Biliary physiology and disease: Reflections of a physician-scientist


  • Potential conflict of interest: Nothing to report.


A review is presented of Gustav Paumgartner's five decades of research and practice in hepatology focusing on biliary physiology and disease. It begins with studies of the excretory function of the liver including hepatic uptake of indocyanine green, bilirubin, and bile acids. The implications of these studies for diagnosis and understanding of liver diseases are pointed out. From there, the path of scientific research leads to investigations of hepatobiliary bile acid transport and the major mechanisms of bile formation. The therapeutic effects of the hydrophilic bile acid, ursodeoxycholic acid, have greatly stimulated these studies. Although ursodeoxycholic acid therapy for dissolution of cholesterol gallstones and some other nonsurgical treatments of gallstones were largely superseded by surgical techniques, ursodeoxycholic acid is currently considered the mainstay of therapy of some chronic cholestatic liver diseases, such as primary biliary cirrhosis. The major mechanisms of action of ursodeoxycholic acid therapy in cholestatic liver diseases are discussed. An attempt is made to illustrate how scientific research can lead to advances in medical practice that help patients. (HEPATOLOGY 2010:51:1095–1106.)

Although I am honored by the invitation to contribute to HEPATOLOGY's “Master's Perspective” series, I sometimes feel more like a student than a master. As a physician-scientist, I have lived and worked in two worlds—scientific theory and medical practice—and had to cross the gap between them back and forth many times. In this article, I will reflect on some of my personal experiences on my sometimes circuitous path of scientific discovery and treatment of patients.

First Crossings

Inspired by my father, who was a country doctor in Austria, I decided at an early age to study medicine. However, after receiving a scholarship from Princeton University and spending a year there, I became so fascinated by biology and chemistry that I decided to become a biochemist. This was in 1953, when I wrote my “Junior Paper” entitled “Some Aspects of Cellular Oxidation” and became excited about Michaelis-Menton kinetics of enzymatic reactions. (Many years later, while studying the elimination of substances from the blood by the liver, Michaelis-Menton kinetics would come to my mind again.) Back in Austria, I decided to become a physician after all, but with the intention of applying scientific discoveries to the treatment of patients. This was the first of many crossings back and forth between theory and practice.

After finishing Medical School at the University of Vienna, I joined the Department of Pharmacology there as a postdoctoral fellow. Under Franz Brücke and Otto Kraupp, I learned techniques of paper and thin-layer chromatography and fluorimetric assays to study the metabolism of catecholamines. Although I had enjoyed the work and my mentors invited me to pursue a career in pharmacology, I declined. Sticking to my plan to become a physician, I joined the Department of Medicine II at the University of Vienna as an intern in internal medicine and gastroenterology. Besides my work in the ward, I participated in a clinical study under Georg Grabner, in which we performed hepatic vein catheterizations in patients with liver cirrhosis to measure portal pressure and estimate liver blood flow using sulfobromophthalein.


ABC, adenosine triphosphate–binding cassette; BSEP, bile salt export pump; CDCA, chenodeoxycholic acid; CYP, cytochrome p450; DCA, deoxycholic acid; FGF15, fibroblast growth factor 15; FXR, farnesoid X receptor; ICG, indocyanine green; MDR, multidrug resistance p-glycoprotein; MRP, multidrug resistance-associated protein; NTCP, sodium taurocholate cotransporting polypeptide; OATP, organic anion transporting protein; OST, organic solute transporter; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; SHP, short heterodimer partner; TCDCA, taurochenodeoxycholic acid; TUDCA, tauroursodeoxycholic acid.

Biliary Physiology and Pathophysiology

Excretory function of the liver and bile formation are essential for life. Many endogenous and exogenous substances are eliminated from the body by uptake into the hepatocytes, metabolism to less toxic and more water-soluble compounds, and excretion into the bile. Because I had used sulfobromophthalein and later indocyanine green (ICG) in my studies of liver blood flow, I became interested in dye excretion by the liver and consequently in bile formation. At this time, little was known about the molecular mechanisms of bile formation.

Elimination of ICG and Bile Acids by the Liver.

When Hans Popper visited Vienna in 1964, I had the opportunity to present our findings on liver blood flow in patients with liver cirrhosis to him. He advised me to apply for a fellowship with Carroll Leevy at the New Jersey College of Medicine. In 1965, I started to work with Carroll Leevy and used the opportunity to regularly visit Hans Popper and his group at Mount Sinai Hospital. When I joined Carroll Leevy on a trip to Chicago in 1965 to attend the meeting of the American Association for the Study of Liver Diseases, I met a number of colleagues who had been at the Thorndike Laboratory at Boston together with Carroll Leevy and who had become leading scientists in their specific fields of hepatology. At this time, I received much scientific stimulation. In my perspective, nothing is more important for a young physician-scientist than to have the right mentors and role models at the right time. I was very lucky in this respect (Fig. 1).

Figure 1.

Hans Popper (first row, third from the left) and the author (first row, second from the left) taking notes at the Falk Symposium “Jaundice” in Freiburg, Germany, in October 1967.

Under the mentorship of Carroll Leevy, I correlated hepatic function, namely phagocytic activity assessed by clearance of colloidal carbon and excretory function assessed by clearance of ICG, with morphology in animal models of liver injury induced by CCl4 or ethanol.1 In a project funded by the National Aeronautics and Space Administration, ICG removal rate proved to be a sensitive measure for the detection of proton radiation injury to the liver.2

For clinical studies of liver function, we developed dichromatic ear densitometry to measure, without blood sampling, ICG clearance from the blood in patients with liver disease.3 I was intrigued by the observation that disappearance of ICG from the blood obviously exhibited saturation phenomena, and I started to explore the relationship between ICG dose and disappearance rate from the blood. Primed by my exposure to Michaelis-Menton kinetics at Princeton University, I successfully employed the Michaelis-Menton equation for the description of ICG uptake by the liver.4, 5 It became clear that serum disappearance rate of low doses of ICG was not a sensitive parameter of excretory functional capacity of the liver but rather reflected blood flow. We suggested the use of low and high doses of ICG for assessment of blood flow and excretory capacity, respectively, and to obtain estimates of the apparent Michaelis constant (Km) and the maximal velocity (Vmax) of ICG removal from the blood.5, 6 This approach was theoretically appealing but impractical and did not gain clinical utility. The concept, however, remains useful, especially because noninvasive measurement of ICG plasma disappearance rate has had a kind of renaissance in recent years.7–10 When I left New Jersey to return to Vienna, Carroll Leevy suggested that it would be valuable for my future research to study not only the hepatic removal of dyes but also the elimination of endogenous compounds such as bile acids.

Back in Vienna, I continued my ICG clearance studies in the rat and investigated the effect of bilirubin on ICG clearance. Together with Jürg Huber and Peter Probst, we presented evidence that ICG and bilirubin share a common pathway for uptake by the liver. At this time, knowledge of the molecular mechanisms for elimination of ICG and bilirubin was practically nonexistent. It was thought that bilirubin uptake into the hepatocyte occurs by passive nonionic diffusion. Our finding of competitive inhibition of ICG clearance by bilirubin questioned this concept (Fig. 2).4, 5 Lack of inhibition of ICG clearance by glycocholate in the rat suggested that ICG and bile acids are taken up into the hepatocyte by different mechanisms.4, 5 Together with Wolfgang Horak, we demonstrated that the inhibition of bile flow that occurs when ICG was infused at rates above its biliary transport maximum (Tm) was not caused by inhibition of bile salt excretion but by inhibition of the bile salt–independent fraction of bile flow.11

Figure 2.

Lineweaver-Burk plot showing the kinetics of hepatic uptake of indocyanine green (ICG) in the rat and competitive inhibition of ICG uptake by bilirubin. Dark symbols denote control animals receiving ICG only, and open symbols denote animals receiving 2 mg bilirubin/100 g body weight intravenously 30 seconds before the ICG injection. Uptake velocity of ICG is given in milligrams per 100 g liver/minute and dose of ICG is given in milligrams per 100 g body weight. Reproduced with permission from Paumgartner et al.5

From theory, I went back to practice to study a 40-year-old jaundiced patient on the ward with conjugated serum bilirubin of 2.3 mg/dL in the absence of any signs of liver disease. Liver biopsy revealed the diagnosis of Dubin-Johnson syndrome. We had just established thin-layer chromatography to measure serum bile acids and found normal serum bile acid levels in this patient and in five further patients with Dubin-Johnson syndrome.12, 13 ICG clearance, however, was impaired.14 This observation supported our hypothesis that bile acids and organic anions such as bilirubin and ICG are transported from blood to bile by different mechanisms and that in the Dubin-Johnson syndrome, biliary excretion of bile acids is intact. This early and original work was published in the German language, and thus, it did not reach the relevant scientific community and was lost into oblivion.

I realized that for more detailed studies of hepatobiliary transport, I had to move to a laboratory with more expertise, support, and facilities for basic research. At this time, Rudi Preisig had just returned from Columbia University, New York, where he had worked with Stanley Bradley and Henry Wheeler, and asked me to join his group in the Department of Clinical Pharmacology at the University of Bern. I accepted and wonderful years in Switzerland with excellent fellows followed.

With Jürg Reichen, we used Goresky's multiple indicator dilution technique in the perfused rat liver to demonstrate that uptake of taurocholate by the liver exhibited kinetics of carrier-mediated transport.15 We also studied hepatic bilirubin uptake. Our findings were consistent with the concept that not only uptake of bile acids but also uptake of bilirubin is carrier-mediated. In addition, we confirmed my original hypothesis that different pathways exist for the uptake of bile acids and bilirubin by the liver.16, 17

In the perfused rat liver, we then demonstrated for the first time that uptake of taurocholate is sodium-dependent. Substitution of 94% of the Na+ in the perfusion medium decreased the maximal uptake velocity (Vmax) and the apparent Michaelis constant (Km) of taurocholate uptake by 68% and 55%, respectively (Fig. 3).18 With Elisabeth Minder, we showed in isolated hepatocyes that uptake of taurocholate but not of ICG exhibited Na+-dependence, documenting again that the mechanisms responsible for hepatocellular uptake of conjugated bile acids and anionic dyes differ fundamentally.19

Figure 3.

Plot of uptake kinetics of taurocholate in the isolated perfused rat liver showing sodium dependence of hepatic taurocholate uptake. Perfusion medium containing 140 mM Na+ (open circles). Perfusion medium containing only 8 mM Na+, with the rest being replaced by Li+ (dark triangles). Substitution of Na+ by Li+ reduced apparent Km (intercept on abscissa) and Vmax (reciprocal of slope) leaving the ratio Km/Vmax (intercept on ordinate) virtually unchanged. Reproduced with permission from Reichen and Paumgartner.18

Biliary Secretion and Bile Acid Metabolism.

Our interest in bile acids was stimulated by the fact that hepatocellular bile acid secretion is the most important driving force for bile formation. A number of our studies together with Reinhard Herz and Klaus von Bergmann were aimed at defining and manipulating “bile acid–dependent” and “bile acid–independent” bile flow in the bile fistula rat. For instance, we studied the relationship between biliary taurocholate secretion and bile flow,20 the cholestasis which occurs when taurocholate is infused at rates exceeding its excretory transport maximum (Tm)21 and the stimulation of bile acid–independent bile flow by phenobarbital, spironolactone, and rifampicin.22

We also studied bile secretion and bile acid metabolism in rats with a portacaval shunt. In this experimental model, the diversion of portal blood into the systemic circulation resulted in liver atrophy and diminution of the bile acid–independent fraction of bile flow. However, the synthesis rate of cholate was maintained in spite of elevated bile acid concentrations in arterial blood.23 We were puzzled by these results. At that time, the role of fibroblast growth factor 15 (FGF15; FGF19 in humans) in the regulation of bile acid synthesis had been unknown. Today, we know that bile acids in the terminal ileum, by binding to farnesoid X receptor (FXR), induce FGF15, a hormone-like factor which represses bile acid synthesis via down-regulation of cytochrome p450 7A1 (CYP7A1).24 Regulation of bile acid synthesis by FGF15 in our rats with portacaval shunt probably had been intact, but data to answer this question are still not available.

Since the time of our studies in the late 1970s, breathtaking advances of knowledge have been made (Fig. 4). Today, we know that conjugated bile acids, which represent the major fraction of bile acids in the blood, are transported across the basolateral membrane of the hepatocyte together with sodium by the sodium taurocholate cotransporting polypeptide (NTCP, SLC10A1).25, 26 Unconjugated bile acids and a large variety of other organic anions, including bilirubin, are taken up into the hepatocytes mainly by one or more members of the organic anion transporter superfamily, encoded by SLCO genes, known as the organic anion–transporting polypeptides (OATPs).27 Uptake of unconjugated bile acids and bilirubin occurs mainly by OATP1B1 (encoded by the gene SLCO1B1) and is inhibited by ICG.28 With Gerd Kullak-Ublick, we characterized the human SLCO1A2 gene by mapping it to chromosome 12p12,29 and studied OATPA (now termed OATP1A2 and known to be localized in cholangiocytes) expression in human liver.30

Figure 4.

Hepatobiliary transport of bile acids and organic anions. Asterisk denotes (*) down-regulation and (**) up-regulation of transporter in cholestasis. Down-regulation of the basolateral transporters sodium taurocholate cotransporting polypeptide (NTCP) and organic anion transporting protein 1B1 (OATP1B1) and up-regulation of multidrug resistance-associated protein 3 (MRP3), MRP4, and organic solute transporter α/β (OST α/β) (which provide alternative routes of elimination) reduce the hepatocellular accumulation of bile acids and other cholephiles in cholestasis. The adaptive/compensatory responses in cholestasis are mediated by nuclear receptors, including the farnesoid X receptor (FXR), the pregnane X receptor (PXR), and the constitutive androgen receptor (CAR). The effects of FXR on NTCP and OATP1B1 are mediated through other nuclear receptors such as the short heterodimer partner (SHP) and the hepatocyte nuclear factor 4α (HNF4α). BSEP, bile salt export pump.

The rate-limiting step for the elimination of bile acids and other cholephiles from the blood, under most conditions, is the active transport across the canalicular membrane of the hepatocyte. It is driven by a number of ATP-dependent export pumps (ATP-binding cassette [ABC] transport proteins, also known as ABC transporters). Bile salts are transported by the bile salt export pump (BSEP, ABCB11),31 whereas bilirubin bisglucuronide, glutathione, divalent bile acid conjugates, and a large variety of other conjugated organic anions are transported by the multidrug resistance-associated protein 2 (MRP2, ABCC2).32 In patients with the Dubin-Johnson syndrome, MRP2 is mutated and not functioning,33, 34 which results in a defect of biliary bilirubin glucuronide excretion whereas the secretion of bile acids is normal.

Hepatobiliary transporter proteins are regulated at both transcriptional and post-transcriptional levels.35 A number of nuclear receptors that function as ligand-activated transcription factors36–38 function as positive feed-forward and negative feedback regulators to maintain enterohepatic homeostasis.39 In cholestasis, they are largely responsible for the adaptive coordinated responses in liver, kidney, and the intestine to limit hepatocellular accumulation of potentially toxic biliary constituents.39 As a result of the adaptive transcriptional program in cholestasis, basolateral bile acid uptake systems are down-regulated, while basolateral export pumps such as MRP3 and MRP4 are induced (Fig. 4). Moreover, bile acid hydroxylation and conjugation, which reduce toxicity and increase water solubility for subsequent alternative elimination via the urine, are stimulated.39 The central regulator of these adaptive responses is the nuclear receptor for bile acids, FXR (NR1H4). It reduces bile acid uptake into the hepatocyte by down-regulation of NTCP (Na+-dependent bile acid uptake) and OATP1B1 (Na+-indpendent bile acid uptake).39 It up-regulates BSEP (canalicular bile acid excretion)40 and MRP2 (canalicular excretion of conjugated bilirubin).41 In the intestine, FXR stimulates release of FGF15 (FGF19 in humans) which in the liver down-regulates CYP7A1, the rate-limiting enzyme for bile acid synthesis. FXR in the hepatocyte inhibits CYP7A via the small heterodimer partner (SHP) or, in humans, possibly also via induction of FGF19.42

When in 1979 I became Chairman of the Department of Medicine II at the University of Munich, I had the opportunity to perform measurements of serum bile acid levels in larger numbers of patients with liver disease. Together with Alexander Mannes, we studied serum bile acids as prognostic markers in patients with chronic liver disease and showed that the concentration of bile acids in serum more closely correlated with mortality in alcoholic and posthepatitic cirrhosis than the commonly used laboratory parameters and yielded a prediction of mortality comparable to the Child classification.43

Biliary Disease

Gallstone Disease.

Gallstone disease is one of the most common digestive diseases. About 80% of the gallstones in Western industrialized countries consist mainly of cholesterol.44, 45 The pathogenesis of cholesterol gallstone disease is a multifactorial process, with cholesterol supersaturation of bile, destabilization of supersaturated bile, and gallbladder hypomotility being the main factors.45 Supersaturation of bile with cholesterol results from a disproportion between cholesterol and its solubilizers, phosphatidylcholine and bile acids.

Cholesterol is secreted into the bile via the cholesterol transporter ABCG5/ABCG8,46, 47 phosphatidylcholine by the multidrug resistance p-glycoprotein 3 (MDR3), a canalicular ABC transporter encoded by the ABCB4 gene, and bile acids by the BSEP encoded by the ABCB11 gene (Fig. 5). Cholesterol and phosphatidylcholine in bile form metastable unilamellar vesicles. Under the influence of bile acids, these vesicles are converted into stable mixed micelles consisting of cholesterol, phosphatidylcholine, and bile acids. If there is an excess of cholesterol in relation to phosphatidylcholine and bile acids, cholesterol-rich vesicles remain and fuse into large unstable multilamellar vesicles from which cholesterol crystals nucleate.

Figure 5.

Biliary secretion and solubilization of cholesterol. Cholesterol is secreted by the ATP-binding cassette (ABC) transporter ABCG5/ABCG8, phosphatidylcholine by the multidrug resistance p-glycoprotein 3 (MDR3; ABCB4), and bile acids by the bile salt export pump (BSEP; ABCB11). Cholesterol and phosphatidylcholine reach the bile as metastable unilamellar vesicles, which are converted into water-soluble stable mixed micelles by the bile acids. If the secretion of cholesterol into bile exceeds the solubilizing capacity of bile acids and phospholipids, cholesterol-rich vesicles remain, which aggregate into large unstable multilamellar vesicles from which cholesterol crystals precipitate. These crystals may aggregate and form cholesterol stones. Recent evidence indicates that a variant of the cholesterol transporter gene ABCG8 contributes to the risk of cholesterol gallstone formation. Reproduced with permission from Paumgartner and Greenberger, Gallstone disease. In: Greenberger N, Blumberg RS, Burakoff R, eds. Current Diagnosis and Treatment: Gastroenterology, hepatology, endoscopy. New York: McGraw-Hill; 2009:537–546.

The major cause for cholesterol supersaturation of bile is hypersecretion of cholesterol.48, 49 In addition to overnutrition, low physical activity, obesity, low-fiber diet, prolonged fasting, and rapid weight loss, the metabolic syndrome and insulin resistance have been recognized as important risk factors for supersaturation of bile with cholesterol and cholesterol gallstone formation.45 Although environmental factors, such as those mentioned above, are most important for the development of cholesterol gallstones, genetic factors are thought to contribute about 25% to the gallstone phenotype.50 Recently, a polymorphism of the cholesterol transporter, namely ABCG8D19H, which causes a gain-of-function mutation of cholesterol secretion, has been identified as a genetic risk factor for cholesterol gallstone disease.51, 52

In addition to the above factors, a high proportion of deoxycholic acid (DCA) in the bile acid pool had been postulated to be associated with cholesterol supersaturation of bile and gallstone disease.53 This stimulated our own research in this direction. Together with Frans Stellaard, we developed methods to measure pool sizes and turnover rates of DCA, cholic acid, and CDCA in patients, by using 13C-labeled bile acids.54–56 This method was then used in a study together with Frieder Berr to measure pool size and input rate of DCA in patients with gallstones. We showed that, depending on gallbladder function, DCA input is increased in a subset of patients with cholesterol gallstones, and we argued that this contributes to supersaturation of bile with cholesterol.57 In a follow-up study, Frieder Berr and Phil Hylemon58 demonstrated that increased cholic acid 7-alpha-dehydroxylating activity of the intestinal microflora may be responsible for the increased DCA input into the bile acid pool and thus be a risk factor for cholesterol gallstone formation in these patients. They argued that the manipulation of the intestinal flora may be an approach to reduce the risk for gallstone formation.

Gallstone Dissolution and Shock Wave Lithotripsy.

In 1972, CDCA was introduced into clinical medicine for gallstone dissolution.59 Together with Alan Hofmann, I organized a small international workshop in Frankfurt in 1973 on “Chenodeoxycholic acid therapy of gallstones”, sponsored by Dr. Herbert Falk.60 Among the participants were Rudy Danzinger, Hermon Dowling, Hans Fromm, Hans Popper, Adolf Stiehl, and John Thistle. Although the clinical effectiveness of CDCA for dissolution of gallstones had been demonstrated, there was little knowledge on the pharmacokinetics of CDCA and the optimal dosage. We therefore studied the pharmacokinetics of CDCA given in different formulations such as gelatin capsules or micronized water suspensions. Postingestion patterns of serum bile acids were quite inconsistent when unconjugated CDCA was given in gelatin capsules, but two peaks were revealed when the bile acid was administered in the form of a micronized water suspension. Spill-over of unconjugated CDCA into the systemic circulation was seen with a first peak about 15–30 minutes and a second peak following approximately 2–3 hours after ingestion of CDCA.61 The first peak was increased two-fold to three-fold in patients with liver disease.62 Bile acid analysis in bile showed that unconjugated CDCA at doses employed for gallstone dissolution was completely conjugated with glycine and taurine during one passage through the liver.61, 63

Later, ursodeoxycholic acid (UDCA) was introduced into clinical medicine for gallstone dissolution by Makino et al.64, 65 It had been used in Japan for this indication by M. Shimizu already soon after it became commercially available in Japan in 1957.66 The long time, namely up to 2 years, often required for complete dissolution and failures to dissolve the stones, often caused by calcifications on the stone surface, were drawbacks of this new therapy. Together with Tilman Sauerbruch and Michael Delius, we therefore adapted the technique of extracorporeal shock wave lithotripsy, which had just been developed at our hospital for the fragmentation of kidney stones, for the fragmentation of gallstones and combined this approach with dissolution of stone fragments by bile acid therapy.67 The first treatments were performed in general anesthesia. Those were moments of suspense when we watched the first patient wake up from general anesthesia, but everything went well. Soon, the procedure was performed under mild intravenous analgesia on an outpatient basis68 and was very well accepted by the patients. Our first publication in the New England Journal of Medicine67 received considerable attention, and colleagues from all over the world came to Munich to see and to learn the new technique. On the basis of our results,69, 70 including the finding that gallbladder emptying was very important for fragment clearance after shock wave lithotripsy,71, 72 we established selection criteria (one radiolucent stone with a diameter less <20 mm, normal gallbladder function) for this new treatment modality to optimize the outcome.73

Gallstone Recurrence.

With Michael Sackmann, we soon realized that stone recurrence was a problem74, 75 and together with Frieder Berr and Jürgen Pauletzki, we studied stone recurrence. Major risk factors for stone recurrence were incomplete gallbladder emptying76 and a high proportion of DCA in the bile acid pool.77

Results of other groups suggested that the outcome of shock wave lithotripsy can be improved by better stone fragmentation (called “pulverization”), by employing higher shock wave energies and repeated treatments.78 The degree of fragmentation appeared to be even more important for the success than bile acid therapy. Together with Gerd Sauter, we conducted a randomized, double-blinded multicenter trial to compare the safety and efficacy of repeated high-energy shock wave lithotripsy with and without adjuvant bile acid therapy.79 We found that repeated high-energy shock wave lithotripsy without adjuvant bile acid therapy was safe and effective in patients with small single stones (≤20 mm in diameter) and good gallbladder emptying and that adjuvant bile acid therapy improved the outcome in patients with larger stones and multiple stones.79

In comparison to extracorporeal shock wave lithotripsy, cholecystectomy had the advantage of being successful in practically all patients irrespective of the stone type with no recurrence, but had the disadvantage of a relatively long hospital stay, cosmetically less favorable results, and albeit small mortality. This situation dramatically changed with the advent of laparoscopic cholecystectomy. Soon, this was the preferred therapeutic modality and shock wave lithotripsy of gallbladder stones became obsolete.80 Extracorporal shock wave lithotripsy retained some utility only for bile duct stones81–86 and pancreatic stones.87, 88


Cholangiopathies are a heterogeneous group of diseases characterized by an inflammatory destruction of bile ducts leading to cholestatic liver disease. The two major cholangiopathies are primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC). Although PBC is considered to be immune-mediated, the causes of both diseases are still unknown. Present therapies are therefore directed toward suppression of the pathogenetic processes and are not curative.

UDCA Therapy of Cholestatic Liver Diseases.

We were surprised and skeptical when Leuschner et al.89 in 1985 reported that serum liver tests improved in patients with chronic hepatitis who were treated with UDCA for dissolution of gallstones. First reports about beneficial effects of UDCA in patients with liver disease had been published in the Japanese literature90, 91 but were not noticed outside Japan. UDCA had been isolated in Japan in 1927 from dried bile of the black bear (“Yutan”) which was used in traditional Chinese and Japanese medicine as a remedy for liver and gastrointestinal disorders for centuries.92 In 1987, Poupon et al.93 reported that UDCA improved serum liver tests in patients with PBC. With Ulrich Beuers, we started the first double-blind, randomized, placebo-controlled trial for the treatment of PSC with UDCA in 1988. The results of this relatively small trial using 13–15 mg/kg/day of UDCA were published in 1992 and showed significant improvement of serum bilirubin, alkaline phosphatase, gamma-glutamyltransferase, and aminotransferases after 1 year of treatment94 (Fig. 6). Liver histology improved, as evaluated by a multiparametric score95 but not by disease stage. A trial of 13–15 mg/kg/day of UDCA for 2 years published by Lindor et al.96 in 1997 confirmed the improvement of serum liver tests but did not find improvement of liver histology. Trials with higher doses of UDCA to ensure sufficient enrichment of the bile with UDCA followed. A trial in Scandinavia in more than 200 patients with PSC with a follow-up of 5 years demonstrated a trend toward increased survival in the UDCA-treated group but did not reach statistical significance.97 A recent study by Lindor et al.98 used even higher doses of UDCA, namely 28–30 mg/kg/day, but was terminated because of an enhanced risk in the UDCA group for death or liver transplantation and serious adverse events. These occurred mainly in patients with advanced liver disease. This is not completely surprising because for every bile acid, UDCA included, there will be a dose that cannot be handled by the diseased liver. Doses above this threshold will be toxic. Together with Reinhard Herz,21 we had shown in 1976 in the rat that also taurocholate when infused above its transport maximum (Tm) causes cholestasis. Even below this threshold, high doses of UDCA may have adverse effects in patients with PSC who have downstream biliary obstruction.99

Figure 6.

Serum levels of bilirubin, alkaline phosphatase (AP), γ-glutamyltransferase (γ-GT), alanine aminotransferase (ALT), and aspartate aminotransferase (AST) in patients with primary sclerosing cholangitis before and after 6 and 12 months of therapy with ursodeoxycholic acid (dark symbols) or placebo (open symbols). Data are expressed as median percentage deviation from the baseline values. *P < 0.05; **P < 0.001. Reproduced with permission from Beuers et al.94

The UDCA trials in PSC show the difficulties of assessing survival or even progression in a slowly progressive and uncommon disease. The required number of patients is so high and the duration of treatment so long that it is very difficult to perform adequate studies. An additional problem is that PSC may be a mixed bag. Only recently have we learned to regard patients with immunoglobulin G4–associated cholangiopathy as a separate disease, about which we know that it responds to corticosteroids.100 Other still unknown cholangiopathies may be in this category. Therefore, it will be necessary to re-evaluate moderate doses of UDCA in better-selected patients with better-designed trials. After having gone a long way, UDCA therapy in PSC is not much clearer today than at the time when the first controlled trials were performed. Development of therapies for PSC may become easier once we know more about the etiology (or etiologies) of PSC.

For PBC, UDCA (13–15 mg/kg/day) is currently considered the mainstay of therapy and is the only drug approved by the U.S. Food and Drug Administration. Randomized, double-blinded, placebo-controlled trials have consistently shown that UDCA improves serum liver tests, including bilirubin, an important prognostic marker in PBC.95, 101–103 It also reduced elevated concentrations of endogenous bile acids in serum (Fig. 7).104 UDCA delayed the progression of histologic stage when treatment was initiated at stages I to II,105, 106 and it improved survival free of liver transplantation.107 These findings are supported by three recent long-term studies from France,108 Spain,109 and the Netherlands.110 On the basis of all available data, it is currently recommended to treat PBC with UDCA using doses of 13–15 mg/kg/day.100, 111

Figure 7.

Time course of serum bile acid concentrations and distributions in patients with primary biliary cirrhosis treated with ursodeoxycholic acid (UDCA) or placebo. The histogram represents mean bile acid concentrations (μmol/L) at 6-month intervals; the columns show the proportion of primary and secondary bile acids and UDCA relative to the total bile acid concentration. Reproduced with permission from Poupon et al.104

Mechanisms of Action of UDCA in Cholestatic Liver Diseases.

Although beneficial effects had been observed in patients with PBC and PSC under treatment with UDCA, the mechanisms of action remained incompletely understood. In 1987, Poupon et al.93 suggested that qualitative changes in the bile acid pool which resulted in a reduced accumulation of hydrophobic bile acids in the hepatocytes may have been responsible for the observed clinical effects of UDCA. The concept was put forward that UDCA displaces hydrophobic and potentially toxic bile acids from the total bile acid pool. Together with Ulrich Beuers and Frans Stellard, we therefore measured bile acid pool sizes of the hydrophobic bile acids CDCA and DCA with stable isotope-labeled bile acids in patients with PBC or PSC at 4 weeks after start of treatment with UDCA. At this time, the serum liver tests of the patients had already improved. To our surprise, the CDCA and DCA pools did not change in our patients.112 We realized that the enrichment of bile with UDCA was not caused by displacement of hydrophobic bile acids from the enterohepatic circulation.

Stimulated by findings of Kitani92 that tauroursodeoxycholic acid (TUDCA) prevents taurochenodeoxycholic acid (TCDCA)-induced cholestasis in the isolated perfused rat liver by effects on biliary secretion, Ulrich Beuers started a long series of experiments investigating the effects of UDCA on intracellular signaling pathways involved in biliary secretory processes. In the isolated perfused rat liver, he demonstrated that TUDCA activates a complex network of signals in the cholestatic rat liver which stimulates vesicular exocytosis and insertion of transporter proteins into the canalicular membrane of the hepatocyte.113 Activation of protein kinase C (PKC) played a central role in this signaling cascade.114–116 Thus, UDCA enhanced the amounts of the transporter proteins Mrp and Bsep in the canalicular membrane of rat hepatocytes and thereby stimulated bile flow and biliary secretion of hydrophobic bile acids and other potentially toxic compounds.117, 118

Today, we know that adaptive responses occur in cholestasis (see above) by which transporter proteins, such as MRP3, MRP4, and OSTα/OSTβ are up-regulated in the basolateral membrane of the hepatocyte in order to export bile acids and other cholephiles into the systemic blood for excretion via the kidney.39 It has been shown that UDCA up-regulates MRP4 protein in the liver of noncholestatic humans.119 This occurs mainly on a posttranscriptional level. By these mechanisms, the accumulation of toxic compounds in the hepatocytes, and thereby liver injury, is reduced.

We can paint a broader picture of the mechanisms of action of UDCA in cholestatic liver diseases today (Fig. 8).37, 38 Stimulation of hepatobiliary secretion by UDCA is one of several mechanisms. Other important mechanisms are a reduction of the toxicity of bile and an inhibition of bile acid–induced apoptosis. The toxicity of bile is reduced by the enrichment of bile with UDCA, by stimulation of the secretion of phospholipids via MDR3, and by dilution with bicarbonate-rich cholangiocellular bile.38 Depending on the specific cholestatic disease and the stage of the disease, one or the other of these mechanisms, or all in concert, may play a role. When the disease is caused by an immunologic injury of the biliary epithelium such as in PBC, changes in bile composition which reduce the toxicity of bile will be most important, especially in early stages of the disease. In the later stages with cholestasis, stimulation of secretion and antiapoptotic effects will increasingly gain importance. When the primary cause of cholestasis is a defect of biliary secretion, as for instance in intrahepatic cholestasis of pregnancy, stimulation of hepatocellular secretion appears to be of major importance.

Figure 8.

Major targets of ursodeoxycholic acid therapy in cholestatic liver diseases. BA, bile acid.


The path I have taken as a physician-scientist has had many turns. In retrospect, I realize that I could have saved time and effort had I seen some of the problems more clearly and approached them more directly. Sometimes the approach to a problem was influenced too much by the tools available, when it would have been better to first develop more appropriate methods. Sometimes I have felt like the man looking for his keys under the light of a street lamp. Asked why he is searching there and not where he thought he had lost his keys, he replies that only there, in the light of the lantern, can he see anything at all.

As a physician-scientist, I have been fortunate to be able to experience both the thrill of scientific discovery and the satisfaction of helping patients, but I also felt the gap between the two and the difficulty of crossing it nimbly. A physician must do more than apply scientific discoveries to the therapy of patients, for he or she is not only treating diseases, but human beings, with all their sufferings, fears, and hopes. I was reminded of this poignantly by Sir Karl Popper, the great philosopher, who was my patient at one time. One day, after I had examined him, he asked me whether I had found him in good health. When I replied, “Let's first define health,” he looked at me steadily and said “Now is not the time for definitions”. He needed an answer from his doctor at that moment and not a philosophical discussion with a scientist. Once again I found myself staring into the gap between theory and practice—and once again I took the leap across it.


My work described in this article would not have been possible without my mentors, fellows, and students. Not all of them could be mentioned in this article. I am deeply grateful to them. I also acknowledge the continuing support of my family.