Portal de Periódicos da CAPES

Intrahepatic cholestasis: Summary of an American Association for the Study of Liver Diseases single-topic conference

2005; Lippincott Williams & Wilkins; Volume: 42; Issue: 1 Linguagem: Inglês

10.1002/hep.20729

1527-3350

William F. Balistreri, Jorge A. Bezerra, Peter L. M. Jansen, Saul J. Karpen, Benjamin L. Shneider, Frederick J. Suchy,

Liver Disease Diagnosis and Treatment

"The many, often fantastic, theories advanced to explain jaundice without obstruction have long stood as clear evidence of the need of precise information regarding the formation and excretion of the bile …." A. R. Rich, Bulletin Johns Hopkins Hosp 1930;47:338 An American Association for the Study of Liver Diseases (AASLD)-sponsored Pediatric Single-Topic Conference focused on the category of hepatobiliary disease collectively known as syndromes of intrahepatic cholestasis (September 2004). A single-topic conference that focused on this topic was unique and timely because recent advances in molecular genetics have clearly pointed out new directions for investigation and clinical understanding. These disorders, which are individually rare, are common collectively; they are chronic (by definition) and many cases progress to end-stage disease, requiring liver transplantation. More effective treatment options are needed. Prevention and treatment strategies targeted to this age group will be especially cost-effective. It was the goal of the single-topic conference to help to promote further investigation to identify the mechanisms responsible for this subset of diseases that begin during childhood and lead to ongoing liver dysfunction in children and adults. The ultimate objective was to foster research and collaborative ventures in this important area. The conference brought together leading experts from around the world for presentations that highlighted the newest investigative and therapeutic approaches; the discussions drew needed attention to areas requiring further research. The purpose of this report is to briefly summarize this important single-topic conference. AASLD, American Association for the Study of Liver Diseases; PFIC, progressive familial intrahepatic cholestasis; GCT, giant cell transformation; GGT, γ-glutamyltransferase; BSEP, bile salt export pump; OATP, organic anion transport protein; mRNA, messenger RNA; BRIC, benign recurrent intrahepatic cholestasis; HNF, hepatocyte nuclear factor; TGF, transforming growth factor; NR, nuclear receptor; LCA, lithocholic acid. To place the topic in perspective, William F. Balistreri (Cincinnati Children's Hospital Medical Center) emphasized that in the early 1970s, the differential diagnosis of the neonate with conjugated hyperbilirubinemia (cholestasis) was limited. Biliary atresia accounted for approximately 25% of the cases; a small percentage of cases were considered to be caused by viral infections or were the result of a handful of recognizable metabolic or inherited diseases (i.e.,galactosemia, tyrosinemia, cystic fibrosis).1 The vast majority were designated as idiopathic neonatal hepatitis, clearly a default diagnosis, because the underlying pathophysiology was unknown. What was clear was that although many cases were apparently sporadic, others recurred in families. This led to the speculation that cases of familial neonatal hepatitis represented presumed inborn errors or genetic defects in a fundamental process involved in hepatic metabolic or excretory function. It followed that elucidation of their nature would allow a more clear understanding of liver physiology and also to improved therapy. Indeed, over the past 30 years there has been significant progress in dissecting the components of the idiopathic neonatal hepatitis spectrum. Specifically, substantial progress has been made in subcategorizing the range of hepatobiliary diseases that manifest as intrahepatic cholestasis in infants, children, and even adults.2 A major category is that of genetic intrahepatic cholestasis, a heterogeneous subset of diseases that represents specific disorders of abnormal cell fate, canalicular transport (bile acid or phospholipid), or bile acid synthesis, each with different prognostic implications. There are multiple forms with varying clinical features and with a high degree of variability in presentation and prognosis. Certain progressive, familial forms, as in the group known as progressive familial intrahepatic cholestasis (PFIC), are fatal; however, in patients with syndromic paucity of the ducts (Alagille syndrome), the prognosis is more favorable.2 The current system of nomenclature for the various syndromes in which intrahepatic cholestasis is present is imperfect. A proposed classification scheme is shown in Table 1. These syndromes are of more than theoretical interest; detailed study of affected patients has enhanced our understanding of hepatic excretory function and bile acid metabolism. The pathogenetic basis for these diseases has been defined only partially, and the techniques of molecular genetics have been applied only recently to these disorders. Precise terminology of the intrahepatic cholestatic disorders, based on the documented genetic defect, not only will provide clues to the pathophysiology but will also help to establish registries, promote development of new treatments, and allow the institution of valid clinical therapeutic trials. Kevin E. Bove (Cincinnati Children's Hospital Medical Center) emphasized that these genetic defects result in lobular cholestasis, variable giant cell transformation (GCT) of liver cells, isolated hepatocyte necrosis, and a highly variable rate of progression to fibrosis. The degree of hepatocellular injury is least in presumed cases of FIC1 deficiency, intermediate in bile salt export pump (BSEP) deficiency, and most destructive in MDR3 deficiency. In routine practice, the diagnosis is tentative and dependent on clinical parameters (i.e., serum γ-glutamyltransferase [GGT] and bile acid level, phenotype), combined with liver biopsy findings using light and electron microscopy.2 He discussed the predominant liver morphological features of the various forms. In FIC1 deficiency, small tidy hepatocytes with minimal ballooning and bland canalicular cholestasis are present; GCT is uncommon; there is little or no cholangiolopathy (periportal ductular metaplasia) with slow progression of portal fibrosis; and an unusual coarse quality of bile in canaliculi is visualized by electron microscopy. In BSEP deficiency, prominent canalicular cholestasis is often zonal (zone 3 > zone 1); prominent clustered balloon cholestasis of hepatocytes is present; GCT is common and persists beyond infancy; perivenular, pericellular, and periportal fibrosis with progression to cirrhosis is present; mild cholangiolopathy is present; normal interlobular bile ducts are present; and delicate wispy nonspecific quality of bile in canaliculi by electron microscopy is evident. In MDR3 deficiency, there is cholestatic hepatitis with GCT in infants and young children, cholangiolopathy, and progressive injury of the interlobular bile ducts in the early stages of disease. In bile acid synthetic defects, Dr. Bove has noted generalized hepatocyte injury with GCT and periportal inflammatory fibrogenic cholangiolitis; the injury and progressive portal fibrosis with fusion or dysmitotic division of immature hepatocytes is attributed to atypical or toxic bile acids present in the liver in conjunction with reduced concentrations of normal bile acids.3 In Alagille syndrome, he has observed that the smallest ductules either fail to form or rapidly regress, with interlobular bile ducts subject to a smoldering cholangiopathy that results in focal proliferation and progressive destruction at a variable rate. The next session focused on potential mechanisms responsible for the clinical and histological features described above. The predominant mechanism in many forms of intrahepatic cholestasis is altered canalicular transport. The bile canaliculus is surrounded by the apical plasma membrane domains of two or three hepatocytes (Fig. 1).4, 5 The composition of the canalicular membrane is distinct from the basolateral membrane and is enriched in cholesterol and sphingomyelin. Canalicular membrane proteins are targeted to this domain via both direct and indirect pathways. The canalicular membrane is enriched in ABC transporters that function as export pumps for bile acids and a variety of organic solutes. James Boyer (Yale University School of Medicine) pointed out that this membrane domain is metabolically active because it contains a number of active ATP-dependent solute transport proteins as well as GPI-anchored proteins like dipeptidyl peptidase IV, GGT, aminopeptidase N, 5′-nucleotidase, and the protein HA-4. Ion and water channels, ion exchangers, skeletal proteins (villin), vesicle fusion proteins (e.g., SNARE, SNAP, Syntaxin, Rabs) and tight-junction proteins are also present. Most of the solute transport-proteins belong to the superfamily of ABC transporters. The BSEP (ABCB11) is one; BSEP mediates the transport of conjugated bile acids (cholic and chenodeoxycholic acid). Like all other plasma membrane proteins, BSEP is synthesized in the endoplasmic reticulum and after glycosylation in the trans-Golgi network travels through the subapical compartment to the canalicular membrane. There is evidence for an intracellular pool of BSEP protein from which BSEP may be recruited on demand. For retraction and internalization of BSEP, interaction with a protein called HAX-1 seems to be important.4 Other apical ABC transporters include MRP2 (ABCC2), which transports glucuronide and glutathione conjugates; MDR3 (ABCB4), a flippase of phosphatidylcholine; MDR1 (ABCB1), a transporter of chemotherapeutic agents; ABCG5G8, a heterodimeric protein that transports sterols; and BCRP breast cancer resistance protein (ABCG2), a transporter of chemotherapeutic agents, carcinogens, and natural products.5-8 Mutations in several of these ABC transporters (MDR3, BSEP, and MRP2) result in clinical disease. The bile canalicular and the major hepatobiliary transport systems. The basolateral membrane of the hepatocyte contains the solute carrier systems (organic anion transport proteins [OATPs]), whereas the ATP-dependent ABC transporters are mainly located in the canalicular membrane. During cholestasis, multidrug resistance proteins MRP3 and MRP4, two organic anion transporting systems, become expressed at the basolateral membrane. MRP3 mainly transports conjugates such as bilirubin glucuronide; MRP4 transports bile acids together with glutathione. Transport proteins, water and electrolyte channels, and exchangers at the bile duct epithelium indicate that bile composition is modified at this level. The function of transporters in the basolateral membrane is quite different from those at the apical membrane. Uptake from the portal blood into the liver does not require active transport; electroneutrality is sufficient. Some of these transporters are exchangers that mediate organic anion uptake in exchange for bicarbonate or glutathione. These proteins belong to the superfamily of solute carriers (organic anion transport proteins [OATPs]) named OATP-1 (Slc21a1), OATP-2 (Slc21a5), OATP-4 (Slc21a6), and OATP-8 (Slc21a8). They have broad and partly overlapping substrate specificity. NTCP (Slc10a1) is an exception; it transports only bile acids and is a sodium–bile acid cotransporting protein.7 Solute carriers reach the basolateral membrane via a direct sorting route from the trans-Golgi. Their expression is dependent on transcriptional and post-transcriptional regulatory mechanisms (Fig. 2). MRP3 (Abcc3) and MRP4 (Abcc4) in the basolateral membrane act as efflux pumps when canalicular secretion is impaired. MRP4 is a cotransporter of bile acids and glutathione, whereas MRP3 mainly transports conjugates such as bilirubin monoglucuronide and diglucuronide. These transporters have a low expression in normal liver but are upregulated in cholestatic conditions.9, 10 Much remains to be learned about the mechanisms of canalicular membrane formation in development and its maintenance in health and disease. Post-transcriptional regulation of canalicular transporters. Apical transport proteins cycle between the apical membrane and a subapical compartment (SAC). On hypoosmotic conditions and/or stimulation by cyclic-AMP, protein kinase C, phosphatidylinositol-3-kinase, or MAP-kinase, transport proteins are inserted into the apical membrane. On hyperosmotic conditions, LPS administration, or bile duct ligation, they are retrieved into a SAC. The regular half-life of a transport protein is governed by proteosomal degradation after ubiquination. Drugs may influence the activity of transporters by direct interference with their activity. They can do this from the cytoplasmic side (cis-inhibition) or from the canalicular lumenal side (trans-inhibition). Modified from Trauner and Boyer.5 Frederick J. Suchy (Mount Sinai Medical Center, New York) discussed the ontogeny and regulation of these bile acid transporters. At birth, hepatic excretory function is immature, the enterohepatic circulation is inefficient, and bile flow is reduced; immediately after birth, serum bile acid levels are elevated.11 Ntcp and Bsep messenger RNA (mRNA) expression is not fully developed until 1 week after birth (rats). Ntcp protein expression closely follows this mRNA expression. Bsep protein expression is somewhat delayed and is not fully developed until 4 weeks after birth. The same holds true for Mrp2 mRNA and protein.12 The genes encoding these proteins are under transcriptional control of the nuclear hormone receptors RXR, FXR, PXR, CAR, RAR, LXR, and LRH. These are transcription factors that translocate from the cytoplasm to the nucleus on binding a relevant ligand. Bile acids bind to FXR, drugs to CAR and PXR, retinoic acid to RARα,β,γ, and oxysterols to LXRα,β. LRH-1 is an orphan receptor with no known ligand. 9-Cis-retinoic acid binds to RXRα,β,γ. In the nucleus, they interact as heterodimers with palindromic response elements in the promotor regions of target genes. RXRα,β,γ are obligate partners.13 What has been said about the ontogeny of Ntcp, Bsep, and Mrp2 is also true for the nuclear hormone receptors; their expression is significantly less than adult values until 1 week after birth.14 Because these receptors play an important role in protecting the liver against substrate overload, it holds that in the first week of life, the neonate is particularly susceptible to cholestasis and liver damage.2 The exact mechanisms governing the ontogeny of transport proteins and metabolizing enzymes is not known. A complex interaction of nuclear hormone receptors and other transcription factors is likely to play a role. It is clear that nuclear hormone receptors do not act alone but are part of a complex of coactivating proteins. Coactivator-associated arginine methyltransferase (CARM1) acts as a coactivator of FXR, activating BSEP. FXR:RXR and CARM1 are needed for full expression of BSEP mRNA in Hep G2 cells.15 Given the complexity of the structure and function of the bile canaliculus and hepatic excretory systems, it has long been postulated that specific inborn molecular defects could lead to cholestasis. A large number of genes mutated in forms of intrahepatic cholestasis have been identified (Table 2)16-40; however, many as-yet unidentified cholestasis genes likely exist. Laura Bull (University of California San Francisco, San Francisco General Hospital) described genetic approaches to decipher disorders such as intrahepatic cholestasis. These approaches complement functional studies of hereditary diseases. Bull emphasized that a key feature of genetic mapping studies is that they permit identification of the genetic etiology of a disease, even if the identity of the disease gene could not have been predicted. Thus, such studies can facilitate our understanding of disease etiology. For example, it is unlikely that FIC1/ATP8B1 would have been easily identified as a cholestasis gene using a functional, rather than whole-genome screening, approach. Bull has carried out genetic mapping studies to sift through the approximately 3.2 gigabasepairs of DNA in a single copy of the genome to identify mutations that result in disease.41 Roderick Houwen (Wilhelmina Children's Hospital, Utrecht, The Netherlands) discussed the clinical relevance of FIC1 (ATP8B1), the gene involved in various forms of intrahepatic cholestasis associated with FIC1 deficiency (PFIC type 1, Byler's disease, Greenland familial cholestasis, and benign recurrent intrahepatic cholestasis [BRIC] type 1).42, 43 Both PFIC type 1 and BRIC type 1 are autosomal recessive diseases. Nonsense, frameshift, and deletional mutations cause PFIC type 1, whereas missense and splice site mutations cause BRIC type 1. However, identical mutations can cause both PFIC type 1 and BRIC type 1. Houwen emphasized that homozygosity for the common I661T mutation can also occur in healthy persons. This indicates that modifier genes and environmental influences play a role in the expression of PFIC type 1 and BRIC type 1. The pathophysiology of these diseases is not well understood. FIC1 encodes an aminophospholipid translocase that has been proposed to flip phosphatidylserine and phosphatidylethanolamine from the outer to the inner layer of the plasma membrane. How a defect of this protein leads to cholestasis is unknown, and studies carried out on FIC1 knockout mice have failed to elucidate the mechanism.44 FIC1 protein expression in the intestine exceeds that in the liver, and therefore defects in both intestine and liver may be involved in the pathogenesis of these diseases. Increased intestinal absorption of bile acids, decreased canalicular secretion of bile acids, or both may be implicated. Partial biliary diversion, in which the gallbladder is externalized via a stoma using a loop of small intestine, is of clinical benefit in approximately half of the patients with PFIC type 1.45 The bile acid pool is reduced via diversion of bile, but how this leads to amelioration of cholestasis is unknown. Richard Thompson (Guy's, King's & St Thomas' School of Medicine, London) discussed BSEP (ABCB11), the major canalicular bile acid transporter. Mutations of the BSEP gene lead to deficiency of canalicular BSEP expression and variable degrees of intrahepatic cholestasis (also known as PFIC type 2 and BRIC type 2). As in FIC1 deficiency (PFIC type 1 and BRIC type 1), serum GGT activity in these patients is not elevated despite the cholestasis. A great number of mutations have been described in patients with PFIC type 2.18, 46 As in PFIC type 1, severe mutations lead to a phenotype of unremitting cholestasis from birth onward. Missense mutations have been shown to cause a milder disease not unlike BRIC; follow-up studies of this group of patients has not been long enough to know if this disease runs the same benign course as BRIC type 1. Thompson also reported eight patients with PFIC type 2 in whom malignancy developed: five with hepatocellular carcinoma, one with hepatoblastoma, and two with cholangiocarcinoma. The pathogenesis of PFIC type 2 seems clear; deficiency of the major canalicular bile salt transporter leads to severe cholestasis. Therefore, it was surprising that Bsep knockout mice are viable and in fact have a very mild phenotype.47, 48 Studies on these mice made it clear that mice have compensatory pathways to eliminate or detoxify bile acids, or both; for example, tetrahydroxy bile acids are produced that can be eliminated via alternative routes. Humans have the ability to upregulate the expression of basolateral MRP4 as a hepatic escape mechanism. Treatment of PFIC type 2 is mainly supportive, followed by liver transplantation in many cases. The role of partial biliary diversion has to be established; some patients respond well to this therapy, particularly those with the D482G or E297G mutation.20, 42 Ronald Oude Elferink (Academic Medical Center, Amsterdam, The Netherlands) was involved in the discovery of MDR3/Mdr2 (ABCB4/Abcb4) as a phospholipid translocase that flips phosphatidylcholine from the inner to the outer layer of the canalicular membrane.23, 49 Mutations of the MDR3 gene lead to the formation of phosphatidylcholine-poor bile with a normal bile acid concentration. The high content of bile acids in the absence of phospholipid may make this bile toxic to hepatocytes and to the bile duct epithelium. Patients and mice with a defective canalicular MDR3/Mdr2 expression show severe hepatocyte and bile duct damage, periductular inflammation, portal and periportal fibrosis, and eventually (in mice) bile duct carcinoma. The phenotypic variation of this disease is remarkable and includes intrahepatic gallstone formation, biliary fibrosis, and intrahepatic cholestasis of pregnancy. In humans, the cholestatic syndrome that is associated with MDR3 mutations is called MDR3 deficiency or PFIC type 3. In contrast to PFIC type 1 and PFIC type 2, PFIC type 3 is characterized by increased serum GGT levels. A number of PFIC type 3 patients, presumably with partial residual MDR3 function, respond well to ursodeoxycholate therapy.50 Certain forms of intrahepatic cholestasis have been ascribed to altered embryogenesis; the mechanisms and features of these developmental abnormalities of the biliary system were discussed. The most common cause of chronic cholestasis during childhood associated with specific phenotypic features is the Alagille syndrome. There is wide variability in the extent of phenotypic expression; affected patients may present with any combination of five major features: (1) chronic cholestasis resulting from a paucity of interlobular bile ducts, (2) cardiovascular malformations, such as peripheral pulmonic stenosis or more severe lesions as in tetralogy of Fallot, (3) vertebral arch defects, (4) posterior embryotoxon and other ocular abnormalities, and (5) unique facial features.51 In addition, the kidney, pancreas, and cerebrovascular system may be involved. The disease is often manifest as neonatal cholestasis, with pruritus and xanthomas becoming prominent symptoms in later phases of disease; hepatic and cardiac manifestations are responsible for most morbidity and mortality. The genetic defect in patients with the Alagille syndrome has been shown to reside in mutations of the Jagged-1 gene.27, 28 Nancy Spinner (The Children's Hospital of Philadelphia) reviewed how mutations in the Jagged-1 gene correlate with the variable manifestations in patients with the Alagille syndrome. Since the initial identification of mutations capable of disrupting the product of Jagged-1 in affected patients,27, 28 several studies have explored the extent to which mutations segregate with disease phenotypes. Jagged-1 belongs to the Jagged family of genes that encode cell surface proteins that interact with Notch receptors, generating signals to regulate cell fate during embryogenesis. To date, genetic screening techniques have detected Jagged-1 mutations in 60% to 70% of individuals with hepatic, cardiovascular, and abnormalities in other systems.52, 53 However, higher detection rates are now possible. These mutations are spread out across the coding region, and include total gene deletions, protein truncating mutations (caused by nonsense, insertion, and deletion mutations), splicing mutations, and missense mutations. Mutations are de novo in more than 50% of the cases; although most mutations occur in the extracellular domain of the JAG1 protein, there is no predominant mutational "hot spot," making it difficult to routinely use genetic testing in the evaluation of patients with the Alagille syndrome. At the molecular level, haploinsufficiency (decreased gene dosage) seems to be the mechanism of clinically relevant mutations. Notably, missense mutations, which account for 10% to 15% of all mutations, lead to intracellular trafficking defects, with trapping of the mutant protein in the perinuclear region, thus preventing the proper membrane anchoring at the cellular surface. Despite the greater understanding of the molecular consequences of mutations, the search for phenotype-causing mutations has been elusive, with no correlation between mutations and the phenotypic expression in patients with the Alagille syndrome. One exception is the high association of the JAG1-G274D mutation with a cardiac specific phenotype, supporting the concept that specific Jagged-1 mutations may have greater penetrance in the developing heart.54, 55 Specific Jagged-1 mutations that segregate exclusively with hepatic abnormalities have not yet been reported. This notwithstanding, an important role for the JAG–Notch pathway in the hepatobiliary system is supported by findings of abnormal morphogenesis of the intrahepatic bile ducts in experimental models. For example, mice with dual heterozygosity for the Jagged-1 gene and of its receptor Notch-2 exhibit a phenotype akin to the Alagille syndrome.56 Interestingly, inhibition of gene expression leading to interruption in the JAG-mediated Notch signaling was recently shown to also result in abnormal development of the intrahepatic biliary system.57 Frederick Lemaigre (Universite Catholique de Louvain, Brussels, Belgium) and Ai-Xuan Holterman (University of Illinois at Chicago) reviewed recent data identifying additional molecular networks that play key regulatory roles in the morphogenesis of the biliary system. Abnormalities in morphogenesis restricted to intrahepatic bile ducts (ductal plate) have been demonstrated in mice with disrupted JAG-Notch signaling, as discussed above.56, 57 In contrast, developmental abnormalities restricted to the extrahepatic biliary system have been found in Hes1-deficient mice.58 Hes1 encodes the basic helix-loop-helix protein Hes1, which is expressed in the extrahepatic biliary epithelium throughout development. Interestingly, Hes1 expression is controlled by the Notch pathway; however, in contrast to the predominantly intrahepatic phenotype when the JAG-Notch pathway is disrupted, in vivo loss of Hes1 results in gallbladder agenesis and severe hypoplasia of extrahepatic bile ducts. Despite the distinct endodermal origin of the intrahepatic and extrahepatic biliary systems, the molecular forces driving tissue embryogenesis are shared, at least in part, by common pathways, as demonstrated by the phenotypic analysis of mice carrying the inactivation of genes encoding members of two groups of transcription factors. The first is the hepatocyte nuclear factors (HNF)-6 and HNF1β. HNF6, a member of the onecut family of homeoproteins, is expressed in intrahepatic biliary cells and gallbladder primordium. When the HNF6 gene was inactivated in mice, development of the ductal plate was impaired, biliary cysts developed, the gallbladder did not form, and extrahepatic bile ducts were replaced by an enlarged tubular structure that connected the liver to the duodenum.59 HNF6 is known to control the expression of HNF1β. The functional relationship between these factors and relevance to biliary development were further established by the findings of paucity of small intrahepatic bile ducts, dysplasia of larger intrahepatic bile ducts, and abnormal gallbladder and cystic duct in mice lacking HNF1β.60 These findings led to the proposal that the HNF6 → HNF1β pathway is essential for biliary development. It is likely that the application of similar experimental strategies to other transcription factors that are functionally related and expressed in the developing biliary system, such as the factor Onecut-2, will further define how HNFs work in concert to promote biliary development. The second group of transcription factors is the Forkhead Box (Fox) family, known to participate in a broad range of physiological processes such as proliferation, differentiation, transformation, metabolism, and development. In the hepatobiliary system, the factor Foxf1 is expressed in stellate cells and in mesenchymal cells surrounding the intrahepatic and extrahepatic bile ducts. The importance of this expression pattern to biliary development became apparent in gene targeting experiments. Although homozygous inactivation of the Foxf1 gene was embryonic lethal owing to impaired extraembryonic tissue maturation, some heterozygous littermates completed gestation and were later found to have normal intrahepatic bile ducts, but diminutive or absent gallbladder.61 Conversely, targeting of the related gene Foxm1b resulted in impaired development of intrahepatic bile ducts, possibly because of the inability of embryonic hepatoblasts to differentiate toward biliary epithelial cell lineage.62 Collectively, the ability to inactivate genes in vivo and at different phases of embryonic development has facilitated the initial dissection of the molecular circuits controlling morphogenesis of the biliary system (Table 3). The implication of an increased understanding of the molecular basis of biliary development to clinical practice is highlighted by the direct relationship of the JAG-Notch circuit with the Alagille syndrome. In the future, analysis of these molecules in children with syndromes of intrahepatic and extrahepatic cholestasis will explore their potential role as cause or modifiers of disease pathogenesis. Further understanding of the disorders of intrahepatic cholestasis will require insight from global genomics and phenotypic expression. Jorge Bezerra (Cincinnati Children's Hospital Medical Center) discussed the use of molecular profiling through gene microarrays to study liver biology and pathobiology in specific clinical phenotypes. To date, more than 400 such studies have focused on subjects such as liver regeneration, development, carcinogenesis, inflammation, metabolism, and viral hepatitis. Experimental studies have also been carried out on the topic of cholestasis in both humans and animal models. Although this experimental approach has extraordinary potential in elucidating the genetic basis of disease as well as the transcriptional regulation of biological processes, it was emphasized that very sophisticated bioinformatic support is essential to analyze the very lar