Introduction

Cystic fibrosis (CF) is a genetic disorder characterized by epithelial electrolyte transport abnormalities, elevated sweat Cl^– concentrations, pancreatic insufficiency, and chronic lung disease in most patients. It is the most common potentially fatal genetic disorder in the Caucasian population, affecting 1 in 2,400–3,500 live births [1, 2]. It is an autosomal recessive disorder caused by a mutation in the gene CFTR encoding the cystic fibrosis transmembrane conductance regulator (CFTR), a membrane channel protein. The clinical significance of hepatobiliary disease in CF has not been well characterized primarily because of two factors: (1) pulmonary involvement leads to early mortality in a majority of patients, and (2) the clinical identification of CF-associated liver disease has been difficult because, although it is progressive, liver involvement is often asymptomatic until the appearance of end-stage complications. Recently, with improved pulmonary treatments, median life expectancy now exceeds 40 years and CF-associated hepatobiliary disease is recognized and characterized more comprehensively. Liver disease is now the third major cause of death in CF (after pulmonary disease and complications of lung transplant). In recent years, advances in our understanding of the function of CFTR in bile duct epithelia have provided a stronger scientific basis for the pathogenesis of the disease, leading to insights concerning potentially novel therapeutic approaches.

The earliest reports of CF probably date to the Middle Ages, with reports of malnourished and “sickly” children that tasted “salty” when kissed. In 1905, Landsteiner published the first description of an abnormal pancreas and meconium ileus in CF, although it was Anderson’s description in 1938 that gave us a more modern description of “cystic fibrosis of the pancreas” [3]. In 1953, di Sant’Agnese described the abnormal sweat electrolyte concentrations forming the basis for the diagnostic sweat test, which served as the basis of the diagnosis until the recent availability of genetic analysis. On a cellular level, meticulous studies of the sweat duct led Quinton to describe the Cl^– transport defect [4]. Finally, it was the discovery of the responsible gene in 1989 by Riordan, Tsui, and Collins that permitted critical breakthroughs in the understanding of CF pathogenesis [5]. It was hoped that the gene discovery would herald a quick and forthcoming cure for CF, and although this dream has not been realized to date, the intense study of the role of CFTR in cell and organ function has advanced our knowledge of basic cellular physiology enormously.

Cystic Fibrosis Transmembrane Conductance Regulator

Action as a Channel

The cystic fibrosis transmembrane conductance regulator is located on the long arm of chromosome 7. It contains 250,000 base pairs with 27 exons and encodes a polypeptide product of 1,480 amino acid residues CFTR. The protein belongs to a family of transmembrane proteins known as ATP-binding cassette (ABC) proteins, which all contain transmembrane sequences and hydrolyze ATP for activation. CFTR contains two domains, capable of spanning the membrane six times, separated by regulatory cytoplasmic domains consisting of two consensus nucleotide-binding domains and an intervening regulatory domain (Figure 26.1). It is now well established that CFTR functions as a cyclic adenosine monophosphate (cAMP)-dependent Cl^– channel in the apical membrane of secretory epithelia. CFTR-associated Cl^– channels have a small unitary conductance of approximately 8 picosiemens and a linear current-voltage relation. Under normal conditions, cAMP-dependent protein kinase A phosphorylates CFTR, causing channel opening and transport of Cl^–.

Figure 26.1 Putative monomeric structure for the cystic fibrosis transmembrane conductance regulator (CFTR). The 1,480 amino acid residues are arranged into 12 membrane-spanning domains. The first six transmembrane domains are followed by an intracellular regulatory domain, or R-domain, containing phosphorylation sites for protein kinase A. Transmembrane segments 6 and 12 are followed by sequences containing ATP-binding domains or nucleotide-binding domains (NBD). The most common CF-associated mutation (ΔF508) results from a base pair deletion in exon 10, causing a deletion of phenylalanine at position 508 in the first NBD region. The C-terminal end of the protein contains PDZ-domains that may facilitate protein–protein interactions.

Role in Liver Function

In human liver, CFTR is expressed on the apical membrane of bile duct cells (cholangiocytes) and gallbladder epithelia but is not expressed in hepatocytes or other cells of the liver [12]. In the mouse (and probably humans), CFTR is found predominantly in the medium- to large-sized intrahepatic bile ducts, but not in the small intrahepatic ducts. Its location on the apical (luminal) membrane of cholangiocytes, as well as the large increase in cAMP-stimulated CFTR activity observed with secretory agonists, suggests an important role of CFTR in bile formation [13].

The formation of bile by the liver depends on complementary interactions between hepatic parenchymal cells (hepatocytes) and intrahepatic bile duct cells (cholangiocytes). Both of these cell types work in a complementary manner to initiate and modify bile flow. Bile formation is initiated at the hepatocyte canalicular membrane through the transport of bile acids, organic and inorganic solutes, electrolytes, and water. Subsequently, bile flows into the lumen of the intrahepatic bile ducts where it undergoes alkalinization and dilution as a result of cholangiocyte secretion. Although cholangiocytes constitute <5% of the nuclear mass of the liver, they form an extensive branching network and account for approximately 40% of bile flow in humans; demonstrating a prodigious capacity for secretion.

Studies in isolated cholangiocytes, biliary epithelial monolayers, and isolated bile duct segments have helped to elucidate the basic mechanisms of constitutive and stimulated secretion in biliary epithelium [14]. One of the current working models is shown in Figure 26.2. Intracellular Cl^– accumulation occurs via uptake of Cl^– at the basolateral membrane by a bumetadine-sensitive Na⁺/K⁺/2Cl^– cotransporter. While the intracellular accumulation of Cl^– leads to values above the electrochemical equilibrium, the Cl^– permeability of the apical membrane under basal conditions is low. However, exposure to agonists such as secretin, that increase intracellular cAMP levels, leads to a rapid series of events including (1) opening of Cl^– channels in the apical membrane and efflux of Cl^– into the duct lumen; (2) an increase in Cl^––HCO₃^– exchange activity with a resultant increase in ductal HCO₃^– concentration; and (3) movement of water out of the cell through water channels or aquaporins. The findings that CFTR is localized to the apical membrane of cholangiocytes and secretin-stimulated Cl^– channels have properties analogous to CFTR support a working model that postulates a role for CFTR in the regulation of ductular secretion. According to this model, the secretin-stimulated increase in Cl^– and HCO₃^– permeability is through protein kinase A-dependent activation of CFTR. The generation of a lumen-negative potential favors movement of Na⁺ into the bile duct through a paracellular pathway and water through aquaporins. Additionally, two K⁺ channels (SK2, IK1) have been identified in the basolateral membrane of cholangiocytes and play an important role in maintaining the membrane potential difference necessary for continued transepithelial secretion [15]. Therefore, CFTR contributes to normal bile formation and alkalinization through the regulation of Cl^–, HCO₃^–, and water transport (Figure 26.2). Although this model implies a prominent role for CFTR in normal bile formation, it fails to explain why only a minority of patients with CF develop liver disease despite the fact that they all have abnormal or absent CFTR in bile duct epithelia. This observation suggests that CFTR may not be the predominant pathway for Cl^– secretion in cholangiocytes.

Figure 26.2 Model of cholangiocyte bile formation highlighting channels involved in secretion. Stimulation of basolateral receptors by secretin results in increases in cAMP and protein kinase A-dependent stimulation of Cl^– efflux through CFTR. The transmembrane Cl^– gradient drives Cl^–/HCO₃^– exchange. Alternatively, HCO₃^– may enter bile through a conductive manner or through CFTR itself. Water is transported via aquaporin (AQP) proteins. The increase in HCO₃^– and water secretion leads to alkalinization and dilution of bile. Other Cl^– channels, including volume-sensitive, P2 receptor-linked, and TMEM16A, a Ca²⁺-activated Cl^– channel, have been identified. Lumenal ATP and bile acids may also stimulate Cl^– efflux. An apical transporter for bile acids has been identified (ASBT). On the basolateral membrane, Na⁺/H⁺ exchange, Na⁺-dependent Cl^–/HCO₃^– exchange, and Na⁺/HCO₃^– symport help to maintain intracellular pH and HCO₃^– concentrations. Uptake of Cl^– is mediated by a Na⁺/K⁺/2Cl^– cotransporter. The Ca²⁺-activated K⁺ channels (SK2, IK1) have been identified in the basolateral membrane and work in parallel with apical Cl^– channels to hyperpolarize the membrane and provide the driving force for continued secretion. See text for details. P2: purinergic receptor; * location (apical vs. basolateral) not definitively established.

Alternative Channels and Cholangiocyte Secretion

In addition to CFTR, cholangiocytes express several other Cl^– channels, including a Ca²⁺-activated Cl^– channel, a purinergic receptor-linked Cl^– channel, and a volume-stimulated Cl^– channel (Figure 26.2) [16]. However, their regulation and overall contribution to biliary secretion is largely unknown at present. It is attractive to speculate, however, that these alternative Cl^– channels may compensate for the CF secretory defect in the liver in CF. In fact, in the Cftr^–/– mouse model, increased expression of Ca²⁺-activated Cl^– channels in tracheal epithelial cells is associated with mild pulmonary disease [17]. Recently, a Ca²⁺-activated Cl^– channel, TMEM16A, has been identified in liver, where it is localized predominantly on the apical membrane of cholangiocytes [8]. This channel is regulated independently from CFTR and therefore may be an attractive target for new choleretic agents. While CFTR is regulated by secretin binding basolateral membrane receptors, TMEM16A is regulated through ATP binding purinergic receptors (P2) on the apical cholangiocyte membrane [19]. ATP is present in bile, and binding of ATP to P2 receptors increases Cl^– efflux from isolated cholangiocytes and dramatically increases transepithelial secretion in biliary epithelial monolayers [20]. Indeed, the magnitude of the secretory response to ATP is two- to three-fold greater than that observed with cAMP. With recent studies demonstrating that the mechanical effects of fluid-flow and/or shear stress at the apical membrane of biliary epithelial cells is a robust stimulus for ATP release, a model emerges in which mechanosensitive ATP release and Cl^– secretion is a dominant pathway regulating biliary secretion [21]. If these observations apply to in vivo conditions, a decrease in bile flow associated with CF may be accompanied by alterations in these mechanosensitive pathways, which may further exacerbate abnormalities in Cl^– secretion and bile formation.

The Cystic Fibrosis Transmembrane Conductance Regulator as a Regulator of Epithelial Innate Immunity.

Recently, it has been shown in mouse and human induced pluripotent stem cells (iPS)-derived cholangiocytes that CFTR may regulate immune responses [22, 23]. A loss of normal CFTR protein on the apical membrane, which occurs in CF, may lead to a disruption of the actin cytoskeleton, activation of kinase signalling pathways, and elaboration of pro-inflammatory cytokines. Thus, a normal CFTR protein in the membrane of cholangiocytes may serve to suppress an active pro-inflammatory response. These are intriguing findings and more work needs to be done to elucidate the role of CFTR in normal immune functioning.

Summary

In summary, CFTR is a cAMP-dependent Cl^– channel expressed on the apical membrane of cholangiocytes that contributes to ductular secretion. However, the possible role of other membrane Cl^– channels is yet to be determined. Intriguing studies have established that CFTR, in addition to its role as a Cl^– channel, is in fact a “transmembrane regulator” modulating other membrane channels and transporters, signaling pathways, and may be an important component of the innate immune response. Further study of CFTR function and regulation may help to elucidate the mechanisms of cholangiocyte function and bile formation. The remainder of this chapter focuses on the hepatobiliary effects of abnormal CFTR function, namely CF-associated liver disease.

Cystic Fibrosis Transmembrane Conductance Regulator Mutations

There are now more than 2,000 recognized mutations in CFTR (CFTR Mutation Data Base: www.genet.sickkids.on.ca/cftr). Worldwide, the ΔF508 mutation accounts for 66% of the described mutations while G542X and G551D, the next two most common mutations, account for 2.4% and 1.6%, respectively (CFTR Mutation Data Base). The incidence of CF in the Caucasian population corresponds to a carrier frequency of approximately 5%. This high carrier frequency in a lethal genetic disease suggests the possibility of a survival advantage for heterozygotes. In fact, it has been suggested that the absent or unresponsive Cl^– channel associated with CFTR mutations may have protected infants during epidemics of cholera, which causes secretory diarrhea through toxin-mediated, cAMP-dependent activation of Cl^– channels. The Cftr^–/– mouse has been shown to be resistant to the effects of cholera infection, providing some evidence for this theory [24].

Mutations are classified into five groups according to their effect on CFTR protein function (Figure 26.3). Class I mutations (such as G542X and R553X) cause impairment of CFTR mRNA production. Class II mutations result in defective processing or trafficking of CFTR protein to the apical membrane. The most common mutation, ΔF508, is of this class and results in a base pair deletion in exon 10 with a consequent deletion of phenylalanine at position F508 in the first nucleotide-binding domain of the protein [5]. The F508– CFTR protein does not fold correctly and is subsequently diverted from normal trafficking to the apical membrane and degraded by the ubiquitin–proteosome pathway. Class III mutations (G551D and others) are associated with defective regulation of CFTR, which locates correctly to the apical membrane but does not respond to cAMP agonists. Class IV mutations (R117H and others) demonstrate some residual Cl^– conductance but at a significantly decreased amplitude. Class V mutations (A455E, P574H, and others) lead to abnormal splicing of CFTR with partial reduction in the number of functioning Cl^– channels. Mutations of classes I, II, and III are considered severe because they result in an absence of functioning CFTR at the plasma membrane, whereas class IV and V are “mild” mutations with some residual CFTR activity demonstrated. The report from the Cystic Fibrosis Foundation (CFF) registry, which records data from USA CF centers, reveals that ΔF508/ΔF508 homozygotes account for 50.6% and F508/other heterozygotes account for 37.9% of the mutations reported in the registry.

Figure 26.3 Classification of CFTR mutations. Class I mutations (nonsense and frameshift) result in abnormal mRNA production and no CFTR protein. Class II mutations (amino acid deletion, missense) result in abnormal CFTR protein trafficking with subsequent degradation. Class III mutations (missense) result in a mature CFTR protein that is refractory to normal activation. Class IV mutations (missense) result in a CFTR protein that localizes normally, but with a reduction in single-channel conductance. Class V mutations (alternative splicing, missense) result in a decreased full-length mRNA and a decrease in the number of functional CFTR channels at the apical membrane. ER: endoplasmic reticulum.

Overall, although specific gene mutations have been associated with the severity of pancreatic involvement, there is no correlation between specific genotype and clinically detectable liver disease in patients with CF. However, there appears to be a lower frequency of liver disease in pancreatic-sufficient patients, who generally have milder mutations. Because all patients with CF have abnormal CFTR in the biliary tree, it is unclear why significant liver disease does not develop in all patients. Because patients with CF and identical CFTR mutations exhibit variable onset and severity of liver disease, it is postulated that there are other modifying genetic or environmental factors that determine whether clinically significant hepatobiliary involvement will occur.

Pathogenesis of Liver Injury

The pathophysiology underlying the development of CF-associated liver disease is still only speculative. Definitive studies directly assessing the effects of abnormal CFTR in the liver are lacking. Several proposed pathways in the pathogenesis of CF liver disease are shown in Figure 26.4. One leading hypothesis is that impaired secretory function of cholangiocytes results in a decrease in bile flow (cholestasis) and thickened, inspissated secretions in the bile ductules. The subsequent bile duct obstruction leads to liver cell injury and the development of fibrosis and cirrhosis. The histologic finding of inspissated eosinophilic material in bile ducts, a pathognomonic lesion in CF, provides some morphologic evidence for this “bile duct plugging” theory. Theoretically, abnormal viscosity of bile could result from several factors including defective transport of Cl^–, HCO₃^–, and mucins; Na⁺ reabsorption; altered composition of the bile acid pool; or a combination of these. Bile duct plugging would be anticipated to initiate a series of secondary steps including cholangiocyte injury, release of inflammatory mediators, and stellate cell activation with subsequent deposition of collagen, ultimately leading to fibrosis and cirrhosis.

Figure 26.4 A proposed model of the pathogenesis of cystic fibrosis liver disease.

Alternatively, the initiating step may be direct cholangiocyte injury from an abnormal CFTR protein. As mentioned above, the most common mutation ΔF508 results in protein misfolding and subsequent degradation by the ubiquitin– proteosome pathway. The misfolded protein can potentially form aggresomes, which may lead to cellular injury as seen in other diseases. This suggests a possible role of chaperone proteins, which are responsible for quality control mechanisms in the cell by targeting and degrading abnormal or misfolded proteins in the disease pathogenesis.

Additionally, as described above, CFTR may play an important role in the innate immune response within the liver [23]. In human iPS-derived cholangiocytes from a patient with CF and the Δ508 mutation, the loss of normal CFTR on the cholangiocyte membrane led to disruption of the actin cytoskeleton and the auto assembly of the Src tyrosine kinase. The subsequent activation of Src leads to TLR-4 signaling and the elaboration of pro-inflammatory cytokines [23]. This pro-inflammatory response would lead to the recruitment of inflammatory cells and presumably to cholangiocyte and hepatocyte injury, however more work is required to define the role of CFTR in the immune response. Although the initiating event in the development of liver disease is unknown, it appears that a progressive fibrogenic process ultimately leads to cirrhosis in a subset of patients. It is felt that this continuum, from cholestasis to focal biliary obstruction and ultimately to cirrhosis, may progress over many years. Genetic and environmental factors may modify any and all components of the pathway and may explain the heterogeneity in the liver response to abnormal CFTR mutation function. Several potential factors that have been proposed to contribute to, or modify, the liver injury in CF include altered mucin secretion, an altered intestinal microbiome, accumulation of toxic bile acids, circulating cytokines, stellate cell activation, and fat accumulation (steatosis).

Mucins

Hypersecretion of mucus is a major contributor to the lung pathology in CF. If similar factors are operable in the bile duct, then one potential etiology of increased bile viscosity would be altered mucin secretion. The viscous properties of mucus are determined in large part by mucin glycoproteins. The role of mucins in the bile duct is not defined, and there is conflicting evidence linking CFTR and mucin secretion. However, CFTR may contribute to normal mucin gel formation through effects on HCO₃^– secretion. The normal expansion of mucins, once secreted from the cell, is dependent on HCO₃^– and, therefore, the abnormal HCO₃^– transport associated with CF may promote mucin aggregation and plugging of the lumen [25].

Altered Intestinal Microbiome

Data continues to accumulate which highlights important interactions between the intestinal microbiome and the initiation and propagation of liver disease. Mouse and human studies have both demonstrated that the intestinal microbiome is altered in CF [26, 27]. In general, there appears to be less diversity in bacterial species and a shift to more pathogenetic bacterial strains within the intestine. The overall effect is a more inflammatory intestinal phenotype [28]. This is interesting in light of the finding that the induction of intestinal inflammation is a key factor in the development of liver disease in the Cftr^–/– mouse [22, 29]. A recent study of children with CF demonstrated that those with liver disease (defined by cirrhosis) had increased intestinal mucosal lesions and an altered fecal microbiome (increased Clostridium and decreased Bacteroides species) compared to those without liver disease [30].

Toxic Bile Acids

Patients with CF may have an altered bile acid pool, with an increase in hydrophobic and a decrease in hydrophilic bile acids. In fact, a study examining bile acid profiles in patients with CF found a higher level of endogenous biliary ursodeoxycholic acid (UDCA) in patients with CF but no liver disease compared with those with CF-associated liver disease [31]. The authors suggested that the elevated UDCA in these patients without liver disease may play a possible protective role. In addition to a decrease in hydrophilic bile acids, retaining hydrophobic bile acids may be responsible for subsequent hepatocyte injury, as seen in other cholestatic disorders.

Circulating Cytokines

Cytokines elaborated from other organs (e.g., lung and/or intestine) may have profound effects on biliary secretion and liver function. Elevations of many different cytokines have been reported in patients with CF during both baseline health and during respiratory exacerbations [32, 33, 34], which may directly effect cholangiocyte ion channel expression and function. For example, IL-5, IL-6, and IFNγ have all been shown to decrease expression of CFTR and cAMP-dependent transport in biliary epithelium [35, 36, 37]. Additionally, TGFβ has been shown to decrease expression of both CFTR and TMEM16A in airway and intestinal epithelial cells [38]. If similar findings occur in cholangiocytes, then a decrease in cytokine-mediated Cl^– efflux along the bile duct would be expected to result in a decrease in biliary secretion and bile flow.

Stellate Cells

Hepatic stellate cells have been implicated in the pathogenesis of CF and may play a role in the progressive fibrosis characteristic of this disorder [39]. Potentially, activation of stellate cells could occur through direct cholangiocyte or indirect hepatocyte injury with subsequent release of pro-inflammatory cytokines. In liver biopsy specimens of patients with CF, stellate cells have been found located in the periportal regions of the liver and their activation has been correlated with areas of collagen deposition and fibrosis [23]. Additionally, the acquisition of a robust actin-based cytoskeleton in activated stellate cells imparts a contractile phenotype and appears to contribute to the pathogenesis of portal hypertension, a prominent clinical manifestation of liver disease in CF. Elucidation of the role of stellate cells in the progressive fibrosis associated with liver disease may suggest novel therapies targeting these cells.

Steatosis

Hepatic steatosis is a common finding in CF, although it is unclear whether it results from the abnormal CFTR and the cholangiocyte transport defect or represents a separate, secondary entity. Several factors may contribute to hepatic steatosis in CF, including malnutrition, essential fatty acid deficiency, and elevated circulating levels of cytokines. The mechanism by which intracellular fat accumulation causes liver disease is an ongoing area of investigation. It may be that fat accumulation provides an increased substrate for lipid peroxidation and oxidative injury. It is unclear whether hepatic steatosis progresses to fibrosis or multilobular cirrhosis; however, the progressive nature of other disorders associated with fat accumulation (non-alcoholic fatty liver disease) suggests that steatosis may not be as benign a condition in CF as once thought.

Summary

Further studies into the pathogenesis of CF liver disease are clearly needed. It is hoped that the use of novel cell, epithelial, and animal models will help to further our knowledge. Unfortunately, the mouse model has been problematic for the study of liver disease as hepatobiliary complications are inconsistent, affected by the background strain, and may require an inflammatory environment [40]. Larger animal models (e.g., pig, sheep) seem to develop liver disease similar to that observed in humans and may provide important insight into the factors responsible for the development and progression of CF liver disease.

Potential Genetic Modifiers

Family studies suggest that factors independent of CFTR contribute to the development of liver disease and have led to intense scrutiny of the role of modifier genes in the pathogenesis of CF liver disease. Several associations or risk factors for the development of liver disease in CF have been described, including a history of meconium ileus, human leukocyte antigen (HLA) type, and heterozygosity for mutations in genes responsible for other liver diseases or in genes involved in mediating the response to liver injury.

Meconium Ileus

Several studies have suggested that a history of meconium ileus as an infant is a risk factor for the subsequent development of liver disease. However, in a large study of genetic modifiers of liver disease in CF, no association between meconium ileus and the development of liver disease was observed [41]. In fact, the prevalence of meconium ileus in patients with liver disease was similar to the prevalence of those with pancreatic insufficiency and, therefore, may not represent an independent risk factor.

Immune System and Antioxidant Status

Several studies have shown an association between certain HLA types and susceptibility for liver disease in CF. A higher frequency of HLA-DQw6 has been reported in British patients with CF liver disease compared with those without liver disease. Duthie et al. [42] found HLA-DQ6 in 66% of patients with liver disease but in only 33% without it. Two other antigens, HLA-DR15 and HLA-B7, with linkage disequilibrium with this locus, were also significant risk factors. It is interesting to note that the association was greater in males than in females, and only in males when the phenotype was restricted to portal hypertension (representing more severe disease). These findings suggest a possible immune contribution to the pathogenesis of hepatobiliary injury or, alternatively, another susceptibility gene linked with specific haplotypes lies at or near the HLA-DQ locus.

Other Ion Channels

It has been proposed that other Cl^– channels may modify phenotypic expression of CF. Several mouse models of CF have demonstrated the importance of other Cl^– channels in modulating organ-level disease expression. Clarke et al. [17] demonstrated that expression of Ca²⁺-activated Cl^– channels in the lung of Cftr^–/– mice may explain the mild pulmonary disease in these animals; while modulation of the severity of the gastrointestinal disease may also be affected by alternative channels on intestinal enterocytes. The role of other ion channels and transporters in normal bile formation as well as the pathogenesis of CF liver disease remains to be established.

Other Genetic Modifiers

The Z allele of SERPINA1 (encoding α₁-antitrypsin) has been identified as a marker for the development of severe liver disease in patients with CF [41]. The α₁-antitrypsin protein is mainly expressed in the liver and is a serine protease inhibitor. Deficiency of the α₁-antitrypsin is associated with an increased risk of chronic lung disease. In the liver, the normal protein (coded by the M allele) is secreted into the plasma, while the abnormal protein (coded by the Z allele) folds abnormally and accumulates in the endoplasmic reticulum. The accumulation of this abnormal protein leads to hepatocyte apoptosis, fibrosis, and ultimately cirrhosis. Therefore, the mechanism by which α₁-antitrypsin modifies liver disease in CF is likely by rendering the hepatocyte more susceptible to injury in the presence of abnormal CFTR and ongoing biliary obstruction. While smaller cohort studies have suggested possible associations of polymorphisms in other genes and CF liver disease, including genes for transforming growth factor-β, mannose-binding lectin, and glutathione S-transferase, larger screening studies have not been able to confirm these associations.

Clinical Manifestations of Cystic Fibrosis Liver Disease

The two most common clinical presentations of liver disease in CF are (1) an abnormal physical examination (hepatomegaly, small hard liver, splenomegaly) or (2) elevated serum liver enzymes on routine screening. A small, hard liver on examination and splenomegaly are signs of portal hypertension suggesting cirrhosis. It should be noted that patients with CF can present with end-stage liver disease and even cirrhosis with few, if any, outward signs to suggest chronic liver disease. Hepatomegaly suggests steatosis or focal biliary cirrhosis but may also suggest congestive hepatopathy associated with cor pulmonale and right heart failure. These distinct clinical manifestations including hepatic steatosis, neonatal cholestasis, focal biliary cirrhosis, and multilobular cirrhosis have been described based on clinical or histologic criteria and have variable prevalence rates (Table 26.1).

Congestive Hepatopathy

Although hepatic congestion is not a direct result of defective CFTR protein in bile duct epithelial cells, it nonetheless should be considered a clinically significant cause of hepatomegaly in CF. Chronically elevated right-sided heart pressures or cor pulmonale may lead to congestive hepatopathy through increased hepatic vein and sinusoidal pressures. The elevated sinusoidal pressure is thought to cause hepatocyte injury and necrosis. Eventually this can progress to “cardiac cirrhosis” with the development of bands of fibrosis extending between centrilobular areas with intervening normal portal areas. The diagnosis should be considered in those patients with CF with chronic lung disease, clinical signs of cor pulmonale, a large liver (and sometimes tender) on examination, and dilated hepatic veins on ultrasound examination. Ultrasonography with Doppler and echocardiography are, therefore, the main modalities that aid in the diagnosis. Additionally, Doppler ultrasonography or angiography can be helpful to exclude other vascular complications such as thrombosis of hepatic veins or inferior vena cava. Biochemical analysis usually reveals serum aminotransferase levels that are only mildly elevated (<2× to 3× times normal), and a prothrombin time that is normal or only slightly prolonged (<5 seconds prolonged). If aminotransferases are >3× normal, consideration should be given to the co-occurrence of other liver disorders. It is important to exclude hepatic congestion as a cause of hepatomegaly before any consideration of a percutaneous liver biopsy because the dilated hepatic veins associated with this condition increase the risk of bleeding; consequently, a transjugular or surgical approach should be considered in that circumstance. Treatment of this disorder relies on improving the underlying cardiac or lung disease.

Neonatal Cholestasis

Prolonged neonatal cholestasis may be quite common in newborns with CF. In one CF report, 35% of infants with CF had evidence of hepatomegaly or cholestasis within the first few months of life, and autopsy data has revealed histologic evidence of obstructive cholestasis in 38% of patients younger than three years with CF. Infants with CF and cholestasis should be evaluated thoroughly to exclude other cholestatic liver diseases, such as biliary atresia. On examination, hepatomegaly may be present and biochemical evaluation reveals elevation of serum direct bilirubin and gamma-glutamyltransferase concentrations. Stools may be acholic as in other cholestatic disorders, such as biliary atresia. Co-existent factors such as abdominal surgery, parenteral nutrition, or infection may contribute to prolonged cholestasis. There does not appear to be an increased risk for the long-term development of cirrhosis with an early history of neonatal cholestasis. It is important, however, to consider the diagnosis of CF in infants who present with cholestasis. A sweat test or genotype testing should, therefore, be performed as part of the evaluation of neonatal cholestasis.

Steatosis

Hepatic steatosis is characterized by a large, soft liver on palpation and other signs of chronic liver disease or portal hypertension are usually not present. Histologic examination reveals hepatic parenchymal cells filled with micro- and macrovesicular fat (Figure 26.5A). Ultrasound has been used to aid in the diagnosis; however, the true sensitivity or specificity of this modality to determine the presence of steatosis and exclude other causes is still unknown. Other imaging studies such as CT or MRI have also been used to reveal fat density of the liver, although once again the sensitivity of these modalities to diagnose steatosis in CF is unknown. This lesion is relatively common in CF, occurring in 20–60% of affected patients depending on the study. It is unclear whether this is a result of abnormal CFTR function in cholangiocytes or reflects a secondary effect of malnutrition or deficiency of trace element or minerals. Steatosis may resolve with improved nutritional status and correction of trace mineral, vitamin, or fatty acid deficiencies. Although this lesion is felt to be benign and non-progressive, the recent interest in non-alcoholic steatohepatitis as a cause of cirrhosis in adults may lead to a reappraisal of this belief.

Figure 26.5 (A) Steatosis in cystic fibrosis. Cut section of the liver revealing extensive fatty infiltration. Photomicrograph of a histologic section from the liver. Histology reveals extensive micro- and macrovesicular fat (hematoxylin & eosin stain).

(B) Multilobular cirrhosis in cystic fibrosis. Cut section of the liver revealing significant lobulation, fibrosis, scarring, and cirrhosis.

(C) Focal biliary cirrhosis in cystic fibrosis. In image (C1), the gross surface of the liver displays focal areas of scarring and furrowing. Large areas of normal preserved hepatic architecture are present. (C2) Photomicrograph of a histologic section from liver above. The portal tract is expanded with bile duct proliferation and plugging of ducts with “eosinophilic material.” Cholestasis and significant bands of fibrosis are present. (Hematoxylin & eosin staining.)

(Courtesy of Dr. Arthur Weinberg, University of Texas Southwestern Medical Center, Dallas, Texas.)

Focal Biliary Cirrhosis and Multilobular Cirrhosis

Focal biliary cirrhosis is characterized histologically by focal areas of portal inflammation and fibrosis, bile duct obstruction and proliferation, and the inclusion of eosinophilic material in bile ductules (Figure 26.5C2). This lesion is considered pathognomonic of CF liver disease. The focal areas of fibrosis may give the liver a furrowed appearance (Figure 26.5C1). The pink-staining eosinophilic material seen in the bile ductules, as well as the focal nature of the lesion, provides more evidence for the “bile duct plugging” hypothesis as contributing to the pathogenesis of this disease. The clinical diagnosis of focal biliary cirrhosis is difficult as both the physical examination and biochemical evaluation may be normal.

Autopsy studies indicate an increasing incidence of this lesion with increasing age, and focal biliary cirrhosis may progress into the more severe multilobular cirrhosis with portal hypertension or liver failure. It is not known why a subset of patients with focal biliary cirrhosis will progress to more severe liver disease and eventually multilobular cirrhosis. Multilobular cirrhosis is characterized histologically by extensive, broad bands of fibrosis extending between portal areas (Figure 26.5B). The liver is extensively lobulated, and within the individual lobules, both focal areas of scarring and intervening areas of normal hepatocyte parenchyma are present. Physical examination reveals a multilobulated and firm liver; in fact, the extensive lobulation is characteristic of this lesion. Signs of chronic liver disease such as clubbing, spider angiomata, and palmar erythema may be present. The identification of splenomegaly or ascites may herald the development of portal hypertension. Prospective studies have suggested prevalence rates of multilobular cirrhosis as high as 17%, and the majority of patients identified were younger than 14 years of age [48]. Patients with multilobular cirrhosis are at risk for complications of end-stage liver disease and portal hypertension, including esophageal varices, ascites, encephalopathy, fatigue, splenomegaly, hypersplenism, and coagulopathy. In fact, complications from portal hypertension cause the majority of morbidity with this liver lesion.