Animal models of biliary injury and altered bile acid metabolism

Abstract

In the last 25 years, a number of animal models, mainly rodents, have been generated with the goal to mimic cholestatic liver injuries and, thus, to provide in vivo tools to investigate the mechanisms of biliary repair and, eventually, to test the efficacy of innovative treatments. Despite fundamental limitations applying to these models, such as the distinct immune system and the different metabolism regulating liver homeostasis in rodents when compared to humans, multiple approaches, such as surgery (bile duct ligation), chemical-induced (3,5-diethoxycarbonyl-1,4-dihydrocollidine, DDC, α-naphthylisothiocyanate, ANIT), viral infections (Rhesus rotavirustype A, RRV-A), and genetic manipulation (Mdr2, Cftr, Pkd1, Pkd2, Prkcsh, Sec63, Pkhd1) have been developed. Overall, they have led to a range of liver phenotypes recapitulating the main features of biliary injury and altered bile acid metabolisms, such as ductular reaction, peribiliary inflammation and fibrosis, obstructive cholestasis and biliary dysgenesis. Although with a limited translability to the human setting, these mouse models have provided us with the ability to probe over time the fundamental mechanisms promoting cholestatic disease progression. Moreover, recent studies from genetically engineered mice have unveiled ‘core’ pathways that make the cholangiocyte a pivotal player in liver repair. In this review, we will highlight the main phenotypic features, the more interesting peculiarities and the different drawbacks of these mouse models.

Cholestatic liver disorders are chronic disease conditions, resulting from a functional impairment of bile formation and/or bile flow due to defective secretion by hepatocytes or cholangiocytes, or to mechanical obstruction of bile flow through intra- or extrahepatic bile ducts [1–3]. From a clinical point of view, a common end-point of many forms of cholestasis is the progression to biliary fibrosis with development of portal hypertension not necessarily associated with cirrhosis and, eventually, to end-stage liver disease independently of etiology [2–4]. In the last 25 years, a number of different animal models have been established and have laid the basis for the development of novel pathogenetic concepts and, consequently, for testing in vivo new treatment strategies for this group of liver diseases. Among several vertebrate animals that have been variably considered, rodents, and particularly mice, have drawn special interest from basic and clinical researchers because of the many experimental advantages they can offer. They are small, have short life span and gestational period, and are therefore easy to maintain and breed in captivity [5,6]. Furthermore, the remarkable genetic similarity between mice and humans, combined with the possibility and convenience of genetic manipulation, provide us with the possibility to study the implication of specific genes/signaling pathways and relatively, to identify potential therapeutic targets [7]. Unfortunately, given the multiple differences between human and mouse, including liver morphology, no single mouse model can strictly reproduce all the features of human cholestatic liver diseases to date. Indeed, large variations in responses to noxious agents exist between humans and mice regarding pathogenicity, timing, and immuno-inflammatory reactions [5–8].

In the present review, we will systematically summarize the most commonly used rodent models available to study cholestatic liver disease and, in particular, the cholangiopathies (Table 1). In the first part of this review, we will briefly highlight the main features and the more interesting peculiarities of the current animal models in use for biliary injuries and altered bile acid metabolism. In the second part, we will focus on the genetically modified mice that target specific gene products affecting cholangiocyte functions essential for biliary repair. The fundamental histological biliary lesions variably reproduced by the different mouse models are summarized in Fig. 1.

Table 1.

Rodent models of biliary injury^*.

	Disease model	Rodent model	Clinical phenotype
Surgical
Bile Duct Ligation	Obstructive cholestasis	Surgical bile duct ligation and dissection rats or mice	Jaundice, rapid portal fibrosis
Chemical
DDC	Chronic cholestasis (PSC)	3,5-Diethoxycarbonyl-1,4-dihydrocollidine treated mice	Segmental bile duct obstruction, pericholangitis involving extrahepatic bile ducts
ANIT	Acute and chronic cholestasis (PBC)	α-Naphtylisothiocyanate-treated rats or mice	Cholangitis confined to intrahepatic bile ducts with hepatocellular damage
Infective
RRV-A	BA	Rhesus rotavirus type A (RRV)-treated mice	Progressive extrahepatic ductal obstruction
Genetic
Mdr2	PFIC-3, ICP, PSC	Abcb4−/−	Non-purulent cholangitis, biliary cirrhosis
BSEP(Spgp)	PFIC-2, BRIC2	Abcb11−	None
CFTR	CFLD	CFTRtm1UNC	Ductal plugging, focal biliary cirrhosis
		F508del
PKD1	PLD-ADPKD	Pkd1(flox/):pCxCreER	Liver cysts non communicating with the biliary tree
PKD2		Pkd2(flox/):pCxCreER
PRKCSH	ADPLD	Prkcsh(flox/):pCXCreER	Liver cysts non communicating with the biliary tree without kidney involvement
SEC63		Sec63(flox/): pCXCreER
PKHD1	ARPKD,CHF, Caroli syndrome	Pkhd1del4/del4 PCK	Progressive bile duct enlargement with portal fibrosis; renal involvement only in PCK rat

PSC, Primary sclerosing cholangitis; PBC, Primary biliary cholangitis; BA, Biliary atresia; PFIC-3, Progressive familial intrahepatic cholestasis type 3; ICP, Intrahepatic cholestasis of pregnancy; PFIC-2, Progressive familial intrahepatic cholestasis type 2; BRIC2, Benign recurrent intrahepatic cholestasis type 2; CFLD, Cystic fibrosis liver disease; PLD-ADPKD, Polycystic liver disease associated with autosomal dominant polycystic kidney disease; ADPLD, autosomal dominant polycystic liver disease; ARPKD, Autosomal recessive polycystic kidney diseases; CHF, Congenital hepatic fibrosis.

^*The table includes only the animal models discussed in this review.

Fig. 1.

Fundamental phenotypes of biliary injury reproduced by a range of mouse models of different origin. The diagram illustrates the diverse biliary lesions recapitulated by each animal model discussed in the review, behaving as critical determinant of disease progression in cholestasis and cholangiopathies.

1. Bile duct ligation (BDL)

Bile duct ligation (BDL) is the most widely and the longest used experimental model for cholestasis because of its high reproducibility. Originally developed in rats, this model has been then successfully adapted for mice [9,10]. This technique requires a midventral laparotomy and isolation of the common bile duct above the duodenum, followed by double ligation of the bile duct and dissection between the ligatures, thus generating a model of obstructive cholestasis [11]. The structural and functional changes induced by BDL have been extensively analyzed and largely described in the literature. Briefly, surgical BDL induces strong proliferation of cholangiocytes in conjunction with variable activation of oval cells, i.e. hepatic progenitor cells, depending upon additional liver injuries, resulting in extensive ductular reaction, cholestasis, portal inflammation and rapid establishment of biliary fibrosis [9,11,12]. Although experimentally convenient, the flipside of the fast time course of biliary lesions is the fact that development of liver cirrhosis after only 30 days is clearly at odds with the slow progression of most chronic liver diseases, particularly cholangiopathies, usually spanning over decades [13]. More faithfully, this model reproduces acute obstructive biliary lesions, which, however, are rarely seen in human pathology, and are related to specific clinical settings, such as biliary atresia and choledocolithiasis [9,11,12,14]. Therefore, the BDL model has been predominantly used in studies evaluating cholangiocyte proliferation, apoptosis and portal fibrosis due to extrahepatic cholestasis. It is also useful in studies assessing the therapeutic effects of biliary drainage on hepatic blood flow and portal pressure [9]. Finally, additional issues are the relatively high mortality rates due to bile leakage and rupture of a biliary cyst (or gallbladder in mice) that may occur when performing BDL [11].

2. 3,5-Diethoxycarbonyl-1,4-dihydrocollidine (DDC)

Based on these considerations, to study the early pathological alterations occurring in chronic cholestatic diseases, such as primary biliary cholangitis (PBC) and primary sclerosing cholangitis (PSC), is essential to rely on an animal model with slow development of biliary injury [15–17]. Chronic DDC-feeding in mice is a well-established model of cholestatic liver injury originally proposed to study Mallory-Denk body formation, which is specifically associated with metabolic liver injury, as observed in alcoholic and nonalcoholic steatohepatitis, based on DDC ability to induce chronic oxidative cell stress [18–20]. Conversely, cholestatic effects exerted by DDC depends upon the ability to stimulate biliary porphyrin secretion that after a 4-week treatment leads to the generation of intraductal pigment plugs. Due to these properties, chronic DDC feeding has been proposed also as a model for xenobiotic-induced cholangiopathy, featuring ductular proliferation, intense pericholangitis associated with onion skin-type periductal fibrosis, which slowly progressed over time, leading to portal-portal bridging, and extending to the large bile ducts, thus recapitulating specific pathological hallmarks of human PSC [15,19]. This type of progressive biliary injury is initially characterized by specific transporter abnormalities, involving canalicular expression of Sodium/Taurocholate Cotransporter (Ntcp), organic anion transporting polypeptide (Oatp4), and Multidrug Resistance-Associated Protein 2 (Mrp2), responsible for reduced biliary excretion of glutathione (GSH) and phospholipids, occurring before phenotypic changes of cholangiocytes (‘reactive cholangiocyte’) [19]. Reactive cholangiocytes show increased expression of profibrogenic and proinflammatory cytokines, including osteopontin and tumor necrosis factor-alpha (TNF-α), associated with infiltration of neutrophils, and then with activation of periductal myofibroblasts, around both large and small bile ducts [21]. Recent morphological evidence showed that DDC-induced liver injury also leads to hepatocellular necrosis and, consequently, activation of Kupffer cells and compensatory hepatocyte proliferation [22–24]. Therefore, it is largely recognized the usefulness of this model to investigate the mechanisms of chronic cholestasis targeting both cholangiocytes and hepatocytes, and to test novel therapeutic approaches for these conditions [19,21].

3. α-Naphthyl-isothiocyanate (ANIT)

Similar to DDC, ANIT-feeding in mice provides a valuable model to investigate the molecular mechanisms of chemically induced cholangiopathy, with both acute and chronic injury [25]. ANIT is metabolized by hepatocytes and after conjugation with GSH, secreted via MRP2 into bile, whereby it exerts its toxic effects on cholangiocytes. Since ANIT-GSH complex is unstable in the bile, free ANIT undergoes absorption and metabolism recycling cycles, leading to progressively higher biliary concentrations, which become toxic [26]. Persistent exposure to toxic concentrations of unstable ANIT conjugates results in both cholangiocyte and hepatocyte damage and necrosis [27]. Low dose administration of ANIT over the course of 8 weeks in mice leads to bile duct proliferation, periportal inflammation, coupled with progressive biliary fibrosis and mild hepatocellular injury with increased transaminase levels [28–31]. On the other hand, administration of a single, high dose of ANIT (300 mg/kg body weight) to mice leads to rapid (15 to 24 h) cholestasis caused by severe destruction of cholangiocytes often extending to periportal hepatocytes [32]. TGF-β and α_Vβ₆ integrin are intracellular signaling pathways involved in the mediation of obstructive cholestatic injury [29,33,34]. Notably, as compared with the DDC model, large extrahepatic bile ducts mostly remained unaffected, thus the biliary phenotype in ANIT-treated rodents resembles more the pathological findings observed in human PBC.

4. Infected RRV type A mice

In contrast to PBC and PSC, biliary atresia (BA), a multifaceted pediatric colangiopathy of heterogeneous pathogenesis, shows a more rapid progression to biliary cirrhosis due to a fibro-inflammatory obstruction of the extrahepatic bile ducts, associated to a marked proliferation of the intrahepatic ductal component developing together with variable inflammation and fibrosis at the portal level [35]. Several experimental models, including invertebrate animals (sea lampreys and zebrafish) have been proposed to investigate BA [36]. However, to date, mice infected with Rhesus rotavirus type A (RRV) in the first two postnatal days (exploiting the time-restricted susceptibility of cholangiocytes to infections after birth) represent the most accepted model since it captures several fundamental features of the human condition [37–39]. In particular, intrahepatic bile duct hyperplasia associated with portal inflammation dominated by a type I cell infiltrate, are faithfully reproduced in this model [40,41]. Thus, the RRV model provides unique opportunities: a) to probe the progression of extrahepatic bile duct obstruction over the course of the disease, and b) to unravel the role of the immune cell response in promoting epithelial injury and ductal obstruction [40–42]. However, given the early mortality of these mice at 2 weeks of age (thus before cirrhosis may develop), this model is unfitting to study mechanisms linking biliary injury to peribiliary fibrosis.

5. Mdr2-KO mice

The Mdr2 knockout (Mdr2-KO) mouse is one of the best-characterized models of fibrosing cholangiopathy (although caused primarily by hepatocyte dysfunction), generated by Smit and colleague [43,44] which in the last few years has gained considerable interest as surrogate model of PSC [15,16]. The multidrug resistance gene (Mdr2 in rodents/MDR3 in humans) belongs to the group of ABC transporters (ATP binding cassette subfamily B member 4 or Abcb4) [45] and encodes a canalicular flippase (translocase) expressed by hepatocytes, mediating the transport of biliary phospholipids, mainly phosphatidylcholine, into the outer leaflet of the canalicular cell membrane, which facilitates their subsequent excretion into the bile [44,46]. When Mdr2 is defective, lack of phospholipids in the bile profoundly affects its chemical composition, with increased biliary concentration of non-micellar-bound free bile acids, which exert detergent effects on the cell membrane, mainly at the cholangiocyte level [43,45,46]. Therefore, disruption of tight junctions and of the basement membrane beneath the biliary epithelium causes bile leakage to the portal tract, which induces inflammation, brisk ductular proliferation, ultimately leading to peribiliary deposition of fibrotic tissue. Obviously, the Mdr2-KO mouse has important implications for understanding the pathophysiology of human diseases derived from MDR3 defects, which are involved in a wide spectrum of cholestatic liver disorders, ranging from progressive familial intrahepatic cholestasis type 3 (PFIC-3), intrahepatic cholestasis of pregnancy (ICP), to adult biliary cirrhosis [47–19]. Moreover, as a highly reproducible biliary fibrosis model, with well-established and easily detectable readouts, the Mdr2-KO mouse has been often used to test innovative pharmacological strategies aimed at hindering scarring progression. From this point of view, nor-ursodeoxycholic acid (norUDCA) or UDCA treatment as well as integrin blockade have already proven to ameliorate fibrosis in Mdr2-KO mice, though some concerns have been raised with respect to their applicability to the human context [50–53]. Interestingly, at the age of 4–6 months, liver disease related to Mdr2 defects evolve to macroscopically visible tumor nodules which, quite surprisingly, are histologically compatible with hepatocellular carcinoma, rather than cholangiocarcinoma [22,54,55]. Thus, the Mdr2-KO mouse represents also a valuable model to investigate the long-term pro-tumorigenic functions played by a peribiliary chronic inflammatory response induced by cholestasis.

6. BSEP(Spgp)-KO mice

The bile salt export pump (BSEP) or sister of P-glycoprotein (Spgp) encoded by ABCB11, is the primary efflux transporter (member of the ATP binding cassette-ABC superfamily) of bile acids localized at the canalicular membrane of hepatocytes [56]. When BSEP function is impaired, defective bile acid secretion leads to progressive severe cholestasis. BSEP deficiency is associated with progressive familial intrahepatic cholestasis type 2 (PFIC-2), benign recurrent intrahepatic cholestasis type 2 (BRIC2), and several forms of acquired cholestasis [49,57–59]. Interestingly, rodent models lacking Bsep do not develop severe cholestasis probably due to the replacement of the bile acid pool with more hydrophilic bile acids, including muricholic acid and atypical bile acid species, which instead are not produced in humans [19,56,60]. Consequently, these animals do not show any histopathologic sign of liver injury unless treated with cholic acid [61].

Overall, the above-mentioned animal models have some major drawbacks in terms of translatability into clinic, particularly with respect to the stark differences in severity and celerity from human cholestasis. Therefore, mice harboring genetic inactivation of functional proteins normally expressed by cholangiocytes may offer a more fascinating prospect to unravel the complex scenario of biliary injury as it really occurs in humans. Moreover, targeting specific proteins involved in biliary ontogenesis enables us to probe selective morphogenetic pathways that are recapitulated in biliary repair.

7. Genetic mouse models of cholangiocyte dysfunction

7.1. Cftr-KO mice

Cystic fibrosis liver disease (CFLD) is a severe chronic cholangiopathy characterized by progressive peribiliary fibrosis that develops in less than 5% of patients with cystic fibrosis (CF) but, if present, represents the major cause of morbidity/mortality in these patients. Since the defective protein, cystic fibrosis transmembrane conductance regulator (CFTR), is a membrane channel selectively expressed by ductal epithelia, including cholangiocytes [62–64] where it regulates bile flow, Cl⁻ and HCO₃⁻ secretion, vesicle-mediated fluid secretion and apical release of ATP, CFLD has been classically considered a consequence of the impaired bile flow and biliary alkalinization [63,65]. Almost 2000 different CFTR mutations have been identified and are grouped into five classes based on their functional defect and decreased severity [66,67]. Among them, deletion of a phenylalanine residue at position 508 (F508del, class II) is the most common mutation being present in 80% to 90% of patients with CF [66] and generation of F508del mice provide new tools to identify genes that are associated with residual ΔF508CFTR activity, leading to new therapeutic approaches [68–70].Adifferent mutation, S489X, blocking Cftr transcription, is harbored by congenic C57BL/6J-Cftr^tm1Unc (Cftr-KO) mice. Studies performed in these mice have highlighted the concept that loss of CFTR function is not sufficient by itself to cause liver disease [69,71–73]. Thus, other factors, including modifier genes or environmental factors, might play a pathogenic role. Indeed, Cftr-KO mice express the secretory defect but develop liver disease under certain specific inbreeding conditions, or very late in life. More recent studies showed that in CF cholangiopathy the genetic defect is linked with a ‘second hit’ generated by an altered biliary innate immunity after exposure to potentially hepatotoxic conditions [74–76]. Indeed, dextran sodium sulfate (DSS)-feeding in Cftr-KO mice induced biliary damage and portal inflammation [75]. Noteworthy, despite its channel function, CFTR regulates TLR-mediated pro-inflammatory responses, by negatively controlling the activation of Src kinase, the non-receptorial tyrosine kinase responsible for phospho-TLR4 (Tyr674). When CFTR function is defective in cholangiocytes, Src is free to target TLR4 and increase its response to endotoxins, i.e. LPS [75,77]. In addition, aberrant activation of Src decreases the epithelia barrier function by destabilizing cell junctional complexes [77,78].

7.2. Pkd1- and Pkd2-KO mice and mice harboring other genetic defects related to polycystic liver disease

Polycystic liver disease associated with autosomal dominant polycystic kidney disease (PLD-ADPKD) is part of a spectrum of inherited cystic diseases that also includes autosomal dominant polycystic liver disease (ADPLD). Patients with PLD-ADPKD develop bilateral fluid-filled cysts in the kidney accompanied, in approximately 90% of cases, by multiple cysts scattered throughout the liver parenchyma without connection with the biliary tree. PLD-ADPKD is caused by mutations in either PKD1 or PKD2, the genes encoding for polycystin-1 (PC1) and polycystin-2 (PC2), respectively [79–82]. PC1 and PC2 are membrane proteins localized in the ductal epithelia, both in kidney and liver, where they regulate signaling pathways involved in epithelial cell morphogenesis, differentiation and proliferation [80,81]. Liver disease caused by defects in PC1 and PC2 are characterized by the formation of a large number of bile duct-derived liver cysts throughout the liver parenchyma without connection with the biliary tree, that may reach remarkable sizes and lead to major complications (mass effect, rupture, bleeding) [81,82]. Using the tamoxifen-induced Cre-mediated recombination for gene inactivation, Pkd1-KO and Pkd2-KO mice develop a liver phenotype closely resembling the human PLD-ADPKD [79,81,83,84]. In these mice, bile ducts are normal before the tamoxifen-induced gene inactivation, subsequently develop cystic lesions evident 4 weeks after induction, which progressively enlarge through maturation [85]. The mechanisms underpinning biliary cystogenesis have been recently elucidated. In Pkd2-KO mice, the cystic epithelium was shown to produce the vascular endothelial growth factor (VEGF) and to express the cognate receptor VEGFR2, leading to an autocrine loop characterized by cAMP/PKA-dependent activation of the extracellular signal-regulated kinase 1/2 (ERK1/2) pathway that, by stimulating the cystic epithelium proliferation, sustains the progressive cyst enlargement [85,86]. Likewise, VEGF secreted by the cystic biliary epithelium exerts a paracrine effect on the pericystic vasculature, further supporting the cyst growth, like observed in human patients [87,88]. Of note, the perturbation of the cAMP signaling observed in the cystic epithelia associates to changes in the intracellular Ca²⁺ homeostasis and to the involvement of Ca²⁺-inhibited adenylate cyclase isoform 5 (AC5) via the activation of an alternative pathway (store-operated cAMP production-SOcAMP) [89–91].

Liver cysts developing in ADPLD manifest a clinical phenotype indistinguishable from that of ADPKD. The two known disease causative genes for ADPLD are PRKCSH and SEC63, which encode for non-cilial, ER-localized proteins involved in protein biogenesis and post-translational modifications [92–94]. PRKCSH, also known as hepatocystin, encodes the non-catalytic β-subunit of glucosidase II (GIIβ). Glucosidase II acts in the calnexin-calreticulin cycle of folding and quality control for integral membrane and secreted proteins passing through the ER membrane [93,95]. The SEC63 gene product, SEC63p, is a 83-kDa protein that works in concert with the SEC61/62 ER translocon and BiP to facilitate the co-translational transport process across the endoplasmic reticulum (ER) membrane of nascent polypeptides destined to become either secreted or membrane-inserted proteins [92,96]. Mouse models based on tissue selective and inducible expression of Cre have been generated as orthologous gene models for human ADPLD. Induction of PDL genes inactivation in kidney and liver in adult Prkcshflox/flox; pCX-CreER mice (Prkcsh-KO) and Sec63flox/flox; pCX-CreER mice (Sec63-KO) results in the formation of either bile duct or kidney tubule cysts (unlike the human condition where no renal phenotype has been reported) [97]. Prkcsh- and Sec63-KO mice have been used to unravel the role of the integral ciliary proteins involved in the pathogenesis of ADPKD (i.e. PC1 and PC2) and autosomal recessive polycystic kidney disease (ARPKD) (i.e. PKHD1, see below) [97]. Studies from Fedeles and coll. Showed that loss of either GIIβ or Sec63p results in reduced levels of functional PC1-PC2 complex, with PC1 acting as the rate-limiting factor that halts the severity of the polycystic phenotype [98]. Furthermore, PC1 levels can be increased following proteasome inhibition leading to significant amelioration (as increase in apoptosis and decrease in proliferation of the cyst-lining epithelia) of the cystic phenotype of a model of ADPLD with Prkcsh inactivation [97,98].

7.3. Pkhd1^del4/del4 mice and PCK rat

ARPKD, Caroli’s disease and congenital hepatic fibrosis (CHF) are rare diseases of the renal tubular and bile duct epithelia, caused by mutations in the PKHD1 gene [99,100]. PKHD1 encodes for fibrocystin (FPC), also known as polyductin, a plasma membrane protein located in the primary cilium that has a morphogenetic role in tubulogenesis and in maintaining the architecture of the epithelial duct lumen (‘planar cell polarity’) [100,101]. An experimental CHF mouse model has been generated by disrupting the exon 4 on Pkhd1 (Pkhd1^del4/del4 mice) [102]. Pkhd1^del4/del4 mouse develops intrahepatic bile duct dysgenesia (‘biliary microhamartoma’), which progressively evolves to a cyst-like configuration, but no renal involvement. Dysgenetic bile ducts can be observed as early as 2 weeks after birth, and their progression to cysts is paralleled by a progressive peribiliary fibrogenesis, detectable after 3 months, whilst hepatic function is normally preserved for up to 12 months. The presence of splenomegaly in more than 50% of the mice already at 6 months of age indicates clinically relevant portal hypertension, a feature mimicking the human disease behaviour, where portal hypertension represents the main determinant of the disease complications [102]. This model has been suitable to uncover some crucial aspects of biliary fibrogenesis in CHF. FPC-defective cholangiocytes displayed a hectic pro-inflammatory phenotype with abundant chemokine secretion (CXCL1, CXCL10, CXCL12), dependent upon an over-activation of the β-catenin signaling, enabling them to recruit macrophages and to respond to macrophage-derived cytokines (TNF-α, TGF-β) by up-regulating αvβ6 integrin expression. αvβ6 integrin is the local activator of latent TGF-β1, the most potent fibrogenic mediator in the liver [103,104]. Biliary fibrosis sped up when portal myofibroblasts were also recruited in the portal infiltrate, and this phase associated with development of portal hypertension. Accordingly, inhibition of macrophage infiltration by clodronate was relevant to impede progression of portal fibrosis [103].

Other rodent models have been fundamental in understanding the cellular alterations in cystic epithelia and in evaluating treatment efficacy. Along with the Pkhd1^del4/del4 mouse, the PCK rat carrying a spontaneous splicing mutation in the Pck gene (PKHD1 in human), represents a well-recognized animal model of ARPKD [105,106]. Sanzen and coll. First described the liver phenotype in the PCK rats up to 4 months of age, characterized by progressive liver enlargement due to multiple saccular and segmental dilatations of the intrahepatic bile ducts. Further experimental studies on PCK rats highlighted the cyst disconnection from the biliary system with advancing age, consistent with the human disease features [107]. However, in contrast with the Pkhd1^del4/del4 mouse, the PCK rat shows some critical differences, including a more pronounced development of the hepatic cysts, the renal involvement with cyst lesions affecting the outer medullary-collecting ducts, and the mild degree of portal fibrosis, without formation of fibrous septa or development of portal hypertension [108,109]. Overall, these characteristics make the PCK rat a less coherent model of the human CHF, at least with respect to the liver phenotype.

8. Conclusions

Generation of experimental animal models has provided fundamental clues to unveil many pathophysiological mechanisms triggered by biliary and cholestatic injury as well as to validate their clinical correlates. In the last decades, the main steps forward in modeling choletastic liver diseases have been reached with the help of mouse models of intrahepatic cholestasis (endotoxin-induced and drug-induced cholestasis) and extrahepatic biliary obstruction (BDL) [110–112]. However, although extensively characterized, these animal models have left several gaps in knowledge, given the difficulties to target selectively the biliary epithelium without the interference of additional (confounding) factors and to reproduce the smoldering clinical course peculiar of these conditions. On the other hand, several morphogenetic pathways whereby cholangiocytes regulate liver repair are increasingly recognized [62,113,114]. From this perspective, congenital cholangiopathies may represent a valuable asset to understand how a single genetic defect may affect specific cholangiocyte functions essential to mount a correct response to damage. Indeed, thanks to a variety of genetically engineered mouse models, different phases of biliary injury, starting from the earliest stages, can be probed with high accuracy, then followed through their progression, and eventually targeted by novel therapeutic approaches. In humans, these tasks cannot be usually performed given the long course of the disease and the limited possibility to realize sequential studies by collecting samples at different time points. Moreover, when human tissue is obtained from liver explants at the time of transplant, histological alterations are too advanced and generally less specific as dominated by the extensive liver damage. Given the broad role played by cholangiocytes in liver repair, these experimental models might provide in the near future a wealth of information extending well beyond the boundaries of these rare diseases.