffic. Although this hypothesis is consistent with their data, the converse situation could also be true. It is plausible that dense apical tubule function and organization are disrupted by colchicine treatment, leading to inhibition of traffic out of the dense apical tubule accompanied by vesiculation and shrinkage of existing dense apical tubules. This effect in turn might trigger reductions in endocytosis, through a feedback mechanism or by sequestration of factors required for endocytosis. This study does demonstrate that changes in the MT array lead to inhibition of endocytosis, but it fails to show whether the inhibitory effect is direct or indirect via coupling to a reduction in exocytic traffic.

New studies have suggested that MT and microfilaments may both participate in apical endocytosis, at least of albumin. Proximal tubular fluid contains high concentrations of albumin from plasma; however, only a small percentage of the albumin is actually excreted in urine, with the remainder recovered by endocytosis at the apical surface of proximal tubule epithelia (104). The mechanism of albumin endocytosis at the apical surface was probed in depth in OK cells, a model system for proximal tubule (47). Exposure of OK cells to nocodazole reduced apical albumin endocytosis by ~50%, consistent with the studies described in the previous section (41). Cytochalasin D exposure, which promotes disassembly of F-actin, completely inhibited albumin endocytosis in OK cells, suggesting that F-actin is essential for apical endocytosis. Additional studies have suggested that disruption of F-actin prevents the pinching off of coated pits from the apical membrane (58). This study suggests that the MT-dependent trafficking from the apical membrane may be downstream of the initial internalization event, possibly involved in the sorting of internalized albumin-containing vesicles to internal sites rather than the actual formation of the endocytic vesicles.

In agreement with findings that MT participate in apical endocytosis, Verrey et al. (163) found that nocodazole was able to significantly reduce ADH-induced apical endocytosis in epithelia from the distal nephron. However, Hansch et al. (68) did not observe nocodazole-induced inhibition of PTH-mediated internalization of the Na⁺-P_i cotransporter in low-P_i adapted cells. Lotscher et al. (103) also observed no effect of colchicine on the acute downregulation of Na⁺-P_i cotransporter in cells exposed to high P_i after adaptation to low P_i. It is conceivable that more than one endocytic pathway exists at the apical plasma membrane; this proposal becomes more attractive particularly when considering the requirements of the different trafficking pathways utilized for endocytosis of albumin versus endocytosis of channels sensitive to the extracellular environment. Some proteins may be retrieved through a MT-dependent pathway and recovered in the dense apical tubules, whereas others may enter a MT-independent pathway that bypasses the dense apical tubules.

Microtubule-dependent vesicle transport is clearly a contributor to the constant membrane trafficking that occurs between the dense apical tubules and/or apical vesicles and the apical surface, which is modulated by extracellular signals including ions and hormones. This function parallels the demonstrated role for MT-based vesicle transport in regulation of canalicular liver channels previously discussed. A summary of the channels and transporters that may utilize MT-based vesicle transport for apical targeting and insertion is shown in Table 1. Microtubule disassembly induced by either colchicine or nocodazole is associated with reduced insertion of transporters and other membrane proteins into the apical plasma membrane from dense apical tubules. Also, endocytosis at the apical plasma membrane is reduced by colchicine- or nocodazole-induced MT disassembly. The evidence supports both the MT-dependent vesicle transport of traffic derived from the dense apical tubules as well as the MT-dependent retrieval of apical membranes. However, endocytosis and exocytosis at the apical plasma membrane may be very tightly regulated, leading to the suggestion that a direct block in traffic in one direction due to inhibition of MT-based vesicle transport may decrease transport in the other direction through an unknown mechanism. The model in Figure 3 incorporates these data into a working model for MT-dependent apical traffic in renal epithelia.

View larger version (39K): [in this window] [in a new window]   FIG. 3.   Microtubules facilitate insertion and retrieval of apical membranes in renal epithelia. A: ion and water channels stored in apical tubules utilize microtubule (MT)-dependent vesicle transport to reach membrane after stimulation with appropriate signal (fluid or electrolyte imbalance). B: in absence of microtubules, insertion of channels does not occur, accumulation of small vesicles at apical surface is observed, and endocytosis is also blocked. APM, apical plasma membrane; DAT, dense apical tubule.

To study how organic anions were transported from the basolateral surface to the apical surface before secretion into the lumen, Miller et al. (111) used confocal fluorescence microscopy to examine the role of MT-dependent vesicle transport in the movement of fluorescein across teleost renal proximal tubules. After incubation with fluorescein, fluorescence microscopy revealed a diffuse cytoplasmic pattern with additional staining localized to discrete vesicles ~2-3 µm in diameter. The dynamics of these vesicle movements were studied; fluorescein-labeled vesicles normally moved at 0.83 µm/min throughout the cell. Exposure to nocodazole significantly reduced the amount of punctate fluorescence in these cells and blocked the movement of fluorescein-containing vesicles to the apical surface. Also, a decreased recovery of fluorescein and the organic anion p-aminohippurate at the apical lumen was measured. This study reveals that organic anion transport to the apical plasma membrane has a MT-dependent component. Moreover, the reduced punctate fluorescence after nocodazole treatment suggests the existence of a nocodazole-sensitive step involved in organic anion compartmentation. Subsequent studies in the same system correlated the nocodazole-induced reduction in fluorescein secretion with a loss of MT structure (113).

Microtubule polarity in renal epithelia has been defined through studies of the MDCK cell line (11). These epithelia contain MT with their minus ends organized at the apical plasma membrane; traffic to the apical surface from the dense apical tubules is expected to utilize a cytoplasmic dyneinlike motor protein, if this step is MT dependent. Likewise, the nocodazole-sensitive transcytosis of organic anions from the basolateral surface to the apical surface should also utilize a cytoplasmic dyneinlike motor protein.

Although studies of the regulation of many apical proteins, channels, and apical secretion in renal epithelia support a major role for MT-dependent vesicle transport, fewer studies have indicated a role for MT in regulation of the basolateral transporter or channel content of the plasma membrane. However, an elegant study by Brown et al. (21) has examined the distribution of H⁺-ATPase in proximal tubule and in the intercalated cells of the collecting duct. The H⁺-ATPase normally exhibits localization to basolateral membranes in proximal tubule, to apical membranes in the A-intercalated cells, and a punctate vesicular distribution near basolateral and apical surfaces in the B-intercalated cells. The distributions of H⁺-ATPase in the plasma membranes of A and B cells are modulated by changes in acid-base conditions and, in turn, govern the acidification of the urine. Surprisingly, colchicine treatment of rats resulted in recovery of H⁺-ATPase in a punctate pattern throughout the cytoplasm in both proximal tubule (normally basolateral) and in the A-intercalated cells (normally apical). Also, brush-border membrane vesicles and basolateral membrane vesicles purified from colchicine-treated rats revealed a significant reduction in H⁺-ATPase activity, verifying the results obtained by fluorescence microscopy. The findings in A-intercalated cells are consistent with previous work showing that MT are essential in maintaining channels and pumps in the apical membrane. Because the normal basolateral distribution of H⁺-ATPase in the proximal tubule is also altered by colchicine, these findings suggest that the role of MT in membrane protein targeting is not restricted to the apical membrane in specialized populations of kidney cells.