ORIGINAL ARTICLE—LIVER, PANCREAS, AND BILIARY TRACT

Cryptogenic cholestasis in young and adults: ATP8B1, ABCB11,ABCB4, and TJP2 gene variants analysis by high-throughputsequencing

Giovanni Vitale1 • Stefano Gitto1 • Francesco Raimondi3,4 • Alessandro Mattiaccio2 •

Vilma Mantovani2 • Ranka Vukotic1 • Antonietta D’Errico5 • Marco Seri1 •

Robert B. Russell3,4 • Pietro Andreone1,6

Received: 25 October 2017 / Accepted: 4 December 2017

� Japanese Society of Gastroenterology 2017

Abstract

Background Mutations in ATP-transporters ATPB81,

ABCB11, and ABCB4 are responsible for progressive

familial intrahepatic cholestasis (PFIC) 1, 2 and 3, and

recently the gene for tight junction protein-2 (TJP2) has

been linked to PFIC4.

Aim As these four genes have been poorly studied in

young people and adults, we investigated them in this

context here.

Methods In patients with cryptogenic cholestasis, we

analyzed the presence of mutations by high-throughput

sequencing. Bioinformatics analyses were performed for

mechanistic and functional predictions of their conse-

quences on biomolecular interaction interfaces.

Results Of 108 patients, 48 whose cause of cholestasis was

not established were submitted to molecular analysis.

Pathogenic/likely pathogenic mutations were found in ten

(21%) probands for 13 mutations: two in ATP8B1, six in

ABCB11, two in ABCB4, three in TJP2. We also identified

seven variants of uncertain significance: two in ATP8B1,

one in ABCB11, two in ABCB4 and two in TJP2. Finally,

we identified 11 benign/likely benign variants. Patients

with pathogenic/likely pathogenic mutations had higher

levels of liver stiffness (measured by FibroScan�) and bile

acids, as well as higher rates of cholestatic histological

features, compared to the patients without at least likely

pathogenic mutations. The multivariate analysis showed

that itching was the only independent factor associated

with disease-causing mutations (OR 5.801, 95% CI

1.244–27.060, p = 0.025).

Conclusions Mutations in the genes responsible for PFIC

may be involved in both young and adults with cryptogenic

cholestasis in a considerable number of cases, including in

heterozygous status. Diagnosis should always be suspected,

particularly in the presence of itching.

Keywords Progressive familial intrahepatic cholestasis �Cryptogenic disease � Pathogenic mutations � Geneticvariants � Bioinformatics analysis

Abbreviations

PFIC Progressive familial intrahepatic cholestasis

TJP2 Tight junction protein-2

FIC1 Familial intrahepatic cholestasis 1

BSEP Bile salt export pump

MDR Multidrug resistance P-glycoprotein 3

Electronic supplementary material The online version of thisarticle (https://doi.org/10.1007/s00535-017-1423-1) contains supple-mentary material, which is available to authorized users.

& Pietro Andreonepietro.andreone@unibo.it

1 Department of Medical and Surgical Sciences, University of

Bologna, Bologna, Italy

2 Center for Applied Biomedical Research (CRBA), University

Hospital, Bologna, Italy

3 CellNetworks, Bioquant, Heidelberg University, Im

Neuenheimer Feld 267, 69120 Heidelberg, Germany

4 Bioochemie Zentrum Heidelberg (BZH), Heidelberg

University, Im Neuenheimer Feld 328, 69120 Heidelberg,

Germany

5 Addari Institute of Oncology and Transplant Pathology,

Policlinico S. Orsola-Malpighi, University of Bologna,

Bologna, Italy

6 Department of Medical and Surgical Sciences and Research

Center for the Study of Hepatitis, University of Bologna,

Italy, Via Massarenti 9, 40138 Bologna, Italy

123

J Gastroenterol

https://doi.org/10.1007/s00535-017-1423-1

http://orcid.org/0000-0002-4794-9809https://doi.org/10.1007/s00535-017-1423-1http://crossmark.crossref.org/dialog/?doi=10.1007/s00535-017-1423-1&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1007/s00535-017-1423-1&domain=pdfhttps://doi.org/10.1007/s00535-017-1423-1

GGT Gamma-glutamyl-transpeptidase

AP Alkaline phosphatase

BRIC Benign intrahepatic cholestasis

LPAC Low-phospholipid-associated cholelithiasis

ICP Intrahepatic cholestasis of pregnancy

DIC Drug-induced cholestasis

HTS High-throughput sequencing

NGS Next-generation sequencing

PSC Primary sclerosing cholangitis

BA Bile acids

MAF Minor allele frequency

SIFT Sorting Intolerant From Tolerant

HGMD Human Gene Mutation Database

ACMG American College of Medical Genetics and

Genomics

P Pathogenic

LP Likely pathogenic

VUS Variants of uncertain significance

LB Likely benign

B Benign

SD Standard deviation

CI Confidence interval

SNP Single-nucleotide polymorphism

ALT Alanine aminotransferase

OR Odds ratio

Introduction

Progressive familial intrahepatic cholestasis (PFIC) is a

group of autosomal recessive cholestatic diseases that

affects especially newborns and children, and represents a

consolidated indication for liver transplantation [1]. These

disorders are rare and unavoidably progressive to liver

cirrhosis and portal hypertension. However, their incidence

is hard to establish because of difficulties in diagnosis [1].

Mutations in four genes have been linked to PFIC. For

instance, mutations in ATP8B1, which encodes the

aminophospholipid flippase familial intrahepatic cholesta-

sis-1 protein (FIC1), are responsible for PFIC1 [2]. Muta-

tions in ABCB11, encoding a protein that functions as a bile

salt export pump (BSEP), are responsible for the PFIC2 [3].

Mutations in ABCB4, which codes for the multidrug

resistance P-glycoprotein 3 (MDR3), a flippase that medi-

ates the outflow of phosphatidylcholine into the bile, are

linked to PFIC3 [4]. Here, levels of gamma-glutamyl-

transferase (GGT) are always increased [5], in contrast to

very low values that characterize PFIC1 and PFIC2, in

which high levels of alkaline phosphatase (AP) are instead

present. Mutations in a fourth gene, TJP2 (coding for tight

junction protein-2) were linked to intrahepatic cholestasis

with low GGT (PFIC4) [6].

Variants in the same genes cause several other liver

diseases. Benign intrahepatic cholestasis (BRIC) is linked

to mutations in ATP8B1 and ABCB11, characterized by

intermittent cholestasis without progression to cirrhosis.

Low-phospholipid-associated cholelithiasis (LPAC) is

related to mutations in ABCB4 with symptomatic intra-

hepatic lithiasis. Intrahepatic cholestasis of pregnancy

(ICP) is a reversible pregnancy-specific cholestasis char-

acterized by pruritus, elevated liver enzymes, and increased

serum bile acids involving mutations in ATP8B1, ABCB11,

and ABCB4. Mutations in ABCB11 and ABCB4 are asso-

ciated with drug-induced cholestasis (DIC), a disorder

induced by certain drugs. These liver diseases are often

associated to heterozygous mutations [7, 8].

Only heterozygous status for ABCB4 mutations has been

well studied in children with cholestatic liver disease

[9, 10] while Dröge and colleagues [11] recently revealed a

high number of different genetic variants by sequencing

FIC1, BSEP, and MDR3 in a large cohort of patients with

supposed genetic cholestasis.

To date, no authors have investigated the four genes

linked to FIC in young persons and adults, with and

without progressive forms of liver failure.

High-throughput sequencing (HTS), including next-

generation sequencing (NGS), is a technology proposed for

the molecular diagnosis of PFIC, based on the massive

parallel sequencing of specific genomic loci, whole exome

or genomes. Compared to classic Sanger, NGS allows rapid

sequencing with more information at lower costs

[6, 12, 13].

We aimed to develop a targeted NGS panel to investi-

gate the presence of mutations in ATP8B11, ABCB11,

ABCB4, and TJP2 in a population with cryptogenic

cholestasis and related them to the corresponding pheno-

types and risk factors.

Materials and methods

Patients

From May 2013 to November 2016, all outpatients with

cholestatic disease aged[ 6 years were enrolled consecu-tively at the Department of Medical and Surgical Sciences,

a tertiary Italian referral center. We excluded other causes

of cholestasis as follows; primary biliary cholangitis, pri-

mary sclerosing cholangitis (PSC), overlap syndromes,

IgG4-cholangiopathy and obstructive jaundice were

excluded by demonstration of a normal anatomy of the

biliary tree and negative specific serological tests. Viral

hepatitis, alcohol abuse, hemochromatosis, Wilson disease,

J Gastroenterol

123

alfa1-antitrypsin deficiency were excluded too. Chronic

idiopathic cholestasis was defined as GGT e/o AP persis-

tently C 1.5-fold the upper normal values in at least two

tests or as history of itching combined with elevated serum

bile acids (BA) concentration ([ 10 lmol/l) for more than6 months. This study was conducted in accordance with

ethical guidelines of the World Medical Association’s

Declaration of Helsinki and patients or their legal guar-

dians provided written informed consent. Subjects affected

by cryptogenic cholestasis underwent laboratory analysis,

including BA serum concentration, liver fibrosis evaluation

by transient elastography (FibroScan�) and, if indicated,

liver biopsy within 6 months from the execution of the

genetic tests. Histological bile duct alteration was defined

as lobular cholestasis (ductal hepatocyte metaplasia, ductal

proliferation and immunohistochemistry for bile duct

cytokeratin 7, anti-BSEP, and anti-MDR3 antibodies). An

expert senior pathologist performed the liver histology.

Clinical variables considered were history of DIC or

itching, family history of cholestasis, personal or family

history of ICP, neonatal jaundice, juvenile cholelithiasis

(defined as history of gallstones\ 40 years).We compared patients with disease-causing variants

with the remaining population with idiopathic cholestasis.

NGS analysis

We extracted DNA from peripheral blood using the Max-

well 16 blood DNA purification kit (Promega, Madison,

WI, USA). Comprehensive molecular analysis of ABCB11,

ATP8B1, ABCB4, and TJP2 genes was performed by

multiplex targeted amplicon-based sequencing approach

using the Ion Torrent technology (Thermo Fisher Scien-

tific, Waltham, MA, USA). Primers were designed by Ion

AmpliSeqTM Designer tool to cover the entire coding

region of the four genes plus 50 bp of intronic flanking

regions. Target coverage of 98.9% was obtained through

194 amplicons for 42 kb. Libraries were performed by

AmpliseqTM Library Kit 2.0 and concentrations were

evaluated by the Ion Library TaqManTM quantitation kit

using real-time PCR. Emulsion-PCR was performed by Ion

PGMTM Hi-Q OT2 kit and the enrichment was realized by

Ion OneTouchTM ES. NGS was performed by the Ion

Torrent PGMTM System according to the manufacturer’s

procedures (Life Technologies, CA, USA). Molecular

analysis was simultaneously carried out in 12 barcoded

samples in an Ion 318 Chip v2 for each run.

The read files obtained from sequencing were mapped to

the GRCh37/hg19 assembly and the sequence variants

were identified by Variant Caller and Ion Reporter

software.

Allelic variants were reported according to Human

Genome Variation Society guidelines (http://www.hgvs.

org/content/guidelines). NGS uncovered regions as well as

the potential pathogenic variants were confirmed by Sanger

sequencing using ABI PRISM 3730 Genetic Analyzer

(Thermo Fisher Scientific, Waltham, MA, USA).

Prediction of functional consequences of variants

and classification

The filtering step of sequence variants was carried out

according to default criteria for germline mutations:

insertions or deletions, non-sense variants, splicing-site

variants, and missense with minor allele frequency

(MAF) B 0.05 were considered. Three bioinformatics tools

were used to predict the role of missense variants. Sorting

Intolerant From Tolerant (SIFT) (http://www.jcvi.org/cms/

home/), PolyPhen-2 [14] and Mutation Taster [15] software

can evaluate whether an amino-acid substitution influences

protein structure and function, according to physical

modifications and the degree of conservation of protein

sequence among species. Variants that meet at least one of

the following criteria were also considered: ever described

before, reported in Human Gene Mutation Database

(HGMD), predicted not benign by at least one of the

bioinformatic tools. According to American College of

Medical Genetics and Genomics (ACMG) standards [16],

we classified the filtered variants into five categories:

pathogenic (P); likely pathogenic (LP); variants of uncer-

tain significance (VUS); likely benign (LB); and benign

(B).

Variants passing the above filtering process were map-

ped to Uniprot canonical sequences and subjected to

Mechismo [17] analysis to predict their functional conse-

quences at biomolecular interaction interfaces. The

approach matches protein sequence amino-acids to posi-

tions within structures and identifies sites affecting inter-

actions with other proteins, DNA/RNA or small molecules.

We considered low confidence predictions including

known structures or close (C 30% sequence identity)

homologs and only very confident, physical protein–pro-

tein interactions (as defined by Mechismo based on a

benchmark for the accuracy of perturbed interfaces). In

case of ATP8B1, no homolog template structure was

available in Mechismo, so we used a homology model rom

ModBase [18].

We annotated mutations and PTMs (phosphorylations

and acetylations) from Phosphosite [19] of the considered

genes on corresponding Uniprot canonical sequences using

lollipop diagrams [20].

Comparison with an international database

Allele frequency (AF) of common single-nucleotide poly-

morphism (SNPs) was matched with data reported in the

J Gastroenterol

123

http://www.hgvs.org/content/guidelineshttp://www.hgvs.org/content/guidelineshttp://www.jcvi.org/cms/home/http://www.jcvi.org/cms/home/

international Genome Aggregation Database (gnomAD),

Cambridge, MA (http://gnomad.broadinstitute.org/;37)

[allele frequencies accessed October 2017]. We compared

AF of SNPs in our cohort (with and without P/LP mutation)

to AFs of European (non-Finnish), East Asian, and

worldwide population.

Statistical analysis

Categorical variables are expressed as number (%), and

quantitative variables as mean ± standard deviation or as

median (range). Chi-square or Fisher’s exact test was used

to compare categorical variables, while for quantitative

variables the t test or Mann–Whitney’s test (unpaired data)

or the t test or Wilcoxon’s test (paired data) were used.

Binary logistic regression was performed for univariate and

multivariate analyses to identify predictors of causative

variants (variables were included if p\ 0.1 and removed ifp C 0.05). A p\ 0.05 was considered significant for alltests. The statistical software SPSS version 21.0 (�SPSS

Inc., Chicago, IL, USA) was used for statistical analyses.

Results

Patient characteristics

We evaluated consecutively 108 cholestatic patients and 48

fulfilled the established criteria (Supplementary Fig. 1).

We listed the main laboratory patterns in Table 1. There

was a slight bias for males (58%) and mean age at time of

the genetic test was 42 years; the adult population

(C 18 years) was 93.8%. A history of familial cholestatic

diseases and DIC was present in 35%, itching in 27%,

neonatal jaundice in 21%, juvenile cholelithiasis in 17%,

personal or family history of ICP in 13%. Histologic fea-

tures of cholestasis were present in 65% of cases (overall

44 patients agreed to a liver biopsy). Gene analysis

revealed the presence of P/LP mutations in about one-

fourth of subjects; clinical features and type of mutations

found in the 11 patients (ten unrelated probands and one

affected sister) were reported in Table 2.

NGS results and variants

NGS protocol provided an average sequencing depth

of * 10009, 98.78% of reads having 209 coverage, with99% reads on target and a coverage uniformity of 95.98%.

Thirty-one variants that satisfied our filtering criteria are

reported in Supplementary Table 1. Among these, 13 were

P/LP mutations, five were previously undescribed: three

were missense, one was a frameshift, and one was a non-

sense mutation.

Among P/LP variants, two were in ATP8B1, six in

ABCB11, two in ABCB4 and three in TJP2. We identified

seven VUS, two in each of ATP8B1, ABCB4 and TJP2, one

in ABCB11. Finally, we recorded 11 benign mutations, two

in each of ATP8B1, ABCB11 and ABCB4, five in TJP2.

ATP8B1 variants

Six variants were identified in the coding region of

ATP8B1, reported below; according to ACMG standards,

two LP: c.3655G[C (p.D1219H) and c.68C[T (p.P23L);the first was combined with heterozygous LP mutation

c.1057C[T on TJP2 gene in a 71-year-old woman (case 1).The second is a novel mutation found in a 31-year-old male

(case 2) who experienced neonatal jaundice and juvenile

cholelithiasis, with high GGT and bilirubin levels. We

classified the missense c.134A[C (p.N45T) and c.607A[G(p.K203E) as VUS and each was previously described as

risk allele in ICP [21]. We found these variants in com-

pound heterozygous in a 36-year-old female who showed a

Table 1 Baseline features of the entire study population

N 48

Male N (%) 28 (58.3)

Adult population C 18 years N (%) 45 (93.8)

Age (years, mean ± SD)

At time of genetic test 42 ± 15

At presentation 32 ± 15

Risk factor for cholestasis N (%)

DIC history 17 (35.4)

Neonatal jaundice 10 (20.8)

Itching history 13 (27.1)

ICP history 6 (12.5)

Juvenile cholelithiasis 8 (16.7)

Familiarity 17 (35.4)

Laboratory (median, range)

GGT (UI/l) 139 (5–597)

FA (UI/l) 283 (122–1012)

ALT (UI/l) 45 (9–387)

Bilirubin (mg/dl) 0.7 (0.4–11.2)

Bile acids (lmol/l) 11 (2.3–403)

Cholesterol (mg/dl) 211 (87–328)

Albumin (g/dl) 4.3 (2.9–5.2) 4.3 (2.9–5.2)

Platelets (n 9 103/ll) 229 (61–417)

FibroScan (kPa) (median, range) 5.3 (3.1–35.8)

Histologic features of cholestasis N (%) 31 (64.6)

N number, SD standard deviation, DIC drug-induced cholestasis, ICP

intrahepatic cholestasis of pregnancy, GGT gamma-glutamyltrans-

ferase, AP alkaline phosphatase, ALT alanine transaminases, kPa

kiloPascal

J Gastroenterol

123

http://gnomad.broadinstitute.org/%3b37

Table 2 Baseline features of the population with likely pathogenic (LP) and pathogenic (P) mutations

ID Sex LP and P mutations Additional variants Age

(years)

GGT

(U(L)

PA

(U/l)

Bilirubin

(mg/dl)

FibroScan

(kPa)

Notes

1_17421 F ATP8B1:

p.[D1219H]; [=]

TJP2:

p.[R322W]; [=]

ATP8B1:

p.[R952Q]; [=]

ABCB11:

p.[V444A(;)M677V]

TJP2:

p.[M668I]; [=]

71 121 394 1.2 5.5 No affected relatives

Segregation not

assessed

2_15710 M ATP8B1:

p.[P23L]; [=]

ABCB11:

p.[V444A(;)M677V]

31 325 138 3.3 7.4 Familiarity

Juvenile

cholelithiasis

Neonatal jaundice

The mutation was

inherited from the

mother and

segregates in both

brothers and in a

sister

3_16545 F ABCB11:

p.[Y93S(;)V597L(;)R1128]

ABCB11:

p.[V444A]; [=]

ABCB4:

p.[I237=]; [=]

29 5 466 1.2 12 DIC

Neonatal jaundice

Itching

Juvenile

cholelithiasis

ICP

Segregation not

assessed

4_15502 F ABCB11:

p.[A523G]; [A523G]

ABCB11:

p.[V444A];

[V444A]

ATP8B1:

p.[R952Q]; [=]

ABCB4:

p.[I237=]; [=]

20 15 672 11.2 20.9 Familiarity

DIC

Neonatal jaundice

itching

The mutations are

inherited from

healthy parents

4a_15505 F ABCB11:

p.[A523G]; [A523G]

ABCB11:

p.[V444A];

[V444A]

ABCB4:

p.[I237=]; [=]

7 13 1012 3.5 10.1 Familiarity

DIC

Neonatal jaundice

itching

Affected sister of

case 4

The mutations

segregate in both

affected sisters

5_15507 M ABCB11:

p.[E135K]; [S1100Qfs*38]

ABCB11:

p.[V444A];

[V444A]

16 8 467 2.2 11.5 Neonatal jaundice

Itching

The mutations are

inherited from

healthy parents

6_17248 F ABCB4:

p.[K672*]; [=]

ABCB11:

p.[V444A];

[V444A]

ABCB4:

p.[I237=]; [=] TJP2:

p.[T1124=]; [=]

39 363 562 0.4 7.7 Familiarity

DIC

Itching

The mutation was

inherited from the

affected father

J Gastroenterol

123

cholestatic disease characterized by normal GGT and high

BA levels, neonatal jaundice, juvenile cholelithiasis com-

plicated by LPAC, and a history of ICP.

We found two B/LB variants: c.913T[A (p.F305I) andc.2855G[A (p.R952Q). The SNP p.R952Q was found infive cases and in two patients was combined with LP

mutations, one on ABCB11 and one on TJP2, respectively.

AF of p.R952Q was significantly higher in comparison

with AF of East Asian population (10.4 vs. 0.03%:

p\ 0.0001) and comparable with European and worldwidedata (Table 3).

Analysis of variants in the context of sequence and

structural annotations on structures or homology models

revealed that two of them, i.e., p.K203E and p.F305I are

found in spatial proximity within the ATPase catalytic

domain (Fig. 1a), suggesting potential converging effect on

the enzymatic activity, while the p.R952Q is found at the

transmembrane (i.e., phospholipid translocating) domain.

The LP variants, p.P23L and p.D1219H, reside at the N-

and C-terminals of the protein, outside the structured por-

tion (Fig. 1b).

ABCB11 variations

Nine variants were identified in the coding region of

ABCB11 and six resulted in P/LP mutations: three causa-

tive mutations, c.278A[C (p.Y93S), c.1789G[C

(p.V597L) and c.3382C[T (R1128C) were combined in ayoung female (29 years) presenting with a history of

neonatal jaundice, DIC, ICP, juvenile cholelithiasis, itch-

ing, normal GGT levels (case 3). Two sisters (20 and

7 years) were carriers of the LP homozygous mutation

c.1568C[G (p.A523G): both had history of neonataljaundice, itching and DIC, (cases 4–4a). Last two P/LP

mutations, c.3297delC (p.S1100Qfs*38) and c.403G[A(p.E135K), were responsible for a double heterozygosis in

a boy with history of itching and neonatal jaundice, listed

for liver transplantation (case 5); c.3297delC was recently

described [22] while c.403G[A mutation was linked byAnzivino et al. to ICP conditions [23]. The VUS

c.1268A[G (p.H423R) was found in a 43-year-old womanwho presented only high GGT without fibrosis.

Benign mutations detected were among those previously

described [7, 11, 24]: c.2029A[G (p.M677V), detected innine patients, and c.1331T[C (p.V444A), found in 40cases (83.3% of population studied). Of note, SNP

c.1331T[C was reported associated to DIC and ICP inprevious reports [7, 24] and recently to cholestatic phe-

notype in patients without disease-causing mutations in the

respective gene [11]. AF of p.V444A was significantly

more frequent in our cohort of patients without P/LP

mutations in comparison with European and worldwide

populations (81 vs. 60%: p = 0.008; 81 vs. 57%:

p = 0.002). We also observed a higher AF of p.M677V in

Table 2 continued

ID Sex LP and P mutations Additional variants Age

(years)

GGT

(U(L)

PA

(U/l)

Bilirubin

(mg/dl)

FibroScan

(kPa)

Notes

7_15873 M ABCB4:

p.[A364V]; [=]

ABCB4:

p.[I237=]; [=] TJP2

gene:

p.[Q128K(;)M668I]

57 379 606 0.4 35.8 DIC

HCC

No affected relatives.

Segregation not

assessed

8_17362 F TJP2:

p.[T62M]; [=]

ABCB11:

p.[V444A(;)M677V]

ABCB4:

p.[I237=]; [I237=]

51 84 205 0.6 4.4 DIC

ICP

No affected relatives.

Segregation not

assessed

9_17374 M TJP2:

p.[T62M]; [=]

ABCB11:

p.[V444A(;)M677V]

50 130 175 0.6 3.3 No affected relatives.

Segregation not

assessed

10_16643 M TJP2:

p.[I875T]; [=]

ABCB11:

p.[V444A];

[V444A]

ABCB4:

p.[T775M]; [=]

TJP2:

p.[R12H]; [R12H]

37 217 399 0.6 8.1 Familiarity

Both mutations

segregate in an

affected brother

with history of high

GGT

ID 4 and 4a are affected sibling. Sequence variants are reported according to HGVS recommendations (http://varnomen.hgvs.org)

J Gastroenterol

123

http://varnomen.hgvs.org

our cohort with and without P/LP mutations respect to all

other populations (13.5 vs. 1.76% European, 0% East

Asian, 2.7% worldwide: all p\ 0.0001) (Table 3).All nine reported mutations for ABCB11 were mapped

to homolog structures through Mechismo (Fig. 1a). P/LP

variants were found to be located either at ATP-binding

domain, like p.A523G, p.V597L, p.S1100fs and p.R1128C,

or at transmembrane domain, like p.E135K or p.Y93S,

which are also found in spatial proximity (Fig. 1b).

Moreover, the variant p.H423R was found at the interface

with the ATP-binding pocket (Supplementary Table 2).

This overall suggests that deleterious mutations on

ABCB11 might similarly affect the ATP-binding cassette

(ABC) transporter’s functionality by impairing either ABC

catalytic activity, which couples ATP-hydrolysis to

transport, or by perturbing transmembrane region, which

mediates ligands transport to the extracellular side [25].

ABCB4 variations

We identified six variants in the coding region of ABCB4

and two were P/LP mutations. Disease-causing variants

were: c.2014A[T (p.Lys672*), previously undescribed,was found in a 39-year-old woman with familiarity for

cholestatic diseases (mutation was inherited from the

affected father with high GGT), history of DIC and itching

(case 6) while c.1091C[T (p.A364V), described byDegiorgio et al. [26], was present in a 57-year-old cirrhotic

patient with DIC and hepatocellular carcinoma (case 7).

Two VUS, p.L73V (c.217C[G) and c.2324C[T

Table 3 Allele frequencies of detected SNPs

Gene Mutation Protein dbSNPs AF full

cohort

(a) (%)

AF Pts w/o

P/LP

mutations

(b) (%)

AF

EU

(1)

(%)

AF East

Asia (2)

(%)

AF

world

(3) (%)

p values f

ATP8B1 c.2855G[A p.R952Q rs12968116 10.4 10.8 12 0.03 8.3 a-1 = 0.9151 b-1 = 0.8297a-2 < 0.0001 b-2 < 0.0001

a-3 = 0.5888 b-3 = 0.5744

ABCB11 c.1331T[C p.V444A rs2287622 83 81 60 73 57 a-1 < 0.0001 b-1 = 0.008a-2 = 0.1329 b-2 = 0.3318

a-3 =

E135K Y93S

A523GH423RV444A V597L

M677V R1128C

S1100Qfs

Pholip_ATPase

ATP8B1 ABCB4ABCB11

cytosol

extracellular

ABC_tran

PDZ

PDZ

PDZ

SH3_2Guanylate_kin

TJP2

ATP8B1P23

LN45

TK20

3EF30

5IR95

2Q D1219H

PhoLip_ATPase E1-E2 ATPase Cation_ATPase Phospholipid-translocating P-type ATPase C-terminal

ABCB11Y93

SE13

5KH42

3R V444A

A523G

V597L

M677V

S1100Q

fs*38R11

28C

ABC_membrane ABC_tran ABC_membrane ABC_tran

0 93 135 371 423 523 586 677 755 858 1029 1100 12141247 1321

ABCB4L73

V A36

4V R652G L67

2*

T775M

ABC_membrane ABC_tran ABC_membrane ABC_tran

0 73 121 237 364 424 524 561 624 672 711 775 985 1052 1210 1286

TJP2

V3LR12H

T62M Q12

8KR32

2W

M668I

I875T

PDZ PDZ PDZ SH3_2 Guanylate_kin

3 33 62 91 319 668 875 1190

2338 66 145 172 203 305 413 440 476 532 632 658667 779789 902924952 992 11731203 12170

R4E

(L)B or VUS(L)PPhosphorylation siteAcetylation site

R952Q

K203E

F305IE1_E2_ATPase

(P23L)(N45T)

(D1219H)

ABC_membraneL73V

A364V

R652G

T775M

(K672*)

T62M

Q128K

R322W M668I

I875T

A256V

(V3L)(R4E)(R12H)

(A256V)

(a)

(b)

J Gastroenterol

123

(p.T775M), were identified, previously associated to dif-

ferent PFIC3 phenotypes [10, 26].

The missense variant p.T775M was detected together

with LP c.2717T[C (p.I906T) in exon 18 of TJP2 in a37-year-old male presenting a familiarity for cholestasis

(case 3), while the variant p.L73V was detected in a

23-year-old female with high GGT levels, without evi-

dence of significant liver fibrosis at FibroScan�. The

known LB c.711A[T (p.I237=) [11, 27] and c.1954A[G(p.R652G) [11, 28] mutations were found in 16 and four

cases, respectively. The SNP p.I237= was associated in

previous reports with elevated GGT-cholestasis, ICP, and

gallstones [11].

AF of p.I237= was significantly more common in our

cohort of patients with disease-causing mutations than in

European and worldwide population (33 vs. 18%:

p = 0.004; 33 vs. 21%: p = 0.033; Table 3).

AF of p.R652G was similar between our cohort and both

European and worldwide population while it appeared to

be more common in East Asian subgroup (10.8 vs. 27%;

p = 0.00106).

Similarly to ABCB11, ABCB4 non-synonymous muta-

tions were mapped to homolog structures with exception of

the stop-gain p.L672*. Mutations were essentially located

in transmembrane region (Fig. 1a), suggesting a perturba-

tion of ligand transport mechanisms. The stop-gain

p.L672*, located in the middle of the gene (Fig. 1b), would

result in a halved transcript and protein sequence, lacking

one ATP-binding cassette and one transmembrane domain,

further stressing its high severity.

TJP2 variations

Variants identified in the coding region of TJP2 were ten:

three LP mutations, two VUS, and five B/lB variants.

The LP mutations were: c.278C[T (p.T62M) in twocases (cases 8 and 9), one who experienced DIC and ICP,

c.1057C[T (p.R322W), combined to LP heterozygousp.D1219H on ATP8B1 in a 71-year-old man presenting

only high GGT (case 1), and c.2717T[C (p.I875T) asso-ciated with VUS p.T775M in ABCB4 in one patient with

familiarity for cholestatic diseases (case 10).

Differently from what has been reported in a previous

report [6], all three patients presented high GGT. Two VUS

were c.43G[C (p.V3L) and c.46A[G (p.R4E) while B/LBmutations were c.71G[A (p.R12H), c.475C[A (p.Q128K),c.860C[T (p.A256V), c.2097G[A (p.M668I), andc.3465G[AT (p.T1124=).

VUS p.R4E was present in a 31-year-old female asso-

ciated to benign C1954A[G and C711A[T mutations inABCB4: the patient had a history of ICP, juvenile

cholelithiasis and recurrent DIC to different drugs used for

multiple sclerosis. Finally, a 31-year-old man with high

GGT and BA levels presented another VUS p.V3L.

SNP p.M668I was significantly more frequent in our

cohort of patients with disease-causing mutation compared

to European, East Asian, and worldwide population (14.6

vs. 6.3%: p = 0.0385, 0.03%: p\ 0.001, and 5.3%:p = 0.0043) (Table 3).

Most deleterious TJP2 mutations were found at the level

of PDZ, SH3, and guanylate kinase domains (Fig. 1a, b).

This suggests a potential perturbation of TJP2’s function,

i.e., organizing tight and adherent junctions by binding to

the cytoplasmic C termini of junctional transmembrane

proteins and linking them to the actin cytoskeleton [29].

Indeed, one variant (p.R322W) is predicted to affect the

homo- and hetero-dimerization interface with TJP1

(Fig. 2).

Summary mutation profile for FIC patients

Excluding the common SNP V444A on ABCB11, (present

in 83.3% of cases) and I237= on ABCB4 (33%), 26/48

patients (58.3%), have at least one mutation in one or more

genes: eight patients had only one mutated allele while 18

patients had two or more mutated alleles. ABCB11 and

TJP2 resulted most affected genes and they had mutual

exclusive but non-significant tendency. In particular,

ABCB11/M677V, TJP2/M668I, TJP2/Q128K, and TJP2/

R12H had a tendency to affect patients in a mutual

exclusive way between each other (but in combination,

within the same patient, with various mutations). Figure 3

summarizes mutations discovered contemporary, showing

allele variants in the four genes, according to pathogenicity

prediction (P/LP vs. B, LB, and VUS mutations), zygosity,

and the clinical data (age, FibroScan�, GGT, neonatal

jaundice, itching history, bile acids and cholestasis at

histology).

It is possible that patients with multiple alleles mutated

have more severe phenotypes by a synergistic effect at

different bile transporters’ sites as predicted by Mechismo.

However, patients having C 2 non-synonymous mutations

(with at least one predicted P/LP), tend to have higher liver

stiffness comparing to patients with\ 2 non-synonymousmutations (kPa, 7.9 [3.3–35.8] vs. 4.9 [3.1–29.9],

bFig. 1 a Protein sequence annotations of FIC-selected non-synony-mous mutations displayed as pink and red lollipops to indicate

respectively B/LB or VUS and P/LP predicted consequences. Circle

diameters are proportional to the number of patients affected. Blue

and green lollipops indicate phosphorylation and acetylation sites.

Protein sequence regions corresponding to conserved domains are

highlighted by colored boxes and are indicated by their respective

Pfam names; b non-synonymous mutation annotations on available3D structures as assessed through Mechismo, are displayed following

the same coloring scheme as in a. Labels in parenthesis indicatemutations with no structural information

J Gastroenterol

123

p = 0.028), indicating a potential role of these variants as

disease modifiers.

Subanalysis of patients with P and LP mutations vs.

remaining cholestatic population

Significant differences between subjects with and without

P/LP mutations were not observed in sex, age at symptoms

presentation, and the following risk factors for PFIC: DIC

history, family or personal history of ICP, juvenile

cholelithiasis and family history for cholestatic diseases

(Table 4). However, at least one of these risk factors was

present in 9/11 patients with P/LP mutations (Table 2).

Patients with P/LP mutations had more frequently neonatal

jaundice (45.5 vs. 13.5%, p = 0.036) and itching (54.5 vs.

18.9%, p = 0.029).

Regarding laboratory tests, significant differences were

observed between the two subgroups only in terms of BA

concentration (23.8 [4.4–403] vs. 8.8 [2.3–114],

p = 0.003).

Liver stiffness was greater in subjects with P/LP muta-

tions in comparison with the others (8.1 [3.3–35.8] vs. 4.8

[3.1–29.9], p = 0.009). All subjects of the first group

showed histological features of intrahepatic cholestasis

while 72% of patients without P/LP mutations presented it

(p = 0.018). The univariate analysis for predictors of P/LP

mutations indicated that AP (odds ratio [OR] 0.995, 95%

CI 0.990–0.999, p = 0.019), liver stiffness (OR 0.921,

95% CI 0.841–1.010, p = 0.081), itching (OR 5.143, 95%

CI 1.214–21.795, p = 0.026), neonatal jaundice (OR

5.330, 95% CI (1.172–24.277, p = 0.030) showed a

p\ 0.1 (Supplementary Table 3). Multivariate analysis(Supplementary Table 3) showed that only itching was an

independent predictor of P/LP mutations in patients with

cryptogenic cholestasis (OR 5.801, 95% CI 1.244–27.060,

p = 0.025).

Fig. 2 a Prediction of the functional consequences of TJP2’sp.R322W variant through Mechismo. Green and orange arrows,

respectively, indicate an enabling and mixed (i.e., both enabling and

disabling) effect of the p.R322W mutation towards interactors.

b Structural details of the TJP2 dimerization interface predicted to beaffected by the p.R322W mutation

*(L)P(L)B or VUSHM

Age at testFibroscan

High GGT Yes No

7 71

3.1 35.8

Patient ID

NA NA NA

Bile acidsNeonatal jaundice

Itching

Cholestasis at histology

2.3 403

Yes No

Yes No

Yes No

Fig. 3 Bi-dimensional representation showing genes affected by non-synonymous mutations (row) in each patient (column). Mutated genes

are sorted according to their mutation rate and each mutation is

colored according to pathogenicity prediction (P/LP in red, B/LB and

VUS in pink). Homozygous mutations are indicated by an asterisk.

For each patient, the following clinical data are also displayed: age of

patient, FibroScan� values, GGT levels, neonatal jaundice, itching

history, bile acids and cholestasis at histology

J Gastroenterol

123

Discussion

ABCB11 and ABCB4 are ATP-binding cassette proteins

and members of MDR/TAP subfamily. They are membrane

proteins, with a long intracellular domain, which makes

them partly similar to ATP8B1, though it shares no close

homology with either of them. These three genes are pre-

dicted, but equivalents in non-human species, to interact

both physically and functionally [30]. Specifically,

ATP8B1 homolog in Drosophila MRP has been seen to

interact with ABCB11/B4 homolog CG31729 and yeast

equivalents YOR1 and DNF3 are genetic interactors. The

majority of mutations described in these three genes lie in

or near the intracellular domains in these three proteins.

TJP2, while lacking trans-membrane domains is never-

theless a peripheral membrane protein attached (when not

in the nucleus) via lipid head groups to the cytoplasmic

side of the membrane. Functionally, the formation of tight-

junctions is thought to be related to separation of bile from

plasma (which in itself is intricately linked to their trans-

port) [31]. It is possible that some or all of these proteins

form a complex at some point during the production or

transport of liver metabolites.

PFIC are considered pediatric diseases related to liver

failure. Mutations in ATP8B1, ABCB11, ABCB4 and TJP2

have been associated to a plethora of cholestatic disorders;

hepatocellular carcinoma and cholangiocarcinoma (to

ABCB11 and TJP2), ICP (linked to ATP8B1, ABCB11 and

ABCB4), LPAC (to ABCB4), DIC (to ABCB11 and ABCB4)

and BRIC (to ATP8B1 and ABCB11), are related and may

coexist in the same patient. Furthermore, all four genes may

be responsible for progressive forms of cholestasis, tradi-

tionally considered exclusive of childhood [5, 6, 32, 33].

Of interest, we found a mean age of 37 years in subjects

with disease-causing variants to confirm that P/LP muta-

tions are not present only in children. Few studies linked

mutations in PFIC genes with non-progressive diseases,

especially in heterozygous subjects. In our cohort of

patients with at least one disease-causing mutation, only

four patients were homozygous or compound heterozygous

for disease-causing variants in the same gene.

Colombo et al. [10] described pathogenic mutations in

ABCB4 in about a quarter of asymptomatic children where

cholestatic disease was incidentally discovered via liver

enzyme abnormalities and, not surprisingly, some of these

patients carried a single heterozygous mutation. Gordo-

Table 4 Main features of patients according to presence of P/LP mutations

Pts with P/LP mutations N = 11 Pts without P/LP mutations N = 37 p value

Male N/Tot N (%) 5/11 (45.5) 23/37 (62.2) 0.260

Age: (years, mean ± SD)

At time of genetic test 37 ± 19 44 ± 14 0.244

At symptoms presentation 27 ± 12 33 ± 15 0.308

Risk factor of cholestasis N/Tot N (%)

DIC history 6/11 (54.4) 11/37 (29.7) 0.126

Neonatal jaundice 5/11 (45.5) 5/37 (13.5) 0.036

Itching history 6/11 (54.5) 7/37 (18.9) 0.029

ICP history 2/11 (18.2) 4/37 (10.8) 0.420

Juvenile cholelithiasis 3/11 (27.3) 5/37 (13.5) 0.259

Family history 5/11 (45.5) 12/37 (32.4) 0.327

Laboratory, median (range)

GammaGT (UI/l) 121 (5–379) 148 (18–597) 0.425

AP (UI/l) 467 (138–1012) 280 (122–610) 0.061

ALT (UI/l) 41 (9–387) 46 (9–358) 0.608

Bilirubin (mg/dl) 1.2 (0.4–11.2) 0.7 (0.4–6.2) 0.484

Bile acids (lmol/l) 23.8 (4.4–403) 8.8 (2.3–114) 0.034

Cholesterol (mg/dl) 182 (132–303) 215 (87–328) 0.484

Albumin (g/dl) 4.3 (2.9–5.2) 4 (3.6–4.7) 4.3 (2.9–5.2) 0.126

Platelets (103/lL) 223 (128–412) 233 (61–417) 0.873

FibroScan (kPa), median (range) 8.1 (3.3–35.8) 4.8 (3.1–29.9) 0.009

Histologic features of cholestasis N/Tot N (%) 10/10 (100) 21/34 (61.8) 0.018

N number, Tot N total number, DIC drug-induced cholestasis, SD standard deviation, P/LP mutations pathogenic and likely mutations, ICP

intrahepatic cholestasis of pregnancy, AP alkaline phosphatase, ALT alanine transaminases, kPa kiloPascal

J Gastroenterol

123

Gilart et al. [9] reported similar results in another cohort of

pediatric patients in whom defects in a single allele of

ABCB4 were identified in 9/67 subjects. Some authors

suggested that cryptogenic cholestasis in adults should be

added to the spectrum of conditions associated with

ABCB4 mutations [34–36]. Furthermore, adults with PSC

had early onset of disease if MDR3 deficiency was found

[34].

In a recent study analyzing 427 cholestatic patients, 149

subjects presented at least one disease-causing variant on

ATP8B1, ABCB11 and ABCB4. Surprisingly, 25 patients

with only one heterozygous variant experienced symptoms

before first year of life, suggesting the presence of muta-

tions in other genes, epigenetic changes or environmental

factors responsible of cholestatic phenotype severity [11].

However, adult population data lack clinical significance

of heterozygosity in ABCB11, ATP8B1, and TJP2, though

the first two genes have been correlated to BRIC when only

one heterozygous variant was present. In some instances,

these diseases may evolve towards progressive forms (e.g.,

PFIC), thus representing a clinical continuum [22, 37].

Our study represents the first report of simultaneous

sequencing of ATP8B1, ABCB11, ABCB4 and TJP2 in

patients with cryptogenic cholestasis.

In our cohort, 21% of patients had at least one P/LP

mutation and 17% with a disease-causing mutation pre-

sented a liver disorder in adulthood ([ 18 years), sug-gesting once again that variants in PFIC genes are not

exclusive to cholestatic diseases in childhood.

Patients with P/LP mutations had higher liver stiffness

and BA levels than patients without disease-causing vari-

ants and all exhibited histological evidence of lobular

cholestasis, confirming a more aggressive phenotype in

subjects with causative mutations.

Many patients had multiple variants in these four genes.

It is thus tempting to hypothesize a synergistic effect in

determining different cholestasis phenotypes and/or speci-

fic interactions with environmental factors, in particular

certain drugs. A reduction of bile flow represents a

pathophysiological state and can be critical for metabolism

of many drugs. This is supported by MDR3 and BSEP

deficiencies being known to be associated to different

levels of DIC [7, 38]. Examples are SNPs V444A/M677V

in ABCB11 and I237= in ABCB4. We found V444A in

83.3% of patients (38% in homozygous state): this benign

variant was reported with AF in the general population of

56.9 and of 65.9% in cholestatic patients with no BSEP

disease-causing mutations [11]. It is associated with DIC,

ICP and reduced expression of BSEP levels in previous

studies [24, 39–41]. Specifically in DIC, patients carrying

V444A were at increased risk of drug-induced liver injury,

when taking drugs containing a carbocyclic system with

aromatic rings [42].

SNP M677V occurred in 13.5% of our cohort of cho-

lestatic patients without P/LP mutations, more significantly

frequent than in the other zones (AF in word 2.73%).

In our cohort, SNP p.I237= has an allele frequency of

25%, while Dröge et al. [11] reported 15.4% in an assumed

genetic cholestatic population. Our AFs were significantly

increased respect to AFs reported in worldwide and

European population. This ABCB4 polymorphism increa-

ses the risk of development of cholestatic disease, ICP and

LPAC in previous reports [11, 43].

Common SNPs V444A/M677V/I237= were detected in

a very high number of cases, with or without disease-

causing mutations, suggesting that these common variants

might contribute to cholestasis development or worsening.

The integration of standard variant prediction and clas-

sification tools with more advanced bioinformatic analyses,

taking into account biological pathways and/or biomolec-

ular structures, is becoming increasingly used to get a

deeper understanding on the role of genetic variation from

large NGS datasets, as we previously showed for genetic

diseases [44] and cancer [45]. This is also witnessed by

recent work by Dröge et al. [11] for a larger panel of

cholestatic liver disease patients. For example, we showed

that the common SNP p.V444A of ABCB11 is found at the

catalytic domain of the ABC transporter (Fig. 1) in close

proximity but not in direct physical contact with the ATP-

binding cassette. We thus speculate that while it is unlikely

that this variant directly perturbs the enzymatic activity of

the transporter, it might still induce subtle structural rear-

rangements on the catalytic domain leading to a slightly

perturbed, but still functional enzyme. Such a genotype

would give rise to a disease phenotype when in co-presence

with other genetic or environmental factors.

Disease-causing mutations in TJP2 are also well estab-

lished. Sambrotta et al. [6] associated them with progres-

sive forms of normal GGT cholestasis in pediatric patients,

while our four patients had heterozygous status, different

ages of onset, and high GGT levels. Our results highlight

the variability of clinical presentation in patients having

mutations in ATP8B1, ABCB11, ABCB4, and TJP2, espe-

cially when only one allele was involved. Our structural

analysis, while corroborating pathogenicity predictions by

standard bioinformatic tools, also provided novel mecha-

nistic insights on the functional consequences of these

variants. For example, the TJP2 p.R322W mutation, pre-

dicted to be pathogenic, was shown by Mechismo to affect

interfaces mediating homo- and hetero-dimers (Fig. 2),

thus suggesting to likely impact TJP2’s physiological

function.

Similarly to other studies, we found it difficult to assess

the causative role of many missense variants. We classified

the variants detected in our cohort according to ACMG

criteria, and some missense resulted as VUS. The patients

J Gastroenterol

123

showing only VUS were not included in our analyses of

mutated patients (Table 4 and Supplementary Table 1) but

some of these variants might be re-classified as likely

pathogenic when related to the clinical phenotype.

However, in our cohort of patients, the etiology of the

cholestatic disease remains elusive, with only 21%

attributable to mutations in the known causative genes.

Mutational combinations at different loci as well as the

several VUS detected in one or more FIC-genes and their

potential synergistic effect could also contribute to the

disease onset.

The main limit of our study is that we have considered

four genes responsible for intrahepatic cholestasis, which

does not exclude that other genes could potentially be

associated with this condition. Other studies proposed

multi-gene panels able to facilitate genetic diagnosis, but

only in children with intrahepatic cholestasis [13, 46, 47].

For instance, the recent discovery of new genes

responsible of FIC: myosin-5B gene, linked to microvillus

inclusion disease and NR1H4 gene, encoding farnesoid

X-receptor, a bile acid-activated nuclear hormone receptor

that regulates bile acid metabolism and suppression of bile

acid production. Both genes are able to lead to progressive

cholestatic liver disease [48–50].

Nevertheless, the highest rates of BA levels, liver

fibrosis, itching, and bile histologic alterations detected in

the patients with P/lP mutations make it possible to classify

the pathogenicity of variants in accordance with the

ACMG standards [16].

In conclusion, the remarkable number of cases with

causative mutations found in adults with cryptogenic

cholestasis confirms the usefulness of mutational screening

in ATP8B1, ABCB11, ABCB4, and TJP2, especially in

those cases with itching history, regardless of age of

cholestasis onset. Remains to be evaluated the synergistic

role of mutations at different loci and of VUS in these four

genes. Further observational studies are needed to under-

stand the long-term clinical significance of mutations in

genes responsible for FIC. Finally, the high proportion of

unsolved cases suggests novel genetic etiologies that

remain to be elucidated.

Author contributions GV and PA designed the study and collecteddata. AM, VM, and MS performed the DNA sequencing and applied

prediction tools; AD supervised the histological evaluations, FR and

RBR performed protein modeling by Mechismo; SG, AM, VM, GV,

RV, and PA analyzed the patients’ data. GV wrote the manuscript; all

authors critically revised the manuscript.

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict ofinterest.

Financial support No grants and other financial support werereceived.

References

1. Hori T, Nguyen JH, Uemoto S. Progressive familial intrahepatic

cholestasis. Hepatobiliary Pancreat Dis Int. 2010;9:570–8.

2. Paulusma CC, Elferink RP, Jansen PL. Progressive familial

intrahepatic cholestasis type 1. Semin Liver Dis.

2010;30:117–24.

3. Lam P, Soroka CJ, Boyer JL. The bile salt export pump: clinical

and experimental aspects of genetic and acquired cholestatic liver

disease. Semin Liver Dis. 2010;30:125–33.

4. Smit JJ, Schinkel AH, Oude Elferink RP, et al. Homozygous

disruption of the murine mdr2 P-glycoprotein gene leads to a

complete absence of phospholipid from bile and to liver disease.

Cell. 1993;75:451–62.

5. Davit-Spraul A, Gonzales E, Baussan C, et al. The spectrum of

liver diseases related to ABCB4 gene mutations: pathophysiology

and clinical aspects. Semin Liver Dis. 2010;30:134–46.

6. Sambrotta M, Strautnieks S, Papouli E, et al. Mutations in TJP2

cause progressive cholestatic liver disease. Nat Genet.

2014;46:326–8.

7. Pauli-Magnus C, Meier PJ, Stieger B. Genetic determinants of

drug-induced cholestasis and intrahepatic cholestasis of preg-

nancy. Semin Liver Dis. 2010;30:147–59.

8. Poupon R, Rosmorduc O, Boëlle PY, et al. Genotype-phenotype

relationships in the low-phospholipid-associated cholelithiasis

syndrome: a study of 156 consecutive patients. Hepatology.

2013;58:1105–10.

9. Gordo-Gilart R, Hierro L, Andueza S, et al. Heterozygous

ABCB4 mutations in children with cholestatic liver disease. Liver

Int. 2016;36:258–67.

10. Colombo C, Vajro P, Degiorgio D, et al. SIGENP Study Group

for Genetic Cholestasis. Clinical features and genotype-pheno-

type correlations in children with progressive familial intrahep-

atic cholestasis type 3 related to ABCB4 mutations. J Pediatr

Gastroenterol Nutr. 2011;52:73–83.

11. Dröge C, Bonus M, Baumann U, et al. Sequencing of FIC1, BSEP

and MDR3 in a large cohort of patients with cholestasis revealed

a high number of different genetic variants. J Hepatol.

2017;S0168–8278:32147–55.

12. Xuan J, Yu Y, Qing T, et al. Next-generation sequencing in the

clinic: promises and challenges. Cancer Lett. 2013;340:284–95.

13. Herbst SM, Schirmer S, Posovszky C, et al. Taking the next step

forward—diagnosing inherited infantile cholestatic disorders

with next generation sequencing. Mol Cell Probes.

2015;29:291–8.

14. Adzhubei IA, Schmidt S, Peshkin L, et al. A method and server

for predicting damaging missense mutations. Nat Methods.

2010;7:248–9.

15. Schwarz JM, Cooper DN, Schuelke M, et al. MutationTaster2:

mutation prediction for the deep-sequencing age. Nat Methods.

2014;11:361–2.

16. Richards S, Aziz N, Bale S, et al. ACMG Laboratory Quality

Assurance Committee. Standards and guidelines for the inter-

pretation of sequence variants: a joint consensus recommendation

of the American College of Medical Genetics and Genomics and

the Association for Molecular Pathology. Genet Med.

2015;17:405–24.

17. Betts MJ, Lu Q, Jiang Y, et al. Mechismo: predicting the

mechanistic impact of mutations and modifications on molecular

interactions. Nucleic Acids Res. 2015;43:e10.

J Gastroenterol

123

18. Pieper U, Webb BM, Dong GQ, et al. ModBase, a database of

annotated comparative protein structure models and associated

resources. Nucleic Acids Res. 2014;42(Database issue):D336–46.

19. Hornbeck PV, Chabra I, Kornhauser JM, et al. PhosphoSite: a

bioinformatics resource dedicated to physiological protein phos-

phorylation. Proteomics. 2004;4:1551–61.

20. Jay JJ, Brouwer C. Lollipops in the CLINIC: information dense

mutation plots for precision medicine. PLoS ONE.

2016;11:e0160519.

21. Painter JN, Savander M, Ropponen A, et al. Sequence variation in

the ATP8B1 gene and intrahepatic cholestasis of pregnancy. Eur

J Hum Genet. 2005;13:435–9.

22. Vitale G, Pirillo M, Mantovani V, et al. Bile salt export pump

deficiency disease: two novel, late onset, ABCB11 mutations

identified by next generation sequencing. Ann Hepatol.

2016;15:795–800.

23. Anzivino C, Odoardi MR, Meschiari E, et al. ABCB4 and

ABCB11 mutations in intrahepatic cholestasis of pregnancy in an

Italian population. Dig Liver Dis. 2013;45:226–32.

24. Lang C, Meier Y, Stieger B, et al. Mutations and polymorphisms

in the bile salt export pump and the multidrug resistance protein 3

associated with drug-induced liver injury. Pharmacogenet

Genomics. 2007;17:47–60.

25. Jin MS, Oldham ML, Zhang Q, et al. Crystal structure of the

multidrug transporter P-glycoprotein from Caenorhabditis ele-

gans. Nature. 2012;490:566–9.

26. Degiorgio D, Colombo C, Seia M, et al. Molecular characteri-

zation and structural implications of 25 new ABCB4 mutations in

progressive familial intrahepatic cholestasis type 3 (PFIC3). Eur J

Hum Genet. 2007;15:1230–8.

27. Müllenbach R, Linton KJ, Wiltshire S, et al. ABCB4 gene

sequence variation in women with intrahepatic cholestasis of

pregnancy. J Med Genet. 2003;40:e70.

28. Liu C, Aronow BJ, Jegga AG, et al. Novel resequencing chip

customized to diagnose mutations in patients with inherited

syndromes of intrahepatic cholestasis. Gastroenterology.

2007;132:119–26.

29. Fanning AS, Anderson JM. Zonula occludens-1 and -2 are

cytosolic scaffolds that regulate the assembly of cellular junc-

tions. Ann NY Acad Sci. 2009;1165:113–20.

30. Szklarczyk D, Franceschini A, Wyder S, et al. STRING v10:

protein–protein interaction networks, integrated over the tree of

life. Nucleic Acids Res. 2015;43:D447–52.

31. van Mil SW, Houwen RH, Klomp LW. Genetics of familial

intrahepatic cholestasis syndromes. J Med Genet.

2005;42:449–63.

32. Alissa FT, Jaffe R, Shneider BL. Update on progressive familial

intrahepatic cholestasis. J Pediatr Gastroenterol Nutr.

2008;46:241–52.

33. Zhou S, Hertel PM, Finegold MJ, et al. Hepatocellular carcinoma

associated with tight-junction protein 2 deficiency. Hepatology.

2015;62:1914–6.

34. Degiorgio D, Crosignani A, Colombo C, et al. ABCB4 mutations

in adult patients with cholestatic liver disease: impact and phe-

notypic expression. J Gastroenterol. 2016;51:271–80.

35. Gotthardt D, Runz H, Keitel V, et al. A mutation in the

canalicular phospholipid transporter gene, ABCB4, is associated

with cholestasis, ductopenia, and cirrhosis in adults. Hepatology.

2008;48:1157–66.

36. Ziol M, Barbu V, Rosmorduc O, et al. ABCB4 heterozygous gene

mutations associated with fibrosing cholestatic liver disease in

adults. Gastroenterology. 2008;135:131–41.

37. Van Ooteghem NA, Klomp LW, Van Berge-Henegouwen GP,

et al. Benign recurrent intrahepatic cholestasis progressing to

progressive familial intrahepatic cholestasis: low GGT cholesta-

sis is a clinical continuum. J Hepatol. 2002;36:439–43.

38. Padda MS, Sanchez M, Akhtar AJ, et al. Drug-induced

cholestasis. Hepatology. 2011;53:1377–87.

39. Meier Y, Zodan T, Lang C, et al. Increased susceptibility for

Intrahepatic cholestasis of pregnancy and contraceptive-induced

cholestasis in carriers of the 1331T[C polymorphism in the bilesalt export pump. World J Gastroenterol. 2008;14:38–45.

40. Pauli-Magnus C, Lang T, Meier Y, et al. Sequence analysis of

bile salt export pump (ABCB11) and multidrug resistance p-

glycoprotein 3 (ABCB4, MDR3) in patients with intrahepatic

cholestasis of pregnancy. Pharmacogenetics. 2004;14:91–102.

41. Dixon PH, van Mil SW, Chambers J, et al. Contribution of variant

alleles of ABCB11 to susceptibility to intrahepatic cholestasis of

pregnancy. Gut. 2009;58:537–44.

42. Ulzurrun E, Stephens C, Crespo E, et al. Role of chemical

structures and the 1331T[C bile salt export pump polymorphismin idiosyncratic drug-induced liver injury. Liver Int.

2013;33:1378–85.

43. Gudbjartsson DF, Helgason H, Gudjonsson SA, et al. Large-scale

whole-genome sequencing of the Icelandic population. Nat

Genet. 2015;47:435–44.

44. Boldt K, van Reeuwijk J, Lu Q, et al. UK10K Rare Diseases

Group. An organelle-specific protein landscape identifies novel

diseases and molecular mechanisms. Nat Commun.

2016;7:11491. https://doi.org/10.1038/ncomms11491.

45. Raimondi F, Singh G, Betts MJ, et al. Insights into cancer

severity from biomolecular interaction mechanisms. Sci Rep.

2016;6:34490. https://doi.org/10.1038/srep34490.

46. Wang NL, Lu YL, Zhang P, et al. Specially designed multi-gene

panel facilitates genetic diagnosis in children with intrahepatic

cholestasis: simultaneous test of known large insertions/deletions.

PLoS ONE. 2016;11:e0164058.

47. Togawa T, Sugiura T, Ito K, et al. Molecular genetic dissection

and neonatal/infantile intrahepatic cholestasis using targeted

next-generation sequencing. J Pediatr. 2016;171:171–4.

48. Gomez-Ospina N, Potter CJ, Xiao R, et al. Mutations in the

nuclear bile acid receptor FXR cause progressive familial intra-

hepatic cholestasis. Nat Commun. 2016;18(7):10713.

49. Qiu YL, Gong JY, Feng JY, et al. Defects in myosin VB are

associated with a spectrum of previously undiagnosed low c-glutamyltransferase cholestasis. Hepatology. 2017;65:1655–69.

50. Gonzales E, Taylor SA, Davit-Spraul A, et al. MYO5B mutations

cause cholestasis with normal serum gamma-glutamyl transferase

activity in children without microvillous inclusion disease.

Hepatology. 2017;65:164–73.

J Gastroenterol

123

https://doi.org/10.1038/ncomms11491https://doi.org/10.1038/srep34490

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具

http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/vip.htmlhttp://www.xuebalib.com/db.phphttp://www.xuebalib.com/zixun/2014-08-15/44.htmlhttp://www.xuebalib.com/

Cryptogenic cholestasis in young and adults: ATP8B1, ABCB11, ABCB4, and TJP2 gene variants analysis by high-throughput sequencingAbstractBackgroundAimMethodsResultsConclusions

IntroductionMaterials and methodsPatientsNGS analysisPrediction of functional consequences of variants and classificationComparison with an international databaseStatistical analysis

ResultsPatient characteristicsNGS results and variantsATP8B1 variantsABCB11 variationsABCB4 variationsTJP2 variationsSummary mutation profile for FIC patientsSubanalysis of patients with P and LP mutations vs. remaining cholestatic population

DiscussionAuthor contributionsReferences

学霸图书馆link:学霸图书馆