- Jan 20, 2003
- Views: 68
- Page(s): 9
- Size: 1.44 MB
1 TRANSACTIONS OF THE AMERICAN CLINICAL AND CLIMATOLOGICAL ASSOCIATION, VOL. 107, 1995 A NEW VOYAGE OF DISCOVERY: TRANSPORT THROUGH THE HEPATOCYTE JOHN L. GOLLAN and by invitation STEPHEN D. ZUCKER BOSTON, MASSACHUSEITS One of the vital functions of the liver is the biotransformation and detoxification of small hydrophobic molecules into more hydrophilic species (e.g., sugar, amino acid or sulfate conjugates), which ultimately are excreted into bile or urine. The mechanisms whereby the hepato- cyte efficiently transports the vast array of endogenous and xenobiotic compounds and drugs from the sinusoidal plasma membranes to the intracellular sites of metabolism are poorly defined. In fact, the rapid trafficking of substrates in the liver cell may involve the integrated function of a number of different mechanisms and cellular components. Potentially, the action of cytosolic and membrane proteins, organellar membranes, vesicular transport pathways, the cytoskeleton and cyto- plasmic flow may be involved to a varying extent. The likely complexity of intracellular transport is no wonder, if one considers the intricate three-dimensional structure of the hepatocyte, which is composed of a dense dynamic meshwork of membranous barriers suspended in a viscous cytosolic soup (Figure 1). Utilizing high-resolution technology, we have undertaken a series of investigations to define the potential role of intracellular membranes and cytosolic binding proteins in the transport of bilirubin in the hepatocyte. We postulated that intracel- lular membranes may play a pivotal role in the binding, transport and metabolism of small hydrophobic molecules. Bilirubin, the orange-colored end product of heme degradation, is a commonly utilized model organic anion, since it is produced endog- enously and has well-characterized physicochemical and physiological properties. Unconjugated bilirubin has an aqueous solubility of less than 100 nM at neutral pH. Circulating bilirubin, bound to albumin, dissociates from this protein and is efficiently taken up and trans- ported into the hepatocyte via an organic anion transport protein (OATP) in the plasma membrane. It then moves to the endoplasmic reticulum and is conjugated with glucuronic acid by the enzyme UDP- From the Division of Gastroenterology, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts. Address to which requests for reprints should be sent: John L. Gollan, M.D., Ph.D., Director, Gastroenterology Division, Brigham and Women's Hospital, 75 Francis Street, Boston, MA 02115. 48
2 INTRACELLULAR TRANSPORT IN THE HEPATOCYTE 49 .^ S. ..4 or d .t a - P. *;j i- W'v :\ -% Per :I . .'r" V: PC' VI' ; . .4E r. FIG. 1. Intracellular Ultrastructure of a Normal Hepatocyte. The complex meshwork of organellar membranes is readily apparent (x36,000). The electron micrograph illustrates the following structures: endoplasmic reticulum (rer), Golgi apparatus (G), mitochondria (m), peroxisomes (p), primary (plys) and secondary (*) lysosomes, lipid droplets (L), and glycogen rosettes (gly). With permission from MJ Phillips, S Poucell, J Patterson, P Valencia. The Liver: An Atlas and Text of Ultrastructure Pathology. New York, Raven Press, 1987;16.
3 50 JOHN L. GOLLAN glucuronosyltransferase. The resultant hydrophilic bilirubin glucu- ronides are excreted into the bile canalicular lumen via the multispe- cific organic anion transporter (cMOAT) in the canalicular membrane. It has long been presumed that the cytosolic protein, glutathione S-transferase B (GST-B) or ligandin, serves as the principal trans- porter of bilirubin from the plasma membrane to the endoplasmic reticulum in the hepatocyte. However, some years ago it was demon- strated that at neutral pH, unconjugated bilirubin partitions into phospholipid membranes, rather than the bulk aqueous phase. Indeed, it was estimated that almost one-half of the unconjugated bilirubin in the hepatocyte is membrane-bound (1). Moreover, prior experimental evidence from our laboratory has demonstrated that bilirubin glucu- ronidation occurs more efficiently when bilirubin is presented to he- patic microsomal UDP-glucuronosyltransferase associated with unila- mellar phospholipid vesicles (liposomes), rather than bound to cytosol or purified binding proteins (2,3). Since the network of membranes, tubules, vesicles and lamellae which constitute the endoplasmic retic- ulum occupies -15% oftotal hepatocyte volume (4), we thus postulated that intracellular membranes play a significant role in the transport of small hydrophobic compounds, such as bilirubin. Hence, we performed a series of investigations designed to examine the relative contribu- tions of cytosolic binding proteins and cellular membranes in the intracellular transport of bilirubin in the hepatocyte. Small unilamellar vesicles of varying lipid composition were pre- pared by sonication. Large unilamellar vesicles of graded size were formed by aqueous injection of an ethanolic lipid solution, with vesicle diameter and homogeneity assessed by quasielastic light scattering. Rat liver microsomes, basolateral plasma membranes, canalicular plasma membranes, and ligandin (YaYc fraction) were isolated, and the purity confirmed by enzymatic assay, Western blot analysis, and SDS- PAGE. Using stopped-flow fluorescence techniques and curve-fitting analysis, kinetic and thermodynamic parameters for the transfer of bilirubin between various model and native hepatocyte membrane populations were determined, both in the presence and absence of ligandin. These data were then applied to define a physiological model of bilirubin transport. In our initial studies (5), bilirubin was added to donor vesicles labeled with the fluorescent phospholipid probe, dansyl-PE. When bound to the donor vesicles, bilirubin quenches the dansyl probe fluo- rescence, and hence the movement of bilirubin to the unlabeled accep- tor vesicles can be monitored directly by the reemergence of fluores- cence over time. Using a high-resolution, stopped-flow device, the
4 INTRACELLULAR TRANSPORT IN THE HEPATOCYTE 51 transfer from donor to acceptor vesicles was shown to be a single- exponential function, with a mean half-time of 2.0 msec at 370C. No increase in bilirubin transfer rate was observed with increasing accep- tor vesicle concentration, indicating that membrane-to-membrane transfer of bilirubin occurs via aqueous diffusion, rather than mem- brane collisions. In addition, the activation energy for bilirubin disso- ciation from membrane vesicles was low, suggesting that bilirubin is associated with phospholipid bilayers at the membrane-water inter- face (despite its marked hydrophobicity), thereby resulting in the rapid rate of dissociation. Thus, these data suggest that the intracellular transport of bilirubin to its site of glucuronidation in the endoplasmic reticulum involves rapid, spontaneous intermembrane transfer through the aqueous phase (5). The rate of bilirubin dissociation from model phospholipid vesicles declined asymptotically with increasing donor vesicle diameter, due primarily to a decrease in the entropy of activation for the larger vesicles (6). The incorporation of phosphatidylethanolamine and phos- phatidylserine, at physiologic concentrations, significantly enhanced the dissociation of bilirubin from phosphatidylcholine vesicles. Choles- terol induced a biphasic effect on the transfer rate constant. A decrease in the bilirubin transfer rate from 248 to 217 sect1 was observed with increasing cholesterol:phospholipid ratios up to 20 mol%. This was followed by a dramatic rise in rate to 312 sect1 as the cholesterol concentration increased beyond this level, to a maximum of 70 mol%. As predicted by our analysis of the data obtained using model phos- pholipid vesicles, the bilirubin dissociation rate from native rat liver endoplasmic reticulum (9.1 sect1) was significantly slower than for both basolateral and canalicular plasma membranes, which exhibited rate constants of 11.7 and 25.8 sect1, respectively. The transfer of bilirubin from ligandin to membranes was described by a single exponential, and was -25-fold slower than that observed for intermembrane transfer (Figure 2). Arrhenius plots of the rate constants demonstrated a 3-fold greater activation energy for the dis- sociation of bilirubin from ligandin compared with phospholipid vesi- cles. The transfer of bilirubin from ligandin to acceptor vesicles oc- curred at a progressively slower rate as the acceptor concentration was increased, asymptotically approaching a mean rate of 8.1 sect (7). Similarly, the rate of bilirubin transfer from model donor vesicles to ligandin was dependent on the donor:acceptor molar ratio, with in- creasing ligandin concentration. We next studied the transfer of bilirubin from isolated plasma mem- branes (250 Ag protein/ml) to fluorescent-labeled acceptor vesicles (500
5 52 JOHN L. GOLLAN 1.0 PC Vesicles 0.8 11 ~0.6 0) U.. 0.2 ferase B (GSTL~~ST or (igandin).Fursecinestwamesrdoraprodf02 c 0.0 0.1 0.2 0.3 0.4 0.5 Time (sec) FIG. 2. Bilirubin Dissociation from Membrane Vesicles versus Glutathione S-trans- ferase B (GST or Ligandin). Fluorescence intensity was measured over a period of 0.2 sec and 0.5 sec for phosphatidyicholine (PC) vesicles and GST, respectively, using a stopped- flow device. The time-dependent reemergence of fluorescence is due to the transfer of bilirubin from the donor PC vesicles or GST to acceptor vesicles. The transfer rate from GST was -25-fold slower than that for intermembrane transfer. gM) in the presence of ligandin, over a range of molar ratios simulating those present in the hepatocyte. A progressive decline in the mean membrane transfer rate was observed, from 6.2 sec1 in the absence of ligandin to 1.3 sec 1 in the presence of 3 ,tM ligandin, indicating that ligandin slows the rate of intermembrane bilirubin transfer. Again, these data are consistent with a diffusional mechanism of intracellular bilirubin transport. Collectively, these findings indicate that bilirubin transfer between membrane vesicles occurs via aqueous diffusion, by a series of inter- membrane "jumps", and that the cholesterol:phospholipid ratio repre- sents the principal determinant of the bilirubin dissociation rate. Moreover, bilirubin dissociation from the cytosolic protein ligandin also occurs via a diffusional, rather than a collisional mechanism, and is considerably slower than from plasma membrane vesicles (Figure 3). Hence, no direct membrane-protein interaction occurs in the process of ligandin-mediated bilirubin transfer. This type of kinetic behavior is inconsistent with that described for other intracellular transport pro- teins, a conclusion which is supported by the observed decline in the bilirubin intermembrane transfer rate with increasing ligandin concentration.
6 INTRACELLULAR TRANSPORT IN THE HEPATOCYTE 53 Collisional Model Sinusoidal Membrane Endoplasmic Reticulum FIG. 3. Kinetic Models of Glutathione S-transferase-Mediated Bilirubin Transport in the Hepatocyte. This schematic diagram depicts models for both collision-mediated and diffusion-mediated transfer of bilirubin (B) between basolateral plasma membranes and acceptor vesicles, in the presence of glutathione S-transferase (GST). The collisional model (upper panel) assumes that bilirubin transfer occurs via direct interactions (collisions) between GST and the membranes. Hence, GST functions as a bilirubin "shuttle", and the transfer rate is predicted to increase proportionately with the protein concentration. The diffusional model (lower panel) postulates that the interaction of bilirubin with GST occurs only following bilirubin dissociation from the membrane. This model predicts an asymptotic decline in the rate of intermembrane bilirubin transfer as the concentration of GST rises, as was observed experimentally. We postulate that the inherent cellular membrane-cholesterol gra- dient in the hepatocyte creates a directed flux of bilirubin from the plasma membrane to the endoplasmic reticulum, and represents a key driving force for intracellular bilirubin transport. Thus, the mem- branes of the endoplasmic reticulum likely function as an enormous bilirubin sink, with the endogenous flux of bilirubin directed from the periphery (i.e. plasma membrane) of the hepatocyte to the endoplasmic reticulum. While the role of cytosolic proteins in the intracellular transport of small hydrophobic compounds remains speculative, our data suggest that regulation of ligandin (glutathione S-transferase B) concentrations may represent a mechanism whereby the hepatocyte is able to modulate the rate of ligand presentation to the endoplasmic reticulum for metabolism. We further hypothesize that ligandin and potentially other cytosolic binding proteins may function as an intra- cellular reservoir which facilitates hepatocellular uptake and protects
7 54 JOHN L. GOLLAN organellar membranes from high concentrations of hydrophobic ligands. SUMMARY In summary, hepatocellular membranes likely play an essential role in the binding and directed trafficking of unconjugated bilirubin, and potentially of a variety of other small hydrophobic molecules. Target- ing of these substrates to the endoplasmic reticulum is determined by membrane cholesterol content, surface area and integral protein bind- ing and enzyme activity. The rate of intracellular transport potentially may be modulated by the concentration of cytosolic binding proteins, but, at least for ligandin, this protein does not appear to function primarily as an intracellular bilirubin transporter. ACKNOWLEDGMENTS This work was supported by NIH Research Grants DK-36887, DK-02047, DK-43955, DK-34854 and a Harvard Digestive Diseases Center Pilot/Feasibility Award (SDZ). The authors gratefully acknowledge the technical contributions of Wolfram Goessling, Julianne Narciso and Emma Bootle. REFERENCES 1. Tipping R, Ketterer B, Christodoulides L. Interactions of small molecules with phos- pholipid bilayers: Binding to egg phosphatidylcholine of some organic anions (bro- mosulphophthalein, oestrone sulphate, haem and biliburin) that bind to ligandin and aminoazo-dye-binding protein A. Biochem J 1979;180:327-337. 2. WVhitmer DI, Ziurys JC, Gollan JL. Hepatic microsomal glucuronidation of bilirubin in unilamellar liposomal membranes: Implications for intracellular transport of lipophilic substrates. J Biol Chem 1984;259:11969-11975. 3. Whitmer DI, Russell PE, Ziurys JC, Gollan JL. Hepatic microsomal glucuronidation of bilirubin is modulated by the lipid microenvironment of membrane-bound sub- strate. J Biol Chem 1986;261:7170-7177. 4. DePierre JW, Andersson G, Dalner G. Endoplasmic reticulum and Golgi complex. In: The Liver: Biology and Pathology, Ed. 2, edited by Arias IM, Popper H, Shafritz DA, Jakoby WB, Schachter D. New York, Raven Press; 1988. pp 165-188. 5. Zucker SD, Storch J, Zeidel ML, Gollan JL. Mechanism of spontaneous transfer of unconjugated bilirubin between small unilamellar phosphatidylcholine vesicles. Biochemistry 1992;31:3184-3192. 6. Zucker SD, Goessling W, Zeidel ML, Gollan JL. Membrane lipid composition and vesicle size modulate bilirubin intermembrane transfer: Evidence for membrane- directed trafficking of bilirubin in the hepatocyte. J Biol Chem 1994;269:19262- 19270. 7. Zucker SD, Goessling W, Ransil BJ, Gollan JL. Influence of glutathione S-trans- ferase B (ligandin) on intermembrane transfer of bilirubin. Implications for the intracellular transport of nonsubstrate ligands in hepatocytes. J Clin Invest 1995;96:1927-1935.
8 INTRACELLULAR TRANSPORT IN THE HEPATOCYTE 55 DISCUSSION George Schreiner, Washington, D.C.: I refer to your experiments with albumin as the carrier of ligands. There are multiple sites on the albumin molecule for lipid particles, which may be half full or empty. I like to think of it as a series of trucks that are either big trucks or little trucks. Because it is possible to get defatted albumin, have you ever had a chance to see whether the albumin-carrying capacity or transport capacity changes with the lipid load that is bound to the albumin? Gollan: I think that is a very good question. Obviously, competition for binding sites on albumin is a potential issue, depending on the extent to which fatty acids and other molecules share binding sites. The albumin we used was, in fact, predominantly defat- ted. It was as defatted as one can get by the usual kind of procedures, including the use of activated charcoal. What we did show was that when you increased the concentration of substrate, more and more was bound by HDL, and by such lipid particles, rather than by the binding protein. Certainly for bilirubin, I would imagine that in a jaundiced patient with a serum bilirubin concentration at about 10 mg. %, that as much as 50% may be bound to HDL. Now I don't want to imply that albumin does not play a role in plasma transport. I think that it is very important, and I believe that the binding proteins probably serve an extremely important function. If I could kind of just let the hypothesis roll with only limited data to support the comment, what I think they do is this: The binding proteins act as a kind of sink, but this sink is very critical in terms of potentially regulating the rate at which molecules get to the plasma or intracellular membranes. Once they are on the membranes, it is all action and presumably these small molecules bounce around rapidly until they actually make a hit on a catalytic site of an enzyme, where they undergo biotransformation. I think that the soluble binding proteins are extremely important and are probably going to be one of the important regulatory components of this whole system. Arnold Weissler, Rochester, MN: We often see patients with atherosclerotic disease who do not have an abnormal HDL content. We know that other mechanisms may be promoting the atherosclerosis. We have heard earlier today about homocystine. Is it possible that the transport mechanism involving HDL can be altered in disease and that this would account for the abnormalities in atherosclerosis beyond the abnormality related to its content in the blood? Gollan: I compliment you on your comment. In fact, Dr. Steve Zucker, who has been an NIH physician scientist recipient during these experiments, has just submitted a grant application which concerns exactly this topic: potentially whether these small hydrophobic molecules that may bind to HDL in the circulation may modify its behavior. I think this is very pertinent. There is, in fact, some quite indirect evidence suggesting that bilirubin levels may actually protect in terms of atherogenesis. Moreover, if biliru- bin acts as an antioxidant, which is suggested in the literature, you would imagine that people with Gilbert's syndrome (who exhibit elevated serum unconjugated bilirubin levels), if there were such protection, would live forever; but it doesn't seem to be so from the literature. Gray, Stanford: John, I was wondering about the movement of smaller ligands that are polar, rather than being attracted to membranes they would be rejected by them. How are they going to move around in the liver? Gollan: Some of the preliminary data we have obtained show that the glucuronides, BMG and BDG, which are the monoglucuronides and the diglucuronide of bilirubin, actually bound to a higher extent than the hydrophobic unconjugated bilirubin mem- branes. It doesn't make a whole lot of intuitive sense; however, it is going to depend very much on the biophysical properties of the ligand, of the molecule that we are looking at.
9 56 JOHN L. GOLLAN It is going to relate to size, charge, and hydrophobicity: how lipid-soluble it is. It is quite possible that the determinants of transport are going to be a mixture of these things. It also may be just transported by rapid diffusion of the molecule itself within the aqueous, since this is water-soluble now and may not need a protein to bind it. Whereas, if you have a big protein glob, one of our original assumptions was that these binding proteins are relatively slow to move 3-dimensionally compared to a single small molecule; so it is quite possible that small polar molecules are actually just whipping through the aqueous.Load More