A frequent variant in the human bile salt export pump gene ABCB11 is associated with hepatitis C virus infection, but not liver stiffness in a German population

Hepatitis C virus (HCV) is one of the most common causes of fibrotic liver disease. It is estimated that there are 5 million carriers in Europe, and up to 500.000 chronically HCV-infected people in Germany [1]. These individuals are at high risk of developing progressive liver disease, including cirrhosis and hepatocellular carcinoma [2, 3]. Genetic factors influence susceptibility to HCV infection but also progression of the disease. Due to the viral life cycle and its reliance on the infected cell, host genetic factors are likely to exert a strong influence on these processes. This has been shown for example in a study compromising Irish women infected with an HCV-contaminated anti-D immunoglobulin acquired from the same acutely infected person [4, 5]. The analysis of liver samples acquired in average almost two decades after the infection showed a highly variable disease progression, which was apparently not affected by the type of the hepatitis C virus itself but by the host genetic background and lifestyle [4, 5]. A number of genetic studies have reported associations of variants in candidate genes such as tumor necrosis factor [6‐8], interleukin 10 [9], HLA subtypes [10‐12] and CC chemokine receptor 5 [13] with susceptibility and progression of HCV infection. Huang et al. [14] used a machine learning approach and two independent Caucasian cohorts to identify a seven-gene signature that allows better prognosis of fibrosis progression in HCV infected individuals than established clinical parameters. This study and the follow-up by Marcolongo et al. in an Italian population [15] illustrate the feasibility of genetic risk assessment for cirrhosis prediction. They also accentuate the need to identify additional genetic factors that contribute towards HCV susceptibility and subsequent fibrotic progression. A comprehensive map of cirrhosis-associated variants might permit verification of interacting and modifying loci in selected patient groups, which would in turn improve the predictive values of polygenic cirrhosis risk scores.

Previously a common, non-conservative polymorphism c.1331 T > C (p.V444A) in the hepatobiliary bile salt transporter ABCB11 (also known as bile salt export pump, BSEP) was identified as a risk factor for cholestatic liver diseases, in particular drug-induced cholestasis [16] and intrahepatic cholestasis of pregnancy [17, 18]. The human bile salt export pump represents the central canalicular transporter of bile salts from hepatocytes into the bile canaliculus [19, 20]. In normal conditions, ABCB11 together with other hepatobiliary proteins (i.e. ABCB4 and ABCG5/8 transporting phospholipids and sterols into bile, respectively) maintain proper concentrations of biliary constituents. This, in turn, preserves bile flow and minimises the toxic properties of biliary components. Interestingly, the analysis of hepatobiliary transporter expression has demonstrated that the functional effect of the p.A444V polymorphism is due to markedly lower expression of ABCB11 [21, 22] rather than impaired function of the protein [16].

Lately two reports have underscored an association between the c.1331 T > C variant and both response to antiviral treatment [23] and the presence of cirrhosis in patients with chronic HCV infection [24]. The analysis of 151 individuals treated with ribavirin and pegylated interferon demonstrated that those carrying the genotype [TT] and infected with HCV genotypes 2 or 3 show increased sustained virological response (SVR) rates as compared to [CC] individuals [23]. Additionally, the [CC] genotype has been reported to increase the risk of developing liver cirrhosis in chronic HCV infection [24].

To validate and extend previous results concerning possible associations of the common polymorphism c.1331 T > C in the bile salt export pump with HCV infection and fibrosis progression, we analysed its genotype distribution in a larger cohort of 649 patients with chronic HCV infection, in 444 of whom liver fibrosis was staged non-invasively by transient liver elastography.

The present study shows that a common variant in the human bile salt export pump ABCB11 is overrepresented in patients with HCV infection. The reason underlying this observation is speculative, although it might be caused by differences in spontaneous recovery from infection and sustained virological response (SVR) rates, as was observed in patients bearing HCV genotypes 2 and 3 in a previous study: Iwata et al. [23] were able to show that the effect might be mediated by serum bile salt concentrations, with low levels (< 8 μM) being associated with SVR in their cohort. This is not the first time that bile salts or genes involved in their metabolism have been implicated in the pathogenesis of HCV infection. Ursodeoxycholic acid has been previously proven in various trials to be beneficial in patients suffering from liver damage due to chronic infection, albeit not via modulation of SVR[27‐29]. Contrastingly, increased serum levels of bile salts are associated with negative outcomes in HCV therapy [30]. The observed association between a polymorphism that impinges bile salt transport from the hepatocyte into the canaliculus offers an explanation of this conundrum: Viral replication might be boosted inside the hepatocyte by relatively high bile salt concentrations [30]. A variant of the bile salt export pump with lower expression level resulting in reduced transport activity, such as the common [C] allele of c.1331 T > C, might lead to transient increases in hepatocellular bile salt concentration, depending on the efficiency of the sinusoidal import.

Ursodeoxycholic acid as rather weak FXR agonist is known to increase the expression and activity of bile salt transporters by transcriptional as well as multiple post-transcriptional mechanisms and to contribute to SHP-mediated inhibition of bile salt synthesis [31, 32]. In vitro experiments using HCV replicon-harboring cells have shown that the impact of bile salts on HCV replication is mediated by the action of FXR rather than via direct effects. FXR antagonisation by guggulsterone blocked the bile salt-induced up-regulation of virus replication; furthermore, guggulsterone inhibited basal levels of HCV replication [30, 33]. Hence, it is possible that HCV "hijacks" or uses transcriptional activation via FXR. Whatever the mode of action by which HCV takes advantage of FXR signalling, it appears to be advantageous for the hepatotropic virus to use central functions of hepatocytes for its own end.

HCV replication is tightly linked to lipid metabolism, and bile salts are crucially involved in fat absorption and transport. This reinforces the key roles of bile salt responsive regulatory molecules such as FXR and TGR5 as potential host factors modulating HCV replication, not necessarily through direct effect of bile salts but via feedback regulation of lipid metabolism. In infected cells, HCV core protein accumulates on the surface of lipid droplets [34, 35]. Other functional proteins, such as RNA binding replicase NS5A, are targeted to lipid droplets, and this targeting appears to be crucial for HCV assembly [36]. A recent combined transcriptome and proteome analysis revealed the impact of HCV infection on previously unknown metabolic pathways, particularly an increase in cellular cholesterol and free fatty acid levels [37]. One possible reason for this "metabolic reprogramming" might be the dependency of HCV replication on lipid metabolism as outlined above. Considering the central role of ABCB11 in bile flow, it is not inconceivable to conclude that common variants of this transporter are associated with variations in plasma cholesterol levels. In fact, a meta-analysis of genome-wide association scans supports this reasoning [38], thus providing further evidence for another potential connection between bile salt transport and HCV replication. The results from our study are indicative of a role for the major allele of the ABCB11 c.1331 variant in conveying susceptibility to persistent HCV infection, enhancing this link.

Iwata and colleagues observed an overrepresentation of the [C] allele at ABCB11 c.1331 in HCV patients (n = 206) compared with controls (n = 110) in the Swiss Hepatitits C cohort (allelic frequency = 62.9%), although no significant difference in median bile salt serum concentration could be detected between bearers of homozygous [C] or [T] alleles [24]. However, a comparison of fibrosis scores from 178 patients revealed a skewed distribution: Carriers of homozygous [C] allele were overrepresented among cirrhotic patients, and the [CC] genotype was an independent risk factor for cirrhosis in multivariate analysis (p = 0.010).

Our investigations into a connection between ABCB11 c.1331 T > C and liver stiffness as measured by TE (Fibroscan) did not reveal any significant associations. This may be due to methodolocial differences in assessing fibrosis/cirrhosis [25, 26]. Scatterplot analysis of TE results from the different genotypes seems to indicate a prevalence of lower values in patients bearing the [TT] genotype (Figure 1), which would be in agreement with the data from the Swiss cohort, but the numbers of homozygous carriers of the minor allele in our cohort were too low to reach statistical significance.

Cirrhosis was significantly associated with the age of patients (p = 0.006), which may reflect longer duration of chronic HCV infection. However, since the exact time point of infection was unknown in the majority of our patients, we used age as surrogate marker in our statistical analyses.

Interestingly, when comparing allelic distribution in our cohort with figures from the Swiss Hepatitis C cohort [24], ABCB11 c.1331 T > C is equally skewed towards an overrepresentation of the [C] allele among British patients with non-alcoholic fatty liver disease in the study by Iwata et al., pointing towards a potential role of ABCB11 via its impact on lipid metabolism rather than viral replication. The development of diet-induced obesity and hypercholesterolemia following hepatocye-specific overexpression of ABCB11 in mice is a pointer in the same direction [39].

Based on the assumption that populations from both the Swiss study and our cohort are from a similar ethnic background, we speculate that the major [C] allele of ABCB11 c.1331 T > C might be a high-frequency low-risk contributor towards susceptibility for various complex liver diseases triggered by external stimuli and involving lipid metabolism [40].