REVIEWS IN BASIC AND CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Robert F. Schwabe and John W. Wiley, Section Editors
Inherited Disorders of Bilirubin Transport and Conjugation: New Insights Into Molecular Mechanisms and Consequences Serge Erlinger,1 Irwin M. Arias,2 and Daniel Dhumeaux3 1 University of Paris 7, Paris, France; 2National Institutes of Health, Bethesda, Maryland; and 3Henri Mondor Hospital, Créteil, University of Paris-Est, Créteil, France
Inherited disorders of bilirubin metabolism might reduce bilirubin uptake by hepatocytes, bilirubin conjugation, or secretion of bilirubin into bile. Reductions in uptake could increase levels of unconjugated or conjugated bilirubin (Rotor syndrome). Defects in bilirubin conjugation could increase levels of unconjugated bilirubin; the effects can be benign and frequent (Gilbert syndrome) or rare but severe, increasing the risk of bilirubin encephalopathy (Crigler–Najjar syndrome). Impairment of bilirubin secretion leads to accumulation of conjugated bilirubin (Dubin–Johnson syndrome). We review the genetic causes and pathophysiology of disorders of bilirubin transport and conjugation as well as clinical and therapeutic aspects. We also discuss the possible mechanisms by which hyperbilirubinemia protects against cardiovascular disease and the metabolic syndrome and the effects of specific genetic variants on drug metabolism and cancer development. Keywords: Crigler–Najjar Syndrome; Hepatic Storage Disease; Glucuronosyl Transferase; Bile Secretion; Kernicterus.
H
ereditary hyperbilirubinemias range from benign to lethal, and all are caused by defective bilirubin transport or conjugation by the liver. Bilirubin is the end product of heme catabolism. It belongs to the superfamily of tetrapyrrolic compounds, one of the most highly conserved groups of molecules in living organisms. Bilirubin is poorly water soluble. In blood, it circulates bound to serum albumin, presumably to prevent the toxicity of free (unbound) bilirubin. Unbound bilirubin is rapidly and selectively taken up by hepatocytes and then conjugated to glucuronic acid into bilirubin glucuronides by uridine diphosphate (UDP)glucuronosyl transferase before being secreted into bile through the hepatocyte canalicular membrane via an adenosine triphosphate (ATP)-dependent transporter. In clinical practice, hereditary hyperbilirubinemias can be separated into predominantly unconjugated and predominantly conjugated forms. These conditions result from mutations of transporters or enzymes involved in the hepatic bilirubin elimination pathway. The aim of this review is to describe these inherited disorders, with a particular focus on their molecular mechanisms. Studies of bilirubin metabolism have broader implications, showing the beneficial effects of moderate hyperbilirubinemia (due to the antioxidant
properties of bilirubin) and other consequences of mutations on drug metabolism and cancer susceptibility. We will not discuss hereditary hemolytic unconjugated hyperbilirubinemia, which is caused by bilirubin overproduction, or several genetically mediated cholestatic diseases such as progressive familial intrahepatic cholestasis or Alagille syndrome, which can also lead to hyperbilirubinemia.
Bilirubin Transport and Conjugation by the Liver Uptake by Hepatocytes: Passive Diffusion or Active Transport? Unconjugated bilirubin is lipid soluble and should thus readily cross biological membranes. However, passive diffusion alone would not account for the remarkable specificity of hepatic uptake. One possible explanation for this specificity is the presence in hepatocytes of cytoplasmic proteins with a higher affinity than albumin for bilirubin. One such protein was identified and characterized in the late 1960s and early 1970s by Arias et al and was named Y protein or ligandin.1 Kinetic studies suggested that bilirubin binding to this protein was not involved in initial cellular uptake but rather reduced bilirubin efflux from the cytosol back into the space of Disse, thus resulting in intrahepatocytic bilirubin accumulation. More recent studies have attempted to identify bilirubin transport proteins in the hepatocyte basolateral membrane, particularly among the organic anion transport proteins (OATPs). They belong to the OATP superfamily, which is also called the solute carrier organic anion transporter (SLCO) superfamily.2 The human SLCO superfamily comprises 11 members grouped
Abbreviations used in this paper: ATP, adenosine triphosphate; BSP, bromosulphophthalein; CN, Crigler–Najjar; DJS, Dubin–Johnson syndrome; GS, Gilbert syndrome; ICG, indocyanine green; MRP2, multidrug related protein 2; OATP, organic anion transport protein; OMIM, Online Mendelian Inheritance in Man; RS, Rotor syndrome; SLCO, solute carrier organic anion transporter; UDP, uridine diphosphate; UGT1A1, uridine diphosphate glucuronosyl transferase 1A1. © 2014 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2014.03.047
REVIEWS AND PERSPECTIVES
Gastroenterology 2014;146:1625–1638
1626 Erlinger et al
REVIEWS AND PERSPECTIVES
into 6 families encoded by SLCO genes. Bilirubin is a substrate for OATP1B1 (Online Mendelian Inheritance in Man [OMIM]*604843) and OATP1B3 (OMIM*605495).2 Human OATP1B1 and OATP1B3 can transport conjugated and, possibly, unconjugated bilirubin in vitro.3,4 Studies in humans, and particularly genome-wide association studies, suggest that polymorphisms that reduce OATP1B1 or OATP1B3 activity are associated with higher serum levels of both conjugated and unconjugated bilirubin.5,6 Oatp1a and Oatp1b knockout mice lacking the Oatp1a and 1b transporters have serum bilirubin levels more than 40 times higher than those of their wild-type counterparts.7 Serum bilirubin in these mice is mostly conjugated, probably (see the following text) because of defective reuptake of bilirubin glucuronide.8 However, serum levels of unconjugated bilirubin are 2-fold higher in Oatp1a/1b knockout mice than in their wild-type counterparts.7 This suggests that Oatp1a/1b proteins may contribute to unconjugated bilirubin uptake by hepatocytes. Human embryonic kidney cells (HEK293) permanently expressing recombinant OATP1B1 (formerly OATP2) showed uptake of [3H]monoglucuronosyl bilirubin, [3H]bisglucuronosyl bilirubin, and [3H]sulfobromophthalein, with Km values of 0.10, 0.28, and 0.14 mmol/L, respectively.3 However, this observation could not be reproduced with unconjugated bilirubin in HeLa or HEK293 cells transfected with OATP1B1.9 Further studies are thus needed to clarify the respective roles of passive diffusion and carriermediated transport in unconjugated bilirubin uptake by hepatocytes.
Conjugation After its uptake and binding to ligandin, bilirubin is transferred to the smooth endoplasmic reticulum, where it is conjugated into bilirubin glucuronides by UDP–glucuronosyl transferase 1A1 (UGT1A1). The process of bilirubin conjugation and the function of UDP–glucuronosyl transferases have been extensively reviewed recently.10–12 UGT1A1 (OMIM*191740) (Figure 1) appears to be the only enzyme that glucuronidates bilirubin. Indeed, mutations that completely suppress UGT1A1 activity result in a total absence of bilirubin glucuronides.11 UGT1A1 is a transmembrane protein located mainly on the smooth endoplasmic reticulum. It has a binding site for bilirubin and another one for glucuronic acid, with both sites located on the luminal face of the endoplasmic reticulum membrane.11 Glucuronic acid is derived from uridine diphospho-glucuronic acid, which itself is derived
Gastroenterology Vol. 146, No. 7
from UDP glucose. The location of the binding sites implies that both bilirubin and glucuronic acid are transported from the cytosol into the lumen of the endoplasmic reticulum. UGT1A1 catalyzes the conversion of bilirubin to bilirubin monoglucuronide and then to bilirubin diglucuronide. Bilirubin glucuronides then move to the cytoplasm, probably through a specific endoplasmic reticulum membrane transporter (Figure 2, inset), where they bind to ligandin, albeit with far lower affinity than unconjugated bilirubin.
Secretion Into Bile and Plasma Once back in the cytosol, bilirubin diglucuronide can diffuse toward either the canalicular pole or the sinusoidal pole of the hepatocyte. At the canalicular pole, it is efficiently secreted into bile, mostly by the ATP-dependent MRP2/ABCC2 transporter.13,14 This protein mediates the canalicular secretion of several organic anions, including bilirubin glucuronides, dyes such as sulfobromophthalein (BSP) and indocyanine green (ICG), divalent bile salts, and reduced glutathione.13,14 Other transporters, particularly Abcg2,15 may be involved in bilirubin secretion across the canalicular membrane. Interestingly, a substantial fraction of bilirubin glucuronide is rerouted to the sinusoidal pole and secreted back into plasma by another transporter, Abcc3.8 From there, it can be taken up again by hepatocytes via Oatp1b1/3.8 It has been proposed that this reuptake process may take place in downstream hepatocytes (hepatocytes located near the central vein) to prevent saturation of the biliary secretory capacity of upstream hepatocytes (hepatocytes located near the portal tracts).8 A schematic representation of bilirubin transport and conjugation by hepatocytes is provided in Figure 2.
Hereditary Hyperbilirubinemias Hereditary hyperbilirubinemias may be caused by increased bilirubin production, mostly as a result of hyperhemolysis, or decreased bilirubin clearance. This review will be limited to conditions associated with decreased bilirubin clearance. Decreased bilirubin clearance may be caused by (1) defective bilirubin uptake by hepatocytes, leading to unconjugated hyperbilirubinemia; (2) defective conjugated bilirubin reuptake, such as in Rotor syndrome (RS) (also known as hepatic uptake and storage disease; see the following text); (3) defective bilirubin conjugation, such as in Gilbert syndrome (GS), Crigler–Najjar (CN) syndrome, neonatal transient familial hyperbilirubinemia, and breast
Figure 1. Schematic representation of the UGT1A1 locus and UGT1A1 protein.
Hereditary Hyperbilirubinemia 1627
REVIEWS AND PERSPECTIVES
June 2014
Figure 2. Schematic view of bilirubin transport by the hepatocyte. (Inset) Localization of UGT1A1 in the reticulum endoplasmic membrane and transport of uridine diphosphoglucuronic acid (UDPGA), bilirubin, and bilirubin glucuronide across the endoplasmic reticulum membrane. UCB, unconjugated bilirubin; Alb, albumin; BG, bilirubin glucuronide.
milk jaundice; or (4) defective bilirubin canalicular secretion, such as in Dubin–Johnson syndrome (DJS).
Defective Uptake One might expect defective bilirubin uptake to result systematically in unconjugated hyperbilirubinemia. Surprisingly, however, only a few cases of unconjugated hyperbilirubinemia have been attributed to defective uptake. Most defects of bilirubin uptake result in predominantly conjugated hyperbilirubinemia (RS). The reason for this was recently identified, as explained in the following text.
Unconjugated hyperbilirubinemia associated with an uptake defect? Rare cases of unconjugated hyperbilirubinemia have been reported in which plasma clearance of cholephilic dyes such as BSP and ICG was markedly impaired,16–18 suggesting a role of impaired hepatic bilirubin uptake. Among the 39 cases of unconjugated hyperbilirubinemia that we investigated, 3 patients had a significant reduction in the BSP plasma disappearance rate. Interestingly, the patients were 3 brothers (suggesting familial transmission), and liver UGT1A1 activity measured in one of these patients was normal, ruling out GS (if defined by defective conjugation).17 Genetic and molecular analyses suggest that such cases of unconjugated hyperbilirubinemia could be linked to polymorphisms in SLCO1B1 and SCLO1B3.5,6,8,19 However, given the modest role of OATPs in bilirubin uptake,8 the role of these polymorphisms remains to be confirmed.
Although this type of unconjugated hyperbilirubinemia is probably rare, its prevalence may nonetheless have been underestimated; in the absence of BSP plasma kinetics and liver UGT1A1 activity measurement, many cases may have been erroneously classified as GS. Indeed, although glucose6-phosphate dehydrogenase deficiency and other causes of unconjugated hyperbilirubinemia could not be ruled out, Skierka et al recently showed that 39% of their patients with unconjugated hyperbilirubinemia did not have identifiable UGT1A1 variants that explained the disorder.20
Rotor Syndrome: Conjugated Hyperbilirubinemia Due to Defective Reuptake? History and presentation. The disorder first described by Rotor et al in 1948 is a rare and benign disease (OMIM*237450) characterized by low-grade (40–100 mmol/L), chronic or fluctuating, predominantly conjugated hyperbilirubinemia.21 It is a familial disorder with autosomal recessive transmission.22 Except for jaundice when apparent, there are no symptoms and clinical findings are normal. In cases with apparent jaundice, RS is usually detected shortly after birth or during childhood. Other cases are discovered fortuitously or after diagnosis of another family member. Apart from the predominantly conjugated hyperbilirubinemia, liver test results are normal. Urinary excretion of coproporphyrins is markedly elevated because of increased urinary excretion of isomer I and, to a lesser extent, isomer III,22 following a shift from the biliary to the urinary route of coproporphirin excretion. Liver biopsy is not required. When performed, it is normal and shows no abnormal pigment deposits,
1628 Erlinger et al
REVIEWS AND PERSPECTIVES
contrary to DJS.23 The prognosis is excellent, and no treatment is necessary. Because of the presence of conjugated bilirubin in blood, RS was initially attributed to defective biliary excretion and was considered a variant of DJS. In 1975, one of us reported a case of hereditary, benign, predominantly conjugated hyperbilirubinemia associated with very slow plasma clearance of BSP and other cholephilic dyes such as dibromosulphthalein and ICG.24 The hepatic relative storage capacity of BSP and dibromosulphthalein was markedly impaired, whereas the biliary transport maximum was only slightly affected.24 After this description (hepatic uptake and storage disease; OMIM*237550), several groups, including ours, reexamined patients with RS and found that they all had defective hepatic uptake and storage, mainly characterized by decreased BSP clearance and storage capacity.25–27 From that time onward, hepatic uptake and storage disease and RS were considered to be the same entity.23 Impressive progress has since been made in our understanding of hepatic transporters and the pathophysiology of hereditary hyperbilirubinemias. In the absence of mutations in ABCC2 (the canalicular export pump for bilirubin glucuronide and other organic anions), it has been confirmed that RS is not caused by defective biliary excretion.28 However, despite these advances, it remained unclear why serum bilirubin is predominantly conjugated in patients with RS. Genetic and molecular basis. The recent work of van de Steeg et al probably solves this mystery.8 These investigators generated knockout mice in which the Slco1a/b genes (encoding the sinusoidal transporters Oatp1a/b, functionally close to human OATP1B1/3) were inactivated. They showed that (1) these knockout mice exhibited marked conjugated hyperbilirubinemia and (2) sinusoidal Oatps in the normal mouse function in “tandem” with the sinusoidal efflux transporter Abcc3 to successively mediate hepatic efflux (by Abcc3) and reuptake (by Oatps) of bilirubin glucuronides. They also found that transgenic expression of human OATP1B1 or OATP1B3 restored the liver-blood shuttle in Oatp1a/1b-deficient mice, indicating that, in humans, both OATP1B1 and OATP1B3 effectively reuptake bilirubin glucuronide from plasma to liver, in line with their bilirubin glucuronide uptake capacity in vitro.7 The same investigators scanned the whole genome and mapped candidate gene intervals in 11 patients with RS from 8 different families. Homozygosity mapping identified a single genomic region on chromosome 12 for which 8 of the tested index patients and none of the healthy siblings were homozygous, suggesting inheritance of both mutated alleles from a common ancestor. Sequence analysis revealed predictably pathogenic mutations affecting both SLCO1B3 and SLCO1B1 in each of the tested subjects. Autosomal recessive segregation was found in all the investigated RS families. The severity of the mutations was supported by the near absence of OATP1B immunostaining in liver biopsy specimens. Although these experimental results need to be confirmed, mouse experiments and studies of patients with RS strongly suggest that homozygous mutations of the human OATP genes SLCO1B1 and SLCO1B3 on chromosome 2, inducing complete OATP1B1 and OATP1B3 deficiency,
Gastroenterology Vol. 146, No. 7
disrupt hepatic reuptake of bilirubin conjugates and lead to accumulation of conjugated bilirubin in blood, thus causing RS. OATP1B1/3 deficiency also explains the marked impairment of BSP uptake and storage capacity in patients with RS (BSP is a substrate of these transporters5), as well as their increased urinary coproporphyrin excretion, in line with the interaction of several porphyrins with OATP1B1.29 Finally, the obligatory deficiency in 2 different genes explains the rarity of RS, which has an estimated frequency of about 1 in 106 overall, although it may be several times lower or higher in specific populations.8 Drug interactions. OATP1B1 plays a major role in drug detoxification. Indeed, reduced-activity OATP1B1 polymorphisms have been shown to reduce drug transport and increase plasma and tissue concentrations of drugs such as anticancer agents, methotrexate, and statins, potentially resulting in toxicity.30–33 Even if no drug accumulation or toxicity has so far been reported in patients with RS, it is advisable to take these findings into account, especially in patients with RS who have jaundice. Some drugs, such as high-dose cyclosporin A and antiviral agents such as alisporivir, can increase the plasma concentration of conjugated bilirubin, with no other evidence of liver damage. Until now, this was believed to be mediated primarily by inhibition of the canalicular transporter ABCC2. However, cyclosporin A and other drugs also inhibit OATP1B,7 and this might be an additional cause of drug-induced conjugated hyperbilirubinemias. Animal model. The hepatic transport abnormalities seen in RS resemble those of mutant Southdown sheep.34 Indeed, like patients with RS, these sheep are characterized by chronic hyperbilirubinemia, impaired plasma clearance of BSP and ICG, and a reduction in the BSP hepatic relative storage capacity with a subsequent reduction in transport maximum.34 Serum bilirubin is predominantly unconjugated in this model, originally suggesting defective uptake. The Oatp1b1/3 functional status of these animals has not been tested.
Defective Conjugation Gilbert Syndrome. First described at the turn of the 20th century,35 GS (OMIM*143500) is characterized by mild, predominantly unconjugated hyperbilirubinemia without hyperhemolysis that usually occurs in young adults with otherwise normal liver test results, including normal serum bile acid levels. Patients sometimes describe right upper abdominal discomfort or dyspepsia, but the relationship with hyperbilirubinemia is unclear. Except for jaundice when apparent, clinical findings are normal, as is liver histology when performed. The serum bilirubin level rarely exceeds 70 mmol/L in these patients. Fasting increases serum bilirubin levels, while phenobarbital, a microsomal enzyme inducer, lowers serum bilirubin levels, often to normal values. Treatment with this drug is rarely necessary, however. The prevalence of GS in the general population is approximately 8%. The condition is benign, and the prognosis is excellent. UGT1A1 mutations. A total of 130 UGT1A1 mutations have been identified to date.36,37 The most common molecular defect in GS is the addition of an extra dinucleotide
sequence, TA, to the promoter TATA box of the conjugating enzyme UGT1A1.38 The resulting genotype is designated A(TA)7TAA (instead of the normal A(TA)6TAA) or UGT1A1*28. Much less common A(TA)5TAA or A(TA)8TAA mutations are also found.39 UGT1A1 activity in those homozygous for one of these mutations is approximately 10% to 35% of normal, owing to decreased synthesis of the functionally normal enzyme. Bilirubin in bile is mostly diglucuronidated, but the proportion of monoglucuronide is >20% instead of the normal 7%.12 Homozygosity for this promoter mutation is necessary but not sufficient for hyperbilirubinemia to occur; population studies suggest that only 40% of A(TA)7TAA homozygotes have hyperbilirubinemia.38 Additional variables such as mild hemolysis (reported in up to 50% of patients with GS),40 dyserythropoiesis, additional defects in bilirubin uptake, or perhaps other mutations41 may be necessary for biochemical or clinical expression of the mutation. The existence of an uptake defect in GS (in addition to UGT1A1 deficiency) has been repeatedly suggested, mostly because the clearance of diagnostic dyes such as BSP and ICG is delayed in some patients.16,18 Unconjugated hyperbilirubinemia without overt hemolysis (the clinical definition of GS) could be attributable to 3 mechanisms: (1) decreased UGT1A1 activity with no uptake defect, (2) both decreased UGT1A1 activity and defective uptake, and (3) defective bilirubin uptake with normal UGT1A1 activity17 (see the preceding text). Although the 3 patient categories were probably covered by the original description of the disease,35 we propose that only patients with decreased UGT1A1 activity and the A(TA)7TAA genotype should be diagnosed with GS, whether or not they have an additional uptake defect. Interestingly, a patient with mutations in both the UGT1A1 promoter and ABCC2 (the bilirubin canalicular transporter gene) was found to have a phenotype corresponding to GS/DJS,42 illustrating the genetic complexity of these disorders. Several mutations in the coding regions of the UGT1A1 gene (rather than the promoter region) have been shown to result in proteins with only mildly reduced enzymatic activity. Such missense mutations resulting in a GS phenotype have thus far been reported only in Japan.43 Patients with GS are either homozygotes or compound heterozygotes, and GS is now regarded as an autosomal recessive disorder.44 More than 100 mutations have so far been reported, and their frequencies differ among countries.45 For example, UGT1A1*28 (þ/þ) is found in 12% of Scottish people, 16% of European people, 12% of Indian people, 8% of Egyptian people, and 23% of black people.45 Frequencies of UGT1A1*28 (þ/þ) are much lower in Asia. Some variants are found in geographically distinct populations. Haplotypes including more than one of these genetic variants have also been described; in particular, variants of the promoter TATA box can coexist with coding region mutations.45 This explains in part why GS is observed in only approximately 8% of the white general population despite a 10% to 16% prevalence of UGT1A1*28 homozygosity. GS and CN syndrome were once considered to be distinct genetic and pathophysiological entities, with GS considered
Hereditary Hyperbilirubinemia 1629
autosomal dominant and CN syndrome considered autosomal recessive. Both of these entities are now attributed to homozygous mutations of the same enzyme, UGT1A1, but with quantitatively different consequences. Molecular studies have clearly established that a single normal UGT1A1 allele is sufficient to maintain a normal plasma bilirubin concentration and that almost all cases of both GS and CN syndrome are autosomal recessive. The earlier belief that transmission of GS was dominant was probably attributable to the high frequency of the disease and, hence, the common observation of 2 successive affected generations. Diagnosis and management. The diagnosis of GS is usually based on mild prolonged unconjugated hyperbilirubinemia, without overt hemolysis and with otherwise normal liver tests and normal clinical and hepatic ultrasonographic findings. Hyperbilirubinemia is often discovered on a routine blood test performed for another reason and rarely because of clinical jaundice. A definitive diagnosis can be made by determining the A(TA)7TAA genotype, but such a sophisticated method should only be used in clinical practice when the serum bilirubin level is >70 mmol/L or when treatment with irinotecan is planned (see the following text). The patient should be reassured as to the benign course of the condition and informed that the metabolism of some drugs may be affected (see the following text). Phenobarbital (or another microsomal enzyme inducer) should only be prescribed when the bilirubinemia is high enough to cause overt, unsightly jaundice. GS and drug toxicity. Hepatic handling of a variety of drugs metabolized by glucuronidation may be affected in patients with GS, including menthol, estradiol benzoate, ethinyl estradiol, lamotrigine, tolbutamide, rifamycin SV, acetaminophen, nonsteroidal inflammatory drugs, statins and gemfibrozil,45 and human immunodeficiency virus protease inhibitors.45,46 The human immunodeficiency virus protease inhibitors indinavir and atazanavir produce hyperbilirubinemia by inhibiting UGT1A1.47 This hyperbilirubinemia is more pronounced in patients with preexisting GS. Sorafenib, used in patients with various types of cancer and approved for hepatocellular carcinoma, also inhibits UGT1A1, despite being metabolized by UGT1A9.48,49 It increases the serum bilirubin concentration in patients with GS who are homozygous for A(TA)7TAA as well as in A(TA)7TAA heterozygotes.48 Clinically significant toxicity is rarely ascribed to these pharmacokinetic abnormalities. An increased risk of antitubercular drug toxicity has been reported in people with a compound UGT1A1*27 and UGT1A1*28 genotype.50 The main drug-related danger in adults with GS comes from the antitumor agent irinotecan (CPT-11), the active metabolite of which is glucuronidated by UGT1A1. Administration of CPT-11 to patients with GS has resulted in severe toxicity, including intractable diarrhea and myelosuppression with severe neutropenia. European and American patients at risk are mostly UGT1A1*28,51,52 whereas most affected Japanese patients are UGT1A1*6.53 This field is rapidly expanding, and other drug toxicities may be discovered in the future.
REVIEWS AND PERSPECTIVES
June 2014
1630 Erlinger et al
REVIEWS AND PERSPECTIVES
The possibility of genotype-based dosing is being actively explored. GS and atherosclerosis: a genetic advantage? It is becoming clear that unconjugated bilirubin has a protective effect on cardiovascular disease and also perhaps on cancer. Long considered at best a waste product and at worst toxic in certain circumstances, a moderately elevated serum unconjugated bilirubin level is now known to have a number of beneficial effects on oxidative stress-mediated diseases, particularly vascular disease, diabetes, metabolic syndrome, and obesity.54 This appears to be largely due to its antioxidant action (for review, see Rigato et al55) and antiinflammatory properties.56 These epidemiological observations are supported by experimental evidence obtained both in vitro and in vivo. The negative relationship between serum bilirubin levels and coronary artery disease was first reported in 1994.57 Numerous studies have since confirmed the protective effect of moderately increased serum bilirubin concentrations (particularly in patients with GS) on both coronary and peripheral atherosclerotic disease.56,58–62 It appears that each 1-mmol/L increment in serum bilirubin level is associated with a 6.5% decrease in cardiovascular disease.61 Supporting these epidemiological observations, pathological studies have shown an inverse relationship between serum bilirubin concentration and both coronary artery calcification63 and ischemic stroke64 as well as markedly slower progression of carotid intimomedial thickness in patients with GS compared with normobilirubinemic subjects.65 The benefits of moderately increased serum bilirubin concentrations also extend to diabetes and metabolic syndrome.54,66,67 It appears that UGT1A1 variants resulting in hyperbilirubinemia may confer a strong genetic advantage with respect to major causes of death, including cardiovascular disease as well as overweight, obesity, and metabolic syndrome, the prevalence of which is increasing worldwide. Modulation of bilirubin levels may prove to be an attractive intervention for cardiovascular disease and metabolic syndrome. It is conceivable that the high worldwide allelic frequency of homozygous genetic variants of the UGT1A1 gene might be due in part to an evolutionary advantage. GS and cancer: good or bad?. As previously mentioned, a decrease in UGT1A1 activity impairs the glucuronidation not only of bilirubin but also of a variety of other compounds, including estrogen and benzopyrene metabolites.68 It is thus conceivable that reduced glucuronidation of estrogen and mutagens in tissues carrying UGT1A1*28 or other promoter variants could influence the development of hormone-dependent and carcinogen-induced tumors. In a Chinese study and in a European study, UGT1A1*6 and UGT1A1*7 variants were associated with an increased risk of colorectal cancer.69,70 Antioxidant mechanisms might be expected to protect against cancer initiation. Indeed, some studies have shown a decreased risk of colorectal cancer in patients with GS and in patients with moderate hyperbilirubinemia.71,72 Zucker et al found an inverse correlation between the prevalence of colorectal cancer and serum bilirubin concentrations
Gastroenterology Vol. 146, No. 7
among 176,748,462 American subjects, with each 10-mg/L (17-mmol/L) increase in the bilirubin concentration associated with a significant decrease in the prevalence of colorectal cancer.71 Jirásková et al detected a significant association between the risk of colorectal cancer and UGT1A1*28 allele carrier status.72 This issue clearly needs to be revisited. In a case-control study, low UGT1A1 activity due to UGT1A1*28 was associated with a lower risk of endometrial cancer.73 Low UGT1A1 activity may also improve the prognosis of Hodgkin lymphoma.74 Results for breast cancer are also controversial, probably because of geographic variability.45 A Chinese study showed an increase in the risk of breast cancer among women younger than 40 years of age carrying UGT1A1*28, but other variants were not examined.75 This link was not confirmed in a large recent case-control study from the United States.76 A single variable is unlikely to be associated with a clinically significant risk, and the issue clearly requires further haplotype and multigene analyses. This is true, for instance, of the claimed association between UGT1A1 variants and gallstones; an additional risk factor for gallstones, such as cystic fibrosis, spherocytosis, or sickle cell disease, was found in all reported cases.45 Crigler–Najjar Syndrome. CN syndrome is a rare recessive disorder that was first described in 1952.77 It is characterized by major unconjugated hyperbilirubinemia (100–750 mmol/L) due to UGT1A1 mutations. The estimated prevalence is 0.6 per million.78 There are 2 types of CN syndrome; in CN syndrome type I, bilirubin levels are 350 to 750 mmol/L and UGT1A1 mutations result in a complete or near-complete loss of UGT1A1 enzyme activity. In CN syndrome type II, bilirubin levels are 100 to 400 mmol/L and UGT1A1 activity is
Inherited Disorders of Bilirubin Transport and Conjugation: New Insights Into Molecular Mechanisms and Consequences Serge Erlinger,1 Irwin M. Arias,2 and Daniel Dhumeaux3 1 University of Paris 7, Paris, France; 2National Institutes of Health, Bethesda, Maryland; and 3Henri Mondor Hospital, Créteil, University of Paris-Est, Créteil, France
Inherited disorders of bilirubin metabolism might reduce bilirubin uptake by hepatocytes, bilirubin conjugation, or secretion of bilirubin into bile. Reductions in uptake could increase levels of unconjugated or conjugated bilirubin (Rotor syndrome). Defects in bilirubin conjugation could increase levels of unconjugated bilirubin; the effects can be benign and frequent (Gilbert syndrome) or rare but severe, increasing the risk of bilirubin encephalopathy (Crigler–Najjar syndrome). Impairment of bilirubin secretion leads to accumulation of conjugated bilirubin (Dubin–Johnson syndrome). We review the genetic causes and pathophysiology of disorders of bilirubin transport and conjugation as well as clinical and therapeutic aspects. We also discuss the possible mechanisms by which hyperbilirubinemia protects against cardiovascular disease and the metabolic syndrome and the effects of specific genetic variants on drug metabolism and cancer development. Keywords: Crigler–Najjar Syndrome; Hepatic Storage Disease; Glucuronosyl Transferase; Bile Secretion; Kernicterus.
H
ereditary hyperbilirubinemias range from benign to lethal, and all are caused by defective bilirubin transport or conjugation by the liver. Bilirubin is the end product of heme catabolism. It belongs to the superfamily of tetrapyrrolic compounds, one of the most highly conserved groups of molecules in living organisms. Bilirubin is poorly water soluble. In blood, it circulates bound to serum albumin, presumably to prevent the toxicity of free (unbound) bilirubin. Unbound bilirubin is rapidly and selectively taken up by hepatocytes and then conjugated to glucuronic acid into bilirubin glucuronides by uridine diphosphate (UDP)glucuronosyl transferase before being secreted into bile through the hepatocyte canalicular membrane via an adenosine triphosphate (ATP)-dependent transporter. In clinical practice, hereditary hyperbilirubinemias can be separated into predominantly unconjugated and predominantly conjugated forms. These conditions result from mutations of transporters or enzymes involved in the hepatic bilirubin elimination pathway. The aim of this review is to describe these inherited disorders, with a particular focus on their molecular mechanisms. Studies of bilirubin metabolism have broader implications, showing the beneficial effects of moderate hyperbilirubinemia (due to the antioxidant
properties of bilirubin) and other consequences of mutations on drug metabolism and cancer susceptibility. We will not discuss hereditary hemolytic unconjugated hyperbilirubinemia, which is caused by bilirubin overproduction, or several genetically mediated cholestatic diseases such as progressive familial intrahepatic cholestasis or Alagille syndrome, which can also lead to hyperbilirubinemia.
Bilirubin Transport and Conjugation by the Liver Uptake by Hepatocytes: Passive Diffusion or Active Transport? Unconjugated bilirubin is lipid soluble and should thus readily cross biological membranes. However, passive diffusion alone would not account for the remarkable specificity of hepatic uptake. One possible explanation for this specificity is the presence in hepatocytes of cytoplasmic proteins with a higher affinity than albumin for bilirubin. One such protein was identified and characterized in the late 1960s and early 1970s by Arias et al and was named Y protein or ligandin.1 Kinetic studies suggested that bilirubin binding to this protein was not involved in initial cellular uptake but rather reduced bilirubin efflux from the cytosol back into the space of Disse, thus resulting in intrahepatocytic bilirubin accumulation. More recent studies have attempted to identify bilirubin transport proteins in the hepatocyte basolateral membrane, particularly among the organic anion transport proteins (OATPs). They belong to the OATP superfamily, which is also called the solute carrier organic anion transporter (SLCO) superfamily.2 The human SLCO superfamily comprises 11 members grouped
Abbreviations used in this paper: ATP, adenosine triphosphate; BSP, bromosulphophthalein; CN, Crigler–Najjar; DJS, Dubin–Johnson syndrome; GS, Gilbert syndrome; ICG, indocyanine green; MRP2, multidrug related protein 2; OATP, organic anion transport protein; OMIM, Online Mendelian Inheritance in Man; RS, Rotor syndrome; SLCO, solute carrier organic anion transporter; UDP, uridine diphosphate; UGT1A1, uridine diphosphate glucuronosyl transferase 1A1. © 2014 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2014.03.047
REVIEWS AND PERSPECTIVES
Gastroenterology 2014;146:1625–1638
1626 Erlinger et al
REVIEWS AND PERSPECTIVES
into 6 families encoded by SLCO genes. Bilirubin is a substrate for OATP1B1 (Online Mendelian Inheritance in Man [OMIM]*604843) and OATP1B3 (OMIM*605495).2 Human OATP1B1 and OATP1B3 can transport conjugated and, possibly, unconjugated bilirubin in vitro.3,4 Studies in humans, and particularly genome-wide association studies, suggest that polymorphisms that reduce OATP1B1 or OATP1B3 activity are associated with higher serum levels of both conjugated and unconjugated bilirubin.5,6 Oatp1a and Oatp1b knockout mice lacking the Oatp1a and 1b transporters have serum bilirubin levels more than 40 times higher than those of their wild-type counterparts.7 Serum bilirubin in these mice is mostly conjugated, probably (see the following text) because of defective reuptake of bilirubin glucuronide.8 However, serum levels of unconjugated bilirubin are 2-fold higher in Oatp1a/1b knockout mice than in their wild-type counterparts.7 This suggests that Oatp1a/1b proteins may contribute to unconjugated bilirubin uptake by hepatocytes. Human embryonic kidney cells (HEK293) permanently expressing recombinant OATP1B1 (formerly OATP2) showed uptake of [3H]monoglucuronosyl bilirubin, [3H]bisglucuronosyl bilirubin, and [3H]sulfobromophthalein, with Km values of 0.10, 0.28, and 0.14 mmol/L, respectively.3 However, this observation could not be reproduced with unconjugated bilirubin in HeLa or HEK293 cells transfected with OATP1B1.9 Further studies are thus needed to clarify the respective roles of passive diffusion and carriermediated transport in unconjugated bilirubin uptake by hepatocytes.
Conjugation After its uptake and binding to ligandin, bilirubin is transferred to the smooth endoplasmic reticulum, where it is conjugated into bilirubin glucuronides by UDP–glucuronosyl transferase 1A1 (UGT1A1). The process of bilirubin conjugation and the function of UDP–glucuronosyl transferases have been extensively reviewed recently.10–12 UGT1A1 (OMIM*191740) (Figure 1) appears to be the only enzyme that glucuronidates bilirubin. Indeed, mutations that completely suppress UGT1A1 activity result in a total absence of bilirubin glucuronides.11 UGT1A1 is a transmembrane protein located mainly on the smooth endoplasmic reticulum. It has a binding site for bilirubin and another one for glucuronic acid, with both sites located on the luminal face of the endoplasmic reticulum membrane.11 Glucuronic acid is derived from uridine diphospho-glucuronic acid, which itself is derived
Gastroenterology Vol. 146, No. 7
from UDP glucose. The location of the binding sites implies that both bilirubin and glucuronic acid are transported from the cytosol into the lumen of the endoplasmic reticulum. UGT1A1 catalyzes the conversion of bilirubin to bilirubin monoglucuronide and then to bilirubin diglucuronide. Bilirubin glucuronides then move to the cytoplasm, probably through a specific endoplasmic reticulum membrane transporter (Figure 2, inset), where they bind to ligandin, albeit with far lower affinity than unconjugated bilirubin.
Secretion Into Bile and Plasma Once back in the cytosol, bilirubin diglucuronide can diffuse toward either the canalicular pole or the sinusoidal pole of the hepatocyte. At the canalicular pole, it is efficiently secreted into bile, mostly by the ATP-dependent MRP2/ABCC2 transporter.13,14 This protein mediates the canalicular secretion of several organic anions, including bilirubin glucuronides, dyes such as sulfobromophthalein (BSP) and indocyanine green (ICG), divalent bile salts, and reduced glutathione.13,14 Other transporters, particularly Abcg2,15 may be involved in bilirubin secretion across the canalicular membrane. Interestingly, a substantial fraction of bilirubin glucuronide is rerouted to the sinusoidal pole and secreted back into plasma by another transporter, Abcc3.8 From there, it can be taken up again by hepatocytes via Oatp1b1/3.8 It has been proposed that this reuptake process may take place in downstream hepatocytes (hepatocytes located near the central vein) to prevent saturation of the biliary secretory capacity of upstream hepatocytes (hepatocytes located near the portal tracts).8 A schematic representation of bilirubin transport and conjugation by hepatocytes is provided in Figure 2.
Hereditary Hyperbilirubinemias Hereditary hyperbilirubinemias may be caused by increased bilirubin production, mostly as a result of hyperhemolysis, or decreased bilirubin clearance. This review will be limited to conditions associated with decreased bilirubin clearance. Decreased bilirubin clearance may be caused by (1) defective bilirubin uptake by hepatocytes, leading to unconjugated hyperbilirubinemia; (2) defective conjugated bilirubin reuptake, such as in Rotor syndrome (RS) (also known as hepatic uptake and storage disease; see the following text); (3) defective bilirubin conjugation, such as in Gilbert syndrome (GS), Crigler–Najjar (CN) syndrome, neonatal transient familial hyperbilirubinemia, and breast
Figure 1. Schematic representation of the UGT1A1 locus and UGT1A1 protein.
Hereditary Hyperbilirubinemia 1627
REVIEWS AND PERSPECTIVES
June 2014
Figure 2. Schematic view of bilirubin transport by the hepatocyte. (Inset) Localization of UGT1A1 in the reticulum endoplasmic membrane and transport of uridine diphosphoglucuronic acid (UDPGA), bilirubin, and bilirubin glucuronide across the endoplasmic reticulum membrane. UCB, unconjugated bilirubin; Alb, albumin; BG, bilirubin glucuronide.
milk jaundice; or (4) defective bilirubin canalicular secretion, such as in Dubin–Johnson syndrome (DJS).
Defective Uptake One might expect defective bilirubin uptake to result systematically in unconjugated hyperbilirubinemia. Surprisingly, however, only a few cases of unconjugated hyperbilirubinemia have been attributed to defective uptake. Most defects of bilirubin uptake result in predominantly conjugated hyperbilirubinemia (RS). The reason for this was recently identified, as explained in the following text.
Unconjugated hyperbilirubinemia associated with an uptake defect? Rare cases of unconjugated hyperbilirubinemia have been reported in which plasma clearance of cholephilic dyes such as BSP and ICG was markedly impaired,16–18 suggesting a role of impaired hepatic bilirubin uptake. Among the 39 cases of unconjugated hyperbilirubinemia that we investigated, 3 patients had a significant reduction in the BSP plasma disappearance rate. Interestingly, the patients were 3 brothers (suggesting familial transmission), and liver UGT1A1 activity measured in one of these patients was normal, ruling out GS (if defined by defective conjugation).17 Genetic and molecular analyses suggest that such cases of unconjugated hyperbilirubinemia could be linked to polymorphisms in SLCO1B1 and SCLO1B3.5,6,8,19 However, given the modest role of OATPs in bilirubin uptake,8 the role of these polymorphisms remains to be confirmed.
Although this type of unconjugated hyperbilirubinemia is probably rare, its prevalence may nonetheless have been underestimated; in the absence of BSP plasma kinetics and liver UGT1A1 activity measurement, many cases may have been erroneously classified as GS. Indeed, although glucose6-phosphate dehydrogenase deficiency and other causes of unconjugated hyperbilirubinemia could not be ruled out, Skierka et al recently showed that 39% of their patients with unconjugated hyperbilirubinemia did not have identifiable UGT1A1 variants that explained the disorder.20
Rotor Syndrome: Conjugated Hyperbilirubinemia Due to Defective Reuptake? History and presentation. The disorder first described by Rotor et al in 1948 is a rare and benign disease (OMIM*237450) characterized by low-grade (40–100 mmol/L), chronic or fluctuating, predominantly conjugated hyperbilirubinemia.21 It is a familial disorder with autosomal recessive transmission.22 Except for jaundice when apparent, there are no symptoms and clinical findings are normal. In cases with apparent jaundice, RS is usually detected shortly after birth or during childhood. Other cases are discovered fortuitously or after diagnosis of another family member. Apart from the predominantly conjugated hyperbilirubinemia, liver test results are normal. Urinary excretion of coproporphyrins is markedly elevated because of increased urinary excretion of isomer I and, to a lesser extent, isomer III,22 following a shift from the biliary to the urinary route of coproporphirin excretion. Liver biopsy is not required. When performed, it is normal and shows no abnormal pigment deposits,
1628 Erlinger et al
REVIEWS AND PERSPECTIVES
contrary to DJS.23 The prognosis is excellent, and no treatment is necessary. Because of the presence of conjugated bilirubin in blood, RS was initially attributed to defective biliary excretion and was considered a variant of DJS. In 1975, one of us reported a case of hereditary, benign, predominantly conjugated hyperbilirubinemia associated with very slow plasma clearance of BSP and other cholephilic dyes such as dibromosulphthalein and ICG.24 The hepatic relative storage capacity of BSP and dibromosulphthalein was markedly impaired, whereas the biliary transport maximum was only slightly affected.24 After this description (hepatic uptake and storage disease; OMIM*237550), several groups, including ours, reexamined patients with RS and found that they all had defective hepatic uptake and storage, mainly characterized by decreased BSP clearance and storage capacity.25–27 From that time onward, hepatic uptake and storage disease and RS were considered to be the same entity.23 Impressive progress has since been made in our understanding of hepatic transporters and the pathophysiology of hereditary hyperbilirubinemias. In the absence of mutations in ABCC2 (the canalicular export pump for bilirubin glucuronide and other organic anions), it has been confirmed that RS is not caused by defective biliary excretion.28 However, despite these advances, it remained unclear why serum bilirubin is predominantly conjugated in patients with RS. Genetic and molecular basis. The recent work of van de Steeg et al probably solves this mystery.8 These investigators generated knockout mice in which the Slco1a/b genes (encoding the sinusoidal transporters Oatp1a/b, functionally close to human OATP1B1/3) were inactivated. They showed that (1) these knockout mice exhibited marked conjugated hyperbilirubinemia and (2) sinusoidal Oatps in the normal mouse function in “tandem” with the sinusoidal efflux transporter Abcc3 to successively mediate hepatic efflux (by Abcc3) and reuptake (by Oatps) of bilirubin glucuronides. They also found that transgenic expression of human OATP1B1 or OATP1B3 restored the liver-blood shuttle in Oatp1a/1b-deficient mice, indicating that, in humans, both OATP1B1 and OATP1B3 effectively reuptake bilirubin glucuronide from plasma to liver, in line with their bilirubin glucuronide uptake capacity in vitro.7 The same investigators scanned the whole genome and mapped candidate gene intervals in 11 patients with RS from 8 different families. Homozygosity mapping identified a single genomic region on chromosome 12 for which 8 of the tested index patients and none of the healthy siblings were homozygous, suggesting inheritance of both mutated alleles from a common ancestor. Sequence analysis revealed predictably pathogenic mutations affecting both SLCO1B3 and SLCO1B1 in each of the tested subjects. Autosomal recessive segregation was found in all the investigated RS families. The severity of the mutations was supported by the near absence of OATP1B immunostaining in liver biopsy specimens. Although these experimental results need to be confirmed, mouse experiments and studies of patients with RS strongly suggest that homozygous mutations of the human OATP genes SLCO1B1 and SLCO1B3 on chromosome 2, inducing complete OATP1B1 and OATP1B3 deficiency,
Gastroenterology Vol. 146, No. 7
disrupt hepatic reuptake of bilirubin conjugates and lead to accumulation of conjugated bilirubin in blood, thus causing RS. OATP1B1/3 deficiency also explains the marked impairment of BSP uptake and storage capacity in patients with RS (BSP is a substrate of these transporters5), as well as their increased urinary coproporphyrin excretion, in line with the interaction of several porphyrins with OATP1B1.29 Finally, the obligatory deficiency in 2 different genes explains the rarity of RS, which has an estimated frequency of about 1 in 106 overall, although it may be several times lower or higher in specific populations.8 Drug interactions. OATP1B1 plays a major role in drug detoxification. Indeed, reduced-activity OATP1B1 polymorphisms have been shown to reduce drug transport and increase plasma and tissue concentrations of drugs such as anticancer agents, methotrexate, and statins, potentially resulting in toxicity.30–33 Even if no drug accumulation or toxicity has so far been reported in patients with RS, it is advisable to take these findings into account, especially in patients with RS who have jaundice. Some drugs, such as high-dose cyclosporin A and antiviral agents such as alisporivir, can increase the plasma concentration of conjugated bilirubin, with no other evidence of liver damage. Until now, this was believed to be mediated primarily by inhibition of the canalicular transporter ABCC2. However, cyclosporin A and other drugs also inhibit OATP1B,7 and this might be an additional cause of drug-induced conjugated hyperbilirubinemias. Animal model. The hepatic transport abnormalities seen in RS resemble those of mutant Southdown sheep.34 Indeed, like patients with RS, these sheep are characterized by chronic hyperbilirubinemia, impaired plasma clearance of BSP and ICG, and a reduction in the BSP hepatic relative storage capacity with a subsequent reduction in transport maximum.34 Serum bilirubin is predominantly unconjugated in this model, originally suggesting defective uptake. The Oatp1b1/3 functional status of these animals has not been tested.
Defective Conjugation Gilbert Syndrome. First described at the turn of the 20th century,35 GS (OMIM*143500) is characterized by mild, predominantly unconjugated hyperbilirubinemia without hyperhemolysis that usually occurs in young adults with otherwise normal liver test results, including normal serum bile acid levels. Patients sometimes describe right upper abdominal discomfort or dyspepsia, but the relationship with hyperbilirubinemia is unclear. Except for jaundice when apparent, clinical findings are normal, as is liver histology when performed. The serum bilirubin level rarely exceeds 70 mmol/L in these patients. Fasting increases serum bilirubin levels, while phenobarbital, a microsomal enzyme inducer, lowers serum bilirubin levels, often to normal values. Treatment with this drug is rarely necessary, however. The prevalence of GS in the general population is approximately 8%. The condition is benign, and the prognosis is excellent. UGT1A1 mutations. A total of 130 UGT1A1 mutations have been identified to date.36,37 The most common molecular defect in GS is the addition of an extra dinucleotide
sequence, TA, to the promoter TATA box of the conjugating enzyme UGT1A1.38 The resulting genotype is designated A(TA)7TAA (instead of the normal A(TA)6TAA) or UGT1A1*28. Much less common A(TA)5TAA or A(TA)8TAA mutations are also found.39 UGT1A1 activity in those homozygous for one of these mutations is approximately 10% to 35% of normal, owing to decreased synthesis of the functionally normal enzyme. Bilirubin in bile is mostly diglucuronidated, but the proportion of monoglucuronide is >20% instead of the normal 7%.12 Homozygosity for this promoter mutation is necessary but not sufficient for hyperbilirubinemia to occur; population studies suggest that only 40% of A(TA)7TAA homozygotes have hyperbilirubinemia.38 Additional variables such as mild hemolysis (reported in up to 50% of patients with GS),40 dyserythropoiesis, additional defects in bilirubin uptake, or perhaps other mutations41 may be necessary for biochemical or clinical expression of the mutation. The existence of an uptake defect in GS (in addition to UGT1A1 deficiency) has been repeatedly suggested, mostly because the clearance of diagnostic dyes such as BSP and ICG is delayed in some patients.16,18 Unconjugated hyperbilirubinemia without overt hemolysis (the clinical definition of GS) could be attributable to 3 mechanisms: (1) decreased UGT1A1 activity with no uptake defect, (2) both decreased UGT1A1 activity and defective uptake, and (3) defective bilirubin uptake with normal UGT1A1 activity17 (see the preceding text). Although the 3 patient categories were probably covered by the original description of the disease,35 we propose that only patients with decreased UGT1A1 activity and the A(TA)7TAA genotype should be diagnosed with GS, whether or not they have an additional uptake defect. Interestingly, a patient with mutations in both the UGT1A1 promoter and ABCC2 (the bilirubin canalicular transporter gene) was found to have a phenotype corresponding to GS/DJS,42 illustrating the genetic complexity of these disorders. Several mutations in the coding regions of the UGT1A1 gene (rather than the promoter region) have been shown to result in proteins with only mildly reduced enzymatic activity. Such missense mutations resulting in a GS phenotype have thus far been reported only in Japan.43 Patients with GS are either homozygotes or compound heterozygotes, and GS is now regarded as an autosomal recessive disorder.44 More than 100 mutations have so far been reported, and their frequencies differ among countries.45 For example, UGT1A1*28 (þ/þ) is found in 12% of Scottish people, 16% of European people, 12% of Indian people, 8% of Egyptian people, and 23% of black people.45 Frequencies of UGT1A1*28 (þ/þ) are much lower in Asia. Some variants are found in geographically distinct populations. Haplotypes including more than one of these genetic variants have also been described; in particular, variants of the promoter TATA box can coexist with coding region mutations.45 This explains in part why GS is observed in only approximately 8% of the white general population despite a 10% to 16% prevalence of UGT1A1*28 homozygosity. GS and CN syndrome were once considered to be distinct genetic and pathophysiological entities, with GS considered
Hereditary Hyperbilirubinemia 1629
autosomal dominant and CN syndrome considered autosomal recessive. Both of these entities are now attributed to homozygous mutations of the same enzyme, UGT1A1, but with quantitatively different consequences. Molecular studies have clearly established that a single normal UGT1A1 allele is sufficient to maintain a normal plasma bilirubin concentration and that almost all cases of both GS and CN syndrome are autosomal recessive. The earlier belief that transmission of GS was dominant was probably attributable to the high frequency of the disease and, hence, the common observation of 2 successive affected generations. Diagnosis and management. The diagnosis of GS is usually based on mild prolonged unconjugated hyperbilirubinemia, without overt hemolysis and with otherwise normal liver tests and normal clinical and hepatic ultrasonographic findings. Hyperbilirubinemia is often discovered on a routine blood test performed for another reason and rarely because of clinical jaundice. A definitive diagnosis can be made by determining the A(TA)7TAA genotype, but such a sophisticated method should only be used in clinical practice when the serum bilirubin level is >70 mmol/L or when treatment with irinotecan is planned (see the following text). The patient should be reassured as to the benign course of the condition and informed that the metabolism of some drugs may be affected (see the following text). Phenobarbital (or another microsomal enzyme inducer) should only be prescribed when the bilirubinemia is high enough to cause overt, unsightly jaundice. GS and drug toxicity. Hepatic handling of a variety of drugs metabolized by glucuronidation may be affected in patients with GS, including menthol, estradiol benzoate, ethinyl estradiol, lamotrigine, tolbutamide, rifamycin SV, acetaminophen, nonsteroidal inflammatory drugs, statins and gemfibrozil,45 and human immunodeficiency virus protease inhibitors.45,46 The human immunodeficiency virus protease inhibitors indinavir and atazanavir produce hyperbilirubinemia by inhibiting UGT1A1.47 This hyperbilirubinemia is more pronounced in patients with preexisting GS. Sorafenib, used in patients with various types of cancer and approved for hepatocellular carcinoma, also inhibits UGT1A1, despite being metabolized by UGT1A9.48,49 It increases the serum bilirubin concentration in patients with GS who are homozygous for A(TA)7TAA as well as in A(TA)7TAA heterozygotes.48 Clinically significant toxicity is rarely ascribed to these pharmacokinetic abnormalities. An increased risk of antitubercular drug toxicity has been reported in people with a compound UGT1A1*27 and UGT1A1*28 genotype.50 The main drug-related danger in adults with GS comes from the antitumor agent irinotecan (CPT-11), the active metabolite of which is glucuronidated by UGT1A1. Administration of CPT-11 to patients with GS has resulted in severe toxicity, including intractable diarrhea and myelosuppression with severe neutropenia. European and American patients at risk are mostly UGT1A1*28,51,52 whereas most affected Japanese patients are UGT1A1*6.53 This field is rapidly expanding, and other drug toxicities may be discovered in the future.
REVIEWS AND PERSPECTIVES
June 2014
1630 Erlinger et al
REVIEWS AND PERSPECTIVES
The possibility of genotype-based dosing is being actively explored. GS and atherosclerosis: a genetic advantage? It is becoming clear that unconjugated bilirubin has a protective effect on cardiovascular disease and also perhaps on cancer. Long considered at best a waste product and at worst toxic in certain circumstances, a moderately elevated serum unconjugated bilirubin level is now known to have a number of beneficial effects on oxidative stress-mediated diseases, particularly vascular disease, diabetes, metabolic syndrome, and obesity.54 This appears to be largely due to its antioxidant action (for review, see Rigato et al55) and antiinflammatory properties.56 These epidemiological observations are supported by experimental evidence obtained both in vitro and in vivo. The negative relationship between serum bilirubin levels and coronary artery disease was first reported in 1994.57 Numerous studies have since confirmed the protective effect of moderately increased serum bilirubin concentrations (particularly in patients with GS) on both coronary and peripheral atherosclerotic disease.56,58–62 It appears that each 1-mmol/L increment in serum bilirubin level is associated with a 6.5% decrease in cardiovascular disease.61 Supporting these epidemiological observations, pathological studies have shown an inverse relationship between serum bilirubin concentration and both coronary artery calcification63 and ischemic stroke64 as well as markedly slower progression of carotid intimomedial thickness in patients with GS compared with normobilirubinemic subjects.65 The benefits of moderately increased serum bilirubin concentrations also extend to diabetes and metabolic syndrome.54,66,67 It appears that UGT1A1 variants resulting in hyperbilirubinemia may confer a strong genetic advantage with respect to major causes of death, including cardiovascular disease as well as overweight, obesity, and metabolic syndrome, the prevalence of which is increasing worldwide. Modulation of bilirubin levels may prove to be an attractive intervention for cardiovascular disease and metabolic syndrome. It is conceivable that the high worldwide allelic frequency of homozygous genetic variants of the UGT1A1 gene might be due in part to an evolutionary advantage. GS and cancer: good or bad?. As previously mentioned, a decrease in UGT1A1 activity impairs the glucuronidation not only of bilirubin but also of a variety of other compounds, including estrogen and benzopyrene metabolites.68 It is thus conceivable that reduced glucuronidation of estrogen and mutagens in tissues carrying UGT1A1*28 or other promoter variants could influence the development of hormone-dependent and carcinogen-induced tumors. In a Chinese study and in a European study, UGT1A1*6 and UGT1A1*7 variants were associated with an increased risk of colorectal cancer.69,70 Antioxidant mechanisms might be expected to protect against cancer initiation. Indeed, some studies have shown a decreased risk of colorectal cancer in patients with GS and in patients with moderate hyperbilirubinemia.71,72 Zucker et al found an inverse correlation between the prevalence of colorectal cancer and serum bilirubin concentrations
Gastroenterology Vol. 146, No. 7
among 176,748,462 American subjects, with each 10-mg/L (17-mmol/L) increase in the bilirubin concentration associated with a significant decrease in the prevalence of colorectal cancer.71 Jirásková et al detected a significant association between the risk of colorectal cancer and UGT1A1*28 allele carrier status.72 This issue clearly needs to be revisited. In a case-control study, low UGT1A1 activity due to UGT1A1*28 was associated with a lower risk of endometrial cancer.73 Low UGT1A1 activity may also improve the prognosis of Hodgkin lymphoma.74 Results for breast cancer are also controversial, probably because of geographic variability.45 A Chinese study showed an increase in the risk of breast cancer among women younger than 40 years of age carrying UGT1A1*28, but other variants were not examined.75 This link was not confirmed in a large recent case-control study from the United States.76 A single variable is unlikely to be associated with a clinically significant risk, and the issue clearly requires further haplotype and multigene analyses. This is true, for instance, of the claimed association between UGT1A1 variants and gallstones; an additional risk factor for gallstones, such as cystic fibrosis, spherocytosis, or sickle cell disease, was found in all reported cases.45 Crigler–Najjar Syndrome. CN syndrome is a rare recessive disorder that was first described in 1952.77 It is characterized by major unconjugated hyperbilirubinemia (100–750 mmol/L) due to UGT1A1 mutations. The estimated prevalence is 0.6 per million.78 There are 2 types of CN syndrome; in CN syndrome type I, bilirubin levels are 350 to 750 mmol/L and UGT1A1 mutations result in a complete or near-complete loss of UGT1A1 enzyme activity. In CN syndrome type II, bilirubin levels are 100 to 400 mmol/L and UGT1A1 activity is
