Fibrosis in Autoimmune and Cholestatic Liver Disease

Abstract

Autoimmune and cholestatic liver disease account for a significant part of end-stage liver disease and are leading indications for liver transplantation. Especially cholestatic liver diseases (primary biliary cirrhosis and primary sclerosing cholangitis) appear to be different from other chronic liver diseases with regards to pathogenesis. Portal fibroblasts located in the connective tissue surrounding bile ducts appear to be different from hepatic stellate cells with regards to expression of marker proteins and response the profibrogenic and mitogenic stimuli. In addition there is increasing evidence for a cross talk between activated cholangiocytes and portal myofibroblasts. Several animal models have improved our understanding of the mechanisms underlying these chronic liver diseases. In the present review, we discuss the current concepts and ideas with regards to myofibroblastic cell populations, mechanisms of fibrosis, summarize characteristic histological findings and currently employed animal models of autoimmune and cholestatic liver disease.

Abbreviations

ECM

Extracellular matrix

PBC

primary biliary cirrhosis

PSC

primary sclerosing cholangitis

HSC

hepatic stellate cells

STAP

stellate cell activation-associated protein

CRBP

cellular retinol-binding proteins

LRAT

lecithin-retinol acyltransferase

GFAP

glial fibrillary acidic protein

PDGF

platelet-derived growth factor

FGF-2

fibroblast growth factor-2

TGFβ

transforming growth factor β

HGF

hepatocyte growth factor

VEGF

vascular endothelial growth factor

CTGF

connective tissue growth factor

IGF1

insulin-like growth factor 1

NGF

nerve growth factor

MCP-1

monocyte chemoattractant protein-1

EMT

epithelial to mesenchymal transition

VDR

farnesoid X receptor (FXR), vitamin D receptor

LXR

liver X receptor

AMAs

antimitochondrial antibodies

PDC-E2

E2 subunit of pyruvate dehydrogenase complex

UDCA

ursodeoxycholic acid

AE2

anion exchanger 2

KO

knockout

pANCA

perinuclear antineutrophil cytoplasmatic antibody

ERCP

endoscopic retrograde cholangio-pancreaticography

MRCP

magnetic resonance cholangio-pancreaticography

CFTR

cystic fibrosis transmembrane conductance regulator

DCC

3,5-diethoxycarbonyl-1,4-dihydrocollidine

ANIT

alpha-naphthylisothiocyanate

AIP

autoimmune pancreatitis

Introduction

Chronic liver disease is characterized by continuous liver cell injury followed by compensatory proliferation, inflammatory cell infiltration and deposition of extracellular matrix (ECM) proteins by myofibroblasts. Cirrhosis is the end result of this wound healing response and is characterized by the presence of fibrosis and nodular regeneration throughout the liver. At this stage of disease, replacement of liver tissue by connective scar tissue is associated with loss of normal liver function. Immune mediated and cholestatic liver disease progress to liver cirrhosis and end-stage liver disease if not treated. Indeed, cholestatic liver diseases are a frequent cause for liver transplantation (∼15%) [1].

In this article we will discuss the cellular sources and mediators of fibrosis in cholestatic and autoimmune liver disease and highlight primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC) and their animal models as paradigms of cholestatic disorders, which can serve as model diseases for understanding the mechanisms of biliary fibrosis.

Cellular sources and mediators of fibrosis in cholestatic and autoimmune liver disease

Myofibroblastic cell populations in liver fibrosis with special emphasis of cholestasis

Hepatic myofibroblasts are a heterogeneous population of cells characterized by expression on α-smooth muscle actin and might be discriminated by tissue localization, expression of different markers and response to stimuli (Fig. 1). It is common consent that hepatic stellate cells (HSC) are the main source of collagen in most fibrotic liver diseases. In the normal liver, HSC reside in a quiescent state in the space of Dissé and are the major storage site of vitamin A in the human body. Following liver injury HSC undergo a phenotypic transformation and switch to a proliferative, contractile, migratory, inflammatory and most importantly fibrogenic phenotype [2]. However, the simple paradigm that all collagen found in cirrhotic liver is produced by activated HSC has grown far more multifaceted. It has become increasingly clear that myofibroblasts are a heterogeneous population of cells and do not derive entirely from quiescent HSC [3]. Portal fibroblasts located in the connective tissue around portal tracts appear to be of particular importance in ischemic and cholestatic liver disease and are different from HSC with regards to expression of marker proteins and response the profibrogenic and mitogenic stimuli (Fig. 2) [4]. In this is respect it was shown that isolated portal fibroblasts express elastin but not desmin, whereas HSC consistently express desmin but never elastin [5]. In a comparative proteomic study several proteins were identified that discriminate between HSC and myofibroblasts. Cytoglobin (also known as stellate cell activation-associated protein – STAP) appears to be the best marker for distinguishing HSC from portal myofibroblasts [6]. HSC were first successfully isolated based on their characteristic low density due to cytoplasmic lipid droplets using a discontinuous gradient of arabinogalactan [7]. These lipid droplets are the major site for storage of vitamin A and its metabolites in the body of vertebrates and cause characteristic autofluorescence of retinoids observed in HSC cultures and in unstained frozen sections. Proteins involved in vitamin A metabolism, including cellular retinol-binding proteins (CRBP) or lecithin-retinol acyltransferase (LRAT) are therefore potential markers for specific identification of HSC [5,8,9]. Accordingly, portal myofibroblasts do also not form lipid droplets when retinol is added to the culture media [5]. Other markers with variable expression in portal myofibroblasts are neurotrophin NT-3 and fibroblast marker TE-7 [4,10]. Vimentin, vinculin, synemin, glial fibrillary acidic protein (GFAP), neurotrophin receptor p75 appear to be a marker expressed by both cell types [10,11]. However, only a minority of these markers was studied thoroughly both in vitro and in vivo. In addition there are reports indicating that other cells also express some of these markers, e.g. cellular retinol binding proteins, cytoglobin, p75.

Fig. 1.

Myofibroblastic cell populations in the liver. Myofibroblasts in the liver differ with regards to their localization within a liver lobule (A). HSC are located in the space of Dissé (B). Portal fibroblasts surround cholangiocytes in the portal tracts (C). Smooth muscle cells in vessel walls and surrounding centrolobular veins might also contribute to ECM deposition (D). In fibrotic liver, interface myofibroblasts are located at the border between fibrotic septa and neighboring parenchyma called have been identified (E).

Fig. 2.

Differential expression of marker proteins by portal myofibroblasts and HSC. Whereas portal myofibroblasts and hepatic stellate cells share expression of some marker proteins, some proteins appear to be specific for each particular cell type.

In addition to differences in marker expression there also appears considerable functional differences between portal myofibroblasts and HSC. Whereas, platelet-derived growth factor (PDGF) is the most potent mitogenic stimuli for HSC it inhibits portal fibroblasts proliferation and myofibroblastic differentiation [5,12]. In contrast, fibroblast growth factor-2 (FGF-2) causes portal fibroblast proliferation. Portal myofibroblasts produce significant amounts of transforming growth factor β2 (TGFβ2) and, unlike activated HSC express all three TGFβ receptors and their proliferation is inhibited by TGFβ1 and TGFβ2 [12]. Myofibroblastic differentiation of portal fibroblasts was shown to depend on both TGFβ and matrix stiffness [5]. Interestingly, increasing the stiffness of the culture matrix is associated with increased myofibroblastic phenotype suggesting that portal myofibroblasts differentiate as a function of mechanical tension [5].

The term interface myofibroblasts is used by some authors to specifically refer to myofibroblasts at the edge of between fibrotic septa and neighboring parenchymal cells [13]. In addition, smooth muscle cells in vessel walls (pericytes) and myofibroblasts surrounding centrolobular veins might contribute to ECM deposition in chronic liver disease [14]. At present there are no protocols established for selective isolation of these cells. However, both cell types were reported to express vinculin, p75, synemin and vimentin [10].

Cholangiocytes – victims and culprits in cholestatic liver disease

There also appears to be extensive paracrine cross-talk between cholangiocytes (bile duct epithelial cells) and portal fibroblasts in cholestatic liver disease (Fig. 3). In many forms of liver injury especially in cholangiopathies an increased number of proliferating small bile ducts can be observed which is referred to as ‘ductular reaction’. These newly formed, ‘reactive’ cholangiocytes are characterized by increased expression of anti-apoptotic genes (Bcl-2), neuroendocrine molecules (chromogranin A, NGF), adhesion molecules (NCAM, ICAM), co-stimulatory molecules (MHC II, CD40), cytokines (TNFα, IL-1, IL-6, IL-8, IFNγ), chemokines (MCP-1), growth factors (HGF, VEGF, CTGF, IGF1, NGF) and profibrogenic stimuli (PDGF, TGFβ2, endothelin-1) which have autocrine and paracrine effects on myofibroblasts activation, migration and proliferation [15–26]. In addition some of these factors might also affect Kupffer cells, hepatocytes or endothelial cells functions. Furthermore, in normal healthy liver, portal fibroblasts express ecto-nucleotidase NTPDase2, which converts ATP to AMP. Following cholestatic liver injury portal myofibroblasts loose expression of NTPDase2 allowing activation of basolateral P2Y receptors expressed by bile duct epithelia cells resulting in MAPK mediated proliferation of cholangiocytes [27]. Bile duct epithelial cells are also as source of chemokines. In this respect it was shown that the monocyte chemoattractant protein-1 (MCP-1)/CCL2 is upregulated in biliary fibrosis by cholangiocytes and MCP-1 increases αSMA and α(1)-procollagen production in portal myofibroblasts [28]. The tight relationship between cholangiocyte proliferation and fibrogenesis has lead to the suggestion that the ductular reaction is the ‘pacemaker of portal fibrosis’. Importantly, ‘quiescent’ cholangiocytes do not produce these peptides.

Fig. 3.

Cross-talk between cholangiocytes and myofibroblasts. Following cholestatic liver injury an increased number of bile ducts can be observed. These cholangiocytes are characterized by increased expression of neuroendocrine molecules, adhesion molecules, co-stimulatory molecules, cytokines, chemokines, growth factors and profibrogenic stimuli which have autocrine and paracrine effects on myofibroblasts activation, migration and proliferation.

Epithelial to mesenchymal transition as a potential source of myofibroblasts

It was suggested that myofibroblasts might also originate from epithelial cells via epithelial to mesenchymal transition (EMT) in liver fibrosis. According to this concept cholangiocytes might serve as another source of myofibroblasts via EMT to contribute to portal tract fibrosis in cholestatic liver injury [29–32]. However, recent evidence suggests that EMT is a cell culture phenomenon, which does not occur in vivo and that initial studies relied on inappropriate methodology leading to interpretational pitfalls [33–35].

Role of bile acids and activated nuclear receptors driving fibrosis in cholestatic liver disease

Several key transcriptional regulators including ligand-activated nuclear receptors orchestrate the transdifferentiation of HSC from a quiescent to an activated myofibroblastic-like phenotype producing ECM in the fibrotic liver. Factors increasing as a results cholestasis might also be pro-fibrogenic themselves. In this respect it was shown that bile acids bind to the epidermal growth factor receptor on HSC activating the protein kinase C/extracellular signal-regulated kinase/p70S6K-dependent pathway causing cell proliferation [36]. Stimulation of the nuclear bile acid/farnesoid X receptor (FXR) has been suggested to play a key role since a series of studies suggested that FXR can modulate HSC activity by restoring peroxisome proliferator-activated receptor γ (PPARγ) signalling and by FXR-SHP-dependent inhibition of AP-1 signalling on downstream profibrogenic targets [37]. In contrast to these findings, a recent study in different mouse models of biliary type and non-biliary type of fibrosis revealed that lack of FXR significantly reduces fibrosis of the biliary type, while having no impact on non-cholestatic liver fibrosis [38]. Notably, FXR expression was virtually undetectable in mouse HSC and myofibroblasts and only minimal in human HSC [38]. Except vitamin D receptor (VDR), which can also be activated by hydrophobic bile acids, and liver X receptor (LXR), expression of other nuclear receptors is rather low in activated mouse and human HSC [38]. Despite these recent molecular insights into bile acid signalling, the direct role of bile acids in driving liver fibrosis still remains to be defined. Rather, bile acids appear to promote fibrosis indirectly through hepatocyte injury [39].

Primary Biliary Cirrhosis

Primary biliary cirrhosis is a slowly progressive cholestatic autoimmune liver disease of unknown etiology [40]. It affects individuals of all ethnicities and accounts for ∼1% of deaths from cirrhosis worldwide [41]. Despite being a rather rare cause of cirrhosis, PBC represent a leading indication for liver transplantation [42]. Interestingly, the incidence of PBC is increasing worldwide and PBC is typically diagnosed in women in their 5^th-7^th decade. Antimitochondrial antibodies (AMAs) against the inner lipoyl domain of the E2 subunit of pyruvate dehydrogenase complex (PDC-E2) are the diagnostic hallmark [43]. Histologically, PBC is characterized by chronic, nonsuppurative cholangitis and can be divided into four distinct stages: Stage I is characterized by inflammatory cell infiltrate consisting of lymphocytes, plasma cells and eosinophil granulocytes restricted to the portal triad. Eventually inflammation extends into hepatic parenchyma referred to as interface hepatitis (stage II). Lymphocytes also form bridge-like extensions into the lobes and may be the forerunners of fibrous septa. Fibrotic septa extend from the portal tracts and eventually link portal tracts to each other forming the so-called ‘bridging fibrosis’ (stage III). Complete distortion of the liver architecture by numerous fibrotic septa then leads to cirrhosis with the existence of regenerative nodules (stage IV) [44,45]. Three independent studies have shown that patients with PBC histologically progress by one stage every 1.5 years [44,46,47]. These stages are not always easy to discriminate in biopsy specimen, because lesions are not evenly distributed throughout the liver and partly because the stages may overlap. PBC- typical granulomas are present in ∼15% of patients but are frequently missed in small percutaneous biopsy specimens. Sites of bile duct destruction are marked by aggregates of lymphoid cells and bile duct proliferation might also be observed at histological examination. The median survival for symptomatic patients with PBC ranges from 5 to 8 years from the onset of symptoms [48]. Secondary biliary cirrhosis is most often caused by a previous surgical procedure resulting in bile duct injury. However, other causes of bile duct obstruction such as stones, malignancies, cysts, or parasites must also be considered.

Mouse Models of PBC

NOD.c3c4 mice

The NOD.c3c4 mouse was the first spontaneous mouse model of PBC. It is characterized by destructive cholangitis, granuloma formation and eosinophilic infiltration. In contrast to human PBC the extrahepatic biliary ducts are also affected. Livers are characterized by biliary cyst formation and adjacent lymphocytic infiltrates. These mice also develop PDC-E2 antibodies, which precedes the development of liver abnormalities and ANA [49]. These findings are equivalent to the occurrence if anti-PDC-E2 in humans several years before the overt onset of PBC. Importantly, depletion of T-cells protects from biliary disease indicating that CD4+ and CD8+ T-cells infiltrating the biliary epithelium play a crucial role in the pathogenesis of this disease. Accordingly, adoptive transfer of splenocytes or CD4+ T-cells into NOD.c3c4-scid mice causes autoimmune biliary disease, indicating that T-cells play a crucial role in this animal model of PBC [50].

dnTGFβRII mice

Studies in TGFβ1 knockout (KO) mice indicated that TGFβ1 is one of the key negative regulators of immune homeostasis and its absence causes autoimmune disease in several organs. To delineate the role of TGFβ1 in T cell homeostasis, mice expressing a dominant-negative (dn) TGFβ1 receptor type II under the control of the CD4 promoter were generated resulting in specific abrogation of TGFβ1 signaling in CD4 positive T cells. These mice also develop autoimmune inflammatory disease in multiple other organs including the lungs and intestine [51]. Remarkably, dnTGFβRII mice mimic several characteristic features of human PBC, including lymphocytic cell infiltration with periportal inflammation analogous to histological findings in patients, a serum cytokine profile typical for PBC characterized by increased levels of IFNγ, TNFα, IL-6 and IL-12p40, and most importantly spontaneous production of AMAs directed to the same mitochondrial autoantigens [52].

IL2R-α KO mice

The IL-2R consists of three subunits: α (CD25), β (CD122), and γ (CD132). IL-2Rα transduces a crucial signal for T cell proliferation and is responsible for the development, activity and expansion of the CD4+CD25+ regulatory subset of T cells (Tregs), which reduces proliferation, and activation of effector T cells [53–55]. Tregs play a pivotal role in the adaptive immune system and limit autoreactive T cell responses in many different models of autoimmunity. IL2Rα KO mice show many characteristic features like those in human PBC: IL- increased serum levels of IgG and IgA, mild interface hepatitis and bile duct injury [56]. However, these mice also develop an ulcerative colitis-like disease, usually associated with PSC but not PBC in humans.

Ae2a,b KO mice

The Cl-/HCO3- anion exchanger 2 (AE2) mediates Cl⁻/HCO3⁻ exchange across the plasma membrane and plays a critical role in the regulation of in intracellular pH levels and transepithelial acid-base transport, including biliary bicarbonate excretion stimulated by secretin. Notably, AE2 expression and activity is reduced in liver and mononuclear cells of patients with PBC [57–59]. It was therefore proposed that reduced expression of AE2 and resulting changes in bile composition promote bile duct injury [57–59]. Interestingly, AE2 deficiency in mice alters pH homeostasis in several cells and results in increased perinatal mortality, azoospermia, decreased gastric acid secretion, bone abnormalities, growth retardation and deafness [60,61]. AE2 KO mice display splenomegaly, elevated pH in splenocytes, increased expression of IFNγ and IL12 as wells as increased number of CD8(+) T cells but reduced numbers of CD4/FoxP3(+)T cells. Most AE2 deficient mice also show increased serum levels of IgM, IgG and alkaline phosphatase and develop AMA. However, only one third of mice display extensive portal inflammation and fibrosis of the portal tracts [62].

Chemical xenobiotics as initiators of autoimmune liver disease

Accumulating evidence suggests that environmental factors play a key role in the pathogenesis of PBC. It was suggested that replacement of the lipoyl domain of PDC-E2 by a chemical xenobiotic mimic would be sufficient to break self-tolerance. In a screen more than 100 potential xenobiotics were coupled to PDC-E2, spotted on microarray slides and were assayed for Ig reactivity using sera from patients with PBC. This led to the identifiaction of 2-octynoic acid which commonly occurs in perfumes, soaps, detergents, lipsticks, toilet waters, facial creams, and many common food flavorings as a potential xenobiotic [63]. Remarkably, C57BL/6 mice immunized with 2-octynoic acid coupled to bovine serum albumin develop antibodies to PDC-E2 and increase serum levels of TNFα and IFNγ. Histologically these mice are characerized by inflammatory cell infiltration and the development of biliary disease [64].

Autoimmune Hepatitis

Autoimmune hepatitis (AIH) is characterized by elevated serum transaminase levels, hypergammaglobulinemia, presence of autoantibodies, and chronic hepatitis [65]. Similar to PBC it mostly affects women (∼75%). The histological hallmark of AIH is interface hepatitis, which however is not diagnostic for AIH as it can also be found in other chronic liver diseases (e.g. viral hepatitis). Hepatocellular damage and a mixed inflammatory cell infiltrate consisting of lymphocytes, plasma cells and sometimes eosinophiles can be seen in the portal tracts, at the interface, but also within the parenchyma. Bridging necrosis is common and surviving hepatocytes often form gland-like structures called ‘hepatitic rosettes’. The degree of inflammation may vary and cholestasis, bile duct damage or even ductopenia might be present in some patients. Furthermore, in rare instances the principal features at initial presentation are centrilubular necrosis or inflammation in perivenular areas, typically seen in acute hepatitis [66]. Cirrhosis is observed in one third of patients at initial diagnosis but is reversible following immunosuppressive treatment. Patients with AIH usually respond rapidly to corticosteroid treatment and follow-up biopsies show varying degrees of resolution of the necro-inflammatory process. Fibrosis and other structural changes may also improve. Importantly, histological features can overlap with those of PSC or PBC and a careful histological examination is therefore mandatory [67,68].

Mouse Model of AIH

TGFβ1 KO mice

TGFβ1 plays a key role in many cellular functions, including the control of cell growth, proliferation, differentiation and apoptosis and therefore affects biological processes such as tissue remodeling and wound healing, carcinogenesis, and immune responses. Most importantly, TGFβ1 is a negative regulator of survival, activation, proliferation and effector functions of several cells of the immune system. Accordingly, mice lacking TGFβ1 display excessive inflammation characterized by inflammatory lesions primarily in the heart and lungs resembling autoimmune disorders [69].

Breeding TGFβ1 KO mice into the BALB/c background dramatically modified the phenotype of TGFβ1 KO mice and resulted in the development of spontaneous hepatitis resembling human AIH. Accordingly, BALB/c-TGFβ1 KO mice develop lethal inflammatory hepatitis that was not observed in other genetic backgrounds. Livers of BALB/c-TGFβ1 KO mice show accumulation of activated CD4(+) T cells characterized by high expression of IFNγ. TGF-β1/IFN-γ double KO mice do not develop necroinflammatory hepatitis suggesting that IFNγ plays a critical role in this mouse model [70]. These findings suggest that TGFβ1 is a critical immunosuppressive cytokine that prevents development of autoimmune liver disease.

Primary Sclerosing Cholangitis

Primary sclerosing cholangitis is characterized by chronic inflammation and obliterative fibrosis of the intra- and/or extra-hepatic biliary tree. PSC typically presents during the 3^rd-4^th decade [71], and the most recent epidemiological study indicates that PSC is as common as PBC and AIH [72]. Autoimmune features are common in PSC, antinuclear antibodies are present in up to 50% and perinuclear antineutrophil cytoplasmatic antibody (pANCA) in 80% of cases [73]. Seventy to 80% of patients with PSC have or will develop IBD, mainly ulcerative colitis, and ∼15% of patients develop cholangiocarcinoma [74–76]. PSC mostly affects medium- and large- sized bile ducts, which are typically not seen in percutaneous liver biopsy specimen. Accordingly, endoscopic retrograde cholangio-pancreaticography (ERCP) and more recently magnetic resonance cholangio-pancreaticography (MRCP) are considered the gold standard for the diagnosis of PSC.

Small duct PSC refers to a variant of PSC that is characterized by cholestasis and histological features typical for PSC but mostly normal bile ducts on cholangiography [77]. In contrast to PSC, small duct PSC can only be diagnosed by percutaneous liver biopsy, as small ducts are typically present in biopsy specimen. Small duct PSC has been reported to be less rapidly progressive than large duct PSC with a lower incidence of cholangiocarcinoma. Notably 20% of patients develop regular, large duct PSC [78].

The characteristic histological feature of PSC is concentric periductal fibrosis (‘onion-skinning’) that gradually progresses and eventually leads to obliteration and loss of interlobular bile ducts. Portal inflammation and periductal edema are other characteristic histological features. The duct epithelium may show various degrees of atrophy and might even disappear, leaving a characteristic fibro-obliterative scar. Major bile ducts are often inflamed, ulcerated or dilated. Ludwig proposed a histological staging system based on characteristic features as well as extent of inflammation and fibrosis. The stages are similar to those of PBC: Stage I is marked by bile duct injury and portal inflammation with minimal fibrosis; Stage II by expansion of portal tracts, periportal fibrosis and further inflammation; Stage III by fibrous septa, bridging fibrosis, and progressive ductopenia; Stage IV by cirrhosis.

Animal models of PSC

Rat model of small bowel overgrowth

Bacterial cell wall components have been suggested to play a key role in the pathogenesis of PSC by triggering the innate immune system. Surgical creation of blind loops of the jejunum in rats promotes bacterial overgrowth associated with bacterial translocation and exposure of Kupffer cells to bacterial cell wall components, especially peptidoglycan-polysaccharides [79]. Histologically this animal model is characterized by portal inflammation, bile duct proliferation, interlobular bile duct destruction and biliary fibrosis. Cholangiography shows strictures of bile ducts with intervening areas that are irregulary tortuous and dilated [80]. This appears to be TLR4 independent, since treatment with gentamicin or polymyxin B, targeting lipopolysaccharide does not affect development of liver disease. In contrast, metronidazole or tetracycline treatment prevents hepatic [81].

Mdr2 KO mouse

The multidrug resistance gene (mdr2) encodes a P-type glycoprotein critical for the transport of phosphatidylcholine into the outer leaflet of the canalilcular cell membrane, facilitating the subsequent extraction of bile acids and excretion into the bile. Accordingly, mdr2 KO mice lack phosphatidylcholine excretion into bile, which renders bile toxic because of an increased concentration of free non-micellar bile acids [82]. Mdr2 KO mice display accumulation of toxic bile salts in the intrahepatic biliary system, which eventually disrupts tight junctions and basement membranes of bile ducts. This causes regurgitation of bile acids into portal tracts leading to non-suppurative inflammatory cholangitis, onion-skin type periductal fibrosis and ductular proliferation. Moreover, this results in activation of periductal myofibroblasts, periductal fibrosis causing obliterative cholangitis owing to atrophy and death of bile duct epithelial cells at 2 months of age [83]. At the age of 4–6 months mdr2 KO mice start to develop multiple foci in the liver parenchyma, culminating in macroscopically visible tumor nodules. Microscopically, these lesions consist of sheets of atypical hepatocytes with loss of orderly structured architecture and reticulin pattern [84]. The mdr2 KO mouse therefore represents a valuable model to investigate the development of HCC in an inflammatory and fibrotic setting [85].

CFTR

As mentioned earlier, changes in bile composition as a consequence of altered pH for instance might affect progression of disease. In this respect, an increased prevalence of variations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene was noted in a subset of patients with IBD and PSC [86]. Interestingly, deletion of exon 10 of the CFTR gene in mice does not result in liver disease. However, induction of colitis in these mice causes bile duct injury characterized by portal tract and bile duct infiltration with mononuclear cells and bile duct proliferates [87].

Chemical xenobiotics as initiators of sclerosing cholangitis liver disease

Feeding of 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DCC) represents another valuable mouse model to investigate the mechanisms of xenobiotic-induced cholangiopathies and subsequent biliary fibrosis. DCC feeding causes increased biliary secretion of porphyrin with subsequent inflammation associated with expression of TNFα, VCAM and osteopontin by bile duct epithelial cells. This leads to pericholangitis, ductular reaction and biliary fibrosis caused by activation of periductal myofibroblasts [88]. Litocholic acid feeding in mice leads to segmental bile duct obstruction, destructive cholangitis and periductal fibrosis. At ultrastructural level small bile ducts are frequently obstructed by crystals, suggesting that mechanical injury and subsequent obstruction is crucial for the development of bile duct injury [89]. Alpha-naphthylisothiocyanate (ANIT) is another drug, which is toxic to bile ducts and induces ductular proliferation. Oral administration of ANIT to rats causes inflammatory infiltrates in portal tracts progressing to portal inflammation and extensive fibrosis over the course of 2 weeks [90].

IgG4-related sclerosing cholangitis

Sclerosing cholangitis is sometimes also observed in patients with autoimmune pancreatitis (AIP) and was initially referred to as atypical PSC. IgG4-antibody stained plasma cells can be found in various organs including the liver and serum IgG4-levels are increased in patients with AIP are useful markers to distinguish AIP from other pancreatic diseases. Accordingly sclerosing cholangitis associated with AIP is now referred to as IgG4-related sclerosing cholangitis and can be discriminated from PSC by quantification of serum IgG4 levels and IgG4 immunostaining of biopsy specimen. IgG4-related sclerosing cholangitis shows cholangiographic features similar to PSC but in contrast to PSC both AIP and IgG4-related sclerosing cholangitis respond well to steroid therapy and biliary drainage[ [91–93]. Histological findings of IgG4-related sclerosing cholangitis include portal inflammation, large bile duct damage, portal sclerosis, lobular hepatitis, and canalicular cholestasis [94].

Overlap syndromes

In the majority of cases the so-called ’overlapping-syndromes’ occur between AIH and PBC or AIH and PSC. Depending upon diagnostic criteria and patient selection 2–19% of patients with PBC and 7–14% of patients with PSC have features that overlap with those of AIH. There are no standardized criteria for definition of overlapping syndromes and overlapping features include symptoms, biochemical tests, a variety of immunological findings as well as histology. Accordingly, patients with overlap syndromes may present histological features of different immune-mediated liver diseases.

Outlook

There has been substantial increase in our understanding of the mechanisms underlying autoimmune and cholestatic liver disease in the past years. However, only little is known about disease specific mechanisms with regard to myofibroblast activation and ECM deposition. For instance, the relative contribution of HSC to cholestatic liver disease remains elusive. Whereas, most animal models of autoimmune and cholestatic liver disease have provided significant insights into the mechanisms leading to autoimmunity and liver injury, only little is known about the pathways causing myofibroblast activation and ECM deposition. The successful establishment of isolation protocols for different myofibroblast cell types and disease specific mouse models will help to address this and other questions in the future.

Practice points.

• •Autoimmune and cholestatic liver disease are leading indications for liver transplantation
• •PBC is typically diagnosed in women in their 5^th-7^th decade
• •Antimitochondrial antibodies against the E2 subunit of pyruvate dehydrogenase complex are the diagnostic hallmark of PBC
• •Granulomas are present in ∼15% of patients with PBC
• •AIH mostly affects women
• •Interface hepatitis and hepatitic rosettes are typical histological features of AIH
• •Cirrhosis is observed in one third of patients with AIH but is reversible following immunosuppressive treatment
• •PSC typically presents during the 3^rd-4^th decade
• •Other autoimmune diseases like inflammatory bowel disease are common in patients with PSC
• •Antinuclear antibodies are present in ∼50% and perinuclear antineutrophil cytoplasmatic antibody in ∼80% of patients with PSC
• •Magnetic resonance cholangio-pancreaticography is the gold standard for the diagnosis of PSC
• •IgG4-related sclerosing cholangitis can be observed in patients with autoimmune pancreatitis
• •IgG4-related sclerosing cholangitis is characterized by increased IgG4 serum levels and presence of IgG4-positive B-cells in biopsy specimen
• •Overlapping-syndromes between AIH and PBC or AIH and PSC should be considered in the differential diagnosis
• •Overlapping features include symptoms, biochemical tests, immunological findings and histology

Research agenda.

• •Evaluation of the quantitative contribution of portal fibroblasts and hepatic stellate cells to fibrosis in PBC and PSC
• •Investigate long-term outcomes and evolution of fibrosis in existing animal models of PBC, PSC and AIH
• •Establishment of new animal models of PBC and AIH that allow investigating myofibroblast activation and mechanisms of fibrosis in these diseases
• •Establishment of mouse models (GFP reporter mice, HSV-TK transgenic mice, Cre mice) that allow to specifically follow, ablate and manipulate portal myofibroblasts and hepatic stellate cells in experimental liver fibrosis

Conflict of interest statement

All authors have nothing to declare.