Perspective: Five decades of cholestasis research and the brave new world


  • Irwin M. Arias

    Corresponding author
    1. National Institutes of Health, National Institute of Child Health and Human Development, Bethesda, MD
    • National Institutes of Health, NICHD, 31 Center Drive, Bethesda, MD 20892
    Search for more papers by this author
    • fax: 301-402-0078

  • Potential conflict of interest: Nothing to report.

Although honored to be invited to prepare a “Perspective” on cholestasis, a topic that has occupied my research for 50 years, I cannot live up to the true meaning of the word, which is derived from the Latin “perspicere,” meaning “to see clearly.” If anything, the quest to understand has led to increasing complexity and challenge … which summarizes the theme of this report.


ABC, adenosine triphosphate–binding cassette; AMPK, adenosine monophosphate–activated protein kinase; ATP, adenosine triphosphate; NIH, National Institutes of Health; MT, microtubules; MDR, multidrug-resistant; NR, nuclear receptor; Pgp, permeability glycoprotein; RNAi, RNA interference.

Having been told that this manuscript does not have to adhere to all of the rigors of a HEPATOLOGY original manuscript, an increasingly challenging process since my days as Founding Editor of the journal, I will begin, as requested, by relating the origin of my scientific orientation, changing concepts in understanding cholestasis, two major paradigm shifts, and conclude by expressing views as to where the quest to understand basic mechanisms and potential therapies of cholestasis appears to be heading.

Since high school I have been interested in science, and I worked in a unique International Business Machines after-school New York City laboratory that, in its one year of existence, spawned three Nobel Laureates (Blumberg, Lederberg, and Glauber) and many other leading biological scientists and engineers. Medical school followed college and a year of graduate study in protein chemistry. My interest in liver disease research was cultivated by being a student, house officer, and fellow at the Boston City Hospital, where I was mentored by William Castle, Max Finland, Charles Davidson, and Ed Kass. In 1957, I joined Irving London's Department of Medicine at the newly created Albert Einstein College of Medicine and remained in that exciting environment until 1983, when I became Chairman of the Department of Physiology at Tufts. In 2001, I joined the National Institutes of Health (NIH) as Senior Scientist/NICHD and Assistant to the Director, Intramural Program/NIH. Einstein provided an incredible array of mentors, including Alex Novikoff, Harry Eagle, Irving London, and Sam Seifter, who encouraged me to enter the rapidly emerging field of biochemistry as applied to physiology and medicine. This was the “golden age” for becoming a physician–scientist. Believe it or not, what was rate limiting was not money but ideas.

Success with initial experiments on bilirubin conjugation defined the biochemical basis of Crigler-Najjar and Gilbert's disease. Because defects in bilirubin metabolism result in a readily detected phenotype, search for families with inheritable jaundice was productive and exciting, and it introduced colleagues who were outstanding pediatricians, geneticists, and biochemists, particularly Geoffrey Dutton, the discoverer of uridine 5′ diphosphoglucuronic acid, the substrate for bilirubin and other glucuronyl transferases. Advances in biochemistry and cell biology facilitated study of bile pigment uptake and conjugation and secretion by the liver, which proved to be unique processes having greater clinical importance regarding drugs and endogenous substrates than bile pigment. Bilirubin had to be taken up into the hepatocytes and conjugated to be secreted in bile. The secretory step proved rate limiting, which was a prelude to subsequent interest in cholestasis. Under the tutelage of experts, my fellows and I entered into the realms of drug metabolism, kernicterus, hormonal regulation, phylogeny, ontogeny, and cell biology. These “bridge-building” experiences have been critical in my career and continue to bring exciting new colleagues and dimensions to our research. The Einstein Liver Research Center was formed in 1974 and remains dedicated to bridging basic science and study of liver structure, function, and disease. By 1977, more than 30 established investigators in nine departments collaborated to link cellular, molecular, and immunobiology into the rapidly growing field of liver research. This rich environment continuously stimulates my research, which has also been immeasurably aided by a disappearing fringe benefit of academic medicine … the sabbatical. At approximately seven-year intervals, I spent productive years with Bob Schimke at Stanford studying protein turnover; Nate Kaplan, Jack Kyte, and Gordon Sato at the University of California San Diego, working in protein chemistry and cell biology; Bob Adelstein at NIH, learning about cytoskeletal biology; Guido Guidotti at Harvard, studying membrane and molecular biology, and Jennifer Lippincott-Schwartz at NIH, learning cell biology and quantitative live cell imaging. Each sabbatical resulted in new concepts and methods for our research.

Serendipity has also played an important role. For example, in 1948, I was frustrated in medical school and considered returning to graduate studies. During a train ride to Boston, a chance conversation with John Enders, who won the Nobel Prize for polio virus research a few years later, advised me to change schools and get involved in research, which changed my career. In 1987, fortuitous attendance at a seminar by Ira Pastan on Pgp (Permeability glycoprotein), the first multidrug resistance protein to be identified in cancer cells, prompted a major advance in hepatobiliary physiology, which we and many others have been exploring ever since.

After the seminar, I inquired as to whether Pgp is present in normal cells. No one knew. After acquiring normal human tissues (without maiming anyone), we discovered that Pgp (now called MDR1 or ABCB1) is expressed in the apical domain of hepatocytes, intestinal cells, proximal tubular cells, and others. Using rat liver canalicular membranes (techniques developed by Inoue and Kinne at Einstein), Kamimoto discovered that adenosine triphosphate (ATP) hydrolysis drives transport of organic cationic anticancer drugs out of the canalicular vesicles.1 At this time, we and others believed (and published) that bile acid secretion was driven by the electrochemical gradient (−35 mV) generated by the Na+/K+ adenosine triphosphatase. In retrospect, all investigators should have realized that −35 mV could not generate the large concentration gradients for biliary components between hepatocyte and bile. However, in those days ATP hydrolysis in mammalian membranes was believed to be restricted to transport of ions and not organic substrates. Had we been microbiologists, the problem would have been solved at least 10 years earlier, because microbes were long known to couple ATP hydrolysis with membrane transport of organic substrates.

Shortly after Kamimoto's experiments, we pondered whether bile acids and other organic anions could be transported across the canalicular membrane by ATP-dependent processes. We were not alone in realizing the biologic implications of the canalicular Pgp studies, and shortly thereafter, several laboratories, including ours, demonstrated ATP-dependent canalicular transport of bile acids and a separate mechanism for ATP-dependent transport of organic anions, including bilirubin glucuronide.2–4 A new paradigm of hepatobiliary secretion emerged. Within a few years, the genes responsible for these individual processes were identified as members of the large ATP-binding cassette (ABC) superfamily of membrane transporters.5 Pgp (ABCB1) transports organic cations, bile salt export pump (ABCB11) transports bile acids, ABCB4 “flips” phosphatidylcholine, MRP2/cMOAT (ABCC2) transports organic anions, and the heterodimer ABCG5/ABCG8, which is defective in sitosterolemia, is postulated to transport sterols. The field rapidly expanded to include basic and clinical studies of mechanisms, regulation, and genetics.

The next major advance was identification of molecular defects in specific canalicular ABC transporters as the basis for inheritable liver diseases, mainly cholestatic, which had previously been clinically characterized. Mendelian recessive point mutations resulting in stop codons were identified for 3 “canalicular” diseases followed by generation of single and multiple knockout mice, thereby permitting further studies. Discovery of the inheritable defects came through different routes. Pgp (MDR1) was first identified by a multidrug-resistant phenotype in cancer cells.6 However, in rodents, three MDR genes were identified. MDR2 (ABCB4) had high sequence identity with MDR1 (ABCB1), but did not confer drug resistance when expressed in cell lines. Peptide antibody studies showed that ABCB4 is the most abundant MDR-type protein in the bile canaliculus. Identification of its natural substrate was due to serendipidity. Investigators in the Netherlands Cancer Center generated mice in which MDR2 was removed by homologous recombination. No cancer phenotype was observed; however, some homozygous deficient mice became jaundiced and manifested small bile duct damage followed by more extensive liver cell damage. Oude Elferink performed thin-layer chromatography of bile from mutant mice and observed virtual absence of phosphatidylcholine.7 Subsequent studies confirmed that MDR2 uses ATP to transfer (or “flip”) phosphatidylcholine from the inner to the outer leaflet of the canalicular membrane, from which it is probably removed by bile acids and forms mixed micelles.8 Small bile duct damage likely results from the detergent action of bile acids on the small bile duct epithelium. Thus, a protective function for biliary phosphatidylcholine was discovered.

Shortly thereafter, Deeley and Cole identified an MDR-like protein (MRP1/ABCC1) in the plasma membrane of a lung cancer.9 Cloning showed that it is a member of a large family including MRP 1-9 (ABCC 1-9), the functions of which have been delineated, except for MRP6 (ABCC6).5 ABCC2 transfers organic anions including eicosanoids, bilirubin glucuronide, and other sulfate, glucuronide, and glutathione conjugates from the liver cell into bile.10, 11 ABCC2 is mutated in the Dubin-Johnson syndrome in humans and rats and also occurs in Corriedale sheep and Golden Lion Tamarind monkeys, which are characterized by plasma retention of these organic anionic substrates, but not bile acids, which are secreted by a different transporter. The basolateral transporters ABCC3 and ABCC4 are induced when biliary secretion is impaired and facilitate transfer of bile acid and bilirubin anionic conjugates from the hepatocyte into plasma, thereby providing a partial, albeit inadequate, hepatic defense mechanism in bile secretory failure.12 ABCC6 presents a different challenge because major expression is in the hepatocyte basolateral membrane and the protein is mutated in pseudoxanthoma elasticum, a disease characterized by widespread defects in elastic tissue.13 What the liver secretes into blood, which may regulate elastase and possibly other serine proteases, is not known.

Hans Popper, a mentor and friend, coined the term “cholestasis” based on the pathological finding of bile stasis in dilated canaliculi and small bile ducts. From Virchow in the 19th century to the 1970s, mechanical explanations for intrahepatic cholestasis were proposed for “cholestatic” drugs, viruses, and toxins. Canalicular bile plugs were originally proposed to be the cause of cholestasis; however, as Gerald Klatskin presciently noted, the canalicular network is like a sponge and you cannot obstruct a sponge. Slowly the concept emerged that cholestasis is a disorder of bile secretion by the hepatocyte.

In 1970, Simon did the first studies seeking defects in canalicular membrane composition and protein turnover in experimental cholestasis.14 Looking back on those studies, I realize that we were on the right track but, lacking awareness of canalicular transporters, we were searching in the dark. The critical advance was expression cloning and functional analysis by Gerloff and colleagues of the bile salt excretory protein (later renamed ABCB11).15 ABCB11 mutations resulted in progressive familial cholestasis, type 2, thereby confirming that ABCB11 is the major, if not only, physiologically important canalicular bile acid transporter.16

Based on the concept that bile acid retention is the cause of hepatocellular injury in cholestatic disease, we focused on the biology of ABCB11. The questions that emerged involved molecular and cell biology and concerned how ABCB11 is transcribed, regulated, trafficked, targeted, inserted into the plasma membrane, and degraded. We also pondered whether defects in these processes are responsible for acquired cholestasis associated with viruses, drugs, toxins, hypoxia, pregnancy, and hyperalimentation. What determines why certain ABC transporters ultimately reside in the canalicular membrane, whereas others are restricted to the basolateral plasma membrane? What is the relationship between these processes and cellular polarity? These are major cell biological questions, which we have been exploring for several years using ABCB11 as a probe.

In the 1980s, cell biologists prepared antibodies against canalicular proteins without much concern as to their function but as markers to study trafficking mechanisms. Fortuitously, each laboratory generated antibodies against the same potent antigen, cCAM105. Using these antibodies, they observed that canalicular membrane proteins traffic from the Golgi to the basolateral plasma membrane, from which they undergo vesicular transcytosis to the canalicular membrane.17 This somewhat circuitous route was proposed to be the mechanism whereby all canalicular proteins are targeted. However, the earlier studies were performed before ABC transporters were identified. Using antibodies against ABCB11, ABCB1, and ABCC2, we noted no reaction with basolateral membranes and proposed that canalicular ABC transporters may use a direct trafficking route from Golgi to the apical membrane. Subsequent pulse-chase studies in rats by Kipp revealed that, in contrast to single transmembrane and glycosyl inositol phospholipid–linked canalicular proteins, the ABC transporters traffic directly from Golgi to the canalicular membrane.18 In addition, Kipp discovered that some, and possibly all, ABC transporters are present in a large intracellular pool(s) from which they can be directed to the apical membrane by agonists, such as taurocholate or cyclic adenosine monophosphate.19 Each of these physiological stimuli selectively increases the amount of ABC transporters in the canalicular membrane independently of protein synthesis. When both agonists are combined, the amount of canalicular ABCB11 increased 5-fold to 6-fold over basal levels, indicating recruitment from intracellular pool(s). Subcellular fractionation identified the intracellular pool(s) as being enriched in several small rab guanosine triphosphatases, particularly rab 11a, which is a marker for the recycling endosome system.

Precise identification and visualization of the intracellular pool(s) and mechanisms for transfer of ABCB11 to the canalicular membrane required use of a cell line because primary cultures of rat hepatocytes could not be maintained for sufficient time for biochemical or imaging studies. The WIF-B9 cell line, developed by Cassio and further characterized by Hubbard, proved useful.20 These polarized cells are a hybrid between a rat hepatoma and human fibroblasts, form a bile canalicular sphere into which secretion of fluorescent substances can be quantified, and are ideal for live cell imaging. Because WIF-B9 cells are difficult to transfect by conventional methods, Wakabayashi infected them with a green fluorescing protein–ABCB11 adenoviral construct and performed biochemical and live cell imaging studies.21 In the steady state, most ABCB11 was in a large intracellular endosomal pool identified by rab 11a and colocalized with a unconventional motor protein, myosin Vb. Live cell fluorescence recovery after photobleaching and fluorescence loss after photobleaching studies demonstrated that ABCB11-containing vesicles are transferred along microtubules from the Golgi to the canalicular membrane from which they cycle between a large rab 11a-recycling endosome pool and the canalicular membrane. Within the apical plasma membrane, diffusion of the transporter is limited by the tight junctional ring around the canaliculus. These studies set the stage for seeking chaperones, cofactors, coactivators, and specific sites in the pathway at which dominant negative or mutated constructs could be tested.22 Unfortunately, WIF-B9 cells lack the ability to respond to taurocholate or cyclic adenosine monophosphate as secretory agonists observed in rat and human liver. This regulation is now being explored in rat and human hepatocytes, which, because of technological advances,23, 24 can be maintained for 2 to 6 weeks with no appreciable loss in discernible structure, function, or gene expression.

These studies began to elucidate complex intracellular trafficking pathways and their regulation,22 a pursuit that will continue for several years and is accelerated by advances in imaging methods, proteomics, and introduction of inhibitory and small RNAs and genome-wide RNA interference (RNAi) viral vectors. The experiments support the hypothesis that acquired cholestasis results from an intracellular traffic jam, which ultimately impairs ABCB11 synthesis, packaging, trafficking, or degradation. The concept of intracellular traffic jams is “in”. For example, Kai Simons25 recently proposed a traffic jam concept in studies of raft-associated transport in relation to physiology and disease.

Now the problem is to identify the highways and shortcuts, motors, and regulators, as well as transcriptional regulation. The concept is analogous to a freight train, which requires engineers, engine, tracks, conductors, stations, and so forth. Vesicles containing ABC transporters, particularly ABCB11, move along microtubular tracks throughout the cell but eventually only fuse with the apical plasma membrane, indicating that the attachment mechanism is downstream and not at the Golgi or endoplasmic reticulum.22 Possible sequence signals in ABCB11 and other ABC transporters have been explored, such as PDZ domains, N-glycosylation, and others; however, there is no general theorem here. PDZ domains are essential for ABCC2 and cystic fibrosis transmembrane conductance regulator (ABCC7) targeting but not for other hepatocyte ABC transporters. N-glycosylation of ABCB11 requires at least two of four possible sites for apical targeting. Soluble N-ethylmaleimide-sensitive fusion protein attachment receptor molecules, particularly syntactin 3, are putative receptors in the canalicular membrane. Specificity and mechanism of vesicle attachment and cargo delivery are challenging cell biological problems under investigation.

The search for proteins that bind to ABCB11 used several methods, including yeast 2 hybrid screens by which Ortiz identified Hax-1 and non-muscle myosin 2 light chain 2a as binding proteins and demonstrated the effect on ABCB11 trafficking in Madin-Darby canine kidney cells.26, 27 Hax-1 links the rab 11a associated recycling endosome to the actin mechanism. Blocking Hax-1 and actin components (in other words, clathrin, eps11, cortactin, and others) increased the amount of ABCB11 in the apical membrane, indicating that Hax-1, clathrin, and the actin cytoskeleton participate in the endocytic portion of the cycling pathway. Myosin 2 light chain 2a is required for the exocytic portion of the cycling pathway; inhibition of myosin light chain kinase reduced the amount of ABCB11 in the apical membrane as well as transport activity. Figure 1 illustrates currently identified proteins required for ABCB11 secretory and cycling pathways. Evidence for their involvement is the demonstration of inhibited ABCB11-mediated taurocholate secretion (bile secretory failure) after the addition of chemical inhibitors or expression of dominant negative or RNAi constructs. Many other proteins and lipids undoubtedly participate in the cycling system. Identification of these partners requires proteomic and lipidemic analysis of pure recycling endosomes, a difficult task not yet achieved.

Figure 1.

Intracellular trafficking of ABCB11 (and other ABC transporters) in mammalian hepatocyte. Factors identified, thus far, as being required for specific steps in this process are indicated. Abbreviations: TGN, trans-Golgi; ER, endoplasmic reticulum; MLC, myosin light chain.

Transporter proteins are transferred to the Golgi for posttranslational modification and proper folding, which permits their budding from the trans-Golgi into complex interactions between the cytoskeleton, targeting processes, and tight junctional components. Several mutated ABC transporters remain within the Golgi because of improper folding and release. Attempts to modify misfolded proteins by small molecules, such as glycerol and 4-phenyl butyrate have yielded encouraging experimental with cells expressing mutated ABCB11 and ABCC7.28 The mechanisms are complex, and whether these effects are clinically beneficial remains to be determined.

Trafficking of endosomes containing ABCBll and other proteins to the canalicular membrane requires a functioning microtubular system, which, in a manner not fully known, intersects with the pericanalicular actin network before endosomal fusion with the canalicular membrane and delivery of cargo. Microtubules (MT) do not directly attach to the plasma membrane. Recent studies show that the plus end of dynamic microtubules attaches to actin and associated proteins (IQGAP, Rac, APC, and EB1), which surround the bile canaliculus. These studies indicate that dynamic MTs are required for canalicular targeting of ABCB11 and possibly other apical membrane proteins, and may be the long sought link between MT and actin-based endosome trafficking systems that surround the bile canaliculus.29

How actin participates in ABCB11 apical trafficking, cycling, and endocytosis is unclear; however, several observations support a role for actin. Rab guanosine triphosphatases interact with the actin cytoskeleton. Cytochalasin and latrunculin A inhibit bile acid and cyclic adenosine monophosphate–mediated increase in bile acid secretion, and ABCB11 endocytosis from the canalicular membrane. The most direct studies implicating actin in targeting and cycling of apical ABC proteins involve mice in which radixin, the dominant ezrin-radixin-moesin protein in liver, was eliminated by targeted mutation. The ezra-radixin-moesin family of proteins cross-links actin filaments and integral membrane proteins. Removal of radixin, which is concentrated in the canalicular domain, resulted in progressive dilatation of the canaliculus, decreased microvilli, jaundice resulting from impaired apical targeting of ABCC2, and subsequent disappearance of other canalicular ABC transporters.30, 31

The canalicular surface area would double every 15 to 30 minutes if not compensated by internalization and degradation of its components.32 ABCB11 in the canalicular membrane is determined by its half-life, internalization, and degradation. In the steady state, its half-life is approximately 3 days.19 The major retrieval mechanism is endocytosis and delivery of membrane components to lysosomes and, possibly, proteosomes, particularly for misfolded or ubiquinated membrane proteins. In support of this view, clathrin-coated profiles, actin-associated proteins, soluble N-ethylmaleimide-sensitive fusion protein attachment receptors, Rabs, annexins, and other components of the endocytic mechanism are enriched in the canalicular domain. Treatment of hepatocytes with PI3-kinase inhibitors results in accumulation of apical plasma membrane proteins in lysosomes.33 Whether the endogenous recycling process for canalicular ABC transporters is entirely clathrin-dependent has not been tested; however, caveolae do not participate. Some ABCB11 mutations, which produce progressive familial intrahepatic cholestasis type 2 and benign recurrent cholestasis, result in misfolded or unstable proteins, which are degraded from either Golgi or the apical domain by entry into the proteosome.34, 35

Another important arena of increasingly understood canalicular physiology involves lipids; however, questions far outweigh answers. In comparison with the endoplasmic reticulum, the canalicular membrane is enriched in sterols, phospholipids, glycolipids, sphingomyelin, and aminophospholipids. How lipid asymmetry is created and maintained within cell organelles and different domains of the plasma membrane in polarized cells is a challenging problem. The rate of lipid exchanges varies with lipid classes. These dynamic events constitutively occur in a plasma membrane, which appears morphologically unchanged.

The “raft” hypothesis of Schuck and Simons36 postulates that specific proteins are targeted to the plasma membrane by virtue of specific associations with sphingomyelin-cholesterol “rafts” that facilitate their delivery to the plasma membrane.36 Canalicular “raft” proteins include single transmembrane ectoenzymes and glycosyl inositol phospholipid–linked proteins. Apical ABC transporters may also be associated with rafts. The original “raft” hypothesis has been modified to accommodate the finding that they consist of many different lipids and are dynamic structures that cluster to form platforms. Simons et al.37 recently developed methods for immunoisolation of pure organelles in yeast; subsequent high-resolution mass spectroscopy of lipid species revealed more than 2000 lipid species in cell lysates and differences in different organelles.37 These major advances are being applied to mammalian cells and permit quantitative analysis of cellular lipids and their dynamic changes. Here, too, the situation is rendered complex because the inner and outer leaflets of the plasma membrane have different lipid composition, which results from specific “flippases” in the membrane. For example, ABCB4 “flips” PC from the inner to the outer leaflet of the plasma membrane. Simultaneously, FIC1, the mutated product of which causes Byler's disease and benign recurrent cholestasis,38 is an aminophospholipid adenosine triphosphatase that transfers phosphatidylethanolamine and phosphatidylserine from the outer to the inner leaflet of the canalicular membrane,39 an activity that is somehow associated with FXR and ABCB11 expression.40 These rapid phospholipid transfer activities are constitutive and possibly influence the stability and composition of transmembrane proteins, including canalicular ABC transporters. 3′ Phosphoinositide products of PI-3 kinase activity are also required for trafficking ABCB11 from Golgi to the canalicular membrane41 and for degradative traffic from the plasma membrane to lysosomes. In addition, PI 3 kinase is associated with the canalicular membrane. Suppression by wortmannin inhibited ATP-dependent taurocholate transport in canalicular membrane vesicles, perfused rat liver, and transfected WIF-B9 cells.42 Inhibition was overcome by addition of 3′ but not 4′ phosphoinositides, which are the substrates for PI 3-kinase. A decapeptide that facilitates binding of the enzyme to its 4′ phosphoinositide substrates enhances taurocholate secretion in canalicular membrane vesicles,42 WIFB9 cells and perfused rat liver.

Eventually it will be possible to identify all major components and quantify cascades of their interrelations as well as mechanisms of agonists and antagonists. Currently, molecular or chemical manipulation of 9 separate components produces experimental bile secretory failure. Principles learned from these studies will be applicable to similar processes in other polarized epithelial cells.

The second major paradigm shift resulted from discovery of hormone nuclear receptors (NRs), largely by Ronald Evans.43 In a prescient article entitled “Little is known about Bileology,” Evans echoed the molecular biologists' desire to understand cellular physiology and the translational importance of their studies. These bridges have been increasingly built, and research in this area is proceeding in several directions.

Nuclear receptors (NR) are pleotrophic transcription factors (LXR, FXR, CAR, PXR, RXR) that are critical for liver function by regulating the activity of key genes controlling the synthesis and metabolism of an increasingly diverse group of lipophilic molecules including fatty acids, bile acids, and cholesterol.44 These receptors regulate bile acid, cholesterol, and drug uptake, secretion, and metabolism and are coordinately expressed and function in the liver, intestine, kidney, and other organs. Recent studies reveal important roles in an expanding repertoire of sensor-coupled defense mechanisms, which maintain hepatic integrity despite cytotoxic exposure to bacterial toxins and xenobiotics. Expression profiling of the enlarging family of NRs suggests important roles in metabolic regulation, circadian rhythm, and nutrition.45 These advances provide the basis for interest in developing pharmaceuticals related to hypercholesterolemia, fibrosis, and acquired and inheritable cholestatic disease. How NRs are coordinately expressed and regulated is a major challenge.

Because NRs are critical for physiological regulation of bile acid metabolism and bile secretion, genetic and acquired changes in NRs function may be important in the pathogenesis and treatment of cholestasis.46 Although primary NR gene defects have not been identified as causing cholestasis, polymorphisms have been described in patients with primary biliary cirrhosis, sclerosing cholangitis (PXR), cholestasis of pregnancy (FXR), and PFIC1 (FXR); their pathobiological importance is not known. In response to impaired biliary secretion of bile acids and other organic anions, FXR, PXR, and CAR are activated and orchestrate downstream responses, including repression of hepatobiliary uptake mechanisms and bile acid synthesis as well as induction of bile acid detoxification and apical (ABCB11) and basolateral efflux pumps (ABCC 3 and 4) for conjugated bile acids and bilirubin. This response has been proposed to protect hepatocytes against bile acid accumulation, which may, conceptually, be therapeutically enhanced in cholestatic disease. Many currently used pharmaceuticals, including chenodeoxycholate and rifampin, affect these mechanisms. Selective targeting of these processes by pharmaceuticals acting on NRs is an exciting prospect.

Serendipity lead us to explore an unexpected relation between trafficking of ABC transporters and apical polarity in hepatocytes. Polarization of epithelial cells is required for directional transport and absorption, which are essential for survival. The complex components and mechanisms responsible for apical polarization of hepatocytes, including generation of canalicular microvilli, which increase secretory surface area, are gradually being unravelled largely as a result of studies performed in polarized cell lines other than hepatocytes and are discussed elsewhere.47, 48 They involve complex interactions between the cytoskeleton, tight and adherent junctions, signal transduction, specific polarity proteins, lipid rearrangements, ion flux, and trafficking systems. Maintenance of polarity requires other components affecting protein and lipid turnover. Given our observations on the role of rab 11a and myosin Vb in apical trafficking of ABCB11, we anticipated that dominant negative expressions or RNAi knockdown would impair ABCB11 trafficking resulting in cholestasis. To our surprise, stable knockdown of Rab11a by RNAi or expression of a Rab11a guanosine diphosphate–locked mutant prevented polarization in WIF-B9 cells.49 Overexpression of the tail domain of myosin Vb, which competes with full-length myosin Vb for binding to Rab 11a, also prevented bile canalicular formation. The absence of Rab 11a or myosin Vb caused ABC transporters, including ABCB11, to remain intracellular and colocalized with transcytosis membrane proteins, which were also transported to the plasma membrane. These observations suggest that sorting and trafficking pathways may be reprogrammed during differentiation and polarization and that the rab 11a–myosin V compartment, which traffics ABCB11 and other ABC cargo to the apical domain, may contain cues for polarization or activate substrates at the apical plasma membrane that are required for polarization. The rab 11a-recycling compartment appears to be essential for creation and possibly maintenance of canalicular polarization. If so, this organelle is a likely target for cholestatic-producing agents (in other words, viruses, drugs) as well as potential therapeutics.

More recent studies reveal that polarization of tight junctions is also regulated by AMP-activated protein kinase (AMPK), thus linking it to metabolic control and glucose utilization.50 AMPK is the major metabolic sensor in the cell and is influenced by nutrition, reactive oxygen, circadian rhythm, stress, and glucose. In addition, AMPK is related to polarity in drosophila, worms, and mammalian cells.51 The dynamic nature of plasma membrane components and trafficking and docking are likely to be influenced by cellular metabolism as regulated by AMPK.52 These observations may be relevant to cholestasis in hyperalimentation, starvation, and prematurity.

Much as the two major paradigm shifts in biliary secretion and cholestasis have advanced understanding of these processes, they have, like all good science, prompted more questions into complex cellular and organ functions and regulation. The challenge is to bridge the incredible advances in basic biology with physiology and pathophysiology. This phase of the chase is just beginning. Many new tools are becoming available, permitting examination of ever more complex biological phenomena.

If I may predict, advances beyond our current imagination will rapidly increase understanding of cholestasis and many other pathophysiological processes. Studies of cell biological complexity, signal cascades, genomic polymorphism, transcriptional physiology, and intracellular dynamic trafficking and regulation require collaboration with other disciplines, including mathematics, bioengineering, and developmental biology. Although these ideas may appear as “things dreams are made on,” our current dreams are likely shortsighted. The future will be more exciting than the past and offers great opportunity for those interested in bridging science with medicine.

The author appreciates the opportunity to have worked with many colleagues who have shared our journey, in particular: Charles Cornelius, Gerald Fleischner, Lawrence Gartner, Zenaida Gatmaitan, Michael Gottesman, Matt Harris, Masayasu Inoue, Peter Jansen, Tatehiro Kagawa, Yoshiyuki Kamimoto, Rolf Kinne, Ralph Kirsch, Tsuneo Kitamura, Cynthia Leveille-Webster, A. Jonathan Levi, Alec Mowat, Toshirou Nishida, Daniel Ortiz, Humberto Reyes, Jayanta and Namita Roy-Chowdhry, Franz Simon, Peter Ujhazy, Yoshiyuki Wakabayashi, Allan Wolkoff, and Lyuba Varticovski. The National Institutes of Health (NIH) has generously supported our research since 1958, for which we are all immensely appreciative.