Intrahepatic cholestasis of pregnancy (ICP) is a pregnancy-specific disorder characterised by skin pruritus and mild to moderate biochemical cholestasis, which appears mainly in the late-second or third trimester of gestation and rapidly resolves after delivery (Refs Reference Arrese and Reyes1, Reference Reyes2, Reference Lammert, Marschall and Matern3). ICP is associated with increased perinatal morbidity and mortality, which are mainly due to unexplained foetal death and premature delivery, and patients are considered to have a high-risk pregnancy (Refs Reference Rioseco4, Reference Fisk and Storey5, Reference Glantz, Marschall and Mattsson6). The aetiopathogenesis of ICP is not completely understood, although it is believed to be multifactorial, with an interaction of genetic, hormonal and environmental factors (Ref. Reference Lammert, Marschall and Matern3). Recent advances in the understanding of hepatobiliary transport mechanisms and their alterations in experimental settings and some forms of human cholestasis (Refs Reference Boyer7, Reference Wagner and Trauner8) has led to the development of new hypotheses of the pathogenesis of ICP and other forms of cholestasis presenting during pregnancy (Refs Reference Arrese and Reyes1, Reference Arrese9). This review summarises current knowledge of the potential mechanisms involved in ICP at the molecular level. General concepts of bile secretion are discussed, followed by a description of the effects of pregnancy and hormones on hepatic transport function, as well as the effects of the genetic defects identified in patients with ICP.
Hepatobiliary transport and cholestasis: a brief overview
Basic concepts of bile formation
Bile formation is a vital function of the liver, which contributes to the absorption of lipids from the intestinal lumen and also serves as an important excretory route for many exogenous and endogenous compounds (Refs Reference Hofmann10, Reference Arrese and Trauner11). Bile is produced as a consequence of an osmotic filtration process resulting from the active transport of osmotically active compounds into the canalicular space by hepatocytes, which is followed by the passive movement of water through tight junctions. Bile salts are the main solutes in bile and are thought to be the major osmotic driving force in the generation of bile flow, although bile-salt-independent processes also contribute to bile production (Ref. Reference Arrese, Ananthananarayanan and Suchy12). Bile salts and other biliary constituents are transported from the blood to bile in a vectorial process that includes uptake from sinusoidal blood, intracellular transport through the hepatocyte – with or without the occurrence of metabolic biotransformation – and canalicular secretion against concentration gradients, all of which require energy-dependent active transport (Ref. Reference Arrese and Trauner11).
Once the bile is generated at the canalicular level, it undergoes further modification along the biliary tree as a result of the active secretion of bicarbonate, and is then reabsorbed and secreted by biliary epithelial cells (Ref. Reference Banales, Prieto and Medina13). Finally, after reaching the intestine, most bile salts undergo enterohepatic recirculation via the active transport mechanism located in the apical pole of enterocytes at the terminal ileum (Ref. Reference Shneider14). This allows bile salt retrieval from the intestinal lumen to the portal circulation and to the liver for uptake and re-secretion.
Molecular basis of hepatobiliary transport
In the past decade, the molecular identity of distinct hepatic transport systems involved in the hepatic handling of bile salts and cholephilic organic anions has been elucidated, and the physiology of bile secretion is therefore now better defined. Figure 1 shows the main transport systems of the hepatocyte and the cholangiocyte. Some of these transporters are also expressed in tissues other than the liver, where they play specific physiological roles. The following sections briefly summarise current information on the transport systems for cholephilic compounds of the hepatocyte. For further information on this topic, the reader is referred to several recent comprehensive reviews (Refs Reference Hofmann10, Reference Geier15, Reference Trauner and Boyer16).
Hepatobiliary transport proteins involved in bile formation. Transporters depicted in blue correspond to proteins belonging to the solute carrier (SLC) family, whereas transporters depicted in pink belong to the ATP-binding cassette (ABC) transporter family. The ‘half’ ABC transporter is depicted as a semi-circle. The hepatic Na+/K+ ATPase of the liver cell is represented as a green circle. The Na+ taurocholate co-transporter (NTCP, SLC10A1) at the basolateral membrane of hepatocytes mediates bile salt (BS) uptake, whereas Na+-independent organic anion transporting polypeptides (OATPs) transport not only bile salts but also various organic anions (OA). Organic cations (OC) are transported by an OC transporter (OCT), and certain organic anions (e.g. drugs) are transported by the organic anion transporters (OATs). Bile salts are excreted via a canalicular bile salt export pump (BSEP, ABCB11), whereas glutathione (GSH) and organic anions (e.g. conjugated bilirubin) are excreted via a canalicular conjugate export pump [multidrug-resistance-associated protein 2 (MRP2, ABCC2)]. In addition, the canalicular membrane contains export pumps for organic cations [multidrug-resistance 1 (MDR1)], phospholipids (PL) (MDR3 or ABCB4 in humans, Mdr2 in rodents) and sulphate conjugates [breast-cancer-related protein (BRCP, ABCG2)]. The twinned canalicular half-transporters ABCG5 and ABCG8 transport the majority of sterols into bile and play an important role in the regulation of biliary cholesterol secretion. MATE1 and aquaporin 8 (AQP8) facilitate the transport of cations and canalicular secretion of water, respectively. MRP3 (ABCC3) and MRP4 (ABCC4) are expressed in the basolateral domain of hepatocytes and may facilitate extrusion of biliary constituents as they accumulate in the cytoplasm. Although MRP3 can transport organic compounds conjugated to glutathione, sulphate or glucuronate, MRP4 extrudes sulphates and bile salts from the liver, but does not transport glucuronides. The heterodimeric organic solute transporter OSTα/β also participates in the extrusion of bile salts and other organic anions to the portal blood. The familial intrahepatic cholestasis 1 (FIC1) protein is a P-type ATPase and functions as a flippase for phosphatidylserine within the canalicular membrane and maintains phospholipid membrane symmetry, which influences bile salt transport. Mutations of the gene encoding this transporter (ATPB8B1) are present in certain forms of genetic cholestasis. For details on transporter nomenclature see the text. Abbreviation: SBS, sulphated bile salts.
Sinusoidal uptake of biliary solutes
The sinusoidal membrane of hepatocytes contains a number of transport systems that mediate the entry of bile salts and organic substances – including physiological substrates, drugs and exogenous compounds – into the liver. Bile salts are taken up from portal blood at the basolateral plasma membrane of the hepatocytes, mainly via a sodium-dependent mechanism mediated by the Na+/taurocholate-cotransporting polypeptide (NTCP, official GenBank gene symbol SLC10A1). NTCP is strictly localised on the basolateral membrane of hepatocytes, is exclusively expressed in the liver and is probably the predominant Na+-dependent bile salt transporter of the hepatocyte as it provides more than 80% of conjugated bile salt uptake (Ref. Reference Hagenbuch and Dawson17). NTCP transports bile salts by taking advantage of the transmembrane sodium gradient, which is maintained by the Na+/K+ ATPase of the hepatocyte. Sinusoidal Na+-independent transport of bile salts and organic anions is mediated by several members of a superfamily of polyspecific transporters called organic anion transporting polypeptides [OATPs or solute-carrier organic anion transporter family (SLCO)] (Ref. Reference Hagenbuch and Meier18). OATPs exhibit overlapping transport activities for a number of compounds including some bilirubin conjugates, thyroid hormones, and numerous drugs and other xenobiotics. OATP isoforms present in the liver and involved in bile formation include OATP-C/1B1 (SLCO1B1) and OATP8/1B3 (SLCO1B3). These are predominantly expressed in the liver. Recent evidence suggests that although OATP1B1 may be involved in uptake processes, OATP8/1B3 may additionally mediate the extrusion of organic anions (Ref. Reference Briz19). Thus, some OATPs might work as bidirectional transporters at high intracellular substrate concentrations, acting as an extrusion system of potentially dangerous compounds when necessary. Several OATPs are expressed in extrahepatic tissues, particularly in the intestine, kidney, brain and placenta, underscoring their role in the removal of amphipatic compounds (Ref. Reference Hagenbuch and Meier18).
Additional transport proteins, such as OAT2 [the liver homologue of the kidney organic anion transporters (OATs)] and OCT1 (a polyspecific organic cation transporter), which belong to the organic anion and organic cation transporter family of solute carriers (SLC22), have been recently identified at the sinusoidal domain of hepatocytes (Ref. Reference Koepsell and Endou20). These transporters may play a role in drug handling by the liver but their exact role in hepatic uptake of foreign compounds remains to be established.
Intracellular transport of cholephilic compounds
Conjugated bile salts are rapidly transported to the canalicular pole of hepatocytes and are probably bound to intracellular transport proteins (Ref. Reference Agellon and Torchia21). By contrast, some other biliary solutes undergo chemical modification during their transcellular residence. This consists of oxidation/reduction reactions (phase I) and/or conjugation with polyatomic groups (phase II). Phase I reactions are mainly catalysed by the cytochrome P450 enzymes, whereas phase II reactions are mediated by uridine glucuronyl transferases and sulphotransferases. Reactions in phase II increase the molecular weight and water solubility of a given compound, which facilitates its excretion in bile and urine. Sulphation, glucuronidation and glutathione conjugation represent the three most prevalent classes of phase II metabolism, whereas other phase II reactions (e.g. amino acid conjugation, acetylation and methylation) are less-common metabolic pathways for xenobiotics (Ref. Reference Zamek-Gliszczynski22).
Canalicular transport of biliary solutes
Several unidirectional ATP-dependent export pumps known as ABC (ATP-binding cassette) transporters are key proteins in the secretion of cholephilic compounds across the canalicular membrane of the hepatocyte. ABC transporters known to be present in the canalicular membrane include: (Reference Arrese and Reyes1) the bile salt export pump (BSEP; official gene symbol ABCB11), which mediates bile salt secretion into the canaliculus (Refs Reference Stieger, Meier and Meier23, Reference Arrese and Ananthanarayanan24); (Reference Reyes2) multidrug-resistant-associated protein 2 (MRP2; ABCC2), which transports anionic conjugates of many lipophilic substances and reduced glutathione, providing the main driving force for bile-salt-independent bile flow (Ref. Reference Nies and Keppler25); (Reference Lammert, Marschall and Matern3) multidrug-resistant 1 (MDR1; ABCB1) gene product (P-glycoprotein), which acts as a transporter of bulky cationic compounds and steroids and is considered one of the major determinants of drug disposition and toxicity (Ref. Reference Trauner26); and (Reference Rioseco4) MDR3 (ABCB4) gene product, which acts as a phosphatidylcholine translocator from the inner to the outer leaflet of the canalicular membrane (Ref. Reference Trauner, Fickert and Wagner27). Three additional ‘half’ ABC transporters – which contain six rather than the twelve transmembrane domains typical of ABC transporters and form homo- or heterodimers as functional entities – are present in the canalicular membrane of hepatocytes: the twinned sterol half-transporters ABCG5 and ABCG8, which play an important role in the regulation of biliary cholesterol secretion (Refs Reference Oude Elferink, Paulusma and Groen28, Reference Kusuhara and Sugiyama29, Reference Hazard and Patel30) and the breast-cancer-resistance protein BCRP (ABCG2), a polymorphic efflux transporter protein that is able to transport sulphated bile-salt conjugates and other cytotoxic compounds (Ref. Reference Hardwick, Velamakanni and van Veen31). Additional transporters in the canalicular membrane include anion exchanger 2 (AE2 or SLC4A2), which excretes bicarbonate into bile (Ref. Reference Banales, Prieto and Medina13) and a polyspecific organic cation exporter, MATE1 (multidrug and toxin extrusion 1), which mediates H+-coupled electroneutral excretion of toxic organic cations (Ref. Reference Otsuka32). In addition, there are several integral membrane channel proteins or aquaporins (AQPs), which facilitate rapid passive movement of water, are present in the canalicular membrane and are thought to contribute to the bile-salt-independent fraction of bile flow (Ref. Reference Calamita33). Finally, a canalicular protein named FIC1 (familial intrahepatic cholestasis 1) encoded by ATP8B1 is a putative flippase for phosphatidylserine, which seems to play an important role in bile secretion by regulating transporter function via an undefined mechanism (Refs Reference van Mil34, Reference Eppens35, Reference Paulusma36).
Basolateral efflux of biliary constituents
The basolateral hepatocyte membrane localises additional ABC transporters that have been implicated in the cellular efflux of several compounds, such as bile salts and compounds conjugated with glutathione, glucuronic acid or sulphate (Refs Reference Pauli-Magnus and Meier37, Reference Kruh38). These transporters belong to the MRP family and include MRP1 (official gene symbol ABCC1), MRP3 (ABCC3), MRP4 (ABCC4), MRP5 (ABCC5) and MRP6 (ABCC6). Most of these transporters, with the exception of MRP4, exhibit low levels of expression in the liver (Refs Reference Kruh38, Reference Borst, de Wolf and van de Wetering39). Moreover, other transporters not directly activated by ATP are also located at the basolateral membrane of hepatocytes. This is the case for the heterodimeric protein organic solute transporter OSTα/β, which transports oestrone 3-sulphate, dehydroepiandrosterone sulphate, taurocholate, digoxin and prostaglandin E2 (Ref. Reference Ballatori40). It is thought that these transporters might favour the elimination of potentially toxic substances from liver cells.
Transport systems in cholangiocytes
Bile may be significantly modified by an array of absorptive mechanisms on the apical membrane of cholangiocytes. A number of specific transport proteins are located in biliary epithelia, including anion exchangers and water transporters (for a review see ref. Reference Glaser41).
Cholestasis: transporter changes and adaptive responses
Cholestasis is a condition defined by the impairment of normal bile flow resulting in the hepatic and systemic retention of products normally excreted into bile (Refs Reference Boyer7, Reference Elferink42). The underlying causes of cholestasis include a broad range of disorders ranging from functional defects at the level of the hepatocyte (due to drugs, genetic disorders or cytokine/toxin-induced impairment of bile secretion) or a mechanical obstruction at the bile duct level (due to tumours, stones and bile-duct-destructing diseases, such as primary biliary cirrhosis and sclerosing cholangitis). The increased knowledge of hepatobiliary transport mechanisms summarised in the previous sections has led to major advances in the understanding of the molecular events occurring in the setting of cholestasis (Ref. Reference Wagner and Trauner8). Thus, data generated from patients with genetic cholestasis, as well as from animal models of cholestasis and genetically engineered mice, have led to the following general concepts: (Reference Arrese and Reyes1) altered expression and/or function of membrane transporters underlie some forms of cholestasis providing a ‘proof of principle’ that a defective transporter may directly cause cholestasis; (Reference Reyes2) during cholestasis, accumulation of toxic compounds normally excreted into bile triggers adaptive anticholestatic responses, aiming to limit hepatocellular accumulation of potentially toxic biliary constituents, promoting detoxification of these substances and providing alternative excretory routes for accumulating cholephilic substances in cholestasis (Refs Reference Boyer7, Reference Wagner and Trauner8). Several alterations in transporter expression have been documented, which are mainly secondary (rather than primary or causative), but nevertheless, may explain and maintain the ongoing functional impairment of bile secretion in cholestasis (Ref. Reference Geier15); (Reference Lammert, Marschall and Matern3) a set of several ligand-activated transcription factors belonging to the family of nuclear receptors regulates the expression of hepatobiliary transport proteins and some detoxification enzymes in hepatocytes, and is critical in orchestrating the adaptive responses seen in cholestasis (Ref. Reference Zollner43). The main nuclear receptors involved in these responses are the farnesoid X receptor FXR (NR1H4), the steroid X receptor (SXR) (or pregnane X receptor (PXR) in rodents) (NR1I2), the retinoid X receptor (RXR; NR2B1), the retinoid acid receptor (RAR; NR1B), the liver X receptor (LXRα and LXRβ, encoded by NR1H3 and NR1H2, respectively), the constitutive androstane receptor CAR (NR1I3) and the small heterodimer partner (SHP; NR0B2) amongst others. These transcription factors act in a complex network, regulating hepatic bile salts and other organic solute uptake, as well as bile-salt synthesis and detoxification (Refs Reference Boyer7, Reference Zollner43).
Effect of pregnancy on bile secretory function
Pregnancy is a physiological state known to affect bile secretory function (Ref. Reference Reyes and Simon44). Experimental studies carried out in pregnant rodents have consistently found a decreased uptake of bile acids (Refs Reference Ganguly, Hyde and Vore45, Reference Ganguly46), as well as reduced maximal transport capacity of organic anions in the liver (Ref. Reference Arrese47). In addition, reduced uptake of several steroid glucuronides and biliary excretion of many organic anions, such as dibromosulphthalein and glucuronide conjugates of drugs have been reported to be significantly decreased in pregnant rats (Ref. Reference Auansakul and Vore48). Although the available data suggest that pregnancy affects bile secretory function in the majority of species studied, some differences may exist, in both the intensity and nature of the pregnancy-induced alterations. For example, decreased bile flow has been shown in pregnant hamsters (Ref. Reference Reyes and Kern49), but not pregnant rats (Ref. Reference Reyes50). Earlier studies performed in humans have also shown a reduced clearance of bromosulphthalein (Ref. Reference Combes51) and a decrease in the fractional turnover rate of primary bile acids (Ref. Reference Kern52).
The precise mechanisms underlying pregnancy-induced changes of bile secretory function are not known. Recent studies have assessed the expression of some hepatobiliary transporters in the pregnant rat (Refs Reference Ganguly46, Reference Arrese47, Reference Cao53). In one study, a significant downregulation of mRNA and protein levels of the principal bile salt importer, Ntcp (Slc10a1) and reduced Mrp2 (Abcc2) protein levels were seen during late pregnancy, whereas no changes were seen in the expression of Oatp1 or the canalicular bile salt export pump Bsep (Abcb11) (Ref. Reference Arrese47). Another study showed unchanged Ntcp levels during pregnancy and a mild reduction of Oatp2 and downregulation of Mrp2 and Mrp3 (Ref. Reference Cao53). Mechanistic data indicate that pregnancy seems to affect transcription of Ntcp and protein degradation of Mrp2 in rats (Refs Reference Arrese47, Reference Jones54). Assessment of bile secretory function showed decreased hepatic bile salt transport capacity (Bsep function) and a reduction of biliary excretion of 2,4-dinitrophenyl-glutathione during pregnancy (Refs Reference Arrese47, Reference Cao53). Although this indicates a disturbed canalicular transport function, no changes have been seen in the sinusoidal export pump Mrp4 (Ref. Reference Geier15), whereas levels of Mrp3, whose expression is already low, is further reduced during pregnancy (Ref. Reference Geier15). Collectively, these findings suggest that pregnancy in rodents induces a significant perturbation of bile secretory function. Ntcp downregulation may explain the elevated levels of bile salts (Refs Reference Castano55, Reference Heikkinen56), and decreased Mrp2 levels might contribute to the reduced clearance of bromosulphthalein observed during pregnancy (Ref. Reference Combes51). Thus, if similar changes in hepatic transporter expression occur in humans, they might explain the functional alterations described in pregnant women and would support the concept that pregnancy is a physiological cholestatic-prone condition. This concept is in line with the existence of a relatively frequent (~10% of normal pregnancies) condition seen in pregnant women, called asymptomatic hypercholanaemia of pregnancy (AHP) (Ref. Reference Pascual57). AHP results in subclinical increased serum bile salt levels in the third trimester and is thought to reflect a continuum between normal pregnancy and ICP.
The mechanisms of the molecular alterations highlighted above are not well understood, but are probably related to the increased levels of sex hormones, mainly oestrogen, during pregnancy. This can be inferred from the similarities of the observed changes and those seen in the experimental model of ethinyloestradiol-induced cholestasis in rats (see below).
Pathogenetic factors in ICP: sex hormones, genes and environment
As mentioned above, the cause of ICP remains unknown, although its pathogenesis appears to be related to the effects of sex hormones in the liver of genetically predisposed women (Ref. Reference Lammert58). Expression of the disease may be modulated by environmental factors. Available information on the role of sex hormones, genetic and environmental factors is summarised below.
Influence of sex hormones
Role of oestrogens
Several lines of evidence support the notion that oestrogens are involved in the pathogenesis of ICP (Refs Reference Reyes and Simon44, Reference Arrese, Leuschner, Berg and Holtmeier59). First, the disease appears mainly in the last third of gestation when oestrogen production reaches its maximum. In addition, the prevalence of ICP is higher in twin pregnancies when levels of oestrogen are higher than in single pregnancies (Ref. Reference Gonzalez60). The oestrogen conjugates whose serum levels increase most during pregnancy are oestriol-3-sulphate, oestriol-16-glucuronide, oestriol-3-glucuronide and oestriol-3-sulphate-16-glucuronide (Ref. Reference Levitz, Kadner and Young61). Second, ICP closely resembles the cholestatic picture seen in some women using oral contraceptives with a high oestrogen content (Ref. Reference Reyes and Simon44). And third, administration of an oestrogen derivative to experimental animals induces cholestasis, which appears to be related to impaired expression and/or function of specific transporter proteins such as Ntcp, Bsep and Mrp2 (Refs Reference Geier15, Reference Arrese47). Ntcp is strongly downregulated after oestrogen administration in the rat, which parallels a reduction of Na+-dependent uptake of bile salts (Refs Reference Simon62, Reference Geier63). In addition, Bsep protein expression is reduced significantly to about 60% of normal levels (Ref. Reference Lee64) and the function of Bsep is disturbed by oestrogens (Ref. Reference Arrese65), probably by internalisation of the protein (Ref. Reference Crocenzi66). Protein levels of Mrp2 are consistently reduced after oestrogen treatment (Ref. Reference Trauner67), which seems to be related to post-transcriptional mechanisms, such as oestrogen-induced endocytic internalisation (Ref. Reference Mottino68). It is also likely that altered tight junction integrity may contribute to impaired bile secretion and may be causally related to changes in the expression of Mrp2, and possibly Bsep (Ref. Reference Mottino69). Interestingly, the changes observed in transporter expression in oestrogen-induced cholestasis are very similar to those observed during normal pregnancy in the rat (Ref. Reference Arrese47). Oestrogens also downregulate AQP8 protein expression, impairing osmotic membrane water permeability and contributing to bile secretory dysfunction in oestrogen-induced cholestasis (Ref. Reference Carreras70). In addition, a recent study indicates that ethinyloestradiol increases expression and activity of rat liver Mrp3, which may represent an adaptive response to prevent intracellular accumulation of common Mrp substrates (Ref. Reference Ruiz71). Finally, oestrogens can also cause cholestasis by indirect actions: as oestrogen can induce modifications of plasma membrane fluidity (Ref. Reference Reyes and Simon44), this may play a role in changing the efficacy of hepatic transporters.
Role of progesterone
Progesterone metabolites have been also postulated to play a role in the pathogenesis of ICP. Administration of progesterone in early pregnancy may be a risk factor for ICP (Ref. Reference Bacq72). In addition, alterations in progesterone metabolism have been shown to occur in ICP patients that exhibit increased serum levels and urinary excretion of total sulphated progesterone metabolites (Refs Reference Reyes and Sjovall73, Reference Meng74, Reference Meng75, Reference Meng76). Interestingly, serum levels of progesterone disulphates decrease when patients with ICP are treated with ursodeoxycholic acid (UDCA) – a hydrophilic bile acid that stimulates bile secretion and ameliorates pruritus (Ref. Reference Glantz77). Thus, it is possible that a selective defect in the secretion of sulphated progesterone metabolites into bile may underlie ICP. However, it is not clear whether these changes are of pathogenic relevance or represent secondary changes (Ref. Reference Poupon78). In the experimental setting, administration of progesterone to rodents has little effect on bile secretion, but in vitro data suggest that several sulphated progesterone metabolites induce transinhibition of the canalicular bile salt transporter Bsep (Refs Reference Vallejo79, Reference Stieger80). However, transinhibition of Bsep requires the prior secretion of sulphated progesterone into the canalicular lumen via Mrp2, which is reduced in pregnancy (Refs Reference Arrese47, Reference Cao53). Clearly, further studies are necessary to better define the role of progesterone metabolites in the pathogenesis of ICP.
Genetic factors
The existence of an underlying genetic predisposition to ICP is suggested by the presence of familial clustering (Refs Reference Reyes, Ribalta and Gonzalez-Ceron81, Reference Holzbach, Sivak and Braun82), a higher prevalence of ICP in Mapuche Indians (Ref. Reference Reyes83) and the exaggerated response to an oestrogen challenge observed in nulliparous sisters and brothers of ICP patients (Ref. Reference Reyes84). The search for a specific genetic defect has been addressed by different research groups in recent years, focusing on the role of mutations in genes encoding biliary transport proteins, which are involved in several forms of familial intrahepatic cholestasis (Ref. Reference Oude Elferink, Paulusma and Groen28). Thus, mutations in genes encoding MDR3, the aminophospholipid transporter ATP8B1/FIC1 and BSEP have been found in patients diagnosed with ICP (Refs Reference Painter85, Reference Lang86, Reference Pauli-Magnus87, Reference Dixon88, Reference Schneider89). In addition, polymorphisms of ABCC2 (Ref. Reference Castano90) and functional variants of the bile salt sensor FXR (Ref. Reference Van Mil91) have been found in patients with ICP. A brief summary of currently available information is discussed below.
Mutations/variants in ABCB4
Lack of ABCB4 gene product as a result of homozygous inactivating mutations results in defective functioning of the phospholipid export pump and in impaired biliary secretion of phosphatidylcholine (Ref. Reference Trauner, Fickert and Wagner27). This genetic defect has been linked to a broad spectrum of hepatobiliary disorders, ranging from progressive familial intrahepatic cholestasis in children to ICP, adulthood biliary cirrhosis and cholelithiasis (Ref. Reference Jacquemin92). The first report that a heterozygous nonsense mutation of the MDR3 gene was present in some patients that suffered from cholestasis during pregnancy was published in 1999 (Ref. Reference Jacquemin93). Since then, single cases or small series of patients harbouring mutations in ABCB4 and suffering from severe cholestasis during pregnancy have been published (Refs Reference Pauli-Magnus87, Reference Jacquemin93, Reference Lucena94). However, caution should be exercised in interpreting these data, because patients included in these studies belong to a heterogeneous population, which may not represent true cases of ICP. In fact, some reports include patients with uncommon features of ICP, such as elevated serum levels of γ-glutamyl transpeptidase (GGT) and very early disease presentation (i.e. first trimester) and even patients who went on to develop chronic liver disease (Refs Reference Pauli-Magnus87, Reference Lucena94). Thus, cholestasis during pregnancy in patients with MDR3 deficiency may represent a hitherto underlying genetic defect that is unmasked by pregnancy rather than true ICP, which is a well-defined benign clinical entity (Refs Reference Arrese9, Reference Arrese and Accatino95). In fact, larger studies carried out in a Finnish population have failed to find a strong association between mutations in ABCB4 and ICP (Ref. Reference Savander96). Moreover, one report from the UK suggests that mutations in the coding region of ABCB4 are not a common cause of ICP (Ref. Reference Mullenbach97) and another from Italy (Ref. Reference Floreani98) found such mutations in only a small percentage of their ICP patients. However, a recent study from Sweden studying patients with well-defined ICP showed that common variants of ABCB4 may represent a genetic risk factor for the severe form of ICP defined by bile acid levels above 40 µmol/l (Ref. Reference Wasmuth99).
Mutations/variants in ABCB11
ABCB11 is located on chromosome 2q24 and encodes the bile salt exporter BSEP (Ref. Reference Thompson and Strautnieks100). Patients with mutations in this gene develop a progressive form of familial cholestasis (PFIC2) (Refs Reference Strautnieks101, Reference Jansen102) or more benign recurrent cholestatic episodes depending on the extent of BSEP deficiency (Ref. Reference van Mil103). Data on the relationship of mutations or variants of ABCB11 with ICP are somewhat conflicting. Although some reports suggest that mutations in ABCB11 cause enhanced susceptibility to ICP (Refs Reference Eloranta104, Reference Keitel105), other studies indicate that common ABCB11 variants do not play a role in the pathogenesis of ICP (Refs Reference Pauli-Magnus87, Reference Wasmuth99, Reference Painter106). A recent study suggests that heterozygosity for the common ABCB11 mutations only accounts for 1% of Caucasian ICP cases, and that the V444A polymorphism, which is associated with reduced levels of BSEP, is a risk factor for developing the disease (Ref. Reference Dixon88).
Mutations/variants in ATPB8B1
Although its exact function is not known, deficiency of the ATPB8B1 gene product FIC1 causes either progressive cholestasis in infancy or recurrent cholestasis in adults (Refs Reference Bull107, Reference Bull108). Current information on the role of ATPB8B1 in ICP indicates that it is probably not a major contributing factor. Although mutations or variants in ATPB8B1 have been identified in patients exhibiting cholestasis during pregnancy, this was an infrequent finding: ATPB8B1 mutations were found in only four out of 182 cases in a British study (Ref. Reference Mullenbach109) and less than 10% of cases in a study from Finland (Ref. Reference Painter85).
Mutations/variants in ABCC2
MRP2, the product of gene ABCC2, is a multispecific organic anion transporter responsible for the secretion of a large variety of xenobiotic and endogenous compounds – including oestrogen and progesterone metabolites and bile acid sulphates and glucuronides – into bile. Mutations in ABCC2 cause Dubin-Johnson syndrome (Ref. Reference Paulusma110), which is characterised by hyperbilirubinaemia, deposition of melanin-like pigment in otherwise normal liver cells and prolonged retention of bromosulphthalein (Ref. Reference Arias111). ABCC2 is a good candidate gene for ICP because protein levels of this transporter are reduced during pregnancy in rodents and also in experimental oestrogen-induced cholestasis (Refs Reference Arrese47, Reference Trauner67). However, only one study has assessed the possibility that genetic variation at this ABC transporter may influence ICP susceptibility (Ref. Reference Sookoian112). In this report, four single nucleotide polymorphisms were investigated in 40 patients with ICP. An association between a variant in exon 28 of ABCC2 and ICP was found. Although further studies are required to confirm this observation, this result is in line with recent findings indicating that genetic variations of MRP2 are an important predisposing factor for drug-induced toxic liver injuries (Ref. Reference Choi113).
Mutations/variants in NR1H4
The NR1H4 gene product is FXR, which acts as an intracellular bile salt sensor and is thought to be of pivotal importance in the adaptive response of hepatocytes in cholestasis, controlling bile salt synthesis, metabolism and transport (Ref. Reference Zollner43). A recent case-control study of allele frequencies assessed the possibility that genetic variation in FXR confers susceptibility to ICP in two independent cohorts of Caucasian ICP patients (Ref. Reference Van Mil91). Four novel heterozygous FXR variants were detected and one of them (M173T) was found to be significantly associated with ICP (Odds ratio: 3.2). This variant was shown to have functional consequences because it influences ABCB11 transcription in vitro. Interestingly, some women harbouring FXR variants had pruritus when not pregnant, at specific stages of the menstrual cycle, which is an uncommon feature in patients that have previously suffered ICP.
Other potential candidate genes
Genetic variants or mutations in other genes could hypothetically be involved in ICP pathogenesis. Since no specific studies have been performed so far, the role of other proteins or transporters remains speculative. Based on their importance in drug disposition or in the control of cellular responses to cholestasis, we postulate that other candidate genes may include: (Reference Arrese and Reyes1) the polyspecific xenobiotic transporter ABCG2; (Reference Reyes2) the steroid and xenobiotic receptor SXR/PXR; and (Reference Lammert, Marschall and Matern3) genes controlling biotransformation, such as cytochrome P450 enzymes and uridine glucuronyl transferases. Genetic polymorphisms of ABCG2 are common and may influence the removal of xeno- and endobiotics as well as drugs, which may be relevant during pregnancy when the excretory capacity of the liver is already reduced (Refs Reference Hardwick, Velamakanni and van Veen31, Reference Lee114, Reference Wang115). SXR/PXR is a transcription factor that mediates protection against bile-salt-induced liver injury in cholestatic animal models by regulating several genes encoding proteins critical to drug transport and metabolism, including cytochrome P450 3A4 (CYP3A4) and drug efflux pumps such as MDR1 (ABCB1). Interestingly, some reports suggest that functional SXR gene variants influence the disease course in primary sclerosing cholangitis – a prototypical cholestatic disease (Ref. Reference Tom116). No information is available on the role of SXR/PXR in ICP at present. Interestingly, one drug used to treat ICP is ursodeoxycholic acid (UDCA) (Refs Reference Zapata117, Reference Glantz118), which is a potent stimulator of this nuclear receptor in human hepatocytes (Refs Reference Boyer119, Reference Schuetz120). Finally, polymorphisms in genes controlling biotransformation, such as cytochrome P450 enzymes and uridine glucuronyl transferases, may also play a role in ICP, as they do in other forms of liver injury (Refs Reference Daly121, Reference Burchell122).
Environmental factors
The observed geographical and seasonal variations of ICP suggest the involvement of environmental factors in the occurrence of ICP in genetically susceptible individuals (Refs Reference Arrese and Reyes1, Reference Reyes2, Reference Berg123). Although several factors have been investigated, including the long-chain monounsaturated fatty acid erucic acid and selenium, there is no definitive evidence that they play a pathogenetic role in ICP (Refs Reference Reyes50, Reference Reyes124). A recent report suggests that a ‘leaky gut’ may be present in ICP patients, which would enhance the absorption of bacterial endotoxin and potentially other xenobiotics present in the diet (Ref. Reference Reyes125). Further research is needed to elucidate environmental factors involved in the pathogenesis of ICP.
An integrative view of ICP pathogenesis
Based in the available information summarised above, an integrative view of ICP pathogenesis may be attempted (Fig. 2). Of course, this endeavour has a number of limitations because the weight of scientific evidence for each factor is not the same. However, some principles can be delineated. First, pregnancy seems to be a physiological cholestatic-prone condition with functional alterations in bile secretory function, probably related to downregulation of key hepatobiliary transporters, such as NTCP and MRP2, and functional impairment of BSEP. This is likely to be related to the effects of oestrogens in the liver, although transinhibition of BSEP by progesterone metabolites may play a role. Second, genetic variants, rather than mutations, in either canalicular transporters (mainly ABCB4, ABCB11 and ABCC2) or bile salt sensors (FXR) may provide a genetic background that exaggerates the cholestatic-prone state and determines the development of a phenotypic continuum ranging from subclinical cholestasis (i.e. AHP) to cholestatic jaundice. And third, modulation of disease expression is related to one or more unidentified environmental factors.
Integrative view of intrahepatic cholestasis of pregnancy (ICP) pathogenesis. Pregnancy is a physiological cholestatic-prone condition with functional alterations in bile secretory function related to downregulation of key hepatobiliary transporters, such as NTCP (SLC10A1) and MRP2 (ABCC2), and functional impairment of BSEP (ABCB11). These changes are probably due to the effects of oestrogens in the liver, although transinhibition of BSEP by progesterone metabolites may play a role. Genetic variants, rather than mutations, in canalicular transporters (mainly ABCB4, ABCB11 and ABCC2), bile salt sensors (NR1H4; FXR) or other genes, may provide a genetic background that exaggerates the cholestatic-prone state and determines the development of a phenotypic continuum, ranging from subclinical cholestasis (asymptomatic hypercholanemia of pregnancy [AHP]) to cholestatic jaundice. Modulation of disease expression is related to unidentified environmental factors. Abbreviations: PL, phospholipids; BS, bile salts; FIC1, familial intrahepatic cholestasis 1; OA, organic anions; BSEP, bile salt export pump; FXR, Farnesoid X receptor. ABCB11 and ABCB4 are depicted in bold to indicate that evidence for a role of these transporters is more abundant. Data on ATPB8B1, ABCC2 and NR1H4 are more limited or negative. The role of ABCG2, PXR/SXR and other proteins is hypothetical.
Pathogenesis of perinatal consequences of ICP
ICP is associated with preterm delivery and poor perinatal outcome, including intrauterine foetal death (Refs Reference Reyes2, Reference Nichols126). The mechanisms involved in the pathogenesis of premature labour, as well as foetal distress and stillbirths in ICP are unclear (Ref. Reference Germain127). Elevated bile salt levels can contribute to preterm delivery by increasing oxytocin bioaction (Ref. Reference Germain128). In addition, bile salts may also have a role in unexplained intrauterine foetal death by inducing vasoconstriction of chorionic veins (Ref. Reference Sepulveda129), oxidative stress and increased apoptosis in the placenta and foetal liver (Ref. Reference Perez130) or impairing foetal cardiomyocyte function (Ref. Reference Williamson131). Elevated bile salt levels in the foetus result from increased bile salt levels in the maternal serum and the impaired ability of the placenta to carry out vectorial bile acid transfer from the foetus to the mother (Refs Reference Macias132, Reference Serrano133). Moreover, immature hepatic transport systems can contribute to impaired handling of bile salts and other organic anions in a cholestatic setting. Experiments carried out in a model of experimental cholestasis in pregnant rats have shown that a reversible impairment of neonatal hepatobiliary function occurs in the offspring of cholestatic mothers (Refs Reference Arrese134, Reference Serrano135).
Research in progress and outstanding research questions
As mentioned above, ICP seems to represent an exaggerated response to the high levels of sex hormones present in the final third of pregnancy, probably associated with a genetic defect. However, the precise nature of the contribution of specific genes is unclear. Although genes relevant for hepatobiliary transport have been investigated in patients with ICP, it would be of great help for clinicians if future investigations assess whether one or more steps of drug and other xenobiotic processing by the liver are involved in the pathogenesis of ICP. Another important problem with the genetic studies of ICP conducted so far is the heterogeneous nature of the populations of patients included. Thus, important aspects of future research in the field are the screening of large, well-characterised populations of different ethnic backgrounds and the use of a genome-wide association approach that allows the study of hundreds or thousands of single nucleotide polymorphisms at the same time.
A new nomenclature for cholestatic episodes occurring during pregnancy may be needed to better characterise the underlying defects in each patient. The general term of ‘obstetric cholestasis’ could be adopted, and subtypes according to the underlying defect defined. This is relevant because it is likely that each subtype has a different prognosis. For instance, obstetric cholestasis with high GGT levels may carry a risk of chronic liver disease (Ref. Reference Lucena94) and will require a closer follow-up than is required for patients with normal GGT levels. In addition, because the classic form of ICP usually resolves within a few days after delivery, persistent alterations in liver function may indicate the presence of a genetic liver disease, which should be investigated.
Identification of the environmental factors affecting women with enhanced susceptibility to ICP would help to prevent this disease. The marked variation in prevalence of the disease in countries such as Chile (Ref. Reference Reyes124) poses a key unsolved research question. A more-detailed study of the consequences of maternal cholestasis on both placental transport function (Ref. Reference Marin136) and hepatic transport capacity of foetal liver (Ref. Reference Marin137) would help to prevent foetal distress and intrauterine death in the clinical setting of ICP.
Finally, current therapeutic strategies of ICP mainly include the use of the hydrophilic bile acid UDCA. Based on results from several randomised controlled trials (Refs Reference Palma138, Reference Kondrackiene, Beuers and Kupcinskas139), UDCA is widely regarded as the first line of treatment for ICP because it improves pruritus and liver enzymes in ICP patients. UDCA also prolongs the duration of pregnancy without any adverse effect on the foetus. The proposed mechanism of action of UDCA includes stimulation of targeting of canalicular transporters, mainly Bsep and Mrp2 (Refs Reference Beuers140, Reference Dombrowski, Stieger and Beuers141, Reference Beuers142, Reference Haussinger143, Reference Kurz144, Reference Kubitz145). Since the clinical phenotypes of AHP, pruritus of pregnancy and jaundice of pregnancy may correlate with the amount of BSEP or MRP2 in the canalicular membrane, UDCA is a safe and useful drug in this scenario. New strategies aiming to increase the abundance of these transporters are emerging, as the role of nuclear receptors in cholestasis is clarified (Refs Reference Boyer7, Reference Zollner43, Reference Karpen146). Thus, one can envisage that the use of nuclear receptor ligands (FXR or PXR/SXR agonists in particular) could also be useful drugs for the treatment of ICP (Ref. Reference Fiorucci147). Other molecules that increase the abundance of canalicular proteins by post-transcriptional mechanisms (Refs Reference Hayashi and Sugiyama148, Reference Ananthanarayanan149) may also hold promise as therapeutic agents in ICP.
Acknowledgements and funding
This work was partially supported by a grant from the Fondo Nacional de Ciencia y Tecnología (FONDECYT 1050780 to MA) and by the Junta de Castilla y Leon (Grants SA013/04 and SA059A05), Spain; Ministerio de Ciencia y Tecnologia, Plan Nacional de Investigacion Cientifica, Desarrollo e Innovacion Tecnologica (Grant BFI2003-03208), Spain; Fundacion Investigacion Medica Mutua Madrileña (Conv-III, 2006); Instituto de Salud Carlos III, FIS (Grants CP05/00135 and PI051547). Authors from Spain are members of the Network for Cooperative Research on Membrane Transport Proteins (REIT), co-funded by the Ministerio de Educacion y Ciencia, Spain and the European Regional Development Fund (ERDF) (Grant BFU2005-24983-E/BFI) and belong to the ‘Centro de Investigacion Biomedica en Red’ (CIBERehd) for Hepatology and Gastroenterology Research, Instituto de Salud Carlos III, Spain. We thank the peer reviewers of this article for their helpful comments that contributed significantly to the final version of the manuscript.
