Professional Documents
Culture Documents
Edited by
James Neuberger
University Hospital Birmingham
Queen Elizabeth Hospital
Edgbaston
Birmingham, UK
Gideon M. Hirschfield
University of Toronto
Toronto
Ontario, Canada
This edition first published 2020 © 2020 by John Wiley & Sons Ltd
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted,
in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as
permitted by law. Advice on how to obtain permission to reuse material from this title is available at
http://www.wiley.com/go/permissions.
The right of James Neuberger, Gideon M. Hirschfield to be identified as the author of editorial work has been
asserted in accordance with law.
Registered Office(s)
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office
9600 Garsington Road, Oxford, OX4 2DQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products visit us at
www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears
in standard print versions of this book may not be available in other formats.
10 9 8 7 6 5 4 3 2 1
v
Contents
Preface xv
Contributors xvii
Abbreviations xxi
Anti‐mitochondrial Antibody 78
History 78
Immunofluorescence Reactivities, Antigenic Targets, and their Clinical
Significance 78
Anti‐asialoglycoprotein Receptor Antibody 79
History 79
Immunofluorescence Reactivities, Antigenic Targets, and their Clinical
Significance 79
Indications for Autoimmune Liver Serology Testing 79
Concluding Remarks 81
References 82
Section II Autoimmune Liver Diseases and Their Clinical Correlation 103
Malignancy 149
Cholangiocarcinoma 149
Gallbladder Cancer 150
Colorectal Cancer 150
Prognosis 150
Risk Stratification: Clinical Characteristics 150
Natural History Models 151
ALP and Bilirubin 151
Non‐invasive Evaluation of Fibrosis 151
Treatment 152
Medical Treatment 152
Bile Acid Therapy 153
Microbiota Modulators 153
Anti-inflammatory Treatment 154
Antibiotic Therapy 154
Endoscopic Treatment 154
Surveillance for Malignancy 154
Colorectal Cancer: Colonoscopy 155
Gallbladder Cancer: Ultrasound 155
Hepatocellular Cancer: Ultrasound 155
Cholangiocarcinoma 155
Bone Mineral Density 155
Juvenile Sclerosing Cholangitis 155
Epidemiology in Pediatric Disease 155
Autoimmune Sclerosing Cholangitis 156
Diagnosis in Children 156
Clinical Features in Pediatric Disease 158
Treatment of Pediatric Disease (ASC and Juvenile PSC) 158
Prognosis in Pediatric Disease 159
Conclusion 159
References 160
Section IV Transplantation and Its Role in Autoimmune Liver Disease 233
Index 301
xv
Preface
Over the last three decades there have been pathological and clinical aspects of AILD but
many advances in the understanding of auto- also stimulate further research to provide
immune liver disease (AILD). The advances effective therapies. We are indebted to the
in understanding the pathogenesis of these authors who have worked so hard to summa-
diseases is starting to lead to more specific rize complex fields. They represent some of
therapies both for slowing or arresting pro- the world leaders in this area and we are
gression and for treatment of the symptoms. grateful to them for their time. We know
Many people with AILD see delays in diag- there is some overlap and minor conflicts
nosis and are misdiagnosed; treatments tend between chapters but we have retained these
at best to control progression rather than as we would like chapters to be comprehen-
induce resolution. Symptoms are often not sive and self‐contained.
treated because of either poor knowledge or We would also like to thank Deirdre Barry,
inadequate therapies. Liver transplantation, Yogalakshmi Mohanakrishnan and Baskar
a highly effective therapy for those with end‐ Anandraj at Wiley for their support, patience
stage disease, has its inherent risks and com- and editorial skills and, most importantly, our
plications and is really a reflection of a failure families who have been patient during this time.
to treat the original disease. Recurrence of Finally, we thank you for reading this
AILD after transplantation, which manifests volume and hope you find it stimulating and
despite sufficient immunosuppression to provoking. Whilst autoimmune liver dis-
prevent allograft rejection and in a foreign eases are rare, they are impactful, and every
HLA milieu, must provide clues to some of opportunity to raise awareness, standards
the processes that lead to AILD but the and knowledge is an opportunity for better
lessons remain obscure. patient care.
We have been pleased to edit this volume
James Neuberger
which we hope will not only provide a bal-
Gideon M. Hirschfield
anced and comprehensive background to the
xvii
Contributors
Eleanor Barnes PhD, FRCP, F Med Sci M. Eric Gershwin MD, MACR, MACP
Professor of Hepatology and Experimental Distinguished Professor of Medicine, Chief,
Medicine, Translational Gastroenterology Division of Rheumatology, Allergy and
Unit, Nuffield Department of Medicine, Clinical Immunology, The Jack and
University of Oxford; and Oxford NIHR Donald Endowed Professor,
Biomedical Research Centre, Oxford, UK University of California School of Medicine,
Davis, California, USA
Kirsten Muri Boberg MD, PhD
Professor Norwegian PSC Research Alessio Gerussi MD
Center and Section for Gastroenterology, Research Fellow, Division of
Department of Transplantation Medicine, Gastroenterology, University of Milan
Division of Surgery, Inflammatory Medicine Bicocca, Milan, Italy
and Transplantation, Oslo University
Hospital Rikshospitalet; and Institute of Núria Guañabens MD
Clinical Medicine, Faculty of Medicine, Professor of Medicine, University of
University of Oslo, Oslo, Norway Barcelona; and Senior Consultant Physician,
Department of Rheumatology,
Marco Carbone MD, PhD
Metabolic Bone Diseases Unit,
Clinical Lecturer in Gastroenterology,
Hospital Clínic, Institut d’Investigacions
Program for Autoimmune Liver Diseases,
Biomèdiques August Pi Sunyer (IDIBAPS),
Department of Medicine and Surgery,
Centro de Investigación Biomédica en Red
University of Milan Bicocca, Milan, Italy
de Enfermedades Hepáticas y Digestivas
(CIBERehd), Barcelona, Spain
Olivier Chazouillères MD
Director Department of Hepatology,
AP‐HP, Hôpital Saint‐Antoine, Service Aliya Gulamhusein MD, MPH
d’Hépatologie; Reference Center for Toronto Centre for Liver Diseases, Toronto
Inflammatory Biliary Diseases and General Hospital, University Health
Autoimmune Hepatitis; INSERM, UMR‐S Network; and Division of Gastroenterology,
938, CDR Saint‐Antoine; and Sorbonne University of Toronto,
Université, Paris, France Toronto, Ontario, Canada
Emma L. Culver PhD, MRCP Peter C. Hayes MBChB, BMSc, MD, PhD
Consultant Hepatologist and Senior Head of Department, Professor of Medicine
Lecturer, Translational Gastroenterology and Consultant Physician,
Unit, Nuffield Department of Medicine, Department of Hepatology,
University of Oxford; and Oxford NIHR Royal Infirmary of Edinburgh,
Biomedical Research Centre, Oxford, UK Edinburgh, UK
xviii Contributors
Michael A. Heneghan MD, MMedSc, FRCPI Piotr Milkiewicz MD, DSc, MRCP(UK)
Clinical Director for Liver Services and Professor of Medicine/Consultant Physician,
Consultant Hepatologist and Professor of Head, Liver and Internal Medicine Unit,
Hepatology, Institute of Liver Studies, King’s Medical University of Warsaw, Warsaw,
College Hospital, Denmark Hill, London, UK Poland
Introduction
The Paradigm and Paradox of Liver Autoimmunity
M. Eric Gershwin
Division of Rheumatology, Allergy and Clinical Immunology, University of California School of Medicine, Davis, CA, USA
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
2 Introduction
devoted to autoimmunity was not published In this volume, Neuberger and Hirschfield
until 1963, or over 60 years after Ehrlich’s have taken up the challenge of providing the
initial observations. This book, authored by scientific basis for liver autoimmunity. Liver
Mackay and Burnett, was published in a autoimmunity has often been a stepchild
series of monographs entitled American compared with other autoimmune diseases.
Lectures in Living Chemistry (Charles The pharmaceutical industry has made enor
Thomas Publishers, Springfield, Illinois). mous efforts in helping patients with
This textbook contains 12 chapters devoted rheumatoid arthritis, with psoriasis, and of
to specific autoimmune diseases. The con course with human viral hepatitis, but the
cept of the forbidden clone was discussed commitment to liver autoimmunity, not only
and, of particular interest to liver specialists, in industry but in academia, has been modest
the terms active chronic and lupoid hepatitis at best. And yet the same generic factors that
were introduced. Absent from this book was lead to both organ‐specific and systemic auto
of course the concept of lymphoid subpopu immune diseases are at play in the liver auto
lations, the genetics of the immune response, immune syndromes, since the development of
major histocompatibility complex (MHC) disease requires a genetic predisposition and
restriction and, especially, discussion of an environmental stimulus. This is a feature
maternal–fetal immune homeostasis. These that I have coined “bad genes and bad luck.”
omissions are easily forgiven when it is real However, this volume goes beyond descriptive
ized that the tools used by immunologists in immunobiology. It provides a rigorous anal
the 1960s had not changed significantly in ysis of multiple critical areas of why patients
over three decades and relied particularly on become susceptible and why their natural his
diffusion and electrophoresis. tory of disease may vary, including for example
This issue of maternal–fetal immune the roles of genetics and epigenetics, environ
homeostasis is critical in the introduction of mental exposures, the specific and usual
liver autoimmunity. Whether it is a human or clinical challenges faced by patients with liver
a mouse, the fetus is a privileged site and disease, and of course helping the particularly
spared from any immune attack from the difficult and challenging patient (i.e. liver dis
mother. The basis for this tolerance resides in ease during pregnancy). This volume also fills
the liver. The liver was known even in ancient a major void in explaining the role of liver
Greek mythology to be an organ capable of transplantation by placing not only its
regeneration. Prometheus stole from the scientific significance in perspective, but also
Gods, giving it away to mortals. As a result, he tackling the difficult subject of disease recur
was chained to a rock by Zeus, where each day rence following liver transplant. There remain
an eagle pecked out his liver. Presumably, many controversies and multiple unanswered
Prometheus survived because of liver regen questions, and the book ends with an attempt
eration. However, with that singular exception, to make sense of overlap and crossover syn
and the use of liver in the everyday kitchen, dromes, and the relationship to extrahepatic
the liver as an organ responsible for immu abnormalities in patients with autoimmune
nity was ignored. Indeed, a major basis liver disease. Finally, it focuses on the
for maternal–fetal tolerance is the portal most important element, which is clinical
circulation. It is therefore ironic that the organ management and how we may help our
which is most important for the critical patients. At the end of the day, whatever
element of immune tolerance can itself the scientific interest may be, the only
become a victim of autoimmunity, manifest in fundamental question of value is whether our
such diseases as autoimmune hepatitis, pri work and our publications will have a positive
mary biliary cholangitis, primary sclerosing impact on the health of people. This volume
cholangitis and IgG4‐associated disease. should fill such a need.
3
Section I
Abstract
The liver has many functions including metabolic homeostasis, disposal of endotoxins and
xenotoxins, metabolism of bilirubin and urea, and bile formation and secretion. The process of bile
formation depends on the liver synthesis and the canalicular secretion of bile acids. Besides their
roles in dietary lipid absorption and cholesterol homeostasis, bile acids also play a key role as
signaling molecules in the regulation of hepatic metabolism and energy homeostasis. The regener‑
ative ability of the liver lies in the multiple niches of biliary tree stem cells.
Keywords bile formation; carbohydrate metabolism; hepatocyte; cholangiocytes lipid
etabolism; metabolic zonation; protein metabolism; bile acid.
m
Key Points
●● The liver is largely composed of hepato‑ ●● Besides their roles in dietary lipid absorption
cytes and biliary epithelial cells or cholan‑ and cholesterol homeostasis, bile acids
giocytes; both hepatocytes and intrahepatic (BAs) also play a key role as signaling mole‑
cholangiocytes differentiate from bipotent cules in the regulation of hepatic metabo‑
liver progenitors, the hepatoblasts. lism and energy homeostasis.
●● The liver has many functions, among which ●● BAs also have hormonal signaling
metabolic homeostasis, disposal of endo‑ functions and interact with dedicated
toxins and xenotoxins, metabolism of bili‑ receptors such as the nuclear receptor
rubin and urea, and bile formation and farnesoid X receptor and G protein‐coupled
secretion are just examples. The liver also receptors, which regulate BA homeostasis
produces fundamental circulating proteins and BA‐induced injury and/or
and clotting factors and hormones. In inflammation.
addition, the liver receives and processes ●● Multiple niches of biliary tree stem/
the blood coming from the intestine and has progenitor cells reside in different locations
a fundamental role in innate immunity. along the human biliary tree and within the
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
6 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
liver parenchyma and have a key role in rise to the ductular reaction, a major driver
regeneration of the liver. of the progression of hepatic fibrosis.
●● Cholangiocytes modify the primary bile ●● Cholangiocytes also contribute to the
by secretion of chloride and bicarbonate immune response through antigen pre‑
fluid. This is a major protective mecha‑ sentation to immune cells, being a target
nism for the biliary tree. of immune‐mediated aggression or being
●● Cholangiocytes, a barrier and secretory epi‑ the initiators of an inflammatory reaction
thelium in normal conditions, activate and/ that then progresses to adaptive immune
or proliferate following a liver insult and give activation.
the total population of fixed macrophages in bone marrow cells. Bilirubin formed in the
the body. These are macrophages attached to monocytic–macrophage cell system of liver,
the endothelial lining of the sinusoid, in spleen and bone marrow and some of the bil‑
greater numbers in the periportal areas. irubin formed in the hepatocytes from
They are responsible for removing old and hepatic heme are released into plasma where
damaged blood cells or cellular debris, bilirubin is bound to albumin at high‐affinity
bacteria, viruses, parasites, and tumor cells. binding sites. In normal conditions, only a
The HSCs, also known as Ito cells or lipo‑ minimal amount of bilirubin is unbound in
cytes, are mesenchymal cells and represent plasma. An increase in free bilirubin would
the major storage site of retinoids in healthy allow the pigment to enter tissues where it
livers. They lie within the subendothelial can have toxic effects; this is what is observed
space, and their long cytoplasmic extensions in neonates with defective conjugation and in
have close contact with parenchymal cells Crigler–Najjar syndrome, when diffusion of
and sinusoids, where they may regulate unbound bilirubin into the brain can cause
blood flow and hence influence portal kernicterus.
hypertension. HSC activation is the central In normal conditions, bilirubin is effi‑
event in hepatic fibrosis. During hepatocyte ciently taken up by the liver whereas the
injury, HSCs proliferate, migrate to zone albumin remains in plasma. In the liver,
three of the acinus, change to a myofibro‑ bilirubin is bound initially to glutathione‐
blast‐like phenotype, and produce collagen S‐transferase, then glucuronidated and
and laminin. excreted into bile. Microsomal bilirubin
Finally, the pit cells represent the natural uridine diphosphate glucuronosyltransfer‑
killer (NK) cells of the liver and are located ase (UGT) is the enzyme that converts
within the sinusoids. Pit cells show unconjugated bilirubin to conjugated bili‑
spontaneous cytotoxicity against tumor‐ and rubin monoglucuronide and diglucuronide.
virus‐infected hepatocytes. Biliary excretion of the glucuronide is
mediated by the adenosine triphosphate
(ATP)‐dependent multidrug resistance
Hepatic Metabolism protein (MRP)‐2 and this is the rate‐lim‑
iting factor in the transport of b ilirubin
The liver is the site of many metabolic path‑ from plasma to bile. Bilirubin diglucuro‑
ways. It stores and makes available many nide is not reabsorbed from the small
nutrients as energy sources. In turns, the intestine; in the colon it may be hydrolyzed
metabolic function of the liver is regulated by by bacterial β‐glucuronidases, producing
hormones secreted by the pancreas, adrenal urobilinogens and urobilin, which are
gland, and thyroid. excreted in the stool or urine.
Bilirubin is an end‐product of heme catabo‑ The liver plays a key role in carbohydrate
lism. Two enzymes are involved in bilirubin metabolism. It maintains carbohydrate stores
formation: the microsomal heme oxygenase by synthesizing glycogen and generating
converts heme to biliverdin and a cytosolic glucose from precursors such as lactate, pyru‑
reductase subsequently reduces biliverdin to vate, and amino acids. Glycogen stored in the
bilirubin. liver is the main source of rapidly available
The majority (up to 85%) of heme is derived glucose for the whole organism. In the fed
from hemoglobin and only a small fraction state, glycogen synthesis occurs preferentially
from other heme‐containing proteins such as in the perivenous hepatocytes whereas in the
cytochrome P450, myoglobin and immature fasting state, glucose release via glycogenolysis
8 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
and gluconeogenesis initially occurs in peri‑ lesterol into the plasma. In patients with
portal hepatocytes. The liver glycogen stores severe chronic cholestasis who are however
contain up to a 2‐day supply of glucose, after malnourished, for example in premature
which glucogenesis occurs mainly from lac‑ ductopenic variant of primary biliary cholan‑
tate. In acute liver failure, the blood glucose gitis (PBC) or in carcinomatous biliary
level may drop whereas this is infrequent in obstruction, the serum cholesterol may be
chronic liver disease, where it is more common lower. Elevated cholesterol levels are clini‑
to observe hyperglycemia and insulin resis‑ cally associated with skin xanthomas and
tance. This may be related to decreased xanthelasma. Hypercholesterolemia is not
glucose uptake by muscle and reduced gly‑ consistently associated with subclinical ath‑
cogen storage in the liver and muscle. erosclerosis in PBC, but should be treated if
other risk factors for cardiovascular disease
are also present.
Lipid Metabolism
In addition, the liver has other roles in lipid
A major role of the liver in lipid metabolism metabolism such as oxidizing triglycerides to
is to synthesize large quantities of cholesterol produce energy; synthesizing lipoproteins;
and phospholipids, many of which are pack‑ and converting excess carbohydrates and
aged with lipoproteins and made available to proteins into fatty acids and triglyceride,
the rest of the body. Free cholesterol also which are then exported and stored in
derives from the uptake of chylomicron rem‑ adipose tissue.
nants and lipoproteins from the circulation.
BAs are synthesized from free cholesterol,
Protein Metabolism
and both BA and cholesterol are secreted
into bile. Bile provides a major route for Amino acids are derived from the diet and
cholesterol excretion. from tissue breakdown and they reach the
The rate‐limiting step of cholesterol liver via the portal vein. Many critical aspects
synthesis is the conversion of 3‐hydroxy‐3‐ of protein metabolism occur in the liver, such
methylglutaryl‐CoA (HMG‐CoA) to meval‑ as the deamination and transamination of
onate by the enzyme HMG‐CoA reductase, amino acids, followed by conversion of the
which is located almost exclusively in peripor‑ non‑nitrogenous parts of these molecules to
tal cells. Synthesis is increased in biliary duct glucose or lipids. Several of the enzymes used
obstruction, terminal ileal resection, biliary or in these pathways, e.g. alanine and aspartate
intestinal lymph fistula, and with medications aminotransferases, are commonly measured
such as colestyramine, corticosteroids and in serum to assess liver damage.
thyroid hormones. Cholesterol synthesis is The liver is also responsible for a number
inhibited by BAs, cholesterol feeding, fasting, of vital metabolic processes, including the
and medications such as fibrates, nicotinic removal of ammonia (an important factor in
acid, and statins. the development of hepatic encephalopathy)
During chronic cholestasis, such as in via the synthesis of urea; the synthesis of
primary biliary cholangitis and primary
non‐essential amino acids; and the synthesis
sclerosing cholangitis (PSC) but also in acute of most plasma proteins such as albumin
cholestasis, cholesterol serum levels are (the major plasma protein), fibrinogen, α1‐
increased. This is mainly secondary to reten‑ antitrypsin, haptoglobin, ceruloplasmin,
tion of cholesterol normally excreted in bile transferrin, and several coagulation factors.
but also to increased hepatic synthesis of
cholesterol, reduced plasma lecithin‐choles‑
Metabolic Zonation
terol acyltransferase activity, and regurgita‑
tion of biliary lecithin, which produces a shift The multiple functions of the liver are
of cholesterol from pre‐existing tissue cho‑ facilitated by a functional specialization
Chapter 1 Physiology, Immunology and Pathology of the Liver and Biliary Tree 9
Alternative BA export
BA BA BA
MRP4
MRP3
OSTα
OSTβ
CAR
PXR
GR
RAR
MRP2 Bili-glu
CAR PPARα
OATP1B1 HNF4 BSEP BA
BA
FXR MDR2/3 PC
BA Uptake Biliary secretion
PPARα
BA ABCG8
HNF4
NTCP SHP PXR Chol
Na+ GR ABCG5
LXR
Cl–
GR AE2
HCO3–
PPARα CYP7A1
BA synthesis
Hepatocyte
Figure 1.1 Transcriptional regulation of hepatocellular bile formation. Expression of hepatobiliary transporters in
hepatocytes determines hepatic bile acid (BA) flux and hepatocellular concentrations of these potentially toxic
metabolites. To ensure the balance between synthesis, uptake and excretion, expression of hepatobiliary
transporters is tightly regulated by nuclear receptors (NRs). NRs provide a network of negative feedback and
positive feed‐forward mechanisms for the control of intracellular concentration of biliary constituents, which are
often also ligands for these NRs. BA‐activated FXR is a central player in this network, that represses hepatic BA
uptake (NTCP) and (via SHP) synthesis (CYP7A1), promotes bile secretion via induction of canalicular transporters
(BSEP, MRP2, ABCG5/8, MDR3), and induces BA elimination via alternative export systems at the hepatic
basolateral (sinusoidal) membrane (OSTα/β). Several NR pathways converge at the level of CYP7A1 as the rate‐
limiting enzyme in BA synthesis. CAR and PXR facilitate adaptation to increased intracellular BA concentrations by
upregulation of alternative hepatic export routes (MRP3 and MRP4) and induction of detoxification enzymes (not
shown). Together with RAR, these xenobiotic receptors also regulate the canalicular expression of MRP2.
Cholesterol sensor LXR promotes biliary cholesterol excretion via ABCG5/8. Stimulation of AE2 expression by GR
stimulates biliary bicarbonate secretion thus reducing bile toxicity. Green arrows indicate stimulatory and red
lines suppressive effects on target genes. Bili‐glu, bilirubin glucuronide; BSEP, bile salt export pump; CAR,
constitutive androstane receptor; CYP7A1, cholesterol 7α‐hydroxylase; FXR, farnesoid X receptor; GR,
glucocorticoid receptor; HNF4, hepatocyte nuclear factor 4; LXR, liver X receptor; MDR3, multidrug resistance
protein 3, phospholipid flippase; MRP2, multidrug resistance‐associated protein 2; MRP3, multidrug resistance‐
associated protein 3; MRP4, multidrug resistance‐associated protein 4; NTCP, sodium taurocholate cotransporting
polypeptide; OSTα/β, organic solute transporter alpha and beta; PC, phosphatidylcholine; PXR, pregnane X
receptor; PPARα, peroxisome proliferator‐activated receptor alpha; RAR, retinoic acid receptor; SHP, small
heterodimer partner. Source: Halilbasic et al. [3]. Reproduced with permission of Elsevier.
promoting the hepatocellular clearance of leukotriene C4, and divalent bile salt conju‑
xenobiotics during cholestasis. gates, as well as drug substrates such as
In addition to its role in the uptake of chemotherapeutic agents and antibiotics.
conjugated BAs, NTCP also plays a key role MDR1 contributes to the canalicular excre‑
in hepatitis B and hepatitis D virus entry into tion of drugs and other xenobiotics into bile,
hepatocytes; and, recently, NTCP has also although its exact contribution has yet to be
been shown to modulate hepatitis C virus established. Its broad substrate specificity
infection of hepatocytes by regulating innate and its physiologic expression in various tis‑
antiviral immune responses in the liver. As sues with excretory and protective functions
such NTCP has been established as a novel make MDR1 one of the major determinants
antiviral target. of drug disposition and toxicity. Substrates
are neutral and positively charged organic
compounds and include various chemother‑
Apical (Canalicular) Transporters
apeutic and immunosuppressant agents,
Various ATP‐binding cassette (ABC) trans‑ antiarrhythmic drugs, HIV protease inhibi‑
porters mediate the secretion of bile salts tors, and antifungals.
and xenobiotics across the canalicular Transcriptional regulation of BSEP and
membrane of hepatocytes. These include MDR3 is mediated by FXR and their activation
members of the family of multidrug leads to increased bile salt efflux and the
resistance (MDR) P‐glycoproteins such as formation of mixed micelles in the biliary tree
MDR1 (ABCB1), MDR3 (ABCB4), and the during cholestatic episodes, thereby preventing
bile salt export pump (BSEP or ABCB11). the toxic effects of bile salts on hepatocytes and
Within the family of MDR proteins, BSEP cholangiocytes. In addition, FXR has been
and MDR3 are two highly conserved mem‑ shown to induce MRP2 expression, which
bers that are involved in the secretion of might constitute another compensatory mech‑
cholephilic compounds from the liver cell anism during cholestasis. In contrast, MDR1 is
into the bile canaliculus. upregulated via the pregnane X receptor (PXR),
BSEP constitutes the predominant bile salt which in addition to endogenous ligands is
efflux system of hepatocytes and mediates the activated by different xenobiotics. This pathway
cellular excretion of numerous conjugated bile is part of a general cellular mechanism of detox‑
salts such as taurine‐ or glycine‐conjugated ification, because MDR1 is the key transporter
cholate, chenodeoxycholate, and deoxycho‑ protein involved in the cellular efflux of
late. MDR3 works as an ATP‐dependent numerous drugs and xenobiotics.
phospholipid flippase, translocating phospha‑
tidylcholine from the inner to the outer leaflet
of the canalicular membrane. Canalicular Drug Metabolism
phospholipids are then solubilized by canalic‑
ular bile salts to form mixed micelles, thereby The liver has a major role in drug metabo‑
protecting cholangiocytes from the detergent lism. The main hepatocyte enzymes
properties of bile salts. involved in metabolism belong to the
In addition to these processes, MRP2, the cytochrome P450 group, a large family of
only canalicular member of the MDR‐associ‑ related enzymes housed in the smooth
ated protein family, mediates the canalicular endoplasmic reticulum of the hepatocyte.
transport of glucuronidated and sulfated bile Metabolism is often divided into two phases
salts. MRP2 is the main driving force for bile of biochemical reaction. Phase 1 involves
salt‐independent bile flow through canalic‑ reduction, hydrolysis or oxidation of the
ular excretion of reduced glutathione. drug, the latter being the most common
Furthermore, MRP2 transports a wide process. After phase 1 reactions, the
spectrum of organic anions, including bili‑ resulting drug metabolite is still often chem‑
rubin diglucuronide, glutathione conjugates, ically active. Phase 2 metabolism involves
12 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
broad physiologic effects in the liver and secondary bile salt species are contained in
intestine, drugs targeting FXR and TGR5 there‑ human bile, although primary bile salts are
fore open important perspectives for pharma‑ usually predominant.
cotherapy of cholestatic and metabolic liver
disorders, including the complications of liver
Enterohepatic Bile Acid Circulation
cirrhosis such as portal hypertension and
hepatocellular carcinoma (HCC). BAs undergo an enterohepatic circulation that
In addition, BAs stimulate glucagon‐like depends on active transport systems in the
peptide (GLP)‐1 production via TGR5 liver and the intestine (Figure 1.2). More spe‑
activation. GLP‐1 is known to promote cifically, BAs are excreted from hepatocytes
insulin secretion and thus regulate glucose into bile through BSEP/ABCB11 at the bile
homeostasis. Because GLP‐1 mimetics and canaliculus, reabsorbed in the ileum by the
receptor agonists are currently under clinical apical sodium‐dependent bile salt transporter
development and have shown promise in (ASBT/SLC10A2), and return through the
improving glucose homeostasis in diabetes, portal blood to the liver, where they are taken
BA‐based TGR5 agonists may be a potential up by hepatocytes via the basolateral transport
therapeutic to stimulate GLP‐1 secretion in systems NTCP/SLC10A1 for conjugated BAs
diabetic patients. and OATPs/SLCO/SLC21 family for unconju‑
gated BAs, thus limiting the amount of BA
spillover into the systemic circulation. BAs
Bile Acid Synthesis and Metabolism
complete the enterohepatic circulation six to
BAs are synthesized from cholesterol. In eight times a day and are highly efficiently con‑
humans, the “primary” BAs are cholic acid served. BAs that escape ileal reabsorption
(CA) and chenodeoxycholic acid (CDCA). reach the colon, where they are deconjugated
Before secretion into the bile, both CA and and metabolized (e.g. dehydroxylated) by gut
CDCA are conjugated to the amino group of microbiota to secondary BAs, which can still
taurine or glycine. Conjugation enhances the be passively absorbed as unconjugated BAs in
hydrophilicity of the BA, the major function the colon. Unconjugated BAs are partially
of this being to decrease the passive diffusion reconjugated (and rehydroxylated) during
of BAs across the cell membranes during their passage through the liver before being
their transit through the biliary tree and excreted into bile again, which completes their
intestine. Therefore, conjugated BAs are enterohepatic cycle. In addition, BAs are
absorbed only if a specific membrane carrier filtered by the glomeruli and then reabsorbed
is present. The process of bile formation in renal tubules, again limiting their renal loss.
depends on the liver synthesis and the cana‑ BAs may also cycle between cholangio‑
licular secretion of BAs. The active transport cytes and hepatocytes through a cholehepatic
of BAs across the canalicular membranes of shunt pathway. Unconjugated BAs induce a
hepatocytes is a primary driving force for bile greater degree of bile flow per BA molecule
flow. The majority of the BAs in the intestine excreted in bile. To account for this hyper‑
are absorbed intact. Approximately 15% are choleretic effect, it was proposed that uncon‑
deconjugated by the bacterial flora in the jugated BAs may be passively absorbed by
distal small intestine, with the production of bile ducts, enter the peribiliary plexus
“secondary” BAs by the conversion of CA to adjacent to intrahepatic bile ducts, and then
deoxycholic acid and of CDCA to lithocholic forwarded to the hepatic sinusoids to be
acid. Most of the conjugated and deconju‑ returned to cholangiocytes by hepatocyte
gated BAs are reabsorbed in the distal secretion. Cholehepatic shunting initiated by
intestine and undergo enterohepatic passive absorption of non‐ionized bile salt
circulation that maintains the BA pool. Thus, results in the generation of HCO3–‐rich
at least 12 major conjugated primary and hypercholeresis [4,5].
14 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
Liver Cholesterol
CYP7A1
NTCP
Bile acids
Systemic
circulation BSEP
Active
transport Kidney
Enterohepatic
circulation Bile
Intestine
ASBT
Urine
Feces
Figure 1.2 Enterohepatic circulation of bile acids (BAs). After hepatic synthesis and biliary secretion, BAs
undergo enterohepatic circulation. The BSEP (ABCB11) is the main canalicular transporter for BAs. After
reabsorbtion by ASBT/SLC10A1 in the ileum, BAs (and exit through OSTα/β from enterocytes; not shown) are
transported back to the liver through the portal blood. Reuptake of conjugated BA from portal blood through
NTCP/SLC10A1 (and OATPs for unconjugated BAs; not shown) into the hepatocytes completes the
enterohepatic circulation. ASBT, apical sodium‐dependent bile salt transporter; BSEP, bile salt export pump;
NTCP, sodium taurocholate cotransporting polypeptide; OATPs, organic anion transporters; OST, organic solute
transporter α/β. Source: Trauner et al. [5]. Reproduced with permission of John Wiley and Sons.
Death and Regeneration
recognition. Apoptosis is essential to avoid the
of Hepatocytes outflow of intracellular contents and to limit
the immunologic response against intracel‑
Cell Death
lular autoantigens. Nevertheless, apoptotic
Hepatocytes can die because of either necrosis bodies and fragments can under some cir‑
or apoptosis. Necrosis is the loss of plasma cumstances constitute a major source of
membrane integrity with release of the cel‑ immunogens in autoimmune diseases that
lular contents locally, which triggers an involve the targeting of ubiquitous autoanti‑
inflammatory response. Apoptosis is a highly gens. This has been described in PBC. In the
regulated process in which cells that are dam‑ BECs of patients with PBC there is increased
aged, senescent or deregulated self‐destruct DNA fragmentation, implying increased
with a lower release of inflammatory prod‑ apoptosis, when compared with normal con‑
ucts. Dying cells undergo morphologic modi‑ trols. While mitochondrial proteins are pre‑
fications including chromatin condensation, sent in all nucleated cells, in PBC there is a
nuclear fragmentation, and generation of apo‑ highly specific multilineage immune response
ptotic bodies. Furthermore, they express sig‑ directed to the nominal mitochondrial auto‑
nals on the cell surface that allow macrophage antigenic epitope, the inner lipoyl domain of
Chapter 1 Physiology, Immunology and Pathology of the Liver and Biliary Tree 15
the E2 subunit of the pyruvate dehydrogenase which far exceeds the capacity of remaining
complex (PDC‐E2) of the BECs. Apoptosis of healthy hepatocytes to replicate and restore
BECs has been proposed as a potential source liver function, resident liver progenitor cells
of “neoantigens” that could be responsible for (i.e. oval cells) are activated to support or take
activation of autoreactive T lymphocytes or a over the role of regeneration. In adults, there
target for effector cells or antibodies. PDC‐E2 are multiple niches of biliary tree stem/pro‑
is not only immunologically intact during genitor cells (BTSCs) residing in different loca‑
apoptosis in BECs, but it localizes in the apo‑ tions along the human biliary tree and niches
ptotic bodies of BECs where it is accessible to found within the liver parenchyma, with a key
recognition by anti‐mitochondrial antibodies. role in regeneration of the liver. Figure 1.3
shows the location of stem/progenitor cell
niches in the human biliary tree. Canals of
Liver Regeneration
Hering harbor hepatic stem/progenitor cells
Liver possesses a unique capacity to replace its (HpSCs), while peribiliary glands (PBGs) con‑
mass after tissue injury or loss. The regenera‑ stitute the niche for BTSCs.
tive process involves a cascade of events that Those within the biliary tree are found in
moves cells from their resting G0 phase PBGs and contain especially primitive stem
through G1, S phase (DNA synthesis), and G2 cell populations, expressing endodermal tran‑
to M phase (mitotic cell division). A typical scription factors relevant to both liver and
example is hepatic growth after partial hepa‑ pancreas, pluripotency genes, and even
tectomy. The majority of research on liver markers indicating a genetic signature over‑
regeneration has focused on cytokine‐ and lapping with that of intestinal stem cells [7].
growth factor‐ mediated pathways involved in The distribution of PBGs is not uniform,
initiation and progression through the cell varying along the biliary tree: PBGs are mostly
cycle. During more extensive acute liver injury, found in the hepatopancreatic ampulla and
Figure 1.3 Stem/progenitor cell niches in the human biliary tree. Canals of Hering harbor hepatic stem/
progenitor cells (HpSCs), while peribiliary glands (PBGs) constitute the niche for biliary tree stem/progenitor
cells (BTSCs). ASH, alcoholic steatohepatitis; BA, bile acid; CCA, cholangiocarcinoma; NAFLD, non‐alcoholic fatty
liver disease; NAS, non‐anastomotic strictures; PSC, primary sclerosing cholangitis. Source: Overi et al. [6].
16 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
are less numerous in the common bile duct; follows paracrine cross‐talk mediated by the
they are not present in the gallbladder, but a ability of RDCs to secrete profibrotic and
stem cell‐like compartment is located in the proinflammatory growth factors, and cross‐
epithelial crypts. Biliary progenitors support talk with cells of mesenchymal origin, in
the renewal of large intrahepatic and extrahe‑ particular Kupffer cells and portal fibro‑
patic bile ducts. Stem cells are present in the blasts, which are the main effectors of
canals of Hering, and participate in the renewal fibrosis, as stimulators of the deposition of
of the small intrahepatic bile ducts and in the extracellular matrix by activated myofibro‑
regeneration of liver parenchyma. Small hepa‑ blasts. In addition, RDCs also establish para‑
tocytes located in pericentral positions are also crine communications with endothelial cells
believed to act as progenitor cells on certain that provide the vascular support necessary
occasions. Distinct subpopulations of mature for the growth and arborization of the ductal
hepatocytes and stem/progenitor cell com‑ structures themselves [8].
partments are differentially activated depend‑
ing on the nature and duration of the liver
damage versus different human pathologies. rotective Role of Biliary
P
HCO3− Secretion
Cholangiocyte Reaction The cholangiocyte is exposed to millimolar
to Biliary Damage concentrations of hydrophobic bile salts, which
are toxic to other cells such as hepatocytes
BECs are usually quiescent, but following a at micromolar levels. Resistance against these
liver insult they activate and/or proliferate. A noxious compounds and their cytolytic poten‑
typical element of the repair response to liver tial is therefore essential. One of the strategies
damage is the ductular reaction (DR), a ste‑ that cholangiocytes have developed to survive is
reotyped histopathologic lesion of the biliary the biliary HCO3− umbrella.
epithelium that plays a fundamental role in Biliary HCO3− secretion sustains bile flow
the progression of hepatic fibrosis. The DR and confers its appropriate viscosity, gener‑
is characterized by a marked proliferation ates part of the alkaline tide necessary for
of cholangiocytes, leading to formation of optimal digestion of various nutrients within
reactive ductular cells (RDCs), with poor
the intestine, and protects the apical surface
cytoplasm and arranged in cell cords without of cholangiocytes against protonated apolar
a lumen or in richly anastomosed small‐ hydrophobic BA monomers by maintaining
diameter ducts (<10 μm) with almost unrec‑ an alkaline pH above the apical membrane.
ognizable lumens. RDCs are activated Isoforms of the Cl−/HCO3− exchanger, AE2,
epithelial cells that secrete a vast array of are responsible for the vast majority of biliary
factors, including cytokines, chemokines, HCO3− secretion. Dysfunction of any of the
growth factors, and angiogenic factors. They elements involved in HCO3− formation
may derive from hepatocytes undergoing a might weaken the biliary HCO3− umbrella
process of ductular metaplasia, or from and contribute to the development of chronic
activation of the hepatic progenitor cell cholestatic liver disease such as sclerosing
compartment and/or from proliferation and cholangitis.
dedifferentiation of preexisting cholangio‑
cytes. The increase in RDCs is generally
associated with a significant increase in Cholangiocytes and Immunity
inflammatory infiltrate and portal fibrosis.
RDCs are considered the major driver of BECs are a first line of defense in liver innate
portal fibrosis during parenchymal and/or immunity: they can present antigen to
biliary injury. The deposition of fibrosis immune cells, be a target of immune‐mediated
Chapter 1 Physiology, Immunology and Pathology of the Liver and Biliary Tree 17
may result from (i) hepatocellular and/or chol‑ failure can result in elevation of the intestinal
angiocellular secretory defects or (ii) obstruc‑ ALP isoenzyme. In patients with raised ALP,
tion of bile ducts by bile duct lesions, stones or hyperthyroidism should be ruled out. Rarely,
tumors, but may also be related to mixed ALP can be identified in patients with under‑
mechanisms in conditions such as PBC or PSC. lying malignancy not involving either liver or
ALP and GGT are markers of cholestasis. bones. This is the Regan isoenzyme, bio‑
ALP is a ubiquitous membrane‐bound glyco‑ chemically different from the liver isoen‑
protein that catalyzes the hydrolysis of phos‑ zyme, that has been described in lung cancer,
phate monoesters at basic pH values. Liver Hodgkin disease, and renal cell carcinoma.
and bone are the major source of serum ALP. Finally, ALP should be tested after fasting
The liver isoenzyme is located on the cana‑ since its level can rise after a fatty meal.
licular side of the hepatocyte plasma mem‑ GGT is an enzyme that can be induced by
brane and the luminal surface of bile duct several stimuli such as drugs and alcohol. It
epithelium. Serum ALP elevation more than is mainly localized in hepatocytes and bil‑
three times normal strongly suggests chole‑ iary epithelia, and is also present in
stasis if bone disease is absent and GGT is extrahepatic tissues such as kidney, spleen,
elevated. In patients with cholestasis, the pancreas, heart, lung, and brain, but not
ALP elevation is triggered by increased syn‑ bone. The lack of GGT in bone can be used
thesis and release of the enzyme into serum to distinguish a liver source from a bone
rather than impaired biliary secretion. BAs source of a raised ALP. GGT is more liver
build up in hepatocytes and solubilize the specific than ALP, although during chole‑
plasma membrane, thereby resulting in stasis is less specific since it can be influ‑
release of ALP. The half‐life of serum ALP is enced by other factors such as alcohol, fat,
5–7 days, and therefore serum ALP remains and drugs. A GGT/ALP ratio over 2.5 may
elevated for several days after resolution of point to alcohol abuse, although up to one‐
the biliary obstruction. ALP is not used as a third of those who abuse alcohol (>80 g/day)
marker of cholestasis in adolescent and preg‑ have a normal GGT. A normal GGT in
nant women since ALP in these conditions patients with elevated liver ALP isoenzyme
can be raised as a consequence of rapid bone should raise the suspicion of benign recur‑
growth and placental growth. Chronic renal rent intrahepatic cholestasis.
20 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
References
1 Ober, E.A. and Lemaigre, F.P. (2018). 6 Overi, D., Carpino, G., Cardinale, V. et al.
Development of the liver: insights into organ (2018). Contribution of resident stem cells
and tissue morphogenesis. J. Hepatol. 68: to liver and biliary tree regeneration in
1049–1062. human diseases. Int. J. Mol. Sci. 19 (10):
2 Roskams, T.A., Theise, N.D., Balabaud, C. 2917. https://doi.org/10.3390/ijms
et al. (2004). Nomenclature of the finer 19102917.
branches of the biliary tree: canals, ductules, Cardinale, V., Wang, Y., Carpino, G. et al.
7
and ductular reactions in human livers. (2011). Multipotent stem/progenitor cells in
Hepatology 39: 1739–1745. human biliary tree give rise to hepatocytes,
3 Halilbasic, E., Claudel, T., and Trauner, M. cholangiocytes, and pancreatic islets.
(2013). Bile acid transporters and regulatory Hepatology 54: 2159–2172.
nuclear receptors in the liver and beyond. J. 8 Fabris, L., Spirli, C., Cadamuro, M. et al.
Hepatol. 58 (1): 155–168. (2017). Emerging concepts in biliary repair
4 Boyer, J.L. (2013). Bile formation and and fibrosis. Am. J. Physiol. Gastrointest.
secretion. Compr. Physiol. 3: 1035–1078. Liver Physiol. 313: G102–G116.
5 Trauner, M., Fuchs, C.D., Halilbasic, E., and 9 Strazzabosco, M., Fiorotto, R., Cadamuro,
Paumgartner, G. (2017). New therapeutic M. et al. (2018). Pathophysiologic
concepts in bile acid transport and signaling implications of innate immunity and
for management of cholestasis. Hepatology autoinflammation in the biliary epithelium.
65 (4): 1393–1404. Biochim. Biophys. Acta 1864: 1374–1379.
21
Abstract
Key Points
●● Autoimmunity results from the complex autoreactive T‐cell receptors, and auto
interplay of genetic, epigenetic, immuno reactive B cells to develop a specific
logic, and environmental factors. autoimmune disease.
●● Environmental triggers initiate loss of tol ●● Subsequently, additional immunologic
erance to autoantigens in genetically sus mechanisms perpetuate chronic progres
ceptible individuals. sive inflammatory disease.
●● Susceptible individuals must have ●● Novel therapies based on immunopatho
inflammatory innate immune responses, genic mechanisms in autoimmunity are in
human leukocyte antigen (HLA) alleles development.
capable of autoantigen presentation to
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
22 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
Autoimmune diseases
Cumulative
adaptive
immune
repertoire
Genetics Environment
Thymic B/T cell selection Microbiome
PAMPs, DAMPs
Sex Gut
Pathogens
HLA alleles Skin
Xenobiotics/drugs
Non-HLA genes Nasopharynx
Hepatic micro-
Epigenetic enhancers Genitourinary
environment
Immune
regulatory
capacity
Figure 2.1 Interaction of factors required for generation of autoimmune diseases. Autoimmune diseases,
including AILDs, result from the complex interaction among factors involving genetic susceptibility, the status
of the cumulative immune response repertoire, an individual’s immune regulatory capacity, environmental
triggers, and the influence of the microbiome.
Chapter 2 Concepts of Autoimmunity Relevant to Autoimmune Liver Diseases 23
(PAMPs) and damage‐associated molecular instead are active participants in the immu
patterns (DAMPs) associated with microbes nopathogenesis of their own injury and
and injured or dying cells, respectively [3]. In demise.
contrast, the adaptive immune system
responds to specific peptide antigens through
Innate Immunity
antigen‐specific T‐cell responses and B‐cell
production of antigen‐specific antibodies [4]. The innate immune system responds imme
It is now clear that cytokines produced by an diately to PAMPs, comprising microbial con
innate immune response dictate both the stituents, and to DAMPs, comprising cellular
type and magnitude of adaptive responses constituents from stressed, neoplastic or
and that both innate and adaptive immunity dying cells (Table 2.2). Both PAMPs and
play important roles in the immunopatho DAMPs bind to pattern recognition recep
genesis of all autoimmune diseases (Table 2.2 tors (PRRs) and other receptors expressed on
and Figure 2.2). neutrophils, dendritic cells (DCs), activated
Innate immune inflammasome responses, macrophages (including Kupffer cells), and
especially those mediated by nucleotide‐ natural killer (NK) and natural killer T (NKT)
binding oligomerization domain (NOD)‐like cells [3]. NK cells express killer receptors for
receptor P3 (NLRP3), occur in both innate stressed, dying or neoplastic cells and Fc
immune cells and non‐immune cells, receptors that mediate antibody‐dependent
including hepatocytes, cholangiocytes, and cell‐mediated cytotoxicity (ADCC) of target
hepatic stellate cells (HSCs) [5]. Thus, clini cells coated with antibodies [3]. Antigen‐pre
cians must embrace a new paradigm: cells or senting cells (APCs) activate NKT and γδT
tissues are not passive “targets” of autoim cells by invariant HLA class I‐like molecule
mune diseases (including AILDs), but (CD‐1) presentation of lipid or glycolipid
24 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
Characteristics
Onset Rapid responses mediated by Delayed due to requirement for antigen
preformed PRRs for microbial activation, clonal proliferation and maturation
PAMPs and DAMPs of effector cell functions
Specificity PAMPs, DAMPs, ROS, Epitopes of peptide antigens recognized by
activated C′ proteins, TCRs or B‐cell immunoglobulins
apoptotic bodies
Genetics Restricted, germline‐encoded Complex with TCRs and antigen‐binding
domains of immunoglobulins produced by
somatic recombination of gene segments
Diversity Limited Virtually infinite due to TCR rearrangement
Evolutionarily conserved
Memory No Yes. Memory T‐ and B‐cell responses capable
of anamnestic reactivation
Self‐tolerance Discriminates pathogens and Positive and negative selection of TCR and
endogenous DAMPs. Not B‐cell immunoglobulin reduce autoreactive
responsive to autoantigens capacity
Components
Physical barriers Skin, mucosal epithelia of Intraepithelial lymphocytes within gut, Peyer’s
oropharynx, gut and patches
reproductive tract,
antimicrobial proteins
Cells Dendritic cells, monocytes, Professional APCs, αβT cells, γδT cells,
macrophages (Kupffer cells), natural and inducible Treg cells, B cells,
neutrophils, NK cells, NKT plasma cells
cells
Proteins C′ proteins, IFN‐α, ‐β, ‐γ, T‐cell cytokines
cytokines, chemokines IgM, IgG, IgA, IgE antibodies
Role of liver as Yes Yes
an immunologic DCs, neutrophils, Kupffer Antigen presentation by Kupffer cells, LSECs,
organ cells, excessive concentrations HSCs, cytokines stimulate hepatocytes and
of NK cells, NKT cells cholangiocytes, PD‐L1/2 induce senescent
IFN‐α, ‐β, ‐γ, balance between T‐cell apoptosis and inhibition of activated
proinflammatory and CD8 T cells
immunosuppressive
chemokines
C′, complement; DAMPs, damage‐associated molecular patterns; HSC, hepatic stellate cell; IFN, interferon; LSEC,
liver sinusoidal endothelial cell; NK, natural killer; PAMPs, pathogen‐associated molecular patterns; PD‐L1/2,
programmed death ligands 1 and 2; PRRs, pattern recognition receptors; ROS, reactive oxygen species; TCR, T cell
receptor.
Examples of common PAMPs: (i) lipopolysaccharide (LPS), a cell wall constituent of all Gram‐negative bacteria;
(ii) lipotechoic acid, a cell wall constituent of all Gram‐positive bacteria; (iii) peptidoglycans, essential cell wall
components of all bacteria; (iv) viral proteins; and (v) unmethylated bacterial CpG dinucleotide motifs.
Examples of common DAMPs: (i) nuclear proteins; (ii) cytosolic proteins and lysosomal enzymes; and
(iii) non‐proteins, such as ATP, DNA, RNA, mitochondrial DNA, U1 ribonucleoprotein (U1RNP) and uric acid.
Chapter 2 Concepts of Autoimmunity Relevant to Autoimmune Liver Diseases 25
antigens [3]. γδT cells protect from autoim secrete collagen, resulting in fibrosis. PAMPs
mune diseases and microbial infections and and DAMPs also induce NLRP3 inflamma
participate in mucosal immunity and tumor some responses [5] in hepatocytes, cholan
surveillance. giocytes and HSCs, resulting in their
Microbial PAMPs bind predominantly to secretion of proinflammatory cytokines and
PRRs called Toll‐like receptors (TLRs). chemokines.
DAMPs bind to PRRs and receptors for The reticuloendothelial function of the
chemical ligands or nuclear constituents. liver normally prevents PAMPs, DAMPs and
Still other innate receptors bind apoptotic microbial antigens in portal venous blood
bodies and activated complement (C′) mole from entering the systemic circulation. This
cules. PAMPs and/or DAMPs activate DCs barrier function of the liver also plays a role
and macrophages (including Kupffer cells), in oral tolerance. Hepatic lymph formed in
increasing phagocytic activity and inducing the space of Disse flows retrograde into
secretion of chemokines and proinflamma portal tract lymphatics. Thus, hepatic lymph
tory cytokines interleukin (IL)‐1β, IL‐6, exposes portal bile ducts, arterioles and
IL‐12, and IL‐18 and tumor necrosis factor portal vein branches to variable amounts of
(TNF)‐α, as well as the immunosuppressant PAMPs, DAMPs, cytokines, chemokines,
cytokine IL‐10 [3]. The relative amounts of proteins, lipoproteins, and lipids.
specific PAMPs, DAMPs, cytokines and che
mokines creates a dynamic balance between
Adaptive Immunity and Adaptive
proinflammatory and immunosuppressive
Immune Functions of the Liver
cytokines that determines whether the
adaptive immune response promotes immu As shown in Figure 2.2, the adaptive immune
nopathology (Figure 2.2). system in autoimmunity generates multiple,
delayed, autoantigen‐specific effector immune
responses [4]. Most adaptive immune
Liver as an Innate Immune Organ
responses occur in lymph nodes and spleen;
The liver is an essential innate immune organ thus, effector cells must circulate to tissues
that must either tolerate or respond to expressing autoantigens to mediate immuno
microbial PAMPs, toxins, xenobiotics, food pathology. Class II HLA‐DR, ‐DQ or ‐DP
antigens, microbial pathogens, and neo molecules present peptide autoantigens to T‐
plastic cells [6]. The liver contains special cell receptors (TCRs) of autoreactive CD4 T
ized liver sinusoidal endothelial cells (LSECs) cells. In contrast, class I HLA‐A, ‐B or ‐C mol
and large quantities of Kupffer cells, DCs and ecules present peptide autoantigens to TCRs
NK, NKT and γδT cells. Most intrahepatic B of autoreactive cytotoxic CD8 T cells, also
cells are B1 cells, an innate B‐cell subset that known as cytotoxic T lymphocytes (CTLs).
secretes low‐affinity IgM antibodies against Polymorphisms in HLA class I and II alleles
glycoproteins, called “natural antibodies.” determine which autoantigens can be
The liver also synthesizes C′ proteins, presented to T cells, which explains the strong
acute‐phase reactant proteins, and multiple association of autoimmunity with specific
cytokines. PAMP and DAMP activation of HLA alleles [2]. Polymorphic class III HLA
TLRs on Kupffer cells, LSECs, hepatocytes, alleles encode immune response proteins,
DCs and HSCs primarily generates IL‐10, including TNF‐α/β, complement (C′) factors
providing an immunosuppressive environ and heat‐shock proteins. The balance bet
ment. However, high quantities of PAMPs or ween proinflammatory and anti‐inflammatory
DAMPs induce the proinflammatory cyto cytokines produced by local innate immune
kines IL‐1β, IL‐6, IL‐12, IL‐18 and TNF‐α. responses determines the polarization of
HSCs normally produce matrix proteins, but cytotoxic mechanisms of the adaptive immune
once transformed into myofibroblasts they response (Figure 2.2).
IFNy
CD8 TNFα
Thymus CTL
IL-2
nTreg
IL-12 CD4
IFNγ Mechanisms of
Th1
IFNγ Cytotoxicity
Macrophage TNFα
nTreg nTreg HLA IL-4 IL-9 Cell-Mediated
CD4 IL-1β
APC Class II IL-10
TGFβ Th9
IL-21 IL-23 • CD8 cytotoxic T cells
Costimulation:
• NK or NKT cels
CD80/86 CD40 Fc • Macrophages
Ag NK
HLA Receptor
TCR Class I CD4
cell
CD28 CD154 IL-4 IL-4 Antibody-Mediated
Th2
Naive
CD8 IL-10 B Plasma Target
CD4 • IgG, IgM, or IgA
CTL IL-13 cell cell cell
Th0 • Complement
CD4 IL-21
IL-6 NKT • Immune complexes
Tfh
• Antibody-Dependent
cell cellular cytotoxicity
Autoantibodies Complement (ADCC) mediated by NK
IL-23 activation Fc receptors
IL-4 IL-12 CD4
IL-6
IL-10 IL-6
Anti-inflammatory
cytokines
CD4 iTregs to
CD4 Th
CD4 Th17 cells
subsets
CD4
TGFβ iTreg
IL-23
Local innate CD4
IL-6
CD4
immune responses Tr1 Th3 IL-1β
TGFβ
Figure 2.2 Adaptive immune responses generate multiple mechanisms of cytotoxicity in autoimmunity. Failure of thymic nTregs allows professional APC
presentation of peptide autoantigens to autoreactive naive CD4 Th0 and CD8 T cells. APC class II HLA molecules, encoded by HLA‐DR, −DQ, and ‐DP alleles, present
peptide autoantigens to TCRs of CD4 Th0 cells. APC class I HLA molecules, encoded by HLA‐A, ‐B and ‐C alleles, present different peptide autoantigens to TCRs of
naive CD8 CTLs. APC processing of autoantigen–immunoglobulin immune complexes dramatically increases autoantigen‐specific activation of CD8 CTLs.
Polymorphisms in HLA class I and II alleles determine which autoantigens are presented to TCRs. Differentiation of functional subsets of autoantigen‐specific CD4 Th
cells and cytotoxic CD8 T cells depends on two factors: costimulation of HLA–antigen activated T cells and the composition of the cytokine milieu produced by local
innate immune responses. Potent costimulation is mediated by binding of CD80/CD86 expressed on APCs to CD28 expressed on T cells or binding of CD40 expressed
on APCs to CD154 (aka CD40 ligand, CD40L) expressed on T cells. If costimulation is absent or inadequate, antigen‐activated CD4 Th0 and CD8 T‐cell clones become
anergic to subsequent reexposures to the activating peptide antigens. When local innate immune responses produce abundant proinflammatory cytokines IL‐12,
IL‐6, IL‐1β and IFN‐γ, activated CD4 Th0 cells predominantly differentiate into Th1 cells, rather than Th2 cells. Secretion of other cytokines, acting alone or in
combination, promotes CD4 Th0 differentiation into other CD4 T‐cell subsets. CD4 Th1 cytokines and growth factors drive the proliferation and differentiation of
cytotoxic antigen‐specific CD8 CTLs. Th1 cytokines also activate tissue macrophages and induce B‐cell secretion of C′‐fixing IgG2a. IL‐4 produced by local innate
immune responses favors CD4 Th0 differentiation into Th2 cells that secrete IL‐4, IL‐5, IL‐10, and IL‐13 and induce B‐cell secretion of IgG, IgM, IgA, and IgE. CD4 Th1
and Th2 cells achieve a dynamic balance by secreting signature cytokine profiles that inhibit the proliferation and cytokine secretion of the other subset. TGF‐β and
IL‐6 play important roles, both alone and in combination, in the differentiation of CD4 Tfh, Th17, and iTreg cells. IL‐21, IL‐23 and RORγ, RORα promote differentiation
of Th17 cells, which intensify and perpetuate tissue inflammation. Tfh and Th9 cells secrete the pluripotent cytokine IL‐21. IL‐21 is a pivotal cytokine in autoimmune
disease pathogenesis mediated by autoantibodies and cytotoxic cells and cytokines. Specifically, IL‐21 induces B‐cell differentiation into plasma cells, increases
proliferation of CD4 subsets, especially Th17 cells, increases resistance to iTregs and proliferation and cytokine production of NKT cells, increases NK cell cytotoxicity
and ADCC of immunoglobulin‐coated target cells, and increases IFN‐γ secretion and proliferation and cytotoxicity of CD8 CTLs. IL‐9 stimulates cell proliferation and
inhibits apoptosis, while IL‐13 is an anti‐inflammatory cytokine that induces IgE and enhances APC survival. Tfh cells localize within B‐cell follicles in lymph nodes and
Peyer’s patches where they promote selection and survival of B‐cell clones by secreting IL‐4 and IL‐21. Tfh cells play a central role in the formation of tertiary germinal
centers characteristic of autoimmune diseases. CD4 iTreg cells suppress antigen‐specific T‐cell responses of all CD4 T‐cell subsets by secreting immunosuppressive
IL‐10 and TGF‐β. Local secretion of IL‐23, Il‐6, IL‐1β, and TGF‐β by cytokine‐stimulated epithelial cells (e.g. cholangiocytes) transform iTregs into proinflammatory Th17
cells, defeating the immunoregulatory function of iTregs. Innate NK, NKT cells and activated macrophages and adaptive immune effector B and T cells act together to
mediate target cell cytotoxicity. NK cells express killer inhibitory receptors (KIRs) that inhibit killing of normal cells but target abnormal cells requiring elimination.
Activated NK cells secrete abundant IFN‐γ, a synergizing cytokine in the proinflammatory cytokine milieu. NKT cells react to lipid or glycolipid antigens presented by
CD1d molecules and secrete IFN‐γ, IL‐2, IL‐4, TNF‐α, and granulocyte–macrophage colony‐stimulating factor (GM‐CSF). NK group 2 member D (NKG2D) receptors on
activated macrophages, NK, NKT and γδT cells induce apoptosis by engaging MICA and MICB ligands produced by HLA class I‐like genes in abnormal cells. Activated
macrophages and NK and NKT cells also express Fc receptors that mediate antibody‐dependent cellular cytotoxicity.
28 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
After exiting lymph nodes, circulating Furthermore, binding of Kupffer cell PD‐L1
activated T cells and B cells undergo transen to PD‐1 expressed by activated CD4 T cells
dothelial migration to enter inflamed tissues induces immunosuppressive IL‐10 and
expressing chemokines [1]. Activated T cells, reduces apoptosis of CD4 T regulatory cells
B cells, and other leukocytes expressing che (Tregs).
mokine receptors migrate through the tight Mucosal invariant T (MAIT) cells have
junctions of endothelial cells, and their che characteristics of both innate and adaptive
mokine receptors mediate migration toward immune cells that are relevant in AILDs [7].
the cellular sources of chemokine secretion. MAIT cells have invariant TCR α‐chains that
This results in mixed inflammatory infil respond to bacterially processed vitamin B
trates that mediate cytotoxicity (Figure 2.2). antigens presented by unique major
histocompatibility complex (MHC) class I‐
related molecules (MR‐1) on APCs. Normally
Role of the Liver as an Adaptive
found only in blood, gut mucosa and mesen
Immune Organ
teric lymph nodes, MAIT cells congregate in
In addition to the innate immune functions, the portal tracts of patients with chronic liver
the liver has important roles in adaptive diseases, including AIH, PBC, PSC, alcoholic
immunity [6]. The normal liver is a distinct and non‐alcoholic fatty liver diseases, and
“lymphoid” compartment containing CD8 αβ chronic hepatitis C virus (HCV) infection.
T cells, activated CD4 and CD8 T cells, γδT Proinflammatory cytokines IL‐12 and IL‐18
cells, memory T cells and B cells. Indeed, the activate MAIT cells to display dual charac
proportions of highly activated T cells in the teristics of CD4 Th1 and Th17 cells by
liver exceed those in blood. In contrast, secreting interferon (IFN)‐γ, TNF‐α and
normal liver contains scant numbers of naive IL‐17. Up to 25% express cytotoxic granzyme
T cells and B cells. B, a feature of cytotoxic NK, NKT, and CD8
Most often, antigen‐activated DCs migrate T cells. Importantly, MAIT cell cytokines
to local lymph nodes to activate non‐hepatic stimulate cholangiocyte secretion of cyto
T cells, which subsequently circulate and kines capable of transforming CD4 iTregs
enter tissues through transendothelial migra into proinflammatory CD4 Th17 cells
tion. However, hepatic DCs, LSECs, and (Figure 2.2).
HSCs also function as APCs for intrahepatic
T‐cell activation of adaptive immunity.
Hepatocytes and cholangiocytes may also Generation and
serve as APCs after cytokine stimulation.
Maintenance of Tolerance
CD4 T cells activated by hepatic APCs pri
marily differentiate into Th2 cells, secreting to Self‐antigens
immunosuppressive IL‐10 and IL‐4.
Overview
Frequently, CD8 T cells activated in the liver
have functional deficiencies leading to pre Immune tolerance is the ability of the
mature apoptosis. However, LSECs can immune system to identify and respond to
directly activate functional antigen‐specific foreign or non‐self‐antigens, while aborting
CD8 CTLs in the liver by cross‐presentation or controlling potentially deleterious
of soluble exogenous antigens in class I HLA responses to autoantigens [1]. Despite ge
molecules. netic variability within the population and
Hepatic Kupffer cells regulate T‐cell substantial differences in environmental
homeostasis by inducing apoptosis of senes exposures that shape an individual’s immune
cent T cells expressing CD95 (Fas), TNF‐ repertoire, the vast majority of people main
apoptosis‐inducing ligand (TRAIL), and tain immune tolerance to self‐antigens. As a
programmed cell death 1 (PD‐1) receptors. result, autoimmune diseases are rare, even
Chapter 2 Concepts of Autoimmunity Relevant to Autoimmune Liver Diseases 29
may manifest as more than one clinical dis HLA‐DRB1*03:01 haplotypes associated
ease (Figure 2.1) [1]. Overall, environmental with autoimmune diseases. It is notable that
exposures are more important than genetics even healthy persons with the 8.1 haplotype
in shaping the immune repertoire. Thus, a exhibit an enhanced immunologic pheno
person’s “exposome” – the lifelong sum of all type with autoantibodies, circulating
endogenous and exogenous environmental immune complexes, activated T cells, and
exposures interacting with genetic and increased apoptosis of lymphocytes. The
epigenetic factors – can favor autoimmunity HLA‐DRB1 gene is of interest because it is
or protect against it. the most polymorphic gene within the HLA
Genome‐wide association studies (GWAS) class II region. Thus, it is not surprising that
in autoimmune diseases, including AILDs, PBC is associated with the DR allele HLA‐
have demonstrated shared associations with DR8. Future studies are needed to define how
HLA alleles and single nucleotide polymor different HLA alleles contribute to a break in
phisms (SNPs) of genes related to innate tolerance and development of specific auto
immunity and adaptive immunity [15]. immune diseases.
These include genes encoding interleukins,
interleukin receptors, proinflammatory Non‐HLA Gene Associations
cytokines and their receptors, cytotoxic T Interpreting the risks for autoimmune dis
lymphocyte antigen (CTLA)‐4 (a mediator eases conferred by non‐HLA genes is prob
of senescence and apoptosis of activated T lematic because the odds ratios for risk of
cells), signal transducer and activator autoimmune diseases are far lower than
(STAT)‐4, chemokine receptors and CD69 those for HLA alleles [15]. Polymorphisms
(involved in migration of lymphocytes from in genes encoding CTLA‐4, TNF‐α, Fas
lymph nodes and development of immuno (CD95), TNF‐induced protein 3 (TNFAIP3),
logic “memory”). However, odds ratios for macrophage migration inhibitory factor
each gene’s contribution to risk for autoim (MIF), and SH2B adapter protein 3 (SH2B3)
munity are quite small, suggesting that mul have been implicated in susceptibility to
tiple SNPs are required for development of autoimmune diseases, including AILDs.
one or more of the 80 known autoimmune CTLA‐4 is critical for downregulating
diseases. effector immune responses and producing
T‐cell senescence. TNF‐α is a potent
HLA Risk Alleles proinflammatory effector cytokine and
The strongest HLA‐associated risks for auto therapeutic target in specific autoimmune
immune diseases, including AIH and PSC, diseases. Dysfunction of genetic variants or
but not PBC, lie within the evolutionarily deficient levels of TNFα‐induced protein
conserved 8.1 ancestral haplotype: HLA‐A1, 3‐interacting protein 1 (TNIP1) predispose
Cw7, B8, TNFAB*a2b3, TNFN*S, C2*C, Bf*s, normal innate cells to produce abnormal
C4A*Q0, C4B*1, DRB1*03:01, DRB3*01:01, inflammatory responses to innocuous TLR
DQA1*05:01, DQB1*02:01 [2]. The extended ligands. TNFAIP3 is a modifying enzyme for
8.1 haplotype is the result of linkage disequi ubiquitin, which may influence antigenicity
librium, indicating an evolutionary advantage of self‐proteins. CD95 (Fas) and CD95L
to sequestration of these alleles within class I, (FasL) variants promote survival of autore
II and III HLA loci. These alleles are impor active T and B cells and modify apoptosis of
tant for robust CD4 and CD8 T‐cell activation target cells. MIF is a proinflammatory cyto
and generation of TNF‐α/β and C′ factors. kine in both innate and adaptive immunity.
The 8.1 haplotype is associated with AIH and SH2B3 suppresses cytokine‐induced inflam
PSC. Most Caucasians with HLA‐B8, mation by downregulating Janus kinases
DR3 have the 8.1 haplotype; however, (JAKs) and receptor tyrosine kinases (RTKs).
recombinations have produced additional
Several autoimmune diseases also are linked
Chapter 2 Concepts of Autoimmunity Relevant to Autoimmune Liver Diseases 33
to SNPs in cytokine and cytokine receptor immune cells. lncRNAs promote suscepti
genes. A pertinent example is the IL23R bility to autoreactivity and perpetuation in
gene, which encodes a protein that forms a autoimmune diseases, including systemic
dimer with IL‐12Rβ1 to create the IL‐23 lupus erythematosus (SLE), rheumatoid
receptor complex required for IL‐23 arthritis (RA), multiple sclerosis (MS), and
induction of proinflammatory Th17 cells. psoriasis.
MicroRNAs
Critical Role of Epigenetics
MicroRNAs (miRNAs) function epigeneti
Transcription Factor Enhancers cally as posttranscriptional and posttransla
and Super Enhancers tional regulators of gene expression [17].
Gene transcription is an essential process in miRNAs and TFs regulate each other’s
autoimmunity and autoimmune disease expression in feedback loops: TFs are nega
specificity. Emerging evidence indicates that tively regulated by miRNAs, while miRNAs
the generation and perpetuation of autoim are positively regulated by TFs. Compelling
munity is primarily driven by epigenetics evidence indicates that miRNAs control the
[16]. Over 90% of the thousands of disease‐ activation state of innate macrophages and
associated SNPs detected by GWAS are DCs, activation of T cells, apoptotic elimina
localized within non‐coding regions of DNA, tion of activated T and B cells, and the critical
that produce regulatory enhancers, some balance between iTregs and Th17 cells in
times occurring in clusters spanning 50 kb of sites of inflammation. To date multiple
DNA, called super enhancers (SEs). Multiple miRNAs have been implicated in the
enhancers and SEs control the expression of immunopathogenesis of autoimmune dis
protein‐coding genes. These enhancers, eases, including AILDs.
especially SEs, are especially important
drivers of cell‐ and tissue‐specific genes.
Sex and Sex Hormones
Enhancer functions are mediated through
transcription‐factor (TF) binding motifs that Most autoimmune diseases have a female
require TF recruitment and nucleation of predilection [1,18]. However, PSC is a
transcriptional machinery. Thus, enhancer notable exception among AILDs (Table 2.1).
activity depends on the local structure of Sex is genetically binary and imprinting of
chromatin responsible for the packaging of XX female and XY male genomes occurs in
DNA in association with histone proteins. all cells from embryogenesis throughout
Since tightly packed DNA is inaccessible to life [18]. Both innate and adaptive immune
DNA‐binding TFs, additional epigenetic cells have sex hormone receptors, and sex
mechanisms are required to render DNA hormones contribute to the development
accessible to enhancer TFs. These processes and activity of the immune repertoire
include methylation of DNA and posttran (Figure 2.1). Yet it is unclear whether female
scriptional modification of histones through predominance in autoimmune diseases
methylation, acetylation, or ubiquitination. reflects regulatory mechanisms related to
Epigenetic proteins also interact directly sex‐linked genes, sex‐specific hormones or
with transcriptional mechanisms and a combination of both. Females normally
enhancer DNA can also generate enhancer produce higher titers of antibodies, more
RNA (eRNA) to increase the amount of autoantibodies and greater cell‐mediated
enhancer proteins. Epigenetic long non‐cod immunity to infections and immunization
ing RNAs (lncRNAs) are expressed within than do normal males. In addition, there is
innate immune cells and CD4 Th1, Th2, a gender bias in AIRE expression in the
Th17, Tregs, CD8 and B cells. lncRNAs thymus, which may impact deletion of
directly regulate activation and functions of autoreactive clones and formation of nTregs
34 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
and Bregs. Estrogen levels are also immu c orrelation between the incidence of autoim
noregulatory: high levels inhibit CD4 Th1‐ mune disease and increasing latitude.
mediated cellular responses and promote SNPs in two VDR genes, BsmI and TaqI,
CD4 Th2 antibody‐mediated responses. are associated with AILDs. SNPs in these
Prolactin, progesterone and testosterone genes and FokI are associated specifically
also regulate immune responses by modu with AIH. Dietary vitamin D and vitamin D
lating expression of estrogen receptors and synthesized in response to ultraviolet‐B
altering cytokine secretion. However, the radiation in sunlight determine serum
effects of estrogen and other sex hormones vitamin D levels. Sequential hydroxylation
in adult women do not explain the female in the liver and kidney produce D3, the most
predilection for autoimmunity in children potent ligand for VDRs. Innate and adaptive
or elderly adults. Postulated explanations immune cells express VDRs that induce
include skewed X chromosome activation, multiple immunologic effects. For example,
X monosomy, and microchimerism. D3 regulates production of antimicrobial
Beginning during embryogenesis and peptides and reactive oxygen species in
continuing through life, either the maternal innate immune cells and promotes NKT cell
or paternal X chromosome within each cell immunomodulatory functions. D3 also
is inactivated, forming a Barr body adjacent downregulates immunopathogenic CD4
to the nucleolus. Thus, females are a mosaic Th1 and Th17 cells, while upregulating
of two cell types, based on the parental immunosuppressive CD4 Th2 cells and
source of the active X chromosome. iTregs. By reducing Th1 secretion of IL‐2,
Skewing of X inactivation varies widely and D3 not only inhibits proliferation of CD4
increases with age. Thus, some X‐linked and CD8 effector T cells but also generates
autoantigens might theoretically escape low IL‐2 levels, favoring proliferation of
central or peripheral tolerance. The functional iTregs.
proximity of Barr bodies to the nucleolus
suggested that they may interact to expose
cryptic autoantigens. X monosomy might L oss of Immune Tolerance to
promote autoimmunity by preventing tol Autoantigens and Perpetuation
erance to autoantigens encoded by both X
chromosomes. Entry of fetal hematopoietic
of Autoimmune Diseases
stem cells into the maternal circulation
Overview
during pregnancy can cause fetal–maternal
microchimerism that could promote loss of Autoimmune diseases are uncommon,
self‐tolerance. despite the high frequency of genetic suscep
tibility and environmental exposure to
microbial pathogens and xenobiotics,
Vitamin D and Sunlight Exposure
including drugs. Autoimmunity requires
In addition to its roles in bone mineralization four essentials. First, environmental triggers
and calcium homeostasis, vitamin D modu are required to initiate loss of tolerance to
lates immune reactions and risk for autoim autoantigens in susceptible individuals
munity [19]. Both vitamin D deficiency and (Figure 2.1). Second, the individual must
genetic polymorphisms in the vitamin D have HLA alleles that facilitate autoantigen
receptor (VDR) increase the incidence and activation of T cells. Third, T cells must
severity of multiple autoimmune diseases. express autoreactive TCRs. Fourth, the
Conversely, high vitamin D levels reduce the individual must fail to immunoregulate the
risk of MS. Reduction in sunlight exposure response to autoantigens. Evidence that low
and synthesis of 1,25‐dihydroxyvitamin‐D3 levels of autoimmune reactions are common
(D3) might explain, in part, the strong in healthy people highlights the fact that
Chapter 2 Concepts of Autoimmunity Relevant to Autoimmune Liver Diseases 35
these initial steps are common but only dele As discussed earlier, bacterial metabolites of
terious when not immunoregulated. GWAS vitamin B are the activating antigens for
indicates that failure to immunoregulate an MAIT cells. Thus, the gut microbiome and
autoimmune reaction is dictated by epige intestinal immune responses appear to
netic SNPs for enhancers and SEs of gene influence systemic immunity, and risk for
expression more than SNPs for HLA and autoimmunity, through the innate and
non‐HLA proteins. Moreover, these SNPs adaptive immune responses of the liver.
are critical drivers of the cell and tissue spec Altered composition of the gut microbiota,
ificity of autoimmune diseases. Multiple referred to as “dysbiosis,” has been reported
factors and mechanisms to break tolerance in autoimmune diseases, including AILDs.
to self‐antigens have been identified. Whether dysbiosis is the cause or the effect
The increasing incidence of autoimmunity of autoimmunity remains a key unresolved
and inflammatory diseases observed world issue. Female sex hormones can also alter the
wide is correlated with changes in environ gut microbiome, which may contribute to
mental factors, including a more modern the female predilection in autoimmunity. If
lifestyle, improved hygiene, a Western diet, an altered gut microbiota or increased gut
use of antibiotics, and elimination of permeability were causally related to autoim
childhood parasitic infections. Each of these mune diseases, then restoration of a normal
affect the composition and function of the microbiota and mucosal integrity could be
gut microbiota. Although the proposed therapeutic.
causal link between the gut microbiome and
autoimmunity has not been proved, available
Mechanisms of Loss of Tolerance
data indicate that interplay between the gut
to Autoantigens
microbiota and the innate and adaptive
immune systems of the intestine and liver Bacterial and Viral Infections
plays key roles in both normal and dysfunc Bacterial, fungal or viral infections can insti
tional systemic immunity. gate innate and adaptive immune responses
that result in autoimmunity [21]. Migration
Role of the Microbiome of neutrophils to sites of infection or sterile
The gut microbiome has been implicated in inflammation mediated by IL‐8, IFN‐γ, C3a,
autoimmunity (Figure 2.1) [20]. The gut C5a, and leukotriene B4 plays a key role.
microbiota change dynamically under the Neutrophil phagocytosis of microbial organ
influence of aging, sex hormones, diet, water isms and particles leads to release of soluble
purity, hygiene, sanitation, pollution, antimicrobials and formation of neutrophil
exposure to pathogens, use of antibiotics, extracellular traps (NETs) comprising
and the integrity of the gut–blood barrier. chromatin and proteases that enmesh and
Gut microbiota are composed of commensal kill microbes. The proteolytic function of
bacteria, fungi and viruses, as well as poten NETs also can generate neoantigens or
tially pathogenic bacteria and fungi. In microbial peptide molecular mimics of auto
addition, the microbiota process food, antigens (see next section).
xenobiotic pollutants and drugs to generate Superantigens are proteins produced by
micronutrients and metabolites. Intestinal pathogenic bacteria and viruses that trigger
immune responses and the leakiness of the non‐antigen‐specific polyclonal T‐cell
mucosa dictate the types and concentrations activation of up to 20% of the host’s T cells.
of PAMPs, DAMPs, cytokines, microbes, These activated T cells create a cytokine
and food antigens entering the portal vein for storm of secreted cytokines, especially IFN‐γ,
processing in the liver. PAMPs and DAMPs which activates macrophages to secrete
stimulate enterocyte inflammasomes leading proinflammatory IL‐1β, IL‐6 and TNF‐α.
to secretion of proinflammatory cytokines. This proinflammatory environment could
36 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
defects in Fas (CD95), Fas‐L (CD95L) and RAS T‐cell Receptor Revision
pathways cause lymphoproliferative disorders in the Periphery
and autoimmunity due to inability to eliminate The fact that the TCR repertoire selected in the
autoreactive effector cells. thymus can be altered in the periphery helps
explain the presence of rogue autoreactive T
Immune Deviation of Activated T Cells cells [24]. Surface expression of the costimula
Immune deviation refers to the evolution of tory molecule CD40, thought to be restricted
dominant populations of effector T cells, to APCs, has been observed recently on neural
which alter local immune responses and com cells, endothelial cells, adipocytes, and subsets
promise tolerance [1]. Activated CD4 Th0 of CD4 and CD8 T cells. Activation of CD40
differentiate into multiple functional Th1 expressed by the mature CD4 Th subset,
subsets differing in cytokine production designated Th40 cells, induces RAG‐1/RAG‐2‐
(Figure 2.2). CD4 Th1 cells secrete IL‐2, the mediated rearrangement of both TCR α‐ and
mitogen for all activated CD4 and CD8 T β‐chains in the periphery. This refutes early
cells, as well as IFN‐γ and TNF‐α. Th1 cells dogma of the immutability of the TCR reper
activate macrophages and stimulate B cells to toire and its antigen specificity after T cells
secrete C′‐fixing IgG2. In contrast, Th2 cells emigrate from the thymus. The fact that Th40
secrete IL‐4, IL‐5, IL‐10, and IL‐13 that stim cells can undergo multiple cycles of TCR revi
ulate B cells to secrete IgG, IgM and IgA anti sion multiplies the risk of developing autoreac
bodies, while immunosuppressing the effects tive TCRs, including ones recognizing unique
of Th1 cytokines. Th1 predominance pro self‐neoantigens. The finding that adoptively
duces greater immunopathology associated transferred Th40 cells induced antigen‐specific
with autoimmune diseases that cause tissue type 1 diabetes mellitus (T1DM) in non‐obese
damage. The signature cytokines of Th1 and diabetic severe combined immunodeficiency
Th2 cells inhibit the proliferation and secre mice confirms their immunogenic function.
tion of the cytokines of each other, resulting Thus, TCR revision to recognize autoantigens
in a dynamic balance. Skewing of this balance represents a new paradigm, explaining autore
contributes to either the maintenance or the active peripheral T cells.
loss of tolerance. The T follicular helper (Tfh)
cell secretes IL‐21, which is best known for
Perpetuation of Autoimmune
inducing differentiation of activated B cells
Diseases
into memory B cells and plasma cells.
Additional pluripotent effects of IL‐21 impact The primary factor in perpetuation of autoim
both innate and adaptive immune functions mune diseases is failure to immunoregulate
to increase immunopathology. In innate and terminate the initial activation of B and
immunity, IL‐21 increases antigen processing T cells to autoantigens or their mimics [1].
and presentation by APCs, a ctivates macro Multiple factors contribute to perpetuation.
phages to chemoattract neutrophils, increases
NK cell cytotoxicity, including ADCC and Epigenetics
secretion of IFN‐γ, and induces NKT cell pro Enhancers and Super Enhancers
liferation and secretion of IFN‐γ, IL‐2, IL‐4, As discussed earlier, epigenetics plays a crucial
IL‐13, and IL‐17A. In adaptive immunity, role in regulating TFs in innate and adaptive
IL‐21 induces proliferation and differentiation immune responses involved in autoimmune
of Th17 cells and increases both the cytotox diseases and determining specificity for differ
icity of CD8 CTLs and their secretion of ent cell types and tissues [16]. While the mech
IFN‐γ and TNF‐α. Thus, polarization of anisms already discussed focus on factors
immune responses toward Th1 and Tfh cells leading to autoantigen recognition, it is now
greatly increases the consequences of autore clear that epigenetics dictates the subsequent
active T and B cell activation. obligatory failure to regulate and terminate
38 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
epithelial cells and result in bilateral stimula ation of autoimmunity provide conceptual as
tion. These cytokines have a dual potential well as realistic targets for interventions to
to immunoregulate innate and adaptive prevent and treat autoimmune diseases [1].
immune responses or to promote immuno Clearly, genetic and epigenetic SNPs and envi
pathogenesis. These cytokines are most often ronmental exposures cannot be eliminated as
pathogenic in autoimmune diseases, perpet risk factors. However, it may become possible
uating chronic tissue inflammation. to identify children at risk of autoimmunity
and develop strategies to reduce their risk of
Tertiary Lymphoid Structures autoimmune diseases. The development of
and Germinal Centers new immunosuppressive medications, inhibi
Tertiary lymphoid structures (TLSs) frequently tion of cytokine production and function, and
develop in autoimmune diseases that contain epigenetic inhibitors increase the probability
T‐cell and B‐cell zones capable of producing of controlling a variety of autoimmune dis
adaptive cellular and humoral immune eases in the near future.
responses [27]. TLSs portend poor prognosis
due to increased autoactivation of T‐ and B‐cell Strategies to Prevent Autoimmunity
clones. Excessive production of autoantibodies
against initiating autoantigens and those Vitamin D3
involved with epitope spreading augment anti Vitamin D deficiency is epidemiologically
body‐mediated cytotoxicity during the evolu associated with risk of autoimmunity [19].
tion of autoimmune diseases (Figure 2.2). In Achieving and maintaining high normal
addition, APC phagocytosis and processing of serum levels of vitamin D3 is a realistic and
immune complexes composed of autoantigens achievable goal, which could reduce the
and autoantibodies dramatically increase the incidence and severity of autoimmune
quantity of autoantigen‐specific CD8 CTLs. diseases.
Tfh cells and their signature cytokine IL‐21
drive B‐cell production of high‐affinity autoan Gut Microbiota Manipulation
tibodies, differentiation of memory B cells, and in Pregnancy and Infancy
APC functions of activated B cells within the It remains unclear if changes in the fecal
expanded number of germinal centers in TLSs. microbiota associated with specific autoim
mune diseases represent causes or effects
Epithelial Cell‐induced Transformation [20]. Were causal relationships identified for
of iTreg to Th17 Cells either initiation or perpetuation of autoim
As discussed earlier, cytokines produced by munity, several strategies theoretically could
MAIT cells can stimulate epithelial cells, decrease the risk of autoimmunity, espe
including cholangiocytes, to secrete IL‐6, IL‐1β, cially if used during pregnancy and infancy.
IL‐23 and TGF‐β [7]. This combination of cyto These include probiotics to shape the evolu
kines can transform iTregs into activated Th17 tion of the gut microbiota and sustain the
cells that promote a proinflammatory Th17‐ mucosal barrier, fecal microbiota transplan
mediated inflammation (Figure 2.2). tation to create a preventive gut microbi
ome, and either deliberate infection with
non‐pathogenic helminths or ingestion of
revention of Autoimmunity
P specific parasite peptides to promote a
systemic immunosuppressive Th2 environ
and Therapeutic Control
ment [28].
of Autoimmune Diseases
Oral Tolerance
Overview
Ingestion of antigens leads to absorption and
Our current understanding of the mecha interaction with the reticuloendothelial
nisms involved in the generation and perpetu system of the liver. Repeated ingestion can
40 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
result in oral tolerance, defined as the T1DM, HCV‐associated vasculitis, and SLE.
inability to respond to the ingested antigen The second strategy involves generating
when given parenterally [29]. Low doses of antigen‐specific iTregs from peripheral
antigen induce iTregs, while higher doses blood mononuclear cells ex vivo for infusion.
create anergic tolerance. Studies involving iTregs specific for CYP2D6 autoantigens in
the induction of oral tolerance to glutamic type 2 AIH have been produced, but not yet
acid decarboxylase, the instigating autoanti infused.
gen in T1DM, are ongoing. The applicability
of oral tolerance depends on identification of Epigenetic Enhancer Regulation
disease‐specific initiating autoantigens; thus, SNPs within enhancers, especially SEs, play
it would be currently relevant in AILDs only dominant roles in autoimmune pathogenesis
for PBC and type 2 AIH. (see section Epigenetics) [16]. The theoret
ical fear that targeting epigenetic regulatory
proteins might cause severe toxicity has
Strategies to Treat Established
Autoimmune Diseases
not been observed with first‐generation
inhibitors. Among SE proteins regulating
At the time of diagnosis, the initiating events genes associated with activation, stress and
of most autoimmune diseases likely occurred differentiation, the bromodomain and extra‐
months to years earlier and mechanisms of terminal (BET) family has emerged as a
perpetuation are well established. This poses prime target for inhibition. BET inhibition
challenges for clinical management and decreased macrophage expression of
focuses attention on therapeutic strategies to inflammatory genes induced by lipopolysac
control inflammation, modify symptoms and charide by preferentially inhibiting de novo
signs, and retard progression. Progress in expression of SE genes. BET inhibitors also
understanding the pathogenesis of autoim blocked de novo SE gene expression in endo
mune diseases provides rationales for addi thelial cells, markedly reduced CD4 Th cells
tional conventional and novel therapies differentiation to polarized subsets, and pre
(Table 2.3). Two therapeutic approaches vented effector cytokine production by
warrant additional comment. already polarized Th cells. BET inhibitors
also impacted B cells by inhibiting prolifera
Inducible T Regulatory T Cells tion and ability to switch the isotype of their
Studies of two strategies to produce iTregs antibodies. More importantly, BET inhibi
specific for autoantigens are in progress tors have reduced inflammation and pro
(Table 2.3) [1]. The first strategy is to increase tected from disease in animal models of
the proliferation of existing iTreg popula T1DM, MS, RA, and psoriasis. These data
tions by infusion of low doses of IL‐2, which indicate that BET‐specific, and more gener
is being studied in hematopoietic stem cell ally SE‐related, mechanisms are promising
transplantation, graft‐versus‐host disease, therapeutic targets.
Table 2.3 Current and future therapeutic approaches to control autoimmune diseases.
Increase regulatory Infuse low‐dose IL‐2 Expand existing autoantigen‐specific iTregs in vivo POC established. Clinical trials ongoing
control of Infuse iTregs generated ex Ex vivo generation of disease and autoantigen‐ POC of iTreg generation established. No
autoimmune vivo to control Th cell and specific iTregs infusion trials to date. Viability, function
responses cytokine responses and distribution after infusion unknown
Bromodomain and Inhibit disease‐specific epigenetic enhancers, POC established. Clinical trials ongoing
extra‐terminal (BET) super enhancers and eRNA production
family of proteins
Mesenchymal stem cells Inhibit innate immune cells, effector T cells. POC established. Clinical trials ongoing
Induction of iTregs. Reduction of
proinflammatory TNF‐α
Decrease numbers Immunosuppressive Reduce functional mass of activated T cells using SOC in multiple autoimmune diseases.
and/or functions of drugs: CNI, mTOR, immunosuppressive drugs Active research to refine target specificity
autoimmune antiproliferative agents and reduce toxicities. Combinations of
effector cells and drugs working at different sites of action
pathogenic using subtoxic doses preferred
autoantibodies Anti‐CD20 B‐cell depletion SOC and off‐label uses
Anti‐BAFF Initial B‐cell depletion followed by mobilization of SOC in SLE. Broad future potential,
memory B cells. Concurrent inhibition of BAFF especially in combination regimens
signaling in T cells
Anti‐BAFF, followed by Depletion Sequential use to eliminate mobilized
anti‐CD20 of memory B cells mobilized by anti‐BAFF memory B cells to increase efficacy
Anti‐CD40 Block CD40–CD40L (CD154) costimulation of T POC. Clinical trial initiated in
cells and B cells transplantation
Block IgG recycling and First in class antibody fragment to block FcRn POC to reduce pathogenic autoantibodies
increase IgG clearance (efgartigimod) and formation of immunoglobulin–
autoantigen complexes
Prevent egress of activated Block sphingosine‐1‐phosphate receptors SOC in MS, new agents in development
T cells from lymph nodes
MDSCs Inhibit activation and proliferation of autoreactive POC in preclinical models. Plans for
T cells clinical trials in RA
(Continued )
Table 2.3 (Continued)
Decrease Anti‐TNF‐α or TNF‐α Decrease TNF‐α mediated tissue injury and SOC in multiple autoimmune diseases
proinflammatory receptor proinflammatory signaling
cytokines Anti‐IL‐6 or anti‐IL‐6R Decrease pathological consequences of SOC in RA, clinical trials ongoing
proinflammatory IL‐6 signaling in innate and
adaptive immune response
Anti‐IL‐12 Decrease pathological consequences of Monoclonal antibody against p40 subunit.
proinflammatory IL‐12 signaling in innate and SOC in psoriasis and Crohn’s disease.
adaptive immune response Also blocks IL‐23 signaling
Anti‐IL‐23 Decrease pathological consequences of Monoclonal antibody against p40 subunit.
proinflammatory IL‐23 SOC in psoriasis and Crohn’s disease.
Blocks IL‐23 stimulation of Th17 cells
Anti‐IL‐17 Decrease pathological consequences of Th17 SOC for psoriasis and psoriatic arthritis.
production of IL‐17 Clinical trials designed
Anti‐IL‐21 Decrease multiple pathological consequences of Clinical trials in RA, T1DM, Crohn’s
IL‐21 in innate and adaptive immune responses disease
Inhibition of IL‐2 Decrease proliferation of activated CD4 and CD8 SOC CNI and mTOR inhibitors. POC
proinflammatory T cells JAK3 inhibition
cytokine signaling IL‐6 Decrease proinflammatory IL‐6R‐mediated POC and SOC indications for JAK1/2
signaling inhibition
IL‐12/IL‐23 Decrease proinflammatory IL‐12 and IL‐23 POC and SOC indications for JAK2
signaling that polarize to a Th1 response, increase inhibition
secretion of IFN‐γ and TNF‐α, increase
cytotoxicity of NK cells and CD8 CTLs and drive
differentiation of pathogenic Th17 cells
IFN‐α/IFN‐β Decrease gene expression induced by type 1 IFNs POC and SOC indications for JAK1
inhibition
IFN‐γ Decrease proinflammatory actions of IFN‐γ POC and SOC indications for JAK1/JAK2
produced by NK, NKT, CD4 and CD8 T cells inhibition
Approaches Goal Status
AIH, autoimmune hepatitis; BAFF, B‐cell‐activating factor; CNI, calcineurin inhibitor; ERCP, endoscopic retrograde cholangiopancreatography; IL, interleukin; iTregs,
inducible T regulatory cells; JAK, Janus kinase; MDSCs, myeloid‐derived suppressor cells; MS, multiple sclerosis; mTOR, mechanistic target of rapamycin; PIF, pre‐
implantation factor; POC, proof of concept; PSC, primary sclerosing cholangitis; RA, rheumatoid arthritis; rHuIL‐10, recombinant human IL‐10; SLE, systemic lupus
erythematosus; SOC, standard of care; T1DM, type 1 diabetes mellitus; Th, T helper cell; TNF, tumor necrosis factor.
44 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
References
1 Devarapu, S.K., Lorenz, G., Kulkarni, O.P. 10 Theofilopoulos, A.N., Kono, D.H., and
et al. (2017). Cellular and molecular Baccala, R. (2017). The multiple pathways to
mechanisms of autoimmunity and lupus autoimmunity. Nat. Immunol 18 (7): 716–724.
nephritis. Int. Rev. Cell Mol. Biol 332: 11 Dominguez‐Villar, M. and Hafler, D.A.
43–154. (2018). Regulatory T cells in autoimmune
2 Gambino, C.M., Aiello, A., Accardi, G. et al. disease. Nat. Immunol 19 (7): 665–673.
(2018). Autoimmune diseases and 8.1 12 Taher, T.E., Bystrom, J., Ong, V.H. et al.
ancestral haplotype: an update. HLA 92 (3): (2017). Intracellular B lymphocyte
137–143. signalling and the regulation of humoral
3 Toubi, E. and Vadasz, Z. (2019). Innate immunity and autoimmunity. Clin. Rev.
immune‐responses and their role in driving Allergy Immunol 53 (2): 237–264.
autoimmunity. Autoimmun. Rev 18 (3): 13 Phillips, B.E., Garciafigueroa, Y., Engman,
306–311. C. et al. (2019). Tolerogenic dendritic cells
4 Caso, F., Costa, L., Nucera, V. et al. (2018). and T‐regulatory cells at the clinical trials
From autoinflammation to autoimmunity: crossroad for the treatment of autoimmune
old and recent findings. Clin. Rheumatol 37 disease: emphasis on type 1 diabetes
(9): 2305–2321. therapy. Front. Immunol. 10: 148.
5 Shen, H.H., Yang, Y.X., Meng, X. et al. 14 Lee, G.R. (2018). The balance of Th17
(2018). NLRP3: a promising therapeutic versus Treg cells in autoimmunity. Int. J.
target for autoimmune diseases. Mol. Sci 19 (3): 730.
Autoimmun. Rev 17 (7): 694–702. 15 Inshaw, J.R.J., Cutler, A.J., Burren, O.S.
6 Grant, C.R. and Liberal, R. (2017). Liver et al. (2018). Approaches and advances in
immunology: how to reconcile tolerance the genetic causes of autoimmune disease
with autoimmunity. Clin. Res. Hepatol. and their implications. Nat. Immunol 19
Gastroenterol 41 (1): 6–16. (7): 674–684.
7 Chiba, A., Murayama, G., and Miyake, S. 16 Tough, D.F. and Prinjha, R.K. (2017).
(2018). Mucosal‐associated invariant T cells Immune disease‐associated variants in
in autoimmune diseases. Front. Immunol. 9: gene enhancers point to BET epigenetic
1333. mechanisms for therapeutic intervention.
8 Cheng, M. and Anderson, M.S. (2018). Epigenomics 9 (4): 573–584.
Thymic tolerance as a key brake on 17 Di Marco, M., Ramassone, A., Pagotto, S.
autoimmunity. Nat. Immunol 19 (7): et al. (2018). MicroRNAs in autoimmunity
659–664. and hematological malignancies. Int. J. Mol.
9 Ray, A. and Dittel, B.N. (2017). Mechanisms Sci 19 (10): 3139.
of regulatory B cell function in autoimmune 18 Moulton, V.R. (2018). Sex hormones in
and inflammatory diseases beyond IL‐10. acquired immunity and autoimmune
J. Clin. Med 6 (1): 12. disease. Front. Immunol. 9: 2279.
Chapter 2 Concepts of Autoimmunity Relevant to Autoimmune Liver Diseases 45
19 Dankers, W., Colin, E.M., van Hamburg, how unique costimulatory molecules
J.P., and Lubberts, E. (2016). Vitamin D in contribute to diabetogenesis. Front.
autoimmunity: molecular mechanisms Endocrinol 8: 208.
and therapeutic potential. Front. Immunol. 25 Molodtsov, A. and Turk, M.J. (2018). Tissue
7: 697. resident CD8 memory T cell responses in
20 De Luca, F. and Shoenfeld, Y. (2019). The cancer and autoimmunity. Front. Immunol.
microbiome in autoimmune diseases. Clin. 9: 2810.
Exp. Immunol 195 (1): 74–85. 26 Chen, J., Caspi, R.R., and Chong, W.P.
21 Rojas, M., Restrepo‐Jimenez, P., Monsalve, (2018). IL‐20 receptor cytokines in
D.M. et al. (2018). Molecular mimicry and autoimmune diseases. J. Leukoc. Biol 104
autoimmunity. J. Autoimmun 95: 100–123. (5): 953–959.
22 White, K.D., Chung, W.H., Hung, S.I. et al. 27 Pipi, E., Nayar, S., Gardner, D.H. et al.
(2015). Evolving models of the (2018). Tertiary lymphoid structures:
immunopathogenesis of T cell‐mediated autoimmunity goes local. Front. Immunol.
drug allergy: the role of host, pathogens, 9: 1952.
and drug response. J. Allergy Clin. Immunol 28 Wang, M., Wu, L., Weng, R. et al. (2017).
136 (2): 219–234. quiz 235. Therapeutic potential of helminths in
23 Abdolmaleki, F., Farahani, N., Gheibi autoimmune diseases: helminth‐derived
Hayat, S.M. et al. (2018). The role of immune‐regulators and immune balance.
efferocytosis in autoimmune diseases. Parasitol. Res 116 (8): 2065–2074.
Front. Immunol. 9: 1645. 29 Commins, S.P. (2015). Mechanisms of oral
24 Wagner, D.H. Jr. (2017). Overlooked tolerance. Pediatr. Clin. North Am 62 (6):
mechanisms in type 1 diabetes etiology: 1523–1529.
47
Abstract
There have been major advances in understanding the human leukocyte antigen (HLA) associations
with autoimmune liver disease using serotyping, phenotyping and sequencing based HLA association
studies, and more from imputation of classical HLA alleles from genomewide association studies
(GWAS) or iChip genotype data. This chapter summarizes genetic associations that have been identi-
fied largely as a result of GWAS and related study designs. GWAS and related study designs have sub-
stantially improved knowledge of the genetic architecture of primary biliary cholangitis (PBC) and
primary sclerosing cholangitis (PSC), especially outside the HLA region. Numerous genomewide
significant non-HLA risk loci for PBC or PSC have been identified. Although genetic studies of PBC or
PSC have successfully identified numerous risk loci, it is estimated that less than 20% of PBC heritability
and less than 10% of variance of PSC liability has been explained, the so-called ‘missing heritability’.
Key Points
●● There have been major advances in under- ●● Many of these HLA associations are
standing the human leukocyte antigen shared with other autoimmune diseases.
(HLA) associations with autoimmune liver ●● In addition to HLA‐associated genes,
disease using serotyping, phenotyping and non‐HLA gene loci have been identi-
sequencing‐based HLA association studies, fied in association with presence of
and more recently from imputation of disease.
classical HLA alleles from genome‐wide ●● Only a relatively small proportion of
association studies or iChip genotype data. overall genetic risk of developing autoim-
●● For primary biliary cholangitis (PBC), pri- mune liver diseases is explained by the
mary sclerosing cholangitis (PSC) and new loci determined so far.
autoimmune hepatitis (AIH), both risk ●● Such loci, particularly in the case of PBC,
and protective HLA associations have do however support an autoimmune eti-
been identified. ology for disease.
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
48 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
Table 3.2 Genome‐wide association studies (GWAS) or related study designs of primary biliary
cholangitis (PBC).
Hirschfield et al. (2009) [9] GWAS European 505 1507 526 1206
a
Liu X et al. (2010) [10] GWMA European 958 2452 NA NA
Mells et al. (2011) [11] GWAS European 1840 5163 620 2514
Nakamura et al. (2012) [12] GWAS Japanese 487 476 787 615
Liu JZ et al. (2012) [13] iChip European 2861 8514 NA NA
Juran et al. (2012) [14] iChip European 2426 5731 NA NA
Cordell et al. (2015)b [15] GWMA European 2764 10 475 3716 4261
Paziewska et al. (2017) [16] GWAS Polish 443 934 NA NA
Kawashima et al. (2017) [17] GWAS Japanese 1381 1505 512 6512
Qiu et al. (2017) [18] GWAS Chinese 1122 4036 907 2127
a
GWMA of discovery panel from the GWAS by Hirschfield et al. [9] and a GWAS dataset consisting of 453 Italian
cases and 945 Italian controls.
b
GWMA of discovery panels from the GWAS by Hirschfield et al. [9], Liu et al. [10] and Mells et al. [11].
GWMA, genome‐wide meta‐analysis; iChip, Illumina Immunoarray association study.
Table 3.3 Genome‐wide association studies (GWAS) or related study designs of primary sclerosing
cholangitis (PSC).
Discovery Validation
Karlsen et al. (2010) [19] GWAS Norwegian 285 298 766 2935
Melum et al. (2011) [20] GWAS European 715 2962 1025 2174
Ellinghaus et al. (2013)a [21] GWAS European 392 2977 1012 11 659
Liu et al. (2013) [22] iChip European 3789 25 079 NA NA
Ji et al. (2016) [23] GWAS European 2871 12 019 1925 7936
Ellinghaus et al. (2016)b [24] iChip European 3408 34 213 NA NA
Paziewska et al. (2017) [16] GWAS Polish 120 934 NA NA
a
In this study, Ellinghaus et al. [50] undertook combined analyses of GWAS for PSC and ulcerative colitis (UC)
comprising 392 PSC cases, 987 UC cases and 2977 controls and followed up top association signals in 1012 PSC
cases, 4444 UC cases and 11 659 controls.
b
In this study, Ellinghaus et al. [42] undertook cross‐phenotype subset‐based meta‐analysis of Illumina
Immunoarray (iChip) genotype data from 52 262 cases with ankylosing spondylitis (8726), Crohn’s disease (19 085),
psoriasis (6530), PSC (3408) or UC (14 513), and 34 213 healthy controls.
iChip, Illumina Immunoarray association study.
50 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
Table 3.4 Genome‐wide association studies (GWAS) or related study designs of type 1 autoimmune hepatitis.
Discovery Validation
summarizes genetic associations that have from imputation of classical HLA alleles
been identified largely as a result of GWAS from GWAS or iChip genotype data [29].
and related study designs in the last decade. Several HLA associations with PBC, PSC or
AIH have been identified (Table 3.5).
Consistent with other autoimmune condi-
tions, HLA effects in PBC, PSC or AIH are
HLA Associations large, with ORs up to approximately 3.
in Autoimmune Liver In PBC, the most robust associations in
Disease populations of European origin are with the
risk haplotypes DRB1*08:01‐DQA1*04:01‐
The human leukocyte antigen (HLA) com- DQB1*04:02 and DRB1*04:04‐DQB1*03:02,
plex is located on the short arm of and the protective haplotypes DRB1*11:01‐
chromosome 6 and harbors genes encoding DQA1*05:01‐DQB1*03:01 and DRB1*15:01‐
the HLA molecules that are essential for DQA1*01:02‐DQB1*0602. In the study by
antigen presentation. It is characterized by Invernizzi et al. [30], conditional analysis
clusters of highly polymorphic genes that suggested that these haplotype associations
are related by sequence homology and/or were most likely driven by their respective
function: class I genes (HLA‐A, ‐B and ‐C) DRB1 alleles, i.e. the risk alleles DRB1*08 and
and class II genes (HLA‐DP, ‐DQ and ‐D2) DRB1*14 and the protective allele DRB1*11.
encode proteins involved in antigen presen- In PSC, key risk haplotypes include Ancestral
tation, while class III genes encode a large Haplotype 8.1 (AH8.1), which carries the B*08
number of other immune proteins, such as and DRB1*03:01 alleles, and the DRB3*01:01‐
tumor necrosis factor (TNF)‐α [27]. The DRB1*13:01‐DQA1*01:03‐DQB1*06:03 haplo-
HLA complex is characterized by strong type [31–33]. In admixed African‐American
LD extending across large fractions of the populations, LD between PSC associated
region, giving rise to extended HLA haplo- HLA‐B and DRB1 alleles is low and association
types [28]. This means that when HLA is reported for HLA‐B*08 but not DRB1*03,
associations are detected, it may be difficult suggesting that HLA‐B*08 accounts for the
to determine which allele within an AH8.1 association [34]. More generally, it high-
extended haplotype accounts for the lights that class I effects are important in PSC.
association. Patterns of LD across the HLA However, class II effects are also important,
region are different in different ethnicities, shown for example by association with the
so a potential solution to this problem is to protective haplotypes DRB1*04‐DQB1*03:02
study HLA associations across different and DRB1*07:01‐DQB1*03:03 [33, 35]. Strong
ethnic groups. LD across the class II region makes it difficult
Knowledge of HLA associations with auto- to identify which allele drives association with
immune liver disease comes from serotyp- these haplotypes, but circumstantial evidence
ing, phenotyping and sequencing‐based suggests that DRB1 alleles may account for
HLA association studies, and more recently protective effects [36].
Table 3.5 HLA locus associations in primary sclerosing cholangitis (PSC), primary biliary cirrhosis (PBC)
and autoimmune hepatitis (AIH).
PSC R B*08
B*07
C*06
A*01‐C*07‐B*08‐DRB3*01:01‐DRB1*03:01‐DQA1*05:01‐DQB1*02:01
DRB1*13:01‐DQA1*01:03‐DQB1*06:03
DRB1*15:01‐DQB1*06:02
rs116212904/A at NOTCH4
P DRB1*04‐DQB1*03:01
DRB1*04‐DQB1*03:02
DRB3*03:01‐DRB1*13:02‐DQB1*06:04
DRB1*07:01‐DQB1*03:03
DRB1*13:XX‐DQB1*06
HLA‐DPB1 amino acid substitutions that small (OR ~1.5). Non‐HLA risk loci for AIH
confer PBC susceptibility or protection. have not been identified at genome‐wide
level of significance, but GWAS of AIH have
so far been comparatively underpowered.
Non‐HLA Associations Although genetic studies of PBC or PSC
in Autoimmune Liver Disease have successfully identified numerous risk
loci, it is estimated that less than 20% of
GWAS and related study designs have sub- PBC heritability and less than 10% of vari-
stantially improved knowledge of the ge- ance of PSC liability has been explained, the
netic architecture of PBC and PSC, so‐called “missing heritability” [13, 45]. It is
especially outside the HLA region. plausible that better‐powered studies of
Numerous genome‐wide significant non‐ PBC or PSC would identify additional risk
HLA risk loci for PBC or PSC have been loci for these conditions, but it may be
identified (Tables 3.6 and 3.7). Consistent impossible to achieve the sample sizes
with other complex disorders, the effects of required to identify additional risk loci with
non‐HLA risk loci in these conditions are ever‐smaller effects [46–49].
54 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
1p36.32 Liu et al. (2013) [22] rs3748816 7 × 10−12 1.21 MMEL1, TNFRSF14
2q13 Melum et al. (2011) [20] rs6720394 4 × 10−8 1.29 BCL2L11
−20
2q33.2 Liu et al. (2013) [22] rs7426056 2 × 10 1.30 CD28, CTLA4
2q36.3 Ellinghaus et al. (2016) [24] rs7556897 5 × 10−9 1.18 CCL20
−9
2q37.3 Ellinghaus et al. (2013) [21] rs4676410 2 × 10 1.38 GPR35
3p21.31 Melum et al. (2011) [20] rs3197999 1 × 10−16 1.39 MST1
−15
3p13 Ji et al. (2017) [23] rs80060485 3 × 10 1.44 FOXP1
4q24 Ellinghaus et al. (2016) [24] rs17032705 4 × 10−10 1.18 NFKB1
−13
4q27 Liu et al. (2013) [22] rs13140464 9 × 10 1.30 IL2, IL21
6q15 Liu et al. (2013) [22] rs56258221 8 × 10−12 1.23 BACH2
−17
10p15.1 Liu et al. (2013) [22] rs4147359 8 × 10 1.24 IL2RA
11q13.1 Ji et al. (2017) [23] rs663743 2 × 10−13 1.20 CCDC88B
−9
11q23.1 Liu et al. (2013) [22] rs7937682 3 × 10 1.17 SIK2
12q13 Liu et al. (2013) [22] rs11168249 5 × 10−9 1.15 HDAC7
−9
12q23.3 Ellinghaus et al. (2016) [24] rs12369214 1 × 10 1.18 RFX4, RIC8B
12q24.12 Liu et al. (2013) [22] rs3184504 6 × 10−11 1.18 SH2B3, ATXN2
−11
16p13.13 Ellinghaus et al. (2016) [24] rs11649613 1 × 10 1.19 CLEC16A, SOCS1
18q21.2 Ellinghaus et al. (2013) [21] rs1452787 3 × 10−8 1.33 TCF4
−8
18q22.2 Liu et al. (2013) [22] rs1788097 3 × 10 1.15 CD226
19q13.32 Liu et al. (2013) [22] rs60652743 6 × 10−10 1.25 PRKD2, STRN4
−17
21q22.2 Liu et al. (2013) [22] rs2836883 3 × 10 1.28 PSMG1
21q22.3 Ji et al. (2017) [23] rs1893592 2 × 10−12 1.22 UBASH3A
−8
22q11.23 Ji et al. (2017) [23] rs145832854 3 × 10 1.87 SGSM1
Candidate genes for PBC and PSC are listed AIH are briefly described in Boxes 3.1–3.3,
in Tables 3.6 and 3.7; however, it should be respectively. The reader is reminded that the
noted that GWAS and related studies identify potential role of these genes in the pathogen-
disease‐associated SNPs, but do not identify esis of autoimmune liver disease is obvious
causal variants or causal genes (Figure 3.1). but also speculative.
Causal variants and genes may be prioritized The mechanism by which causal variants
through functional annotation of risk loci but lead to disease is unknown and may be differ-
laboratory experiments linking genotype and ent for each risk locus. It has been observed,
molecular phenotype are required to confirm however, that many candidate causal variants
(and explain) the contribution of the variant for complex disorders are regulatory vari-
and gene to disease pathogenesis [50]. ants, meaning that they overlap regulatory
Nevertheless, based on functional annotation, elements (e.g. promoter regions) active in
risk loci for PBC, PSC and AIH appear to be relevant tissues. In the genome‐wide meta‐
enriched for gene products involved in innate analysis of PBC by Cordell et al. [15], for
or acquired immune responses, consistent example, up to one‐third of candidate causal
with an autoimmune component to patho- variants were regulatory, overlapping with
genesis. Candidate genes for PBC, PSC and regulatory elements active in cells of the
Chapter 3 Genetics and Risk of Autoimmune Liver Diseases 55
Plotted SNPs
25 r2 100
chr16:11188930
0.8
15 60
10 40
5 20
0 0
Figure 3.1 LocusZoom plot of the PBC risk locus 16p13.13, illustrating that, while the index variant may
strongly suggest a causal gene (in this case, CLEC16A), this cannot be assumed: at any risk locus, the index
variant is the variant most strongly associated with disease, while the causal variant is an unknown variant
with functional effects that influence disease expression. Association is detected with the index variant owing
to correlation with the causal variant. The index variant is correlated to numerous other variants, however,
making it difficult to know which one is causal. Risk loci encompass the index variant and neighboring variants
correlated to it. They may harbor several genes (in this case, DEXI, CLEC16A, SOCS1, TNP2, PRM3, and so forth),
any of which might be the gene involved in causing disease. Bioinformatics approaches are helpful for
prioritizing candidate genes but functional studies are ultimately required to confirm causality and elucidate
mechanism.
immune system. This suggests that many clinically associated autoimmune disorders –
variants contribute to disease by influencing ulcerative colitis (UC), Crohn disease (CD),
the regulation of gene expression. Consistent type 1 diabetes mellitus, celiac disease, psori-
with this, many candidate causal variants are asis (PS), rheumatoid arthritis and sarcoid-
correlated with gene expression (expression osis – suggesting close similarity in the
quantitative trait loci or eQTLs) or methyla- genetic architecture of PSC and each of these
tion (methylation quantitative trait loci or conditions [45]. This was further explored in
mQTLs). the study by Ellinghaus et al. [24], who under-
A striking observation from GWAS is the took a cross‐phenotype, subset‐based meta‐
extent to which risk loci are shared by mul- analysis of iChip genotype data from 52 262
tiple autoimmune conditions. Pervasive cases with ankylosing spondylitis (8726), CD
sharing of genetic effects has been formally (19 085), PS (6530), PSC (3408) or UC
demonstrated in PSC. In the iChip study by (14 413), and 34 213 healthy controls. They
Liu et al. [22], pleiotropy analysis revealed identified 244 independent multidisease sig-
strong positive correlation between associa- nals and showed that comorbidities among
tions with PSC and associations with seven the five disorders were best explained by
56 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
Interleukin‐12 subunit alpha (IL12A, 3q25.33) ment and cell motility protein family. It inter-
is the 35‐kDa subunit of IL‐12, a cytokine that acts with dedicator of cytokinesis 1 (DOCK1)
functions as a growth factor for activated T and protein to promote phagocytosis of apoptotic
NK cells, enhances the lytic activity of NK cells cells and cell migration.
and lymphokine‐activated killer (LAK) cells,
Interferon regulatory factor 5 (IRF5, 7q32.1)
and stimulates production of IFN‐γ. IL‐12 sig-
is a transcription factor involved in the
naling is mediated by STAT4.
induction of interferons IFNA and INFB and
Nuclear factor kappa B p105 subunit inflammatory cytokines following viral infec-
(NFKB1, 4q24) is the 105‐kDa subunit of NFκB, tion. It is activated by Toll‐like receptor (TLR)7
a pleiotropic transcription factor expressed in or TLR8 signaling.
most cell types. NFκB functions in the tran-
Tumor necrosis factor ligand superfamily
scriptional regulation of diverse biological
member 15 (TNFSF15, 9q32) is the ligand for
processes, including inflammation, immunity,
TNFRSF25 and TNFRSF6B. It mediates
differentiation, cell growth, tumorigenesis,
activation of NFκB and mitogen‐activated pro-
and apoptosis.
tein kinases (MAPK), promotes activation of
Interleukin‐21 (IL21, 4q27) is a cytokine that caspases and apoptosis, and may inhibit endo-
functions in both innate and adaptive immune thelial cell proliferation.
responses by inducing the differentiation, pro-
Coiled‐coil domain containing 88B
liferation and activity of multiple target cells
(CCDC88B, 11q13.1) is a member of the hook‐
including macrophages, NK cells, B cells, and
related protein family. It is required for several
cytotoxic T cells.
CD4+ and CD8+ T‐cell functions, including
Interleukin‐7 receptor subunit alpha (IL7R, expression of cell surface markers of activation,
5p13.2) is a subunit of the receptor for IL‐7, a proliferation and cytokine production in
hematopoietic growth factor that is important response to specific or non‐specific stimula-
for B‐ and T‐cell development. IL7–IL7R is tion. It enhances NK cell cytotoxicity and sup-
necessary for V(D)J rearrangement of the TCRβ presses endoplasmic reticulum stress‐induced
(TCRB) during early T‐cell development. It may apoptosis.
also be involved in regulation of intestinal
POU domain class 2‐associating factor 1
mucosal lymphocytes.
(POU2AF1, 11q23.1) is a transcriptional coacti-
Interleukin‐12 subunit beta (IL12B) is the 40‐ vator that specifically associates with either
kDa subunit of the cytokines IL‐12 (see above) OCT1 or OCT2, which is essential for the
and IL‐23, a heterodimeric cytokine which response of B cells to antigens and required for
binds to a receptor complex composed of the formation of germinal centers.
IL12RB1 and IL23R, activating a Jak–Stat sig-
C‐X‐C chemokine receptor type 5 (CXCR5,
naling cascade. IL‐23 signaling stimulates
11q23.3) is a cytokine receptor that binds to
memory T cells and promotes production of
B‐lymphocyte chemoattractant (BLC). It is
proinflammatory cytokines.
involved in B‐cell migration into B‐cell follicles
Tumor necrosis factor alpha‐induced pro- of spleen and Peyer’s patches.
tein 3 (TNFAIP3, 6q23.3) is a ubiquitin‐editing
Tumor necrosis factor receptor superfamily
enzyme involved in cytokine‐mediated
member 1A (TNFRSF1A, 12p13.31) is the
immune and inflammatory responses, which
receptor for TNF‐α and lymphotoxin‐α. Binding
inhibits NFκB activation and TNF‐mediated
of membrane‐bound TNF‐α to membrane‐
apoptosis.
bound TNFRSF1A induces receptor trimeriza-
Engulfment and cell motility protein 1 tion and activation, which plays a role in cell
(ELMO1, 7p14.1) is a member of the engulf- survival, apoptosis, and inflammation.
58 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
Tumor necrosis factor receptor superfamily protein kinase that is involved in various cel-
member 3 (LTBR, 12p13.31) is the receptor for lular processes, including regulation of the B‐
the heterotrimeric lymphotoxin containing cell receptor (BCR) signalosome and oxidative
LTA and LTB, and for TNFS14/LIGHT. It pro- stress‐induced apoptosis.
motes apoptosis and may play a role in the
Interleukin‐21 receptor (IL21R, 16p12.1) is
development of lymphoid organs.
the receptor for IL‐21. Ligand binding of IL21R
SH2B adapter protein 3 (SH2B3, 12q24.12) is leads to activation of multiple downstream
a member of the SH2B adaptor family of pro- signaling molecules, including JAK1, JAK3,
teins, which are involved in a range of signaling STAT1, and STAT3. Signaling via IL21R is impor-
activities by growth factor and cytokine recep- tant for the proliferation and differentiation of
tors. SH2B3 links TCR activation signal to phos- T cells, B cells, and NK cells.
pholipase Cγ1, GRB2 and phosphatidylinositol
Coiled‐coil domain‐containing protein 113
3‐kinase. It is a key negative regulator of cyto-
(CCDC113, 16q21) is a component of centrio-
kine signaling and plays a critical role in
lar satellites contributing to primary cilium
hematopoiesis.
formation.
Leukemia‐associated protein 1 (DLEU1,
Interferon regulatory factor 8 (IRF8, 16q24.1)
13q14.2) is an RNA gene that may act as a
is a transcription factor of the interferon
tumor suppressor.
regulatory factor (IRF) family that binds to the
DNA repair protein RAD51 homolog 2 IFN‐stimulated response element (ISRE) and
(RAD51B, 14q24.1) is involved in the homolo- regulates expression of genes stimulated by
gous recombination repair pathway of double‐ type I IFNs (i.e. IFN‐α and IFN‐β). IRF8 has a
stranded DNA breaks arising during DNA negative regulatory role in cells of the immune
replication or induced by DNA‐damaging system, and is involved in CD8+ DC
agents. differentiation.
Tumor necrosis factor alpha‐induced pro- Zinc finger protein Aiolos (IKZF3, 17q21.1) is
tein 2 (TNFAIP2, 14q32.32) can be induced by a transcription factor involved in lymphocyte
TNF‐α and may be a mediator of inflammation differentiation. It plays an essential role in reg-
and angiogenesis. ulation of B‐cell differentiation, proliferation
and maturation to an effector state. It is
Interleukin‐16 (IL16, 15q25.1) is a proinflam-
involved in regulating BCL2 expression and
matory cytokine that signals via CD4. It is che-
controlling apoptosis in T cells in an IL2‐
motactic for CD4+ T lymphocytes, monocytes,
dependent manner.
and eosinophils; primes CD4+ T cells for IL‐2
and IL‐15 responsiveness; upregulates IL‐2 Microtubule‐associated protein tau (MAPT,
receptor and HLA‐DR4 expression; and inhibits 17q21.31) promotes microtubule assembly
TCR/CD3‐dependent activation. and stability, and might be involved in the
establishment and maintenance of neuronal
C‐type lectin domain family 16 member A
polarity.
(CLEC16A, 16p13.13) is involved in mitophagy
through the upstream regulation of the AT‐rich interactive domain‐containing pro-
RNF41/NRDP1‐PRKN pathway, mitophagy tein 3A (ARID3A, 19p13.3) is a transcription factor
being a selective form of autophagy necessary which may be involved in B‐cell differentiation.
for mitochondrial quality control.
Non‐receptor tyrosine‐protein kinase (TYK2,
Protein kinase C beta type (PRKCB, 16p12.2) 19p13.2) is a member of the Janus kinase (JAK)
is a calcium‐activated, phospholipid‐ and diac- family of protein. TYK2 associates with the
ylglycerol (DAG)‐dependent serine/threonine cytoplasmic domain of type I and type II cytokine
Chapter 3 Genetics and Risk of Autoimmune Liver Diseases 59
receptors and propagates cytokine signals by duce large amounts of interferon and block
phosphorylating receptor subunits. It is a com- viral replication. SPIB may be required for BCR
ponent of both the type I and type III inter- signaling, which is necessary for normal B‐cell
feron signaling pathways. development and antigenic stimulation.
Transcription factor Spi‐B (SPIB, 19q13.33) is Synaptogyrin‐1 (SYNGR1, 22q13.1) is an inte-
a transcriptional activator which binds to the gral membrane protein associated with pre-
PU‐box, a purine‐rich DNA sequence (5′‐ synaptic vesicles in neuronal cells, which may
GAGGAA‐3′) that acts as a lymphoid‐specific play a role in synaptic‐like microvesicle
enhancer. SPIB promotes development of plas- formation and/or maturation, and may play a
macytoid DCs, which have the capacity to pro- role in regulated exocytosis.
Box 3.3 Candidate Genes for Autoimmune leiotropy analysis has not been undertaken
p
Hepatitis in PBC or AIH, but it is anticipated that
findings would be similar. Pervasive sharing
SH2B adaptor protein 3 (SH2B3, 12q24.12). of risk variants of genes across a variety of
See Box 3.1. autoimmune diseases is relevant because
cross‐phenotype analyses might therefore
Caspase recruitment domain family
be helpful in identifying molecular targets
member 10 (CARD10, 22q13.1) belongs to
for therapies with widespread applications
the membrane‐associated guanylate kinase
(although this cannot be assumed).
(MAGUK) family and activates NFκB via
Conversely, pervasive sharing of genetic
BCL10 and IKK.
effects across multiple autoimmune condi-
tions emphasizes that GWAS of PBC, PSC
biological pleiotropy rather than heteroge- and AIH have not identified candidates that
neity, suggesting, for example, that the would explain the tissue specificity of these
strong comorbidity between PSC and diseases. Paradoxically, genetic studies of
inflammatory bowel disease (IBD) may rep- autoimmune liver disease highlight the likely
resent a unique disease, genetically distinct importance of environmental factors in
from classical IBD phenotypes. Formal shaping disease.
Chapter 3 Genetics and Risk of Autoimmune Liver Diseases 61
Consistent with this, some of the strongest inflammatory risk loci). The pleiotropy of
associations in GWAS of immune‐mediated the GWAS outcomes is poorly understood
conditions identify risk loci directly involved and varies between different geographic and
in gene–environment interactions (e.g. HLA‐ ethnic groups. Importantly, only a fraction
DQB1 in celiac disease and NOD2 in Crohn of the heritability of liability is explained by
disease) [51, 52]. It is conceivable that common loci determined to date, partly because
causal variants for complex disorders have study panels have been relatively modest
escaped neutral processes (e.g. genetic drift) given the low prevalence of autoimmune
because of selective pressure exerted by envi- liver diseases. However, the latter may also
ronmental influences. GWAS findings may suggest that interacting genetic or non‐
therefore represent “hotspots” where the genetic factors (environmental factors in
gene–environmental homeostasis of an particular) are involved in causing each
individual is disturbed. Interpretation of ge- specific clinical condition on the background
netic findings may therefore require better partly determined by GWAS. Following
knowledge of the environmental factors GWAS, there is therefore a strong need to
involved in disease, justifying ongoing efforts search for these factors, along with further
to identify these factors. genetic characterization of risk loci
determined (e.g. by means of sequencing to
determine rare pathogenic variants).
Conclusion Ultimately, the accumulating knowledge
from these studies needs to be explored in
The genetic architecture of PBC, PSC and the context of existing and novel in vitro and
AIH is similar to most autoimmune diseases in vivo model systems with the aim of gener-
(i.e. a strong HLA association accompa- ating a basis for novel therapeutics for
nied by associated groups of general autoimmune liver diseases.
62 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
References
1 Boonstra, K., Beuers, U., and Ponsioen, C.Y. 10 Liu, X., Invernizzi, P., Lu, Y. et al. (2010).
(2012). Epidemiology of primary sclerosing Genome‐wide meta‐analyses identify three
cholangitis and primary biliary cirrhosis: a loci associated with primary biliary
systematic review. J. Hepatol. 56: cirrhosis. Nat. Genet. 42: 658–660.
1181–1188. 11 Mells, G.F., Floyd, J.A., Morley, K.I. et al.
2 Gershwin, M.E., Selmi, C., Worman, H.J. (2011). Genome‐wide association study
et al. (2005). Risk factors and comorbidities identifies 12 new susceptibility loci for
in primary biliary cirrhosis: a controlled primary biliary cirrhosis. Nat. Genet. 43:
interview‐based study of 1032 patients. 329–332.
Hepatology 42: 1194–1202. 12 Nakamura, M., Nishida, N., Kawashima, M.
3 Corpechot, C., Chretien, Y., Chazouilleres, et al. (2012). Genome‐wide association
O., and Poupon, R. (2010). Demographic, study identifies TNFSF15 and POU2AF1 as
lifestyle, medical and familial factors susceptibility loci for primary biliary
associated with primary biliary cirrhosis. cirrhosis in the Japanese population. Am. J.
J. Hepatol. 53: 162–169. Hum. Genet. 91: 721–728.
4 Jones, D.E., Watt, F.E., Metcalf, J.V. et al. 13 Liu, J.Z., Almarri, M.A., Gaffney, D.J. et al.
(1999). Familial primary biliary cirrhosis (2012). Dense fine‐mapping study identifies
reassessed: a geographically‐based new susceptibility loci for primary biliary
population study. J. Hepatol. 30: 402–407. cirrhosis. Nat. Genet. 44: 1137–1141.
5 Bergquist, A., Lindberg, G., Saarinen, S., and 14 Juran, B.D., Hirschfield, G.M., Invernizzi, P.
Broome, U. (2005). Increased prevalence of et al. (2012). Immunochip analyses identify
primary sclerosing cholangitis among first‐ a novel risk locus for primary biliary
degree relatives. J. Hepatol. 42: 252–256. cirrhosis at 13q14, multiple independent
6 Yoshida, O., Abe, M., Furukawa, S. et al. associations at four established risk loci
(2009). A familial case of autoimmune and epistasis between 1p31 and 7q32 risk
hepatitis. Intern. Med. 48: 315–319. variants. Hum. Mol. Genet. 21: 5209–5221.
7 Nozic, D., Dimitrijevic, J., and Knezevic‐ 15 Cordell, H.J., Han, Y., Mells, G.F. et al.
Usaj, S. (2005). Occurrence of autoimmune (2015). International genome‐wide meta‐
chronic hepatitis in siblings. Vojnosanitetski analysis identifies new primary biliary
pregled Military‐Medical and cirrhosis risk loci and targetable pathogenic
Pharmaceutical Review 62: 591–594. pathways. Nat. Commun 6: 8019. https://
8 Ohira, H., Shinzawa, J., Suzuki, T. et al. doi.org/10.1038/ncomms9019.
(1998). Two sister cases of autoimmune 16 Paziewska, A., Habior, A., Rogowska, A.
hepatitis. Fukushima J. Med. Sci. 44: 113–120. et al. (2017). A novel approach to genome‐
9 Hirschfield, G.M., Liu, X., Xu, C. et al. wide association analysis identifies genetic
(2009). Primary biliary cirrhosis associated associations with primary biliary
with HLA, IL12A, and IL12RB2 variants. cholangitis and primary sclerosing
N. Engl. J. Med. 360: 2544–2555. cholangitis in Polish patients. BMC Med.
Chapter 3 Genetics and Risk of Autoimmune Liver Diseases 63
Abstract
Autoimmune liver serology is an essential tool in the diagnosis and management of autoimmune
liver disease. However, autoantibodies are not diagnostic on their own, and the serologic laboratory
reports must always be interpreted in the clinical context of the single patient. In autoimmune
hepatitis (AIH), autoantibodies are a key disease feature: according to the simplified diagnostic cri-
teria published in 2008 by the International AIH Group, a definite diagnosis cannot be made in the
absence of positive autoimmune serology. Antinuclear antibody (ANA), originally named antinu-
clear factor, was the first antibody associated with AIH. Extrahepatic and systemic conditions may
also be associated with autoantibodies positivities, especially ANA, antismooth muscle antibody,
and antineutrophil cytoplasmic antibody (ANCA), thus multidisciplinary patient management is
key to avoid misdiagnosis. ANCA is tested on ethanol-fixed and formaldehyde-fixed human
neutrophils, in contrast to the remaining liver-relevant autoantibodies, which are detected on tri-
ple rodent tissue.
Key Points
●● The recommended first‐line tests for are characterized by anti‐nuclear and/or
autoimmune liver serology are indirect anti‐smooth muscle antibody.
immunofluorescence (IIF) on triple rodent ●● AIH type 2 is characterized by anti‐liver‐
tissue coupled with a solid‐phase assay kidney microsomal type 1 and/or anti‐liver
for anti‐soluble liver antigen antibody. cytosol type 1 antibody.
●● IIF using human neutrophils as a substrate ●● Anti‐mitochondrial antibody is the serologic
should be used to detect anti‐neutrophil hallmark of primary biliary cholangitis (PBC).
cytoplasmic antibody (ANCA). ●● Anti‐nuclear antibody with a rim‐like/
●● Autoimmune hepatitis (AIH) type 1 and membranous or multiple nuclear dots
autoimmune sclerosing cholangitis (ASC) pattern on HEp2 cells is specific for PBC.
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
66 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
rodent tissue (i.e. liver, kidney and stomach), region is added, and the preparation is placed
though some observer‐independent immuno- under an ultraviolet microscope: fluoro-
chemical techniques are either already avail- chrome‐labeled antibodies emit light allowing
able or under development [2]. IIF, a technique the detection of any bound antibody
introduced in 1954 by Weller and Coons, (Figure 4.1). Positive sera should be titrated to
allows the simultaneous detection of ANA, extinction; the process should be repeated,
SMA, AMA, anti‐LKM1, and anti‐LC1, the doubling the dilution until the reactivity disap-
latter specificity being usually masked by pears. Any nuclear reactivity (i.e. sera which
the concomitant presence of anti‐LKM1. stain the cell nuclei on triple rodent tissue)
However, IIF is observer‐dependent and should be further characterized on HEp2 cells,
poorly standardized. derived from a human laryngeal carcinoma,
In an attempt to maximize consistency which have prominent nuclei that allow the
among laboratories, a consensus statement detection of IIF nuclear patterns. These pat-
was published in 2004 by the Committee for terns have high clinical s ignificance and should
Autoimmune Serology of the IAIHG [2] always be reported by the laboratory to the cli-
providing detailed guidelines on substrate nician. An additional very important advantage
preparation, application of test sera, dilution of of IIF is the ability to detect autoantibodies
sera and fluorochrome‐labeled reagents, use of directed against still molecularly undefined
control sera, and interpretation of immunoflu- antigenic targets (see later).
orescence reactivities. The quality of the tissue ANCA are directed against antigens local-
substrate is important: besides appropriate ized in neutrophils, hence human neutrophils
cutting of the kidney tissue to ensure presence are used as an IIF substrate to look for this
of both cortex and medulla on the preparation, specificity.
rodent tissues should be freshly frozen and In adults, the positivity cutoff at IIF on
unfixed. Commercially available slides with triple rodent tissue is 1 : 40, while in chil-
triple rodent tissue are usually treated with fix- dren and adolescents the cutoff is lower,
atives to extend shelf‐life, hampering the inter- being 1 : 20 for ANA and SMA, and 1 : 10
pretation of fluorescence patterns due to for anti‐LKM1 and anti‐LC1. The target
background staining. The patient’s serum, at a antigens of anti‐LKM1, AMA, anti‐LC1,
starting dilution of 1 : 10, is incubated with the and partially ANA, SMA, and ANCA have
tissue sections, which are subsequently washed been identified, leading to the establish-
to remove unbound antibodies. A second ment of solid‐phase immunoassays. Such
fluorochrome‐labeled antibody specifically
assays are based on purified or recombinant
targeting the human immunoglobulin constant antigens coupled to a solid phase; patient
Antibody-containing
patient serum
Fluorochrome-
labeled anti-human
antibody
Figure 4.1 Diluted patient serum is incubated with triple rodent tissue sections, which are subsequently
washed to remove unbound antibodies. A second fluorochrome‐labeled antibody specifically targeting the
human immunoglobulin constant region is added, and the preparation is placed under an ultraviolet
microscope. Fluorochrome‐labeled antibodies emit light allowing the detection of any bound antibodies that
give specific patterns.
68 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
serum is added, and any circulating anti- large variety of reported IIF staining patterns
body targeting the specific antigen binds to and of molecularly identified ANA antigens.
it. A second antibody is added in order to ANA, originally named anti‐nuclear factor,
detect any bound autoantibody; this second was the first antibody associated with AIH. It
antibody may be enzyme‐labeled, radiola- was detected for the first time in 1949 in the
beled, or labeled with a chemiluminescence blood of patients with lupus erythematosus
or fluorescent agent (Figure 4.2). The test (LE), being responsible for the presence of
result may be quantitative or semiquantita- the LE cells, which are neutrophils engulfed
tive, depending on the technique employed. with nuclear debris of damaged cells. The
Enzyme‐linked immunosorbent assay phagocytosis is mediated by ANA. LE cells
(ELISA) is widely used because it can be had been reported by Ian Mackay in 1956 in
fully automated and allows the simulta- the ascites of patients with “chronic hyper-
neous analysis of large numbers of samples, gammaglobulinemic hepatitis,” which led to
in contrast to immunoblot (also called the original naming of AIH as lupoid
Western blot) which is demanding for the hepatitis. Subsequently, it became clear that
laboratory but has the main advantage of LE and lupoid hepatitis (or “chronic hyper-
being highly sensitive and specific. It is gammaglobulinemic hepatitis”) are different
important to note that the AIH‐specific diseases, both being characterized by ANA
anti‐soluble liver antigen antibody (anti‐ positivity. The cumbersome LE test was
SLA) is not detectable by IIF, and therefore replaced by IIF in the early 1960s.
recent international guidelines recommend
combining a solid‐phase assay for anti‐SLA Immunofluorescence Reactivities
with IIF on triple rodent issue in the initial and Antigenic Targets
diagnostic work‐up of patients with sus-
The recommended technique for investi-
pected autoimmune liver disease [4].
gating ANA in the context of liver disease is
IIF on triple rodent tissue (Figure 4.3) [2].
The use of HEp2 cells at a screening level is
Anti‐nuclear Antibody not advisable owing to a reported frequency
of ANA positivity (≥1 : 40) as high as 30%
History
in healthy adults using this very sensitive
ANA, as the name implies, is directed against technique. In IIF on triple rodent substrate,
antigenic targets located in the cell nucleus, ANA stains nuclei in each of the three tissues,
which is a complex structure, explaining the and the staining pattern should be further
Figure 4.2 ELISA (enzyme‐linked immunosorbent assay) as an example of a solid‐phase assay. Purified or
recombinant antigen is coupled to a microplate well. Patient serum is added and any circulating antibody
targeting the specific antigen binds to it. A second, enzyme‐labeled antibody is added in order to detect any
bound autoantibody. In a third step, a chromogenic enzymatic substrate solution is added, which promotes a
color reaction. The intensity of the color produced is proportional to the concentration of antibodies in the
serum sample.
Chapter 4 Autoantibodies and Understanding of Autoimmune Liver Diseases 69
Figure 4.3 Immune serology in autoimmune hepatitis: ANA, anti‐nuclear antibody (rat liver), and/or SMA,
anti‐smooth muscle antibody (rat kidney), define autoimmune hepatitis type 1; anti‐LKM1, anti‐liver
microsomal antibody type 1 (rat liver, bottom; rat kidney, top), and/or anti‐LC1, anti‐liver cytosol type 1 (rat
liver), define autoimmune hepatitis type 2.
characterized on HEp2 cells. A variety of ANA The most frequently detected IIF pattern
patterns can be recognized in IIF: homoge- in AIH type 1 is the homogeneous one,
neous, speckled, nucleolar, membranous/rim‐ being present in some three‐quarters of the
like, and (multiple) nuclear dots. An updated patients, the remainder displaying a speck
tentative international consensus on ANA led or nucleolar pattern. Identified ANA
patterns is available at https://anapatterns. antigenic targets in AIH include double‐ and
org. The extractable nuclear antigens (ENAs) single‐stranded DNA, histones, centromere,
are a heterogeneous group of nuclear antigens chromatin, ribonucleoproteins, and cyclin A.
mostly corresponding to a speckled IIF pattern Commercial kits for a variety of identified
on HEp2 cells. They can be detected by ELISA nuclear antigens are available; however, they
or immunodiffusion. Anti‐ENA antibodies should not be used as a screening tool in AIH
are frequently named after the patients in because about one‐third of ANA‐positive
whose serum they were first identified: anti‐ AIH patients do not react with any of the
Sm for Smith (specific for LE), anti‐Ro for known nuclear molecular targets. Therefore,
Robert (also referred to as anti‐Sjögren syn- ANA in this substantial proportion of
drome antigen A), and anti‐La for Lane (also patients is detected only by IIF.
referred to as anti‐Sjögren syndrome antigen Two IIF ANA patterns are of particular
B). Anti‐ribonucleotide (typically associated importance in the context of liver disease,
with mixed connective tissue disease), anti‐ being pathognomonic for PBC:
Jo‐1 (typically associated with polymyositis),
●● the multiple nuclear dots (MND); and
and anti‐Scl‐70 and anti‐centromere anti-
●● the rim‐like/membranous pattern (Figure 4.4).
bodies (ACA) (both typically associated with
systemic sclerosis) are also part of the anti‐ This highlights the crucial relevance of the
ENA antibodies. The term “anti‐synthetase laboratory reporting the IIF pattern on HEp2
antibodies” is used to identify a subset of ANA cells of any ANA‐positive serum sample.
directed against aminoacyl‐tRNA synthetases Anti‐MND at IIF on HEp2 cells appears as
(including anti‐Jo‐1), which is highly specific 3–20 nuclear dots sized 0.2–1 μm, corres
for a form of immune‐mediated inflammatory ponding to the nuclear bodies. The main
myopathy associated with multiorgan involve- antigenic targets of anti‐MND are the nuclear
ment, primarily interstitial lung disease. proteins referred to as sp‐100 and promyelocytic
70 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
Figure 4.4 Immune serology in primary biliary cholangitis: AMA, anti‐mitochondrial antibody (rat kidney, bottom;
rat stomach, top); RLM, rim‐like membranous antibody (HEp2 cells); MND, multiple nuclear dots (HEp2 cells).
leukemia protein. Recently, sp140 and small mitotic plate. This is in contrast to anti‐
ubiquitin‐related modifiers have also been MND, which do not stain metaphase cells,
reported to be targeted by anti‐MND. Anti‐ allowing differentiation between the two.
rim‐like/membranous ANA stains the The ACA target is the kinetochore, a com-
nuclear envelope, a complex structure con- plex protein localized to the chromosome
sisting of a double bilayer membrane, the centromere, the immunodominant autoanti-
nuclear lamina layer, and the nuclear pore gen being CENP‐B.
complex. It is therefore not surprising that a ANA are frequently detected also in PSC
variety of molecular targets have been identi- patients, with a reported prevalence rang-
fied, including gp210, lamin‐B receptor, and ing from 7.4 to 77%. However, their diag-
nucleoporin p62. ELISA and immunoblotting nostic and prognostic role, as well as
assays based on recombinant gp210 and specific antigenic targets, have been poorly
sp100 have been developed; however, they investigated. Reported antigenic targets in
should not replace IIF as they do not include PSC are double‐stranded DNA, SSA/B,
all antigens targeted by anti‐rim‐like/ ENA, RNP, Scl‐70, single‐stranded DNA
membranous and anti‐MND ANA, respec- and Sm (Table 4.1).
tively. They have a role in confirming IIF
results, as well as in detecting anti‐gp210 and
Clinical Significance in Autoimmune
anti‐sp100, being more sensitive than IIF,
Liver Disease
partially because at IIF concomitant AMA
may mask ANA. Non‐specific ANA IIF ANA characterizes type 1 AIH, being found
patterns detected in PBC patients include in two‐thirds of the patients, where it is asso-
ACA, speckled, homogeneous, and nucleolar ciated with anti‐SMA in half of the cases and
patterns. ACA is best identified on the
is found in isolation in the remainder
Hep‐20‐10 form of HEp2 cells which con- (Tables 4.1 and 4.2). ANA titers do not corre-
tains a high number of mitotic cells and late with disease activity in neither children
appears on interphase cells as multiple dis- nor adults [5,6]. ANA are very rarely found in
crete dots scattered all over the nucleus, association with anti‐LKM1 in type 2 AIH
corresponding to the chromosome centro- [6]. There is no association of specific ANA
meres. ACA stains mitotic cells as well, with IIF patterns or molecular targets in AIH with
a speckled/dot pattern localized to the clinical parameters, with the exception of a
Table 4.1 Clinical significance of anti‐nuclear antibody in autoimmune liver disease.
AIH, autoimmune hepatitis; AMA, anti‐mitochondrial antibody; ASC, autoimmune sclerosing cholangitis; PBC, primary biliary cholangitis; PSC, primary sclerosing cholangitis;
RNP, ribonucleoprotein; Sm, Smith; SMA, smooth muscle antibody; SS, Sjögren’s antigen.
Table 4.2 Main features of the liver‐relevant autoimmune serologic specificities.
Other
Methods of associated
Specificity detection Antigenic targets Disease specificity Frequency liver diseases Comments
SMA IIF Filamentous actin VG and VGT IIF patterns 85% in AIH‐1 DILI V pattern is seen in 20% of type 1
Vimentin specific for AIH‐1 75% in ASC NAFLD AIH patients
Desmin Up to 83% in PSC Viral hepatitis Titers correlate with disease activity
Unknown in PBC Wilson disease
Anti‐actin Molecular‐ Actin Specific for AIH‐1 at high titers 60% in AIH‐1 DILI Sensitivity and specificity depend
antibody based assays NAFLD on the chosen cutoff point. Less
Viral hepatitis specific for AIH than the VG/VGT
Wilson disease IIF pattern
Anti‐LKM1 IIF Epitopes of CYP2D6 Specific for AIH‐2 in absence Up to 90% in Hepatitis C Titers correlate with disease activity
Molecular‐ of hepatitis C AIH‐2
based assays
Anti‐LC1 IIF Formiminotransferase Specific for AIH‐2 in absence 60% of AIH‐2 Hepatitis C Only serologic marker in 10–30%
Molecular‐ cyclodeaminase of hepatitis C of AIH‐2 cases
based assays Titers correlate with disease activity
Anti‐SLA Molecular‐ O‐Phosphoseryl‐tRNA (Sec) Highly specific for AIH. Low Up to 58% in Very rare in Associated with more severe
based assays selenium transferase disease sensitivity when tested AIH‐1, AIH‐2 hepatitis C disease course
with commercial solid phase and ASC
assays
pANNA IIF β‐Tubulin isotype 5; HMG1; Specific for AIH‐1, PSC and 40–96% in AIH‐1 May be the only serologic marker in
HMG2; other yet unknown IBD 26–94% in PSC AIH‐1
autoantigens 75% in ASC IBD to be excluded
Anti‐ IIF ASGPR Liver‐specific but not disease‐ 24–82% in AIH‐1 DILI Titers correlate with disease activity
ASGPR specific 16–57% in AIH‐2 Viral hepatitis
6–100% in PBC NAFLD
33% in PSC Alcoholic liver
disease
AMA IIF E2 subunits lipoyl domains Specific of PBC 95% in PBC Acute liver Predicts PBC if detected in absence
Molecular‐ of PDC, OGDC and failure of liver disease
based assays BCOADC, E3‐binding PBC/AIH
protein of PDC overlap
AIH, autoimmune hepatitis; AMA, anti‐mitochondrial antibody; ASGPR, asialoglycoprotein receptor; BCOADC, branched‐chain 2‐oxo acid dehydrogenase complex; DILI, drug‐
induced liver injury; HMG1, high mobility group non‐histone chromosomal protein; IBD, inflammatory bowel disease; IIF, indirect immunofluorescence; LC1, liver cytosol type 1;
LKM, liver kidney microsomal; NAFLD, non‐alcoholic fatty liver disease; OGDC, 2‐oxoglutarate dehydrogenase complex; pANNA, perinuclear anti‐neutrophil nuclear antibody;
PBC, primary biliary cholangitis; PDC, pyruvate dehydrogenase complex; PSC, primary sclerosing cholangitis; SLA, soluble liver antigen; SMA, smooth muscle antibody; staining
patterns: V, vessel; G, glomerulus; T, tubulus.
Chapter 4 Autoantibodies and Understanding of Autoimmune Liver Diseases 73
VGT pattern corresponds to the F‐actin or from AIH. For this reason, the laboratory
microfilament pattern. These reactivities are should report the SMA IIF pattern on kidney
highly characteristic, but not entirely specific tissue of every positive sample. Nevertheless,
for type 1 AIH. Furthermore, some 20% of the clinician needs to be aware that some 20%
type 1 AIH patients lack the VGT/microfila- of the type 1 AIH SMA‐positive patients have
ment pattern, and this finding should not only the V pattern. Recent data suggest that
preclude a diagnosis of AIH. The observed SMA‐positive patients with VG/VGT pat-
F‐actin IIF reactivity suggested the establish- terns and any increase of transaminase levels
ment of molecular‐based assays using puri- are at risk of developing type 1 AIH on long‐
fied filamentous actin as an antigen source. term follow‐up. Molecular assays based on
These assays are less specific than the VG/ purified actin are available and should be
VGT IIF patterns, their specificity being used as complementary to IIF. Lastly, it should
higher if the anti‐actin antibody is detected be recalled that ASC is characterized by the
at high concentration. Molecularly tested low‐ same serologic profile as type 1 AIH, namely
concentration anti‐actin antibody is found in by SMA and ANA positivity.
a wide variety of conditions, including non‐
alcoholic steatohepatitis, type 2 AIH, PBC,
PSC and, importantly, viral hepatitis. Attempts Anti‐liver‐kidney
have been made to establish higher cutoffs of
the assays, but the improved specificity is
Microsomal Antibody
achieved at the expense of sensitivity. The cur-
History
rently used cutoffs are low, leaving it to the cli-
nician to correctly interpret the laboratory Anti‐LKM was first reported by Mario
results in the clinical context. The molecular Rizzetto in 1973 while working in Deborah
target of VGT SMA is probably part of fila- Doniach’s laboratory in London. He detected
mentous actin, as suggested by the IIF pattern in the serum of 16 patients, the majority hav-
seen on fibroblasts. However, this assumption ing liver disease, an antibody staining liver
is challenged by the fact that about 20% of and kidney tissue of both human and animal
AIH type 1 patients with VG or VGT SMA origin. This antibody was named anti‐LKM
fail to react with purified actin when tested by since the reactivity was abolished by
molecular‐based assays. Possible causes for incubation of the serum with a “microsomal
this incomplete overlap are presence of other, fraction” obtained by ultracentrifugation of a
yet unknown, SMA antigenic targets, or loss liver homogenate. This fraction contains the
of epitopes during antigen purification. endoplasmic reticulum, where the antigenic
SMA titers correlate with disease activity protein is located, as subsequently demon-
both in adults and in children with AIH type strated. The liver disease associated with the
1 and should be used, together with IgG anti‐LKM antibody was characterized by
levels, to monitor treatment response. Daniel Alagille’s group in Paris in 1987, who
In conclusion, SMA characterizes type 1 described this condition affecting mainly
AIH, being detected in some 85% of patients, children. In Alagille’s series, the outcome was
where it is associated with ANA in half of the poor despite treatment with prednisone and
cases. However, it lacks disease specificity, azathioprine. This aggressive juvenile liver
being detected in viral hepatitis, malig- disease was called type 2 AIH.
nancies, rheumatic diseases, DILI, NAFLD, Anti‐LKM was renamed anti‐LKM1 after
and alcohol‐induced liver disease. The IIF the description of antibodies with slightly dif-
patterns on kidney tissue are very helpful to ferent IIF patterns on triple rodent tissue,
the clinician, since the VG and VGT patterns which were referred to as anti‐LKM2 and anti‐
are specific for AIH, while the V pattern is LKM3, respectively. Anti‐LKM2, targeting
seen also in a variety of conditions different CYP2C9, is of historical interest, since it is
Chapter 4 Autoantibodies and Understanding of Autoimmune Liver Diseases 75
associated with hepatitis induced by ticrynafen levels and higher histologic activity has been
(also known as tielinic acid), an uricosuric anti- reported. Interestingly, two case reports have
hypertensive drug withdrawn from the market been published showing AIH type 2
in the 1980s for its hepatotoxicity. Anti‐LKM3, development following spontaneous HCV
also originally reported by Mario Rizzetto’s clearance. Both patients were females; the
group, is detected in about 13% of patients time elapsed from viral clearance to type 2
with chronic hepatitis delta infection and in a AIH diagnosis was 17 months and 8 years,
minority of type 2 AIH patients. respectively. These cases suggest that, though
rarely, HCV may have a role in triggering
AIH in predisposed individuals, a concept
Immunofluorescence Reactivities, supported by the identification of HCV
Antigenic Targets, and Clinical
sequences that are shared in common with
Significance
CYP2D6, possibly leading to cross‐reactivity
Anti‐LKM1 is the serologic hallmark of type of anti‐HCV antibodies with self‐antigens.
2 AIH, provided that hepatitis C virus (HCV) Therefore, LKM1‐positive HCV‐infected
infection is ruled out. At IIF on triple rodent patients may be predisposed to autoimmu-
tissue, it stains brightly hepatocyte cyto- nity. Because of interferon’s potential to
plasm, while on kidney tissue it stains the induce/enhance autoimmunity, anti‐LKM1‐
larger, proximal tubules (Figure 4.3). It does positive HCV patients were a challenging
not stain stomach tissue, at variance with clinical issue until a few years ago. In the
AMA, thus allowing differentiation of the current era of direct‐acting antiviral agents
two reactivities. This difference is the main this is no longer the case, and patients can be
reason why it is not advisable to use slides safely treated for HCV infection, avoiding
with only kidney tissue as an IIF substrate. interferon and its side effects. However, long‐
Indeed, there are fine differences in the IIF term follow‐up data of anti‐LKM1‐positive
pattern on kidney tissue between anti‐LKM HCV patients after successful treatment with
and AMA, but they are appreciated only be direct‐acting antiviral agents are lacking.
experienced observers. Furthermore, the use In addition to type 2 AIH and HCV infec-
of triple tissue substrate allows the detection tion, anti‐LKM1 is also detected in the rare
of anti‐LC1, which stains only liver tissue. patients with AIH type 2‐like hepatitis in the
The antigenic target of anti‐LKM1 is the context of autoimmune polyendocrinopathy/
hepatic cytochrome P4502D6 (CYP2D6). candidiasis/ectodermal dystrophy (APECED),
Reported immunodominant linear epitopes an autosomal recessive disease due to muta-
within this large protein are CYP2D6193–212, tions in the autoimmune regulator (AIRE)
CYP2D6254–271 and CYP2D6321–351. Molecular‐ gene. Despite the fact that on IIF anti‐LKM1
based assays are commercially available and in this condition is indistinguishable from
have been shown to have high sensitivity and anti‐LKM1 seen in AIH type 2, it targets a
specificity. However, they should not be used different cytochrome, namely CYP2A6, in
at a screening level for the previously men- addition to CYP2D6.
tioned advantages of IIF on triple rodent tissue. The target antigen of anti‐LKM3 is the
In type 2 AIH, anti‐LKM1 levels correlate with uridine 5′‐diphosphate glucuronosyltrans-
disease activity and should be tested every 6 ferase 1 (UGT‐1) family. In rare cases of type
months to monitor the treatment response. 2 AIH this is the only serologic positivity.
Anti‐LKM1 is usually not detectable after liver However, it is not routinely tested by IIF
transplantation for type 2 AIH, and its reap- because it stains only human tissue, and
pearance predicts post‐transplant relapse. molecular‐based assays are lacking.
Anti‐LKM1 is detected in up to 13% of Lastly, it is worth mentioning that anti‐
patients with chronic HCV infection, where LKM1 giving an atypical IIF staining pattern
an association with higher transaminase on triple rodent tissue has been reported in
76 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
de novo AIH, a disease entity indistinguish- There are no clinical features associated
able from AIH arising in patients who have with either reactivity. Although more rarely
undergone liver transplantation for liver dis- than anti‐LKM1, anti‐LC1 has been reported
eases different from AIH. Atypical anti‐LKM1 in patients with chronic HCV infection,
stains only kidney tissue. mainly in association with anti‐LKM1,
thus again underlying the crucial impor-
tance of excluding viral hepatitis before
nti‐liver Cytosol Type 1
A diagnosing AIH.
Antibody
History
nti‐soluble Liver Antigen
A
Anti‐LC1 was identified in Paris in 1988 in Antibody
21 patients with chronic hepatitis (age
range 3–21 years), the majority being History
positive also for anti‐LKM1. Anti‐LC1 was
In 1981, Peter Berg in Germany reported an
not found in healthy controls and in a large
antibody reacting with liver and pancreas
cohort of patients with systemic and
rodent homogenates detected in the blood
hepatic well‐characterized conditions,
of 20 patients with an aggressive form of
thus already suggesting its high disease
chronic hepatitis, responding satisfactorily
specificity.
to immunosuppression with steroids and
azathioprine. He suggested naming this anti-
Immunofluorescence Reactivities, body anti‐liver pancreas (LP). The history of
Antigenic Targets, and Clinical
what later became known as anti‐SLA con-
Significance
tinues in Germany, where in 1986 Michael
The IIF pattern of anti‐LC1 on triple rodent Manns reported an antibody reacting with
tissue is one of bright staining confined to rat liver homogenates which he called anti‐
the liver tissue, with a weakening of the SLA, also detected in the blood of patients
staining around the central vein (Figure 4.3). with chronic hepatitis, mostly young women,
As mentioned before, anti‐LKM1 and anti‐ with hypergammaglobulinemia and good
LC1 often coexist in the same patient; in response to immunosuppressive therapy.
these cases, in IIF anti‐LC1 is masked by Neither anti‐LP nor anti‐SLA was detectable
anti‐LKM1, making the availability of molec- by IIF. Soon it became clear that anti‐LP and
ular‐based assays particularly useful. Indeed, anti‐SLA were one and the same. Both
the identification of the liver enzyme original reports proposed that patients
formiminotransferase cyclodeaminase as positive for anti‐SLA/LP represent a third
the anti‐LC1 antigenic target allowed the type of AIH, since this reactivity was the
establishment of such assays, which should only one detected in a substantial proportion
be used to investigate anti‐LC1 in anti‐ of patients with AIH. This proposal was later
LKM1‐positive patients. declined by the IAIHG, since in the original
About two‐thirds of type 2 AIH patients reports the ANA positivity cutoff was greater
are positive for both anti‐LKM1 and anti‐ than 1 : 80, thus significantly higher than the
LC1. Few data are available about the ≥1 : 40 recommended by the IAIHG diag-
proportion of type 2 AIH patients with iso- nostic scoring systems. Therefore, some
lated anti‐LC1 reactivity, but eported anti‐SLA/LP patients in the original reports
frequencies range from 10 to 30%. All type 2 were likely to be positive for ANA as well,
AIH patients are positive for anti‐LKM1 and thus would be classified as type 1 AIH.
and/or anti‐LC1, apart from the very rare The current name has been simplified to
cases positive for anti‐LKM3 in isolation. anti‐SLA.
Chapter 4 Autoantibodies and Understanding of Autoimmune Liver Diseases 77
in PSC patients include H1, bactericidal/ published a case report demonstrating high
permeability increasing protein, lactoferrin, titers of autoantibodies to tissue homoge-
elastase, cathepsin G, catalase, and human nates in a PBC patient. Seven years later, a
lysosomal‐associated membrane protein 2. landmark paper by Sheila Sherlock’s group
The clinician should bear in mind that described a characteristic IIF reactivity in
ANA stains the neutrophil nucleus and the serum of 32 PBC patients, which was
therefore may mask ANCA if both antibodies not seen in patients with cholestatic dis-
coexist in the same serum. Substrate con- eases of other origins. That this autoanti-
taining both HEp2 cells and neutrophils may body was directed against a mitochondrial
be helpful in detecting both specificities. antigen was first suggested in 1967 in a
Some laboratories do not routinely test ANCA paper reporting its reactivity against mito-
by IIF on fixed neutrophils, but in stead use chondrial fraction of liver homogenates
molecular‐based assays containing recombinant from differential centrifugation. Different
or purified antigenic targets, mainly proteinase 3 mitochondrial reactivities were subse-
and myeloperoxidase. pANNA, however, is not quently reported, leading to a nomenclature
detected by such assays. ranging from M1 to M9, proposed by Peter
The reported proportion of type 1 AIH Berg, which is no longer used.
patients with ANCA positivity ranges bet- The identification of the AMA molecular
ween 40 and 96%. Importantly, ANCA may be target as the E2 subunit of the pyruvate
the only serologic positivity in rare type 1 AIH dehydrogenase complex (PDC) protein by
patients. Current guidelines recommend Eric Gershwin in 1987 was a further major
looking for ANCA in patients with suspected step forward [10], allowing the establish-
AIH who test negative in IIF on triple rodent ment of assays based on the recombinant
tissue and with solid‐phase assays for anti‐ antigen.
SLA. Of note, ANCA is absent in type 2 AIH.
ANCA is frequently detected in children
with ASC, being present in some three‐quar- Immunofluorescence Reactivities,
Antigenic Targets, and their Clinical
ters of the cases [9]. There are no associa-
Significance
tions between ANCA and specific clinical
features in type 1 AIH and ASC. In PSC, the Since AMA targets mitochondria, it stains all
reported frequency of ANCA is up to 94%. mitochondria‐rich tissues. On triple rodent
Unfortunately, patients with small duct PSC tissue, it stains the hepatocyte cytoplasm
were excluded from studies, so no data are faintly and the gastric parietal cells brightly,
available on ANCA frequency in this sub- whereas on kidney tissue sections it stains
group. Some studies in adults with PSC have preferentially the distal, mitochondria‐rich
suggested that ANCA‐positive patients may tubules (Figure 4.4). This is in contrast to
have a more benign disease course. anti‐LKM1, which stains the proximal, larger
Because of the strong association of renal tubules. The cytoplasm of HEp2 cells is
pANNA with IBD, its detection should also stained by AMA, and laboratory reports
prompt endoscopic investigations even in often suggest AMA positivity of sera giving
asymptomatic patients. a diffuse, reticular cytoplasmic IIF pattern
on HEp2 cells. However, it is not advisable to
screen for AMA on HEp2 cells, since the IIF
pattern on HEp2 cells of AMA does not
Anti‐mitochondrial Antibody always mirror the IIF results on triple rodent
tissue or solid‐phase assays.
History
PDC is part of the 2‐oxo‐acid dehydrogenase
The first report of autoantibody positivity multienzyme complex, together with 2‐oxo-
in PBC dates back to 1958, when Ian Mackay glutarate dehydrogenase complex (OGDC)
Chapter 4 Autoantibodies and Understanding of Autoimmune Liver Diseases 79
negative
SMA positive:
ANA positive: characterize AMA positive:
characterize LKM1 and/or
pattern on HEp2 cells LC1 positive:
pattern on kidney PBC highly
tissue probable AIH still possible: check
diagnostic for
AIH-2 for anti-SLA (ELISA/line
Multiple nuclear blot) and ANCA (IIF)
dots or
Membranous/rim-
like: PBC highly VG or VGT pattern:
probable highly suggestive of
AIH SLA positive: pANNA positive highly negative
diagnostic of AIH suggestive of AIH-1 or
ASC
Homogeneous :
suggestive of AIH MIT3-based AMA
V pattern: AIH still assay,
possible, particularly if recombinant-
high titer based gp210/
Any positivity: highly sp100 assays
Other patterns: PBC, AIH, suggestive for PBC
PSC still possible
Figure 4.5 Diagnostic algorithm for autoimmune liver serology based on indirect immunofluorescence on
triple rodent tissue. AIH, autoimmune hepatitis; AMA, anti‐mitochondrial antibody; ANA, anti‐nuclear
antibody; ANCA, anti‐neutrophil cytoplasmic antibody; ASC, autoimmune sclerosing cholangitis; LC1, anti‐liver
cytosol type 1; LKM1, anti‐liver kidney microsomal type 1; pANNA, perinuclear anti‐neutrophil nuclear
antibody; PBC, primary biliary cholangitis; PSC, primary sclerosing cholangitis; SLA, soluble liver antigen; SMA,
smooth muscle antibody; staining patterns: V, vessel; G, glomerular; T, tubular.
with unclear liver disease, particularly those tibodies are not diagnostic on their own, and
presenting with acute/fulminant hepatitis, it the serologic laboratory reports must always
is advisable to repeat testing during follow up. be interpreted in the clinical context of the
single patient. Exclusion of viral hepatitis is of
particular importance, since it is a frequent
Concluding Remarks condition and may be associated with positive
autoimmune serology. The clinician should
Autoimmune liver serology is an essential be familiar with the basic principles of the
tool in the diagnosis and management of laboratory techniques, both to order the
autoimmune liver disease. However, autoan- appropriate tests and to interpret the results.
82 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
References
1 Hennes, E.M., Zeniya, M., Czaja, A.J. et al. 6 Couto, C.A., Bittencourt, P.L., Porta, G. et al.
(2008). Simplified criteria for the diagnosis of (2014). Antismooth muscle and antiactin
autoimmune hepatitis. Hepatology 48: antibodies are indirect markers of
169–176. https://doi.org/10.1002/hep.22322. histological and biochemical activity of
2 Vergani, D., Alvarez, F., Bianchi, F.B. et al. autoimmune hepatitis. Hepatology 59:
(2004). Liver autoimmune serology: a 592–600. https://doi.org/10.1002/hep.26666.
consensus statement from the committee 7 Manns, M.P., Czaja, A.J., Gorham, J.D. et al.
for autoimmune serology of the (2010). Diagnosis and management of
international autoimmune hepatitis group. autoimmune hepatitis. Hepatology 51:
J. Hepatol. 41: 677–683. https://doi. 2193–2213. https://doi.org/10.1002/
org/10.1016/j.jhep.2004.08.002. hep.23584.
3 Hov, J.R., Boberg, K.M., and Karlsen, T.H. 8 Mieli‐Vergani, G., Vergani, D., Czaja, A.J.
(2008). Autoantibodies in primary et al. (2018). Autoimmune hepatitis. Nat.
sclerosing cholangitis. World J. Rev. Dis. Primers 4: 18017. https://doi.
Gastroenterol. 14: 3781–3791. https://doi. org/10.1038/nrdp.2018.17.
org/10.3748/wjg.14.3781. 9 Gregorio, G.V., Portmann, B., Karani, J.
4 European Association for the Study of the et al. (2001). Autoimmune hepatitis/
Liver (2015). EASL clinical practice sclerosing cholangitis overlap syndrome in
guidelines: autoimmune hepatitis. J. childhood: a 16‐year prospective study.
Hepatol. 63: 971–1004. https://doi. Hepatology 33: 544–553. https://doi.
org/10.1016/j.jhep.2015.06.030. org/10.1053/jhep.2001.22131.
5 Gregorio, G.V., McFarlane, B., Bracken, P. 10 Gershwin, M.E., Mackay, I.R., Sturgess, A.,
et al. (2002). Organ and non‐organ specific and Coppel, R.L. (1987). Identification and
autoantibody titres and IgG levels as specificity of a cDNA encoding the 70 kd
markers of disease activity: a longitudinal mitochondrial antigen recognized in
study in childhood autoimmune liver primary biliary cirrhosis. J. Immunol. 138:
disease. Autoimmunity 35: 515–519. 3525–3531.
83
Abstract
It is generally accepted that autoimmune liver diseases arise as a result of a complex interaction of
environmental factors on the background of genetic predisposition to disease. This chapter reviews
the research into the environmental triggers linked with primary biliary cholangitis (PBC) and
factors linked with autoimmune hepatitis. Human leukocyte antigen polymorphisms have been
closely linked with the development of autoimmune disease, drug hypersensitivity, and persistence
of viral infection, whereas other polymorphisms provide protection against developing autoim
mune disease, vasculitis, and clearance of viral infection. There are several models and clinical
observations illustrating how an environmental agent can trigger an autoimmune response, some
of which are not mutually exclusive. The involvement of xenobiotics in the pathogenesis of PBC is
supported by epidemiologic clustering studies around coal mines and toxic waste sites as well as by
animal models modified by specific chemical compounds to elicit the PBC specific antimitochon
drial antibody response.
Key Points
●● Autoimmune liver diseases likely arise ●● The possibility of an environmental factor(s)
from a complex interaction between in triggering autoimmune liver diseases is
genetic and environmental factors. suggested by several epidemiologic studies
●● It is possible that the genetic predisposi and laboratory models.
tion provides a relative state of increased ●● Mechanisms by which agents could
risk to specific infectious or otherwise induce loss of tolerance to self include
noxious chemical agents that mediate loss molecular mimicry, presentation of
of tolerance to self and immune mediated cryptic antigens, epitope spreading,
injury. and microbial superantigens triggering
autoimmunity and immunodeficiency.
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
84 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
AMAs play in causing PBC is questionable Accordingly, the immune system can be
because patients with AMA‐negative PBC manipulated into creating organ‐specific dis
have similar disease outcomes as AMA‐ ease. However, this procedure has obvious
positive patients and some AMA‐positive limitations with non‐organ‐specific autoanti
individuals have no disease at all. As a result, gens. For example, the administration of
AMAs appear to fall into the category of PDC‐E2 with adjuvant failed to produce a
pathologically irrelevant but diagnostically PBC‐like disease in rodent models. This
useful antibodies. Furthermore, Witebsky’s then raises the question: how does an autoim
postulates have not been satisfied in PBC for mune response to the PDC‐E2 antigen, which
the characterized mitochondrial antigens, is found in every nucleated cell, cause an
pyruvate dehydrogenase complex (PDC)‐E2, autoimmune disease limited to bile ducts,
and related inner mitochondrial proteins tear ducts and salivary glands [5]?
[3]. Experiments using the passage of AMA,
autoreactive T lymphocytes, or transgenic
Interaction of Genes and Environmental
expression of PDC‐E2 in mice have all failed
Triggers in Autoimmunity
to generate autoimmune biliary disease
(ABD) [2]. Currently, we lack models to demonstrate
It is well known that experimental models reproducibly how any specific environmental
may be manipulated into inducing autoim agent triggers a complex disease on a
mune responses by inoculation of an auto particular genetic background. Koch’s postu
antigen with an adjuvant, such as Freund lates address acute disease and Hill’s criteria
adjuvant containing inactivated mycobacteria chronic disease but both criteria lack tools
emulsified in mineral oil. Stimulants such as for evaluation of the genetic influence [2].
these were referred to as the “immunologist’s Recent genome‐wide association studies
dirty little secret” by the revered immunolo (GWAS) have identified a number of genes
gist Charles Janeway. He noted that “foreign involved in various immune pathways that
antigen alone was insufficient to elicit the contribute to the pathogenesis of autoim
adaptive immune response and that scientists mune liver disease [1]. One common feature
instead had to routinely pepper their experi is the linkage with polymorphisms within the
ments with crude extracts like mineral oil, major histocompatibility complex (MHC) on
mycobacteria, and aluminum hydroxide in the sixth chromosome that contains the
order to get T and B cells to do their bidding” human leukocyte antigen (HLA) class I and
[4]. He questioned why non‐self antigens II regions. HLA polymorphisms have been
alone failed to be recognized and stimulate an closely linked with the development of auto
immune response. His subsequent studies led immune disease, drug hypersensitivity, and
to a long career investigating the need for persistence of viral infection, whereas other
costimulatory molecules and the role of the polymorphisms provide protection against
innate immune system in regulating adaptive developing autoimmune disease, vasculitis,
immune responses. The “secret” was that and clearance of viral infection. However, the
adjuvant was an absolute requirement for role that HLA plays in the development of
breaking tolerance to self to create autoim PBC and PSC remains obscure and fine map
mune responses. Many autoimmune models ping studies are underway to try to identify
have been established using this maneuver. the relevant genetic risk.
For example, autoimmune arthritis models The PBC genomic studies of patients of
have been created by injecting mice with col European descent revealed a strong linkage
lagen and adjuvant. Similarly, experimental with several polymorphisms along the inter
autoimmune allergic encephalomyelitis leukin (IL)‐12 cytokine signaling pathway.
models have been created by inoculation of However, a clinical trial using the IL‐12 inhib
mice with myelin basic p rotein and adjuvant itor ustekinumab failed to show any appre
to produce a multiple sclerosis‐like disorder. ciable biochemical improvement of disease.
86 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
Accordingly, how the genetic risk contrib Notably, microbes share many highly
utes to disease remains the subject of conserved proteins with humans and so
intense investigation and debate, with some microbial molecular mimicry is very common.
studies showing that the risk alleles provide In the case of PBC, many bacteria have the
diminished IL‐12 transcription. Looking at highly conserved proteins that share similar
other idiopathic inflammatory disorders, ities with the PDC‐E2 autoantigens linked
the genetic risk of patients with extreme phe with PBC (Table 5.1). In these studies, addi
notypes appears to be linked with immuno tional proof of infection is often lacking or
deficiency and increased risk of infectious unconfirmed. Furthermore, the antigenic
disease. Notably, children with very early association may be obvious, but proof of an
onset IBD are often found to have monogenic autoimmune disease pathogenesis is not.
disease linked with immunodeficiency When the possibility of molecular mimicry is
syndromes (e.g. IL‐10 deficiency, Wiskott– proposed as a potential mechanism for
Aldrich syndrome, immunodysregulation induction of autoimmunity, discussion
polyendocrinopathy enteropathy X‐linked concerning the role of infection itself is
syndrome, and atypical severe combined seldom, if ever, discussed. Another
immunodeficiency to name a few). Indeed, consideration is that viruses use mimicry to
several of the candidate genes associated with hijack some cellular machinery, but then the
PSC and PBC, such as IL2, IL2RA, IL12B, host may then use the very same mechanism
TNFRSF1A, and TYK2, are all linked with to restrict viral infection. In summary, we lack
primary immunodeficiency syndromes [1]. definitive processes to demonstrate that
Similarly, most of the mouse models that shared host and microbial antigenic determi
develop spontaneous autoimmune liver nants trigger an autoimmune response that
disease are severely immune deficient [1,2]. then causes chronic disease, when microbial
Therefore, we have put forward the model pathology and antimicrobial immune
that the genetic predisposition provides a responses are also present.
relative state of increased risk to specific The presentation of cryptic antigens that
infectious agents, permitting establishment are not usually exposed to the immune
of chronic infection that in turn triggers the system has also been linked with auto
autoimmunity, rather than an over‐exuberant immune responses (Figure 5.1b). A good
response to self [5]. example of this process is the demonstration
of increased and aberrant expression of
PDC‐E2 and other oxo‐acid dehydrogenase
Mechanisms for Triggering Autoimmunity
proteins on the cell surface of cholangiocytes,
There are several different models and and salivary glands, and within perihepatic
clinical observations illustrating how an lymph nodes of PBC patients [3]. Variations
environmental agent can trigger an autoim of this process include the modification of
mune response (Figure 5.1), some of which autoantigens such as neoantigens that may
are not mutually exclusive. One such mecha arise in tumors, for example. A second pro
nism, first conceived over 35 years ago, is posed mechanism is the modification of
molecular mimicry, which is thought to occur enzymes and other autoantigens by xenobi
when viral or bacterial epitopes resemble host otics or drugs that have been shown to trigger
proteins and infection triggers an autoim autoimmune responses.
mune response with loss of tolerance to self Epitope spreading may occur as a result of
(Figure 5.1a) [6]. On first sight, the hypothesis bystander activation. This process of “guilt
appears attractive. Indeed, an exhaustive list by association” occurs when a microbial
of autoantibodies that recognize microbial antigen is copresented with a self‐antigen,
antigens and host autoantigens has accumu with subsequent induction of an antimicro
lated throughout the years. However, this bial and an autoimmune response to self
hypothesis has proven very difficult to confirm. (Figure 5.1c) [6]. Bystander activation has
Chapter 5 Environmental Exposure and Risk in Autoimmune Liver Diseases 87
(a) (b)
(c) (d)
Figure 5.1 Models of microbial induction of autoimmune responses to pyruvate dehydrogenase complex
(PDC)‐E2. (a) The antigen‐presenting cell (APC) is processing bacterial PDC‐E2 (red) to CD4+ lymphocytes that
provides the necessary secondary signaling to create CD8+ and B‐cell immune responses to the microbial
PDC-2 antigen. By the mechanism of molecular mimicry, the T cells and anti‐mitochondrial antibody (AMA)
recognize and break tolerance to self PDC‐E2. (b) A leaky mitochondrial syndrome occurs as a result of
infection or cell damage and the mitochondrial PDC‐E2 antigens (red) usually sequestered in the inner
mitochondrial membrane become exposed or possibly modified as a cryptic antigen to trigger an autoimmune
response. (c) As retroviruses incorporate cellular proteins (red) when budding from the cell surface, APC
present both self (red) and viral (gray) proteins to the adaptive immune system. Bystander activation occurs
when CD4+ lymphocytes reacting to viral proteins provide secondary signals to B cells to produce AMA and to
CD8+ lymphocytes to enable epitope spreading to PDC‐E2. (d) The betaretrovirus‐infected B cell expresses a
superantigen (gray) that binds the class II MHC to the T‐cell receptor outside of the antigenic grove and
activates T lymphocytes of different antigenic specificity. This potent immune‐stimulating mechanism is
hypothesized to trigger an autoimmune response [6].
Table 5.1 Bacteria linked with primary biliary cholangitis (PBC) and the development of anti‐mitochondrial antibody (AMA) through the mechanism of
molecular mimicry.
Bacterial protein ID
(percent of PBC patients AMA reactivity (antigenic
with serologic reactivity to similarity to amino acids Pathogen detection (PBC
Bacteria linked with PBC bacterial protein) 150–173 of human PDC‐E2) vs. other liver disease) Clinical relevance
Bacteria ✓ ✓ ✓ ✓ ✓ ✗ ✗ ✗ ✗ ✗
Xenobiotics NA Bound to ✗ ✓ NA ✗ ✗ ✗ ✗ ✗
hapten
b a
Betaretrovirus 15% to ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
HBRV Env
a
Electron microscopy, proviral integrations, RNA hybridization, in situ hybridization.
b
Lytvyak et al. [9].
92 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
proteins are relocated to the cell surface as a the new risk factors associated with the region
result of apoptosis or xenobiotic modifica they move to. For example, British migrants
tion. However, biliary apoptosis appears to to Australia experience a three to five times
be limited to immune attack, whereas AMA‐ decrease in the risk of developing PBC com
positive biliary epithelium are observed in pared to people in their home country. The
otherwise normally appearing cholangio opposite trend is observed in migrants
cytes without any evidence of apoptosis [3]. from India moving to the UK, where PBC is
Furthermore, xenobiotic modification of far more common. Given these findings, it
PDC‐E2 has been linked with the production reasonable to suspect that exposure to a
of AMA in animal models but not been common environmental factor may be cause
shown to promote cholangiocyte PDC‐E2 of disease.
expression. The most parsimonious explana A substantial contribution to the hypo
tion for cell surface AMA reactivity is that thesis of an environmental trigger for PBC
unmodified oxo‐acid proteins are actually came from a study by David Triger in 1980
aberrantly expressed from leaky mitochon suggesting a waterborne risk factor [8]. The
dria [5]. report noted that Sheffield in northern
England had a particularly high PBC preva
lence of 54 per million, and that PBC clus
Geo‐epidemiology, Clusters,
tered within specific districts throughout the
and Case–Control Studies of PBC
city. While originally thought to be associ
In support of the influence of environmental ated with the coal mining and steel industry,
factors in the pathogenesis of PBC, the dis the source of water supply was noticed to be
ease has been reported to occur in non‐ a common factor linked to the development
related family members and caregivers. of PBC after ruling other potential environ
Furthermore, the genetic input to disease is mental factors. Indeed, individuals receiving
estimated to be less than 25%, implicating a water from the Rivelin reservoir experienced
role for external agents [1]. In agreement a 10‐fold increased risk of PBC. Water quality
with this idea, epidemiologic reports, cluster tests revealed no significant findings other
analyses and case–control studies demon than having soft water and lower concentra
strate geographic differences in disease prev tions of fluoride. Despite the study failing
alence rates, the formation of PBC clusters to narrow down a conclusive environmental
within small regional areas, and potential trigger, the association of PBC clustering
lifestyle variables contributing to PBC [8]. with a water supply system resulted in suc
Foremost, prevalence rates of PBC vary bet cessive studies investigating PBC clusters
ween national cohort studies. Of the coun and potential water‐related risk factors [8].
tries where studies have been conducted, the Clustering of PBC has also been discovered
incidence of PBC is higher in the USA with around areas of former coal mining sites in
400 per million in 2000, and in China with the UK and toxic waste sites in the USA. A
492 cases per million in 2010. In comparison, study analyzing the spatial variation of PBC
27–54 cases per million have been reported around the old coal mining city of Newcastle
in Japan, with the lowest rate of 19 per mil found three definite clusters, where the
lion reported in Australia. In Europe, there is spatial variation could not be accounted for
a greater incidence of PBC in northern by referral bias, familial clustering, or water
regions, with a prevalence of 200 cases per supply. A higher prevalence of men with PBC
million in southern Wales and 240–251 per was seen within each cluster, where coal
million in northern England [8]. mining was a predominant occupation.
Also of interest, observations from inde Other studies also observed geographic clus
pendent migration studies reveal a correla tering of PBC in northeast of England with
tion between the risk of PBC in migrants and evident space–time grouping indicative of
Chapter 5 Environmental Exposure and Risk in Autoimmune Liver Diseases 93
the potential for a transient environmental into a bacterial pathogenesis of PBC has
agent. Similar studies within the USA have yet to emerge.
also reported increased prevalence of PBC in The concept of molecular mimicry has
close proximity to Superfund toxic waste been the most popular mechanism linking
sites [8]. Two of these groups were found to the introduction of foreign microbial proteins
be proximal to sites contaminated with to a loss of tolerance to self‐proteins in PBC
high concentrations of halogenated aromatic (Table 5.1). This hypothesis suggests that
hydrocarbons. microbes presenting epitopes homologous to
While chemical involvement in the patho mitochondrial proteins can break down
genesis of PBC may be inferred from these immune tolerance to PDC‐E2 in the biliary
studies, it is important to note that hidden epithelium through induction of cross‐reac
confounding factors such as socioeconomic tive autoantibodies and effector T cells
status may also be an explanation. Different (Figure 5.1a). Damage of infected bile ducts
risk factors have been uncovered in epidemi by initial T‐cell responses are then believed to
ologic studies from other countries. For release intact PDC‐E2 and mitochondrial
example, a higher risk of PBC was reported unrelated autoantigens, stimulating the initial
among atomic bomb survivors in Japan. This autoimmune attack on bile duct tissue
observation suggests that either radiation (Figure 5.2a). Activation of antigen‐present
may be a trigger for PBC or that patients with ing cells and other facets of the innate immune
depressed immunity had a lower threshold system are thought to propagate the autoim
for infection. The recent Canada‐wide study mune process, where processing of microbial
documenting a higher prevalence of PBC in mimics, PDC‐E2 and other autoantigens may
the eastern provinces is also in keeping with increase cytokine production and expansion
a relationship with PBC and lower socioeco of autoreactive cells [11].
nomic status. Escherichia coli has been studied as a can
didate agent for PBC because of frequent
reports of UTI in PBC patients. There have
Bacterial Infection and PBC
been multiple reports of increased incidence
Supported mainly by frequent concurrence of recurrent UTI in PBC patients, and E. coli
of UTI in PBC patients and experimental has been isolated in many cases. In one study,
AMA and T‐lymphocyte immune responses E. coli was detected in the stool of all PBC
to PDC‐E2‐related microbial proteins, the patients, though also found present in healthy
role of bacterial infection as an environ controls and 25% of patients with other
mental trigger of PBC has been explored for hepatic diseases [10]. Most studies proposed
several decades [10]. For reasons unknown, the mechanism of molecular mimicry to
lipopolysaccharide (LPS) found on the outer be responsible for production of AMA and
membrane of Gram‐negative bacteria has lymphocyte autoreactivity to PDC‐E2, fol
been reported to induce portal lymphocyte lowing several reports of PBC sera cross‐
infiltration and cholangiocyte degeneration reacting to human and E. coli PDC‐E2 [10].
in mice models [11]. Similarly, lipoteichoic Though the affinity of human AMAs to E.
acid from the outer membrane of Gram‐ coli PDC‐E2 is shown to be 100 times higher
positive bacteria has also been found around in PBC patients, a low titer of AMA has also
damaged bile ducts of PBC patients with been observed in women with recurrent
higher IgA titers in sera [10]. Bacterial DNA UTI, independent of a liver disease diag
with unmethylated CpG motifs has also nosis. Notably, very few E. coli proteins unre
been reported to trigger PDC‐E2‐specific lated to human PDC‐E2 are recognized
T‐helper 1 immune response in PDC‐E2 by PBC sera. The cross‐reactive PDC‐E2
immunized mice [11]. Despite the tanta epitope is also recognized by the cellular
lizing clues, further mechanistic insights immune system, suggesting that E. coli and
94 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
(a)
(b)
Figure 5.2 Model for loss of tolerance to PDC‐E2 and PBC pathogenesis. (a) In a two‐hit process: (i) biliary
epithelial cells undergo apoptosis or other modifications, with xenobiotics for example, and present intact
PDC‐E2 on the biliary epithelium; and (ii) bacterial infection results in PDC‐E2 presentation on antigen‐
presenting cells (APC), which breaks tolerance to human PDC‐E2 (arrows represent T‐cell help). The subsequent
autoimmune response homes in on the aberrant expression of PDC‐E2 in biliary epithelium (? represents factors
that may precipitate cell surface PDC‐E2 in (a)). (b) In a single‐hit process, the mouse mammary tumor virus
(MMTV)‐like human betaretrovirus infects biliary epithelium leading to PDC‐E2 expression on the cell surface.
Then the virus either incorporates PDC‐E2 while budding from the cell surface or exits with PDC‐E2 in exosomes
(not shown). APCs then present PDC‐E2 and viral proteins, resulting in a bystander immune response to PDC‐E2
and an immune response to viral and self‐proteins expressed on biliary epithelium. PBC, primary biliary
cirrhosis; PDC, pyruvate dehydrogenase complex. Source: adapted from Wasilenko et al. [12] with permission.
human PDC‐E2 cross‐react at the CD4 and and 2‐oxoglutarate dehydrogenase complex
CD8 T‐cell level. As an example, clones of (OGDC)‐E2 peptides, while T‐cell clones
human PDC‐E2‐reactive T cells were also specific to E. coli OGDC‐E2 responded to
found to be responsive to E. coli PDC‐E2 human PDC‐E2 autoantigens as well [11].
Chapter 5 Environmental Exposure and Risk in Autoimmune Liver Diseases 95
Similar interest has been provoked by initiate bile duct injury if not specifically
mycobacterial species. The Mycobacterium observed in PBC patients nor found in the
gordonae heat shock protein 65 (Hsp65) was liver. Thus, the autoimmune damage may
noted to share similarity with human PDC‐ presumably be represented by a two‐hit
E2, and PBC patient AMAs reacted to the model (Figure 5.2a) incorporating a “hit‐and‐
protein. In small series, 100% of PBC run” process as the bacteria cannot be
patients were found to have serologic reac detected in the majority of PBC patients.
tivity to the M. gordonae Hsp65, while in Notably in the NOD.1011 mouse model, N.
other reports both DNA and proteins from aromaticivorans inoculation induced AMA
M. gordonae were detected in the livers of a production and IgG and IgA antibodies
proportion of PBC patients (Table 5.1). against bacterial and mammalian PDC‐E2
However, studies in the UK only observed [11]. Following autoimmune T‐cell responses
antibody reactivity to Hsp65 in 4% of PBC against small bile ducts, the mice appear to
patients and other follow‐up studies failed develop bile duct lesions resembling PBC.
to confirm the clinical association of PBC The destruction of small bile ducts is seen
with M. gordonae. Of note, similar to obser even in transfer experiments of monocytes
vations with E. coli, approximately 40% of from infected to control mice, and antibiotic
patients with Mycobacterium tuberculosis treatments prevented progression of the dis
were found to have low‐level AMAs without ease. While these studies firmly link bacterial
any additional evidence of having PBC. infection with ABD, it should be noted that
To date, Novosphingobium aromaticivorans the non‐obese diabetic (NOD)‐derived
has been heavily promoted as the most likely mouse models harbor MMTV. This is impor
bacterial candidate [11]. This aerobic Gram‐ tant because viral expression is linked with
negative bacterium has the ability to metabo the mitochondrial phenotype of PBC, the
lize halogenated compounds and estrogens. spontaneous production of AMA as well as
Therefore, it has been hypothesized that this the development of cholangitis. As this
bacterial infection can cause aberrant expres betaretrovirus is linked to the development
sion of PDC‐E2, possibly aided by xenobiotic of PBC, it is difficult to make firm conclu
modification. The metabolism of estrogens sions about the role of other candidate agents
provides a linked with female preponderance using the NOD‐derived mouse models.
in PBC. As a molecular mimic, two proteins The long and extensive search for a bacte
have been discovered to have a high degree of rial mimic as a trigger for PBC has been
homology with the human PDC‐E2 epitope, questioned [12]. While molecular mimicry is
reacting with the sera of all AMA‐positive an attractive hypothesis linking involvement
and some AMA‐negative PBC patients. It has of bacteria to PBC, the mechanism remains a
also been reported that first‐degree relatives theoretical model. Indeed, there has yet to be
of PBC patients in Iceland harbor antibodies concrete evidence showing that human auto
against N. aromaticivorans which reacted immune disorders in general can be triggered
with mitochondrial autoantigens in PBC by molecular mimicry [6]. In fact, only her
patients as well [11]. In comparison with pesvirus‐induced stromal keratitis in a
other candidates, serum reactivity to N. aro- mouse model has been able to demonstrate a
maticivorans was found to be 100–1000 times definitive role for the induction of disease
higher compared with E. coli which was though molecular mimicry. Furthermore, in
already shown to have high affinity to AMAs. that specific model, viral replication was still
However, N. aromaticivorans was not really needed to perpetuate the disease [12].
detected in PBC patients and molecular sur One key criticism brought up previously
veys only observed bacterial DNA in 25% of regarding the use of a complex molecular
both PBC and control fecal samples. mimicry model in PBC is its requirement for
Accordingly, it is unclear how the agent could the combination of two unrelated hits to
96 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
cause PBC. First the bacterium must provoke PBC. As Koch’s postulates relate to linking
the mitochondrial phenotype of PBC with infection with an acute disease process
expression of PDC‐E2 on surface of bile without genetic predisposition, other criteria,
ducts and then it needs to trigger loss of such as Bradford Hill, can be employed to pro
immune tolerance to PDC‐E2 autoantigens vide a causal relationship for infection and dis
(Figure 5.2a) [12]. Notably, no bacterial infec ease. These criteria address the strength,
tion has been found to induce the mitochon consistency and specificity of the relationship,
drial phenotype, central to the development which most bacterial infections fail to meet
of PBC, or has been convincingly detected in for PBC [2]. The plausibility of an agent trig
PBC biliary epithelium. While the model gering disease and the demonstration of a
covers how immune tolerance may be lost, it biological gradient are also factored in, which
does not provide a definitive cause for the again are lacking for the bacterial hypothesis
development of increased extracellular of PBC [2]. Experimental data are also encom
expression of PDC‐E2 in bile ducts. This passed by the Bradford Hill criteria as well as
model also separates the site of infection the coherence of the bacterial hypothesis,
from the onset of autoimmunity in the liver. which is also incoherent at best (Figure 5.2a)
The second issue is that the data linking [12]. Analogy with similar factors can also be
bacterial pathogens with PBC are circum considered for these criteria, but no bacteria
stantial and heavily reliant on positive have been shown to trigger organ‐specific dis
serology to hypothesized bacterial mimics. ease with autoimmune responses to common
Numerous strains of bacteria have been proteins found within all nucleated cells. As a
investigated as potential infectious agents in result, there are very sparse data to support a
PBC (Table 5.1) with little convincing evi bacterial etiology of PBC (Table 5.2).
dence to link infection with disease. It is true
that many of these bacterial candidates dem
Xenobiotics and PBC
onstrate molecular mimicry to the mitochon
drial autoantigen, and cross‐reactivity with The involvement of xenobiotics in the patho
PDC‐E2 antibodies and/or initiation of T‐cell genesis of PBC is supported by epidemiologic
response. Even parasites such as trypano clustering studies around coal mines and
somes and Ascaridia galli have been linked to toxic waste sites as well as by animal models
PBC by molecular mimicry and antibody modified by specific chemical compounds to
cross‐reactivity to human PDC‐E2. Indeed, elicit the PBC‐specific AMA response [10,11].
the detection of low‐level AMAs in patients Xenobiotics are small foreign compounds
with chronic M. tuberculosis and E. coli infec that can bind and modify self‐proteins and
tion without any histologic or biochemical change their molecular structure sufficiently
evidence of PBC argues against the likelihood to induce an autoimmune response. Notably,
of bacteria (or AMA) inducing PBC. xenobiotics can be found in common
These data lead one to question the reli household detergents, preservatives and
ability of determining the causation of PBC pharmaceuticals and they are mainly metab
based on this serologic activity alone, as olized in the liver. One hypothesis suggests
there are evidently many infectious pathogens that xenobiotics induce autoimmunity in
that fit this criterion. Modifications of Koch’s PBC through a direct toxic effect causing cell
postulates in vitro or in animal models are death through apoptosis or necrosis.
required to demonstrate that a specific Autoantigens will then be found concen
microbe can trigger the mitochondrial pheno trated around the surface of apoptotic cells,
type of PBC, for example. Then the infectious which have been found to generate
agent should be directly linked with the autoantibody responses in systemic autoim
development of cholangitis and shown to munity (Figure 5.2a). Experimental models
reproduce the mitochondrial phenotype of have focused on xenobiotic modification of
Chapter 5 Environmental Exposure and Risk in Autoimmune Liver Diseases 97
the PDC‐E2 complex to alter the molecular protein to trigger loss of tolerance. Notably,
structure and trigger loss of self‐tolerance; T xenobiotics have not been found in PBC
cells and B cells primed against these neoan cholangiocytes nor shown to reproduce the
tigens are then hypothesized to cross‐react mitochondrial phenotype specific for PBC.
with mitochondrial autoepitopes [11]. Current support for the proposed effect of
In support of this hypothesis, a halogenated xenobiotics in PBC has mainly been indirect
organic compound attached to the mitochon evidence based on AMA binding to hapten‐
drial epitope was shown to react with PBC bound xenobiotic or chemically modified
sera at a higher affinity than the native form PDC‐E2. Mouse models treated with xenobi
of the antigen in vitro [11]. Additionally, otics demonstrate a degree of liver disease,
patients repeatedly given halothane‐derived but it is unclear whether this is a result of
anesthesia have been shown to develop anti autoimmune damage or merely chemically
bodies cross‐reacting with lipoylated PDC‐ mediated toxic damage to the liver. The latter
E2. Lipoic acid, a critical component for the is an important consideration as some
reactivity of PDC‐E2 complex, is found on the models also develop peritonitis. Future work
exterior of the PDC‐E2 complex, increasing is still required to determine whether the
the susceptibility for chemical modification. identified halogenated compounds that elicit
Accordingly, it has been hypothesized that AMA and liver lesions in mice are truly
organic compounds with similar molecular linked with the development of PBC in any
structure to lipoic acid could independently shape or form (Table 5.2).
trigger an AMA response. For example, 6‐
bromohexonate, a halogenated xenobiotic
Viruses in PBC
complex, was found to induce AMA produc
tion without the adjunction of the peptide It is not intuitive that a retrovirus should
backbone of PDC‐E2 in a rabbit model [11]. trigger a gramulomatous hepatobiliary dis
However, the model failed to produce liver ease but, as discussed, retroviruses including
lesions and the phenotype dissipated once the HIV have been closely linked with autoim
stimulus was removed. However, a more munity and autoimmune diseases in gen
sustained study using guinea pigs reported eral. Viral discovery studies started with
PBC‐like liver lesions. electron microscopy detection of virus‐like
Animal models of PBC have been designed particles in isolated cholangiocytes from
to evaluate the potential role of xenobiotics. PBC patients as well as serologic studies
One such study reported the development of showing antibody reactivity to retroviruses.
a PBC‐like disease in the NOD.1101 and Then, a human betaretrovirus (HBRV)
C57BL/6 models immunized with 2‐octynoic closely related to MMTV was cloned from a
acid that lost tolerance to PDC‐E2 but did PBC biliary epithelium library using primers
not develop progressive disease. Another capable of amplifying any retroviral pol gene
group also tested the effect of 2‐nonynoic sequence. The full‐length proviral genome
acid, and this compound not only shows a was then cloned from a PBC patient’s peri
high affinity to AMAs in PBC sera but also hepatic lymph node [3].
induces cholangiopathy in treated mice. This The female biology of betaretrovirus infec
artificial compound is frequently used in sev tion is of interest. Betaretroviral replication is
eral cosmetic products and nail polish, which turned on by female hormones during preg
suggests this lifestyle factor may be a reason nancy and the virus is produced in large
for observed female predominance. amounts in breast milk. The reliance for
Similar to the bacterial model for PBC, the female hormones on viral production may be
xenobiotic hypothesis relies on external com intimately linked with the female predisposi
pounds molecularly mimicking or altering tion to PBC. Neonatal mice become tolerant
the highly conserved PDC‐E2 mitochondrial to infection when infected at birth from
98 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
infected breast milk and make poor antibody infection (i.e. demonstration of viral integra
responses to MMTV. Initially, B cells become tions into host DNA) showed that the
infected within the gut‐associated lymphoid majority of patients had infection in cholan
tissue and the infected cells produce a supe giocytes. In situ hybridization and RNA
rantigen that stimulates lymphocyte repli measurement assays also clearly showed
cation (Figure 5.2d), which enables MMTV active viral infection in biliary epithelium as
infection as betaretroviruses can only repli well, providing the necessary proof of infec
cate in dividing cells. In mice and humans, tion at the site of disease. In follow‐up
betaretroviruses are cell associated, mainly studies, HBRV was isolated by coculture of
detected in lymphoid tissue, expressed at a PBC lymph nodes with HS578T cells, a
low level, and difficult to detect. human breast cancer cell line.
Because of difficulty with viral detection, One of the most interesting facets linking
the early studies in PBC patients were con betaretrovirus with PBC is that viral infection
sidered somewhat controversial. While is intimately linked with the mitochondrial
HBRV was detected in most perihepatic phenotype [3]. In the first instance, HBRV
lymph nodes, only one‐third of PBC patients and MMTV infection have been shown to
were found to have evidence of viral infec trigger the mitochondrial phenotype and
tion within the liver. Follow‐up studies using have addressed Koch’s postulates in an in
the gold standard for detecting retroviral vitro model (Figure 5.3). In these studies,
PBC
lymph node
γ irradiation
MMTV
Control
lymph node
Control
viruses
Figure 5.3 Primary biliary epithelial cells (BEC) develop the mitochondrial phenotype when incubated with
lymph node homogenates from PBC patients, conditioned supernatants and MMTV. Isolated perihepatic
lymph nodes from PBC and control patients were homogenized and cocultured with normal BEC isolated
from liver transplant recipients. Subsequently, the PBC lymph node‐conditioned BEC demonstrated
anti‐mitochondrial antibody (AMA) reactivity as observed by immunoelectron microscopy and
immunofluorescence, with evidence of human betaretrovirus (HBRV) p27 capsid proteins by
immunofluorescence. Conditioned supernatants had the same effect, whereby only the PBC‐conditioned
supernatants triggered AMA reactivity. This process was abrogated by gamma‐irradiation. Normal BEC
incubated with supernatant from mouse mammary tumor virus (MMTV)‐producing MM5MT cells also
showed AMA reactivity and anti‐p27 capsid immunofluorescence, whereas control viruses had no such
effect. Source: Xu et al. [3]. Copyright 2003 National Academy of Sciences.
Chapter 5 Environmental Exposure and Risk in Autoimmune Liver Diseases 99
lymph node homogenates from PBC patients Western blot and immunoprecipitation
were shown to trigger the appearance of cell studies suggest that betaretroviruses take
surface PDC‐E2 in cultured biliary epithelial up the mitochondrial autoantigens while
cells. The conditioned supernatants could budding from the cell surface, which can
also induce the mitochondrial phenotype and trigger an autoimmune response by epitope
this property could be abrogated by gamma‐ spreading and bystander activation. In
irradiation. Subsequently, the HBRV was addition, betaretroviruses express a
characterized within the conditioned media superantigen that non‐specifically activates
and, as the final experiment, pure isolates of lymphocytes in an unrestricted pattern.
MMTV were shown to trigger the mitochon Notably, a search of the betaretrovirus
drial phenotype in cultured biliary epithelial genome reveals no antigens resembling the
cells (Figure 5.3). Although reported in one mitochondrial oxo‐acid dehydrogenase or
publication, these findings have been repro other known autoantigens associated with
duced in three separate laboratories. PBC. Rather, the autoantigens may be intri
Betaretroviruses have also been linked cately associated with the viral biology, such
with the mitochondrial phenotype in vivo the nuclear pore protein gp210 that may
[2,3]. For example, perihepatic lymph nodes guide a pre‐integration complex into the
from PBC patients have been shown to har nucleus. Accordingly, a different model for
bor monocytes with evidence of HBRV loss of tolerance to PDC‐E2 has been pro
infection and PDC‐E2 expression posed for betaretrovirus infection. HBRV
(Figure 5.4a,b). Similarly, MMTV infection triggers cell surface PDC‐E2 exposure on the
has been linked with the mitochondrial surface of infected cholangiocytes, which is
phenotype in all the spontaneous mouse incorporated into the virion particle that
models of ABD tested (Figure 5.4c–j) [2]. may then trigger AMA production and loss
One of the mouse models, NOD.c3c4, was of tolerance to PDC‐E2 (Figure 5.2b). In this
instrumental for demonstrating a central example, the pathogenesis of PBC no longer
role of MMTV in the development of ABD revolves around autoimmunity, but rather
because the mice had clear evidence of centers on the role of betaretroviral infec
MMTV cholangitis in bile ducts demon tion that masks itself by manipulating
strating the mitochondrial phenotype and the host’s immune system. Indeed, a recent
the disease could be abrogated using report of cellular immune responses
combination antiretroviral therapy. Indeed, involving intrahepatic lymphocytes revealed
a recent clinical trial using the combination that PBC patients harbor a 100 fold excess of
of emtricitabine, tenofovir, and lopinavir lymphocytes reacting with HBRV peptides
boosted with ritonavir demonstrated the versus the characterized immunodominant
utility of this approach for producing sus PDC-E2 epitope.
tained biochemical and histologic responses
for patients who could tolerate therapy.
However, the side‐effect profile of the protease Autoimmune Hepatitis
inhibitors was far in excess of that reported
for the treatment of HIV and enrollment AIH is an idiopathic liver disease character
was terminated prematurely. ized by specific autoantibodies, elevated
Betaretroviruses appear to be capable of IgG levels, histologic evidence of hepatitis,
inducing autoimmunity by several mecha and subsequent response to corticosteroid
nisms (Figure 5.1b–d). First, betaretrovirus treatment following diagnosis. Usually AIH
infection induces changes in metabolism can be diagnosed once other causes of liver
and mitochondrial biogenesis leading disease have been excluded, such as viral
to unveiling of cryptic epitopes usually hepatitis. However, case reports have linked
sequestered within mitochondria. Secondly, AIH with a spectrum of hepatotropic viruses
100 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
(a) (b)
(c) (d)
Figure 5.4 Immunochemistry studies show betaretrovirus proteins in the same tissue distribution as anti‐
mitochondrial antibody (AMA) reactivity. (a) In the PBC lymph node, green anti‐p27CA signal is observed in the
cytoplasm with aberrant expression of PDC‐E2 plasma membrane localization of mitochondrial antigen with
AMA (red) but only mitochondrial staining can be seen in the (b) PSC lymph node (magnification ×400). (c)
Anti‐mouse mammary tumor virus (MMTV) p27CA reactivity was observed in mononuclear cells in the spleen
of the AMA‐producing interleukin‐2 receptor alpha deficient mouse with (d) AMA reactivity observed in a
similar distribution (magnification ×200, with ×600 in insets). (e) Anti‐MMTV gp52SU, (f ) anti‐MMTV p27CA,
and (g) AMA are all observed in an apical distribution in NOD.c3c4 bile ducts (arrowed) but not in veins (*).
In the healthy C57Bl negative control, only background reactivity is seen with each antibody (h, i, and j)
(magnification ×600). PBC, primary biliary cholangitis; PSC, primary sclerosing cholangitis. Source: (a,b)
Xu et al. [3]. Copyright 2003 National Academy of Sciences.
Chapter 5 Environmental Exposure and Risk in Autoimmune Liver Diseases 101
References
1 Hirschfield, G.M., Chapman, R.W., Karlsen, after liver transplantation and effects on
T.H. et al. (2013). The genetics of complex graft and patient survival. Gastroenterology
cholestatic disorders. Gastroenterology 144: 156: 96–107.
1357–1374. 8 Griffiths, L., Dyson, J.K., and Jones, D.E.
2 Sharon, D. and Mason, A.L. (2015). Role of (2014). The new epidemiology of primary
novel retroviruses in chronic liver disease: biliary cirrhosis. Semin. Liver Dis. 34:
assessing the link of human betaretrovirus 318–328.
with primary biliary cirrhosis. Curr. Infect. 9 Lytvyak, E., Montano‐Loza, A.J., Saxinger,
Dis. Rep. 17: 460. L., and Mason, A. (2015). Combination
3 Xu, L., Shen, Z., Guo, L. et al. (2003). Does a anti‐retroviral therapy provides reduction
betaretrovirus infection trigger primary in human betaretrovirus load and durable
biliary cirrhosis? Proc. Natl. Acad. Sci. USA biochemical responses in patients with
100: 8454–8459. primary biliary cirrhosis. Hepatology 62
4 Gayed, P.M. (2011). Toward a modern (Suppl): 528A.
synthesis of immunity: Charles A. Janeway 10 Koutsoumpas, A.L., Kriese, S., and
Jr. and the immunologist’s dirty little secret. Rigopoulou, E.I. (2012). Popular and
Yale J. Biol. Med. 84: 131–138. unpopular infectious agents linked to
5 Mason, A. (2018). Is PBC a viral infectious primary biliary cirrhosis. Autoimmun.
disease? Best Pract. Res. Clin. Gastroenterol. Highlights 3: 95–104.
34–35: 27–39. https://doi.org/10.1016/j. 11 Selmi, C., De Santis, M., Cavaciocchi, F.
bpg.2018.06.001. et al. (2010). Infectious agents and
6 Sfriso, P., Ghirardello, A., Botsios, C. et al. xenobiotics in the etiology of primary
(2010). Infections and autoimmunity: the biliary cirrhosis. Dis. Markers 29: 287–299.
multifaceted relationship. J. Leukoc. Biol. 87: 12 Wasilenko, S.T., Mason, G.E., and Mason,
385–395. A.L. (2009). Primary biliary cirrhosis,
7 Montano‐Loza, A., Hansen, B., Roccarina, bacteria and molecular mimicry: what’s the
D. et al. (2019). Factors associated with molecule and where’s the mimic? Liver Int.
recurrence of primary biliary cholangitis 29: 779–782.
103
Section II
Autoimmune Hepatitis
Aliya Gulamhusein1,2 and Patrick McKiernan3
1
Toronto Centre for Liver Diseases, Toronto General Hospital, University Health Network, Toronto, Ontario, Canada
2
Division of Gastroenterology, University of Toronto, Toronto, Ontario, Canada
3
Children’s Hospital of Pittsburgh, University of Pittsburgh Medical Center, Pittsburgh, PA, USA
Abstract
Key Points
●● Autoimmune hepatitis (AIH) is rare and ●● Diagnosis is clinical and takes into account
characterized by liver‐directed immune serologic profile including anti‐nuclear
injury. and anti‐smooth muscle antibodies in the
●● Clinical presentation can vary from acute case of type 1 AIH, and anti‐liver kidney
severe hepatitis with fulminant liver microsomal and anti‐liver cytosol type 1
failure to an asymptomatic elevation in antibodies in the case of type 2 AIH.
liver tests. In either case a high index of Whether type 3 AIH exists is controver
suspicion must be maintained to allow for sial but the disease‐specific anti‐soluble
confident, expedient diagnosis and timely liver antigen antibody may characterize
delivery of effective therapy. this subtype.
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
106 Section II Autoimmune Liver Diseases and Their Clinical Correlation
being associated with younger age of disease (Treg) cells are of particular interest in their
onset, higher immunoglobulin G (IgG) levels, counteracting functions, with Th17 cells
treatment failure, relapse, and progression to involved in host defense and Treg suppres
liver transplantation, and the latter being sion, and Tregs responsible in large part for
more typical of an older, female cohort with a maintaining immune homeostasis and self‐
more benign course [11]. Indeed, associations tolerance [26,27]. Indeed, increased hepatic
with HLA‐DR3 and ‐DR4 are sufficiently Th17 cells are seen in patients with AIH as
strong to be included in some diagnostic cri compared to viral hepatitis B, with degree of
teria including the revised score developed by infiltration correlating with degree of inflam
the International Autoimmune Hepatitis mation and fibrosis [28]. While Tregs are
Group (IAIHG) [12]. Non‐HLA risk loci have likely to be highly relevant, their precise role
also been identified through GWAS efforts, in disease remains undefined. Early studies
with Src homology 2 adaptor 3 (SH2B3) and demonstrated reduced numbers and dys
the caspase recruitment domain family functional Tregs at diagnosis; however,
member 10 (CARD10) identified as genes subsequent efforts suggested normal levels
potentially implicated in disease [13–15]. of circulating Tregs and intrahepatic enrich
Interestingly, while pan‐ethnic associations ment, though this discrepancy may be attrib
have been seen with the HLA‐DRB1 locus, utable to measurement of different Treg
other HLA subtypes differ between ethnic subsets or to aberrant homing in an inflamed
groups and in a cohort of Japanese patients, liver [22,29,30]. It has been suggested that a
the non‐HLA CARD10 gene variant was not low frequency of CD39+ Tregs with impaired
associated with disease, highlighting the fact immunosuppressive ability and heighted ten
that genetic susceptibility may be vary bet dency to produce interferon (IFN)‐γ and
ween and across ethnicities [16]. interleukin (IL)‐17 on proinflammatory
It is likely that in a genetically susceptible challenge leads to increased conversion of
host, exposure to an as yet undefined Tregs into effector cells, driving the patho
antigenic trigger results in a dysregulated genesis of disease [24,31]. Such lines of evi
liver‐targeted immune response. While the dence emphasize the importance of immune
triggers in AIH are largely unclear, some tar dysregulation in disease and further under
gets are well recognized with cytochrome standing of these biologic mechanisms offers
P450 2D6 as the principal autoantigen recog potential for novel targeted therapeutic
nized by the anti‐LKM1 antibody, sep (O‐ strategies.
phosphoserine) tRNA: Sec (selenocysteine)
tRNA synthetase (SepSecS) as the main
target for anti‐SLA antibody, and formimi Epidemiology
notransferase cyclodeaminase as the
principal target for anti‐LC1 antibodies AIH is rare in both adults and children
[17–19]. Once initiated, both the innate and though epidemiologic data are limited and
adaptive immune responses interact leading may be biased by under‐recognition, particu
to hepatitic liver injury. Several lines of evi larly in underdeveloped countries. That said,
dence highlight the importance of T‐cell‐ in adults the prevalence ranges between 15
mediated immunity in disease [20–24]. and 25 cases per 100 000, though this seems
Presentation of an antigenic peptide bound to be rising [5]. A recent large nationwide
to HLA class II molecules to naive CD4+ T‐ population‐based Danish study noted a dou
helper cells leads to widespread differentiation bling in incidence of AIH between 1994 and
to Th1, Th2, and Th17 subsets which subse 2012 which could not be attributed to diag
quently produce an array of distinct cyto nostic ascertainment [32]. In children, the
kines that contribute to downstream immune reported minimum incidence is 0.2–0.4 per
dysregulation [25]. Th17 and T regulatory 100 000 children and prevalence rates vary
108 Section II Autoimmune Liver Diseases and Their Clinical Correlation
between 3 and 10 per 100 000 based on two fueled the hypothesis of ethnic variability in
recent North American studies [33]. outcome in AIH. That said, such differences
AIH is a female predominant disease, may also reflect distinct environmental expo
occurring in a 3 :
1 ratio [5]. A bimodal sures or intricate gene–environment
distribution of presentation is seen in adults interactions, or be driven by complex social
with peaks in the teenage years and another determinants of health.
in middle age between the fourth and sixth
decade, though recent data suggest an
increasing number of elderly patients diag
nosed beyond 65 years [34,35]. Type 1 AIH is
Presentation
by far most frequent in adults, with type 2
Adult‐onset AIH
AIH accounting for at most 10% of adult AIH
cases [5]. In children, type 1 AIH is at least The spectrum of clinical manifestations in
twice as common as type 2 and females are AIH are widely variable, ranging from
more commonly affected in both types. In asymptomatic to extreme forms of severe
children, the mean age of onset for type 1 acute hepatitis and fulminant hepatic failure.
AIH is between 10 and 11 years of age and for In the latter case, it is important to note
type 2 is younger at 6–7 years of age [36]. that autoantibodies and hypergammaglobu
Concurrent autoimmune diseases in patients linemia may be absent at first screen and thus
and their first‐degree relatives are common, a heightened index of suspicion for the possi
in both childhood and adulthood reported in bility of AIH must be maintained [46].
20–40%. While autoimmune thyroid disease Clinical presentation can be broadly grouped
is most common, a spectrum of manifesta into those with an acute severe presentation,
tions ranging from celiac disease to type 1 those with non‐specific symptoms, and
diabetes, inflammatory bowel disease, and asymptomatic patients with often subclinical
rheumatologic, dermatologic, and hemato disease. Up to 25% of patients present with
logic autoimmune manifestations has been an acute onset and can be further subdivided
described [37,38]. into those with an acute‐on‐chronic presen
While typical AIH is seen in all ethnic tation in contrast to true acute AIH in the
backgrounds, there are insufficient data to absence of clinical or histologic evidence of
determine whether significant inter‐ethnic chronicity [47,48]. In one‐third of patients,
differences exist in disease presentation and/ chronic non‐specific symptoms ranging
or outcome, though some observations have from, but not limited to, fatigue and gener
been reported. For instance, Alaskan natives alized malaise, right upper quadrant pain,
have a high frequency of acute icteric disease anorexia, weight loss, amenorrhea, and poly
at presentation, and a more frequent and arthralgias may be the primary complaints
severe phenotype has been reported in North [49]. Finally, one‐third of patients with AIH
American Aboriginal populations [39,40]. are asymptomatic at presentation with an
African‐American patients fare worse than insidious course typically characterized by
their white American counterparts, with a chronically unexplained elevations in liver
higher frequency of cirrhosis, treatment tests [5,50]. Importantly, up to half of patients
failure, and progression to liver transplanta at diagnosis have histologic evidence of
tion [41,42]. Patients of Hispanic origin may cirrhosis at presentation, attesting to the
also have an aggressive presentation with cir likelihood of long‐standing subclinical dis
rhosis and cholestatic features being reported ease [47,51]. This highlights the importance
more frequently, and Asian patients also tend of timely and confident diagnosis to avoid
toward poorer outcomes [43–45]. These delays in initiation of immunosuppressive
observations are limited to retrospective therapy that may delay or prevent poor
series of small patient numbers but have outcomes (Table 6.1). Physical findings are
Chapter 6 Autoimmune Hepatitis 109
Adults Pediatrics
●● amenorrhea
●● polyarthralgias
often absent except in the case of established Occasionally, children come to attention
cirrhosis where typical findings may be seen, because of an incidental finding of abnormal
including palmar erythema, spider angioma, liver function tests. Disappointingly, some
and sarcopenia. In decompensated disease children will give a history of a previous
with portal hypertension, obvious manifesta unexplained seronegative hepatitis, which
tions of splenomegaly, ascites, pedal edema, emphasizes the importance of identifying an
or encephalopathy may be seen. etiology for childhood hepatitis. Frequently,
there is clinical evidence of chronic liver dis
ease including spider nevi, finger clubbing,
Pediatric‐onset AIH
and palmar erythema. Hepatomegaly is usual
Four basic phenotypes can be distinguished and splenomegaly is common.
in pediatric AIH. The commonest presenta
tion, in about half of cases, is with an acute
hepatitis. Symptoms are similar to acute viral Diagnosis
hepatitis, which is itself now rare in most
developed countries. A small propor tion The diagnosis of AIH is a summation of
may progress to develop encephalopathy and clinical, biochemical, immunologic, imaging,
acute liver failure within the first 2 months of and histologic features once diseases that
presentation. This is commoner is type 2 mimic AIH, including viral hepatitis and
AIH and is often seen in younger children. drug‐induced liver injury, have been
For another 30–40%, there is a longer and excluded. Typical biochemical abnormalities
often non‐specific history including malaise, include at times severe elevation in transam
weight loss, poor growth, and intermittent inase levels and hyperbilirubinemia, often in
jaundice. The remainder of patients pre the absence of significant elevation of chole
sent initially with a complication of chronic static enzymes, though these too may be
liver disease including hematemesis, bruis moderately deranged. Elevated gamma‐glob
ing, or incidentally identified organomegaly. ulin (IgG) levels are found in 80–85% of
110 Section II Autoimmune Liver Diseases and Their Clinical Correlation
patients even in the absence of cirrhosis and possibility of a variant syndrome with PBC
usually with normal corresponding IgA and but this should only be considered in the
IgM levels [52,53]. Concomitant elevation in clinical context of associated cholestasis and
IgA and/or IgM may be suggestive of super the predominant pathology based on clinical
imposed additional disease processes such as phenotype should be treated [65]. Periodic
alcoholic liver disease or primary biliary reassessment of serologic profile may be of
cholangitis (PBC). Isolated elevated IgG value particularly in patients initially diag
levels are distinctive to AIH and are a useful nosed with seronegative disease as antibody
biomarker of treatment response [54,55]. titers may vary over time. While in adults
The absence of viral hepatitis is included as a autoantibody titers correlate weakly with dis
diagnostic criterion but poses challenges in ease activity and clinical course, in children
endemic settings where the two entities may these are useful biomarkers of treatment
coexist. Particularly in the era of direct‐act response and should be followed.
ing antiviral therapy for hepatitis C virus Histologic evaluation is required for the
(HCV) infection, treatment of underlying diagnosis of AIH. As much as possible, it
viral hepatitis with subsequent careful reas should be performed before treatment initia
sessment of biochemical activity, autoanti tion and is useful in guiding treatment
body profile, and histologic findings may aid decisions and staging fibrosis. Hepatitis at the
in diagnosis and treatment decisions. portal–lobular boundary, so‐called “inter
Autoantibodies are an essential compo face” hepatitis, exhibits typical manifestations
nent of diagnostic evaluation. Indirect immu including plasma‐cell rich infiltrates, hepato
nofluorescence is the preferred technique cyte rosettes, emperipolesis, and hepatocyte
[56,57]. ANA and SMA are not disease necrosis; while these can be etiologically non‐
specific and show wide heterogeneity but are specific, in the right clinical context they are
hallmarks of type 1 AIH and are present in 43 highly supportive of the diagnosis (Figure 6.1)
and 40–60% of patients, respectively [66]. Plasma cells are usually abundant,
[56,58,59]. Anti‐LKM1 and/or anti‐LC1 are though their absence does not preclude diag
hallmarks of type 2 AIH, often coexist, and nosis [67]. In the acute setting, severe pan‐
have been reported in 66 and 53% of patients, lobular hepatitis with bridging and/or massive
respectively [58]. Interestingly, these anti necrosis with histologic collapse may be seen
bodies have been described in 5–10% of that may closely resemble drug‐induced toxic
patients with HCV infection, likely due to the injury and thus needs to be considered in the
presence of homologous sequences between correct clinical context [68].
CYP2D6 and HCV proteins [60,61]. Anti‐ A comprehensive scoring system was
SLA is the only disease‐specific autoantibody originally developed by the IAIHG in 1999 to
though is detected in only 30% of patients, define homogeneous cohorts of patients for
and often coexists with other autoantibodies, clinical trials rather than for diagnostic
particularly anti‐Ro52 [62]. As mentioned purposes and utilizes clinical, laboratory,
previously, the presence of anti‐SLA has serologic, and histologic factors along with
been associated with severe disease and response to therapy to classify patients’ proba
poorer prognosis in AIH [63]. In adults, bility of having disease (Table 6.2) [12]. Though
positive titers are considered at levels above robust, this tool is rather cumbersome to use
1 :
40 dilution by immunofluorescence, in clinical practice given its complexity,
whereas in children weaker titers of 1 : 20 for inability to distinguish cholestatic syndromes,
ANA and SMA or 1 : 10 for anti‐LKM are and use of response to therapy as a diagnostic
considered sufficiently positive in the correct criterion. Subsequently, in 2008 a simplified
clinical context [64]. Anti‐mitochondrial scoring system was proposed for everyday
antibodies (AMAs) are seen in 8–12% of clinical practice by the IAIHG and is based
patients with adult AIH and may raise the on four easy‐to‐use parameters including
Panel A Panel B
Figure 6.1 A and B. Panel A ‐ Histologic depiction of marked chronic portal inflammation with extension into
interface and lobules. Panel B – magnified view of severe portal and interface lymphoplasmacytic infiltration Images
provided by: Dr. Oyedele Adeyi, MD. Department of Pathology, Toronto General Hospital, University Health Network
Table 6.2 Revised Original Scoring System of the International Autoimmune Hepatitis Group
Criteria Score
Sex Female +2
AP:AST (or ALT) ratio >3 –2
< 1.5 +2
γ globulin or IgG level > ULN >2 +3
1.5‐2.0 +2
1.0‐1.5 +1
<1 0
ANA, SMA, or anti‐LKM‐1 titres >1:80 +3
1:80 +2
1:40 +1
<1:40 0
AMA Positive –4
Viral markers Positive –3
Negative +3
Drugs Yes –4
No +1
Alcohol <25 g/day +2
>60 g/day –2
HLA DR3 or DR4 +1
Immune Disease Thyroiditis, colitis, etc +2
Other markers Anti‐SLA, anti‐actin, anti‐LC1, pANCA +2
Histologic Features Interface hepatitis +3
Plasmacytic +1
Rosettes +1
None of the above –5
Biliary changes –3
Other features –3
Treatment Response Complete +2
Relapse +3
IgG concentration, autoantibody titers, his profile. Biliary histologic features are com
tologic findings, and absence of viral hepatitis moner in ASC but may be very subtle and
[54]. The diagnostic performance of the sim can also occur in AIH. There are epidemio
plified score is very good, with a sensitivity of logic differences in that ASC shows an equal
95%, specificity of 90%, and diagnostic accu sex distribution and is more frequently
racy of 92% [69,70]. Its ease of use, utility in associated with inflammatory bowel disease.
patients with cholestatic syndromes, and However, in individual cases distinguishing
lack of requirement for treatment response between these entities requires cholangiog
make it a useful tool in the correct clinical raphy. Abdominal ultrasound frequently
context to serve as a guide for the initiation demonstrates a large gallbladder in ASC but
of therapy. The revised IAIHG or simplified intrahepatic bile duct abnormalities are
score is not ideal for use in childhood due to rarely detected. Magnetic resonance cholan
age‐related differences in alkaline phospha giography is the investigation of choice with
tase levels and autoantibody cutoff titers, the endoscopic retrograde cholangiography
inclusion of alcohol and drug intake, and being reserved for cases where a dominant
their lack of utility in acute liver failure. A stricture requires intervention, or if there is
new scoring system has been proposed to need for tissue diagnosis to help exclude
address these concerns but this still requires cholangiocarcinoma.
prospective validation [71].
While variant syndromes, otherwise
referred to as “overlap syndromes,” are dis
cussed in more detail in Chapter 15, a few Treatment
points are worthy of mention. Approximately
10% of adult patients with AIH will present, Long‐term immunosuppression is the corner
usually consecutively and rarely simulta stone of therapy in pediatric and adult AIH
neously, with clinical features of coexistent and requires corticosteroid‐based induction
PBC or primary sclerosing cholangitis (PSC) followed by, ideally, steroid‐sparing mainte
[5]. While widely validated diagnostic cri nance of remission with azathioprine (Figure
teria for these syndromes are lacking, it is 6.2). Treatment is required for all patients
accepted that the revised and simplified with AIH who have symptoms, significant
IAIHG‐developed scores perform poorly in biochemical or histologic activity, and/or
this setting and are not recommended for advanced fibrosis or cirrhosis; the recommen
diagnosis of AIH in these patients. Patients dation for therapy in patients with mild
with AIH and features of PBC should meet activity without fibrosis and in elderly patients
diagnostic criteria for PBC and require a for whom prognostic impact of treatment is
biopsy to establish the presence of moderate less well defined remains unclear [5,72,73].
to severe interface hepatitis. In patients with Despite this, given the potential for unpredict
features of AIH and PSC, diagnostic features able severe fluctuations of disease and the risk
of AIH should be established and cholangio of progression, in the absence of contraindica
graphic features of biliary disease should be tions treatment for AIH is recommended in
demonstrated [5]. all patients. Lifelong immunosuppression is
generally the rule, with durable sustained
remission off therapy being achieved in the
Autoimmune Sclerosing Cholangitis
minority (<20%) of patients [74,75]. The goal
Unique to childhood is the particularly chal is to prevent or delay progressive liver disease
lenging issue of distinguishing between AIH by achieving biochemical, serologic, and his
and autoimmune sclerosing cholangitis tologic remission, defined as normalization of
(ASC). These patients present in a very sim alanine aminotransferase (ALT) and IgG and
ilar manner and have a similar immunologic resolution of interface activity.
Chapter 6 Autoimmune Hepatitis 113
Budesonide 9 mg daily
Predni(so)lone 0.5–1 mg/kg/day may be alternative in mild disease
(without advanced fibrosis)
ASSESS RESPONSE
TREATMENT RESPONDER
SUB-OPTIMAL RESPONSE
(normal ALT and IgG)
CONFIRM ADHERENCE
Taper prednisone and continue
long term azathioprine Verify absence of superimposed drug injury
maintenance Verify absence of superimposed liver disease (NASH,
viral hepatitis)
Assess for variant syndromes with PBC or PSC
Figure 6.2 Suggested treatment algorithm for management of autoimmune hepatitis in adults.
steroid‐based therapy though, importantly, or advanced fibrosis who wish for a trial off
histologic remission lags behind biochemical therapy, a pretreatment withdrawal biopsy
normalization by months [77]. should be performed; in patients with active
Once remission is achieved, the target dose interface hepatitis or an hepatic activity index
of azathioprine should be optimized to 2 mg/ (HAI) above 3, the risk of relapse exceeds 90%
kg and subsequently prednisone should be and in this case treatment should be continued
gradually tapered off. With this approach [55]. If histologic remission is confirmed
steroid‐free relapse is rare after 1 year, and and treatment is stopped, patients should be
greater than 80% remission rates can be closely monitored for relapse; this is most
expected after 67 months of follow‐up [80,81]. likely to occur within the first 12 months and
Long‐term maintenance with azathioprine warrants resumption of therapy which after
monotherapy is optimal, though low‐dose relapse should be lifelong [74].
prednisone therapy (<10 mg) may be consid
ered but is a less favorable option. Safety and Second‐line Options
efficacy data for maintenance budesonide Second‐line therapy may need to be consid
therapy in adults and children alike are lacking ered in 10–15% of patients as a result of
and this approach is discouraged. Tolerability either intolerance or suboptimal response to
to standard regimens remains problematic, standard regimens. After confirmation of
with 10–15% of patients discontinuing therapy adherence and exclusion of alternative etiol
due to side effects that can result in clinical ogies of liver injury, true treatment failure is
relapse warranting reinduction therapy. On‐ seen in less than 5% of patients and in this
therapy relapse warrants a detailed history of case second‐line therapy may need to be
adherence, which is the most likely reason for considered. The most frequently considered
loss of disease control. Cosmetic, psychologic, agents include 6‐mercaptopurine (6‐MP),
and metabolic toxicity of corticosteroids can mycophenolate mofetil (MMF), and calci
be particularly troublesome with long‐term neurin inhibitors such as ciclosporin or
use. Monitoring while on azathioprine is tacrolimus. Data for use of biologic agents
required, particularly for features of bone including rituximab and anti‐tumor necrosis
marrow suppression. An increased risk of factor (TNF) agents are limited to very small
malignancies has been reported, though this case series. In patients intolerant to azathio
remains controversial; however, this may be of prine, use of 6‐MP has been trialed with
relevance with lifelong use [82,83]. response rates of up to 75% from small series
being reported [84]. As 6‐MP is a downstream
Withdrawal of Therapy metabolite of azathioprine, manipulation of
Durable remission off therapy is unusual, the same biologic pathway is a rational option
with relapse rates of 50–90% widely reported. particularly when azathioprine use is limited
For all patients, maintenance treatment by gastrointestinal complaints or myelosup
should be continued for at least 3 years with pression. There are limited data of patients
at least 2 years of complete remission (as tolerating 6‐MP in the setting of azathio
evidenced by normal serum ALT and IgG) prine‐related hepatotoxicity and if this is
before a trial of withdrawal may be consid considered careful monitoring of liver tests
ered [78]. In patients with advanced fibrosis needs to be ensured with prompt discontinu
or cirrhosis, prior episodes of relapse, or ation if drug injury is noted. MMF is most
intolerance to induction therapy, continuing widely used as a second‐line agent at doses of
maintenance agents without a trial of discon 1.5–2 g daily in divided doses, with reported
tinuation is acceptable after discussion of rates of remission ranging between 31 and
risks and benefits, and this may be the 73% in both poor responders and those intol
preferred option for both the patient and erant to azathioprine [85,86]. Data on the
practitioner. In patients without prior relapse efficacy of ciclosporin in adults is limited
Chapter 6 Autoimmune Hepatitis 115
whereas good biochemical response has been and 6‐methylmercaptopurine levels can be
observed in children. Use of tacrolimus as a useful, especially where there is hypersplen
steroid‐sparing second‐line option has also ism with baseline neutropenia and thrombo
been variably reported; however, a recent ret cytopenia and in order to demonstrate
rospective study of 201 patients treated with compliance. An alternative strategy is to
second‐line therapy in AIH showed no reserve azathioprine for delayed steroid
difference in remission rates between MMF‐ response and/or steroid side effects.
treated and tacrolimus‐treated patients who Normal biochemistry can be expected to be
had failed azathioprine (69.4 vs. 72.5%) [87]. achieved in more than 80% of cases by 6
Doses of tacrolimus used generally range months of treatment and a complete response
between 1 and 8 mg daily with target trough is defined as normal biochemistry, normal
blood levels of 4–6 ng/ml being reasonable immunoglobulins, undetectable autoanti
but clearly empiric. For refractory disease, bodies, and absence of histologic activity.
salvage options of therapy with rituximab Maintenance therapy with low‐dose cortico
have been reported in small series demon steroids and azathioprine should be continued
strating biochemical improvement and for at least 2–3 years assuming a complete
improved histologic inflammation, though response. Repeat liver biopsy should be sched
only small numbers of patients have been uled after at least 2 years of biochemical
exposed in this setting [88]. response and only if it would be appropriate to
Use of infliximab has also been trialed with consider relaxing or withdrawing treatment.
moderate success in a series of 11 patients,
though this approach is controversial due to risk Withdrawal of Therapy
of infections noted and reports of anti‐TNF‐ Where there has been complete response the
related immune‐mediated hepatitis [89,90]. options of reduction/discontinuation of
treatment should be discussed with the young
person and their family. This should be con
Therapeutic Approach in Children
sidered at a time of physiologic stability, spe
First‐line Therapy cifically avoiding the rapid growth phase of
As with adults, where there is a possible drug early adolescence. The relapse rate is particu
trigger for AIH, this should be discontinued larly high in type 2 AIH. Many young people
immediately. Induction treatment starts with may consider continuing long‐term low‐dose
prednisolone 2 mg/kg daily up to a maximum monotherapy with azathioprine rather than
of 60 mg. This should be combined with an risk relapse with the subsequent necessity for
antacid in the first instance. This dose should reintroduction of steroids. During relaxation/
be maintained for at least 2 weeks and subse withdrawal of treatment, frequent monitoring
quently reduced over the next 4–12 weeks, should be undertaken including liver bio
according to response, to a maintenance dose chemistry and autoantibody levels.
of 0.1 mg/kg daily. This should be guided by Relapses are relatively common and gener
weekly, or at minimum fortnightly, labora ally associated with non‐compliance. They
tory monitoring. should be treated similarly to initial presenta
Having excluded thiopurine S‐methyl tion with sufficient steroids to induce remis
transferase (TPMT) deficiency, most centers sion and subsequent gradual relaxation to
now introduce azathioprine electively, maintenance treatment according to response.
unless there is severe cholestasis. Once there
has been an initial biochemical response, Second‐line Options
azathioprine should be introduced in a dose First‐line treatment fails or is associated with
of 1 mg/kg daily and if tolerated subse significant side effects in 10–15% of cases.
quently increased to 1.5–2 mg/kg daily over Subsequent choices would include either
the next month. Monitoring 6‐thioguanine MMF or one of ciclosporin/tacrolimus [87].
116 Section II Autoimmune Liver Diseases and Their Clinical Correlation
There is insufficient evidence to make a rec and chronic viral hepatitis in the era of such
ommendation between these agents. The studies. Suffice to say that untreated
choice will often depend on the experience moderate to severe AIH carries a very poor
and familiarity with these drugs of the treating prognosis [91,92]. The survival benefit
physician and family preference. The greatest associated with immunosuppression using
experience is with MMF, commencing at a prednisone with or without azathioprine
dose of 20 mg/kg daily increased as tolerated has been demonstrated in a number of
to 30–40 mg/kg daily. Remission rates of randomized controlled trials with four to
greater than 80% are reported. Particular side seven-fold reduction in mortality in treated
effects include abdominal pain and constipa patients, emphasizing the importance of
tion and neutropenia. A significant concern timely diagnosis and prompt institution of
with this drug is that it is teratogenic and is therapy [80]. Liver‐related morbidity and
contraindicated in pregnancy. transplantation may be avoided with early
Tacrolimus is a potent calcineurin inhib effective intervention and maintenance of
itor and is the default immunosuppressive remission, though in a minority of cases
agent following solid organ transplantation transplantation will be required [80,91]. In
in childhood. It is usually commenced twice North America and Europe, approximately
daily in a lower dose than used in transplan 4% of liver transplant activity is due to AIH
tation and guided by therapeutic monitoring, [21]. Treatment failure is more likely in
aiming for trough levels of approximately severe acute presentations, particularly in
5 ng/ml. Side effects include nephrotoxicity, sub‐fulminant disease; prognosis is particu
hypertension, headache, and diabetes. larly poor in this setting, with mortality
Ciclosporin has a similar mode of action and ranging from 19 to 45% and requirement for
side‐effect profile to tacrolimus. However, liver transplantation in 9–81% [93]. Liver
ciclosporin has proved to be particularly dif transplantation can be a highly effective sal
ficult to use in AIH due to toxicity, and vage option; however, recurrence of disease
increasingly pediatricians are unfamiliar has been reported in up to 36% of patients at
with using and monitoring this drug. 5 years, though late acute cellular rejection
For children resistant or intolerant to sec episodes are also more frequent in patients
ond‐line treatment, there are few proven with AIH and how clinically distinct these
options. There is anecdotal positive experi two entities are remains somewhat unclear
ence with rituximab, infliximab, and siroli [94,95]. Indeed some groups report a lower
mus. When embarking on such treatment, it frequency of recurrent AIH post transplan
should ideally be done as part of a multidisci tation with long‐term prolonged use of low‐
plinary team and following appropriate con dose prednisone post transplantation [96].
sultation with locally available and recognized
expertise. Randomized controlled trials of
Childhood
such treatment are probably impractical but
there is a need for at least a register of such Transplantation will be necessary in approxi
treatment protocols so that knowledge can mately 10–15% of young people with AIH,
be accumulated and shared. with indications falling into two broad groups:
(i) those presenting with acute liver failure
and encephalopathy who do not respond rap
idly to immunosuppressive treatment; and (ii)
Prognosis the more common scenario where transplant
is required for a combination of chronic liver
Adulthood
disease and intolerance of immunosuppres
The prognosis of untreated disease is diffi sive treatment. The hepatic indications for
cult to determine historically in part because transplantation are similar to any chronic
of difficulties with distinction between AIH liver disease. For some young people with
Chapter 6 Autoimmune Hepatitis 117
autoimmune liver disease, myalgia and aner Recurrent disease remains the commonest
gia is a persistent and dominant symptom. cause of graft loss in this group. Factors con
The prognosis after liver transplantation is tributing to recurrence include uncontrolled
excellent, with the expectation of in excess of inflammation at transplant, manifesting
90% high‐quality long‐term survival. The HLA‐DRB1*03, and discontinuing cortico
absolute rate of recurrence following liver steroids [97]. The impact of HLA matching
transplantation is high, with published rates between donor and recipient is unclear. In
ranging from 30 to 80% with increasing inci recognition of this, most programs issue
dence over time. specific immunosuppression regimens for
Severe recurrent disease requiring retrans AIH including long‐term low‐dose cortico
plantation is less common, although a steroids, maintenance azathioprine, and
significant burden at 15–25% overall. calcineurin inhibitors.
118 Section II Autoimmune Liver Diseases and Their Clinical Correlation
References
1 Manns, M.P., Czaja, A.J., Gorham, J.D. et al. 10 Liberal, R., Krawitt, E.L., Vierling, J.M. et al.
(2010). Diagnosis and management of (2016). Cutting edge issues in autoimmune
autoimmune hepatitis. Hepatology 51 (6): hepatitis. J. Autoimmun. 75: 6–19.
2193–2213. 11 Donaldson, P.T. (2004). Genetics of liver
2 van Gerven, N.M., Verwer, B.J., Witte, B.I. disease: immunogenetics and disease
et al. (2014). Epidemiology and clinical pathogenesis. Gut 53 (4): 599–608.
characteristics of autoimmune hepatitis in 12 Alvarez, F., Berg, P.A., Bianchi, F.B. et al.
the Netherlands. Scand. J. Gastroenterol. 49 (1999). International Autoimmune
(10): 1245–1254. Hepatitis Group Report: review of criteria
3 Czaja, A.J. (2016). Diagnosis and management for diagnosis of autoimmune hepatitis. J.
of autoimmune hepatitis: current status and Hepatol. 31 (5): 929–938.
future directions. Gut Liver 10 (2): 177–203. 13 de Boer, Y.S., van Gerven, N.M., Zwiers, A.
4 Mackay, I.R. (2008). Historical reflections on et al. (2014). Genome‐wide association
autoimmune hepatitis. World J. study identifies variants associated with
Gastroenterol. 14 (21): 3292–3300. autoimmune hepatitis type 1.
5 European Association for the Study of the Gastroenterology 147 (2): 443–452 e5.
Liver (2015). EASL clinical practice 14 Liu, J.Z., Almarri, M.A., Gaffney, D.J.
guidelines: autoimmune hepatitis. J. Hepatol. et al. (2012). Dense fine‐mapping study
63 (4): 971–1004. identifies new susceptibility loci for
6 Muratori, P., Lalanne, C., Fabbri, A. et al. primary biliary cirrhosis. Nat. Genet. 44
(2015). Type 1 and type 2 autoimmune (10): 1137–1141.
hepatitis in adults share the same clinical 15 Zhernakova, A., Stahl, E.A., Trynka, G.
phenotype. Aliment. Pharmacol. Ther. 41 et al. (2011). Meta‐analysis of genome‐wide
(12): 1281–1287. association studies in celiac disease and
7 Liberal, R., Longhi, M.S., Mieli‐Vergani, G., rheumatoid arthritis identifies fourteen
and Vergani, D. (2011). Pathogenesis of non‐HLA shared loci. PLoS Genet. 7 (2):
autoimmune hepatitis. Best Pract. Res. Clin. e1002004.
Gastroenterol. 25 (6): 653–664. 16 Migita, K., Jiuchi, Y., Furukawa, H. et al.
8 Doycheva, I., Watt, K.D., and Gulamhusein, (2015). Lack of association between the
A.F. (2019). Autoimmune hepatitis: current CARD10 rs6000782 polymorphism and
and future therapeutic options. Liver Int. 39 type 1 autoimmune hepatitis in a Japanese
(6): 1002–1013. population. BMC Res. Notes 8: 777.
9 van Gerven, N.M., de Boer, Y.S., Zwiers, A. 17 Manns, M.P., Johnson, E.F., Griffin, K.J.
et al. (2015). HLA‐DRB1*03:01 and HLA‐ et al. (1989). Major antigen of liver kidney
DRB1*04:01 modify the presentation and microsomal autoantibodies in idiopathic
outcome in autoimmune hepatitis type‐1. autoimmune hepatitis is cytochrome
Genes Immun. 16 (4): 247–252. P450db1. J. Clin. Invest. 83 (3): 1066–1072.
Chapter 6 Autoimmune Hepatitis 119
18 Wies, I., Brunner, S., Henninger, J. et al. T‐cells on CD8 T‐cell function in patients
(2000). Identification of target antigen for with autoimmune hepatitis. J. Autoimmun.
SLA/LP autoantibodies in autoimmune 25 (1): 63–71.
hepatitis. Lancet 355 (9214): 1510–1515. 30 Peiseler, M., Sebode, M., Franke, B. et al.
19 Lapierre, P., Hajoui, O., Homberg, J.C., and (2012). FOXP3+ regulatory T cells in
Alvarez, F. (1999). Formiminotransferase autoimmune hepatitis are fully functional
cyclodeaminase is an organ‐specific and not reduced in frequency. J. Hepatol.
autoantigen recognized by sera of patients 57 (1): 125–132.
with autoimmune hepatitis. 31 Grant, C.R., Liberal, R., Holder, B.S. et al.
Gastroenterology 116 (3): 643–649. (2014). Dysfunctional CD39(POS)
20 Hardtke‐Wolenski, M. and Jaeckel, E. regulatory T cells and aberrant control of
(2010). Mouse models for experimental T‐helper type 17 cells in autoimmune
autoimmune hepatitis: limits and chances. hepatitis. Hepatology 59 (3): 1007–1015.
Dig. Dis. 28 (1): 70–79. 32 Gronbaek, L., Vilstrup, H., and Jepsen, P.
21 Manns, M.P., Lohse, A.W., and Vergani, D. (2014). Autoimmune hepatitis in Denmark:
(2015). Autoimmune hepatitis: update incidence, prevalence, prognosis, and
2015. J. Hepatol. 62 (1 Suppl): S100–S111. causes of death. A nationwide registry‐
22 Longhi, M.S., Ma, Y., Bogdanos, D.P. et al. based cohort study. J. Hepatol. 60 (3):
(2004). Impairment of CD4(+)CD25(+) 612–617.
regulatory T‐cells in autoimmune liver 33 Jimenez‐Rivera, C., Ling, S.C., Ahmed, N.
disease. J. Hepatol. 41 (1): 31–37. et al. (2015). Incidence and characteristics
23 Ferri, S., Longhi, M.S., De Molo, C. et al. of autoimmune hepatitis. Pediatrics 136
(2010). A multifaceted imbalance of T cells (5): e1237–1248.
with regulatory function characterizes type 34 Schramm, C., Kanzler, S., zum
1 autoimmune hepatitis. Hepatology 52 (3): Buschenfelde, K.H. et al. (2001).
999–1007. Autoimmune hepatitis in the elderly. Am.
24 Taubert, R., Hardtke‐Wolenski, M., Noyan, J. Gastroenterol. 96 (5): 1587–1591.
F. et al. (2014). Intrahepatic regulatory T 35 Chen, J., Eslick, G.D., and Weltman, M.
cells in autoimmune hepatitis are (2014). Systematic review with meta‐
associated with treatment response and analysis: clinical manifestations and
depleted with current therapies. J. Hepatol. management of autoimmune hepatitis in
61 (5): 1106–1114. the elderly. Aliment. Pharmacol. Ther. 39
25 Longhi, M.S., Ma, Y., Mieli‐Vergani, G., and (2): 117–124.
Vergani, D. (2010). Aetiopathogenesis of 36 Mieli‐Vergani, G., Vergani, D., Baumann,
autoimmune hepatitis. J. Autoimmun. 34 U. et al. (2018). Diagnosis and management
(1): 7–14. of pediatric autoimmune liver disease:
26 Korn, T., Bettelli, E., Oukka, M., and ESPGHAN hepatology committee position
Kuchroo, V.K. (2009). IL‐17 and Th17 cells. statement. J. Pediatr. Gastroenterol. Nutr.
Annu. Rev. Immunol. 27: 485–517. 66 (2): 345–360.
27 Seki, E. and Schwabe, R.F. (2015). Hepatic 37 Rigopoulou, E.I., Dalekos, G., and
inflammation and fibrosis: functional links and Bogdanos, D.P. (2007). How common are
key pathways. Hepatology 61 (3): 1066–1079. connective tissue disorders in patients
28 Zhao, L., Tang, Y., You, Z. et al. (2011). with autoimmune hepatitis? Semin.
Interleukin‐17 contributes to the Arthritis Rheum. 36 (5): 332. author
pathogenesis of autoimmune hepatitis reply 333.
through inducing hepatic interleukin‐6 38 Teufel, A., Weinmann, A., Kahaly, G.J. et al.
expression. PLoS One 6 (4): e18909. (2010). Concurrent autoimmune diseases
29 Longhi, M.S., Ma, Y., Mitry, R.R. et al. in patients with autoimmune hepatitis.
(2005). Effect of CD4+ CD25+ regulatory J. Clin. Gastroenterol. 44 (3): 208–213.
120 Section II Autoimmune Liver Diseases and Their Clinical Correlation
39 Hurlburt, K.J., McMahon, B.J., Deubner, H. patients with autoimmune hepatitis. Exp.
et al. (2002). Prevalence of autoimmune Ther. Med. 7 (1): 145–148.
liver disease in Alaska Natives. Am. J. 50 Feld, J.J., Dinh, H., Arenovich, T. et al.
Gastroenterol. 97 (9): 2402–2407. (2005). Autoimmune hepatitis: effect of
40 Minuk, G.Y., Liu, S., Kaita, K. et al. (2008). symptoms and cirrhosis on natural history
Autoimmune hepatitis in a North and outcome. Hepatology 42 (1): 53–62.
American Aboriginal/First Nations 51 Miyake, Y., Iwasaki, Y., Terada, R. et al.
population. Can. J. Gastroenterol. 22 (10): (2006). Clinical characteristics of
829–834. fulminant‐type autoimmune hepatitis: an
41 Verma, S., Torbenson, M., and Thuluvath, analysis of eleven cases. Aliment.
P.J. (2007). The impact of ethnicity on the Pharmacol. Ther. 23 (9): 1347–1353.
natural history of autoimmune hepatitis. 52 Al‐Chalabi, T., Underhill, J.A., Portmann,
Hepatology 46 (6): 1828–1835. B.C. et al. (2008). Impact of gender on the
42 Lim, K.N., Casanova, R.L., Boyer, T.D., and long‐term outcome and survival of patients
Bruno, C.J. (2001). Autoimmune hepatitis with autoimmune hepatitis. J. Hepatol. 48
in African Americans: presenting features (1): 140–147.
and response to therapy. Am. J. 53 Floreani, A., Niro, G., Rosa Rizzotto, E.
Gastroenterol. 96 (12): 3390–3394. et al. (2006). Type I autoimmune hepatitis:
43 Wong, R.J., Gish, R., Frederick, T. et al. clinical course and outcome in an Italian
(2012). The impact of race/ethnicity on the multicentre study. Aliment. Pharmacol.
clinical epidemiology of autoimmune Ther. 24 (7): 1051–1057.
hepatitis. J. Clin. Gastroenterol. 46 (2): 54 Hennes, E.M., Zeniya, M., Czaja, A.J. et al.
155–161. (2008). Simplified criteria for the diagnosis
44 Levy, C., Naik, J., Giordano, C. et al. (2014). of autoimmune hepatitis. Hepatology 48
Hispanics with primary biliary cirrhosis are (1): 169–176.
more likely to have features of autoimmune 55 Luth, S., Herkel, J., Kanzler, S. et al. (2008).
hepatitis and reduced response to Serologic markers compared with liver
ursodeoxycholic acid than non‐Hispanics. biopsy for monitoring disease activity in
Clin. Gastroenterol. Hepatol. 12 (8): autoimmune hepatitis. J. Clin.
1398–1405. Gastroenterol. 42 (8): 926–930.
45 Zolfino, T., Heneghan, M.A., Norris, S. 56 Vergani, D., Alvarez, F., Bianchi, F.B. et al.
et al. (2002). Characteristics of (2004). Liver autoimmune serology: a
autoimmune hepatitis in patients who are consensus statement from the committee
not of European Caucasoid ethnic origin. for autoimmune serology of the
Gut 50 (5): 713–717. International Autoimmune Hepatitis
46 Czaja, A.J. (2011). Autoimmune hepatitis in Group. J. Hepatol. 41 (4): 677–683.
special patient populations. Best Pract. Res. 57 Bogdanos, D.P., Invernizzi, P., Mackay, I.R.,
Clin. Gastroenterol. 25 (6): 689–700. and Vergani, D. (2008). Autoimmune liver
47 Stravitz, R.T., Lefkowitch, J.H., Fontana, serology: current diagnostic and clinical
R.J. et al. (2011). Autoimmune acute liver challenges. World J. Gastroenterol. 14 (21):
failure: proposed clinical and histological 3374–3387.
criteria. Hepatology 53 (2): 517–526. 58 Muratori, P., Granito, A., Quarneti, C. et al.
48 Ferrari, R., Pappas, G., Agostinelli, D. et al. (2009). Autoimmune hepatitis in Italy: the
(2004). Type 1 autoimmune hepatitis: Bologna experience. J. Hepatol. 50 (6):
patterns of clinical presentation and 1210–1218.
differential diagnosis of the “acute” type. 59 Frenzel, C., Herkel, J., Luth, S. et al. (2006).
QJM 97 (7): 407–412. Evaluation of F‐actin ELISA for the
49 Peng, M., Li, Y., Zhang, M. et al. (2014). diagnosis of autoimmune hepatitis. Am. J.
Clinical features in different age groups of Gastroenterol. 101 (12): 2731–2736.
Chapter 6 Autoimmune Hepatitis 121
80 Lamers, M.M., van Oijen, M.G., Pronk, M., 89 Weiler‐Normann, C., Schramm, C., Quaas,
and Drenth, J.P. (2010). Treatment options A. et al. (2013). Infliximab as a rescue
for autoimmune hepatitis: a systematic treatment in difficult‐to‐treat autoimmune
review of randomized controlled trials. J. hepatitis. J. Hepatol. 58 (3): 529–534.
Hepatol. 53 (1): 191–198. 90 Rodrigues, S., Lopes, S., Magro, F. et al.
81 Stellon, A.J., Keating, J.J., Johnson, P.J. et al. (2015). Autoimmune hepatitis and anti‐
(1988). Maintenance of remission in tumor necrosis factor alpha therapy: a
autoimmune chronic active hepatitis with single center report of 8 cases. World J.
azathioprine after corticosteroid Gastroenterol. 21 (24): 7584–7588.
withdrawal. Hepatology 8 (4): 781–784. 91 Cook, G.C., Mulligan, R., and Sherlock, S.
82 Meyer zum Buschenfelde, K.H. and Lohse, (1971). Controlled prospective trial of
A.W. (1995). Autoimmune hepatitis. N. corticosteroid therapy in active chronic
Engl. J. Med. 333 (15): 1004–1005. hepatitis. Q. J. Med. 40 (158): 159–185.
83 Ngu, J.H., Gearry, R.B., Frampton, C.M., 92 Soloway, R.D., Summerskill, W.H.,
and Stedman, C.A. (2012). Mortality and Baggenstoss, A.H. et al. (1972). Clinical,
the risk of malignancy in autoimmune liver biochemical, and histological remission of
diseases: a population‐based study in severe chronic active liver disease: a
Canterbury, New Zealand. Hepatology 55 controlled study of treatments and early
(2): 522–529. prognosis. Gastroenterology 63 (5):
84 Hubener, S., Oo, Y.H., Than, N.N. et al. 820–833.
(2016). Efficacy of 6‐mercaptopurine as 93 Yeoman, A.D., Westbrook, R.H., Zen, Y.
second‐line treatment for patients with et al. (2014). Prognosis of acute severe
autoimmune hepatitis and azathioprine autoimmune hepatitis (AS‐AIH): the role
intolerance. Clin. Gastroenterol. Hepatol. of corticosteroids in modifying outcome. J.
14 (3): 445–453. Hepatol. 61 (4): 876–882.
85 Wolf, D.C., Bojito, L., Facciuto, M., and 94 Montano‐Loza, A.J., Mason, A.L., Ma, M.
Lebovics, E. (2009). Mycophenolate mofetil et al. (2009). Risk factors for recurrence of
for autoimmune hepatitis: a single practice autoimmune hepatitis after liver
experience. Dig. Dis. Sci. 54 (11): transplantation. Liver Transpl. 15 (10):
2519–2522. 1254–1261.
86 Jothimani, D., Cramp, M.E., and Cross, T.J. 95 Sebode, M., Hartl, J., Vergani, D., and
(2014). Role of mycophenolate mofetil for Lohse, A.W. (2018). International
the treatment of autoimmune hepatitis: an Autoimmune Hepatitis Group.
observational study. J. Clin. Exp. Hepatol. 4 Autoimmune hepatitis: from current
(3): 221–225. knowledge and clinical practice to future
87 Zizzo, A.N., Valentino, P.L., Shah, P.S., and research agenda. Liver Int. 38 (1): 15–22.
Kamath, B.M. (2017). Second‐line agents in 96 Krishnamoorthy, T.L., Miezynska‐Kurtycz,
pediatric patients with autoimmune J., Hodson, J. et al. (2016). Longterm
hepatitis: a systematic review and meta‐ corticosteroid use after liver
analysis. J. Pediatr. Gastroenterol. Nutr. 65 transplantation for autoimmune hepatitis is
(1): 6–15. safe and associated with a lower incidence
88 Burak, K.W., Swain, M.G., Santodomingo‐ of recurrent disease. Liver Transpl. 22 (1):
Garzon, T. et al. (2013). Rituximab for the 34–41.
treatment of patients with autoimmune 97 Stirnimann, G., Ebadi, M., Czaja, A.J., and
hepatitis who are refractory or intolerant to Montano‐Loza, A.J. (2019). Recurrent and
standard therapy. Can. J. Gastroenterol. 27 de novo autoimmune hepatitis. Liver
(5): 273–280. Transpl. 25 (1): 152–166.
123
Abstract
Primary biliary cholangitis (PBC) is a complex disease resulting from the combination of genetic
susceptibility, unknown environmental exposures and epigenetic alterations. The diagnosis is based
on persistently elevated cholestatic liver function tests and PBC-specific serology. The majority of
PBC patients are asymptomatic at diagnosis; when symptomatic, they usually report fatigue, pru-
ritus, and dry eyes and mouth. The introduction of ursodeoxycholic acid (UDCA) has dramatically
changed the disease course of individuals with PBC and reduced the rate of liver transplantation
(LT). PBC has a relatively predictable natural history and accurate prognostic models have been
developed. PBC is a good indication for LT considering the excellent long-term post-transplant
survival, better than for other indications, although the disease recurs almost universally.
Key Points
●● Primary biliary cholangitis (PBC) is a ●● The introduction of ursodeoxycholic acid
complex disease resulting from the (UDCA) has dramatically changed the
combination of genetic susceptibility and disease course of individuals with PBC
as yet unknown environmental exposures and reduced the rate of liver transplanta-
and epigenetic alterations. tion (LT).
●● The diagnosis is based on persistently ele- ●● The majority of patients have improve-
vated cholestatic liver function tests and ment in liver biochemistry under UDCA,
PBC‐specific serology. In those seronega- which translates into good outcomes.
tive, a liver biopsy is usually required for Those with inadequate response to UDCA
diagnostic purposes. now have several licensed/evidence‐based
●● The majority of PBC patients are asymp- options that should be used as add‐on
tomatic at diagnosis; when symptomatic, therapies to UDCA.
they usually report fatigue, pruritus, and ●● PBC has a relatively predictable natural
dry eyes and mouth. history and accurate prognostic models
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
124 Section II Autoimmune Liver Diseases and Their Clinical Correlation
have been developed. These can aid clini- ritus; fatigue is not improved by LT and
cians in the decision‐making process therefore is not an indication.
regarding escalation of management and ●● PBC is a good indication for LT consid-
therapeutic approach. ering the excellent long‐term post‐
●● Indications for LT in PBC are similar to transplant survival, in general better than
those for other chronic liver diseases. An for other indications, although the disease
exceptional indication is intractable pru- recurs almost universally.
Even though there are case reports of been highlighted, all of them located in
patients diagnosed at the age of 15 or 93, the immune‐related regions.
typical disease onset is between 30 and Environmental factors have been linked to
60 years. It is estimated that prevalence of immune tolerance breakdown in PBC. In a
PBC in women over the age of 45 years could genetically predisposed patient, an unidenti-
exceed 1 in 800 individuals. Women who fied environmental agent, such as bacterial
present at a younger age (<45 years) are often infection or xenobiotics, is probably able to
more symptomatic and more likely to fail trigger the pathogenetic cascade of events.
standard treatment with ursodeoxycholic PBC patients show higher prevalence of
acid (UDCA), which may translate into a urinary tract infections, mainly associated
higher risk of disease progression. Men are with Escherichia coli. Molecular mimicry
often diagnosed with PBC later in the disease between epitopes of E. coli and pyruvate
course compared to women, with more dehydrogenase E2 subunit (PDC‐E2), the
advanced disease at presentation, they show dominant mitochondrial autoantigen of
poorer biochemical response to UDCA PBC, has been described [4].
therapy and higher risk of developing hepa- Considering that the liver has a major role
tocellular carcinoma (HCC). in chemical detoxification, hepatocytes and
cholangiocytes are highly exposed to xenobi-
otics, which might affect liver pathobiology
Etiopathogenesis and trigger immune response. Although still
controversial, some epidemiologic studies
PBC is the archetypal autoimmune disease have linked the use of cosmetics and smoking
characterized by the presence of disease‐ with PBC. The 2‐octynoic acid, present in
specific autoantibodies toward a mitochon- cosmetic products, has been shown to induce
drial target [3], a genetic architecture antibodies to PDC‐E2 (AMA) and PBC‐like
pointing to immunoregulatory pathways, a lesions in non‐obese diabetic mice immu-
histologic pattern of chronic granulomatous nized with this substance.
lymphocytic cholangitis of the middle and Whatever the trigger is, it leads to loss of
small-size bile ducts, and a high proportion PDC‐E2 immune tolerance and self‐epitopes
of patients having at least one concurrent on the cholangiocytes are targeted by an
autoimmune condition. On the other hand, it autoimmune reaction. Even though the
is noteworthy that immunosuppressant PDC‐E2 protein is ubiquitous, only cholan-
drugs are not effective in PBC. giocytes are selectively damaged in PBC.
PBC is a complex trait deriving from the Defective clearance of apoptotic bodies
combination of genetic susceptibility, containing intact and antigenically reactive
environmental exposures, and epigenetic
PDC‐E2 released from dying cholangiocytes,
alterations. The prominence of a genetic but not other cell types, has been proposed
component in PBC has been highlighted by as a possible cause of the tissue specificity of
several studies. Twin studies have shown the disease.
higher PBC frequency in monozygotic twins PBC is considered a Th1‐dominant auto-
than in dizygotic ones. Genome‐wide immune response as Th1 effector cells are
association studies (GWAS) have identified critical for the development of cholangitis.
dozens of single‐nucleotide polymorphisms Also described is a skewing toward Th1/
(SNPs) associated with disease suscepti- Th17 activation, seen in the cytokine profile
bility. Among them, the strongest [interleukin (IL)‐12/Th1 and IL‐23/Th17] in
association has been found within the the liver of patients with PBC along with
human leukocyte antigen (HLA) class II downregulation of Th2 markers (e.g. IL‐10).
loci, but many non‐HLA SNPs have also While the IL‐12/Th1 pathway may be a
126 Section II Autoimmune Liver Diseases and Their Clinical Correlation
strong determinant in initiation of the dis- Table 7.1 Symptoms and signs in PBC.
ease process, IL‐23/Th17 pathway plays a
dominant role in the subsequent effector Symptoms Signs
stages of the disease. Cells of innate immu-
Pruritus Skin hyperpigmentation
nity are also involved, such as CD1d‐positive
Fatigue Scratching marks or
natural killer T (NKT) cells, whose fre- Sicca syndrome excoriations
quency in the liver of patients with PBC is (xerostomia and Xanthelasma/xanthoma
higher than normal, or NK cells, that are xerophthalmia) Kayser–Fleischer rings
upregulated as well. The association with Right upper Sclerodactyly
quadrant pain Teleangiectasia
granulomas is also indirect evidence of
Raynaud’s
the role of innate immunity, especially phenomenon
macrophages. Arthralgia
Recently, a novel pathogenic theory pro-
posed that the immune dysregulation in PBC
is a downstream process, which follows pri- strong correlation between severity of disease
mary biliary damage. Decreased expression and symptoms, although asymptomatic
and activity of the anion exchanger 2 (AE2) patients are more likely to have early disease.
on the cholangiocytes might lead to impaired Fatigue is a non‐specific symptom affecting
biliary bicarbonate secretion with conse- 50–78% of patients. It can be debilitating,
quent cholestasis and cholangiocyte injury, with a huge impact on quality of life. Its path-
followed by immune tolerance breakdown ogenesis is still unknown, and it is quite chal-
and autoimmune responses toward the sus- lenging to explore its specificity and to treat.
ceptible cholangiocytes, resulting in pro- Comorbidities that could contribute to, or
found biliary damage. Epigenetic studies are worsen fatigue, such as depression, hypothy-
currently ongoing to test this hypothesis. roidism, anemia, obesity, or medication-side
effects, should be promptly identified and
treated.
Clinical Presentation Pruritus can develop at any stage of the dis-
ease, does not correlate with progression of
Asymptomatic Patients liver disease, and may even improve or disap-
Most patients are currently diagnosed when pear as disease becomes more advanced. It is
asymptomatic, having been referred to the usually reported by over 70% of patients, and
hepatologist for abnormal liver function tests it is typically mild to moderate in intensity.
(LFTs), mostly raised gamma‐glutamyltrans- Given the impact on quality of life and night
peptidase (GGT) or alkaline phosphatase sleep, pruritus is correlated with fatigue.
(ALP), performed for annual screening blood It can rarely be severe, non‐responsive to
tests. Other frequent scenarios include medical therapy, and requiring LT. Pruritus
screening of patients with non‐liver auto- in PBC is characteristically intermittent,
immune diseases (e.g. rheumatoid arthritis) worse at night, and improves during summer.
or investigation of elevated cholesterol, Most patients with xerostomia and
evaluation of itch or unresolved cholestasis xerophthalmia have sicca syndrome rather
postpartum. than primary Sjögren syndrome. Other,
less often referred symptoms incluse non‐
progressive right upper quadrant pain,
Symptomatic Patients
Raynaud phenomenon, and arthralgia.
When symptomatic, PBC patients usually Although decompensated liver disease is
report fatigue, pruritus, and sicca complex. less common now, occasionally patients pre-
Symptoms and signs commonly found in sent with jaundice, ascites, or variceal hem-
PBC are summarized in Table 7.1. There is no orrhage at diagnosis. Jaundice is present only
Chapter 7 Primary Biliary Cholangitis 127
in the later stage of the disease, unless the Triglycerides levels are normal or slightly
patient has a rare form of premature ducto- elevated. It is still unclear whether hyper-
penic variant of disease. lipidemia increases the cardiovascular risk
The physical examination is often unre- in PBC.
markable in early PBC. Signs that should be As disease progresses, hepatosplenomeg-
sought include: aly, spider nevi, palmar erythema, muscle
wasting, ascites, edema, and jaundice can be
●● skin hyperpigmentation, which reflects
found. Of note in PBC, cirrhosis is not neces-
deposition of melanin rather than bilirubin;
sarily required for portal hypertensive
●● scratching marks or excoriations from
complications to develop since this can be
intractable itch, which are more common on
secondary to a pre‐sinusoidal portal hyper-
the trunk and legs, and can lead to scarring;
tension caused by nodular regenerative
●● xanthelasma, subcutaneous deposits of
hyperplasia, although the majority of patients
cholesterol typically found around the eyes
with varices are in fact cirrhotic.
and with a yellow color;
●● more rarely xanthoma, cholesterol deposits
potentially found anywhere in the body,
but more often present around tendons Diagnosis
and bone prominences.
Diagnosing PBC is generally straightforward
Kayser–Fleischer rings are seen very occa- [5]. The foundations for a definite diagnosis
sionally in severe cholestasis and result from are persistently elevated cholestatic LFTs,
copper retention. Signs of associated auto- usually defined as elevated ALP for more
immune conditions such as scleroderma, than 6 months, but patients may just have
sclerodactyly or telangiectasia can be also elevated GGT; positive serology, i.e. AMAs
present. with a titer over 1 : 40 or specific ANAs such
It is reasonable that screening tests for as gp210 and sp100; and liver histology show-
celiac disease, thyroid disorders, and Sjogren ing florid bile duct lesions and, in advanced
syndrome are performed at baseline in all disease, ductopenia. It is necessary to obtain
PBC patients. a patient history excluding recent drug
When prolonged jaundice is present, fat‐ exposure, and cross‐sectional imaging ruling
soluble vitamin (A, D, E, K) deficiency may out bile duct obstruction.
occur, with the main problem being the Given the high specificity of serologic
development of osteoporosis, which is gen- markers, liver biopsy is not necessary for the
erally exacerbated by postmenopausal bone diagnosis of PBC; however, it is still required
loss. Night blindness is now exceptional. when PBC‐specific antibodies are absent or
Steatorrhea has been reported in PBC when coexistent autoimmune hepatitis (AIH)
patients with scleroderma, possibly due to or non‐alcoholic steatohepatitis (NASH) is
decreased bile acid delivery coupled with suspected. Liver biopsy can be useful for
exocrine pancreatic insufficiency, celiac dis- staging the disease for fibrosis and ductope-
ease, and bacterial overgrowth. nia. Finally, it may also be appropriate in the
Hyperlipidemia is another cholestasis‐ presence of other extrahepatic comorbidities.
driven complication, present in up to 85% of
patients at diagnosis. It is characterized by
Biochemical Tests
high levels of total cholesterol, mainly due to
elevation of high‐density lipoproteins (HDLs) The most common scenario is chronic chole-
in the early stages; as disease progresses, HDL stasis in the absence of jaundice, and with
levels fall and low‐density lipoproteins (LDL) transaminases less than three times the
levels rise, due to the pathologic loss of LDL upper limit of normal (ULN). The level of the
receptors and consequent LDL clearance. liver enzymes themselves does not readily
128 Section II Autoimmune Liver Diseases and Their Clinical Correlation
predict severity of the disease, although the techniques are used. Antibodies against the
subsequent biochemical response to therapy major M2 components, i.e. PDC‐E2, 2‐oxo-
is relevant to outcome. glutarate dehydrogenase complex (OGDC),
ALP is more specific for cholestasis than and branched‐chain 2‐oxoacid dehydrogenase
GGT, which can be raised secondary to drug, complex (BCOADC), may then be identified
alcohol or fat. However, when raised, GGT is using MIT3‐based enzyme‐linked immuno-
helpful for confirming that ALP is of biliary sorbent assay (ELISA).
origin (as ALP may also come from bone, Serial monitoring of AMA values is of no
gut, and placenta). clinical value, and AMA titers can fall on
Raised serum transaminases do not neces- treatment over time. AMA has been found in
sarily imply an overlap with AIH, and criteria less than 0.5% of healthy controls, but in up
to diagnose PBC–AIH variant syndrome to 20% of patients with AIH.
have been developed in order to avoid ANAs are present in approximately 30% of
over-diagnosis and over-treatment. Moreover, patients with PBC. The anti‐sp100 (nuclear
it is not rare to see normalization of transami- dots) and anti‐gp210 (perinuclear rims) are
nases together with cholestatic enzymes after specific for PBC (>95%) and are useful in
successful choleretic treatment, revealing how the diagnosis of AMA‐negative patients.
these abnormalities could be part of the Positivity for anti‐gp210 has been associated
biochemical picture of PBC. with aggressive disease and progression to
Another biochemical feature of PBC is liver failure in individuals with PBC.
increased serum IgM concentrations, prob- Other autoantibodies can be found in
ably driven by epigenetic changes, which is PBC, such as anti‐thyroid autoantibodies,
frequent but not specific to PBC. Serum rheumatoid factor, anti‐platelet antibodies,
levels of IgG may also be elevated, even in the anti‐histone antibodies, anti‐centromere anti-
absence of concomitant AIH, more fre- bodies, and anti‐smooth muscles antibodies.
quently in AMA‐negative cases and/or in Their presence is not necessarily related to a
individuals with advanced disease. coexisting autoimmune condition.
Attention should be given to the platelet
count, which is a useful, albeit rough, marker
Liver Biopsy
for advanced liver disease, and can be used to
identify patients who should be considered Liver biopsy for the diagnosis and staging
for esophageal varices screening. of PBC lost favor when it was appreciated
Serum bilirubin is a surrogate marker of that there was an uneven distribution of duct
advanced disease or severe ductopenia. lesions and fibrosis within the liver. The
Raised bilirubin and low albumin levels, widespread availability of non‐invasive mea-
together with prolongation of prothrombin sures of fibrosis means that liver biopsy for
time, are features of advanced cirrhosis. staging of PBC is somewhat obsolete.
Serum bile acids can be elevated but should However, liver biopsy does remain useful
not be determined on a routine basis since in certain settings. The main indications are
their levels do not provide additional value to to confirm the diagnosis of PBC when PBC‐
the diagnostic process. specific antibodies are absent and to confirm
a diagnosis of PBC with AIH features (i.e.
overlap PBC–AIH). Liver biopsy is also use-
Autoantibodies
ful for assessing the relative contribution of
AMA is the hallmark of PBC and is present in each liver injury when a comorbid liver dis-
up to 90–95% of patients. A titer of more than ease such as NAFLD is present. In patients
1 : 40 is considered positive. In up to 10% with inadequate response to UDCA, liver
of patients, AMAs are absent or present biopsy may provide the explanation and
only in low titer when immunofluorescence could undoubtedly inform risk stratification.
Chapter 7 Primary Biliary Cholangitis 129
(a)
Lymphoplasmacytic
portal infiltrate
(b)
Ductular reaction
Figure 7.1 Photomicrograph of stage 2 PBC. (a) Mononuclear inflammatory cells expand the portal tract with
some disruption of the limiting plates (interface hepatitis). The bile duct is surrounded by inflammatory cells.
No evidence of fibrosis (hematoxylin–eosin, ×100). (b) Loss of bile ducts and ductular reaction (cytokeratin 7
staining, ×100).
For example, it may identify a previously ductopenia and fibrosis (Figure 7.1). The
unsuspected variant syndrome, steatohepati- Ludwig and Scheuer scoring systems have
tis, or interface hepatitis of moderate or historically been used to stratify four (1–4)
greater severity. It is also useful in AMA‐ and “stages” of PBC, with stage 4 indicating the
ANA‐specific antibody‐negative cholestatic presence of cirrhosis. In the new system of
patients to indicate an alternative process, Nakanuma, the stage of disease is based on
such as sarcoidosis, small duct primary fibrosis, bile duct loss, and features of chole-
sclerosing cholangitis (PSC), or adult idio- stasis (i.e. deposition of orcein‐positive gran-
pathic ductopenia. ules), whereas the grade of necroinflammatory
PBC is characterized by chronic non‐sup- activity is based on cholangitis and inter-
purative inflammation, which surrounds and face hepatitis. The accumulation of orcein‐
destroys interlobular and septal bile ducts. positive granules occurs evenly across the
Progressive damage of the bile duct leads to PBC liver, which means that staging using
130 Section II Autoimmune Liver Diseases and Their Clinical Correlation
the Nakanuma system is more reliable rejection, ischemic cholangitis, and infectious
regarding sampling variability. cholangitis (e.g. AIDS) should be considered,
although these can be easily excluded based
on the clinical context.
Imaging
1
Probability of remaining free of extensive
0.75
fibrosis or cirrhosis
0.5
0.25
0 2 4 6 8 10
Years
Figure 7.2 Probability of remaining free of extensive fibrosis or cirrhosis in UDCA‐ and placebo‐treated
patients with early‐stage PBC. At 4 years, the probability for UDCA‐treated patients to remain in the early stage
of disease is 76% (95% CI 58–88%), as compared with 29% (95% CI 15–52%) in placebo‐treated patients.
•
Vertical bars indicate 95% confidence interval (CI). (—O—), UDCA; (— —), placebo; (‐ ‐ ‐O‐ ‐ ‐), natural history
(which derives from the study of Locke et al. [6]). Source: Corpechot et al. [7]. Reproduced with permission of
John Wiley and Sons.
0.8
disease stage have been recently developed
to predict prognosis in patients with PBC.
They have been reproduced widely across
0.4
large cohorts and their use is recommended
for all patients after 1 year of UDCA therapy.
However, it is still unclear how to implement
0.0
risk scores in clinical practice.
0 2000 6000 10000 The two most important parameters in eval-
Follow-up (days) uating response to UDCA are ALP and total
bilirubin. Qualitative and quantitative defini-
Figure 7.3 Kaplan–Meier plot for survival according
to Paris I criteria. Kaplan–Meier plots from time‐to‐
tions of UDCA response have been developed,
event analysis in which “failure” is defined as liver based on changes in bilirubin, transaminases
transplant (LT) for PBC, death from PBC‐related liver and ALP after 6–24 months of treatment with
failure, or (for surviving non‐LT recipients) a serum UDCA at 13–15 mg/kg daily. The first criteria
bilirubin level of 100 μmol/l or greater. In the time‐ of UDCA response, proposed in 2006, were
to‐event analysis, the start point is the date of first
presentation and the end point is the date of failure.
the Barcelona criteria, which are the only ones
Surviving non‐LT recipients with bilirubin levels less that consider the change in ALP over
than 100 μmol/l were censored at the date of their treatment. These define response as an ALP
most recent blood tests. The plot shows survival decrease greater than 40% from baseline values
curves for patients who met the Paris I criteria for or normal levels after 1 year of treatment. In
treatment response after 12 months of treatment
with UDCA (blue line) versus those who did not
the Spanish cohort, the observed LT‐free
meet criteria for treatment response (red line). survival was higher than that predicted by the
Source: adapted from Carbone et al. [9]. Mayo model and, for those treated at an early
Chapter 7 Primary Biliary Cholangitis 133
stage of disease, similar to that estimated for ilirubin. A comprehensive list of all available
b
the control population. The Paris I criteria biochemical response criteria is outlined
defined response as an ALP <3 ULN, aspartate in Table 7.2.
aminotransferase (AST) <2 ULN, and normal Such dichotomous criteria are easy‐to‐use
bilirubin after 1 year of UDCA. In the French prognostic tools for clinical practice, but are
cohort, responders had a 10‐year LT‐free simplistic in identifying only two classes of risk
survival of 90% compared with 51% of non‐ (high and low). Furthermore, most of them do
responders. A combination of the Paris‐I cri- not take into account other prognostic vari-
teria and the AST‐to‐platelet ratio index ables, such as markers of disease stage.
(APRI), a surrogate marker of disease stage, A refinement of stratification tools has
applied after 1 year of UDCA improves the been possible thanks to the availability of
predictive power of the UDCA‐response cri- large cohorts, and this has led to the
teria alone. A further evolution were the Paris development of two important continuous
II criteria, which better fit early‐stage disease, scores: the Globe score and the UK‐PBC risk
and define response as ALP <1.5 ULN or score. These scoring systems derive from
AST <1.5 ULN or normal bilirubin. The large multicenter cohorts, are externally vali-
Toronto criteria, developed to predict histo- dated and convey probability of survival/risk
logic stage progression, stratify based on ALP of events on a continuous as opposed to
under treatment (<1.67 ULN), whereas the dichotomous scale, with better predictive
Rotterdam criteria are focused toward liver performances compared with the response
function/stage and include albumin and criteria.
Table 7.2 Biochemical response criteria for risk stratification in UDCA‐treated PBC patients.
ALT, alanine aminotransferase; ALP, alkaline phosphatase; AST, aspartate aminotransferase; ULN, upper limit of
normal.
134 Section II Autoimmune Liver Diseases and Their Clinical Correlation
A drawback of most of these prognostic have relevant prognostic value. The Nakanuma
tools is that they ignore the baseline (pre- system has been shown to reflect liver
treatment) liver biochemistry and what is dysfunction before UDCA treatment and to
defined as treatment response may actually correlate well with the future development of
be just disease severity, with many patients cirrhosis‐related complications. Liver biopsy
having an indolent disease course indepen- may therefore be valuable for risk stratifica-
dently of treatment. tion purposes in PBC. Considering the risk–
More recently, a joint UK/Italian cohort benefit ratio of this invasive procedure, its use
study identified baseline parameters, including may be advocated in patients with poor
younger age, elevated ALP, bilirubin and trans- biochemical response to UDCA.
aminases at diagnosis, the trend of ALP before
therapy, and a longer waiting time before
receiving UDCA, as major predictors of future Treatment
treatment failure to UDCA (Figure 7.4). These
variables were included in a predictive score Medical therapy of PBC targets disease pro-
(UDCA‐response score) that allows risk strati- gression and symptom control. The back-
fication of patients at an early stage, i.e. at bone of treatment for PBC is bile acid. UDCA
diagnosis and provides a basis for exploring has been the only drug available for two
alternative approaches to treatment stratifica- decades, although more recently obeticholic
tion in PBC patients, for example the early acid (OCA), a semisynthetic hydrophobic
identification of patients who would be eligible bile acid analog, has been licensed in patients
for second‐line therapies before waiting for with failure to, or intolerant of, UDCA
treatment failure under UDCA, with potential (Figure 7.5). Several other agents have been
impact on disease course. studied, including immunosuppressants, but
Liver biopsy for diagnostic purposes is no robust evidence of benefit is lacking.
longer recommended, and transient elastog-
raphy can now be used as an alternative option
for detecting severe fibrosis or cirrhosis. Ursodeoxycholic Acid
However, histologic parameters such as UDCA was the first drug to be approved for
fibrotic stage, the degree of lymphocytic inter- use in PBC and remains the first‐line
face hepatitis, and the stage of ductopenia therapy for all patients with PBC [12].
Figure 7.4 Relationship between age at diagnosis and the probability of UDCA response (left) and between
treatment time lag (interval from diagnosis to start of UDCA treatment) and the probability of UDCA response
(right). Source: adapted from Carbone et al. [11].
Chapter 7 Primary Biliary Cholangitis 135
YES
Continue UDCA Biochemical reponse
20–40% NO
Add second-line therapy
Licensed Off-labels
OCA Bezafibrates
Budesonide
UDCA normally accounts for about 4% of Although nearly all patients with PBC have
bile acids, but with pharmacotherapy it some improvement in serum ALP, 20–40%
becomes the predominant bile acid. UDCA have an inadequate biochemical response to
exert its choleretic effect though complex UDCA and remain at risk of disease progres-
posttranscriptional molecular mechanisms sion. Definition of inadequate response to
and the stabilization of the “biliary HCO3– UDCA is based on the level of improvement
umbrella.” UDCA protects cholangiocytes of the LFTs, which can be assessed using
against the cytotoxicity of hydrophobic bile either qualitative definitions (e.g. Barcelona
acids, stimulates the hepatobiliary secretion and Paris criteria) or quantitative scoring
of bile acids, increases the hydrophilicity systems computed from continuous parame-
index of the circulating bile pool, and ters (i.e. Globe and UK‐PBC risk score).
may have immunomodulatory and anti‐ Patients with PBC who have an inadequate
inflammatory effects. UDCA has been response to UDCA or those few (<3%) who
shown to improve the liver biochemistry, are intolerant to UDCA should be candidates
slow down histologic progression, and for second‐line therapies.
improve LT‐free survival.
The optimum dose of UDCA is 13–15 mg/
Obeticholic Acid
kg, which can be given as a single oral daily
dose or in divided doses when tolerability is OCA is approved for patients with an inade-
an issue. Side effects are negligible, mainly an quate response to UDCA or for patients
increase in weight of a few kilograms within unable to tolerate UDCA. OCA is a derivative
the first year and, less frequently, some degree of chenodeoxycholic acid, the naturally
of flatulence, loose stools and hair loss. occurring ligand of the farnesoid X receptor
Concomitant use of bile acid‐binding seques- (FXR), which mediates the synthesis and
trants such as colestyramine or antacids can enterohepatic circulation of bile acids. OCA
alter absorption, and UDCA should be taken is 100 times more potent as an FXR ligand
1 hour before or 4 hours after these agents. compared with chenodeoxycholic acid. In
UDCA is not teratogenic, and even though the liver, FXR activation decreases the
evidence‐based studies of its safety during conversion of cholesterol to bile acids and
pregnancy are lacking, clinicians should not increases the transport of bile acids out of
be worried for patients to continue UDCA hepatocytes. Activation of FXR in the ileum
when used for intrahepatic cholestasis of decreases bile acid reabsorption and
pregnancy (although the drug is not licensed increases expression of fibroblast growth
in pregnancy). UDCA is not effective on factor (FGF)‐19, which acts in the liver to
fatigue, pruritus or bone disease. further decrease bile acid synthesis. OCA
136 Section II Autoimmune Liver Diseases and Their Clinical Correlation
may also have antifibrotic properties and 6 months if the response is judged inade-
may improve portal hypertension. quate by the clinician, provided that the drug
OCA has been evaluated in two Phase II is tolerated. An agreed definition of response
trials and one Phase III trial. The Phase III to OCA has not been produced yet and may
study [13] evaluated the effects of 12 months vary across different countries. The use of
of treatment with OCA in patients with PBC OCA in advanced liver disease needs cau-
who showed an inadequate response to tion. In patients with Child–Pugh score B
UDCA, defined by an ALP 1.67 ULN or and C, a careful risk–benefit evaluation
serum bilirubin above ULN but <2 ULN, or should be performed. The dose should be
were intolerant of UDCA. Patients were reduced to 5 mg per week, to a maximum of
randomized in a double‐blind fashion (1 : 1 : 1) 10 mg twice weekly. Caution is suggested also
to receive placebo daily, OCA 10 mg/day, or in patients with Child–Pugh score A and
OCA 5 mg/day with titration to 10 mg after 6 signs of portal hypertension.
months based on response and tolerability. Exacerbation of pruritus during OCA
The primary end point measured after therapy is dose‐dependent so dose titration is
12 months of treatment was the combination therefore suggested, and the pruritus usually
of ALP <1.67 ULN, with a reduction of at improves with time. A drug holiday for a few
least 15% from baseline, and a normal total weeks and restarting at a lower dose (e.g. 5 mg
bilirubin. More than 90% of the 216 patients every other day) can be helpful. Rifampicin is
with PBC received UDCA as background preferred to colestyramine to avoid reduced
therapy. The primary end point was met by absorption. Modifications in lipid profile are
47 and 46% of patients in the 10 mg and titra- also possible side effects, most frequently
tion groups, respectively, compared with 10% mild reductions in HDL levels, but no data
in the placebo group. There was no change in are available about whether this is associated
non‐invasive measures of fibrosis. Pruritus with a higher likelihood of future cardiovas-
was more common in the OCA group and cular events.
was reported in 68 and 56% of patients in the
10 mg and 5–10 mg OCA‐treatment arms,
Fibric Acid Derivatives
respectively, compared with 38% in the
placebo arm. Discontinuation owing to pru- Fibric acid derivatives, or fibrates, are ago-
ritus occurred in 10% of the 10 mg OCA‐ nists of the peroxisome proliferator‐activated
treated patients compared with 1% in the receptor (PPAR), a nuclear receptor involved
titration group and none in the placebo group. in several metabolic pathways. Fibrates are
After the initial 12 months of the trial, patients licensed for the treatment of hypertriglyceri-
were given the option to continue in an open‐ demia. They also exert potent anticholestatic
label long‐term extension starting at 5 mg effects. Among the fibrates, bezafibrate and
OCA and increasing to 10 mg OCA as toler- fenofibrate, both PPARa selective agonists,
ated. The majority of patients (91%) entered have been extensively studied as therapeutic
the extension phase and results to date agents because of their potential ability to
support ongoing efficacy of OCA through 2 decrease bile acid synthesis and bile acid‐
years. Longer‐term efficacy of OCA and gen- related hepatic inflammation. Numerous
eralisability to the patient population as a small pilot studies and case reports have
whole needs confirmation in prospective shown that fibrates, including fenofibrate in
follow‐up studies. Survival benefit has yet to the USA and bezafibrate in Europe and Japan,
be demonstrated and, for that purpose, a improve liver biochemistries, liver stiffness
long‐term Phase IV randomized trial is cur- measurements, and pruritus in patients with
rently ongoing. PBC. These data have been further sup-
OCA should be started at an initial dose of ported by the recent Phase III randomized
5 mg/day, and increased to 10 mg/day at placebo‐controlled Bezurso trial. Patients
Chapter 7 Primary Biliary Cholangitis 137
with an inadequate response to UDCA by liver histology, the primary end point of the
Paris‐2 criteria were treated with bezafibrate study, were observed. Full trial data are
400 mg/day or placebo for 2 years. The pri- awaited. Caution must be applied when using
mary end point, i.e. normal total bilirubin, budesonide in patients with cirrhosis and
ALP, AST, alanine aminotransferase (ALT), portosystemic shunts. Evidence is lacking as
albumin, and prothrombin time at 2 years, to how to approach long‐term maintenance
was reached more frequently in the bezafi- therapy in overlap patients stepping down
brate group than in the placebo group (30 vs. from corticosteroids. A reasonable approach
0%, respectively). Whether biochemical is to extrapolate from management regimens
improvements translate to significant decrea for pure AIH with the use of antimetabolites.
ses in liver‐related death or need for LT has
not yet been shown.
Liver Transplantation
The safety of fibrates is a matter of con-
cern. Hepatotoxicity is a known side effect Rates of LT for PBC have declined over the
though reversible, reported in 6% of cases in last decades. Indications are similar to those
the Bezurso trial. Elevated creatinine, likely for other chronic liver diseases, i.e. MELD
related to increased release by muscle rather >15 or bilirubin values steadily rising above
than a real worsening in kidney function, is 3–5 mg/dl (40–85 μmol/l). An exceptional
also reversible. Up to 20% of patients com- indication is intractable pruritus, while
plain of cramps and musculoskeletal pain. fatigue is not improved by LT and therefore is
Taken together, these adverse effects repre- not an indication. Outcomes are favorable,
sent a concern for their use in some patients. with 5‐year patient survival rates better than
Although fibrates are currently still used off‐ for most other indications for LT (80–85%).
label as a second‐line treatment in PBC, PBC recurs after LT, with rates that
their efficacy and ability to improve pruritus increase the longer the follow‐up (20–30%
will likely lead to formal approval for this within 10 years, around 50% within 20 years)
indication. [14]. Since AMAs typically persist after LT,
liver biochemistry could be normal, and
abnormal ALP could be due to other post‐
Budesonide
transplant liver complications; the final
Budesonide is currently used as an off‐label diagnosis of PBC recurrence requires his-
medication in PBC. In PBC patients showing tology. However, recurrent PBC infrequently
interface hepatitis on liver biopsy, some leads to graft loss and current evidence
groups have demonstrated the efficacy of does not suggest an impact on graft or
budesonide in improving liver histology and patient survival after LT. The use of UDCA
biochemistry when used in combination with as a preventive measure in this setting is sup-
UDCA. In 1999 a randomized placebo‐con- ported by non‐randomized studies showing
trolled trial showed the efficacy of budesonide lower rate of PBC recurrence.
9 mg on liver biochemistry and liver histology
at 2 years follow‐up. These data have been
Symptom Management
recently challenged by a yet unpublished
Phase III, randomized, double‐blind, placebo‐ All patients should be evaluated for the
controlled trial (NCT00746486), which presence of symptoms, particularly fatigue
showed improved LFTs in the budesonide and pruritus, bearing in mind that the severity
arm (9 mg/day of budesonide as add‐on of symptoms does not correlate with stage of
therapy), with a mean change in ALP disease. Pruritus is a troublesome issue for
from baseline of −94.5% and −8.9% in the many but not all patients with PBC. If skin is
budesonide group and placebo group, respec- dry, moisturizing agents should be used, and
tively. However, no significant changes in patients should be counseled to avoid hot
138 Section II Autoimmune Liver Diseases and Their Clinical Correlation
baths or showers. Clinicians should exclude Others agents include the antihistamines,
secondary causes, like allergies, hypereosino- which sometimes have a non‐specific anti-
philic conditions, or bile duct obstruction. pruritic effect; this may be due to their seda-
First‐line treatment is colestyramine, a tive properties but they are not recommended
non‐absorbable resin and bile sequestrant. as specific therapy.
Even though evidence supporting its use is Physical approaches, such as nasobiliary
weak, clinical experience is solid. The starting drainage, molecular absorbance recirculating
dosage is usually one sachet (4 g) once or system (MARS), and ultraviolet light therapy
twice a day but a total daily dose of 16 g or are experimental approaches with small series
greater may be required. Clinicians must showing benefit. LT is a rare indication for
carefully respect the pharmacokinetics of the intractable cholestatic pruritus.
drug: patients should take colestyramine Fatigue is a non‐specific but often reported
2 hours after or 4 hours before other medica- symptom in PBC, and represents an unmet
tions to avoid preventing their absorption, need since there are no licensed therapies.
and 20 minutes before meals. Tolerability is Since fatigue is not related to the stage of
often an issue, with side effects including disease, and potentially can persist after LT,
bloating and constipation. Colestyramine is its management proceeds in parallel with
best taken with food, and is most effective management of the disease itself. A struc-
when taken with the first meal of the day. tured approach to management – quanti-
Second‐line treatment is rifampicin, which fying fatigue and its impacts (through the
probably acts through its pregnane X use of disease‐specific tools such as the
receptor (PXR) agonist function. There are PBC‐40 quality‐of‐life measures), address-
concerns over potential side effects, including ing contributing and exacerbating factors,
hepatotoxicity and hemolysis and increased and supporting patients to cope with its
international normalized ratio (INR). impact – is effective. Since fatigue in PBC is
Therefore, regular blood tests should be per- not specific, it is important to identify other
formed. Rifampicin is also a potent enzyme disease processes and therapies which may
inducer and caution is required when the be contributing to fatigue, such as hypothy-
drug is given with other medications (such as roidism, depression, anemia, sleep apnea,
warfarin). Also, appropriate consideration and overtreatment with antihypertensive
should be given to balancing the benefits drugs. Drugs such as coenzyme Q and ritux-
against the risks of antimicrobial resistance. imab have been shown to be ineffective. A
Third‐line treatment is oral naltrexone, an graded program of exercise helps some
opiate antagonist, which reduces the sensa- individuals.
tion of itching. It should be started at a Patients with PBC often complain of dry
low dose (12.5 mg/day) and titrated slowly to eyes (xerophthalmia) and/or dry mouth
avoid withdrawal‐like symptoms in the (xerostomia), and sometimes dysphagia and
first few days of treatment. Some patients vaginal dryness, but this is rarely associated
require an intravenous induction stage with with an actual diagnosis of primary Sjögren
naloxone. Long‐term tolerability can be an syndrome. When symptoms are mild,
issue, with many patients having ongoing artificial eye drops and saliva substitutes are
opiate withdrawal‐like reactions or reduced usually sufficient. Patients should be advised
threshold to pain. to undergo regular dental check‐ups, as
Fourth‐line treatment involves the selective they are at risk of caries and oral candidi-
serotonin reuptake inhibitors and gabapen- asis. If xerophthalmia is moderate to severe,
tin, which are used empirically in the an ophthalmologist’s review is suggested.
management of cholestatic itch, typically in Cholinergic agents such as pilocarpine are
patients with pruritus unresponsive to other effective in patients symptomatic despite
agents. first‐line treatments, but usually have
Chapter 7 Primary Biliary Cholangitis 139
Abbreviation: b.i.d., twice a day; q.d., once a day; q.i.d., 4 times a day
Figure 7.6 Practical approach to drug management of most relevant symptoms in PBC.
osteoporosis (generally with a T‐score lower portal hypertension can occur in PBC and
than −1.5). Caution is required when using the possibility of its presence should be
oral bisphosphonates in patients with var- considered in all PBC patients with a gas-
ices because of the risk of superficial erosion trointestinal bleed. The management of
and enhanced bleeding risk. Intravenous gastroesophageal varices and variceal hem-
bisphosphonates can be used if there is orrhage in patients with PBC follows the
clinical concern. Baveno‐VI guidelines.
HCC
Advanced Liver Disease
As with all forms of cirrhosis, patients with
Varices PBC may develop HCC and should therefore
All PBC patients with suspected portal undergo regular screening with cross‐sec-
hypertension should be screened for gas- tional imaging. Male sex and UDCA non‐
troesophageal varices. Patients with PBC response are risk factors for HCC development
can develop varices even in the absence of in PBC. Men without advanced disease who
established cirrhosis, in association with are UDCA non‐responders are at higher risk
nodular regenerative hyperplasia, although of developing HCC than women with cir-
this is relatively uncommon. Non‐cirrhotic rhosis who respond to UDCA.
Chapter 7 Primary Biliary Cholangitis 141
Useful Websites
www.uk‐pbc.com
www.globalpbc.com
www.pbcfoundation.org.uk
References
1 Sherlock, S. (1959). Primary biliary 8 Metcalf, J.V., Mitchison, H.C., Palmer, J.M.
cirrhosis. Chronic intrahepatic obstructive et al. (1996). Natural history of early primary
jaundice. Gastroenterology 31: 574. biliary cirrhosis. Lancet 348: 1399–1402.
2 Boonstra, K., Beuers, U., and Ponsioen, C.Y. 9 Carbone, M., Mells, G.F., Pells, G. et al.
(2012). Epidemiology of primary sclerosing (2013). Sex and age are determinants of the
cholangitis and primary biliary cirrhosis: a clinical phenotype of primary biliary
systematic review. J. Hepatol. 56: cirrhosis and response to ursode
1181–1188. oxycholic acid. Gastroenterology 144 (3):
3 Gershwin, M.E., Mackay, I.R., Sturgess, A., 560–569.e7.
and Coppel, R.L. (1987). Identification and 10 Trivedi, P.J., Corpechot, C., Pares, A., and
specificity of a cDNA encoding the 70 kd Hirschfield, G.M. (2016). Risk stratification
mitochondrial antigen recognized in in autoimmune cholestatic liver diseases:
primary biliary cirrhosis. J. Immunol. 138 opportunities for clinicians and trialists.
(10): 3525–3531. Hepatology 63 (2): 644–659.
4 Joplin, R., Lindsay, J.G., Johnson, 11 Carbone, M., Nardi, A., Flack, S. et al.
G.D. et al. (1992). Membrane (2018). Pretreatment prediction of response
dihydrolipoamide acetyltransferase (E2) to ursodeoxycholic acid in primary biliary
on human biliary epithelial cells in cholangitis: development and validation of
primary biliary cirrhosis. Lancet 339 the UDCA Response Score. Lancet
(8785): 93–94. Gastroenterol Hepatol 3 (9): 626–634.
5 Hirschfield, G.M., Beuers, U., Corpechot, C. 12 Poupon, R.E., Poupon, R., and Balkau, B.
et al. (2017). The diagnosis and management (1994). Ursodiol for the long‐term
of patients with primary biliary cholangitis. treatment of primary biliary cirrhosis. The
J. Hepatol. 67 (1): 145–172. UDCA‐PBC Study Group. N. Engl. J. Med.
6 Locke, G.R. III, Therneau, T.M., Ludwig, J. 330 (19): 1342–1347.
et al. (1996). Time course of histological 13 Nevens, F., Andreone, P., Mazzella, G. et al.
progression in primary biliary cirrhosis. (2016). A placebo‐controlled trial of
Hepatology 23 (1): 52–56. obeticholic acid in primary biliary cholangitis.
7 Corpechot, C., Carrat, F., Bonnand, A.M. N. Engl. J. Med. 375 (7): 631–643.
et al. (2000). The effect of ursodeoxycholic 14 Neuberger, J., Portmann, B., Macdougall,
acid therapy on liver fibrosis progression in B.R. et al. (1982). Recurrence of primary
primary biliary cirrhosis. Hepatology 32 (6): biliary cirrhosis after liver transplantation.
1196–1199. N. Engl. J. Med. 306 (1): 1–4.
143
Abstract
This chapter provides guidance as regards patient management and clinical followup, emphasizing
the peculiarities of pediatric compared to adult presentations of the disease. Adult primary
sclerosing cholangitis (PSC) is a rare disease but occurs in 5–10% of patients with inflammatory
bowel disease (IBD). Cholangiocarcinoma and colorectal cancer are dreaded complications in PSC
and major causes of death; alertness to alarm symptoms and adherence to surveillance strategies
may improve survival. There is no medical therapy with proven effect to halt disease progression in
adult PSC, but ursodeoxycholic acid is often prescribed and immunosuppression may be indicated
in patients with additional features of autoimmune hepatitis. Full colonoscopy with biopsies is
recommended at diagnosis of PSC in all patients without known IBD in order to diagnose subclin
ical colitis. PSC has been denominated a premalignant condition by some, due to the firmly
established and extremely high risk of hepatobiliary as well as colorectal cancer.
Keywords autoimmune hepatitis; cholangiocarcinoma; colorectal cancer; inflammatory bowel
disease; medical therapy; primary sclerosing cholangitis; surveillance strategies
Key Points
●● Adult primary sclerosing cholangitis major causes of death; alertness to alarm
(PSC) is a rare disease but occurs in symptoms and adherence to surveillance
5–10% of patients with inflammatory strategies may improve survival.
bowel disease. ●● There is no medical therapy with proven
●● Cholangiocarcinoma and colorectal can effect to halt disease progression in adult
cer are dreaded complications in PSC and PSC, but ursodeoxycholic acid is often
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
144 Section II Autoimmune Liver Diseases and Their Clinical Correlation
Thus PSC is a complex and heterogeneous PSC is strongly associated with IBD; how
disease, with important knowledge gaps and ever, the frequency of IBD in PSC varies
unmet needs contributing to the challenges across the world, with a similar geographic
in patient care and follow‐up. There are gradient as for the disease itself. The high
important differences between adult PSC est prevalence rates are found in northern
and juvenile sclerosing cholangitis. This Europe and the USA where 60–80% of
chapter provides guidance as regards patient patients have concurrent IBD. Lower
management and clinical follow‐up, empha frequencies are observed in southern
sizing the peculiarities of pediatric compared Europe (40–60%) and Asia (reports of 34%
to adult presentations of the disease. in Japan, 50% in India), possibly due to a
larger fraction of IBD‐unrelated sclerosing
cholangitis (e.g. IgG4‐associated sclerosing
Definition cholangitis) in these areas. In support of the
latter, current data from Japan suggest that
PSC is a chronic and progressive immune‐ among patients diagnosed with PSC there
associated cholestatic liver disease of is a group of elderly patients without IBD
unknown etiology characterized by inflam that is clinically distinct from younger
mation, fibrosis, and multifocal strictures of patients (which show an IBD prevalence
the bile ducts. rate of 60%).
The reported prevalence for PSC among
patients with IBD varies and is probably less
than 10%. PSC was seen more frequently in
Adult PSC extensive rather than distal colitis in a large
population of ulcerative colitis (UC) patients
Epidemiology
from Sweden. There is a seemingly stronger
PSC is a rare disease according to formal def association for PSC with UC compared with
initions, with the highest prevalence and Crohn’s colitis (80% of IBD in PSC is classi
incidence rates in northern Europe and the fied as UC). However, PSC patients with
USA. Population‐based studies are scarce, Crohn’s disease are more likely to be female
but studies from northern Europe and the with small‐duct PSC and a milder (hepato
USA report prevalence rates of about 1 in biliary) disease course and this group may
10 000 and incidence rates of 0.4–2.0 per be underdiagnosed at present. This is sup
100 000 per year. There is a geographic gra ported by a study of a large cohort of IBD
dient with 10–100‐fold lower prevalence patients from Norway where MRC surveil
rates in southern Europe and Asia. Several lance revealed findings consistent with PSC
reports support a rising incidence of PSC. in 7.8%, of which only one‐third were recog
This rise is thought to be real rather than an nized clinically and with similar frequencies
effect of improved diagnostics alone as the of PSC in Crohn’s disease and UC [7].
clinical features of newly diagnosed patients Current data indicate that PSC may be
have not changed over time. Earlier diagnosis occurring equally frequently in men and
of PSC as magnetic resonance cholangiog women but runs a clinically more quiescent
raphy (MRC) has improved and replaced course in females, explaining the tradition
invasive diagnostic methods does not seem ally reported 2 : 1 male to female ratio. Data
to explain fully the increased incidence. from a large multicenter study of 7119 PSC
The parallel increase in incidence in other patients support this, showing significantly
autoimmune and idiopathic inflammatory better outcomes for female patients [8]. In
diseases has fueled hypotheses of a common the largest pediatric series to date, the male
environmental cause driving the increased to female ratio was 1.5 : 2, and disease pro
incidences. gression was similar in boys and girls [6].
146 Section II Autoimmune Liver Diseases and Their Clinical Correlation
Diagnosis
contrast‐enhanced liver magnetic resonance
The diagnosis is made based on characteristic imaging (MRI) within 6 months of diagnosis if
cholangiographic findings, featuring irregular not undertaken as part of the initial work‐up,
bile ducts with multiple alternating strictures with the aim of early detection of cholangio
and dilatations in combination with elevated carcinoma; however, any effect on outcome
cholestatic liver enzymes, i.e. alkaline phos has not been proven and the risk of side
phatase (ALP) and/or gamma‐glutamyltrans effects is debated. ERC is generally reserved
peptidase (GGT). The biliary changes are for cases where intervention (stenting, dilata
most often found both intrahepatically and tion) is required [9].
extrahepatically (50–80%), but may be intra Known causes of secondary cholangitis
hepatic only (<25%) or, more rarely, only should be excluded, for example IgG4‐asso
extrahepatic (<5%) (Figure 8.1). MRC is the ciated cholangitis (IAC), and comorbid liver
imaging modality of choice, with diagnostic diseases identified (viral or AIH, alcohol)
accuracy reaching that of endoscopic retro based on detailed history, extensive blood
grade cholangiography (ERC), with sensi work‐up, and ultrasound. Liver biopsy is not
tivity and specificity of 0.86 and 0.94, part of standard diagnostic work‐up but
respectively, at expert centers. If the initial may be useful in suspected variant PSC with
MRC is inconclusive, showing equivocal concomitant features of autoimmune
minimal changes of the biliary tree, second hepatitis (PSC/AIH; previously denomi
opinion evaluation by an expert center either nated “AIH overlap syndrome” [10]; suspect
of the original images or a repeat MRC may if elevated transaminases five times upper
yield a definitive diagnosis. Some experts in limit of normal, ULN) or small‐duct disease
the field recommend MRC combined with (when a strong suspicion of PSC remains
(a)
(b)
Figure 8.1 (a) Endoscopic retrograde cholangiography (ERC) in a patient with primary sclerosing cholangitis
(PSC) showing multifocal intrahepatic bile duct strictures with intervening dilatations and narrowing of the
entire extrahepatic bile duct (only catheter visible). (b) Magnetic resonance cholangiography (MRC) in a PSC
patient featuring multifocal intrahepatic bile duct strictures with intervening dilatations and a dilated irregular
extrahepatic bile duct. Source: courtesy of Gunter Kemmerich.
Chapter 8 Primary Sclerosing Cholangitis 147
the first sign of infection. Schemes involving the various aspects of the current model for
prescription of rotating antibiotics lack evi PSC pathogenesis such as bile acids (nor‐
dence and risk promoting antibiotic resis UDCA, obeticholic acid), the gut microbiota
tance and cannot be recommended. A and gut–liver axis (vancomycin in particular;
different gut microbiota has been demon fecal transplantation), inflammation (vedoli
strated in PSC compared to healthy controls zumab), and fibrosis (simtuzumab/anti‐
or UC patients and opens up intriguing pos LOXL2; bezafibrate/PPAR agonists). At the
sibilities of using antibiotics, probiotics or moment, drugs aiming at modulating chole
fecal transplantation to manipulate the gut stasis and antibiotics seem most promising.
microbiota with the aim of treating the bil
iary disease. Vancomycin has attracted Bile Acid Therapy
particular attention; however, data are cur Nor‐UDCA is a side‐chain‐shortened
rently insufficient for any form of manipula derivative of UDCA which unlike UDCA is
tion of the gut microbiota in the treatment of secreted unconjugated into the bile and
PSC and further studies are awaited. inhibits toxic bile formation; Phase II trial
results were promising (showing dose‐
Immunosuppression responsive ALP reduction) and a Phase III
There is no indication for the use of study is ongoing. The farnesoid X receptor
immunosuppressant agents to treat regular
(FXR) is one of the master regulators of the
adult PSC. Studies have shown no effect of protective pathways in hepatocyte chole
immunosuppression on transplant‐free sur
stasis, partly inducing its effects through the
vival whereas serious side effects appeared. downregulation of bile salt production.
Preliminary results from retrospective studies Obeticholic acid is an FXR agonist that has
investigating the effect of biological treatment been conditionally approved in PBC with
[tumor necrosis factor (TNF)‐α inhibitors] on promising results from a Phase II study in
PSC are so far disappointing, but data are PSC (safe and showed dose–response effect
scarce. Given the fact that PSC is an autoim on ALP reduction); however, concerns
mune disease as defined genetically and by remain in PSC for potential deleterious
other features, this is somewhat of a paradox. effects of FXR agonism in the presence of
However, in patients with features of AIH (i.e. cholestasis and increased pruritus has been
transaminases >5× ULN, IgG >2× ULN, and reported as a side effect in some patients.
typical or compatible histologic findings Recently, a small retrospective study reported
dependent on the presence or absence of auto biochemical improvement and reduced pru
antibodies, as defined in the Diagnosis section), ritus in patients treated by a combination of
immunosuppressive therapy may be beneficial UDCA and fenofibrate or bezafibrate, sug
and is recommended, with prescription follow gesting a role for PPAR agonists. Like
ing standard guidelines for the treatment of obeticholic acid, fibrates downregulate bile
AIH. The evidence for such treatment is weak acid synthesis, stimulate canalicular secre
and the response to treatment is variable and tion, and upregulate bile acid detoxification.
generally less pronounced than in AIH without A Phase II clinical trial is being planned.
a diagnosis of PSC; hence, the risk of side effects
(e.g. osteopenia) should be recognized and Microbiota Modulators
weighed against treatment effect during close Given the increasing evidence for a role of the
monitoring. intestinal microbiota on PSC pathogenesis,
antibiotic therapy appears attractive. Vancom
Future Drugs ycin is perhaps the best‐studied antibiotic in
Based on new insights into the pathogenesis PSC, with promising results in smaller studies
of PSC, there has been a surge of interest in in adults and children; however, the evidence is
clinical trials in PSC. Several new drugs are currently insufficient to support treatment
being investigated. Approaches vary, t argeting recommendations. A larger clinical trial is
154 Section II Autoimmune Liver Diseases and Their Clinical Correlation
ongoing. Another potential antibiotic therapy MRI/MRC should be performed prior to ERC
for PSC, although with insufficient evidence, is to confirm the indication and identify any
minocycline, whereas rifaximin did not show focal lesions. As further discussed in the EASL
any effect in a clinical trial. Clinical Practice Guidelines describing the
role of endoscopy in PSC [9], the optimal
Anti‐inflammatory Treatment interventional approach has not been resolved
Paradoxically, although PSC is an immune‐ and the choice between balloon dilatation
associated disease, immunosuppression has with or without short‐term stenting for 1–2
not proved effective in the past. The focus weeks, or short‐term stenting only is left to
now is mainly on preventing the potential the operator. Studies including the interna
aberrant “homing” of lymphocytes from the tional multicenter prospective DILSTENT
gut to the liver. Timolumab (BTT1023), an study have shown similar outcome for the two
anti‐VAP‐1 antibody, reduces the homing of alternative approaches, although some data
gut‐activated lymphocytes to the biliary tree suggest a higher rate of procedure‐related
and is currently being investigated in PSC. complications associated with stenting [16].
Current evidence indicates that biological Different stenting durations have not been
therapies used successfully for the treatment formally compared in PSC but short‐term
of IBD are safe but not effective treatments stenting is favored because stents tend to clog
for PSC (including adalimumab, infliximab rapidly in PSC and the efficacy reported in
and vedolizumab). studies of short‐term (1–2 weeks) versus stan
dard (8–12 weeks) stenting was similar.
Antibiotic Therapy Sphincterotomy, although not generally
Results from a Phase II clinical trial investi advised, is recommended following difficult
gating simtuzumab, an anti‐LOXL2 agent tar cannulation or if repeated procedures are
geting fibrosis, disappointingly did not show anticipated to reduce the risk of complications
any effect on event‐free survival, histologic in future procedures. PSC patients carry a
fibrosis scores, or hepatic collagen content. slightly higher risk of complications (cholangi
tis, post‐ERC pancreatitis) related to ERC
Better tools for risk stratification, improving compared to other patients and preprocedure
patient selection for clinical trials, and administration of prophylactic parenteral
surrogate end points to evaluate treatment antibiotics and rectal non‐steroidal anti‐
effect would facilitate the establishment of inflammatory drugs (e.g. 100 mg diclofenac or
effective therapy. It is possible that indometacin) is recommended. In case of a
combination therapy, attacking two or several high risk of post‐ERC pancreatitis, a prophy
components of pathogenesis, may yield lactic pancreatic stent should be considered.
improved therapeutic effect, or that better Endoscopic treatment of clinically significant
characterization or stratification of patients strictures in PSC has been demonstrated to
may direct tailored therapy targeting the dis improve symptoms but the evidence for an
ease‐driving factor in the individual patient. impact on prognosis is less clear, although
some studies have shown increased trans
Endoscopic Treatment plant‐free survival. ERC with brush sampling
Endoscopic treatment plays an important role plays an important role in the surveillance and
in PSC and should be considered whenever a diagnosis of cholangiocarcinoma in PSC.
patient with PSC shows increasing symptoms
(jaundice, pruritus, cholangitis), rapid increase
Surveillance for Malignancy
of cholestatic liver enzymes, or a new or pro
gressing dominant stricture accompanied by Malignancy is a major cause of death in PSC.
appropriate clinical findings [9]. Concomitant Several recommendations regarding surveil
ductal sampling with brush cytology should lance in PSC target the increased risk of hep
always be performed. Most often, combined atobiliary and colorectal cancer and are
Chapter 8 Primary Sclerosing Cholangitis 155
aimed at the early detection of relevant can sensitivity as well as specificity. Combined
cers, focusing on measures with proof of MRI/MRC has the highest sensitivity (89%)
effect on outcome. Complications of chole and specificity (75%) for detection of particu
stasis justify other forms of surveillance. larly small cholangiocarcinomas. Cases with
suspicious findings should be followed up by
Colorectal Cancer: Colonoscopy endoscopic retrograde cholangiopancreatog
The risk of colorectal cancer is fivefold higher raphy (ERCP) with brush cytology. In a posi
than in IBD without PSC. Colonoscopy with tion paper, an international expert panel of
biopsies should be performed at diagnosis of hepatologists and radiologists suggest that
PSC in all patients without a former diagnosis contrast MRI/MRC should be performed
of IBD in order to diagnose subclinical colitis. within 6 months of diagnosis of PSC if not
Surveillance by colonoscopy should be per taken at the time of diagnosis as surveillance
formed regularly (annually or biannually) from for cholangiocarcinoma, as the risk is high in
the time of diagnosis of PSC, including in the first year following diagnosis. ERC with
young adults. Adherence to surveillance has brush cytology should be performed if a new
been demonstrated to increase survival. or progressing dominant stenosis is detected.
100 000 children [4]. Epidemiologic data are Approximately 50% of the patients enrolled
hampered by the limited number of pub had alterations of the bile ducts characteristic
lished series of pediatric sclerosing cholangi of sclerosing cholangitis at presentation and
tis, most being retrospective, single center were diagnosed as having ASC. One‐quarter
and including small patient numbers [4]. of these had no histologic features suggesting
Reported incidences of the various clinical bile duct involvement despite abnormal chol
forms of sclerosing cholangitis differ, angiograms. Serologically, ASC patients were
reflecting different study designs, patterns of indistinguishable from AIH type 1. In contrast
referral and diagnostic protocols. In all retro to AIH, which had a clear female preponder
spective series cholangiographic studies were ance, ASC affected a similar proportion of
prompted by biochemical and/or histologic boys and girls. Surveillance endoscopy to
features of cholestatic disease. Boys are investigate possible IBD was performed in all
reported to be more affected than girls, cases, showing IBD in 45% of ASC and 20% of
20–40% having only intrahepatic cholangi AIH type 1. At presentation, immunologic
opathy. As for adults, there is a strong IBD and biochemical indices did not help in dis
association (60–90% of cases), with over two‐ criminating between AIH and ASC, although
thirds of patients having UC, the others the ALP/AST ratio was significantly higher in
Crohn’s disease or indeterminate colitis. IBD ASC. Biochemical response to predniso
can precede the diagnosis of liver disease by lone with or without azathioprine was similar
many years, be diagnosed at the same time, in AIH and ASC, leading to normalization of
or develop during follow‐up. transaminase levels, although bile duct
damage progressed over the years in half of
the ASC cases.
Autoimmune Sclerosing Cholangitis
In view of the postulated autoimmune
All retrospective studies recognize a form of pathogenesis of adult PSC, the question
sclerosing cholangitis with strong autoim arises as to whether ASC is the early manifes
mune features. However, in some this tation of what will eventually become adult
condition is reported to respond favorably to PSC. Similarities and differences between
immunosuppression, having a better prog PSC and ASC are summarized in Table 8.1.
nosis than PSC; in others the prognosis of
AIH/PSC overlap syndrome is reported to be
Diagnosis in Children
severe and not ameliorated by immunosup
pressive treatment or is similar to that of PSC The diagnosis of sclerosing cholangitis in
irrespective of treatment [4]. Major limita children/adolescents is based on the finding
tions of retrospective studies are uneven of abnormal bile ducts on cholangiography
diagnostic protocols and lack of accurate often, but not always, in association with a
information on IBD treatment before the cholestatic liver profile. Conditions associ
diagnosis of sclerosing cholangitis, as immu ated with bile duct disease or secondary
nosuppression for IBD might affect the liver sclerosing cholangitis should be excluded by
disease. appropriate investigations. Strictly “primary”
Only one study has been conducted pro sclerosing cholangitis is rare [3].
spectively with the aim of establishing the The diagnosis of ASC is based on abnormal
relative incidence of AIH and ASC among cholangiographic findings by MRC at pre
children with liver disease and positive auto sentation in children that fulfill the diagnosis
immune serology, by performing cholangio of AIH. Both intrahepatic and extrahepatic
grams at disease onset, irrespective of bile ducts are most often involved, with iso
biochemical or histologic evidence of chole lated involvement of the intrahepatic bile
static disease. Patients were recruited over ducts in one‐third of the cases. A liver biopsy
16 years [3] and are currently being followed. is mandatory in the diagnostic work‐up of
Chapter 8 Primary Sclerosing Cholangitis 157
Table 8.1 Differences between adult primary sclerosing cholangitis (PSC) and juvenile autoimmune sclerosing
cholangitis (ASC).
ANA, anti‐nuclear antibody; IBD, inflammatory bowel disease; SMA, smooth muscle antibody.
Figure 8.3 Dense portal tract inflammatory infiltrate comprising lymphocyte and plasma cells and invading
the surrounding parenchyma (interface hepatitis) in a child with autoimmune sclerosing cholangitis. The arrow
shows bile duct damage. Hematoxylin and eosin stain.
ASC. The classical histologic changes of hepatitis, similar to what is described in AIH
periductular concentric fibrosis (often (Figure 8.3). Other reported histologic
referred to as “onion skin fibrosis”) are only features include fibro‐obliterative cholangitis
rarely seen in children, probably because and primary ductular involvement [6].
they result from long‐standing disease. Most The International Autoimmune Hepatitis
commonly the histology shows dense mono Group scoring systems, devised for the diag
nuclear inflammatory infiltrate of the portal nosis of AIH in adults, do not discriminate
tract with several plasma cells and interface between AIH and ASC, as they do not include
158 Section II Autoimmune Liver Diseases and Their Clinical Correlation
cholangiographic studies at disease onset. A is higher than in AIH. The immune serologic
position paper of the European Society of profile is most often indistinguishable from
Paediatric Gastroenterology, Hepatology AIH type 1; the great majority of patients are
and Nutrition (ESPGHAN) Hepatology positive for anti‐nuclear and/or anti‐smooth
Committee has recently proposed a diag muscle antibodies, whereas the presence of
nostic scoring system for juvenile autoim anti‐liver kidney microsomal type 1 antibody
mune liver diseases that awaits validation is rare [3]. Perinuclear anti‐neutrophil
[18]. Current EASL guidelines recommend nuclear antibody is more frequently positive
that every child or adolescent diagnosed with in ASC than in AIH, being found in three‐
AIH must undergo MRC in order to actively quarters of cases [3]. Data from the UK show
screen for ASC [19]. All children diagnosed that ASC is associated with HLA‐DRB1*1301
with ASC or PSC should undergo upper gas genotype, at variance with AIH type 1, which
trointestinal endoscopy and full colonoscopy is associated with DRB1*0301, and with AIH
in order to screen for concomitant IBD. type 2, associated with DRB1*0701 [4]. One‐
Juvenile small‐duct sclerosing cholangitis quarter of patients with abnormal cholangio
is a poorly characterized entity. Reliable diag gram do not have histologic evidence of bile
nostic criteria for this condition are lacking. duct injury. Inflammation of the portal tracts
Moreover, genetic defects such as MDR3 is frequent, and interface hepatitis, the histo
deficiency are associated with small‐duct logic hallmark of AIH, is as frequent in ASC
sclerosing cholangitis, and the role of auto as in AIH. Histologically, AIH is associated
immune small‐duct disease is still unknown. with more active inflammation and a higher
Therefore, this condition in children cannot frequency of cholangitis than ASC, but
be directly compared with adult small‐duct overall the histologic picture is similar in the
PSC and deserves further research. two conditions and the final diagnosis is
based on cholangiographic findings. AIH can
evolve to ASC, i.e. AIH patients with normal
Clinical Features in Pediatric Disease
baseline cholangiogram can develop bile
In retrospective studies including various duct disease during follow‐up. This observa
forms of juvenile sclerosing cholangitis most tion raises the question as to whether AIH
patients are males. Symptoms at presenta and ASC belong to the same disease spec
tion depend on the cause underlying the bile trum or are different conditions.
duct damage. ASC affects males and females
equally, the median age at presentation being
Treatment of Pediatric Disease (ASC
12 years. Symptoms and clinical signs at dis
and Juvenile PSC)
ease onset are indistinguishable from those
of AIH, the majority presenting with jaun Treatment of sclerosing cholangitis varies
dice, hepatosplenomegaly and intestinal according to the underlying cause of the bile
symptoms. Extrahepatic autoimmune dis duct injury. For ASC and PSC, the treatment
eases, including IBD, are present in about schedule varies between centers. Randomized
half of the patients. IBD may be asymptom controlled trials are missing. Published ret
atic at presentation and diagnosed at endos rospective series have used uneven diag
copy. Autoimmune disorders in first‐degree nostic criteria and data on treatment
relatives are found in one‐third of the chil regimens are often scarce. Recommended
dren with either ASC or AIH. treatment for ASC is the same as for AIH:
In ASC, the biochemical profile is often steroids with or without azathioprine.
similar to that of AIH, and at least one‐ Additional UDCA (15 mg/kg daily) is widely
quarter of the patients have normal ALP and prescribed even though data on its impact on
GGT levels. Transaminase levels tend to be disease progression and survival are lacking.
lower than in AIH, while the ALP/AST ratio Most ASC patients have an excellent
Chapter 8 Primary Sclerosing Cholangitis 159
References
1 Karlsen, T.H., Folseraas, T., Thorburn, D., 9 Aabakken, L., Karlsen, T.H., Albert, J. et al.
and Vesterhus, M. (2017). Primary (2017). Role of endoscopy in primary
sclerosing cholangitis: a comprehensive sclerosing cholangitis: European Society of
review. J. Hepatol. 67: 1298–1323. Gastrointestinal Endoscopy (ESGE) and
2 Boonstra, K., Weersma, R.K., van Erpecum, European Association for the Study of the
K.J. et al. (2013). Population‐based Liver (EASL) Clinical Guideline. Endoscopy
epidemiology, malignancy risk, and outcome 49: 588–608.
of primary sclerosing cholangitis. 10 Boberg, K.M., Chapman, R.W., Hirschfield,
Hepatology 58: 2045–2055. G.M. et al. (2011). Overlap syndromes: the
3 Gregorio, G.V., Portmann, B., Karani, J. et al. International Autoimmune Hepatitis
(2001). Autoimmune hepatitis/sclerosing Group (IAIHG) position statement on a
cholangitis overlap syndrome in childhood: controversial issue. J. Hepatol. 54: 374–385.
a 16‐year prospective study. Hepatology 33: 11 Boberg, K.M., Bergquist, A., Mitchell, S.
544–553. et al. (2002). Cholangiocarcinoma in
4 Mieli‐Vergani, G. and Vergani, D. (2016). primary sclerosing cholangitis: risk factors
Sclerosing cholangitis in children and and clinical presentation. Scand. J.
adolescents. Clin. Liver Dis. 20: 99–111. Gastroenterol. 37: 1205–1211.
5 Terziroli Beretta‐Piccoli, B., Vergani, D., 12 Schramm, C., Eaton, J., Ringe, K.I. et al.
and Mieli‐Vergani, G. (2018). (2017). Recommendations on the use of
Autoimmune sclerosing cholangitis: magnetic resonance imaging in PSC. A
evidence and open questions. J. position statement from the International
Autoimmun. 95: 15–25. PSC Study Group. Hepatology 66:
6 Deneau, M.R., El‐Matary, W., Valentino, P.L. 1675–1688.
et al. (2017). The natural history of primary 13 Kim, W.R., Therneau, T.M., Wiesner, R.H.
sclerosing cholangitis in 781 children: a et al. (2000). A revised natural history
multicenter, international collaboration. model for primary sclerosing cholangitis.
Hepatology 66: 518–527. Mayo Clin. Proc. 75: 688–694.
7 Lunder, A.K., Hov, J.R., Borthne, A. et al. 14 Vesterhus, M., Hov, J.R., Holm, A. et al.
(2016). Prevalence of sclerosing cholangitis (2015). Enhanced liver fibrosis score
detected by magnetic resonance predicts transplant‐free survival in primary
cholangiography in patients with long‐term sclerosing cholangitis. Hepatology 62:
inflammatory bowel disease. 188–197.
Gastroenterology 151: 660–669.e4. 15 Corpechot, C., Gaouar, F., El Naggar, A.
8 Weismuller, T.J., Trivedi, P.J., Bergquist, A. et al. et al. (2014). Baseline values and changes in
(2017). Patient age, sex, and inflammatory liver stiffness measured by transient
bowel disease phenotype associated with elastography are associated with severity of
course of primary sclerosing cholangitis. fibrosis and outcomes of patients with
Gastroenterology 152: 1975–1984.e8. primary sclerosing cholangitis.
Chapter 8 Primary Sclerosing Cholangitis 161
Abstract
IgG4-related disease (IgG4-RD) is a multisystem disorder associated with inflammation and fibrosis
in affected organs. Almost any organ may be affected, forming part of the IgG4-RD spectrum.
Patients with IgG4-RD may present to hepatologists with symptoms related to focal masses or dif-
fuse swelling of the pancreas, with intrahepatic and/or extrahepatic biliary disease, as inflammatory
pseudotumors of the liver and biliary tree, and as a variant of autoimmune hepatitis. IgG4-RD
involving the pancreas, bile ducts and/or liver may mimic pancreatic cancer, cholangiocarcinoma,
and primary sclerosing cholangitis. Distinguishing IgG4-RD from malignancy is critical, often
requiring biopsy and histologic analysis. IgG4-RD may be associated with progressive organ
dysfunction, organ failure and malignancy. IgG4-RD is usually treated with oral corticosteroids as
first line therapy, but relapse rates are high once steroid treatment is tapered or stopped. Secondline
and steroidsparing agents are commonly used, but the optimal use of these is not well defined in
head to head clinical studies.
Keywords biliary tree; first line therapy; histologic analysis; IgG4-related disease; inflammatory
pseudotumors; pancreatic cancer; primary sclerosing cholangitis; secondline therapy; steroid sparing
agents
Key Points
●● IgG4-related disease (IgG4‐RD) is a multi- focal masses or diffuse swelling of the
system disorder associated with inflam- pancreas (autoimmune pancreatitis), with
mation and fibrosis in affected organs. intrahepatic and/or extrahepatic biliary
Almost any organ may be affected, form- disease (IgG4‐related sclerosing cholan-
ing part of the IgG4‐RD spectrum. It is gitis), as inflammatory pseudotumors of
increasingly diagnosed globally. the liver and biliary tree (IgG4‐related
●● Patients with IgG4‐RD may present to hepatopathy), and as a variant of autoim-
hepatologists with symptoms related to mune hepatitis.
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
164 Section II Autoimmune Liver Diseases and Their Clinical Correlation
a ctivate complement and has a low affinity to fore related to IgG4‐RD underlying processes
Fcγ receptors. Uniquely, IgG4 antibodies are rather than allergy per se [7]. The human
able to exchange half molecules resulting in response to allergens drives both an IgE and
bispecific antibodies that are unable to cross‐ an IgG4 response and both are dependent on
link and therefore result in the formation of interleukin (IL)‐4 as a “class switching factor”
small immune complexes. Because of these but the kinetics and drivers of the responses
properties, IgG4 antibodies are unable to are different, with IgE and a modified Th2
trigger immunologic effector functions and response dominated by IgG1 developing
have a limited role in the clearance of antigens early. Only after repeat antigen exposure
and in inflammation and may actively inhibit does IgG4 start to dominate and it is likely
IgG1 effector functions. IgG4‐positive B cells that additional immune factors, possibly
may be generated in IgG4‐RD as a regulatory those involved in immune tolerance in
response to an underlying inflammatory, response to chronic antigen stimulation such
antigen‐driven process that has yet to be as IL‐10 or transforming growth factor
clearly defined. (TGF)‐β, are required to drive the IgG4
However, studies of IgG4 biology in patients response. These same factors may also drive
with IgG4‐RD have shown that IgG4 anti- the fibrotic responses that ultimately lead to
bodies can form immune complexes via C1q organ failure in IgG4‐RD. Rituximab therapy
binding via both the classical and mannose lec- leads to a major decline in both IgG4 and IgE
tin‐binding pathway. This, together with the levels in IgG4‐RD patients, suggesting that
hypocomplementemia that is found in more both isotypes are produced, at least in part by
than 30% of patients with AIP implicates short‐lived antibody‐secreting B cells [2].
complement activation in the pathogenesis [5]. Some studies suggest that patients with
In addition, immune complex formation (C3c IgG4‐RD and allergy are a specific patient
and IgG4) has been demonstrated in the tissue subset characterized by circulating Th2
in pancreatic and bile ducts in AIP and IgG4‐ memory cells [8] and targeting distinct auto-
related sclerosing cholangitis (IgG4‐SC), and antigens [9]. The Th2 response in IgG4‐RD is
also to colocalize with collagen IV, and we polyclonal, supporting the notion that mul-
have recently shown decreased Fc antibody tiple antigens drive disease [10,11].
galactosylation in AIP/IgG4‐SC patients com-
pared with healthy controls, a post-translational
Antigens That May Drive an IgG4‐RD
modification important in complement
Response
activation [6]. Overall, this may support a
novel role of IgG4 antibodies in IgG4‐RD path- Since IgG4 antibodies can be generated in
ogenesis, and a potential role in fibrosis. response to chronic antigen exposure, efforts
An IgG4 antibody response is also related, have been made to identify self or exogenous
at least in part, to allergic pathways, although antigens that may trigger IgG4‐RD. However,
the role of allergy in IgG4‐RD is currently the IgG4 response in IgG4‐RD is oligoclonal
debated. High levels of polyclonal IgE, eosin- and reactive against a range of antigens,
ophilia and an increase in allergy and/or including food antigens [11], suggesting that
atopy is described in some IgG4‐RD cohorts no single antigen is driving the IgG4 anti-
[1]. One hypothesis is that IgG4‐RD is a two‐ body response seen in IgG4‐RD. IgG1 anti-
hit disease process: the first hit is the bodies against a range of self antigens have
development of allergy and the second hit been associated with IgG4‐RD, including
redirects the immune response to an IgG4‐ lactoferrin, carbonic anhydrase, pancreatic
RD phenotype. Others have suggested that secretory trypsin inhibitor, and heat shock
the increase in IgE and eosinophilia occurs in protein but none are specific for the disease.
the absence of atopy in patients with IgG4‐ Antibodies to galectin‐3 [12], annexin 11
RD and that this immune phenotype is there- [13], and laminin 511 [9] have more recently
Chapter 9 IgG4-Related Liver and Biliary Disease 167
and in the Mayo clinic, USA (the HISORt cri- be evaluated and applied within an appro-
teria) [22] (Table 9.2). More recently, the priate clinical context. In general, the more
Comprehensive Diagnostic Criteria have features of IgG4‐RD that meet the criteria,
been established that can be usefully applied the more secure the diagnosis. Even with
to all manifestations of IgG4‐RD, rather than these criteria, diagnosis is often complex and
AIP alone [23] (Table 9.3). These guidelines is best confirmed in a multidisciplinary team
each focus on a combination of environment, with specialists familiar with
histopathologic findings, imaging, serum
the disease.
IgG4 levels, and response to therapy. No Making the correct diagnosis in IgG4‐RD
single criterion can be used to make the is imperative since the disease requires
diagnosis and the diagnostic criteria need to lifelong follow‐up and/or long‐term
Note that it is important to differentiate IgG4‐RD from malignant tumors of each organ (e.g. cancer, lymphoma) and
similar diseases (e.g. Sjogren syndrome, primary sclerosing cholangitis, Castleman disease, secondary retroperitoneal
fibrosis, Wegener granulomatosis, sarcoidosis, Churg–Strauss syndrome) by additional histopathologic examination.
Even when patients cannot be diagnosed using these criteria, they may be diagnosed using organ‐specific
diagnostic criteria for IgG4‐RD (see Table 9.4).
Source: Deshpande et al. [23]. Reproduced with permission of Springer Nature.
170 Section II Autoimmune Liver Diseases and Their Clinical Correlation
immunosuppressive therapy and obtaining 1) an IgG4+/IgG+ plasma cell ratio of over
tissue at the point of diagnosis is one of the 40% is required to make a histologic diag-
most important components of this. nosis of IgG4‐RD; and
2) the absolute numbers of IgG4‐positive
Histologic Diagnosis of IgG4‐related plasma cells that support an IgG4‐RD
Hepatobiliary Disease diagnosis depends on the organ involved.
Numbers of IgG4‐positive plasma cells
In general, malignancy is high on the list of required to make the diagnosis for each
differential diagnoses and should be actively organ are given within the consensus state-
excluded. Histologic evaluation is often key ment [23]; in brief, for liver and p ancreas
in making the correct diagnosis of IgG4‐RD with at least two classical histopathologic
in excluding malignancy and therefore every features of disease, this is >10/high‐power
effort should be made to obtain biopsy spec- field (HPF) for a biopsy specimen and >50/
imens before treatment is commenced to HPF for a resection specimen compared
establish the diagnosis. As a broad principle, with >20/HPF for lung and >50/HPF for
fine needle aspiration (FNA) of tissue may be lacrimal gland involvement.
useful to exclude malignancy, but is not
sufficient to make a diagnosis of IgG4‐RD For suspected IgG4‐related hepatobiliary
where tissue with intact architecture is disease, ultrasound‐guided FNA or cytology
required. from biliary brushings may be used to diag-
The classic histopathologic features nose malignancy, but intrabiliary, liver and/or
of IgG4‐RD include storiform (swirling) ampullary biopsies will be required to make a
fibrosis, obliterative phlebitis, and a lympho- histologic diagnosis of IgG4‐RD. Even then,
plasmacytic infiltrate with predominance of the quantification of IgG4‐positive plasma
IgG4‐positive plasma cells. However, in cells will support the diagnosis, but additional
some organs affected by IgG4‐RD and in histologic features may well be absent. In
small biopsy samples, this triad of features is patients with advanced fibrotic disease, IgG4‐
frequently not found (Figure 9.1). positive plasma cells may be reduced in num-
Organ‐specific criteria have been established bers though the IgG4/IgG1 ratio may remain
to aid the diagnosis of IgG4‐RD [23] (Table 9.4). elevated. In these cases, the diagnosis will
Two important principles are that in those need to take into account the clinical features,
with at least one classical histopathologic fea- imaging and ultimately response to therapy.
ture of disease: IgG4‐SC usually affects the extrahepatic and
Figure 9.1 IgG4‐related AIP histology at low power (left) showing an IgG immunostain, and at high
power (right) showing plasma cells positive for IgG4 immunostain (>40%). IgG4 cells are positive for
3,3′‐diaminobenzidine (DAB) and therefore appear dark brown.
Chapter 9 IgG4-Related Liver and Biliary Disease 171
Table 9.4 The Boston consensus criteria for the histologic diagnosis (generic and organ‐specific) of IgG4‐RD.
hilar bile ducts, but small intrahepatic ducts found in up to 80% of IgG4‐RD patients, but
may be affected and observed on liver biopsy. an elevated IgG4 serum antibody is neither
IgG4‐related liver disease may show portal sensitive nor specific for IgG4‐RD. In general,
inflammation, large bile duct obstructive fea- in our experience, the higher the IgG4 serum
tures, portal sclerosis, lobular hepatitis, and antibody level, the more likely the patient is to
evidence of cholestasis. IgG4‐RD in the liver have IgG4‐RD, have multiorgan disease, and
may be considered a variant of AIH, with be at risk for disease relapse. IgG4 serum
infiltration of IgG4‐positive plasma cells and levels more than four times upper limits have
without other classical features of AIH [24], a specificity and positive value of almost 100%
or as part of the IgG4‐RD spectrum especially for distinguishing IgG4‐SC from primary
if additional organs are involved. For sclerosing cholangitis (PSC), but the s ensitivity
pancreatic disease, biopsy can be obtained by remains low at 42%. Many patients with histo-
core biopsy percutaneously, but recently logically confirmed IgG4‐RD that fulfills diag-
biopsy needles (SharkCore) have been nostic criteria have normal serum IgG4
adapted to reliably obtain core biopsies using antibody levels, whilst IgG4 antibody levels
endoscopic ultrasound (EUS). Biopsy at EUS are elevated in up to 25% of people with other
is probably a more comfortable and safer inflammatory, malignant and autoimmune
alternative way to obtain tissue [25]. diagnoses and in a small subset (5%) of healthy
people. This has led investigators to seek more
specific criteria that utilize serum IgG4 anti-
Utility of Blood Tests, Including Serum IgG4
body but in a ratio with IgG1, whereby an
Levels in IgG4‐RD Diagnostics
IgG1/IgG4 ratio above 0.24 enhances sensi-
An elevated serum IgG4 antibody (>1.4 g/l) tivity for an IgG4‐SC diagnosis compared to
has utility in the diagnosis in IgG4‐RD and is PSC with elevated serum IgG4 levels (>80%).
172 Section II Autoimmune Liver Diseases and Their Clinical Correlation
Pancreas Pancreas
Figure 9.2 Imaging changes typical in IgG4‐related hepatobiliary disease showing IgG4‐related AIP. (Top left)
Avid uptake in the pancreas on PET‐CT and (top right) “sausage”‐shaped pancreas typical of IgG4‐RD using
conventional CT imaging. (Bottom) Multiple and long extrahepatic and intrahepatic biliary strictures at ERCP.
Chapter 9 IgG4-Related Liver and Biliary Disease 173
immunostaining and cytogenetics supports months) but may fall out s pontaneously. EUS
the diagnosis of IgG4‐RD. may be used to evaluate mass lesions, to
further characterize duct involvement, and to
perform needle aspiration and tissue biopsies.
Radiologic Characteristics
of Hepatobiliary IgG4‐RD
IgG4‐RD and Relationship
Abdominal ultrasound may demonstrate with Malignancy
biliary obstruction and pancreatic enlarge-
ment suggestive of AIP and exclude other Several groups have shown that IgG4‐RD
pathologies such as gallstones. Cross‐sec- may be associated with an increased risk of
tional imaging, usually computed tomog- malignancy in a variety of organs, and not
raphy (CT) as first line, is essential to define just those affected by the IgG4‐RD itself. A
organ involvement (Figure 9.2). CT abdominal history of malignancy has been observed to
imaging in IgG4‐RD may also reveal biliary be as much as threefold higher (16%) in
strictures, thickened bile duct walls and IgG4‐RD patients compared with matched
liver, hilar or pancreatic masses, and a controls [28]. However, other studies that
swollen or atrophic pancreas. We advocate have evaluated similar‐sized cohorts have
a baseline CT chest, abdomen and pelvis at not shown a significantly increased risk of
diagnosis to define subclinical organ involve- prior malignancy. An overall increased can-
ment. Magnetic resonance cholangiopan- cer risk when patients are followed prospec-
creatography (MRCP) will be indicated in tively [4], during initial diagnosis or
IgG4‐related hepatobiliary disease; contin- particularly in the first year after diagnosis
uous bile duct involvement, bile duct walls of IgG4‐RD [29], has also been reported,
thicker than 2.5 mm, and the presence of though again others have failed to confirm
other organ involvement are all characteristic these observations. Though debated, there is
of IgG4‐related hepatobiliary disease [27]. no conclusive evidence that AIP is associ-
18
F‐fluorodeoxyglucose positron emission ated with an increased risk of pancreatic
tomography (PET) is useful for demon- cancer.
strating clinical involvement in other sites. The reasons for the probable association
The role of whole‐body diffusion‐weighted between IgG4‐RD and malignancy are not
magnetic resonance imaging (MRI) without resolved; it may be that the state of chronic
contrast is of particular interest given the inflammation seen in IgG4‐RD is genuinely
radiation dose from repeat CT and PET‐CT. associated with the development of malig-
Endoscopic retrograde cholangio nancy, as has been observed in other settings
pancreatography (ERCP) will be indicated for (e.g. obesity). However, it is also possible that
patients with possible IgG4‐SC in order to malignancy (or the treatment of malignancy)
identify features that may distinguish this may promote an IgG4 response (a paraneo-
from cholangiocarcinoma and PSC, and to plastic syndrome) and is causally linked to
take biliary biopsies and brushings for the development of IgG4‐RD. Alternative
cytology and the evaluation of malignant explanations are that patients with IgG4‐RD
cells. Typical findings in IgG4‐SC include undergo cross‐sectional whole‐body imaging
long and multifocal strictures, diffuse that may lead to an increase in cancer detec-
pancreatic swelling with a thin narrowed tion, or that risk factors predispose to both
pancreatic duct, and mild biliary dilatation. conditions. Careful, prospective studies of
Temporary biliary stenting may be required large cohorts will ultimately be needed to
to relieve biliary obstruction; the stent should resolve these possibilities. Irrespective, the
be removed once the disease has responded importance of differentiating IgG4‐RD from
to therapy (typically in approximately 2 malignant tumors of each organ needs to be
174 Section II Autoimmune Liver Diseases and Their Clinical Correlation
References
1 Culver, E.L., Sadler, R., Bateman, A.C. et al. eosinophilia, and IgE elevation in IgG4‐
(2017). Increases in IgE, eosinophils, and related disease. Allergy 69 (2): 269–272.
mast cells can be used in diagnosis and 8 Mattoo, H., Della‐Torre, E., Mahajan, V.S.
to predict relapse of IgG4‐related disease. et al. (2014). Circulating Th2 memory cells
Clin. Gastroenterol. Hepatol. 15 (9): 1444– in IgG4‐related disease are restricted to a
1452.e6. defined subset of subjects with atopy.
2 Carruthers, M.N., Topazian, M.D., Allergy 69 (3): 399–402.
Khosroshahi, A. et al. (2015). Rituximab for 9 Shiokawa, M., Kodama, Y., Sekiguchi, K.
IgG4‐related disease: a prospective, open‐ et al. (2018). Laminin 511 is a target
label trial. Ann. Rheum. Dis. 74 (6): antigen in autoimmune pancreatitis. Sci.
1171–1177. Transl. Med. 10 (453): eaaq0997.
3 Khosroshahi, A., Carruthers, M.N., 10 Mattoo, H., Mahajan, V.S., Maehara, T.
Deshpande, V. et al. (2012). Rituximab for et al. (2016). Clonal expansion of CD4(+)
the treatment of IgG4‐related disease: cytotoxic T lymphocytes in patients with
lessons from 10 consecutive patients. IgG4‐related disease. J. Allergy Clin.
Medicine (Baltimore) 91 (1): 57–66. Immunol. 138 (3): 825–838.
4 Huggett, M.T., Culver, E.L., Kumar, M. et al. 11 Culver, E.L., Vermeulen, E., Makuch, M.
(2014). Type 1 autoimmune pancreatitis and et al. (2015). Increased IgG4 responses to
IgG4‐related sclerosing cholangitis is multiple food and animal antigens indicate
associated with extrapancreatic organ a polyclonal expansion and differentiation
failure, malignancy, and mortality in a of pre‐existing B cells in IgG4‐related
prospective UK cohort. Am. J. Gastroenterol. disease. Ann. Rheum. Dis. 74 (5): 944–947.
109 (10): 1675–1683. 12 Perugino, C.A., AlSalem, S.B., Mattoo, H.
5 Kawa, S. (2017). The immunobiology of et al. (2019). Identification of galectin‐3 as
immunoglobulin G4 and complement an autoantigen in patients with IgG4‐
activation pathways in IgG4‐related related disease. J. Allergy Clin. Immunol.
disease. Curr. Top. Microbiol. Immunol. 143 (2): 736–745.e6.
401: 61–73. 13 Hubers, L.M., Vos, H., Schuurman, A.R.
6 Culver, E.L., van de Bovenkamp, F.S., et al. (2018). Annexin A11 is targeted by
Derksen, N.I.L. et al. (2018). Unique patterns IgG4 and IgG1 autoantibodies in IgG4‐
of glycosylation in immunoglobulin subclass related disease. Gut 67 (4): 728–735.
G4‐related disease and primary sclerosing 14 Culver, E.L., Smit, W.L., Evans, C. et al.
cholangitis. J. Gastroenterol. Hepatol. (2017). No evidence to support a role for
https://doi.org/10.1111/jgh.14512. Helicobacter pylori infection and
7 Della Torre, E., Mattoo, H., Mahajan, V.S. plasminogen binding protein in
et al. (2014). Prevalence of atopy, autoimmune pancreatitis and IgG4‐related
178 Section II Autoimmune Liver Diseases and Their Clinical Correlation
diagnosis, treatment response, organ rituximab: the Mayo Clinic experience. Gut
involvement, and relapse in a prospective 62 (11): 1607–1615.
IgG4‐related disease UK cohort. Am. J. 6 Wallace, Z.S., Mattoo, H., Mahajan, V.S.
3
Gastroenterol. 111 (5): 733–743. et al. (2016). Predictors of disease relapse
35 Hart, P.A., Topazian, M.D., Witzig, T.E. et al. in IgG4‐related disease following
(2013). Treatment of relapsing autoimmune rituximab. Rheumatology (Oxford) 55 (6):
pancreatitis with immunomodulators and 1000–1008.
181
Section III
10
Abstract
Patients with seronegative or cryptogenic liver disease represent a heterogeneous and challenging
group. Cryptogenic cirrhosis represents the end stage of a chronic liver disease process where
the underlying etiology remains unknown after comprehensive investigation. The incidence
and prevalence of cryptogenic cirrhosis has decreased with recognition of conditions such
as nonalcoholic fatty liver disease and atypical presentations of autoimmune liver disease.
Autoimmune hepatitis has diverse clinical phenotypes, with patient presentations ranging from
fulminant liver failure to asymptomatic, which makes diagnosis and management challenging.
Up to 30% of patients with autoimmune hepatitis present without significant titers of antinuclear
antibodies and smooth muscle antibodies and up to 20% of patients may have no detectable
autoantibodies. Acute liver failure (ALF) is an uncommon but life threatening condition occurring
in patients with no pre-existing liver disease. The use of corticosteroid therapy in ALF remains
controversial and should be decided on a case-by-case basis in conjunction with a transplant center.
Keywords acute liver failure; antinuclear antibodies; autoimmune hepatitis; corticosteroid
therapy; lifethreatening condition; seronegative chronic liver disease; smooth muscle antibodies
Key Points
●● Acute seronegative or “indeterminate” ●● N‐Acetylcysteine is now recommended by
hepatitis refers to a cryptogenic hepatitis the European Association for the Study of
resulting in acute liver failure. It is the the Liver in all cases of acute liver failure,
second most common cause of acute
regardless of etiology, and may improve
liver failure after acetaminophen overdose transplant‐free survival. It should be given
in developed countries and occurs most for a period up to 5 days at the standard
commonly in young females. It is consis- dose for a acetaminophen overdose.
tently associated with a low chance of ●● The administration of corticosteroids
spontaneous recovery. remains controversial in acute liver
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
184 Section III Specific Clinical Challenges
hepatitis A and B), drug‐induced liver injury In 1992, O’Grady et al. [12] modified the
(DILI), and AIH. Despite the varying etiologies, definition of ALF based on the time between
a number of epidemiologic observations have jaundice and the development of hepatic
been made about ALF. Firstly, there is a pre- encephalopathy following observations from a
ponderance among females (67% of cases in a large series of patients with ALF at King’s
large series of 1147 patients) and relatively College. “Hyperacute liver failure” was defined
young patients (mean age 38 years) [5,6]. as the development of encephalopathy within
Transplant‐free survival in ALF was tradition- 7 days of the onset of jaundice. Hyperacute
ally quoted as below 20% [7,8], but more recent liver failure represents the most common
series have shown significantly improved out- form of ALF and is also the most likely group
comes; a recent large series of over 2000 to have an identifiable cause, frequently from
patients with ALF across 31 centers in the USA acetaminophen hepatotoxicity. It is associated
reported an overall survival rate of 71.4%, with with a high incidence of cerebral edema, but
22.3% of patients receiving liver transplantation paradoxically has the best outcomes and
[9]. There is a profound and consistently patients in this group are more likely to sur-
reported disparity in spontaneous recovery vive without liver transplantation. “Acute liver
rates between ALF due to acetaminophen (70% failure” is defined as the development of
transplant‐free survival) and ALF due to non‐ encephalopathy 8–28 days from the onset of
acetaminophen causes (34.5% transplant‐free jaundice; this group also has a high incidence
survival) [9], and acetaminophen‐related ALF of cerebral edema but has a much poorer
is now considered a favorable etiology. prognosis without transplantation. Finally,
Outcomes for ALF have significantly improved “subacute liver failure” pertains to individuals
compared with the pre‐transplant era where with a more insidious onset of encephalop-
mortality rates were at least 80%. athy, which develops 5–12 weeks after the
Acute seronegative (also known as “non‐A, onset of jaundice; in this group the incidence
non‐B, non‐C,” “non A‐E” or “indeterminate”) of cerebral edema is low but prognosis is very
hepatitis refers to a cryptogenic hepatitis poor without transplantation.
resulting in ALF where no identifiable cause Patients with indeterminate or seronega-
is found with currently available diagnostic tive hepatitis most commonly present with
techniques [10]. Historically it was assumed subacute liver failure, which can be difficult
that unidentified pathogenic viruses or hepa- to differentiate clinically from decompen-
totoxic drugs were the most likely cause sated liver cirrhosis or acute‐on‐chronic liver
for the majority of these cases; the term failure. Such cases have poor spontaneous
“seronegative” was introduced more recently, recovery and survival rates and represent a
which recognized the possibility of an unchar- common indication for emergency liver
acterized autoimmune or immune‐mediated transplantation [4,12].
liver injury. A recent systematic review
by Brennan et al. [4] identified 5027 cases
of ALF, of which 20% were estimated to
represent seronegative or indeterminate
otential Causes of Seronegative
P
liver failure, and in the USA it represents ALF and Features that Suggest
the second most common cause of ALF after an Autoimmune Pathogenesis
acetaminophen [11]. Large case series of
seronegative ALF have also consistently dem- Potentially implicated etiologies for seroneg-
onstrated a strong female preponderance ative ALF include viruses, unrecognized or
(59–71%) who present at a young age (37– seronegative autoimmune disease, unidenti-
39 years). In addition, seronegative ALF has fied or missed acetaminophen overdose,
consistently been associated with a poor rate unrecognized drugs, toxins or idiosyncratic
of spontaneous recovery and a poor prognosis drug reactions, and occult metabolic or ge-
in the absence of liver transplantation. netic disorders (Table 10.1).
186 Section III Specific Clinical Challenges
Table 10.3 Clichy criteria for selection of ALF patients patients with non‐acetaminophen ALF,
for liver transplantation. whereas the King’s College criteria outper-
formed MELD in acetaminophen‐induced
Presence of hepatic encephalopathy and ALF. The significant heterogeneity between
Factor V level <20% of normal (in patients <30 years of studies included in the meta‐analysis made
age), or <30% of normal (in patients >30 years of age) conclusions difficult and the authors suggested
that neither scoring system is optimal for all
patients [24].
criteria in other populations, including both
acetaminophen‐ and non‐acetaminophen‐
related ALF [21,22]. anagement of Seronegative
M
Since 2006, the Model for End‐stage Liver ALF
Disease (MELD) score has been proposed as
an alternative risk stratification algorithm to Management of ALF requires an expert mul-
the King’s College criteria. Katoonizadeh et al. tidisciplinary treating team incorporating
[23] investigated the utility of the MELD score hepatologists, transplant surgeons, and
to predict outcomes in 99 patients with non‐ intensive care physicians. All patients with
acetaminophen ALF. A MELD score above 30 ALF should be managed in an experienced
had a similar negative predictive value to the center, which is usually a transplant center;
King’s College criteria (92 and 91%, respec- intensive care unit (ICU) monitoring is
tively), and the optimal cutoff point for differ- mandatory once hepatic encephalopathy is
entiating between survivors and non‐survivors present. The primary aim of management in
was a MELD score above 35, with a sensitivity ALF, irrespective of etiology, is to achieve
and specificity of 86% and 75%, respectively. A metabolic and hemodynamic stability to pro-
recent meta‐analysis by McPhail et al. [24] vide optimal conditions for hepatic regenera-
quantitatively assessed and compared the tion and recovery. Patients often develop
prognostic accuracy of the King’s College hypoglycemia which should be monitored
criteria and MELD scores in patients with ALF. and glucose infusions commenced as appro-
In non‐acetaminophen causes, the King’s priate. Coagulopathy should not be corrected
College criteria demonstrated a sensitivity of unless there is evidence of active bleeding
58%, specificity of 74%, and diagnostic odds and stress ulcer prophylaxis should be
ratio of 4.16; MELD demonstrated a sensitivity provided. Patients should ideally be in a quiet
of 76%, specificity of 73%, and diagnostic odds environment with the head of the bed raised
ratio of 8.42. This study suggested that MELD to at least 30°; sedation, intubation and venti-
may be a better predictor of mortality in lation is recommended if encephalopathy
190 Section III Specific Clinical Challenges
In evaluating the need for transplantation, hepatic encephalopathy is often a very late
multiple factors need to be taken into development in the clinical course of patients
consideration in any patient with ALF, with indeterminate ALF and patients may be
including the following: too sick to survive transplantation by the
time it manifests (particularly if encephalop-
●● Accurate prediction of the likelihood of
athy has been precipitated by infection).
survival without transplantation, thus
Conversely, the improving survival rates for
identifying patients with a poor prognosis
ALF secondary to acetaminophen means that
at an early time point. This minimizes the
a benefit for transplantation is not always pre-
chance the patient will become too unwell
sent and transplantation in this cohort should
to transplant and maximizes the time to
generally be reserved for patients with high‐
find a suitable donor organ.
grade encephalopathy unless there is severe
●● Consideration of survival potential follow-
acidosis that does not correct rapidly after
ing transplantation, including factors such
fluid resuscitation [34]. The current UK list-
as age, comorbidities, and sepsis.
ing criteria are shown in Table 10.4.
●● Psychological assessment of the patient to
assess expected compliance, social supports,
past history of suicide attempts, and high‐
risk substance and alcohol abuse. This is a
n Approach to Cryptogenic
A
complex area and collateral history should Chronic Liver Disease
be obtained from family and friends.
Typically, some of the classical contraindica- Cryptogenic cirrhosis (CC) represents the end
tions to liver transplantation in patients with stage of a chronic liver disease process where
chronic liver disease may be overlooked in the underlying etiology remains unknown even
ALF and thus the impact of psychosocial after comprehensive clinical, serologic, and
factors in influencing the decision to trans- pathologic investigations have been performed
plant requires careful consideration, consis- [1]. This is a heterogeneous condition that may
tency, clear documentation, and involvement be multifactorial in origin. CC traditionally
of all members of the multidisciplinary team. accounted for 5–30% of cases of cirrhosis [35]
but with improvements in viral hepatitis testing,
The decision to list an appropriate patient for the development of serologic markers for auto-
liver transplantation involves determination immune diseases and recognition of NAFLD,
of the etiology of ALF and dynamic applica- the prevalence has decreased to approximately
tion of a prognostic scoring system to identify 5%. Similar to indeterminate ALF, CC is more
patients with a low probability of spontaneous likely to affect females, although the average age
recovery. The presence of encephalopathy is higher at around 60 years [35]. Cryptogenic
has traditionally been an integral component chronic hepatitis (CCH) is the term given to
of both the definition of ALF and the decision patients with persistently elevated serum ami-
to transplant, although it is no longer such a notransferases of unknown etiology [36].
black‐or‐white scenario. In the UK, the cri-
teria used to list patients for emergency liver
transplantation have recently been modified Potential Causes of Seronegative
Chronic Liver Disease and Features that
to reflect the importance of distinguishing
Suggest an Autoimmune Pathogenesis
between favorable and unfavorable etiologies
in patients with non‐acetaminophen ALF. Every patient with CC requires a detailed
Indeterminate ALF is categorized as an unfa- and systematic assessment to exclude
vorable etiology and the new criteria now common and uncommon recognizable dis-
allow emergency liver transplant listing of orders. Possible causes of CC are shown in
such patients in the absence of hepatic Table 10.5. A comprehensive patient history
encephalopathy. This change recognizes that is mandatory, including:
Chapter 10 Managing Acute and Chronic Seronegative Liver Disease 193
1 Acetaminophen pH <7.25 more than 24 hours after overdose and after fluid
resuscitation
2 Acetaminophen Coexisting prothrombin time >100 seconds or INR >6.5, and serum
creatinine >300 μmol/l or anuria, and grade III–IV encephalopathy
3 Acetaminophen Significant liver injury and coagulopathy following exclusion of
other causes of hyperlactatemia after adequate fluid resuscitation:
arterial lactate >5 mmol/l on admission and >4 mmol/l 24 hours
later in the presence of clinical hepatic encephalopthy
4 Acetaminophen Two of the three criteria from category 2 with clinical evidence of
deterioration (e.g. increased ICP, Fio2 >50%, increasing inotrope
requirements) in the absence of clinical sepsis
5 Favorable non‐ The presence of clinical hepatic encephalopathy is mandatory and
acetaminophen (e.g. prothrombin time >100 seconds or INR >6.5, or any three from the
acute viral hepatitis or following: age 40 years, prothrombin time >50 seconds or INR >3.5,
ecstasy/cocaine‐induced any grade of hepatic encephalopathy with jaundice to
ALF) encephalopathy time >7 days, serum bilirubin >300 μmol/l
6 Unfavorable non‐ a. Prothrombin time >100 seconds or INR >6.5
acetaminophen (e.g. b. In the absence of clinical hepatic encephalopathy, an INR >2
seronegative hepatitis or after vitamin K repletion is mandatory and any two from the
idiosyncratic drug following: age >40 years, prothrombin time >50 seconds or INR
reactions) > 3.5; serum bilirubin >300 μmol/l, and if encephalopathy is
present then jaundice to encephalopathy time >7 days
7 Acute presentation of A combination of coagulopathy and any grade of encephalopathy
Wilson disease or
Budd–Chiari syndrome
Table 10.5 Potential causes of cryptogenic cirrhosis. ●● risk factors for viral hepatitis (e.g. prior drug
use, previous travel, blood transfusions);
Established Non‐alcoholic ●● alcohol exposure (including details of
associations steatohepatitis
Seronegative autoimmune previous and cumulative intake), previous
hepatitis blood alcohol levels;
Occult viral hepatitis ●● family history of liver disease;
Concealed ethanol ●● prior obesity, diabetes, dyslipidemia, bariat-
Sarcoidosis ric surgery or previous evidence of fatty liver;
Wilson disease
●● personal or family history of any autoim-
Less established Occult biliary disease mune disorders;
Hepatic vascular disease
Celiac disease ●● occupational history;
●● migratory history (country of birth,
Other Mitochondriopathies
associations Familial Mediterranean fever countries lived in, time in refugee camps,
Systemic lupus etc.);
erythematosus ●● detailed medication history (for drugs such
Alstrom syndrome as methotrexate);
Keratin 18 mutations ●● past unexplained liver function test
Telomerase gene mutations
abnormalities.
194 Section III Specific Clinical Challenges
The patient history may need to be taken on Table 10.6 Features on liver biology/histology that
multiple occasions and collaborative history may suggest an underlying etiology.
is often required from family members.
A full liver screen should be undertaken in Etiology Suggestive features
all patients and previous laboratory tests and
Non‐alcoholic Foci of macrosteatosis
imaging (from the patient’s general practi-
fatty liver Cellular ballooning
tioner or other hospitals) should be sourced disease Glycogenated nuclei
for review. In true CC, the diagnostic yield
Autoimmune Interface hepatitis (with or
from a liver screen is not high, and isolated hepatitis without bile duct injury)
results can be difficult to interpret. For Plasma cell infiltrate
example, ANA can be detected in around one‐ Portal inflammation
third of patients with non‐alcoholic steato- Centrilobular zone 3 necrosis
hepatitis (NASH), generally at a low titer, and Primary biliary Portal stage: infiltration of portal
hyperferritinemia is common in both NAFLD cholangitis tracts by lymphocytes,
neutrophils and eosinophils;
and alcohol‐related liver disease. Quantitative
possible granuloma formation
immunoglobulins may provide helpful clues: Periportal stage: lymphocytic
isolated elevation of IgG is suggestive of auto- cholangitis, mild interface
immune liver disease, isolated IgM is often hepatitis, portal/periportal
present in primary biliary cholangitis (PBC), ductular proliferation
Septal stage: presence of bridging
and isolated elevation of IgA can suggest ste-
fibrosis
atohepatitis (from alcohol and/or NAFLD). Cirrhotic stage: biliary cirrhosis
Other investigations such as IgG4 levels and with nodular regeneration of
celiac serology should also be considered. hepatic parenchyma and
A liver biopsy and histologic assessment prominent ductopenia
Small‐duct PSC
should be undertaken wherever possible in
patients with CC and may be helpful in estab-
lishing the etiology of disease. Often, interpre- diagnosis of NAFLD include foci of macros-
tation of biopsies necessitates close cooperation teatosis, cellular ballooning, and glycoge-
between the hepatologist and pathologist, nated nuclei [35].
looking for residual “histologic footprints” or
pre-existing disorders (Table 10.6) [35].
Wilson Disease
cirrhosis), which can resolve in subsequent while bile duct injury manifested by
histologic examinations; this may represent uctopenia, portal fibrosis and portal edema
d
acute or severe forms of AIH or an acute suggests an overlap syndrome with primary
exacerbation of chronic disease [44]. The sclerosing cholangitis (PSC) [44]. A pro-
presence of centrilobular necrosis may posed algorithm to diagnose seronegative
confer a good response to corticosteroid AIH is shown in Figure 10.1.
therapy [50]. The presence of bile duct injury It is likely that a significant proportion of
in conjunction with interface hepatitis and patients previously labeled with CC actually
otherwise classical features of AIH may also had seronegative AIH and careful reevalua-
suggest an overlap syndrome. Bile duct tion of CC cases should be considered. For
injury including destructive cholangitis, in example, a study in Germany by Heringlake
conjunction with AMA, may represent an et al. [51] of 126 patients with presumed CCH
overlap syndrome between AIH and PBC, found that 34% of patients could be reclassi-
AIH PBC
Changes
Antibodies Interface No Interface
PDA-E2 + F-Actin +, SLA + p-ANCA + suggestive
negative hepatitis hepatitis
of PBC
AMA
Consider Seronegative
PBC AIH negative
AIH or PSC AIH
PBC
Cryptogenic
Probable Consider
chronic
AIH PSC
hepatitis
Figure 10.1 Proposed algorithm to diagnose autoimmune liver diseases in patients with cryptogenic chronic
hepatitis or cryptogenic cirrhosis. AIH, autoimmune hepatitis, PBC, primary biliary cholangitis; PSC, primary
sclerosing cholangitis; ANA, anti‐nuclear antibody; AMA, anti‐mitchondrial antibody; LKM, anti‐liver kidney
microsomal antibody; SLA, soluble liver antigen; p‐ANCA, atypical perinuclear anti‐neutrophil cytoplasmic
antibody; PDA‐E2, PBC‐specific anti‐nuclear antibody.
Chapter 10 Managing Acute and Chronic Seronegative Liver Disease 197
fied as either definite or probable AIH using being reclassified as AMA positive. In addition,
the revised International Autoimmune the majority of truly AMA‐negative patients
Hepatitis (IAH) score. Similarly, two North will test positive to PBC‐specific ANAs such as
American studies concluded that 19–22% of gp210 and sp100 [55]. Such antibodies are
patients with cryptogenic hepatitis could be equivalent to AMA in terms of diagnostic
reclassified as AIH by the revised IAH score accuracy for PBC, and AMA‐negative ANA‐
[46] or by clinical judgment [52]. Lower positive patients do not require biopsy for
frequencies of seronegative AIH have been diagnostic confirmation [54]. Autoantibody‐
reported in other studies and there is likely to negative PBC cannot be diagnosed without a
be significant regional variation. liver biopsy (Table 10.6).
It remains controversial whether seronega- Patients with AMA‐negative PBC should
tive AIH has a different disease course com- be managed in the same way as those with
pared with seropositive cases and few studies AMA‐positive disease. There are conflicting
have examined this. Sonthalia et al. [53] data relating to the clinical course of AMA‐
published an abstract that examined 101
negative patients compared to those with
patients with seropositive or seronegative AMA‐positive PBC. Mendes and Lindor [56]
AIH. The authors found that patients with reviewed five studies that compared AMA‐
seronegative AIH were significantly more negative and AMA‐positive PBC, and
likely to present with severe AIH (defined as reported a similar clinical course, response to
ALF, acute‐on‐chronic liver failure or severe treatment, and prognosis. In addition,
acute hepatitis) (50% vs. 20.27%, p = 0.022), Invernizzi et al. [57] reviewed a historical
although response to treatment and proba- cohort of 297 patients with AMA‐negative
bility of remission was similar between and ‐positive PBC, finding no substantial dif-
groups. Conversely, Mehendiratta et al. [46] ferences in the clinical spectrum or course of
found that the prevalence of ANA or SMA the disease. More recently, some data suggest
did not correlate with the clinical or histo- that patients with PBC‐specific ANAs may
logic severity of AIH at diagnosis when have worse outcomes, with faster progression
studying 52 patients with seropositive or to jaundice, cirrhosis and liver failure [58].
seronegative AIH and treatment response
was similar in both groups.
Conclusion
Seronegative Primary Biliary
Cholangitis
Patients with seronegative or cryptogenic liver
PBC represents an autoimmune liver disease disease represent a heterogeneous and chal-
of immune‐mediated biliary epithelial cell lenging group. A systematic, comprehensive
injury, cholestasis and progressive fibrosis that and multidisciplinary approach is essential in
may progress to end‐stage biliary cirrhosis. It is evaluating such patients and this, coupled
most commonly diagnosed through the with ongoing improvements in diagnostic
combination of cholestatic liver function tests technology, is likely to result in fewer and
and the presence of AMAs [54]. Approximately fewer patients diagnosed with true crypto-
5% of patients with PBC do not have detectable genic liver disease. Atypical presentations of
AMA, a condition termed AMA‐negative autoimmune liver disease undoubtedly repre-
PBC or autoimmune cholangitis. The develop sent the underlying etiology in a significant
ment of complementary assays to detect AMA, proportion of seronegative patients present-
including immunofluorescence, ELISA and ing with both acute and chronic liver disease.
immunoblotting, has resulted in many patients Timely diagnosis is essential and may result in
originally considered to be AMA negative improved outcomes for the patient.
198 Section III Specific Clinical Challenges
References
18 McPhail, M.J.W., Wendon, J.A., and Bernal, improves transplant‐free survival in early
W. (2010). Meta‐analysis of performance of stage non‐acetaminophen acute liver
Kings’s College Hospital criteria in failure. Gastroenterology 137: 856–864.e1.
prediction of outcome in non‐paracetamol‐ 28 Stravitz, R.T., Sanyal, A.J., Reisch, J. et al.
induced acute liver failure. J. Hepatol. 53: (2013). Effects of N ‐acetylcysteine on
492–499. cytokines in non‐acetaminophen acute
19 Bernuau, J., Samuel, D., Durand, F. et al. liver failure: potential mechanism of
(1991). Criteria for emergency liver improvement in transplant‐free survival.
transplantation in patients with acute viral Liver Int. 33: 1324–1331.
hepatitis and factor V (FV) below 50% of 29 Hu, J., Zhang, Q., Ren, X. et al. (2015).
normal: a prospective study. Hepatology 14 Efficacy and safety of acetylcysteine in
(4): Abstract 49A. “non‐acetaminophen” acute liver failure: a
20 Bismuth, H., Samuel, D., Castaing, D. et al. meta‐analysis of prospective clinical trials.
(1996). Liver transplantation in Europe for Clin. Res. Hepatol. Gastroenterol. 39:
patients with acute liver failure. Semin. 594–599.
Liver Dis. 16: 415–425. 30 Darweesh, S.K., Ibrahim, M.F., and El‐
21 Izumi, S., Langley, P.G., Wendon, J. et al. Tahawy, M.A. (2017). Effect of N‐
(1996). Coagulation factor V levels as a acetylcysteine on mortality and liver
prognostic indicator in fulminant hepatic transplantation rate in non‐acetaminophen‐
failure. Hepatology 23: 1507–1511. induced acute liver failure: a multicenter
22 Ichai, P., Legeai, C., Francoz, C. et al. study. Clin. Drug Investig. 37: 473–482.
(2015). Patients with acute liver failure 31 Karkhanis, J., Verna, E.C., Chang, M.S.
listed for superurgent liver transplantation et al. (2014). Steroid use in acute liver
in France: reevaluation of the Clichy‐ failure. Hepatology 59: 612–621.
Villejuif criteria. Liver Transpl. 21: 32 Zhao, B., Zhang, H.‐Y., Xie, G.‐J. et al.
512–523. (2016). Evaluation of the efficacy of steroid
23 Katoonizadeh, A., Decaestecker, J., Wilmer, therapy on acute liver failure. Exp. Ther.
A. et al. (2007). MELD score to predict Med. 12: 3121–3129.
outcome in adult patients with non‐ 33 Ichai, P., Duclos‐Vallée, J.‐C., Guettier, C.
acetaminophen‐induced acute liver failure. et al. (2007). Usefulness of corticosteroids
Liver Int. 27: 329–334. for the treatment of severe and fulminant
24 McPhail, M.J.W., Farne, H., Senvar, N. et al. forms of autoimmune hepatitis. Liver
(2016). Ability of King’s College criteria Transpl. 13: 996–1003.
and model for end‐stage liver disease 34 O’Grady, J. (2014). Timing and benefit of
scores to predict mortality of patients with liver transplantation in acute liver failure.
acute liver failure: a meta‐analysis. Clin. J. Hepatol. 60: 663–670.
Gastroenterol. Hepatol. 14: 516–525.e5. 35 Caldwell, S. (2010). Cryptogenic cirrhosis:
25 Lee, W.M., Stravitz, R.T., and Larson, A.M. what are we missing? Curr. Gastroenterol.
(2011). Introduction to the Revised Rep. 12: 40–48.
American Association for the Study of 36 Gassert, D.J., Garcia, H., Tanaka, K., and
Liver Diseases Position Paper on Acute Reinus, J.F. (2007). Corticosteroid‐
Liver Failure 2011. Hepatology 55: responsive cryptogenic chronic hepatitis:
965–967. evidence for seronegative autoimmune
26 Wendon, J., Cordoba, J., Dhawan, A. et al. hepatitis. Dig. Dis. Sci. 52: 2433–2437.
(2017). EASL clinical practical guidelines 37 Powell, E.E., Cooksley, W.G., Hanson, R.
on the management of acute (fulminant) et al. (1990). The natural history of
liver failure. J. Hepatol. 66: 1047–1081. nonalcoholic steatohepatitis: a follow‐up
27 Lee, W.M., Hynan, L.S., Rossaro, L. et al. study of forty‐two patients for up to 21
(2009). Intravenous N‐acetylcysteine years. Hepatology 11: 74–80.
200 Section III Specific Clinical Challenges
11
Abstract
Menstruation and ovulation is usually impaired in women with autoimmune hepatitis (AIH) but
returns to normal when disease activity is controlled. Pregnancy in women with AIH has been
associated with increased risk of preterm birth and small-for-gestational age neonates, and likely
increased risk of miscarriage in those with cirrhosis. The risk of maternal complications is closely
linked to the presence of cirrhosis and disease activity, and can be reduced with optimization of
medical management. Women with cirrhosis are at risk of decompensation, in particular those
with higher model for end-stage liver disease score at conception. Flare-up of disease activity is
more common in the postpartum period but can also occur during pregnancy, and can precipitate
liver-related complications and adverse maternal and fetal outcomes. The widespread use of
azathioprine in rheumatologic disorders and inflammatory bowel diseases has provided increasing
evidence on the safety of azathioprine in pregnancy.
Keywords autoimmune hepatitis; azathioprine; disease activity; end-stage liver disease;
inflammatory bowel diseases; medical management; preterm birth; small-for-geastational age
neonates
Key Points
●● Menstruation and ovulation is usually ●● The risk of maternal complications is
impaired in women with autoimmune closely linked to the presence of cir-
hepatitis (AIH) but returns to normal rhosis and disease activity, and can be
when disease activity is controlled. reduced with optimization of medical
●● Pregnancy in women with AIH has been management.
associated with increased risk of preterm ●● Flare‐up of disease activity is more
birth and small‐for‐gestational age neo- common in the postpartum period but
nates, and likely increased risk of miscar- can also occur during pregnancy, and can
riage in those with cirrhosis. precipitate liver‐related
complications
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
202 Section III Specific Clinical Challenges
and adverse maternal and fetal women should undergo surveillance and
outcomes. therapy of gastroesophageal varices usu-
●● Corticosteroids and azathioprine therapy ally prior to conception as well as in the
is considered safe and should not be dis- second trimester of gestation.
continued but mycophenolate and mTOR ●● The management of women with AIH can
inhibitors should be avoided because of be challenging and should involve a multi-
teratogenicity. disciplinary team of experienced hepatol-
●● Women with cirrhosis are at risk of ogists and obstetricians.
decompensation, in particular those with ●● Pre‐conception counseling should be
higher Model for End‐stage Liver Disease offered to women considering pregnancy
(MELD) score at conception. These and can improve outcomes.
Preterm
birth
No. of patients Miscarriage (<36–37 Cesarean Congenital
Study or pregnancies Cirrhosis (<20–22 weeks) Stillbirth weeks) section malformations Maternal complications Outcome
(Continued )
Table 11.1 (Continued)
Preterm
birth
No. of patients Miscarriage (<36–37 Cesarean Congenital
Study or pregnancies Cirrhosis (<20–22 weeks) Stillbirth weeks) section malformations Maternal complications Outcome
AIH, autoimmune hepatitis; LT, liver transplantation; NICU, neonatal intensive care unit; VB, variceal bleeding.
Chapter 11 Managing Pregnant Women with Autoimmune Liver Disease 205
In these studies the miscarriage rate, described by Borssen et al. [3], 26% of cir-
defined as fetal loss before weeks 20–22 of rhotic women delivered via CS as opposed to
gestation, ranged between 5.7 and 42.4%. In 11% of non‐cirrhotic women. In the study
the study by Heneghan et al. [2], 18 women that reported the highest CS rate of 58%, all
with AIH had 35 pregnancies; of these preg- 10 women had cirrhosis with thrombocyto-
nancies, there were only two (5.7%) miscar- penia (platelets <100 × 109/l) and evidence of
riages and one termination. The highest portal hypertension (gastroesophageal var-
miscarriage rate was reported among women ices) in 92% [6]. With the increasing screen-
with cirrhosis in the study by Borssen et al. ing for varices prior to conception and during
[3]. In this study, there were 131 pregnancies pregnancy, the practice of CS in cirrhosis is
in 71 women, including 43 pregnancies in 33 anticipated to subside, and CS will be con-
women with cirrhosis. The (self‐reported) fined mainly to women with obstetric indica-
miscarriage rate in cirrhotic women was tions for CS.
42.4%, but was significantly lower (26.3%) in The reported congenital abnormality and
non‐cirrhotic women. There were 13 termi- neonatal death rate was low in the majority
nations of pregnancy in this cohort, several of studies, and did not seem to be associated
of which occurred following clinician advice. with fetal exposure to azathioprine. The
The number of stillbirths, defined as fetal requirement for admission to a neonatal
loss beyond weeks 20–22 of gestation, in all intensive care unit (NICU) is often linked to
studies was very low (0–2). Cirrhosis seems prematurity and perinatal complications. In
to have an impact on the total live birth rate. the study by Westbrook et al. [7], 11 neonates
Westbrook et al. [7] reviewed the outcomes required admission to NICU. The risk of
of 81 pregnancies in 53 women, among which NICU admission was higher in neonates
33 pregnancies occurred in 21 women with born to cirrhotic women (21.1% vs. 5%).
cirrhosis. The live birth rate was 57.6% in cir- With regard to maternal complications,
rhotic and 83.3% in non‐cirrhotic women. Westbrook et al. reported four maternal
The rate of preterm birth (delivery before deaths (one secondary to pulmonary hyper-
weeks 36–37 of gestation) was on average tension/embolism, and three secondary to
20% in the majority of studies [3,4,7,8]. Some liver‐related complications within 12 months
studies reported lower risk of preterm birth postpartum) [7], and Schramm et al. [8]
(6–11.8%) [2,5,9,10], whereas in a single reported one death secondary to sepsis.
study with a limited number of patients pre- Hepatic decompensation, usually in the form
term birth occurred in 5 of 12 pregnancies of VB or ascites, can occur during pregnancy
(42%) [6]. The risk of preterm birth does not and in the postpartum period in women with
seem to be increased in women with cir- cirrhosis and portal hypertension [2,5,7,8].
rhosis. The risk was 21% in non‐cirrhotic and Hepatic decompensation or even acute/
23% in cirrhotic women in the study by acute‐on‐chronic liver failure can be precipi-
Borssen et al. [3], and 17.5 and 26.3%, respec- tated by an AIH flare or can occur in the con-
tively, in the study by Westbrook et al. [7]. text of de novo AIH presentation. Westbrook
Variable rates of Cesarean section (CS) et al. reported six decompensation episodes
have been reported in the literature (6–58%) during pregnancy, five of which occurred in
that likely reflect the variation in practices the context of AIH flares [7]. They concluded
and expertise among different centers in that the risk of serious maternal outcomes
managing women with chronic liver disease. was increased with maternal age, presence of
The concern regarding increased risk of vari- cirrhosis, absence of treatment for AIH, and
ceal bleeding (VB) during vaginal delivery in active disease in the last 12 months prior to
the context of the Valsava maneuver may conception. In the study by Schramm et al.,
provide an explanation for higher rates of CS one woman presented with de novo AIH and
in women with cirrhosis. In the cohort acute liver failure during pregnancy, and
206 Section III Specific Clinical Challenges
required urgent hysterectomy and liver trans- and can be reduced with optimization of
plantation [8]. The implications of AIH flares medical management.
and de novo AIH are discussed in detail later.
All previous studies included small num-
bers of patients given the rarity of the dis- Liver‐related Outcomes
ease, and more importantly did not compare in Pregnancy
pregnancy outcomes to those in the general
population. Two registry studies aimed to AIH is characterized by loss of immunologic
overcome the latter limitation. In a Swedish tolerance to liver autoantigens. The loss of self‐
Patient Register study, 140 women with AIH tolerance is the result of impaired balance bet-
had 171 singleton deliveries between 2006 ween liver‐specific T‐regulatory cells and
and 2011 [12]. Compared to the general effector cells of liver damage [17]. Under
population, they found an increased relative normal circumstances, T‐regulatory cells exert
risk (RR) of preterm birth (RR 3.72), low control over the effector cells ensuring immune
birthweight (RR 2.51), and gestational tolerance. Impaired T‐regulatory cell control
diabetes mellitus (RR 4.35). There was no function or impaired effector cell responsive-
difference in the risk of preeclampsia, CS and ness results in recognition of liver‐specific
congenital malformation. The rate of still- autoantigens, cytokine release and cytotoxic-
birth in this cohort was 1.2%. Of the women ity, leading to liver damage [18]. In active AIH,
with AIH, 48% were on no treatment. T helper 1 (Th1) immune responses predomi-
A Danish nationwide registry‐based cohort nate, with release of tumor necrosis factor
included 721 women with AIH aged under (TNF)‐α and TNF‐β, interleukin (IL)‐1, IL‐2,
50 years who had their first singleton delivery IL‐12, and interferon (IFN)‐γ, whereas in
between 1994 and 2015, and 7210 healthy remission there is a shift toward anti‐
controls [13]. Of the women with AIH, 41% inflammatory Th2 responses, with release of
had cirrhosis and 51% were on immunosup- IL‐4, IL‐5, IL‐10, and IL‐13 [19].
pressive therapy. The risk of preterm birth A number of immunologic changes occur
(RR 3.19) and small‐for‐gestational age neo- in the advent of conception in an effort to
nate (RR 3.2) was increased in women with ensure immune tolerance to paternally
AIH. The risk of miscarriage (RR 1.17) and derived fetal antigens. The immune system
congenital malformations (RR 1.27) was sim- reacts to the recognition of fetal antigens with
ilar to that of the general population. There the development of protective mecha nisms.
were no stillbirths and no maternal deaths up Pregnancy is characterized by a shift from
to 6 months postpartum. Women with AIH proinflammatory Th1 to anti‐inflammatory
had a higher rate of hospital admissions post- Th2 immune responses and cytokine balance.
partum (10% vs. 3%). The risk of intrahepatic Increased levels of cortisol, estrogen and
cholestasis of pregnancy was also higher in progesterone likely play a role in this shift.
the same group compared to the general High corticosteroid levels inhibit IL‐2 syn-
population (11.4% vs. 0.9%). thesis, and progesterone enhances IL‐4 pro-
Pregnancy is feasible in women with AIH duction [19]. Progesterone also inhibits
even in the presence of cirrhosis and portal natural killer (NK) cell activity and prevents
hypertension. Pregnancy in this population abortion [20]. Of particular interest is the
is associated with increased risk of preterm induction of T‐regulatory cells during preg-
birth and small‐for‐gestational age neonates, nancy that facilitate tolerance to fetal
but not congenital malformations. The risk antigens, but which also maintain tolerance
of miscarriage is likely increased mainly in to self‐antigens [21].
women with cirrhosis [6,14–16]. The risk of The immunologic shift in Th1/Th2 balance
maternal complications is closely linked to and the induction of T‐regulatory cells pro-
the presence of cirrhosis and disease activity, vide a potential explanation for the AIH
Chapter 11 Managing Pregnant Women with Autoimmune Liver Disease 207
Table 11.3 Cases of de novo autoimmune hepatitis during pregnancy or postpartum.
De novo De novo
No. of during postpartum
Author patients pregnancy (1–3 months) Presentation Outcome
Danielsson Borssen 2 2
et al. [3]
Terrabuio et al. [5] 1 1 Decompensation Miscarriage
(SBP)
Sato et al. [22] 1 1 Subacute liver CS at 29 weeks
injury (mother and infant
survived)
Westbrook et al. [7] 2 1 1 Good response to
steroids
Efe et al. [23] 3 3 Good response to
steroids
Schramm et al. [8] 6 1 5 1 liver failure 1 hysterectomy and
LT at 18 weeks
Heneghan et al. [2] 2 2 1 CS at 28 weeks: child
decompensation/ with developmental
liver failure problems
Samuel et al. [24] 5 5 Good response to
steroids
CS, Cesarean section; LT, liver transplantation; SBP, spontaneous bacterial peritonitis.
16 weeks required CS at 29 weeks [22]. maternal and fetal mortality [30]. Worsening
Another woman who presented with liver of portal hypertension may occur during
failure required emergency hysterectomy pregnancy as a result of physiologic volume
and liver transplantation at 18 weeks [8]. A expansion and the pressure of the expand-
woman who presented with decompensa- ing uterus on the inferior vena cava and col-
tion/liver failure had CS at 28 weeks and the lateral circulation [29]. These hemodynamic
child had severe developmental prob- changes are more prominent in the second/
lems [2]. De novo AIH is usually responsive thrird trimester when the risk of decompen-
to conventional induction therapy, but occa- sation becomes higher. Initial observations
sionally can be associated with liver failure/ yielded a high rate of VB (18–32%) during
decompensation with deleterious sequelae pregnancy among women with cirrhosis,
for the mother and the fetus. and a rate of up to 78% in those with preex-
isting varices with high associated mortality
[32]. Subsequent studies demonstrated a
Pregnancy in Cirrhosis significantly lower VB rate that likely reflects
improvements in variceal surveillance and
The presence of cirrhosis increases the risk optimal management of gastroesophageal
of both fetal and maternal complications varices. A large study that reviewed the
[29]. Cirrhosis is associated with increased complications of 339 pregnancies in women
risk of miscarriage, preterm birth and small‐ with cirrhosis reported decompensation
for‐gestational age neonate [3,30,31]. The episodes in 15% (ascites 11%, VB 5%). The
primary concern in this population is the maternal mortality rate was 1.8% in cir-
risk of decompensation that increases both rhotic and 0% in non‐cirrhotic women,
210 Section III Specific Clinical Challenges
while the fetal mortality rate was 5.2 and may impair uterine blood flow [38].
2.1%, respectively [30]. The maternal Propranolol is the beta‐blocker of choice and
mortality rate increased to 6% and the fetal is considered generally safe in pregnancy
mortality rate to 12% in cirrhotic women despite concerns regarding decreased pla-
who decompensated during pregnancy. Two cental perfusion, fetal and neonatal brady-
smaller studies in women with cirrhosis cardia, and hypoglycemia (FDA category C).
showed a similar decompensation rate of In the event of VB, EVL, vasoactive agents
10% [33], and 16% (VB 4%) with associated that reduce portal pressure, and broad‐spec-
maternal mortality of 4% [34]. Patients with trum antibiotics are the mainstay of
cirrhosis are at increased risk of infections, treatment. Although terlipressin is the vaso-
which are associated with poor outcomes. A active agent of choice in non‐pregnant
single study in 187 women with cirrhosis patients, octreotide (FDA category B) might
demonstrated an increased risk of peripar- be preferred in pregnancy due to the lack of
tum infections in this population [35]. adequate evidence with regard to safety of
Women with cirrhosis and portal hyperten- the former (potential risk of uterine contrac-
sion are also more likely to develop post- tions and decreased uterine blood flow).
partum bleeding and other complications Transjugular intrahepatic portosystemic
that may require hospital admission [31]. shunt (TIPS) insertion is limited in preg-
The main predictor of pregnancy outcomes nancy due to fetal radiation exposure, but
and liver‐related complications is the severity good outcomes have been reported in a small
of liver disease at conception. Westbrook number of pregnant women with recurrent/
et al. [33] reviewed outcomes of 62 preg- refractory VB [39–41]. Ascites and hepatic
nancies in 29 women with cirrhosis, and encephalopathy are managed with diuretics
found that higher Model for End‐stage Liver and lactulose, respectively, as in non‐preg-
Disease (MELD) and Child–Pugh scores were nant patients.
associated with increased risk of preterm Another consideration in women with cir-
birth. A MELD score of 10 or above was an rhosis and portal hypetension is the optimal
accurate predictor of decompensation, mode of delivery. The increase in intra‐
whereas a MELD score of 6 or less was associ- abdominal pressure during vaginal delivery
ated with no liver‐related complications. A may precipitate rupture of gastroesophageal
further study of 165 pregnancies in 100 varices, which provides an explanation for
women with chronic liver disease (80 preg- the increased frequency of CS in women
nancies in 48 women with cirrhosis) showed with cirrhosis [30,35]. This risk has dramati-
that the albumin/bilirubin ratio is an accurate cally decreased with the implementation of
predictor of live births, and the aspartate ami- variceal screening, and vaginal delivery is
notransferase (AST)/platelet ratio is a good likely safe in women with no or small varices.
predictor of preterm birth [36]. CS should be reserved for women with severe
Screening for varices with esophagogastro- portal hypertension and varices refractory to
duodenoscopy (EGD) prior to (planned) con- medical/endoscopic therapy. Otherwise, the
ception or in the second trimester of decision for CS should be driven by obstetric
pregnancy is indicated in all women with cir- indications.
rhosis [37]. Esophageal varices should be
managed according to their grade and the
presence of high‐risk endoscopic stigmata Safety of Medication
with non‐selective beta‐blockers and/or in Pregnancy
endoscopic variceal ligation (EVL). EGD is
considered safe during pregnancy, with the Prednisolone at high doses is used for
exception of a small risk of fetal hypoxia due induction of remission in AIH, and can also
to sedation and maternal positioning that be used at low doses for maintenance of
Chapter 11 Managing Pregnant Women with Autoimmune Liver Disease 211
remission usually in combination with aza- terminate pregnancy, stemmed mainly from
thioprine. Despite the lack of well‐controlled animal studies that showed risk of teratoge-
data in humans, prednisolone is considered nicity with azathioprine and the lack of con-
safe during pregnancy and lactation as the trolled data in humans. The widespread use
transplacental passage is limited (Table 11.4) of azathioprine in rheumatologic disorders
[42]. There is a slightly increased risk of cleft and inflammatory bowel diseases has
lip/palate in neonates born to women provided increasing evidence on the safety of
exposed to steroids in the first trimester of azathioprine in pregnancy. Despite initial
gestation [43]. Neonates might need to be reports of congenital anomalies and adverse
monitored for adrenal insufficiency when the pregnancy outcomes, these have not been
mother has been exposed to long‐term ste- confirmed in large‐scale studies [45–47].
roid therapy, although this event is rare [44]. None of the previous cohort studies in AIH
Azathioprine and 6‐mercaptopurine have demonstrated increased risk of adverse preg-
been the source of great concern among phy- nancy outcomes or neonatal complications
sicians and women with AIH in the past in women exposed to azathioprine during
because of the equivocal evidence with pregnancy [7]. Rare cases of leukopenia in
regard to their safety in pregnancy. These the neonate have been associated with high
concerns, which led women previously to levels of active metabolites (6‐thioguanine
avoid pregnancy, stop therapy or even nucleotide, 6‐TGN) secondary to low
FDA
pregnancy
Drug category Evidence in pregnancy Evidence in lactation
Pre‐conception/ ●● Amenorrhea and anovulation are not uncommon at the initial presentation of AIH,
conception and usually reverse with initiation of treatment and optimal control of disease
activity
Pregnancy ●● Pregnancy in AIH is associated with increased risk of preterm birth and small‐for‐
outcomes gestational age neonates, but not congenital malformations
●● The risk of miscarriage is increased mainly in women with cirrhosis
Maternal ●● The risk of maternal complications is closely linked to the presence of cirrhosis/
complications severity of liver disease and AIH activity
●● Optimal control of disease activity is an important goal in pregnancy: Prednisolone
and azathioprine are considered safe and should not be discontinued. MMF is
contraindicated in pregnancy
●● Screening for varices prior to conception or in the second trimester is indicated in
all women with cirrhosis
●● Cesarean section should be reserved for women with severe portal hypertension, or
otherwise should be driven by obstetric indications
Postpartum ●● AIH flares can occur in 15% of pregnancies, and more commonly postpartum, and
can be associated with adverse pregnancy outcomes or decompensation episodes
●● Close monitoring is required in the first 3 months postpartum
to a different regimen at least 3 months prior and fetal outcomes. Optimal control of dis-
to conception. ease activity is thus an important goal in
pregnancy. First‐ and second‐line AIH
therapy is considered safe and should not be
Summary discontinued, with the exception of MMF.
Women with cirrhosis are at risk of decom-
Pregnancy in women with AIH has been pensation, in particular those with higher
associated with increased risk of preterm MELD score at conception. Women with cir-
birth and small‐for‐gestational age neonates, rhosis should undergo surveillance and
and likely increased risk of miscarriage in therapy of gastroesophageal varices prior to
women with cirrhosis. The risk of maternal conception and/or in the second trimester of
complications is closely linked with the gestation. The management of women with
presence of cirrhosis and disease activity, and AIH can be challenging and should involve a
can be reduced with optimization of medical multidisciplinary team of experienced hepa-
management. AIH flares are more common tologists and obstetricians. Preconception
in the postpartum period, but can also occur counseling should be offered to women con-
during pregnancy, and can precipitate liver‐ sidering pregnancy and can improve
related complications and adverse maternal outcomes.
214 Section III Specific Clinical Challenges
References
1 Cundy, T.F., Butler, J., Pope, R.M. et al. pregnancy in women with autoimmune
(1991). Amenorrhoea in women with hepatitis. J. Autoimmun. 38: J239–J244.
non‐alcoholic chronic liver disease. Gut 32: https://doi.org/10.1016/j.jaut.2011.1012.
202–206. 1002.
2 Heneghan, M.A., Norris, S.M., O’Grady, J.G. 8 Schramm, C., Herkel, J., Beuers, U. et al.
et al. (2001). Management and outcome of (2006). Pregnancy in autoimmune
pregnancy in autoimmune hepatitis. Gut 48: hepatitis: outcome and risk factors. Am. J.
97–102. Gastroenterol. 101: 556–560. https://doi.
3 Danielsson Borssen, A., Wallerstedt, S., org/10.1111/j.1572‐0241.2006.00479.x.
Nyhlin, N. et al. (2016). Pregnancy and 9 Candia, L., Marquez, J., and Espinoza, L.R.
childbirth in women with autoimmune (2005). Autoimmune hepatitis and
hepatitis is safe, even in compensated pregnancy: a rheumatologist’s dilemma.
cirrhosis. Scand. J. Gastroenterol. 51: Semin. Arthritis Rheum. 35: 49–56. https://
479–485. https://doi.org/10.3109/00365521. doi.org/10.1016/j.semarthrit.2005.1003.
00362015.01115893. 1002.
4 Werner, M., Bjornsson, E., Prytz, H. et al. 10 Buchel, E., Van Steenbergen, W., Nevens, F.,
(2007). Autoimmune hepatitis among fertile and Fevery, J. (2002). Improvement of
women: strategies during pregnancy and autoimmune hepatitis during pregnancy
breastfeeding? Scand. J. Gastroenterol. 42: followed by flare‐up after delivery. Am. J.
986–991. https://doi.org/10.1080/0036552 Gastroenterol. 97: 3160–3165. https://doi.
0601155266. org/10.1111/j.1572‐0241.2002.07124.x.
5 Terrabuio, D.R., Abrantes‐Lemos, C.P., 11 Steven, M.M., Buckley, J.D., and Mackay,
Carrilho, F.J., and Cancado, E.L. (2009). I.R. (1979). Pregnancy in chronic active
Follow‐up of pregnant women with hepatitis. Q. J. Med. 48: 519–531.
autoimmune hepatitis: the disease 12 Stokkeland, K., Ludvigsson, J.F., Hultcrantz,
behavior along with maternal and fetal R. et al. (2016). Increased risk of preterm
outcomes. J. Clin. Gastroenterol. 43: birth in women with autoimmune
350–356. https://doi.org/10.1097/MCG. hepatitis: a nationwide cohort study. Liver
1090b1013e318176b318178c318175. Int. 36: 76–83. https://doi.org/10.1111/
6 Nomuras, R.M., Kleine, R.T., Igai, A.M. et al. liv.12901.
(2013). Clinical and obstetrical management 13 Gronbaek, L., Vilstrup, H., and Jepsen, P.
of pregnant women with autoimmune (2018). Pregnancy and birth outcomes in a
hepatitis complicated by moderate or severe Danish nationwide cohort of women with
thrombocytopenia. Rev. Assoc. Med. Bras. autoimmune hepatitis and matched
(1992) 59: 28–34. population controls. Aliment. Pharmacol.
7 Westbrook, R.H., Yeoman, A.D., Kriese, S., Ther. 48: 655–663. https://doi.org/10.1111/
and Heneghan, M.A. (2012). Outcomes of apt.14925.
Chapter 11 Managing Pregnant Women with Autoimmune Liver Disease 215
14 Szczepanska, M., Horst, N., Mikolajczyk, cirrhosis overlap syndrome: a report of two
M. et al. (2018). The course and delivery of cases. Clin. Res. Hepatol. Gastroenterol.
a pregnancy in a patient with autoimmune 35: 687–689. https://doi.org/10.1016/j.
hepatitis complicated by cirrhosis of the clinre.2011.1006.1003.
liver. Ginekol. Pol. 89: 339–340. https://doi. 24 Samuel, D., Riordan, S., Strasser, S. et al.
org/10.5603/GP.a2018.0058. (2004). Severe autoimmune hepatitis first
15 Orgul, G., Ozkan, E.U., Celik, H.T., and presenting in the early post partum period.
Beksac, M.S. (2017). Autoimmune hepatitis Clin. Gastroenterol. Hepatol. 2: 622–624.
and pregnancy: report of two cases with 25 Aggarwal, N., Chopra, S., Suri, V. et al.
different maternal outcomes. Clin. Exp. (2011). Pregnancy outcome in women with
Hepatol. 3: 212–214. https://doi.org/10. autoimmune hepatitis. Arch. Gynecol.
5114/ceh.2017.71445. Obstet. 284: 19–23. https://doi.
16 Braga, A. and Braga, J. (2016). Successful org/10.1007/s00404‐00010‐01540‐z.
pregnancy with autoimmune cirrhosis. 26 Malhotra, B., Malhotra, N., Deka, D., and
BMJ Case Rep. 2016:bcr2015212501. Takkar, D. (2002). Immunosuppressive
doi: https://doi.org/10.1136/bcr‐2015‐ effect of pregnancy on autoimmune
212501. hepatitis: a case report and review of
17 Wen, L., Peakman, M., Lobo‐Yeo, A. et al. literature. Eur. J. Obstet. Gynecol. Reprod.
(1990). T‐cell‐directed hepatocyte damage Biol. 101: 91–92.
in autoimmune chronic active hepatitis. 27 Colle, I. and Hautekeete, M. (1999).
Lancet 336: 1527–1530. Remission of autoimmune hepatitis during
18 Ferri, S., Longhi, M.S., De Molo, C. et al. pregnancy: a report of two cases. Liver 19:
(2010). A multifaceted imbalance of T cells 55–57.
with regulatory function characterizes type 28 Riely, C.A. (2002). Meet Sarah: a vignette of
1 autoimmune hepatitis. Hepatology 52: autoimmune hepatitis and pregnancy. Am.
999–1007. https://doi.org/10.1002/ J. Gastroenterol. 97: 2945. https://doi.
hep.23792. org/10.1111/j.1572‐0241.2002.07089.x.
19 Muratori, P., Loffreda, S., Muratori, L. et al. 29 Tan, J., Surti, B., and Saab, S. (2008).
(2002). Spontaneous remission of Pregnancy and cirrhosis. Liver Transpl. 14:
autoimmune hepatitis during pregnancy. 1081–1091. https://doi.org/10.1002/
Dig. Liver Dis. 34: 608–609. lt.21572.
20 Szekeres‐Bartho, J. (2002). Immunological 30 Shaheen, A.A. and Myers, R.P. (2010). The
relationship between the mother and the outcomes of pregnancy in patients with
fetus. Int. Rev. Immunol. 21: 471–495. cirrhosis: a population‐based study. Liver
21 Aluvihare, V.R., Kallikourdis, M., and Betz, Int. 30: 275–283. https://doi.
A.G. (2004). Regulatory T cells mediate org/10.1111/j.1478‐3231.2009.02153.x.
maternal tolerance to the fetus. Nat. 31 Puljic, A., Salati, J., Doss, A., and Caughey,
Immunol. 5: 266–271. https://doi. A.B. (2016). Outcomes of pregnancies
org/10.1038/ni1037. complicated by liver cirrhosis, portal
22 Sato, H., Tomita, K., Yasue, C. et al. (2015). hypertension, or esophageal varices.
Pregnant woman with non‐comatose J. Matern. Fetal Neonatal Med. 29:
autoimmune acute liver failure in the 506–509. https://doi.org/10.3109/14767058
second trimester rescued using medical .14762015.11009438.
therapy: a case report. Hepatol. Res. 45: 32 Russell, M.A. and Craigo, S.D. (1998).
349–355. https://doi.org/10.1111/ Cirrhosis and portal hypertension in
hepr.12344. pregnancy. Semin. Perinatol. 22: 156–165.
23 Efe, C., Ozaslan, E., and Purnak, T. (2011). 33 Westbrook, R.H., Yeoman, A.D., O’Grady,
Outcome of pregnancy in patients with J.G. et al. (2011). Model for end‐stage
autoimmune hepatitis/primary biliary liver disease score predicts outcome in
216 Section III Specific Clinical Challenges
cirrhotic patients during pregnancy. Dig. Liver Dis. 40: 387–390. https://doi.
Clin. Gastroenterol. Hepatol. 9: 694–699. org/10.1016/j.dld.2007.1002.1013.
https://doi.org/10.1016/j.cgh.2011.1003. 42 van Runnard Heimel, P.J., Franx, A.,
1036. Schobben, A.F. et al. (2005).
34 Jena, P., Sheela, C.N., Venkatachala, R.P., Corticosteroids, pregnancy, and HELLP
and Devarbhavi, H. (2017). Obstetric syndrome: a review. Obstet. Gynecol. Surv.
outcome in women with chronic liver 60: 57–70; quiz 73–74.
disease. J. Obstet. Gynaecol. India 67: 43 Bermas, B.L. (2014). Non‐steroidal anti
263–269. https://doi.org/10.1007/ inflammatory drugs, glucocorticoids and
s13224‐13016‐10959‐y. disease modifying anti‐rheumatic drugs for
35 Murthy, S.K., Heathcote, E.J., and Nguyen, the management of rheumatoid arthritis
G.C. (2009). Impact of cirrhosis and liver before and during pregnancy. Curr. Opin.
transplant on maternal health during labor Rheumatol. 26: 334–340. https://doi.
and delivery. Clin. Gastroenterol. Hepatol. org/10.1097/BOR.0000000000000054.
7: 1367–1372. 44 Ostesen, M. (1994). Optimisation of
36 Gonsalkorala, E.S., Cannon, M.D., Lim, T.Y. antirheumatic drug treatment in pregnancy.
et al. (2019). Non‐invasive markers (ALBI Clin. Pharmacokinet. 27: 486–503.
and APRI) predict pregnancy outcomes in 45 Francella, A., Dyan, A., Bodian, C. et al.
women with chronic liver disease. Am. J. (2003). The safety of 6‐mercaptopurine for
Gastroenterol. 114: 267–275. childbearing patients with inflammatory
37 Shergill, A.K., Ben‐Menachem, T., bowel disease: a retrospective cohort study.
Chandrasekhara, V. et al. (2012). Gastroenterology 124: 9–17. https://doi.
Guidelines for endoscopy in pregnant and org/10.1053/gast.2003.50014.
lactating women. Gastrointest. Endosc. 76: 46 Alami, Z., Agier, M.S., Ahid, S. et al. (2018).
18–24. https://doi.org/10.1016/j. Pregnancy outcome following in utero
gie.2012.1002.1029. exposure to azathioprine: a French
38 O’Mahony, S. (2007). Endoscopy in comparative observational study. Therapie
pregnancy. Best Pract. Res. Clin. 73: 199–207. https://doi.org/10.1016/j.
Gastroenterol. 21: 893–899. https://doi. therap.2017.1006.1006.
org/10.1016/j.bpg.2007.1005.1007. 47 Saavedra, M.A., Sanchez, A., Morales, S.
39 Ingraham, C.R., Padia, S.A., Johnson, G.E. et al. (2015). Azathioprine during
et al. (2015). Transjugular intrahepatic pregnancy in systemic lupus erythematosus
portosystemic shunt placement during patients is not associated with poor fetal
pregnancy: a case series of five patients. outcome. Clin. Rheumatol. 34: 1211–1216.
Cardiovasc. Intervent. Radiol. 38: https://doi.org/10.1007/s10067‐10015‐
1205–1210. https://doi.org/10.1007/s00270‐ 12987‐x.
00015‐01053‐z. 48 Thomas, C., Monteil‐Ganiere, C., Mirallie,
40 Wood, A.M., Grotegut, C.A., Ronald, J. S. et al. (2018). A severe neonatal
et al. (2018). Identification and lymphopenia associated with
management of abdominal wall varices in administration of azathioprine to the
pregnancy. Obstet. Gynecol. 132: 882–887. mother in a context of Crohn’s disease.
https://doi.org/10.1097/AOG.000000000 J. Crohns Colitis. 12: 258–261. https://doi.
0002805. org/10.1093/ecco‐jcc/jjx1123.
41 Lodato, F., Cappelli, A., Montagnani, M. 49 de Boer, N.K., Jarbandhan, S.V., de Graaf,
et al. (2008). Transjugular intrahepatic P. et al. (2006). Azathioprine use during
portosystemic shunt: a case report of pregnancy: unexpected intrauterine
rescue management of unrestrainable exposure to metabolites. Am. J.
variceal bleeding in a pregnant woman. Gastroenterol. 101: 1390–1392.
Chapter 11 Managing Pregnant Women with Autoimmune Liver Disease 217
12
Abstract
Autoimmune liver diseases include primary biliary cholangitis (PBC), primary sclerosing cholangi
tis, and autoimmune hepatitis (AIH). Osteoporosis is frequently observed in patients with autoim
mune liver diseases, particularly PBC. The pathogenesis of osteoporosis is mainly characterized by
low bone formation, although increased bone resorption has been described in patients with severe
cholestasis and advanced liver disease. Osteomalacia, characterized by poor bone mineralization,
is very uncommon and has only been reported in a few patients with advanced PBC and severe
intestinal malabsorption, and in geographic areas with decreased exposure to sunlight. The preva
lence of bone disease in patients with AIH is uncertain, although it is assumed that osteoporosis is
one of the most frequent complications of corticosteroid treatment. For the prevention and
treatment of osteoporosis, good nutrition, avoidance of smoking and excess alcohol, weightbearing
exercise, treatment of risk factors, and prescription of calcium and vitamin D supplements appear
helpful.
Key Points
●● Osteoporosis is frequently observed in ●● The pathogenesis of osteoporosis is mainly
patients with autoimmune liver diseases, characterized by low bone formation,
particularly primary biliary cholangitis although increased bone resorption has also
(PBC). been described in patients with severe cho
●● Osteoporosis is associated with the severity lestasis and advanced liver disease.
of liver disease, older age, use of corticoste ●● Osteomalacia, characterized by poor bone
roids, and the duration of cholestasis mineralization, is very uncommon and
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
220 Section III Specific Clinical Challenges
has only been reported in a few patients ●● For the prevention and treatment of oste
with advanced PBC and severe intestinal oporosis, good nutrition, avoidance of
malabsorption, and in geographic areas smoking and excess alcohol, weight‐
with decreased exposure to sunlight. bearing exercise, treatment of risk factors,
●● The prevalence of bone disease in patients and prescription of calcium and vitamin
with autoimmune hepatitis is uncertain, D supplements appear helpful.
although it is assumed that osteoporosis is ●● Oral bisphosphonates, including weekly
one of the most frequent complications of alendronate and monthly ibandronate,
corticosteroid treatment. can be effective for increasing bone min
●● Bone density measurement is the usual eral density.
diagnostic procedure. Lateral X‐ray of the ●● The efficacy of parenteral bisphospho
dorsal and lumbar spine should also be nates such as zoledronic acid, as well
carried out to identify vertebral fractures, as other newer agents, in patients at high
and laboratory measurements of risk for developing osteoporotic frac
circulating levels of calcium, phosphorus, tures needs ongoing evaluation in
25‐hydroxyvitamin D, and parathyroid patients with autoimmune liver
hormone can be assessed at diagnosis. diseases.
(a) (b)
Figure 12.1 (a) The diagnosis of osteoporosis is based on densitometric criteria (T‐score equal or below –2.5).
(b) Severe osteoporosis is diagnosed when the densitometric criteria are accompanied by one or more fragility
fractures.
222 Section III Specific Clinical Challenges
Prevalence of Osteoporosis
while the prevalence is higher in advanced
and Fractures stages and before transplantation [13].
The prevalence of fractures ranges from 2
Primary Biliary Cholangitis
to 23% (Table 12.1), being about 22% in
Over the years the prevalence of bone dis patients with advanced disease. In a series of
ease in PBC has changed because patients 185 patients, fractures were detected in 21%
are now diagnosed in the early stages of the of the cases (11% vertebral and 12% peripheral
disease and are mostly asymptomatic [3]. fractures) and associated with osteoporosis
Moreover, in the earlier studies bone disease and the severity of cholestasis. A BMD
was assessed by X‐ray of the spine or by bone threshold that captures most vertebral
biopsy and histomorphometry. Bone mass is fractures has been identified. Thus, the
now best assessed by bone densitometry of patients with a lumbar or femoral BMD
the lumbar spine and proximal femur. T‐score below −1.5 are those with a high risk
The reported prevalence of osteoporosis in for fractures. The clear‐cut correlation bet
patients with PBC varies considerably, from ween vertebral fracture and a T‐score less
20 to 90%. Using bone histomorphometry, than −1.5 observed in these patients may
the prevalence of osteoporosis ranges even indicate that this densitometric measurement
more markedly, from 9 to 90% [1,4,6–15]. In is a useful guide for considering therapy.
the most recent and largest series of patients Data regarding incident fractures are fairly
with PBC, the prevalence of osteoporosis, uncertain. One study reported 5% of incident
assessed by densitometry, is approximately fractures over 2 years, while another reported
31% (Table 12.1), and is significantly higher that 13% of patients with PBC had new radio
than in the age‐matched population. Age graphic vertebral deformities over the same
above 56 years, duration of PBC longer than period. Recent reports are contradictory
4 years, and advanced histologic stage are the with respect to the increased fracture risk in
variables associated with osteoporosis [12]. PBC. Thus, one report did not show an
A lower prevalence of osteoporosis is increase in the fracture risk, while another
observed in patients with less severe disease, population‐based cohort study including
Table 12.1 Prevalence of osteoporosis and fractures according to the severity and stage of primary biliary
cholangitis.
No. of Advanced
Reference cases Females (%) disease (%) Osteoporosis (%) Fractures (%)
930 people with PBC found modest increases a ge‐ and sex‐matched population. About half
in both the absolute and relative fracture risk of patients had early disease, and 22% had
compared with the general population. This cirrhosis. Deficiency of vitamin D and
last study observed an approximately twofold calcium was found in 23% and 4% of patients,
increased risk of any fracture, hip fracture, respectively. The multivariate analyses iden
and ulna/radius fracture among people with tified age older than 54 years, body mass
PBC compared with the general population. index (BMI) below 24.1 kg/m2, and associ
ated IBD for more than 19 years as the
variables associated with osteoporosis.
Primary Sclerosing Cholangitis
Interestingly, osteoporosis was present in
There are fewer reports regarding bone 75% of the patients with these three factors,
health in patients with PSC as compared but in only 3% of patients with none of them.
with PBC. BMD in PSC is lower than There was no significant difference between
expected for age‐ and sex‐matched controls. patients with and without IBD in terms of
Thus, in a cohort of 81 patients included in a baseline bone density or in the rate of bone
randomized trial of UDCA, BMD of the loss of 1% per year, which was significantly
lumbar spine was lower than expected when higher than the bone loss rate expected in a
compared to normal values adjusted for age, matched population. Duration of IBD,
sex and ethnic group at entry and after alkaline phosphatase, aspartate aminotrans
5 years of follow‐up. In this study, 8.6% of ferase, and triglycerides levels, as well as the
patients had lumbar BMD below the fracture Mayo risk score were significantly associated
threshold at entry, and these patients were with a greater rate of bone loss. However, in
older, had a longer duration of inflammatory the multivariate analysis the unique variable
bowel disease (IBD), and more advanced dis that correlated with the rate of bone loss was
ease. Age was the only variable inversely the longer duration of IBD. Another inter
related with baseline lumbar BMD. None of esting finding was the fact that all fragility
the variables predicted progression of the fractures occurred with a BMD T‐score
bone disease. Previous steroid therapy, given lower than −1.5, as previously described for
in 35% of the patients, apparently did not PBC patients. This observation further sup
affect baseline BMD measurements and the ports consideration of treatment in patients
rate of bone loss during the follow‐up. with BMD below this measurement, particu
Moreover, no significant difference in the larly if they have risk factors for osteoporosis
rate of change of BMD was found between or if they are on the waiting list for liver
men and women, neither with respect to the transplantation, since a significant increment
association with IBD. Incident factures were in the risk of fragility fractures is expected
observed in two patients with BMD below after transplantation [13]. Thus, in patients
the fracture threshold. with advanced PSC, the prevalence of osteo
In another more recent bone health evalu porosis was 32.5%, and fractures were present
ation which included 237 patients (42% in 16% of the cases.
female and 58% male) with PSC [5], osteopo
rosis assessed by a BMD T‐score below −2.5
Autoimmune Hepatitis
was found in 15% of patients and occurred
about 25‐fold more frequently than expected There are no confident and conclusive data
for the matched population; 97 patients regarding the prevalence of bone disease in
(41%) had osteopenia (T‐score between −1 patients with AIH, although it is assumed
and −2.5). Severe osteoporosis, defined by that osteoporosis is one of the most frequent
the authors as a Z‐score less than −2, was complications resulting from long‐lasting
found in 33 patients, a prevalence that was corticosteroid therapy. It is estimated that
6.1 times higher than the expected in an osteoporosis and vertebral fracture related
224 Section III Specific Clinical Challenges
(b)
HPT secondary to vitamin D deficiency
Transient
Bilirubin; RANKL/OPG system?
increased bone
resorption Unknown
Figure 12.2 (a) Pathogenesis of osteoporosis in primary biliary cholangitis. Low bone formation is the main
mechanism for osteoporosis, related to the detrimental effects of retained cholestatic substances and
sclerostin. Other factors such as IGF‐1 may influence bone formation. (b) Transient increased bone resorption is
the other mechanism, with vitamin D deficiency and alterations in the RANKL/OPG system. HPT,
hyperparathyroidism. (c) Bone matrix protein alterations related to low vitamin K and some gene
polymorphisms also contribute to osteoporosis.
Recent data have proposed that increased the cause of increased bone turnover in some
sclerostin, a key regulator of the Wnt/β‐ patients. Osteoprotegerin (OPG) and
catenin signaling pathway mainly produced receptor activator of nuclear factor kappa‐B
by osteocytes and which therefore regulates ligand (RANKL) can be involved in the
bone formation, may be involved in the path increased bone resorption, although at pre
ogenesis of osteoporosis in PBC. Vitamin K sent their role remains uncertain.
deficiency may be another additional factor Nevertheless, an increased RANKL/OPG
for osteoporosis based on some studies pro ratio, resulting in high bone resorption and
posing that vitamin K deficiency causes ultimately bone loss, has been observed in
reductions in BMD and increases the frac cirrhotic patients. In addition, serum from
ture risk. Likewise, vitamin K supplementa jaundiced patients upregulates RANKL/
tion has been associated with improvements OPG gene expression, a system intimately
in BMD. The influence of some gene poly involved in osteoblast‐induced osteoclasto
morphisms has also been assessed in patients genesis. The role of some cytokines such as
with PBC. However, these polymorphisms interleukin (IL)‐1 and IL‐6 and tumor
either do not take part in or have a minor necrosis factor (TNF)‐α, which directly or
effect on the development of osteoporosis in indirectly activates the osteoclasts, and the
this cholestatic disease. role of specific adipokines in the pathogen
Elevated bone resorption may also be esis of osteoporosis in patients with PBC and
involved in the pathogenesis of osteoporosis PSC remains to be established, although a
in end‐stage cholestatic diseases. Accor negative correlation between adiponectin
dingly, reduced trabecular thickness and and BMD has been reported.
increased bone turnover proportional to the The pathogenesis of osteoporosis induced
severity of hepatic dysfunction and chole by corticosteroids is mostly relevant for
stasis have been reported. Calcium and patients with AIH. Reduced bone formation
vitamin D deficiencies leading to secondary is the critical process. Even at low doses, cor
hyperparathyroidism have been proposed as ticosteroids have been shown to rapidly
226 Section III Specific Clinical Challenges
s uppress several indices of osteoblast activity, Table 12.2 Risk factors for bone loss.
including serum markers of bone formation
Alcohol abuse
such as serum procollagen type 1 N‐terminal
propeptide (P1NP) and osteocalcin. More Smoking
over, corticosteroid excess inhibits both oste Body mass index <19 kg/m2
oblast differentiation and function while at Male hypogonadism
the same time inducing osteoblast apoptosis, Early menopause
resulting in rapid and deep suppression of
Secondary amenorrhea of more than 6 months
bone formation. Corticosteroids have early
Family history of osteoporotic fracture
effects on bone resorption as well and there
fore they not only compromise the ability of Treatment with corticosteroids (5 mg/day or more
of prednisone for 3 months or longer)
the skeleton to form new bone but also dis
turb the regulatory balance between bone Advanced age
formation and resorption, which finally leads
to a reduction in bone mass and an increase
in fracture risk. In addition to the direct history of bone fragility fractures, and
effects of corticosteroids on osteoblasts, treatment with glucocorticoids (>5 mg of
osteoclasts and osteocytes, they may indi prednisone over more than 3 months)
rectly affect bone metabolism and fracture (Table 12.2).
risk through effects on gonadal hormones Blood levels of calcium, phosphorus,
and the neuromuscular system (as indicated 25‐hydroxyvitamin D and parathyroid hor
by an increased propensity for falls in corti mone should be measured at diagnosis and at
costeroid‐treated patients). The severity and 1‐ or 2‐year intervals. This is important since
advanced stage of AIH can also contribute to vitamin D (25‐hydroxyvitamin D) levels
decreasing bone mass and increasing frac below 20 ng/ml have been reported in
ture risk. 64–92% of patients with liver diseases, pre
dominantly in chronic cholestatic condi
tions, and which are usually inversely
Assessment of Bone Disease correlated with more advanced disease.
Consequently, these measurements should
The gold standard for the diagnosis of osteo be evaluated more frequently in patients with
porosis involves evaluation of BMD by dual‐ severe cholestasis. The biochemical markers
energy X‐ray absorptiometry (DXA) and the of bone turnover can be determined, but
assessment of fractures. A BMD T‐score below they are mainly helpful for monitoring the
−2.5 is the criteria for osteoporosis. Osteopenia response to specific therapy for bone disease.
is diagnosed when the T‐score is between −1 Disturbances in thyroid and gonadal function
and −2.5. Severe or “established” osteoporosis should be ruled out in particular cases
refers to individuals who meet densitometric depending on the medical history.
criteria and have one or more fragility fractures. BMD should be repeated after 2–3 years to
BMD should be performed in all patients with detect bone loss in patients with baseline
autoimmune liver diseases. Lateral X‐rays of normal or nearly normal range. However,
the dorsal and lumbar spine should also be car evaluation should be performed at a shorter
ried out to reveal vertebral fractures. interval of approximately 1–2 years in patients
In all the patients with autoimmune liver with more than one additional risk factor for
diseases, risk factors for low bone mass and osteoporosis in patients with deep chole
fractures should be assessed, including high stasis, and in those treated with high doses of
alcohol intake and tobacco abuse, BMI lower corticosteroids. This strategy is also recom
than 19 kg/m2, early menopause, secondary mended for patients with late‐stage disease
amenorrhea of more than 6 months, family and those eligible for transplantation.
Chapter 12 Bone Health in Patients with Autoimmune Liver Diseases 227
Diagnosis Management
DXA
lumbar spine and hip DXA Treatment
Repeat in 2–3 yrs
Normal
Ca + Vitamin Da
T-score > –1.5
Ca + Vitamin Da
Repeat in 1–2 yrsb
Osteoporosis Bisphosphonates
T-score ≤ –2.5 New agents
Figure 12.3 Diagnosis and management of osteoporosis in primary biliary cholangitis. a Calcium (1000–
1500 mg/day) and vitamin D (400–800 IU/day or 260 μg 25‐hydroxyvitamin D every 2 weeks) to preserve
normal levels. b According to the severity of cholestasis. c Depending on additional risk factors. DXA,
dual‐energy X‐ray absorptiometry.
228 Section III Specific Clinical Challenges
basically because of the low number of trials administration was associated with higher
and the few patients treated (Table 12.3). BMD values in patients with higher indices
Nonetheless, it has been demonstrated that of cholestasis. Ibandronate is also effective at
cyclic administration of etidronate is able to increasing bone mass. Accordingly, in a
prevent bone loss after 2 years of treatment randomized trial comparing monthly iban
and that alendronate increases bone mass in dronate and weekly alendronate for osteopo
patients with PBC, effects which are rosis in PBC, both treatments resulted in a
comparable to those reported in osteopo significant increase in BMD at the lumbar
rosis due to other causes. A placebo‐con spine after 2 years of therapy (Figure 12.4)
trolled trial of alendronate (70 mg/week) [21]. The mean percentage change was 4.5%
indicates that this bisphosphonate is able to and 5.7% at the lumbar spine for alendro
increase bone mass after one year with no or nate and ibandronate, respectively. Changes
minor adverse effects. The BMD of patients in bone markers were similar in both groups
treated with alendronate increased signifi and one patient receiving alendronate devel
cantly at both the lumbar spine and proximal oped a new vertebral fracture. Adherence to
femur compared with placebo. The effects of therapy was higher with ibandronate. Neither
weekly alendronate on BMD were more treatment impaired liver function or chole
patent at the lumbar spine, and greater than stasis. Serious adverse effects were not
with the daily alendronate regimen. Our observed and potential unfavorable effects of
results of once‐weekly alendronate 70 mg in oral bisphophonates, such as esophagitis,
PBC indicate that this regimen is more effec were not reported, likely because the patients
tive and has a better tolerability profile than were carefully instructed to take the medica
daily dosing in the treatment of low bone tion on an empty stomach, swallowing an
mass in patients treated for 1 year. An inter adequate amount of water and remaining in
esting observation was that alendronate an upright position for at least 30 minutes.
Table 12.3 Effect of different oral bisphosphonates on lumbar spine and femoral T‐scores after 1 or 2 years
of therapy in patients with primary biliary cholangitis.
Percent Percent
lumbar femoral
No. of
Dose patients 1 year 2 years 1 year 2 years
Etidronate
Guañabens et al. (1997) [16] 400 mg/day, 13 0.5 1.3
Wolfhagen et al. (1997) [17] cycles 3 months 6 1
Lindor et al. (2000) [18] 29 0.7 1 1.4 0.5
Guañabens et al. (2003) [19] 13 1.9 0.4
Alendronate
Guañabens et al. (2003) [19] 10 mg/day 13 5.8 3.9
Guañabens et al. (2005) [3] 70 mg/week 16 3.3 1.2
Guañabens et al. (2005) [3] 10 mg/day 10 1.2 −0.3
Zein et al. (2005) [20] 70 mg/week 15 10.4 0.4
Guañabens et al. (2013) [21] 70 mg/week 19 4.5 2.7
Ibandronate
Guañabens et al. (2013) [21] 150 mg/month 14 5.7 1.2
Chapter 12 Bone Health in Patients with Autoimmune Liver Diseases 229
5
% change
0
Basal 6 12 18 24
Month
Figure 12.4 Percentage changes in lumbar bone mineral density (BMD) with respect to baseline values for
patients with primary biliary cholangitis treated with ibandronate (broken line) or alendronate (solid line).
No significant differences in BMD change were observed between treatments for each time period, but the
increase in BMD was significant in each arm from 6 months.
There is little information on the effect of patients included in one trial suggest that
parenteral bisphosphonates in PBC. One hormone therapy may worsen cholestasis.
retrospective study assessed the effect of
Despite these fairly positive results, currently
zoledronic acid, pamidronate disodium, or hormone therapy is not considered the most
ibandronate sodium in PBC patients with appropriate treatment for osteoporosis as
osteoporosis. The percentage change of there are other efficacious non‐hormonal
lumbar spine and hip BMD was 2.9% and agents such as bisphosphonates.
0.4%, respectively. No serious adverse events The efficacy and safety of raloxifene, a
were found. Although the number of patients selective estrogen receptor modulator, has
treated was very small and with a diversity of been assessed in a pilot study including
regimens, parenteral bisphosphonates stabi nine postmenopausal women with PBC.
lized BMD in about half of the patients. More Compared to baseline, lumbar spine BMD
prospective studies are needed to evaluate improved significantly after 1 year of
the efficacy of specific parenteral bisphos therapy but not in matched controls, and
phonates in patients with PBC and osteopo no significant effect was observed in femoral
rosis. No information is available regarding neck.
bone disease therapy in patients with PSC Preliminary results with denosumab, a fully
and AIH. human IgG2 monoclonal antibody directed
against RANKL, in 13 patients with PBC indi
Other Agents cate that lumbar spine T‐score significantly
There have been some studies indicating that improved after 12 months, along with changes
hormonal treatment prevents bone loss or in bone alkaline phosphatase and tartrate‐
even increases BMD in patients with PBC. resistant acid phosphatase 5b, as markers of
Other studies have also confirmed that trans bone formation and osteoclast number,
dermal estradiol and medroxyprogesterone respectively. During the treatment period,
plus calcium and vitamin D supplements humeral fracture at 11 months and latent
increase both lumbar spine and femoral hypocalcemia was observed in two patients.
BMD after 2 years, with no severe side Data are very limited for definite recommen
effects. However, the individual data of the dations. The effect of other agents, such as
230 Section III Specific Clinical Challenges
anabolic therapy with teriparatide that is used ndetermined, although it is assumed that
u
in postmenopausal osteoporosis, also war osteoporosis is one of the most frequent
rants additional investigation in patients with complications of corticosteroid treatment.
autoimmune cholestatic disease. DXA is the usual diagnostic procedure
and should be performed in all patients.
Lateral X‐ray of the dorsal and lumbar
Summary spine should also be carried out to disclose
vertebral fractures, and laboratory
Osteoporosis is a complication frequently measurements of circulating levels of calc
observed in patients with autoimmune liver ium, phosphorus, 25‐hydroxyvitamin D
diseases, particularly PBC. The pathogenesis and parathyroid hormone should be
of osteoporosis is mainly characterized by assessed at diagnosis.
low bone formation, although increased For the prevention and treatment of osteo
bone resorption has also been described in porosis good nutrition is required, as are the
patients with severe cholestasis and advanced suppression of risk factors for o steoporosis
liver disease. Osteoporosis is associated with and the administration of calcium and vitamin
the severity of liver disease, older age and the D supplements. There is no specific treatment
duration of cholestasis. Patients with osteo for bone loss, although it has been demon
porosis and those with a lumbar spine T‐score strated that oral bisphosphonates, especially
below −1.5 have a high risk for fractures, par weekly alendronate and monthly ibandronate,
ticularly of the spine. Osteomalacia, charac are effective at increasing BMD. The potential
terized by poor bone mineralization, is very of parenteral bisphosphonates such as zole
uncommon and has only been reported in a dronic acid or the efficacy of newer agents
few patients with advanced PBC and severe that are now prescribed in patients with post
intestinal malabsorption, and in geographic menopausal osteoporosis and at high risk for
areas with decreased sunlight expo sure. developing fractures needs further evaluation
Information regarding the prevalence in patients with autoimmune liver diseases
of bone disease in patients with AIH is and low bone mass.
Chapter 12 Bone Health in Patients with Autoimmune Liver Diseases 231
References
1 Guañabens, N. and Parés, A. (2010). Liver D receptor gene polymorphisms and bone
and bone. Arch. Biochem. Biophys. 503: mass in primary biliary cirrhosis.
84–94. Hepatology 33 (3): 554–560.
2 Leslie, W.D., Bernstein, C.N., Leboff, M.S. 11 Solerio, E., Isaia, G., Innarella, R. et al.
et al. (2003). AGA technical review on (2003). Osteoporosis: still a typical
osteoporosis in hepatic disorders. complication of primary biliary cirrhosis?
Gastroenterology 125: 941–966. Dig. Liver Dis. 35 (5): 339–346.
3 Pares, A. and Guañabens, N. (2018). 12 Guañabens, N., Parés, A., Ros, I. et al.
Primary biliary cholangitis and bone (2005). Severity of cholestasis and
disease. Best Pract. Res. Clin. Gastroenterol. advanced histological stage but not
34–35: 63–70. menopausal status are the major risk
4 Collier, J. (2007). Bone disorders in chronic factors for osteoporosis in primary biliary
liver disease. Hepatology 46: 1271–1278. cirrhosis. J. Hepatol. 42: 573–577.
5 Angulo, P., Grandison, G.A., Fong, D.G. 13 Guichelaar, M.M., Kendall, R., Malinchoc,
et al. (2011). Bone disease in patients with M., and Hay, J.E. (2006). Bone mineral
primary sclerosing cholangitis. density before and after OLT: long‐term
Gastroenterology 140: 180–188. follow‐up and predictive factors. Liver
6 Guañabens, N., Parés, A., Navasa, M. et al. Transpl. 12: 1390–1402.
(1994). Cyclosporin A increases the 14 Guañabens, N., Cerdá, D., Monegal, A.
biochemical markers of bone remodeling in et al. (2010). Low bone mass and severity of
primary biliary cirrhosis. J. Hepatol. 21 (1): cholestasis affect fracture risk in patients
24–28. with primary biliary cirrhosis.
7 Springer, J.E., Cole, D.E., Rubin, L.A. et al. Gastroenterology 138 (7): 2348–2356.
(2000). Vitamin D‐receptor genotypes as 15 Seki, A., Ikeda, F., Miyatake, H. et al.
independent genetic predictors of (2017). Risk of secondary osteoporosis
decreased bone mineral density in primary due to lobular cholestasis in non‐cirrhotic
biliary cirrhosis. Gastroenterology 118 (1): primary biliary cholangitis.
145–151. J. Gastroenterol. Hepatol. 32 (9):
8 Menon, K.V., Angulo, P., Weston, S. et al. 1611–1616.
(2001). Bone disease in primary biliary 16 Guañabens, N., Parés, A., Monegal, A.
cirrhosis: independent indicators and rate et al. (1997). Etidronate versus fluoride
of progression. J. Hepatol. 35 (3): 316–323. for treatment ofosteopenia in primary
9 Newton, J., Francis, R., Prince, M. et al. biliary cirrhosis: preliminary results
(2001). Osteoporosis in primary biliary after 2 years. Gastroenterology 113 (1):
cirrhosis revisited. Gut 49 (2): 282–287. 219–24.
10 Parés, A., Guañabens, N., Alvarez, L. et al. 17 Wolfhagen, F.H., van Buuren, H.R., den
(2001). Collagen type Ialpha1 and vitamin Ouden, J.W. et al. (1997). Cyclical
232 Section III Specific Clinical Challenges
Section IV
13
Abstract
Autoimmune liver diseases (ALIDs) primary biliary cholangitis (PBC), primary sclerosing cholan-
gitis (PSC) and autoimmune hepatitis (AIH) recur after liver transplantation (LT). This chapter
reviews indications for LT in patients with AILD, and the frequency and clinical impact of recur-
rent AILD after LT. Recurrent PBC is diagnosed on the basis of liver histology and may be seen with
normal liver tests; antimitochondrial antibodies recur irrespective of histologic recurrence.
Recurrent PSC is usually diagnosed on the basis of the demonstration of multiple nonanastomotic
strictures in the absence of other risk factors. Recurrent AIH may occur and is diagnosed on the
basis of characteristic autoantibodies and liver histology. Recurrence of AILDs reduces odds of
graft and patient survival, and management strategies are needed to prevent AILD recurrence or
reduce its negative effects.
Key Points
●● Autoimmune liver diseases – primary recurrence. Graft loss is unusual and
biliary cholangitis (PBC), primary sclerosing treatment with ursodeoxycholic acid may
cholangitis (PSC) and autoimmune hepatitis slow progression. Recurrence appears
(AIH) – recur after liver transplantation. more frequent for those on tacrolimus‐
●● The reported prevalence of recurrence var- based immunosuppression compared
ies between series, in part because of varia- with ciclosporin.
tion in detection and diagnostic criteria. ●● Recurrent PSC is usually diagnosed on the
●● Recurrent PBC is diagnosed on the basis basis of the demonstration of multiple
of liver histology and may be seen with non‐anastomotic strictures in the absence
normal liver tests; anti‐mitochondrial of other risk factors (such as any alternate
antibodies recur irrespective of histologic cause of ischemic cholangiopathy).
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
236 Section IV Transplantation and Its Role in Autoimmune Liver Disease
Although controversial, recurrent PSC is ●● Recurrent AIH may occur and is diag-
less likely in those with a colectomy either nosed on the basis of characteristic auto-
before or at the time of transplant. There antibodies and liver histology. The natural
is no effective treatment other than history may be improved by optimisation
retransplantation. of immunosuppression.
Table 13.1 Diagnostic criteria for recurrence autoimmune liver diseases after liver transplantation.
presumed associations are of questionable an earlier and more aggressive course [5].
significance. Azathioprine was found to protect against
Younger age at diagnosis and at the time of rPBC in some studies, though one of them did
LT is associated with higher risk of rPBC [7]. not adjust for confounding variables. There is
Male recipient was significant in one study, but less compelling evidence that steroids may also
this finding could not be replicated in others play a role in preventing rPBC, but his
studies [5,6]. Biochemical markers of chole- association is usually lost in multivariable anal-
stasis within the first year after LT are also ysis after adjusting for ciclosporin (Table 13.2).
associated with rPBC [7]. Patients with overlap
syndrome (PBC/AIH) on the native liver as an
Treatment of PBC Recurrence After LT
indication for LT are also at risk of recurrence.
Regarding human leukocyte antigen (HLA) The standard of treatment for rPBC is
status, there is conflicting evidence as some UDCA. However, unlike in pre‐LT patients,
studies have found an association between the use of UDCA has not been shown to
HLA mismatching or specific HLA and rPBC, modify the natural history of rPBC in the
and others have not. The polymorphism short to middle term, even after improving
rs62270414 at the IL12A locus has been asso- alkaline phosphatase (ALP) levels. Long‐
ciated with a higher risk of rPBC [6]. term studies showing a benefit on progres-
Tacrolimus when compared to ciclosporin sion of rPBC are needed. In addition, we
has been associated with rPBC, and also with lack data regarding the use of second‐line
Table 13.2 Frequency and risk factors for recurrence of autoimmune liver disease after liver transplantation.
Frequency of
Disease recurrence Risk factors for recurrence
Pre-Transplantation Strategies
Figure 13.1 Strategies to reduce the risk of autoimmune liver disease recurrence after liver transplantation.
ALT, alanine aminotransferase; AST, aspartate aminotransferase; UDCA, ursodeoxycholic acid;
IBD, inflammatory bowel disease; EDC, extended donor criteria; DCD, donation after cardiac death.
Source: Montano‐Loza et al. [2]. Reproduced with permission of John Wiley and Sons.
References
1 Adam, R., Karam, V., Delvart, V. et al. 8 Hildebrand, T., Pannicke, N., Dechene, A.
(2012). Evolution of indications and et al. (2016). Biliary strictures and
results of liver transplantation in Europe. recurrence after liver transplantation for
A report from the European Liver primary sclerosing cholangitis: a
Transplant Registry (ELTR). J. Hepatol. retrospective multicenter analysis. Liver
57 (3): 675–688. Transpl. 22 (1): 42–52.
2 Montano‐Loza, A.J., Bhanji, R.A., 9 Trivedi, P.J., Reece, J., Laing, R.W. et al.
Wasilenko, S., and Mason, A.L. (2017). (2018). The impact of ileal pouch–anal
Systematic review: recurrent autoimmune anastomosis on graft survival following
liver diseases after liver transplantation. liver transplantation for primary sclerosing
Aliment. Pharmacol. Ther. 45 (4): 485–500. cholangitis. Aliment. Pharmacol. Ther. 48
3 Singal, A.K., Guturu, P., Hmoud, B. et al. (3): 322–332.
(2013). Evolving frequency and outcomes of 10 Ravikumar, R., Tsochatzis, E., Jose, S. et al.
liver transplantation based on etiology of liver (2015). Risk factors for recurrent primary
disease. Transplantation 95 (5): 755–760. sclerosing cholangitis after liver transplant
4 Webb, G.J., Rana, A., Hodson, J. et al. (2018). ation. J. Hepatol. 63 (5): 1139–1146.
Twenty‐year comparative analysis of 11 Joshi, D., Bjarnason, I., Belgaumkar, A.
patients with autoimmune liver diseases on et al. (2013). The impact of inflammatory
transplant waitlists. Clin. Gastroenterol. bowel disease post‐liver transplantation for
Hepatol. 16 (2): 278–287.e7. primary sclerosing cholangitis. Liver Int. 33
5 Montano‐Loza, A.J., Wasilenko, S., Bintner, (1): 53–61.
J., and Mason, A.L. (2010). Cyclosporine A 12 Lindstrom, L., Jorgensen, K.K., Boberg,
protects against primary biliary cirrhosis K.M. et al. (2018). Risk factors and
recurrence after liver transplantation. Am. J. prognosis for recurrent primary sclerosing
Transplant. 10 (4): 852–858. cholangitis after liver transplantation: a
6 Bosch, A., Dumortier, J., Maucort‐Boulch, Nordic multicentre study. Scand. J.
D. et al. (2015). Preventive administration of Gastroenterol. 53: 297–304.
UDCA after liver transplantation for 13 Trivedi, P.J., Scalera, I., Slaney, E. et al.
primary biliary cirrhosis is associated with a (2017). Clinical outcomes of donation after
lower risk of disease recurrence. J. Hepatol. circulatory death liver transplantation in
63 (6): 1449–1458. primary sclerosing cholangitis. J. Hepatol.
7 Montano‐Loza, A.J., Hansen, B.E., 67 (5): 957–965.
Corpechot, C. et al. (2019). Factors 14 Montano‐Loza, A.J., Mason, A.L., Ma, M.
associated with recurrence of primary et al. (2009). Risk factors for recurrence of
biliary cholangitis after liver transplantation autoimmune hepatitis after liver
and effects on graft and patient survival. transplantation. Liver Transpl. 15 (10):
Gastroenterology 156: 96–107.e1. 1254–1261.
Chapter 13 Recurrent Autoimmune Liver Disease and Its Impact on Clinical Practice 245
15 Krishnamoorthy, T.L., Miezynska‐Kurtycz, J., 17 Satapathy, S.K., Jones, O.D., Vanatta, J.M.
Hodson, J. et al. (2016). Longterm et al. (2017). Outcomes of liver transplant
corticosteroid use after liver transplantation recipients with autoimmune liver disease
for autoimmune hepatitis is safe and using long‐term dual immunosuppression
associated with a lower incidence of recurrent regimen without corticosteroid.
disease. Liver Transpl. 22 (1): 34–41. Transplant. Direct 3 (7): e178.
6 Schramm, C., Bubenheim, M., Adam, R.
1 8 Stirnimann, G., Ebadi, M., Czaja, A.J., and
1
et al. (2010). Primary liver transplantation Montano‐Loza, A.J. (2019). Recurrent and
for autoimmune hepatitis: a comparative de novo autoimmune hepatitis. Liver
analysis of the European Liver Transplant Transpl. 25: 152–166.
Registry. Liver Transpl. 16 (4): 461–469.
247
14
Abstract
The three most common autoimmune liver diseases (ALDs) are autoimmune hepatitis (AIH), pri
mary biliary cholangitis (PBC), and primary sclerosing cholangitis (PSC). This chapter discusses
the incidence, presentation, diagnosis, risk factors, and impact on outcomes of recurrence of PBC,
PSC and AIH after liver transplantation (LT). There is an urgent need to prevent recurrence of
ALD in the graft, especially in PSC and PBC, where recurrence is frequently linked to graft failure
and the necessity of retransplantation. While ALD patients are largely diagnosed when the disease
has progressed, those at the very early stages of disease onset can be followed longitudinally for an
extended period with protocol biopsy and blood sampling before and after LT to access disease
recurrence and perhaps identify biomarkers. The enigma of recurrence is an important immuno
logic issue, not only in ALD but also in other solid organs.
Keywords autoimmune hepatitis; blood sampling; graft failure; immunologic issue; liver
transplantation; primary biliary cholangitis; primary sclerosing cholangitis; protocol biopsy
Key Points
●● Recurrent autoimmune liver disease can ●● Understanding those factors that affect
be a significant cause of graft loss. recurrence have important implications
●● Risk factors for, and treatment of, recur for understanding liver autoimmunity;
rence is not well defined. these include the role of innate immunity,
●● Prospective studies are needed to better target organ expression of host antigens,
define risk factors, biomarkers and bystander mechanisms, and genetic simi
treatment. larities between donor and host.
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
248 Section IV Transplantation and Its Role in Autoimmune Liver Disease
ALDs has been progressing, with a better resentation, diagnosis, risk factors, and
p
understanding of etiopathology [1–5] and impact on outcomes of recurrence of PBC,
long‐term outcomes with medical treatment PSC and AIH after LT.
remarkably improved [6], liver transplanta
tion (LT) remains the only and most effective
treatment option for patients with end‐stage Recurrence of PBC
ALDs. According to the registry data from
the USA, Europe and Japan, the 10‐year PBC is a chronic cholestatic liver disease,
survival rates after LT are similar (71–79%) mainly affecting small‐ to medium‐sized
in PBC, whereas those in PSC varied from intrahepatic bile ducts. Middle‐aged women
83% in the USA to 70% in Europe and 57% in are at the highest risk for developing PBC.
Japan [7–9] (Table 14.1). With regard to AIH, The diagnosis of PBC is made by meeting
the registries contain no available data except two of the following three criteria: (i) chronic
for the European registry, which shows elevation of cholestatic liver enzymes, i.e.
5‐ and 10‐year survival rates of 76% and 67%, alkaline phosphatase (ALP) and gamma‐
respectively [7] (Table 14.1). glutamyltransferase (GGT); (ii) detectable
Despite the high survival rate of ALD anti‐mitochondrial antibodies (AMAs); and
patients with LT, disease recurrence also (iii) characteristic histologic findings of
occurs and negatively affects graft as well as chronic non‐suppurative destructive cholan
overall survival. In particular, recurrence of gitis (CNSDC) and granulomas [13]. PBC is a
PSC develops in at least 25% of patients and slowly progressive disease, which can result
is definitely associated with poor outcome in cirrhosis and liver failure without appro
[10]. Although recurrence of PBC was priate treatment. Although ursodeoxycholic
believed to have little impact on graft failure acid (UDCA) improves LT‐free survival and
and prognosis [11], a recent international is recommended as first‐line treatment for
study with a large cohort demonstrated that PBC by clinical practice guidelines or
graft and patient survival are significantly guidance [13–15], about one‐third of patients
impaired by recurrence of PBC [12]. In this are refractory to UDCA and could progress
chapter, we discuss the incidence, to end‐stage liver disease requiring LT. In
Table 14.1 Patient and graft survival at 5 and 10 years after LT in PBC, PSC and AIHa.
recurrence in PBC. Recurrent PBC is studies from Japan demonstrated that the
diagnosed with PBC‐specific histology, such increased frequency of recurrent PBC is
as florid duct lesions and lymphoplasmacytic associated with initial treatment with ciclo
infiltrates. However, other causes of bile duct sporin after LT [22, 27]. However, these two
injury may also mimic the typical findings of studies were performed by the same study
PBC, including ischemia–reperfusion injury, group, and subjects in these two studies
acute cellular rejection, humoral/chronic partially overlapped. Among 516 patients
rejection, drug‐induced liver injury, vascular who underwent LT for end‐stage PBC from
complications, and graft‐versus‐host disease. 1994 to 2010, living donor liver transplant
In addition, recurrence of PBC may develop (LDLT) and deceased donor liver transplant
few or no symptoms and elevation of (DDLT) of both genders were included in the
cholestatic enzymes may be absent. There first study [22]. However, only LDLT and
fore, protocol liver biopsy rather than event‐ only females were included, and patients who
driven biopsy is required for detection of did not survive for 1 year were excluded, in
recurrent PBC. It may have a huge impact on the second study [27]. The first study also
the difference in incidence of recurrent PBC demonstrated that the switch from tacroli
whether protocol biopsy is scheduled or mus to ciclosporin treatment decreased the
event‐driven biopsy is performed. frequency of recurrence. This finding sug
gests that the timing of selection or change of
calcineurin inhibitor and the way immuno
Risk Factors of Recurrent PBC
suppressive agents are used in clinical prac
Although several studies have reported risk tice may partly account for the difference in
factors associated with recurrence of PBC, the association with recurrence between the
most of them have consistently demonstrated East and the West. On the other hand, it is
that use of tacrolimus is definitely associated well known that genetics plays a crucial role
with an increased risk of recurrence [20, 21, in susceptibility of PBC [36], and this
28, 30–32]. For example, a recent study on difference in role of calcineurin inhibitors
785 PBC patients from North America and suggests that genetics may also contribute to
Europe who underwent LT from February recurrence of PBC after LT, as indicated by
1983 to June 2016 indicated that tacrolimus another previous study suggesting the
was linked to recurrence of PBC. While use of influence of the IL12A locus [20].
ciclosporin was protective, the 5‐year proba
bility of PBC recurrence was reported to be
Impact of Recurrent PBC on Long‐term
28% and 11% in patients receiving tacrolimus
Outcomes
and ciclosporin, respectively (p <0.001) [12].
The reasons for this phenomenon still remain Nevertheless, it is believed that recurrence
unclear. Since infection with bacteria or of PBC does not have a significant impact
viruses has been suggested to be associated on long‐term outcomes, such as overall
with tolerance breakdown against autoanti survival. For instance, a retrospective study
gens [34], the difference in PBC recurrence in the UK on 400 PBC patients who received
rates with these two immunosuppressive LT between 1983 and 1999 demonstrated
drugs is likely due to the more potent immu confirmation of recurrent PBC in 68 (17%)
nosuppressive effect of tacrolimus, which subjects but no effect on patient and graft
might facilitate alloreactivity against graft survival during, on average, 56 months of
liver through infection [35]. follow‐up [28]. Recently, a French–Swiss
On the other hand, the role of tacrolimus collaborative study was conducted on 123
as an agent that causes increasing rates of patients with PBC who underwent LT. Their
PBC recurrence does not seem to be the case data showed that recurrence of PBC was
in other ethnicities. Recently, two cohort observed in 48 subjects (53%) with an
Chapter 14 Recurrent Autoimmune Liver Disease and its Scientific Significance 251
average of 11.7 years of follow‐up; neither features differentiate PSC from PBC [10, 38].
recurrence nor preventive UDCA had a Unlike PBC, PSC is more common in males
significant impact on survival [37]. The two than in females; adolescents and young adults
Japanese studies described previously also are at the highest risk for developing PSC.
suggested that there is no impact of recur Disease‐specific biomarkers such as autoanti
rent PBC on survival [22, 27]. Surprisingly, bodies are not available in PSC (compare
another recent study on 785 PBC patients AMA in PBC). Although onion‐skin periduc
from 13 centers in North America and tal fibrosis is known to be a characteristic his
Europe who received LT with a median tologic finding in PSC, this feature is not
follow‐up of 6.9 years (interquartile range frequently observed in liver specimens, mak
6.1–7.9) reported the opposite and unex ing liver biopsy less valuable for diagnosis of
pected result that disease recurrence was PSC. As a result, the diagnosis of PSC is ulti
found in 240 patients (31%) and, impor mately dependent on cholangiography by
tantly, both graft and patient survival were endoscopic retrograde cholangiopancrea
significantly impaired in those with recur tography (ERCP) or magnetic resonance
rent PBC (p = 0.004 and 0.001, respectively) cholangiopancreatography (MRCP), which
[12]. Although the manner of biopsy (pro sometimes can be challenging. Indeed, the
tocol or event‐driven) depended on the natural course of PSC varies greatly depend
center, the median time for recurrence of ing on cases, a number of prognostic indices
PBC was similar in this study, which is by far have been proposed [10, 38], and no con
the largest cohort with long‐term follow‐up sensus on surrogate end points (ALP, bili
on PBC recurrence after LT. It is necessary rubin, histology, non‐invasive fibrosis
to remember that there might be a number markers, etc.) has been uniformly recom
of variables among centers, such as immu mended [39]. Additionally, cholangiocarci
nosuppressant protocols, diagnosis of noma and inflammatory bowel disease (IBD)
recurrent PBC, and the use of UDCA as are frequently observed as concomitant dis
prevention and/or treatment for PBC, and eases in patients with PSC, making definition
these could affect the conclusion. of surrogate end points more complicated.
Nevertheless, it is imperative to determine Regarding treatment, UDCA is probably used
whether recurrent PBC really has a signi as the first‐line drug in most cases with PSC
ficant impact on patient and graft survival. and indeed the efficacy of UDCA has been
Also, since UDCA treatment seems to be extensively studied, but no single drug
effective in improving markers of chole including UDCA has been proven to prolong
stasis, it will not effectively reduce the fre LT‐free survival of patients. It may be simply
quency and risk of recurrence of PBC after because of the ineffectiveness of UDCA, but
LT [37]; therefore, development of effective may possibly be due to lack of properly
strategies to halt recurrence of PBC is an designed clinical studies.
urgent need. In clinical settings when no medical
treatment is effective in patients with end‐
stage PSC, LT becomes the single treatment
Recurrence of PSC option. According to the registry data, the
long‐term outcomes of LT for patients with
PSC is a chronic cholestatic liver disease, PSC in Europe and the USA are comparable
mainly affecting large‐sized intrahepatic and to those in PBC; the respective 5‐ and 10‐
extrahepatic bile ducts. Although PBC and year patient survival rates are 78% and 70%
PSC share several clinical characteristics as in Europe, and 87% and 83% in the USA
cholestatic liver diseases, including elevation [7, 8]. The 5‐year survival in Europe has
of ALP and GGT and possibility of progres improved to 82% in recent years (1999–
sion to biliary cirrhosis, a number of distinct 2009) [7]. However, there is a striking
252 Section IV Transplantation and Its Role in Autoimmune Liver Disease
difference in the survival rates of Japanese failure thereafter. The percentage of LDLT
LT patients, with 5‐ and 10‐year patient was 100% in the Japanese registry [9], and
survival rates of 73% and 57%, respectively 87% (39/45) in a single‐center report in
[9]. Although graft survival rate was not Japan [40].
reported in this registry data, a recent
single‐center study in Japan reported that
Incidence and Diagnosis of Recurrent PSC
the 5‐year graft survival rate was 55.4% [40],
substantially lower than that in Europe Recurrence of PSC after LT is more fre
(69%) and the USA (81%). This marked quently observed compared to PBC. As
difference is partly explained by the fre shown in Table 14.3, PSC recurs in 8.6–44.7%
quent recurrence of PSC in LDLT and graft of patients after LT within a median of almost
1–5 years. It should be noted that lower for several reasons: sample size, extent and
recurrence rates were reported in the earlier depth of the collected dataset, difference in
studies with relatively shorter follow‐up patient populations, and ethnicities [65].
times [41, 44]. In contrast, the highest recur In Japan, where LDLT is dominant, the
rence rate, 44.7% (21/47), was recently recurrence rate of PSC after LT is relatively
reported from a single center, where the higher compared with other regions
follow‐up period of the patients after LT was (Table 14.3). Egawa et al. [56] identified first‐
a median of 10.2 (5.0–20.8) years, the longest degree relative donor as a significant risk
among all cohorts [62]. factor for recurrence. However, this is not the
The diagnosis of PSC is challenging. As case in North America. In a large‐scale North
described previously, there are no PSC‐ American cohort of patients with PSC
specific biomarkers that are available for (n = 307), it was observed that PSC recur
diagnosis and cholangiography by ERCP or rence rate was not significantly different for
MRCP is the most important tool for pre‐ LDLT versus DDLT recipients (p = 0.36). In
transplant diagnosis. After LT, imaging this study, the risk factors identified were
studies of the biliary tree may reflect a high MELD score, biliary complication, chol
variety of physiologic and pathologic condi angiocarcinoma, and high donor age in both
tions that mimic PSC, including organ pres LDLT and DDLT, while first‐degree relative
ervation, ischemia–reperfusion injury, donor was not significantly associated with
hepatic artery stenosis or thrombosis, recurrence [59]. Another recent cohort study
chronic ductopenia rejection, bile duct anas from Canada on 36 PSC patients who
tomosis, ABO incompatibility, and drug‐ received a graft from a first‐degree living‐
induced liver injury. For diagnosis of related donor, 32 PSC patients who received
recurrent PSC, the criteria proposed by a graft from a living but distinct/unrelated
Graziadei et al. [63] should be strictly used. donor, and 70 PSC patients who received a
Accordingly, (i) the diagnosis of PSC must graft from a deceased non‐related donor
be present before LT; (ii) cholangiographic indicated that first‐degree living‐related
findings (biliary stricturing, beading, and donor did not confer a significant increase in
irregularity lasting more than 90 days) and/ recurrent risk of PSC after LT [60].
or histologic features (fibrous cholangitis The impact of IBD on recurrent PSC is
and/or fibro‐obliterative lesions with or another important issue. Active IBD after
without ductopenia, biliary fibrosis, or bil LT is also associated with increasing recur
iary cirrhosis) are required for diagnosis of rence rate of PSC [40, 53, 57, 58, 66]. A
recurrent PSC; and (iii) other conditions recent study in the Czech Republic reported
mimicking PSC need to be excluded [63]. that de novo colitis was present in 25.5%
(12/47) of PSC patients between 17 and
87 months (median 45 months) post LT and
Risk Factors of Recurrent PSC
a significant number of them (11/12) were
A number of risk factors for recurrent PSC also confirmed with recurrent PSC at a
have been considered, including recipient’s age, median of 72 months post LT. [62]. These
gender, history of cholangiocarcinoma, Model observations surely underline the impor
for End‐stage Liver Disease (MELD) score, and tance of the gut–liver axis in PSC patho
status of IBD; and donor and recipient’s age, physiology, and also point to the
gender/cytomegalovirus mismatch and first‐ recommendation of preventable colectomy
degree relative donors [64, 65]. Acute cellular prior to LT [53, 61]. However, another
rejection (steroid‐resistant) is reported to be recent large‐scale cohort study did not show
another risk for recurrent PSC. It is strongly a preventable effect of colectomy for recur
noted, however, that there is great inconsis rence of PSC [57]. Further prospective
tency in the results of these studies, probably studies are warranted to determine whether
254 Section IV Transplantation and Its Role in Autoimmune Liver Disease
recurrence increases as follow‐up time factor for recurrence in one study [86] but
increases after LT [75, 81, 82]. not in others [79–81, 84]. Withdrawal of cor
In the diagnosis of AIH, the revised AIH ticosteroids has been associated with higher
criteria [70] and the simplified score [71] risk of recurrent AIH [79, 82]. A recent study
have been proposed by the International AIH from the UK demonstrated that the 5‐ and
Group. The former is established for identi 10‐year recurrence rate after LT was 6% and
fying “pure” AIH for clinical studies but 11%, respectively, in a cohort consisting of 69
seems to be too complicated for daily patients with AIH, of which 87% were
practice, while the latter consists of only five receiving long‐term maintenance treatment
criteria and is more suitable in clinical set with corticosteroids after LT [78]. Compared
tings. However, both are not validated for the with the recurrence rate of 27% in their
diagnosis of AIH developed in the graft. previous report in 1999, when patients did
Meanwhile, it is reasonable to assume that not receive long‐term corticosteroid therapy
the criteria for a diagnosis of recurrent AIH [79], the authors concluded that long‐term
are the same as for a diagnosis of AIH before corticosteroid use in combination with
LT: elevation of transaminases, detection of immunosuppressive agents was associated
autoantibodies (ANA, ASMA), elevation of with a lower frequency of recurrence.
IgG, and histologic findings including inter
face hepatitis or perivenular lymphoplasma Impact of Recurrent AIH on Long‐
cytic infiltrates. A comprehensive analysis of term Outcomes
the clinical phenotypes is necessary in order In general, progression to cirrhosis and graft
to diagnose disease recurrence post LT [85]. failure requiring retransplantation is
uncommon even when AIH recurs in the graft
[64]. However, the mechanisms causing recur
Risk Factors of Recurrent AIH
rent AIH after LT are unclear. Furthermore,
A number of factors are reported to be asso there are also substantial differences between
ciated with recurrence of AIH, including adult and pediatric de novo AIH, which sub
severity of pre‐transplant AIH [74, 81]. HLA stantiates the need for more precise diagnostic
locus mismatching is identified as a risk guidelines in this area [87, 88]. When
256 Section IV Transplantation and Its Role in Autoimmune Liver Disease
r ecurrence develops in the graft, the strength Furthermore, these prospective cohorts
of immunosuppression should be reinforced would be valuable resources for further under
with readministration or increased dosing of standing the etiologic mechanisms of disease.
corticosteroids or addition of other immuno While ALD patients are largely diagnosed
suppressive agents. when the disease has progressed, those at the
very early stages of disease onset can be fol
lowed longitudinally for an extended period
Concluding Remarks with protocol (or even event‐driven) biopsy
and blood sampling before and after LT to
The shortage of donor livers is a major issue access disease recurrence and perhaps iden
for patients awaiting LT worldwide [89]. tify biomarkers. Again, the establishment of
Thus, there is an urgent need to prevent well‐designed and organized prospective reg
recurrence of ALD in the graft, especially in istries with clinical information as well as
PSC and PBC, where recurrence is frequently clinical samples are critical not only to
linked to graft failure and the necessity of improve the outcomes of post‐transplant
retransplantation. Furthermore, risk factors patients, but also to unveil the etiologic myths
associated with an increased frequency of of ALD. The enigma of recurrence is an
ALD recurrence are not clearly defined and important immunologic issue, not only in
robust approaches for the prevention of ALD ALD but also in other solid organs. Factors to
recurrence are not available. Large‐scale and be considered include innate immunity, target
multicenter studies with prospective cohorts organ expression of host antigens, bystander
of patients after LT are needed to draw con mechanisms, and genetic similarities between
clusive results, under the auspices of interna donor and host. Further studies are required
tional collaborative consortia. and the mechanisms may vary by disease.
Chapter 14 Recurrent Autoimmune Liver Disease and its Scientific Significance 257
References
16 Nevens, F., Andreone, P., Mazzella, G. liver transplantation in 100 patients. Clin.
et al. (2016). A placebo‐controlled trial of Transpl. 20: 211–220.
obeticholic acid in primary biliary 26 Khettry, U., Anand, N., Faul, P.N. et al.
cholangitis. N. Engl. J. Med. 375: 631–643. (2003). Liver transplantation for primary
17 Harms, M.H., Janssen, Q.P., Adam, R. et al. biliary cirrhosis: a long‐term pathologic
(2019). Trends in liver transplantation for study. Liver Transpl. 9: 87–96.
primary biliary cholangitis in Europe over 27 Kogiso, T., Egawa, H., Teramukai, S. et al.
the past three decades. Aliment. (2017). Risk factors for recurrence of
Pharmacol. Ther. 49: 285–295. primary biliary cholangitis after liver
18 Boonstra, K., Beuers, U., and Ponsioen, transplantation in female patients: a
C.Y. (2012). Epidemiology of primary Japanese multicenter retrospective study.
sclerosing cholangitis and primary biliary Hepatol. Commun. 1: 394–405.
cirrhosis: a systematic review. J. Hepatol. 28 Liermann Garcia, R.F., Evangelista Garcia,
56: 1181–1188. C., McMaster, P., and Neuberger, J. (2001).
19 Bhanji, R.A., Mason, A.L., Girgis, S., and Transplantation for primary biliary cirrhosis:
Montano‐Loza, A.J. (2013). Liver retrospective analysis of 400 patients in a
transplantation for overlap syndromes of single center. Hepatology 33: 22–27.
autoimmune liver diseases. Liver Int. 33: 29 Manousou, P., Arvaniti, V., Tsochatzis, E.
210–219. et al. (2010). Primary biliary cirrhosis after
20 Carbone, M., Mells, G.F., Alexander, G.J. liver transplantation: influence of
et al. (2013). Calcineurin inhibitors and the immunosuppression and human leukocyte
IL12A locus influence risk of recurrent antigen locus disparity. Liver Transpl. 16:
primary biliary cirrhosis after liver 64–73.
transplantation. Am. J. Transplant. 13: 30 Montano‐Loza, A.J., Wasilenko, S., Bintner,
1110–1111. J., and Mason, A.L. (2010). Cyclosporine A
21 Charatcharoenwitthaya, P., Pimentel, S., protects against primary biliary cirrhosis
Talwalkar, J.A. et al. (2007). Long‐term recurrence after liver transplantation. Am.
survival and impact of ursodeoxycholic J. Transplant. 10: 852–858.
acid treatment for recurrent primary 31 Neuberger, J., Gunson, B., Hubscher, S., and
biliary cirrhosis after liver transplantation. Nightingale, P. (2004). Immunosuppression
Liver Transpl. 13: 1236–1245. affects the rate of recurrent primary biliary
22 Egawa, H., Sakisaka, S., Teramukai, S. et al. cirrhosis after liver transplantation. Liver
(2016). Long‐term outcomes of living‐ Transpl. 10: 488–491.
donor liver transplantation for primary 32 Sanchez, E.Q., Levy, M.F., Goldstein, R.M.
biliary cirrhosis: a Japanese multicenter et al. (2003). The changing clinical
study. Am. J. Transplant. 16: 1248–1257. presentation of recurrent primary biliary
23 Guy, J.E., Qian, P., Lowell, J.A., and Peters, cirrhosis after liver transplantation.
M.G. (2005). Recurrent primary biliary Transplantation 76: 1583–1588.
cirrhosis: peritransplant factors and 33 Sylvestre, P.B., Batts, K.P., Burgart, L.J. et al.
ursodeoxycholic acid treatment post‐liver (2003). Recurrence of primary biliary
transplant. Liver Transpl. 11: 1252–1257. cirrhosis after liver transplantation:
24 Hytiroglou, P., Gutierrez, J.A., Freni, M. histologic estimate of incidence and natural
et al. (2009). Recurrence of primary biliary history. Liver Transpl. 9: 1086–1093.
cirrhosis and development of autoimmune 34 Tanaka, A., Leung, P.S.C., and Gershwin,
hepatitis after liver transplant: a blind M.E. (2019). Pathogen infections and
histologic study. Hepatol. Res. 39: 577–584. primary biliary cholangitis. Clin. Exp.
25 Jacob, D.A., Neumann, U.P., Bahra, M. et al. Immunol. 195: 25–34.
(2006). Long‐term follow‐up after 35 Montano‐Loza, A.J. and Mason, A.L.
recurrence of primary biliary cirrhosis after (2017). Recurrence of primary biliary
Chapter 14 Recurrent Autoimmune Liver Disease and its Scientific Significance 259
liver transplantation for autoimmune and 63 Graziadei, I.W., Wiesner, R.H., Batts, K.P.
cholestatic liver diseases: an analysis of the et al. (1999). Recurrence of primary
UNOS database. J. Gastrointest. Surg. 14: sclerosing cholangitis following liver
1362–1369. transplantation. Hepatology 29: 1050–1056.
55 Moncrief, K.J., Savu, A., Ma, M.M. et al. 64 Montano‐Loza, A.J., Bhanji, R.A.,
(2010). The natural history of inflammatory Wasilenko, S., and Mason, A.L. (2017).
bowel disease and primary sclerosing Systematic review: recurrent
cholangitis after liver transplantation: a autoimmune liver diseases after liver
single‐centre experience. Can. J. transplantation. Aliment. Pharmacol.
Gastroenterol. 24: 40–46. Ther. 45: 485–500.
56 Egawa, H., Ueda, Y., Ichida, T. et al. (2011). 65 Visseren, T. and Darwish Murad, S. (2017).
Risk factors for recurrence of primary Recurrence of primary sclerosing
sclerosing cholangitis after living donor cholangitis, primary biliary cholangitis and
liver transplantation in Japanese registry. auto‐immune hepatitis after liver
Am. J. Transplant. 11: 518–527. transplantation. Best Pract. Res. Clin.
57 Ravikumar, R., Tsochatzis, E., Jose, S. et al. Gastroenterol. 31: 187–198.
(2015). Risk factors for recurrent primary 66 Cholongitas, E. and Burroughs, A.K.
sclerosing cholangitis after liver (2012). Recurrence of autoimmune liver
transplantation. J. Hepatol. 63: 1139–1146. diseases after liver transplantation: clinical
58 Hildebrand, T., Pannicke, N., Dechene, A. aspects. Auto Immun. Highlights 3:
et al. (2016). Biliary strictures and 113–118.
recurrence after liver transplantation for 67 Yamada, Y., Hoshino, K., Fuchimoto, Y.
primary sclerosing cholangitis: a et al. (2018). Rituximab induction to
retrospective multicenter analysis. Liver prevent the recurrence of PSC after liver
Transpl. 22: 42–52. transplantation: the lessons learned from
59 Gordon, F.D., Goldberg, D.S., Goodrich, ABO‐incompatible living donor liver
N.P. et al. (2016). Recurrent primary transplantation. Transplant. Direct 4: e342.
sclerosing cholangitis in the Adult‐to‐Adult 68 Manns, M.P., Lohse, A.W., and Vergani, D.
Living Donor Liver Transplantation Cohort (2015). Autoimmune hepatitis: Update
Study: comparison of risk factors between 2015. J. Hepatol. 62: S100–S111.
living and deceased donor recipients. Liver 69 Floreani, A., Restrepo‐Jimenez, P., Secchi,
Transpl. 22: 1214–1222. M.F. et al. (2018). Etiopathogenesis of
60 Aravinthan, A.D., Doyle, A.C., Issachar, A. autoimmune hepatitis. J. Autoimmun. 95:
et al. (2016). First‐degree living‐related 133–143.
donor liver transplantation in autoimmune 70 Alvarez, F., Berg, P., Bianchi, F. et al. (1999).
liver diseases. Am. J. Transplant. 16: International Autoimmune Hepatitis
3512–3521. Group Report: review of criteria for
61 Lindstrom, L., Jorgensen, K.K., Boberg, diagnosis of autoimmune hepatitis.
K.M. et al. (2018). Risk factors and J. Hepatol. 31: 929–938.
prognosis for recurrent primary sclerosing 71 Hennes, E.M., Zeniya, M., Czaja, A.J. et al.
cholangitis after liver transplantation: a (2008). Simplified criteria for the diagnosis
Nordic multicentre study. Scand. J. of autoimmune hepatitis. Hepatology 48:
Gastroenterol. 53: 297–304. 169–176.
62 Bajer, L., Slavcev, A., Macinga, P. et al. 72 European Association for the Study of the
(2018). Risk of recurrence of primary Liver (2015). EASL clinical practice
sclerosing cholangitis after liver guidelines: autoimmune hepatitis.
transplantation is associated with de novo J. Hepatol. 63: 971–1004.
inflammatory bowel disease. World J. 73 Yoshizawa, K., Matsumoto, A., Ichijo, T.
Gastroenterol. 24: 4939–4949. et al. (2012). Long‐term outcome of
Chapter 14 Recurrent Autoimmune Liver Disease and its Scientific Significance 261
Section V
15
Abstract
In a landmark review, J. Woodward and J. Neuberger emphasized that true overlaps should be dif
ferentiated from simple crossover or outlier syndromes. Moderate to severe interface hepatitis is a
fundamental component and histology is vital in evaluating patients with overlap presentation. The
low prevalence of overlap syndrome (OS) has made it impracticable to perform randomized con
trolled trials. For primary biliary cholangitis (PBC)/autoimmune hepatitis (AIH) OS, the European
Association for the Study of the Liver has provided diagnostic criteria and, in most cases, it is pos
sible to define one primary disorder, usually PBC. Primary sclerosing cholangitis/AIH OS is
assumed to exist in a considerable part of mainly young patients with autoimmune liver disease and
longterm progression toward cirrhosis seems to occur in the majority of cases. Treatment of OS is
empiric and includes ursodeoxycholic acid for the cholestatic component and immunosuppressive
agents for the hepatitic component, either simultaneously or sequentially.
Key Points
●● Some patients present with features of ●● The low prevalence of OS (roughly 10% of
both primary biliary cholangitis (PBC) or PBC or PSC) has made it impracticable to
primary sclerosing cholangitis (PSC) and perform randomized controlled trials.
autoimmune hepatitis (AIH) either simul ●● It remains unclear whether this syndrome
taneously or consecutively. forms a distinct entity or is a variant of
●● The term “overlap syndrome” (OS) is used PBC, PSC or AIH.
to describe these settings but lack of ●● Moderate to severe interface hepatitis is a
universal agreement on what precisely fundamental component and histology is
constitutes an OS has generated consider vital in evaluating patients with overlap
able confusion. presentation. Use of the International
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
266 Section V Controversies in Autoimmune Liver Diseases
whether they represent a variant form of diagnostic test (the possible exception being
either disease (PBC, PSC or AIH). The latter PBC). Their diagnosis is based on the
seems to be the most appropriate since a pre presence and relative absence of various
dominant phenotype can be identified in clinical, biochemical, serologic, and histo
most cases. For example, in PBC/AIH overlap, logic markers, with some being less
it has been proposed that overlap represents a categorical and objective than others [1]. As
“hepatitic” form of PBC in genetically suscep a result, there is intrinsic scope for individ
tible individuals (HLA‐B8, ‐DR3 or ‐DR4 uals to present with overlapping features of
positive) [2]. This would fit with the hypo more than one of these conditions although,
thesis that immune‐mediated disease can in most cases, it is possible to define one pri
develop (“secondary” AIH) in any susceptible mary disorder (“dominant” disease). In a
host if, for some reason, the local milieu landmark review, Woodward and Neuberger
becomes proinflammatory. In this regard, the [5] emphasized that “true overlaps” should
name “overlap,” which strongly suggests the be differentiated from simple “crossover” or
presence of two distinct diseases, could “outlier” syndromes (one clear diagnosis
be a misnomer. As a result, according while having one feature associated with
to the European Association for the Study of another). Overlapping presentations include
the Liver (EASL) AIH and PBC guidelines, the following.
the preferred terminology to describe these
●● Biochemical overlap: aspartate aminotrans
conditions is now variant forms, primarily
ferase (AST) or alanine aminotransferase
variant forms of the cholestatic autoimmune
(ALT) more than five times upper limit of
liver diseases PBC or PSC with autoimmune
normal (ULN) in patients with PSC or PBC;
features [3,4]. In contrast, recent British and
or alkaline phosphatase (ALP) more than
US PBC guidelines still use the term “overlap”.
three times ULN in patients with AIH.
Because of the absence of well‐validated
●● Serologic overlap: positive anti‐smooth
diagnostic criteria and publication bias, the
muscle antibody (ASMA) in anti‐mitoch
prevalence of overlap syndromes is difficult
ondrial antibody (AMA)‐positive PBC; or
to ascertain and diagnosis remains a
positive AMA in AIH.
challenge. It should be kept in mind that
●● Histologic overlap: interface hepatitis on
overlap syndrome should not be overdiag
liver biopsy with biliary lesions indicative of
nosed in order not to expose PBC or PSC
PBC or PSC; cholangiographic abnormal
patients unnecessarily to the risk of steroid
ities associated with clinical features of AIH.
side effects. On the other hand, the tragic
●● Varying combinations of the above.
consequences of a missed opportunity of
instituting immunosuppressive therapy in However, these overlap features have var
overlap patients have occasionally been ious significance, the weaker being probably
reported. The low prevalence of overlap immunoserology. Indeed, autoantibody pro
syndromes has made it impracticable to
file should never be used in isolation but
perform randomized controlled trials. As a rather interpreted in conjunction with
consequence, treatment of overlap syndrome biochemical, radiologic, and histologic fea
is largely empiric. tures. For example, patients with histologic
AIH and AMA positivity generally behave
like typical AIH and the same holds true for
General Considerations AMA‐negative but ANA‐ and/or ASMA‐
positive PBC (sometimes described as auto
The main features of the three classical phe immune cholangitis) when compared with
notypes of autoimmune liver disease are typical PBC. Laboratory features lack sensi
indicated in Table 15.1. A key point is that no tivity considering that cholestasis in itself can
autoimmune liver disease has an absolute cause raised ALT levels in the absence of
268 Section V Controversies in Autoimmune Liver Diseases
Gender Female > male (4 : 1) Female > male (9 : 1) Male > female (2 : 1)
Coexisting IBD 3–10% (PSC should Not characteristic Up to 80%a
be excluded)
ANA 70–80% 30–50% (some 30–70%
specific)
ASMA 70–80% May be present: 0–80%
<10%
AMA 5–10% 95% Coincidental
p‐ANCA Up to 90% 0–5% 25–95%
Immunoglobulins IgG elevated IgM elevated in IgG elevated (two‐thirds)
most and IgM (45%) elevated
Cholangiography Usually normal Normal Multifocal stricturing
(not in small‐duct PSC)
Interface hepatitis Characteristic Variably present Variably present
Biliary changes 10% Inflammatory duct Onion‐skin periductal
lesion fibrosis (<30%)
Response to Yes Mild Minimal
immunosuppression
a
Most characteristic features are indicated in bold.
AIH, autoimmune hepatitis; AMA, anti‐mitochondrial antibody; ANA, anti‐nuclear antibody; ASMA, anti‐smooth
muscle antibody; IBD, inflammatory bowel disease; pANCA, perinuclear anti‐neutrophil cytoplasmic antibody; PBC,
primary biliary cholangitis; PSC, primary sclerosing cholangitis.
inflammation and that cirrhosis can lead to reproducible ways to grade interface
high IgG levels in the absence of histologic hepatitis are lacking, although the non‐
hepatitis. In contrast, liver biopsy is the most specific hepatitis activity index (HAI score)
reliable way of diagnosing overlap. is frequently used [3]. A score of more than
4 out of 18 points is observed primarily in
patients with variant syndromes, mainly
Liver Biopsy
due to additional lobular inflammatory
Liver biopsy is considered a prerequisite for activities. Interface hepatitis is not disease
the diagnosis of AIH [3]. Interface hepatitis specific and patients with drug‐related, viral
(hepatitis at the portal–parenchymal inter or immune‐mediated disease may show
face) with dense plasma cell‐rich lympho similar features. Lymphocytic piecemeal
plasmacytic infiltrates is the typical hallmark necrosis (as opposed to biliary piecemeal
of AIH. However, there is no morphologic necrosis that is the consequence of major
feature that is pathognomonic of AIH and a cholestasis) can be seen in about 25% of
degree of interface hepatitis can be observed patients with PBC or PSC and, conversely,
across all liver autoimmune diseases. In some biliary injury is observed in 10–20% of
AIH, the majority of patients will have a patients with AIH [6]. It has been hypothe
significant degree of interface hepatitis, sized that these lesions may represent some
contrasting with the usual low intensity of collateral damage mediated via bystander
interface hepatitis in PBC or PSC mechanisms in the context of a marked
(Figure 15.2) [1]. Unfortunately, robust and inflammation targeted primarily against
Chapter 15 Making Sense of Overlap and Crossover Syndromes 269
Low
Mild Moderate Severe
Interface hepatitis
either hepatocytes or biliary epithelial cells together with associated groups of general
[1]. A clinicopathologic study of PBC with inflammatory risk loci). Information about
interface hepatitis compared to AIH has genetic susceptibility in overlap syndromes is
suggested that, despite similar immunophe limited. The presence of the characteristic
notypes of infiltrating cells, the precise haplotypes of AIH (HLA‐DR3/DR4 and
mechanisms of hepatocellular injuries may HLA‐B8) has been reported in AIH/PBC
not be identical. Nevertheless, interface overlap patients and AIH patients (17/20 vs.
hepatitis is a fundamental component and 18/20, respectively), but these haplotypes
histology is vital in evaluating patients with were seen significantly less often in pure PBC
overlap presentation. It is generally assumed patients (4/20, p <0.05) [2].
that the degree of interface hepatitis can be In clinical practice, overlap syndromes
considered a measure of AIH‐like disease should be considered when a patient with
activity irrespective of the coexistence of autoimmune liver disease deviates from the
underlying cholestatic liver disease [6]. normal clinical phenotype and expected
Finally, interpretation of good‐quality liver response to therapy but it is important not to
biopsy is key and a specialist review of liver overdiagnose nor to frequently use non‐stan
biopsies has a major added value [7]. dard therapy with potential side effects. In
this regard, overlap syndromes should be
diagnosed conservatively and be differenti
AIH Scores ated from simple “crossover” or “outlier” syn
The International Autoimmune Hepatitis dromes that do not require treatment
Group (IAIHG) criteria (either revised or tailoring [5] (see following sections).
simplified) are an attractive tool for making
the diagnosis of overlap syndromes in patients
with an existing diagnosis of PBC or PSC. PBC/AIH Overlap Syndrome
However, although widely used in this setting,
the performance of these criteria appears low With no codified diagnostic approach,
and their use is not recommended for the reported prevalence figures are variable
diagnosis of overlap syndromes [6]. (2.1–19% in PBC patients) but it is generally
assumed that PBC/AIH overlap prevalence is
approximately 8–10% in adult patients with
Genetics
PBC or AIH [6]. The Paris criteria (Box 15.1)
The genetic architecture of PBC and PSC is are currently the most commonly used tool
similar to most autoimmune diseases, for diagnosing PBC/AIH overlap (see also
including AIH (i.e. a strong HLA association Figure 15.3 and Table 15.2).
270 Section V Controversies in Autoimmune Liver Diseases
(92%) and specificity (97%), keeping in mind UDCA, progression of fibrosis, and liver‐
that it is difficult to assess diagnostic related mortality. In a large clinical PBC trial,
performance in the absence of a gold standard. the patients who were identified as having
The EASL guidelines endorsed these diag potential AIH overlap retrospectively had a
nostic criteria but specified that histologic evi clinical outcome similar to “pure” PBC patients
dence of moderate to severe lymphocytic but unfortunately histologic fibrosis course
piecemeal necrosis (interface hepatitis) was was not assessed and no firm conclusions can
mandatory and stated that overlap syndrome be drawn from this study. Although UDCA
should always be considered once PBC has monotherapy may induce biochemical
been diagnosed and in case of poor response response in some patients with PBC/AIH
to ursodeoxycholic acid (UDCA) because of overlap, most patients may require a
potential therapeutic implications [9]. combination of UDCA and immunosuppres
Nevertheless, there are still several areas of sive therapy to obtain a complete biochemical
uncertainty, including the cutoffs for IgG/ response. In our experience of overlap patients
gamma‐globulin and transaminase levels for with histologic follow‐up, the overall fibrosis
indicating liver biopsy and the grade of progression in non‐cirrhotic patients occurred
hepatitis activity for indicating immunosup significantly more frequently under UDCA
pression [4]. Indeed, the recent EASL guide monotherapy (13–15 mg/kg daily) than under
lines on AIH recommend treatment for combined therapy with predniso(lo)ne
patients with AIH at lower cutoffs for trans (0.5 mg/kg daily as starting dose then tapered)
aminase or IgG levels and a histologic HAI and azathioprine (50–100 mg/day) [10].
score as low as 4 [3]. Despite the lack of controlled trials, the 2009
Autoantibodies against soluble liver EASL guidelines, based on the results of small
antigen (SLA/LP) and double‐stranded DNA series, have recommended adding steroids
have been associated with presence of AIH in (eventually budesonide at an initial dose of
patients with PBC and these autoantibodies 9 mg/day) either at the time of diagnosis of
should be tested in the work‐up of PBC overlap syndrome or in case of inadequate
patients with suspected overlap. Additionally, biochemical response after 3 months of
overlap of AMA‐negative PBC with AIH has UDCA [9]. The more recent results of a large
also been reported [8]. retrospective multicenter study (88 patients
The simultaneous presence of features of defined according to Paris criteria) have
both diseases is the usual presentation. In underlined the predictive role of the degree of
most cases, the “dominant” disease pheno interface hepatitis: as first‐line therapy,
type is PBC [6]. Less commonly, the onset of 30 patients received UDCA alone and 58
AIH and PBC is temporally dissociated, usu patients a combination of UDCA and immu
ally with PBC presenting first with a variable nosuppression (prednisone with or without
interval of 6 months to 13 years before the azathioprine). In patients with moderate inter
onset of AIH. The development of AIH could face hepatitis, UDCA alone and combination
not be predicted from baseline characteris therapy had similar efficacy (80%) in terms of
tics and initial response to UDCA. biochemical response, whereas in patients
Alternatively, PBC may (rarely) occur after with severe hepatitis, efficacy of UDCA alone
the diagnosis of AIH. was much lower (14 vs. 71%, respectively).
Patients with PBC/AIH overlap syndrome Second‐line immunosuppressive agents
seem to have a more severe disease compared (ciclosporin, tacrolimus and mycophenolate)
to conventional PBC as illustrated by a higher led to biochemical remission in half of the
frequency of extensive fibrosis at presentation patients who were non‐responders to initial
(despite a younger age in some reports) and immunosuppression [11]. These findings
most series (but not all) support a worse strongly support the use of a combination of
outcome in terms of biochemical response to UDCA and immunosuppression as first‐line
272 Section V Controversies in Autoimmune Liver Diseases
therapy in overlap patients with severe inter features ranged from 7 to 14% [6]. On the
face hepatitis. The aim of the treatment is other hand, cholangiographic abnormalities
complete remission defined by normalization suggesting PSC are found in AIH patients at
of transaminases and IgG levels and minimal various prevalence depending on the age of
inflammation on histology. The optimal dura the patients evaluated: 2–10% in adults (41%
tion of immunosuppression in overlap syn if ulcerative colitis is present) and up to 50%
dromes is unclear. Interestingly, it has been in children [12]. The diagnosis of large‐duct
suggested that, in responders, doses of immu PSC should always be established on typical
nosuppressants in the long term could be cholangiographic findings, bearing in mind
lower and rate of successful withdrawal higher that an intrahepatic biliary tree which simu
than in classical AIH [10,11]. lates a sclerosing pattern can be observed in
Recently, obeticholic acid (OCA) has been any liver disease with extensive fibrosis.
approved as a second‐line therapy for PBC Some cases of small‐duct PSC (normal chol
patients with an inadequate response to angiogram)/AIH overlap syndrome have also
UDCA monotherapy. Impressive results of been reported but it can be argued that
fibrates have also been reported in these approximately 10% of patients with typical
patients. It is important to differentiate AIH, with or without ulcerative colitis, may
patients with “classical” PBC and non‐ have histologic features of bile duct injury. In
response to UDCA from those with overlap children, the hepatitic feature can be very
syndrome who are also non‐responsive to dominant and up to 50% of pediatric AIH
UDCA. Whether the pleiotropic effects of (clinical and/or evidence of liver disease
fibrates or farnesoid X receptor agonists like associated with circulating autoantibodies)
OCA have sufficient immunosuppressive have cholangiographic abnormalities sugges
capacities and could be beneficial for overlap tive of PSC including some (25%) without
syndromes is currently unknown. any histologic features of bile duct injury or
In UDCA‐treated PBC patients devel biochemical cholestasis [12]. Inflammatory
oping AIH (“sequential” overlap syndrome), bowel disease (IBD) was present in 44% of
use of immunosuppressive treatment is these children compared to 20% of those
mandatory [9]. with AIH alone. Thus, the term “autoim
mune sclerosing cholangitis” (AISC) was
introduced by Mieli‐Vergani’s group [12].
PSC/AIH Overlap Syndrome Evolution from AIH to AISC has been docu
mented, supporting the view they could be
PSC/AIH overlap syndrome has been part of the same pathogenic process. It has
described in both children and adults and is been proposed that at least some adult PSC
assumed to exist in a considerable number of cases may represent an advanced, at times
mainly young patients with autoimmune “burnt‐out”, stage of AISC but whether
liver disease. The diagnosis of overlap syn childhood AISC and adult PSC belong to the
drome is made in a patient with overt cholan same disease spectrum remains to be
giographic or histologic features of PSC, established. These findings also suggest the
alongside robust histologic features of AIH need for an investigation of the biliary tree at
concurrently or historically [1,6] (Table 15.2). least with magnetic resonance cholangio
Unfortunately, as opposed to PBC/AIH pancreatography (MRCP) in all children with
overlap syndrome, no precise and strict diag a diagnosis of AIH. At present, this variant
nostic criteria have been proposed. As a seems unique to children, as a prospective
result, reported prevalence figures vary study in adults with AIH was negative and
greatly. When the revised IAIHG criteria thus, in the absence of cholestatic indices,
were applied to large series of PSC patients, MRCP screening does not seem justified in
the prevalence of PSC/AIH overlapping adult‐onset AIH. However, particularly in
Chapter 15 Making Sense of Overlap and Crossover Syndromes 273
young adults with AIH and cholestatic fea is lower than in AIH (100%) [12]. In the series
tures or IBD, and in AIH patients with with the most homogeneous regimen
remaining cholestasis despite adequate (UDCA, prednisolone and azathioprine)
immunosuppression, MRCP for the detec including seven young adults with a mean
tion of possible underlying or coexistent PSC follow‐up of 8 years, the Mayo Risk Score did
is recommended [13]. not increase and transplant‐free survival was
In adults, AIH and PSC may be sequential much better (100%) than that of 34 classical
in their occurrence, typically with AIH pre PSC patients (43%) with the same follow‐up
senting first, as illustrated by a case series of and treated with UDCA. However, in the
AIH patients becoming cholestatic and resis long term (>10 years), progression toward
tant to immunosuppressive therapy. AIH is cirrhosis seems to occur in the majority of
more rarely diagnosed in patients with an patients.
original diagnosis of PSC. Taken together, all these data support the
Similar to PBC/AIH overlap, there are no use of UDCA in combination with an immu
double‐blind, randomized controlled trials in nosuppressive regimen in most patients with
PSC/AIH overlap syndromes. It should be PSC/AIH overlap syndrome despite the lack
kept in mind that although immunosuppres of adequate studies. However, the key point
sants benefit the hepatitic component of is that, even more than in PBC/AIH overlap
AIH, no survival benefit has been demon syndrome, treatments in PSC/AIH overlap
strated with UDCA in PSC. In addition, should be individualized based on bioch
unlike in PBC and AIH, biochemical improve emical, serologic, cholangiographic, and his
ment in PSC does not necessarily translate tologic findings. In patients with severe
into better clinical outcome. Various results interface hepatitis, use of immunosuppres
of therapy (usually prednisolone and azathio sants is mandatory. In other cases (moderate
prine with or without UDCA) have been interface hepatitis), my policy is, at present,
reported in patients with PSC/AIH overlap similar to that in PBC/AIH overlap and to
[6]. It is difficult to draw any firm conclusions start with UDCA monotherapy and add
because of the small number of patients, the immunosuppressants only in case of inade
usually retrospective nature of the studies, quate biochemical response after 3 months
and the heterogeneity of the regimens. The of UDCA.
combination of UDCA and immunosuppres
sive therapy may improve liver biochemistry
and this approach has been advocated by Liver Transplantation
EASL guidelines [9], whereas the American
Association for the Study of Liver Diseases Studies documenting outcomes of patients
(AASLD) guidelines recommend the use of who underwent liver transplantation for
corticosteroids and other immunosuppres overlap syndromes are limited. In the largest
sive agents and the IAIHG position is to con series, patients with overlap (n = 12) had a
sider immunosuppressive treatment with or higher probability of recurrent autoimmune
without UDCA [6]. Unsurprisingly, patients liver disease when compared to those with
with PSC/AIH overlap have a poorer out single autoimmune liver disease (at 5 years,
come when compared to those with (treated) 53% vs. 17%, respectively) together with a
AIH alone. However, their survival is appar shorter median time to recurrence but
ently similar or even better than in classical without lower patient survival [14].
PSC. In the pediatric AISC form treated with Interestingly, two patients had recurrence as
immunosuppressants, liver biopsies may overlap syndrome (one PBC/AIH and one
show improvement in inflammation but PSC/AIH) and experienced a lower graft
cholangiographic appearances may progress survival than those having recurrence as a
and transplant‐free survival at 10 years (65%) single autoimmune liver disease.
274 Section V Controversies in Autoimmune Liver Diseases
References
1 Trivedi, P.J. and Hirschfield, G.M. (2012). 8 Chazouillères, O., Wendum, D., Serfaty, L.
Review article: overlap syndromes and et al. (1998). Primary biliary cirrhosis–
autoimmune liver disease. Aliment. autoimmune hepatitis overlap syndrome:
Pharmacol. Ther. 36: 517–533. clinical features and response to therapy.
2 Lohse, A.W., zum Buschenfelde, K.H., Franz, Hepatology 28: 296–301.
B. et al. (1999). Characterization of the 9 European Association for the Study of the
overlap syndrome of primary biliary Liver (2009). EASL clinical practice
cirrhosis (PBC) and autoimmune hepatitis: guidelines: management of cholestatic liver
evidence for it being a hepatitic form of PBC diseases. J. Hepatol. 51: 237–267.
in genetically susceptible individuals. 10 Chazouillères, O., Wendum, D., Serfaty, L.
Hepatology 29: 1078–1084. et al. (2006). Long term outcome and
3 European Association for the Study of the response to therapy of primary biliary
Liver (2015). EASL clinical practice cirrhosis–autoimmune hepatitis overlap
guidelines: autoimmune hepatitis. J. Hepatol. syndrome. J. Hepatol. 44: 400–406.
63: 971–1004. 11 Ozaslan, E., Efe, C., Heurgue‐Berlot, A.
4 European Association for the Study of the et al. (2014). Factors associated with
liver (2017). EASL clinical practice response to therapy and outcome of
guidelines: the diagnosis and management patients with primary biliary cirrhosis with
of patients with primary biliary cholangitis. features of autoimmune hepatitis. Clin.
J. Hepatol. 67: 145–172. Gastroenterol. Hepatol. 12: 863–869.
5 Woodward, J. and Neuberger, J. (2001). 12 Gregorio, G.V., Portmann, B., Karani, J.
Autoimmune overlap syndromes. et al. (2001). Autoimmune hepatitis/
Hepatology 33: 994–1002. sclerosing cholangitis overlap syndrome in
6 Boberg, K.M., Chapman, R.W., Hirschfield, childhood: a 16‐year prospective study.
G.M. et al. (2011). Overlap syndromes: the Hepatology 33: 544–553.
International Autoimmune Hepatitis 13 Lindor, K.D., Kowdley, K.V., and Harrison,
Group (IAIHG) position statement on a M.E. (2015). ACG clinical guideline:
controversial issue. J. Hepatol. 54: primary sclerosing cholangitis. Am. J.
374–385. Gastroenterol. 110: 646–659. quiz 660.
7 Paterson, A.L., Allison, M.E., Brais, R., and 14 Bhanji, R.A., Mason, A.L., Girgis, S., and
Davies, S.E. (2016). Any value in a Montano‐Loza, A.J. (2013). Liver
specialist review of liver biopsies? transplantation for overlap syndromes of
Conclusions of a 4‐year review. autoimmune liver diseases. Liver Int. 33:
Histopathology 69: 315–321. 210–219.
277
16
Abstract
This chapter presents the knowledge on the role of extrahepatic autoimmunity in autoimmune
liver disease, even though systematic studies on large cohorts of patients are often lacking and
some experience is based only on case series reports. Awareness and early investigation of extrahe-
patic autoimmune diseases (EHAIDs) in patients with autoimmune liver disease is important.
Targeted testing should be considered according to the liver disease; it may be difficult to differen-
tiate between those EHAIDs that can impact the liver and drug-induced liver injury from primary
autoimmune liver disease. In primary sclerosing cholangitis (PSC), ulcerative colitis is a dominating
EHAID. The strong co-occurrence of PSC with inflammatory bowel disease (IBD), as well as
several other unique features of PSC-related IBD, led to the idea that it may be a distinct entity from
IBD, unrelated to PSC in terms of pathogenesis, phenotype and prognosis.
Keywords drug-induced liver injury; extrahepatic autoimmune diseases; inflammatory bowel
disease; primary biliary cholangitis; primary sclerosing cholangitis; ulcerative colitis
Key Points
●● Awareness and early investigation of ●● In autoimmune hepatitis, autoimmune thy-
extrahepatic autoimmune diseases (EHAIDs) roid disease is the most frequent EHAID.
in patients with autoimmune liver disease ●● In primary sclerosing cholangitis (PSC),
is important. ulcerative colitis (UC) is a dominating
●● Targeted testing should be considered EHAID. The clinical presentation of UC in
according to the liver disease; it may be association with PSC differs in many ways
difficult to differentiate between those from UC seen in patients without PSC.
EHAIDs that can impact the liver (such ●● In primary biliary cholangitis, autoimmune
as systemic lupus erythematosus) and thyroid disease, celiac disease, Sjögren syn-
drug‐induced liver injury from primary drome, and systemic sclerosis occur in a
autoimmune liver disease. significant proportion of patients.
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
278 Section V Controversies in Autoimmune Liver Diseases
Inflammatory
Autoimmune Sjögren Systemic Systemic lupus Rheumatoid Celiac bowel
thyroid disease syndrome sclerosis erythematosus arthritis disease disease
(PSC). Patients with both PBC and AIH with Autoimmune thyroid disease per se does
concomitant EHAIDs show significant not seem to significantly influence the natural
female predominance, whereas PSC‐related history of AILDs. Nevertheless, prior exacer-
IBD is diagnosed more commonly in males. bated hypothyroidism may cause a delay in
EHAIDs are more commonly found in AIH diagnosis of chronic liver diseases such as PBC
patients with first‐degree family history of and PSC as both conditions share common
EHAIDs. Concomitant autoimmunity may symptoms including fatigue, metabolic bone
predate the onset of AILDs or occur during disease, hyperlipidemia, and abnormalities in
the course and even years after the diagnosis cholestatic parameters, particularly alkaline
of AILDs. phosphatase (ALP) and bilirubin. On the other
hand, failure to recognize the onset of con-
comitant thyroid dysfunction in patients with
Autoimmune Thyroid Disease established AILD may lead to a misdiagnosis
of liver disease decompensation or lack of
The prevalence of thyroid dysfunction response to therapy. This can potentially lead
among patients with AILD is significantly to improper decisions about liver treatment
higher than in the general population, and or even timing for liver transplantation.
mainly manifests with hypothyroidism. The Therefore, the recognized association between
reported prevalence of abnormal thyroid AILDs and thyroid diseases should warrant
function has been most extensively studied the routine screening of patients with AILDs
in PBC and ranges between 13 and 22%. The for concomitant thyroid dysfunction to make
data among patients with PSC are scarce. A an accurate diagnosis according to their
study by Silveira et al. [5] showed 11% respective diagnostic criteria.
prevalence of thyroid dysfunction in this
group of patients; among them, 6% had
hypothyroidism and 5% hyperthyroidism.
Sjögren Syndrome
Autoimmunity plays a crucial role; other
causes of thyroid disease in AILDs are less SS is a progressive autoimmune exocrinopa-
common. As already mentioned, autoim- thy characterized by a lymphocytic infiltra-
mune thyroid diseases are the most frequent tion of the exocrine glands, leading to ocular
EHAID in patients with AIH, and the sec- dryness (xerophthalmia) and oral dryness
ond most frequent in patients with PBC. In (xerostomia). The disorder can be classified
the spectrum of thyroid autoimmunity, as primary and secondary SS. Primary SS
Hashimoto thyroiditis dominates, with represents a category on its own without
Graves disease being less frequent. The any other underlying condition, whereas
coincidence of autoimmune thyroid dis- secondary SS occurs as a concomitant
eases and AILD may be at least partially condition with another autoimmune disorder.
explained by a common genetic background. Sicca symptoms (i.e. eye and mouth dryness)
Our experience suggests the presence of and difficulty tasting and swallowing are
several single nucleotide polymorphisms reported by 47–73% of patients with PBC.
(located in genes PTPN22, IL2RA, MMEL1, Further investigation, such as Schirmer’s test
CTLA4, and RNASET2) that are associated and/or salivary gland biopsy, leads to the
with shared susceptibility to autoimmune diagnosis of secondary SS in 3.5–27% of them.
thyroid disease and PBC. The possible In fact, this is the most frequent extrahepatic
immunologic mechanism may lead to cross‐ condition seen in PBC [3,6]. The coincidence
reactivity of anti‐thyroid autoantibodies in with other AILDs is much less frequent, with
the presence of autoreactive T cells or sim- the prevalence in AIH being 1.4–2.3%, and
ilar epithelial antigens in both the liver and almost no cases observed in PSC. In
the thyroid. comparison, in the general population the
280 Section V Controversies in Autoimmune Liver Diseases
prevalence of SS ranges from 0.2 to 3%, sim- resulting in fibrosis that can involve the skin
ilar to other autoimmune diseases. The high and other organs but usually not the liver.
coincidence between SS and PBC may be Two major disease subcategories can be dis-
explained by the fact that both diseases share cerned, depending on the extent of skin
some pathogenic mechanisms, with apoptosis involvement: a limited cutaneous form
being likely the important element, and both (lcSSc; formerly CREST syndrome), associ-
are associated with chronic autoimmune epi- ated with limited skin thickening and little
theliitis, leading to destruction of both bile organ involvement; and diffuse cutaneous
ducts and certain exocrine glands [3]. The SSc (dcSSc), involving the skin and multiple
idea is further supported by other similarities internal organs. lcSSc dominates in patients
between both diseases such as female pre- with PBC, as up to 93% of patients with con-
dominance and age of onset in the fifth decade comitant SSc have this subtype. The autoim-
of life. Interestingly, when both diseases mune background of this association remains
coexist, the clinical course is usually mild, as to be elucidated. A positive anti‐centromere
symptoms of SS warrant therapy in only 11% antibody (ACA), a serologic hallmark of SSc,
of cases. Coexistent SS/AIH has been mainly can be observed in 9–30% of patients with
described in a cohort of Asiatic patients, PBC and in 22–25% of all patients with SSc,
among which AIH/PBC was common. The mostly in the limited cutaneous subtype.
clinical significance of the diseases’ Interestingly, patients with PBC and con-
association is not fully understood; however, comitant SSc show higher ACA positivity
anti‐Ro52 positivity has been considered a than in PBC or SSc alone. On the other hand,
negative prognostic factor in these patients. up to one‐fourth of patients with SSc are
On the other hand, among other organs, positive for anti‐mitochondrial antibody
the liver is a common non‐exocrine target in (AMA), although there is no cross‐reactivity
primary SS. Hepatomegaly is reported in between mitochondrial and centromere
2–20% of patients, whereas biochemical liver antigens. Genetic studies have detected asso-
abnormalities, cholestatic rather than hepato- ciations with only a few genes (STAT4 and
cellular, are reported in 49% of patients. These IRF5), which overlap between SSc and PBC
alterations are more common in patients with and which can explain the relative rarity of
positive antibodies to extractable nuclear SSc/PBC coexistence. Another gene,
antigens or evidence of systemic inflamma- PTPN22, that has been associated with SSc,
tion, as shown by an elevated erythrocyte but not with PBC, may infer susceptibility to
sedimentation rate. The abnormalities may the development of SSc in PBC patients.
be either persistent or intermittent but usu- Prognosis in PBC/SSc is not established.
ally have little clinical significance. In a Several studies have shown a significantly
minority of cases, further investigation led to better outcome, slower liver disease progres-
the diagnosis of AILDs or chronic viral sion, and lower rates of liver transplantation
hepatitis C. in PBC/SSc than in PBC alone. Conversely,
an increased mortality due to the coexistence
of SSc in patients with PBC has also been
Systemic Sclerosis reported, mainly due to non‐liver disease
complications. Another study has shown that
Systemic sclerosis (SSc) is an extrahepatic the prevalence of esophageal varices is signif-
manifestation of AILDs with a prevalence icantly higher in non‐cirrhotic patients with
ranging between 1.4 and 12.3% in PBC, and PBC/lcSSc compared with non‐cirrhotic
between 1.2 and 3.5% in AIH, but is seen only patients with PBC alone. In line with these
incidentally in patients with PSC [4]. It is a findings, ACA positivity was highlighted as
systemic CTD characterized by immune‐ a risk factor for an increased rate of
mediated extracellular matrix deposition portal hypertension. Summarizing these
Chapter 16 Role of Extrahepatic Autoimmunity in Autoimmune Liver Disease 281
AIH, autoimmune hepatitis; ANA, anti‐nuclear antibody; anti‐dsDNA, anti‐double‐stranded DNA; anti‐LKM,
anti‐liver kidney microsomal; anti‐SMA, anti‐smooth muscle antibody; SLE, systemic lupus erythematosus.
Source: adapted from Kaw et al. [9].
i ndicated that the prevalence of RA in PBC is disease (lower hemoglobin levels but higher
comparable to that observed in controls. The serum levels of IgG and ALP, and a higher
frequency of concomitant AIH and PSC ratio of RF positivity). However, this issue has
seems to be even lower. Larger studies are not been studied in detail.
needed to determine the real coexistence of
one disease with the other.
Although liver damage is not seen as a Celiac Disease
significant extra‐articular feature of RA,
18–50% of patients with RA have abnormal Celiac disease (CD) is the most common
liver biochemistry, associated with mostly autoimmune enteropathy. As well as clinical
mild and non‐specific changes in liver his- manifestations of a typical malabsorption
tology. In the vast majority of cases, the syndrome, CD may exhibit several extraintes-
underlying cause of these abnormalities is tinal symptoms [8]. The most common
something other than concomitant AILD, hepatic presentation, seen in 40% of patients,
most commonly associated with hepatotox- is isolated hypertransaminasemia with non‐
icity of the administered medications, mainly specific changes in liver histology (celiac
NSAIDs, methotrexate and sulfasalazine. hepatitis). A gluten‐free diet leads to the
DILI is most often manifested by asymptom- normalization of liver biochemistry and his-
atic elevation of aminotransferases. Severe tologic alterations in most cases. On the other
hepatic injury is rather rare; however, pro- hand CD often coexists with a spectrum of
gressive liver damage including liver cir- immunologically mediated liver diseases. The
rhosis has been associated with methotrexate prevalence of AIH in patients with CD is esti-
treatment. The risk of methotrexate‐induced mated at 1.6%, whereas typical CD autoanti-
liver injury is higher in patients with under- bodies are observed in 3.5% of patients with
lying liver disease, in particular those whose AIH in comparison to 0.35% in the general
treatment was prolonged over 2 years due to population. The relationship between CD and
an association with a high drug dosage. PBC is also well established. Epidemiologic
Genetic and genome‐wide association studies indicate that anti‐endomysium anti-
studies have found some common genetic bodies are positive in 10% of patients with
background of PBC and RA, although the PBC. Finally, in patients with PSC, subsequent
vast majority of implicated genes in both dis- studies also showed an increased risk of CD,
eases do not overlap. Overlapping genes with prevalence of 1.6% and, conversely, a
include HLA‐DQB1, STAT4, IRF5, MMEL1, fourfold to eightfold higher risk of PSC deve
CTLA4, and possibly CXCR5. Individuals lopment in patients with CD in comparison
carrying risk alleles may be more susceptible to the general population.
to developing concomitant RA and PBC. On The genetic and immunologic mechanisms
the other hand, patients with a genetic profile behind the AILD/CD association have been
contributing to RA or PBC, in whom genes extensively studied. The genetic link between
do not overlap, probably will not develop CD and AIH has been described by common
concomitant autoimmune disease. This is combinations of genes coding for class II
suggested by epidemiologic data that show a human leukocyte antigen (HLA) molecules
small proportion of patients with concomi- on chromosome 6. The main genetic marker
tant RA and AILD. of CD, HLA‐DQ2.5, has a strong linkage dis-
The data analyzing the influence of con- equilibrium with HLA‐DR4, HLA‐DR52, and
comitant RA in patients with AILD are HLA‐DR3. The last one is also a major HLA
scarce. Some observations suggest that the risk factor for AILDs. Moreover, several other
coexistence of RA may be a negative prog- gene polymorphisms outside the HLA com-
nostic factor for PBC, as PBC/RA patients plex were found in both CD and AILD,
show laboratory indicators of more severe suggesting shared dysregulated immune
284 Section V Controversies in Autoimmune Liver Diseases
responses contributing to the development of the upper limit of normal), laboratory signs
the two diseases. Moreover, the breakdown of of cholestasis, or physical and ultrasono-
gut–liver axis equilibrium has been suggested graphic signs suggesting a liver disorder. In
as a trigger in the immunopathologic pathway such cases, a thorough hepatologic examina-
toward immune‐mediated liver injury in CD. tion is necessary and liver biopsy should be
It is hypothesized that CD‐associated considered in selected cases. An early diag-
increased intestinal permeability allows the nosis with the appropriate treatment can
immunologically active molecules generated improve the outcome of these patients.
from the cross‐linking between tissue trans-
glutaminase and food or bacterial antigens to
reach the liver through the portal circulation. Inflammatory Bowel Disease
Molecular mimicry between bacterial and
hepatic antigens may further induce immune‐ The close association between PSC and IBD
mediated hepatic damage. Aberrant intestinal is well established [10]. The reported preva-
T‐lymphocyte homing to the liver may lence of IBD in PSC patients is usually esti-
further contribute in this process. mated at 70%, but the frequency ranges from
The clinical impact of CD on the progres- 50 to 99% across different studies, depending
sion of AILD remains uncertain. Studies with on the diagnostic criteria for IBD and geo-
a small number of pediatric patients with graphic differences. European and American
AIH/CD suggest that a gluten‐free diet facil- studies report a higher prevalence of IBD in
itates the achievement of treatment‐free PSC patients in comparison to populations
sustained remission of AIH and, conversely, a from Asia. Conversely, in patients with an
gluten challenge may promote AIH relapses established diagnosis of IBD, PSC is found to
in these children. be much less frequent, occurring in 2–8% of
The association between CD and liver dis- patients with UC and 3% of patients with
orders, both celiac hepatitis and concomi- Crohn’s disease. Intriguingly, IBD is not
tant AILD, deserves specific clinical closely associated with other AILDs, as it
attention. Clinicians need to consider CD in occurs in only about 2–8% of patients with
the differential of abnormal liver biochem- both AIH and PBC.
istry. Moreover, routine liver biochemistry is The strong co‐occurrence of PSC with
recommended in all patients with newly IBD, as well as several other unique features
diagnosed CD. Asymptomatic mild hyper- of PSC‐related IBD, led to the idea that it may
transaminasemia, together with a normal be a distinct entity from IBD, unrelated to
physical and ultrasound examination, PSC in terms of pathogenesis, phenotype and
strongly suggests liver manifestation of CD prognosis. In fact, patients with PSC‐related
rather than a coexistent AILD. Because of UC often demonstrate a right‐to‐left gra-
the non‐specific nature of the findings and dient of colonic inflammation, rectal sparing,
invasiveness of the procedure, in such cases and backwash ileitis. This pattern of inflam-
there is usually no need for liver biopsy. The mation is significantly different from individ-
normalization of liver biochemistry after uals with UC alone. Moreover PSC/IBD
6–12 months of gluten elimination confirms patients typically have an increased inci-
the diagnosis. An alternative etiology for dence of extensive colitis and pancolitis, irre-
liver injury should be explored in those spective of IBD subtype (i.e. UC or Crohn’s
patients with persistent hypertransaminase- disease), although reported symptoms are
mia despite gluten exclusion. Moreover, a usually mild (Figure 16.1). Even in patients
clinical profile not consistent with celiac with Crohn’s disease, most common is
hepatitis also requires an extended examina- colonic involvement rather than ileocolic
tion. These particularly include severe and isolated ileum disease localization.
hypertransaminasemia (more than five times Finally, the inflammatory activity in the
Chapter 16 Role of Extrahepatic Autoimmunity in Autoimmune Liver Disease 285
(a) (b)
Backwash Backwash
ileitis ileitis 7%
Right-to-left 51%
gradient Left-to-right
gradient
Rectal
Proctitis
sparing
95%
30%
Figure 16.1 Phenotypic features of (a) PSC‐related ulcerative colitis compared with (b) ulcerative colitis
without concomitant PSC.
intestine is usually low, which corresponds proctocolectomy and, conversely, IBD can
with milder symptoms, lower frequency occur many years after the initial diagnosis of
of strictures and penetrating disease. PSC or even after liver transplantation.
Intriguingly, PSC patients with concomitant In terms of prognosis, the association bet-
CD have a less progressive liver disease, ween PSC and IBD appears to increase the
favorable transplant‐free survival, and lower risk of developing colon cancer compared to
rate of hepatobiliary malignancy, and hence a the patients with IBD only. Colon cancer
better outcome than patients with PSC/UC associated with PSC and IBD shows distinct
or patients with PSC only. features, including right‐sided localization
Because the coexistence of PSC and IBD (up to 76%) and worse prognosis due to more
influences the IBD phenotype, some distinct advanced stage and younger age at diagnosis.
features of PSC may also be noted. Compared Moreover, the prolonged duration of IBD
to PSC patients alone, patients with PSC/IBD may also be associated with an increased risk
demonstrate more common combined intra- of cholangiocarcinoma in PSC patients,
hepatic and extrahepatic biliary involvement. irrespective of colectomy. It is unclear
Moreover, the diagnosis of biliary disease whether the observed elevated cancer risk
tends to be made more often in males and at may be the effect of the severity of intestinal
a younger age, when compared with PSC‐ inflammation.
only patients. The demographic data with The high coincidence of PSC and IBD sug-
respect to clinical onset of IBD related to gests a common underlying pathogenesis,
PSC are inconsistent. Some studies report most likely influenced by genetic predisposi-
that the mean age for diagnosis of IBD is tion, immune‐mediated mechanisms, and
higher in PSC/IBD compared with IBD alone. possibly altered gut microbiota. In fact, both
Conversely, a recent study indicates that diseases share common antibodies. In
PSC/UC patients have the clinical onset of particular, perinuclear anti‐neutrophil cyto-
UC at a significantly earlier age compared plasmic antibody (pANCA) has been
with patients with UC only. Usually, IBD pre- reported in 65–95% of PSC patients, mainly
dates the diagnosis of PSC; however, PSC in the context of concomitant UC. Recently,
may also be recognized many years after IgG to the serine protease proteinase‐3
286 Section V Controversies in Autoimmune Liver Diseases
References
1 Guo, L., Zhou, L., Zhang, N. et al. (2017). 6 Gershwin, M.E., Selmi, C., Worman, H.J.
Extrahepatic autoimmune diseases in et al. (2005). Risk factors and comorbidities
patients with autoimmune liver in primary biliary cirrhosis: a controlled
diseases: a phenomenon neglected by interview‐based study of 1032 patients.
gastroenterologists. Gastroenterol. Res. Hepatology 42: 1194–1202.
Pract. 2017: 2376231. 7 Matsumoto, T., Yoshimine, T., Shimouchi,
2 Teufel, A., Weinmann, A., Kahaly, G.J. K. et al. (1992). The liver in systemic lupus
et al. (2010). Concurrent autoimmune erythematosus: pathologic analysis of 52
diseases in patients with autoimmune cases and review of Japanese autopsy
hepatitis. J. Clin. Gastroenterol. 44: registry data. Hum. Pathol. 23: 1151–1158.
208–213. 8 Marciano, F., Savoia, M., and Vajro, P.
3 Selmi, C., Meroni, P.L., and Gershwin, (2016). Celiac disease‐related hepatic
M.E. (2012). Primary biliary cirrhosis and injury: insights into associated conditions
Sjogren’s syndrome: autoimmune and underlying pathomechanisms. Dig.
epitheliitis. J. Autoimmun. 39: 34–42. Liver Dis. 48: 112–119.
4 Rigamonti, C., Shand, L.M., Feudjo, M. 9 Kaw, R., Gota, C., Bennet, A. et al. (2006).
et al. (2006). Clinical features and Lupus‐related hepatitis: complication of
prognosis of primary biliary cirrhosis lupus or autoimmune association? Case
associated with systemic sclerosis. Gut 55: report and review of the literature. Dig. Dis.
388–394. Sci. 51 (4): 813–818.
5 Silveira, M.G., Mendes, F.D., Diehl, 10 Weismüller, T.J., Trivedi, P.J., Bergquist, A.
N.N. et al. (2009). Thyroid dysfunction et al. (2017). Patient age, sex, and
in primary biliary cirrhosis, primary inflammatory bowel disease phenotype
sclerosing cholangitis and non‐ associated with course of primary
alcoholic fatty liver disease. Liver Int. 29: sclerosing cholangitis. Gastroenterology
1094–1100. 152: 1975–1984.e8.
289
17
Abstract
The chronic and incurable nature of autoimmune liver disease (AILD) needs to be recognized in
approaching the management of patients. This chapter explores the inherent complexity in terms
of patient impact and how this influences approaches to management, with a particular focus on
symptoms and their impact on quality of life. Symptoms can be stage-specific or non-stage-specific.
The commonest symptom in all AILD types is fatigue. Pruritus in cholestatic AILD is the symptom
that is easiest to treat, with sequential intervention approaches that are well described in treatment
guidelines. In AILD, the complex of symptoms related to fatigue has the most challenging effects
on quality of life. The critical issue is that the characteristic symptoms of AILD can each lead to
behavior patterns that can isolate people. There is a real need for the AILD community to explore
new approaches to structured care delivery utilizing digital care pathway approaches.
Keywords autoimmune liver disease; intervention approaches; non-stage-specific symptoms;
patients management; stage-specific symptoms
Key Points
●● Symptoms are an important aspect of important to prevent the former, but may
autoimmune liver disease (AILD) and in have little or no impact on the latter.
some patients are the major problem. They ●● The commonest symptom in all AILD
should always be considered when devel- types is fatigue. This is typically very dif-
oping management plans for individual ficult to treat. There is no single effec-
patients. tive therapy, although a structured
●● Symptoms can be stage‐specific (i.e. asso- supportive approach can improve
ciated with the development of cirrhosis) quality of life.
or non‐stage‐specific (i.e. present at any ●● Pruritus in cholestatic AILD is the
stage of the disease process). Effective symptom that is easiest to treat, with
treatment of the underlying disease is sequential intervention approaches that
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
290 Section V Controversies in Autoimmune Liver Diseases
seen in this disease in practice than might be and language used is understandable to
anticipated given the apparent effectiveness patients and captures the specific impacts of
of the drugs [2]. the relevant disease state. The optimal
approach is a patient‐derived measure vali-
dated within the target population. In the
ymptoms, Quality of Life,
S area of AILD there is a high‐quality patient‐
and Health Utility: Key Concepts derived and validated measure, the PBC‐40
[4], which is widely used in research studies
The study of the impact of disease phenotype and clinical trials (a short‐form version for
on the patient’s quality‐of‐life experiences use in the clinic setting for symptom screen-
invokes the concepts of symptoms them- ing is under development). An equivalent
selves, quality of life assessment, and health measure is under development in PSC. As
utility. In simple terms the former can be yet there has been no work on measures in
regarded as the individual disease phe- AIH that we are aware of. All subjective mea-
nomena and the latter two measures of the sures can be prone to ceiling effects, where a
downstream impact of symptoms on the lives patient’s concept as to the worst symptom
of patients [3]. There is no automatic con- severity they can imagine is dictated by their
nection between symptom presence and experience to date; in simple terms, if a
downstream quality‐of‐life impairment. An patient scores a symptom as 10/10 because
example is pruritus in PBC where frequency they cannot envisage it getting worse, then if
of pruritus report, for example in the setting it does actually get worse it cannot be quanti-
of adverse event reporting, exceeds the take‐ fied using the measure.
up rate for pruritus therapy. Although this Objective symptom measures are highly
reflects, to some degree, poor clinical prac- attractive to the research community and to
tice, a significant element is patients regulators because of the degree to which
describing a symptom which is not severe they seemingly directly ascribe a value to a
enough in their eyes to warrant treatment as disease state. Caution must be applied, how-
it does not impact on their quality of life. It is ever, as the physical measure can be affected
critical to understand this concept to avoid by patient response to the symptom (e.g.
an over‐focus on low‐impact symptoms, the patients with pruritus who manage to control
treatment of which will not actually lead to their scratching, patients with fatigue who
patient improvement. make themselves keep going and thus show
reasonable physical activity levels).
A particular challenge in symptom
Symptoms
assessment relates to those states which can
All patients and clinicians are familiar with be part of normal life (e.g. tiredness or
the concept of symptoms (defined as fatigue) and which may be exaggerated by a
“physical or mental features which are disease state. In this setting, patients are
regarded as indicating a condition of disease, often left making a value judgment as to what
particularly such features that are apparent component of their experience results from
to the patient”). Symptom severity can be the disease and what is part of normal life. In
assessed using subjective and objective contrast, those symptom states which would
approaches. Subjective assessments rely on not be part of normal experience, exempli-
the patient’s stated view about symptom fied by pruritus, can present less of a
severity using simple scales (visual analog challenge. An important corollary of the
scales or numerical rating scales giving a challenge around those symptom states
score out of 10 to symptom severity) and which are and are not part of a continuum or
more complex assessment questionnaires. In normality is that the former can sometimes
all instances it is essential that the approach be seen as less “real” or disease‐related than
292 Section V Controversies in Autoimmune Liver Diseases
the latter. This issue is exemplified by the achieved in the decision‐making process
different perceptions of “reality” and the regarding drug reimbursement. Both PBC
importance of fatigue and pruritus among and AIH have recently undergone health
the PBC clinical community. utility assessments in the UK (PSC remains
to be studied using this approach). Both AIH
and PBC can be associated with normal
Health‐related Quality of Life
health utility (>0.95, 1.0 being deemed
The concept of quality‐of‐life measurement unachievable), suggesting that this should be
can be even more challenging than symptom a reasonable goal for management. Both are,
assessment, including, as it does, elements that typically, associated with significantly
go well beyond the impact of any disease or impaired utility. In the case of PBC the dom-
clinical condition. For this reason, the more inant effect is through non‐stage‐specific
focused concept of health‐related quality of life symptoms (in particular fatigue and
(HRQoL) is used, a standard definition for cognitive symptoms), although where pre-
which is “the measurable impact of a person’s sent end‐stage disease features such as
perception of his or her health and the effect ascites are associated with a substantial
that produces on satisfaction with life and well‐ additive effect. UDCA use, interestingly, is
being.” The more complex, abstract nature of associated with significantly higher levels of
this means that simple numerical scales are not health utility irrespective of UDCA response
helpful (their use would rely on all users status and of any impact on individual symp-
ascribing the same meaning to the term “quality toms. In AIH, although cirrhosis and its
of life”). Well‐validated, generic quality‐of‐life complications again have a marked effect,
measures, such as the SF‐36 and EQ5D, have the largest driver for poor utility is the use of
been used in the study of AILD, and have sug- corticosteroid and calcineurin inhibitor
gested levels of HRQoL impairment similar to therapy (independently of disease activity).
those seen in other chronic disease settings. As
well as being cumbersome to use in practice,
the scoring is not intuitive making judgment as ymptoms and Their Management
S
to the nature of quality‐of‐life impairment dif- in AILD
ficult to reach without detailed analysis.
However, they are the bedrock of quality‐of‐life The symptoms seen in AILD can be divided
assessment as required by regulatory bodies. into stage associated and non‐stage
associated.
Health Utility
Impact of Disease‐modifying Therapy
Health utility represents an analysis of
on Symptoms in AILD
HRQoL measures in which the scores are
converted to a linear scale ranging from 0 Although clearly optimal disease treatment
(equivalent to death) to 1 (perfect health) in AILD is necessary for the best overall out-
[2]. Health utilities make the study of comes, the extent to which it directly
HRQoL less abstract and, critically, allow us improves symptoms is limited; this is
to understand the effectiveness of interven- something which should be considered when
tion in the context of health economic eval- discussing expectations of treatment out-
uation (cost‐effectiveness). One year at a comes with patients. There is no evidence
health utility of 1 represents one quality‐ that UDCA improves either fatigue or pru-
adjusted life‐year (QALY). Reimbursement ritus in PBC [5,6], whilst OCA does not
bodies such as the National Institute for improve fatigue and can worsen pruritus [1].
Health and Care Excellence (NICE) in the Recent data from the UK suggests that
UK will define acceptable costs per QALY although ongoing disease activity is associated
Chapter 17 Symptoms, Chronic Disease, and Patient Management 293
with worse quality of life, so is the use of with the additional specific impact of pru-
long‐term steroid or calcineurin inhibitor ritus in the two cholestatic conditions.
therapy; indeed long‐term steroid therapy is
the major factor associated with impairment Pruritus
of quality of life in this disease [2]. UK‐wide data suggest that up to two‐thirds
of PBC patients will experience pruritus dur-
ing the course of their disease, with around
Stage‐associated Symptoms
half of these experiencing pruritus at any one
In all liver disease states yet studied, symptom time [7]. It is particularly prominent in the
impact is greater and HRQoL lower in cir- typically younger patients with aggressively
rhotic than in pre‐cirrhotic states [3]. The ductopenic disease. Mechanistic studies have
complications of advanced AIH closely pointed to the role of lysophosphatidic acid
mirror those seen in other hepatocellular (LPA) as the direct pruritogen, generated by
diseases, with encephalopathy (characteristic the actions of the enzyme autotaxin (ATX),
cognitive dysfunction and associated fatigue) which is present at elevated levels in patients
and ascites (abdominal discomfort and with cholestatic itch and which has a com-
impaired mobility) to the fore. In the chole- plex functional inter-relationship with bile
static conditions PBC and PSC, symptoms acids. Despite the apparent importance of
from advanced disease are less prevalent. ATX and LPA in the pathogenesis of itch,
This largely reflects the rapid, often understanding of the pathway has not, as yet,
exponential progression seen in cholestatic led to specific therapy. Cholestatic itch can
disease from compensated cirrhosis to death. also occur in PSC but its impact, perhaps
If PBC or PSC patients are seen with decom- paradoxically, appears lower than in PBC.
pensated cirrhosis, then the key question is In PSC, onset or worsening of itch should
whether they are suitable for transplantation always trigger reassessment of the bile duct
(the only effective intervention in this state) given the potential for stricturing as a driver
rather than how to ameliorate individual of cholestasis. The possibility of gallstone
symptoms. disease should also be considered in PBC
Because of the detrimental impact of cir- where there is a specific disease association.
rhosis on quality of life as well as length of Itch assessment in clinical practice can be
life, effective treatment to slow or prevent performed using a numerical rating scale
disease progression is essential in the (rating the impact from 0 to 10). The presence
management of HRQoL in AILD even if, as of pruritus is not always an indication for
already outlined, it may well not directly treatment as many patients with milder itch
improve other symptoms. are not troubled by it. Significant discomfort,
In PSC uniquely, occult cholangitis can be social impacts resulting from embarrass-
a driver of systemic symptoms, particularly ment around scratching in public, or
in older patients, and should always be con- significant night‐time itch (which can be a
sidered and should always be screened for. major contributor to poor sleep and fatigue)
are all indicators of a need for treatment.
The management of cholestatic itch is
Stage‐independent Symptoms
described in detail in national and interna-
The bulk of the HRQoL impairment in AILD tional guidelines ([8] (Table 17.1). There is no
comes not from the features of advanced dis- evidence to suggest that UDCA specifically
ease but from systemic symptoms which can improves pruritus and OCA can worsen pru-
be seen throughout the disease course. ritus or induce it in previously unaffected
Common themes around fatigue, cognitive patients. There are reports that bezafibrate
dysfunction, and emotional and social may improve pruritus, although published
dysfunction emerge across all the diseases, placebo‐controlled trial data with pruritus as
294 Section V Controversies in Autoimmune Liver Diseases
Intervention Notes
Step 1 Imaging to exclude obstruction Especially important in PSC because of structuring risk
to bile duct
Step 2 Bile acid sequestrant Space away from UDCA
(colestyramine 4 g up t.d.s.) Can be poorly tolerated because of taste. Consider
chilling or adding fruit juice
Step 3 Rifampicin 150 mg daily Check LFT after 4 weeks
Stop if significant ALT rise
Increase to maximum 600 mg daily if needed
Step 4 Naltrexone 50 mg daily Can be associated with “opiate withdrawal‐like” response
Cannot use with opiate analgesia (obviously)
Can be a problem in chronic pain states
Step 5 Sertraline 50 mg daily Can be increased if ineffective
Trial evidence positive but experience in practice
suggests limited benefit
Step 6 Experimental drug treatment Opportunity will depend on local activity
Bezafibrate has some promise
Step 7 Physical intervention Plasmapheresis, MARS or nasobiliary drainage
All are effective in the short term but costly, invasive and
with risk
Should only be considered as short‐term “salvage”
procedures
Step 8 Transplantation Instant itch control
Challenge to justify if liver function good
This represents the approach to systematic management of itch as recommended in relevant guidelines. It should
obviously be followed in addition to providing optimal underlying disease management. Steps 3 and 4 can be used
interchangeably based on local preference and experience.
ALT, alanine aminotransferase; LFT, liver function tests; MARS, molecular adsorbent recirculating systems; UDCA,
ursodeoxycholic acid.
an end point are currently lacking. This is key and this mirrors clinical experience of
given the scale of placebo effects in symptom‐ limited if any benefit.
modifying trials in PBC. All second‐line treatments have limitations
Bile acid sequestrants are licensed therapy and it is important to give an adequate trial of
and remain the first line of therapy despite bile acid sequestrants before moving to them.
limitations. Colestyramine has the best sup- Rifampicin, which probably acts through
porting evidence (although this is still very activation of the pregnane X receptor, has the
limited) but is compromised in practice by best supporting evidence (placebo‐controlled
its poor tolerability. Chilling the mixed resin, trials with positive meta‐analysis). However,
or adding fruit juice, can increase accept- tolerability can again be an issue, with rare
ability though many patients still struggle. It cases of significant liver dysfunction and
is also important to space colestyramine commoner gastrointestinal disturbance.
away from UDCA to prevent sequestration Patients should be started at low dose (150 mg
of UDCA in the bowel. Colesevelam, a tablet daily) with liver function monitoring at 2–4
formulation of bile acid sequestrant, offers weeks. If significant elevation of alanine ami-
theoretical advantages. However, a placebo‐ notransferase (ALT) is seen, then the agent
controlled trial failed to show benefit should be discontinued. If well tolerated and
Chapter 17 Symptoms, Chronic Disease, and Patient Management 295
impacted. Even in PBC, where around 50% of improve with exercise therapy; experience
patients experience fatigue, there are very suggests that this can be a challenging area
clearly unaffected patients. Fatigue severity, to obtain patient input.
with the exception of the added element seen ●● The central component is strongly associ-
in end‐stage disease, does not relate to dis- ated with autonomic dysfunction, sleep
ease activity or stage in PBC or PSC; active disturbance, and some features of depres-
inflammatory disease in AIH can be associ- sion; however, fatigue in AILD is not “just
ated with worse fatigue, although patients in depression,” a contention which can harm
biochemical remission can also be impacted, the relationship between patient and clini-
especially if treated with long‐term steroids. cian. There is clear evidence of organic
In each disease state severity of fatigue can CNS abnormality (in PBC in particular,
range from mild, raising questions as to the which is the best characterized of the dis-
extent to which this is disease‐related at all, to eases), with structural and functional
a debilitating symptom with profound impact imaging changes and neurophysiologic
on functional status. It is therefore important abnormality. The functional imaging
to quantify fatigue if it is to be managed effec- changes strongly suggest abnormality in
tively. As is the case with HRQoL in general, connectivity of the hippocampus with
the abstract nature of the symptom means higher brain areas. These findings raise the
that an overarching self‐assessment is possibility that there is a chronic organic
unhelpful and validated measures such as the process impacting on the brains of PBC
PBC‐40 fatigue domain or the Fatigue Impact patients. Recent studies have suggested
Scale should be used. that imaging change is present within
At present there is no consensus as to the months of the diagnosis of disease. As the
mechanism of fatigue in AILD and no specific importance and consequences of this pro-
therapy. The literature supports the follow- cess come to be understood, so it may
ing concepts. impact on disease treatment paradigms in
PBC, with a move toward better treatment
●● There is minimal if any link between earlier in the disease process.
fatigue severity and conventional markers ●● Critically, fatigue is both inherently com-
of disease severity (with the exception of plex (with many potential contributors to
the fatigue worsening seen in advanced the phenotype within and between
disease). patients) and can be sequential in nature:
●● There are both peripheral and central what starts as a central process could
components, with significant variation in become peripheral through the impact of
the relative contribution between patients. deconditioning. This implies that treatment
Central fatigue can overlap with cognitive of the central process, were it to become
symptomatology and includes loss of moti- available, might lose its effectiveness later
vation to undertake activities. Peripheral in the disease process when the secondary
fatigue is characterized by a sensation of deconditioning process has come to
loss of peripheral muscle power with diffi- predominate.
culty sustaining repeat exercise and even
muscle pain. There is currently no specific therapy for
●● The peripheral component is associated fatigue in AILD. Development and evalua-
with a significant bioenergetics abnor- tion of therapies is also challenging given the
mality in muscle, with evidence of over‐ complex phenotype, variation in the contrib-
reliance on anaerobic pathways and uting factors to the phenotype, the challenge
resulting lactic acid build‐up. This compo- in measuring it and, as in the case of pruritus,
nent may represent deconditioning significant placebo effect. However, a sup-
secondary to central fatigue and may portive management approach can help to
Chapter 17 Symptoms, Chronic Disease, and Patient Management 297
Intervention Notes
Step 1 (and Empathize with the Enter into a positive engagement with them. Do
throughout) patient not feel sorry for them but try to understand the
impact the symptom has on their lives. Reassure
them you are there to help them
Step 2 Screen for intercurrent Associated autoimmune conditions (e.g.
or associated conditions hypothyroid disease or autoimmune anemia)
also causing fatigue and Age‐associated disease such as heart failure or type
treat appropriately 2 diabetes mellitus
Step 3 Screen for associated These are features seen in sever fatigue and which
processes worsening worsen the patient experience. Addressing these
fatigue or its impact and will not reverse fatigue but will assist its control
treat appropriately Depression: Typically with SSRI
Autonomic dysfunction: Classically vasomotor
dysfunction (postural dizziness). Look for
inappropriate antihypertensives. Increasing fluid
intake can be helpful. If in doubt seek specialized
autonomic service input
Sleep disturbance: Typically daytime somnolence
that can be severe enough to impact on work.
Consider the possibility of obstructive sleep apnea.
Consider seeking sleep clinic input
Step 4 Help patient with coping It is critical that patients “own” the problem and
strategies the solution
Advise around energy conservation
Advise around scheduling (fatigue is typically
worse later in the day
Advise around maintaining social structures (social
isolation can worsen quality of life)
Advise around the potential value of exercise in
minimizing the peripheral manifestations of
fatigue
Step 5 Experimental drug Opportunity will depend on local activity
treatment
This represents a suggested approach to systematic management of fatigue in PBC. This goes beyond current guideline
recommendations and should therefore be regarded as experimental. There are fewer data around the impact and
pathogenesis of fatigue in PSC and AIH, although the basic approach is still applicable. The approach should again be
followed in addition to providing optimal underlying disease management. Consideration in AIH should be given to
moving away from long‐term steroid maintenance therapy given its association with worsened quality of life (although
clearly change in disease management should not be at the expense of effective disease control). Other than the stage‐
specific fatigue and cognitive symptoms of advanced liver disease, fatigue does not improve with liver transplantation
and this should not be considered as a potential treatment approach without other indications.
are also age‐relevant comorbid conditions positive in their approach. Expressing views
that are fatigue associated (e.g. type 2 suggesting that fatigue is not real, or is a
diabetes). Potential confounding condi- lesser problem than cirrhosis (or indeed pru-
tions should be sought and treated ritus) can lead to a loss of relationship bet-
effectively. ween patient and clinician. Those clinicians
●● Treating associated processes: fatigue is who say they “never see fatigue” might use-
associated with sleep disturbance, fully consider the possibility that patients
autonomic dysfunction and depression. have learned to not mention it because of the
Clearly, cholestatic disease can also be response doing so in the past has elicited.
characterized by pruritus which can dis- One of the single most desirable advances
turb sleep and thus contribute to fatigue. in AILD would be a direct therapeutic agent
All such features should be explored and able to break the cycle of fatigue. At present
treated. Sleep clinic or formal autonomic there is no such licensed agent. The most
assessment (through the “falls and syn- interesting potential therapy is S‐adenosyl-
cope” services which are now widespread) methionine, for which older data suggest a
may be appropriate. One simple step is for beneficial effect in fatigue and for which
the liver clinic to ensure that there are no there are intriguing emerging data to suggest
direct contributing factors to both poor specific deficiency in fatigue‐associated
sleep (reflux, late evening caffeine‐ AILD states. Further trials with mechanistic
containing drinks or pruritus) and as well as symptom evaluation to mitigate the
autonomic dysfunction (excessive antihy- issues with placebo effects are warranted.
pertensive therapy).
●● Coping strategies: ultimately the symptom
Social Isolation Symptoms
is currently untreatable and it is critical that
There are emerging data from PBC to sug-
patients develop strategies that allow them
gest that although individual symptoms
to mitigate the effects. In essence this
clearly impact on patients, there are modi-
revolves around patients “owning the
fiers that limit their overall impact on per-
problem” and, supported by clinicians and
ceived quality of life. The most important of
others, developing coping strategies.
these is social isolation [12]. In simple terms,
Critical elements include pacing them-
highly fatigued PBC patients typically retain
selves (regarding energy as a valuable
good perceived quality of life provided they
resource to be used to its best advantage)
do not have symptoms of social isolation.
and daytime organization (fatigue in AILD
The critical issue is that the characteristic
typically worsens during the course of the
symptoms of AILD can each lead to behavior
day meaning that important activities in the
patterns that can isolate people. Pruritus can
morning should be scheduled for the morn-
result in people avoiding social situations
ing; night shift work should be avoided if at
because of embarrassment about scratching.
all possible). Exercise is beneficial (although
In the case of fatigue, people can look to con-
patients frequently lack the confidence to
serve energy by stopping working or avoiding
do it) and will help mitigate the effects of
social activities (more so if they take place in
deconditioning. As outlined in the follow-
the evening when fatigue might be at its
ing section, social isolation can be a
worse). The net result, however, can be
significant issue and its avoidance through
fatigue benefit at the expense of worsened
coping strategies is key. Patient groups such
overall life quality (“robbing Peter to pay
as the PBC Foundation and LIVErNORTH
Paul”). It is critical, as part of the approach to
in the UK can play a major role.
developing coping strategies, that patients
Behind all of these approaches is a need for are made aware of this risk and that the
clinicians to be empathetic to patients and importance of maintaining social networks is
Chapter 17 Symptoms, Chronic Disease, and Patient Management 299
emphasized. Patient groups may play a key The emerging complexity reflects increasing
role, as can online groups. opportunity to treat patients better. However,
each of the AILDs is rare with only limited
numbers of specialist centers. This raises
ffective Care Delivery
E important questions as to how cutting edge
in AILD therapy reaches all patients in an equitable
fashion. Currently, care delivery largely uses
Ultimately, the AILDs remain non‐curable traditional models. There is a real need for
chronic diseases. We are increasingly able to the AILD community to explore new
control the progression of disease and to approaches to structured care delivery uti-
limit the impact of symptoms. The lizing digital care pathway approaches.
approaches needed to do that are, however, Without a revolution in care delivery to
increasingly complex with multiple treatment mirror the revolution (in PBC at present and
domains to be considered and an increasing AIH in the near future) in treatment options,
number of treatment options in each domain. we are destined to fail to reap the benefits.
300 Section V Controversies in Autoimmune Liver Diseases
References
1 Nevens, F., Andreone, P., Mazzella, G. et al. 7 Hegade, V.S., Mells, G.F., Fisher, H. et al.
(2016). Obeticholic acid in primary biliary (2019). Pruritus in primary biliary cholangitis
cholangitis/cirrhosis: a phase 3 trial. N. Engl. is common and under‐treated: results from
J. Med. 375: 631–643. 2194 patients in the UK‐PBC cohort. Clin.
2 Wong, L.L., Fisher, H.F., Stocken, D.D. et al. Gastroenterol. Hepatol. 17: 1379–1387.e3.
(2018). The impact of autoimmune hepatitis 8 European Association for the Study of the
and its treatment on health utility. Liver (2017). EASL clinical practice
Hepatology 68: 1487–1497. guidelines: the diagnosis and management
3 Orr, J.G., Homer, T., Ternent, L. et al. (2014). of patients with primary biliary cholangitis.
Health related quality of life in people with J. Hepatol. 67: 145–172.
advanced chronic liver disease. J. Hepatol. 9 Goldblatt, J., Taylor, P.J.S., Lipman, T. et al.
61: 1158–1165. (2002). The true impact of fatigue in
4 Jacoby, A., Rannard, A., Buck, D. et al. primary biliary cirrhosis: a population
(2005). Development, validation and study. Gastroenterology 122: 1235–1241.
evaluation of the PBC‐40: a disease specific 10 Hollingsworth, K.G., Newton, J.L.,
health related quality of life measure for Robinson, L. et al. (2010). Loss of capacity to
primary biliary cirrhosis. Gut 54: recover from acidosis in repeat exercise is
1622–1629. strongly associated with fatigue in primary
5 Carbone, M., Mells, G., Pells, G. et al. biliary cirrhosis. J. Hepatol. 53: 155–161.
(2013). Sex and age are determinants of 11 Newton, J.L., Hollingsworth, K.G., Taylor, R.
theclinical phenotype of primary biliary et al. (2008). Cognitive impairment in primary
cirrhosis and response to ursodeoxycholic biliary cirrhosis: symptom impact and
acid. Gastroenterology 144: 560–569. potential aetiology. Hepatology 48: 541–549.
6 Mells, G., Pells, G., Newton, J.L. et al. 12 Dyson, J.K., Wilkinson, N., Jopson, L. et al.
(2013). The impact of primary biliary (2016). The inter‐relationship of symptom
cirrhosis on perceived quality of life: the severity and quality of life in 2055 patients
UK‐PBC national study. Hepatology 58: with primary biliary cholangitis. Aliment.
273–283. Pharmacol. Ther. 44: 1039–1050.
301
Index
Note: Page references in italics refer to Figures; those in bold refer to Tables
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
302 Index