Introduction

Recent studies established the adiponutrin (PNPLA3) variant p.Ile148Met as common genetic risk factor for severe forms of chronic liver diseases1. Indeed, patients with non-alcoholic2,3 and alcoholic fatty liver disease4,5 who carry the p.148Met allele at this locus are prone to progressive liver fibrosis and cirrhosis. In our elastography-based analysis6, we demonstrated that this allele is associated with increased liver fibrosis in patients with chronic liver diseases in general. PNPLA3 has been reported to be a triacylglycerol hydrolase or lysophosphatidic acid (LPA) acyltransferase and the p.Ile148Met variant might affect the remodelling of lipids, including the conversion of LPA into phosphatidic acid1,7. Carriers of the p.148Met allele might have have increased LPA catabolism, which promotes the synthesis of diacylglycerol and modulates the composition of lipid droplets.

To date, the treatment of pruritus in patients with liver diseases remains a troublesome challenge. Although several drugs (ursodeoxycholic acid, rifampicin) and invasive approaches (nasobiliary drainage, plasmapheresis) represent therapeutic options8, they are not effective in many patients. Clinical observations support the notion that patients with comparable grade of cholestasis display different severity of pruritus9. Thus genetic predisposition might affect the degree of itching, but to date no common pruritus-related genes that are associated with treatment-refractory pruritus have been identified. In addition to liver, skin also shows high PNPLA3 expression levels10. Interestingly, LPA has been identified as the critical mediator of cholestatic pruritus11. Indeed, LPA binds to the LPA1 receptor in skin and causes itching11. Patients with cholestatic pruritus are characterized by increased serum concentrations of LPA, which can be generated from lysosphatidylcholine by the enzyme autotoxin in blood11. Therefore we hypothesized that the polymorphism p.Ile148Met of PNPLA3, which catabolizes LPA, affects the severity of cholestatic pruritus and investigated two independent cohorts of patients with cholestatic liver diseases (Tables 1 and 2).

Table 1 Demographic, clinical and laboratory data in PBC patients and controls
Full size table
Table 2 Demographic, clinical and laboratory data in ICP patients and controls
Full size table

Results

As shown in Figure 1 and Supplementary Table 1, the intensity of itch differed significantly between carriers of different genotypes of the PNPLA3 variant p.Ile148Met (ANOVA P = 0.048). Moreover, the PNPLA3 allele p.148Ile conferred a significant risk (OR = 3.94, 95% CI = 0.91–17.00, P = 0.048) to require plasmapheresis (Figure 2, Table 3). On the other hand, this variant did not increase the risk of developing PBC per se (Supplementary Table 2), or cirrhosis in PBC patients (Supplementary Table 3) and was neither associated with laboratory parameters of PBC (Supplementary Table 4) nor with domains of the PBC-40 questionnaires other than the intensity of pruritus (Supplementary Table 1).

Table 3 Distribution of PNPLA3 alleles and genotypes in PBC patients stratified according to the need for plasmapheresis to relieve pruritus
Full size table
Figure 1
figure 1

Itch intensity in PBC patients stratified for PNPLA3 p.Ile148Met genotypes.

Carriers of different PNPLA3 p.Ile148Met genotypes report significantly different intensity of pruritus (ANOVA P = 0.048), as quantified by the itch domain of the PBC-40 questionnaire.

Full size image
Figure 2
figure 2

PNPLA3 p.Ile148Met genotypes in PBC patients who did not and patients who required plasmapheresis for the treatment of pruritus.

Patients who were treated with plasmapheresis carry the PNPLA3 allele p.148Ile significantly (allelic 1-df test P = 0.048) more often than patients who did not require this treatment.

Full size image

Consistent with these observations, we detected a significantly increased frequency of the PNPLA3 allele p.148Ile in a second large independent cohort of patients with ICP who presented with itch during pregnancy as compared to controls (Figure 3, Supplementary Table 5). As illustrated in Supplementary Figure 1, this lead to a departure of the genotype distribution from Hardy-Weinberg equilibrium in ICP patients, supporting the genetic association. Patients carrying the PNPLA3 allele p.148Met were at significantly decreased risk (OR = 0.66, 95% CI = 0.47–0.92, P = 0.013) of presenting with ICP. Since pruritus is the major symptom at diagnosis of ICP, the lower frequency of the p.148Met allele in this cohort replicates the observation that this allele decreases the degree of cholestatic itch.

Figure 3
figure 3

PNPLA3 p.Ile148Met genotypes in ICP patients and sex-matched controls.

Patients diagnosed with ICP present a significantly (allelic 1-df test P = 0.013) higher frequency of the PNPLA3 allele p.148Ile as compared to controls.

Full size image

Discussion

This is the first study demonstrating an association of a genetic polymorphism with cholestatic pruritus. Here we report that the PNPLA3 allele p.148Met decreases itch severity in cholestatic patients. Our results are in line with functional analyses implicating that carriers of this allele display an increased metabolism of LPA7. We hypothesize that increased LPAAT7 activities in liver and/or skin10 of patients carrying the PNPLA3 allele p.148Met might be the mechanism that decreases itching, however further functional analyses are required. Moreover, our study identifies this PNPLA3 variant as potential genetic marker for therapy-refractory pruritus. Interestingly, our recently reported young female patient who developed severe refractory pruritus after acute hepatitis A infection as a result of hepatobiliary transporter variants12 is also a homozygous carrier of the susceptible PNPLA3 allele p.148Ile (M.K. and F.L., data on file).

The association between the PNPLA3 polymorphism and pruritus was independent from potential effects on liver function2,3,4,5,6,13 (Supplementary Tables 3 and 4). The lack of association of the PNPLA3 variant with liver cirrhosis in the PBC cohort indicates that the skin might be the major site where the antipruritic effects are exerted. Hence we conclude that the PNPLA3 variant p.Ile148Met is an example of biological pleiotropy14,15 with influence on more than one liver-related trait: In patients with chronic liver diseases the allele p.148Met is associated with disease progression16, but in case of cholestasis it might protect against itch. However, our results require further replication in additional cohorts. Since the PNPLA3 allele p.148Ile may to a certain extent explain the development of therapy-refractory pruritus, genotyping of the PNPLA3 variant p.Ile148Met might be included in the diagnostic work-up of patients with cholestatic liver conditions.

Patients and Methods

Patients with primary biliary cirrhosis (PBC)

In total, we recruited 187 Polish PBC patients (age range 22–83 years, 166 females). All patients fulfilled the European Association for the Study of the Liver (EASL) criteria for the diagnosis of PBC8. Table 1 presents the detailed description of this cohort. Liver function tests were determined by standard assays in fasted blood samples. In 135 patients, quality of life and intensity of itch were prospectively assessed with the PBC-40 questionnaire17. Sixty-nine patients presented with histological, clinical and/or imaging features characteristic for liver cirrhosis. Among the PBC patients, a total of 13 who did not respond to pharmacological treatment of their pruritus (including ursodeoxycholic acid, colestyramine and rifampicin) were treated with plasmapheresis. The control cohort encompassed 250 healthy blood donors from the National Blood Services (Table 1)18.

Patients with intrahepatic cholestasis of pregnancy (ICP)

A cohort of 201 females with ICP (age range 17–46 years) was recruited between 2000 and 2013. Table 2 summarizes the details of this cohort. The included patients fulfilled the EASL criteria for the diagnosis of ICP8. The control cohort consisted of 198 non-pregnant females (age range 20–60 years) without any documented episodes of cholestasis during pregnancy. The study was conducted according to a study design approved by the local ethical committees and informed consent was obtained from all study participants.

Genotyping

In all individuals, we genotyped the PNPLA3 variant p.Ile148Met (rs738409) as described6. Genomic DNA was isolated from EDTA anticoagulated blood using the membrane-based QIAamp DNA extraction protocol (Qiagen, Hilden, Germany).

Statistical analyses

The consistency of genotype frequencies with Hardy-Weinberg equilibrium (HWE) was tested using an exact test. Allele frequency differences were assessed by 1-df chi2 tests (http://ihg.gsf.de/cgi-bin/hw/hwa1.pl). The study hypothesis was tested at a two-sided p-value of 0.05. For further exploratory analyses, quantitative phenotypic data were analyzed using Mann-Whitney U or Fisher's protected least significant difference (PLSD) tests and qualitative phenotypes were assessed in contingency tables.