Cholangiocyte pathobiology

Abstract

Cholangiocytes, the epithelial cells lining the intrahepatic and extrahepatic bile ducts, are highly specialized cells residing in a complex anatomic niche where they participate in bile production and homeostasis. Cholangiocytes are damaged in a variety of human diseases termed cholangiopathies, often causing advanced liver failure. The regulation of cholangiocyte transport properties is increasingly understood, as is their anatomical and functional heterogeneity along the biliary tract. Furthermore, cholangiocytes are pivotal in liver regeneration, especially when hepatocyte regeneration is compromised. The role of cholangiocytes in innate and adaptive immune responses, a critical subject relevant to immune-mediated cholangiopathies, is also emerging. Finally, reactive ductular cells are present in many cholestatic and other liver diseases. In chronic disease states, this repair response contributes to liver inflammation, fibrosis and carcinogenesis and is a subject of intense investigation. This Review highlights advances in cholangiocyte research, especially their role in development and liver regeneration, their functional and biochemical heterogeneity, their activation and involvement in inflammation and fibrosis and their engagement with the immune system. We aim to focus further attention on cholangiocyte pathobiology and the search for new disease-modifying therapies targeting the cholangiopathies.

Cholangiocytes line a complex network of interconnecting tubes extending from the Canals of Hering in the liver to the duodenum (FIG. 1). In humans, the total length of this network is estimated to be ~1.25 miles (2 km)¹. As with other epithelial cells, cholangiocytes are polarized with distinct apical and basolateral plasma membrane domains and multiple transport functions, many relevant to bile formation. Although cholangiocytes comprise a minority cell population in the liver, they are critical in bile generation, a life-sustaining function of the liver². Bile is a secretory fluid product of the hepatobiliary system containing a variety of components, including bile acids, electrolytes, lipids, proteins and endobiotic and xenobiotic compounds. These factors contribute to health by aiding digestion, maintaining the enterohepatic circulation and helping to eliminate unwanted compounds from the body. The continuous and extensive network of these cells within and outside the liver results in considerable heterogeneity in cholangiocyte function along the biliary tract. The blood supply to cholangiocytes originates from the hepatic artery and forms a peribiliary plexus (PBP) consisting of a 3D network of blood vessels of homogeneous diameter surrounding bile ducts³. The intimate anatomic association of the PBP with cholangiocytes enables crosstalk that probably both helps regulate normal cholangiocyte function and is associated with cholangiocyte malfunction in disease⁴. Under healthy circumstances, cholangiocytes have major physiological functions: bile is modified within the ductal lumen via activities at their apical plasma membrane domain; they form a barrier to potentially damaging molecules and microorganisms in bile via their tight junctions and immunoglobulin A (IgA) secretion; and they enable access to the immune and vascular systems via their basolateral plasma membrane domain. These complex processes are regulated by extracellular signals (for example, peptides, nucleotides, hormones and neurotransmitters), biliary constituents (such as bile acids, glucose and vesicles) and physical forces (including flow and pressure) that are reflected in various intracellular pathways, relying mostly on cAMP and Ca²⁺ signalling as second messengers.

Fig. 1. Ductal bile formation.

Bile produced by hepatocytes (primary or hepatic bile) is delivered into bile ducts. The Canals of Hering provide the continuum between the hepatocyte canaliculus and the ductules or cholangioles, the small bile ducts and the large bile ducts in which hepatic bile is modified to become ductal bile. A hepatic progenitor cell (HPC) niche is also thought to reside at the interface of the cells lining the Canals of Hering and the hepatocyte plate. Active biliary epithelial transport of electrolytes and solutes occurs in small and large bile ducts and determines the vectorial water movement (that is, absorption or secretion) across cholangiocytes, thus altering ductal bile composition and flow. Adapted with permission from REF.⁶, Elsevier.

Cholangiocytes are affected during liver injury and participate in the pathobiology of various liver disease states. Although ample evidence demonstrates that hepatocytes, the predominant epithelial cell type in the liver, regenerate, cholangiocytes can also contribute to liver regeneration when hepatocyte regeneration is impaired⁵. Genetic, infectious, immune-mediated, idiopathic, malignant and vascular diseases can also directly perturb cholangiocyte structure and function, resulting in impaired bile formation (cholestasis), followed by inflammation, fibrosis and liver dysfunction. Although the inflammation and fibrosis might initially be limited to the biliary tract, over time, portal fibrosis worsens and can culminate in hepatic dysfunction, cirrhosis and chronic liver failure. Collectively, these disease syndromes comprise the cholangiopathies, which are unsolved pathophysiological problems and important unmet needs in clinical hepatology⁶ (BOX 1). Although individually they are uncommon or even rare, as a group the cholangiopathies cause considerable morbidity and mortality, and curative therapies are not available; for example, primary sclerosing cholangitis (PSC) is the indication for 6% of liver transplants in the USA, with an approximate yearly cost of $125 million⁷.

Box 1 |. Selected cholangiopathies.

Genetic cholangiopathies

• Alagille syndrome

• ABCB4 deficiency

• Caroli syndrome

• Cystic fibrosis

• Polycystic liver disease (ADPLD, ADPKD and ARPKD)

Infectious cholangiopathies

• Cryptosporidium-associated cholangiopathy

• Recurrent pyogenic cholangitis

• Recurrent cholangitis in patients with a choledochoduodenostomy

Immune-mediated cholangiopathies

• Primary biliary cholangitis

• Primary sclerosing cholangitis

• IgG4-associated cholangitis

• Autoimmune cholangitis

• Graft versus host disease involving the liver

• Eosinophilic or mast cell cholangiopathy

Idiopathic cholangiopathies

• Biliary atresia

• Sarcoidosis

Malignant cholangiopathies

• Cholangiocarcinoma

Vascular cholangiopathies

• Hepatic artery thrombosis after liver transplantation

• Portal hypertensive biliopathy

ADPLD, autosomal dominant polycystic liver disease;

ADPKD, autosomal dominant polycystic kidney disease;

ARPKD, autosomal recessive polycystic kidney disease;

IgG4, immunoglobulin G4.

Evolving information indicates that cholangiocytes are not necessarily innocent victims in these disease processes but might also initiate and/or actively participate in the cholangiopathies. For example, cholangiocytes participate in inflammation by secreting chemokines and cytokines and can directly modulate the biology of myofibroblasts, the cell type responsible for collagen deposition within the liver⁶. Thus, in this Review, we address cholangiocytes in development and liver regeneration, their functional and biochemical heterogeneity, their activation and involvement in inflammation and fibrosis and their engagement with the immune system.

Cholangiocyte heterogeneity

The 3D network of ducts inside and outside the liver provides a very large surface area along which cholangiocytes perform fundamental secretory and absorptive processes that regulate bile flow and composition in its transport to the duodenum^8–12. Throughout the biliary system, cholangiocytes exhibit morphological, biochemical and functional heterogeneity¹³. Immature cholangiocytes within the Canals of Hering, as well as from the intrahepatic and extrahepatic peribiliary glands, are poorly differentiated and are considered progenitor cells that participate in epithelium renewal and tissue regeneration. However, cholangiocytes progressively acquire a greater degree of differentiation along the biliary tree (from small to large bile ducts) in terms of cell polarity, expression of receptors and transporters and response to hormones^13–17. This differentiation might also in part be due to the differences in vascularization of the cholangiocytes along the biliary tree³. Intrahepatic and extrahepatic bile ducts are surrounded by a complex network of vessels and capillaries derived from the hepatic artery and vein. The oxygen tension and metabolite and substrate composition differ along the biliary tree, promoting the differentiated metabolic, secretory and absorptive features of cholangiocytes. Differentiated cholangiocytes have distinct basolateral and apical (luminal) plasma membranes, the latter containing microvilli that provide a fivefold increase in cell surface area^12,18,19 and a single primary cilium able to detect and transmit bile signals and regulate cell function^20,21. Although cholangiocytes represent a minor proportion of all liver cells (3–5%), they are responsible for up to 30% of total bile flow in humans, with the other 70% originating from hepatocyte canalicular secretion¹¹; however, their contribution to this process in rodents is probably less²². Bile flow regulation involves the action of multiple ion carriers (such as transporters, exchangers and channels) strategically distributed along the polarized structure of cholangiocytes, as well as various molecules (including hormones, neurotransmitters, peptides and nucleotides) that tightly regulate bile flow and composition through the interaction with intracellular signalling pathways and cascades (such as cAMP and Ca²⁺) (FIG. 2).

Fig. 2. Molecular mechanisms regulating biliary secretion and absorption.

Cholangiocytes regulate the flow, composition and pH of the primary bile generated at the canaliculi of hepatocytes through different mechanisms, including the absorption of bile acids (BAs), glucose and amino acids (step 1) and the secretion of bicarbonate (HCO₃ −) and water (step 2). Secretin (step 3) stimulates the apical insertion of intracellular vesicles containing anion exchange protein 2 (AE2), cystic fibrosis transmembrane conductance regulator (CFTR) and aquaporin 1 (AQP1), resulting in chloride secretion through CFTR that is exchanged with bicarbonate via AE2. This bicarbonate generates osmotic force for the movement of water via AQP1. Biliary bicarbonate secretion creates the biliary bicarbonate umbrella that protects cholangiocytes against the damaging effect of toxic protonated BAs (BAHs). Hormones such as bombesin and vasoactive intestinal peptide (VIP) stimulate biliary bicarbonate secretion, whereas somatostatin, gastrin and dopamine inhibit this process. Extracellular nucleotides and nucleosides, via P2Y receptors, and acetylcholine (ACh) also promote baseline and secretin-stimulated bicarbonate secretion, respectively (step 4). The cholangiocyte primary cilium acts as a (step 5) mechanosensor (via polycystin 1 (PC1)), (steps 6–8) chemosensor (via G protein-coupled bile acid receptor 1 (TGR5), P2Y purinoceptor 2 (P2Y2) and extracellular vesicle (EV)) and (step 9) osmosensor (via transient receptor potential channel vanilloid subfamily 4 (TrpV4)), detecting signals in bile and subsequently modifying cell biology and bile flow and composition. AC5, adenylyl cyclase type 5; InsP3, inositol 1,4,5-trisphosphate; PKA, protein kinase A; PKC, protein kinase C; TMEM16A, transmembrane protein 16F.

Bicarbonate secretion.

Biliary bicarbonate (HCO₃⁻) secretion is the central event of bile-salt-independent flow, which is stimulated by the gastrointestinal hormone secretin during the postprandial period^11,12,23. Interaction of secretin with its specific G protein-coupled receptor localized to the basolateral membrane of cholangiocytes induces elevated intracellular cAMP levels²⁴ and subsequent protein kinase A activation²⁵. This activation subsequently stimulates the trafficking of intracellular vesicles containing the Cl⁻ channel cystic fibrosis transmembrane conductance regulator (CFTR), the Cl⁻/HCO₃⁻ anion exchange protein 2 (AE2; also known as SLC4A2) and the water channel aquaporin 1 (AQP1) to the apical membrane²⁶. CFTR mediates the apical release of Cl⁻, which is exchanged with HCO₃⁻ by AE2 (REF.²³). Biliary bicarbonate secretion drives the movement of water through AQP1, resulting in the alkalization and fluidization of the bile²³ (FIG. 2). The biliary secretion of bicarbonate creates an alkaline barrier, the so-called ‘biliary bicarbonate umbrella’, that renders bile acids polar, de-protonated and membrane impermeable²⁷. Biliary bicarbonate neutralizes the acidic pH resulting from meal digestion in the duodenum and favours the absorption of nutrients¹². Secretin-stimulated biliary bicarbonate secretion is dependent on the maintenance of the bile acid pool²³ and is inhibited by hormones such as somatostatin^28,29, gastrin³⁰ and endothelin³¹. Of note, this choleretic mechanism can be impaired under certain pathological conditions, as in cholestatic disorders such as primary biliary cholangitis (PBC), an immune-mediated disease that results in damage to intrahepatic interlobular bile ducts. PBC is characterized by the downregulation of AE2 expression in cholangiocytes³², resulting in impaired secretin-stimulated biliary bicarbonate secretion^33,34. Downregulation of AE2 in PBC cholangiocytes probably has a major aetiopathogenic role, as its experimental downregulation in human cholangiocytes in vitro^35,36 or in mice in vivo^37–39 results in the development of multiple PBC-like features, such as periportal lymphocytic infiltrates and bile duct damage; increased serum IgM, IgG and hepatic alkaline phosphatase levels; and spontaneous development of specific anti-mitochondrial autoantibodies. The characteristic downregulation of AE2 in PBC cholangiocytes is linked to miR-506 overexpression^35,36. Additionally, CFTR mutations occurring in cystic fibrosis might also contribute to cholestasis in these patients⁴⁰.

The biliary tree is controlled by neurovegetative innervations. Several neurotransmitters regulate the baseline and/or secretin-stimulated biliary bicarbonate secretion. Gastrin-releasing peptide²⁹ and vasoactive intestinal peptide (VIP)⁴¹ stimulate baseline biliary bicarbonate secretion, whereas acetylcholine (ACh)^11,42 and the α₁-adrenergic agonist phenylephrine⁴³ stimulate secretin-dependent biliary bicarbonate secretion. By contrast, dopamine⁴⁴, α₂-adrenergic agonists⁴⁵ and GABA⁴⁶ all inhibit secretin-dependent biliary bicarbonate secretion (FIG. 2). Several factors present in bile are also able to influence biliary bicarbonate secretion. Extracellular nucleotides and nucleosides interact with P2Y receptors localized to the apical membrane of cholangiocytes and promote an increase in intracellular Ca²⁺ levels and subsequent Cl⁻ secretion through apical Ca²⁺-activated Cl⁻ channels (such as transmembrane protein 16F (TMEM16A; also known as ANO1)) that further promote bicarbonate secretion^11,47–49 (FIG. 2).

Cholangiocyte transport.

Cholangiocytes participate in the absorption of different molecules from bile, including bile acids, glucose, amino acids and ions. Bile acids (that is, steroid acids) exist as either a free acid or conjugated to taurine or glycine. Unconjugated bile salts secreted through the canalicular membrane of hepatocytes can be protonated and passively diffuse across the apical membrane of cholangiocytes. They can then be transported by the basolateral membrane into the PBP, from which they can return to hepatocytes via the cholehepatic shunt, an alternative mechanism to the enterohepatic circulation of bile acids^50,51. In rat cholangiocytes, conjugated bile acids might be absorbed through the apical sodium-dependent bile salt transporter (ASBT)⁵² and subsequently via the basolateral truncated ASBT (t-ASBT)⁵³, multidrug resistance protein 3 (MRP3)^54,55 and/or the organic solute transporters OSTα and OSTβ^12,56–59. Moreover, the vectorial transport of glucose through the apical Na⁺–glucose cotransporter 1 (SGLT1) and the basolateral glucose transporter 1 (GLUT1) provides an osmotic gradient that favours the reabsorption of water from bile⁶⁰. Amino acids resulting from the γ-glutamyltranspeptidase (γGT)-dependent degradation of glutathione (including glutamine, cysteine and glycine) promote canalicular bile-salt-independent bile flow. These molecules can then be reabsorbed by sodium-dependent and sodium-independent mechanisms that generate osmotic gradients that favour water absorption and glutathione resynthesis in hepatocytes^61,62. Additionally, cholangiocytes possess a variety of transporters able to bidirectionally transport various molecules, including organic and inorganic anions and cations as well as proteins (multidrug resistance proteins perform the majority of these transport functions)¹².

Role of the primary cilium.

The primary cilium of cholangiocytes is an antenna-like organelle containing a well-tuned system of receptors and channels able to detect signals in bile that subsequently regulate intracellular signalling mechanisms, ultimately modifying bile flow and/or composition^19,63. This non-motile protuberance functions as a mechanosensor (via polycystin 1 and 2 (PC1 and PC2, respectively))⁶⁴, a chemosensor (via P2Y12 and G protein-coupled bile acid receptor 1 (TGR5; also known as GPBAR1))^65,66 and an osmosensor (via transient receptor potential channel vanilloid subfamily 4 (TrpV4))⁶⁷. Of note, extracellular vesicles (EVs) found in bile can also act as chemosignals that interact with the primary cilium and regulate cell biology⁶⁸. For example, EVs binding to cilia have been shown to inhibit cholangiocyte proliferation, promoting the quiescent status of the biliary system in normal conditions⁶⁸.

Cholangiocytes in development and liver repair

The embryological origin of the liver has been studied for many decades; thus, the developmental events that regulate liver organogenesis are becoming well-understood⁶⁹. However, although the regenerative capacity of hepatocytes has been studied equally as long⁷⁰, the molecular mechanisms governing biliary regeneration have been more clearly defined in the past decade with the advent of lineage-tracing techniques. Importantly, it seems that many of the key developmental pathways governing liver development become active during the processes of biliary regeneration and/or repair⁷¹. This dual role suggests that a better understanding of the developmental biology of the biliary tree might provide insights into therapeutic targeting of these processes during biliary disease. The biliary tree has played a central part in our efforts at understanding liver regeneration as it shares developmental origins with parenchymal hepatocytes, harbours a niche of stem or progenitor-like cells and is activated and expanded in the context of many liver pathologies. Thus, in this section, we briefly outline what is known about human biliary development and how the biliary system is altered during regeneration and repair.

Biliary development.

Around embryonic day (E) 13, bipotent hepatoblasts begin to differentiate towards mature hepatic epithelial cells (hepatocytes or cholangiocytes)⁶⁹. By E15.5, hepatoblasts nearest the portal mesenchyme become cholangiocyte-like and coalesce to form the ductal plate that then gives rise to primitive ductal structures and ultimately the intrahepatic bile ducts⁷². The remaining parenchymal hepatoblasts, which lie further from the influence of the portal mesenchyme, differentiate towards hepatocytes. The adoption of the biliary phenotype is orchestrated through spatiotemporal gradients of Notch⁷³, Wnt⁷⁴, transforming growth factor-β (TGFβ)⁷⁵ and FGF signalling⁷⁶, which arise from endothelial cells and/or mesenchymal cells within the portal tract (FIG. 3a). By contrast, differentiation of hepatoblasts towards mature hepatocytes relies on factors such as oncostatin M generated by haematopoietic cells expanding in the fetal liver⁷⁷, endothelial cell-derived hepatocyte growth factor (HGF)⁷⁸ and tumour necrosis factor (TNF) generated by Kupffer cells^79,80. Jagged 1, a cell surface Notch ligand of principal importance in bile duct development, exerts effects on both differentiation and tubulogenesis⁷² during cell–cell contact between periportal hepatoblasts and the portal mesenchyme. Indeed, Jagged 1 or Notch 2 mutations lead to the biliary abnormalities seen in patients with Alagille syndrome⁸¹, a liver disease associated with ductopenia of intrahepatic interlobular bile ducts.

Fig. 3. Potential sources of cholangiocytes in development and liver regeneration.

a | Cholangiocytes develop via differentiation from hepatoblasts in response to developmental morphogens. b | Cholangiocyte homeostasis is based on self-replication of pre-existing mature cholangiocytes. c | Cholangiocyte regeneration occurs through an accelerated replication in response to regenerative hormones, growth factors and cytokines. d | Cholangiocyte differentiation from hepatic progenitor cells can occur during biliary injury and repair after reactivation of developmental pathways.

Notably, the extrahepatic bile ducts (consisting of the common bile duct, cystic duct, gallbladder and hepatic ducts) have an embryological origin distinct from that of the intrahepatic bile ducts that develops in concert with the ventral pancreatic ductal system and only later anastomoses with the intrahepatic system. At E8.5, a subset of pancreatic endodermal cells that express both SOX17 and PDX1 represent a pancreatobiliary progenitor compartment that ultimately gives rise to the extrahepatic system. SOX17 seems to be critical in this process and works in concert with several other mediators, including HES1, HNF6 (encoded by ONECUT1), HNF1β and homeobox protein HEX (HHEX)^82,83.

Liver regeneration.

In addition to its role in liver development, the biliary system has also been central in discussions related to liver regeneration and repair⁸⁴, partly owing to the recognition that the biliary tree might harbour hepatic progenitor cells (HPCs) in the terminal ductules and the Canals of Hering⁸⁵. In general, however, it is thought that maintenance of the normal liver simply requires occasional self-replication of existing adult epithelial cells (hepatocytes and cholangiocytes) through mitosis as opposed to HPC differentiation⁸⁶. Through this mechanism, a very slow normal rate of liver cell turnover is able to counterbalance occasional apoptotic events to achieve a homeostatic equilibrium (FIG. 3b).

Following partial hepatectomy, a robust liver regeneration programme is activated to accelerate liver cell turnover and restore the lost mass and function. This phenomenon has been recognized for millennia, but the molecular mechanisms have been dissected in more considerable detail only over the past few decades⁸⁷. Our understanding of liver regeneration has centred primarily on hepatocyte biology, but clearly the biliary system must also be regenerated. The presence of multiple temporal waves of DNA synthesis during the regeneration process suggests that other liver cell types, such as cholangiocytes and non-parenchymal cells, undergo a similar type of accelerated replication after resection to hepatocytes⁸⁸. Although the bipotent liver stem cell compartment is known to reside in the periportal region, there does not seem to be substantial activation of these cells in the context of normal liver regeneration^82,86. This observation suggests that normal liver regeneration is achieved primarily through hypertrophy and proliferation of mature liver cells (FIG. 3c), as opposed to expansion and maturation of the biliary compartment of HPCs. However, HPCs are more prominently expanded in the context of liver injury and repair. The existence of these specialized stem-like cells naturally led to the historical view that HPCs can expand and differentiate into both hepatocytes and cholangiocytes when the normal mechanisms of liver regeneration are overwhelmed⁸⁹ (FIG. 3d). These small oval-shaped cells (termed oval cells in rodents) have scant cytoplasm and the ability to differentiate into both hepatocytes and cholangiocytes when isolated in vitro. More recently, additional stem cell niches have been described along the larger bile ducts in the peribiliary glands⁹⁰. Furthermore, most forms of chronic liver disease in humans (especially biliary diseases) and several mouse models of liver disease (especially 3,5-diethoxycarbonyl-1,4-dihidro-collidine (DDC) feeding and the choline-deficient, ethanolamine supplemented diet) are associated with a robust expansion of HPCs⁹¹.

Cholangiocyte proliferation is regulated through complex mechanisms involving the effects of various autocrine and paracrine factors. These molecules include, but are not limited to, growth factors (for example, TGF and TNF), cytokines (such as IL-6), neuropeptides (such as ACh) and hormones (for example, testosterone and oestrogen)⁹². Interestingly, cholangiocytes contain receptors for both male and female sex hormones, which have been shown to promote cholangiocyte proliferation. Oestrogens seem to prevent cholangiocyte apoptosis while also potentiating secretory and proliferative pathways^93,94. Likewise, progesterone binds to specific progesterone receptors on cholangiocytes to increase the biliary mass, and anti-progesterone therapy prevents the cholangiocyte growth caused by bile duct ligation⁹⁵. The role of testosterone was highlighted by work showing that castration or anti-testosterone therapy decreases intrahepatic bile duct mass, reduces secretin-stimulated cAMP levels and blocks ductal secretion in bile-duct-ligated rats⁹⁶.

Notably, many of the developmental morphogens that regulate liver organogenesis also seem to regulate cell fate decisions in the adult organ (for example, Wnt, SHH and Notch)⁹⁷. Many genetic lineage-tracing studies have attempted to reconcile the origin and fate of HPCs with mixed results^98–108. Currently, the prevailing view is that most parenchymal hepatocyte regeneration is hepatocyte-derived and that HPCs themselves might be derived from de-differentiated hepatocytes^109,110, despite their apparent biliary phenotype. However, studies in zebrafish^111,112 and in mice^102,113 demonstrate that biliary-derived cells can expand and differentiate into parenchymal hepatocytes when mature hepatocytes cannot proliferate or are heavily damaged. Similarly, transdifferentiation of hepatocytes can also form biliary structures in mice with developmental disruption of Notch signalling¹¹⁴. Definitive resolution of any apparent discrepancies between these findings is difficult as none of the animal models precisely recapitulate human disease and because Cre–lox-based lineage tracing, although powerful, has some technical limitations¹¹⁵. Overall, it is clear that the liver is unique in its ability to respond to diverse insults, partially as a result of the profound cellular plasticity that it displays in various contexts¹¹⁶. The remaining challenges will be to reconcile which processes are active in various pathological situations in humans and to devise logical therapeutic interventions on the basis of this underlying pathobiology.

Cholangiocytes in inflammation and fibrosis

Cholangiocytes can be activated by a variety of insults, including infections, cholestasis, ischaemia and xenobiotics^{117, 118}. In many cholangiopathies, PSC and PBC included, the activating insult is unknown. Features that characterize the activated cholangiocyte include increased proliferation and pro-fibrotic and pro-inflammatory secretions^119,120. Activated cholangiocytes are also involved in recruitment and crosstalk with immune, vascular and mesenchymal cells, and upon chronic activation the development of biliary fibrosis and cholangiocarcinoma^121,122. The broad changes in protein expression and the activated cholangiocyte secretome make the cholangiocyte an active participant in ongoing immunological reactions to biliary injury through pleiotropic autocrine and paracrine mechanisms¹²³. Furthermore, secondary effects of cholestasis on immune cell function through nuclear receptor signalling (for example, via the FXR pathway and/or the vitamin D receptor) further mould this microenvironment¹²⁴.

Most cholangiopathies share similar pathophysiological mechanisms, including cholestasis, proliferation, apoptosis, inflammation, fibrogenesis and eventually carcinogenesis. At the heart of biliary repair is inflammation. Persistent biliary cell damage and malfunction cause an inflammatory reaction that fuels a pathological reparative reaction, with excessive deposition of scar tissue around the injured ducts and eventually biliary cirrhosis. This complex of inflammatory cells (innate immune cells and T and B cells), mesenchymal cells and activated cholangiocytes is called ductular reaction. Activated cholangiocytes are able to participate in the inflammatory response by secreting chemokines, cytokines and angiogenic growth factors.

The biliary epithelium is exposed to cytokines and chemokines secreted by innate and adaptive immune cells in response to danger-associated molecular patterns (DAMPs), released by damaged liver cells, and/or to pathogen-associated molecular patterns (PAMPs) that originate in the intestine or the bloodstream¹²⁵. In addition to infection and tissue injury¹²⁰, epithelial inflammatory reactions might also be stimulated by autonomous cell mechanisms¹²⁶. In these cases, attempts at restoring normal cell homeostasis sustain a chronic inflammatory response of low-magnitude ‘parainflammation’ (for example, an adaptive response to a persistent cell dysfunction, as shown, for example, for cholangiocytes with defective fibrocystin (congenital hepatic fibrosis)^127–129); if normal biliary homeostasis is not restored, the process becomes maladaptive and stimulates the deposition of scar tissue.

Thus, depending on the type (infectious, toxic, autoimmune or inflammatory) and duration (acute or chronic) of the damage, epithelial cells, inflammatory cells and mesenchymal cells are activated and their crosstalk is orchestrated by a variety of autocrine and paracrine signals mediated by chemokines, cytokines and angiogenetic factors. It is important to remember that what is being repaired is an epithelial wound^130–132 and that the repair process is driven by signals arising from loss of the homeostatic equilibrium in cholangiocytes. These signals are detected by innate inflammatory cells (such as macrophages and neutrophils) and by cells generating the scaffold (myofibroblasts and portal fibroblasts) and its vasculature (endothelial cells). Together, these cells and their corresponding signals comprise the biliary reparative complex, also called the ductular reaction.

All evidence suggests that inflammation is the primer of the reparative response and biliary fibrosis. Signals from the inflamed ducts activate liver mesenchymal cells and attract them to the bile ducts; in turn, cells in the bile ducts respond to soluble factors or vesicles released by the activated mesenchymal cells. This epithelial– mesenchymal crosstalk requires the complementary expression of an array of agonists and their receptors by epithelial and mesenchymal cells. TGFβ1 and TGFβ2 (REFS^133–135), IL-6 and PDGFB¹³⁴ and CC-motif chemo-kine 2 (CCL2) (REF.¹³⁵) are some of the most well-studied factors secreted by reactive ductular cells (RDCs) that stimulate myofibroblast activation.

The epithelial component of ductular reaction — RDCs^136,137 — displays a biliary phenotype and is organized into irregularly shaped structures. These richly anastomosed structures, usually without a recognizable lumen, are located at the periphery of the portal space, eventually extending into the lobule, following a portal-to-portal pattern. RDCs are a distinct cell population from HPCs and possess reparative rather than regenerative functions, as indicated by several studies showing a strong correlation between the amount of RDCs and portal fibrosis^138–140. In addition to the mechanisms described above, RDCs might in part derive from cholangiocytes undergoing senescence, a cell reaction in which the cell is protected from apoptosis and carcinogenesis at the expense of a parainflammatory reaction, or from the ductular metaplasia of periportal hepatocytes, a phenomenon that has been clearly shown for intrahepatic cholangiocarcinomas^{130,136,137, 141}. Thus, RDCs are generated through multiple, highly adaptable mechanisms depending on the nature and intensity of biliary damage. Independently from their histogenesis, RDCs progressively accumulate as an effect of ongoing pathological repair¹³⁷. More than 20 years after Desmet suggested that RDCs are “the pacemaker of biliary fibrosis”¹⁴², several aspects of RDC biology remain elusive. RDCs possess different biological properties from normal cholangiocytes and can acquire a number of morphological and functional features of mesenchymal cells (FIG. 4). De novo expression of epithelial-to-mesenchymal transition (EMT) markers including S100A4, vimentin, Snail and matrix metalloproteinase 2 (MMP2), along with downregulation of the epithelial marker E-cadherin, was observed in RDCs in tissue sections obtained from patients with chronic cholangiopathies¹³¹. This ability of RDCs to express EMT markers and increase mobility is necessary to repair the wound. These features led to the concept of ‘partial EMT’ to underline the phenotypic plasticity of RDCs¹⁴³. Notably, the secretory profile of ductular cells is similar to the senescence-associated secretory response (SASP) of senescent cells. RDCs show heterogeneous expression of senescence markers such as p16 (REF.¹⁴⁴). This observation is functionally important as RDC senescence might be a driver of disease processes via the SASP mechanism. This ability to secrete pro-inflammatory cytokines typical of senescent cells might promote progression of the disease by amplifying the inflammatory and fibrotic responses¹⁴⁵; therefore, targeting of senescent RCD cells might be a viable therapeutic strategy in cholestatic liver diseases¹⁴⁶.

Fig. 4. Ductular reaction and ductular-reactive cells.

Extensive research has identified many of the morphogenetic mechanisms and molecules involved in the complex signalling and cellular crosstalk network of biliary repair. This crosstalk involves many cytokines, chemokines and signalling molecules. Most of these factors have both paracrine and autocrine effects and can act on multiple cell types. For example, vascular endothelial growth factor (VEGF) can autocrinally stimulate cholangiocyte proliferation, as the VEGF2 receptor is also expressed in cholangiocytes, but VEGF has paracrine effects on the endothelial cells (stimulation of neoangiogenesis) and stimulates mesenchymal cells. At the same time, IL-6 can stimulate cholangiocyte growth and the recruitment of neutrophils. The coexistence of reactive ductular cells and a rich mesenchymal and immune infiltrate constitutes the ductular reaction. The signals between the different infiltrating cell types are integrated into morphogenetic cues enabling cholangiocytes to re-create the biliary architecture owing to the re-expression of Wnt, Hedgehog and Notch signalling. CCL2, CC-chemokine ligand 2; CTGF, connective tissue growth factor; CXCL, CXC-chemokine ligands; DAMPs, damage-associated molecular patterns; EGF, epidermal growth factor; HGF, hepatocyte growth factor; IGF1, insulin-like growth factor 1; nitric oxide (NO); PAMPs, pathogen-associated molecular patterns; PDGFBB, platelet-derived growth factor B homodimer B; TGFβ2, transforming growth factor-β2; TNF, tumour necrosis factor.

An important requisite of biliary repair is the ability of RDCs to re-create the biliary architecture. This mechanism is mediated by morphogenetic pathways that are also involved in biliary development. Among them, Notch and the Yes-associated protein (YAP) and transcriptional co-activator with PDZ-binding motif (YAP– TAZ) pathway deserve special consideration because of their known role in maintaining biliary architecture during biliary repair^{140,147–151}.

Notch signalling is heavily involved in biliary repair, specifically in tubulogenesis and biliary transdifferentiation of hepatocytes¹⁴⁰. Direct cell–cell interaction between Notch-expressing HPCs and Jagged-1-expressing portal myofibroblasts induces the conversion of HPCs to RDCs. Notch signalling, particularly the Notch 2 receptor, is involved in the generation of branching tubular structures in bile duct repair^97,141,148. Although defective Notch signalling negatively affects biliary repair, persistent Notch overactivation might result in liver epithelial cell dysplasia and malignant transformation¹⁴⁹. Furthermore, activation of Notch in adult hepatocytes induces their transdifferentiation into biliary phenotype cells that express the biliary markers SOX9 and HNF1β¹⁴⁷.

Immunobiology of cholangiocytes

The liver and the bile ducts comprise complex immunological machinery, closely integrated with the mucosal immune system of the gut. Cholangiocytes contribute to homeostasis in this system through the secretion of IgA and various antimicrobial peptides (for example, β-defensin 2, lactoferrin and cathelicidin) into bile^152–155. As described earlier, cholangiocytes also participate in the response to injury and repair¹³². Driven by developments in the understanding of secretory functions of epithelia^156,157, research in the past few years has focused on activated cholangiocytes and their immunological functions^119,120 (FIG. 5).

Fig. 5. Key aspects of cholangiocyte immunobiology.

Quiescent cholangiocytes secrete antimicrobial molecules into bile (such as immunoglobulin A (IgA)) and express a range of innate immune receptors (for example, Toll-like receptors (TLRs) and NOD-like receptors (NLRs)) that recognize conserved pathogen-associated molecular patterns. The antigen-presenting capacities of cholangiocytes remain disputed regarding class II and T cell receptor (TCR) interactions, but CD1d and MR1 on cholangiocytes have been shown to effectively present lipid antigens and riboflavin derivatives to natural killer T (NKT) cells and mucosa-associated invariant T (MAIT) cells. Activated cholangiocytes engage in extensive paracrine crosstalk with cells of the immune system, including monocytes and macrophages, neutrophil granulocytes and T cells. Furthermore, autocrine signalling loops (for example, IL-6 signalling) provide further stimulation to augment and modify the activated cholangiocyte phenotype. CCL, CC-chemokine ligand; CCR, CC-chemokine receptor; CXCL, CXC-chemokine ligand; CXC-chemokine receptor; ICAM1, intercellular adhesion molecule 1; iTCR, invariant T cell receptor; LPS, lipopolysaccharide; MMP9, matrix metallopeptidase 9; PD-L1, programmed cell death 1 ligand 1; T_H17 cell, T helper type 17 cell; TNF, tumour necrosis factor; T_reg cell, regulatory T cell; VCAM1, vascular cell adhesion molecule 1.

Immunobiology of quiescent cholangiocytes.

Luminal cholangiocyte secretion comprises IgA and a broad range of other proteins with potential antimicrobial and immunological functions¹⁵⁸. In rodents, hepatocytes effectively transport secretory IgA into bile¹⁵⁷, whereas in humans hepatocytes do not express the secretory component and biliary IgA secretion is performed by cholangiocytes¹⁵⁹. Immunoglobulins are the second-most abundant protein fraction in human bile after albumin (which also harbours immune function properties)¹⁶⁰. The biliary immunoglobulins, secretory IgA in particular, contribute to the local antimicrobial defence systems in the bile ducts and upper intestine and might be involved in the clearance of systemic antigens^161,162. Alterations in the hepatobiliary IgA system are observed in chronic liver diseases, in particular chronic alcoholic liver disease¹⁶³. Other luminal secretions from cholangiocytes (such as defensins, mucins, lactoferrin and cathelicidin) contribute to the basic antimicrobial defence systems of bile^152,153,158 and are typically upregulated during infections.

Cholangiocytes constitutively express Toll-like receptors (TLRs) that respond to conserved PAMPs (for example, lipopolysaccharide (LPS) binds TLR4)^164,165. In quiescent cholangiocytes, TLR expression is most pronounced at the luminal membrane¹⁶⁴, and biliary infections lead to upregulation of relevant TLRs^{164,166–168}. TLR signalling pathways (such as NF-κB signalling) have important roles in cholangiocyte activation under these conditions¹¹⁷. Activation of TLRs has also been implicated in a variety of other biliary disease states, from cystic fibrosis and biliary atresia^169,170 (a paediatric liver disease of unknown aetiology characterized by loss of intrahepatic and extrahepatic bile ducts) to PSC and PBC^171–173. In PSC, a predominant hypothesis for disease development has been gut leakage of LPS and other bacterial products into the portal circulation due to concomitant IBD^174,175. Although there is some evidence of TLR activation occurring in PSC^173,176,177, this might also occur as a secondary phenomenon such as bacterial colonization following endoscopic retrograde cholangiography. Cholangiocytes also express cytoplasmic pattern recognition receptors (for example, NOD-, LRR- and pyrin domain-containing 3 (NLRP3)), and the NLRP3 inflammasome has been suggested to have a similar role in cholestatic liver disease as _{the TLRs}^173,178,179.

A potential role of the cholangiocyte as a professional antigen-presenting cell has been much debated. As with other nucleated cells, cholangiocytes constitutively express HLA class I molecules¹⁸⁰. Upon activation, cholangiocytes also express HLA class II molecules^181,182 but lack the B7 family members B7.1 (CD80) and B7.2 (CD86) co-stimulatory molecules that enable interactions with CD28 and CTLA4 on T cells. This finding probably means that the HLA class II expression on activated cholangiocytes is an epiphenomenon, but alternative co-stimulatory or inefficient class II antigen presentations remain theoretical possibilities. More recently, however, it has been shown that cholangiocytes are capable of presenting antigens to unconventional T cells such as mucosa-associated invariant T (MAIT) cells and natural killer T (NKT) cells. These T cells are enriched in compartments proximal to the gut microflora, the intestine and the liver in particular¹⁸³, and their invariable T cell receptors react to antigens presented by pre-set HLA class I-like molecules. The target antigens for MAIT cells are bacterial B vitamins (riboflavin derivatives) presented on the MR1 molecule, whereas NKT cells react to lipids presented by the CD1d molecule. After exposure to bacteria, cholangiocytes activate MAIT cells, causing pro-inflammatory cytokine release¹⁸⁴. Similarly, cholangiocytes have been shown to activate NKT cells upon exposure to relevant lipids¹⁸⁵. Although both NKT and MAIT cells have been suggested to be involved in autoimmune liver diseases^186–189, it is not clear whether this involvement is driven by antigens presented by cholangiocytes.

Immunobiology of the activated cholangiocyte.

Cholangiocytes can be activated by a variety of insults, including infections, cholestasis, ischaemia and xenobiotics^117,118, although in most human cholangiopathies the insult is unclear. Activated cholangiocytes are characterized by increased proliferation and pro-fibrotic and pro-inflammatory secretions^119,120. In this context, cholangiocytes are an active participant in the ongoing immunological processes¹²³. Furthermore, secondary effects of cholestasis on immune cell function also occur¹²⁴.

Crosstalk between activated cholangiocytes and T cells has been explored primarily in the context of autoimmune liver diseases. The interaction involves cholangiocyte expression of relevant adhesion molecules (such as intercellular adhesion molecule 1 (ICAM1) and vascular cell adhesion molecule 1 (VCAM1))^181,190 and other contact-dependent mechanisms (for example, programmed cell death 1 ligand 1 (PD-L1))¹⁹¹, as well as bidirectional cytokine and chemokine communication (for example, activation of CXC-chemokine receptor 6 (CXCR6)-expressing T cells by cholangiocyte-derived CXC-chemokine ligand 16 (CXCL16))¹⁹⁰. The crosstalk recruits T cells to the sites of biliary injury (for example, by CCL20 released by cholangiocytes positioning CC-chemokine receptor 6 (CCR6)-expressing T cells)¹⁹² and modulates relevant T cell activity — for example, by inducing persistence of effector T cells at sites of injury¹⁹³. In PSC and PBC, in which T cell-mediated cholangiocyte destruction has been proposed to be involved in pathogenesis^194,195, the crosstalk between cholangiocytes and T cells might hold clues for therapeutic interventions. Crosstalk between activated cholangiocytes and macrophages is involved in the chemoattraction of monocytes (via signalling molecules such as CCL2, IL-1 and a broad variety of chemokines including CXCL10 and CXCL12)^119,196,197 and the regulation of macrophage effector functions. These functions include cytokine secretions (for example, TNF)¹⁸¹, amplification of apoptotic signalling to cholangiocytes¹⁹⁸ and the resulting signalling cascades that generate a pro-fibrotic peribiliary microenvironment that also involves hepatic stellate cells and portal myofibroblasts^135,199. The presence of proteins including CD14 in bile and immunohistochemistry support the involvement of macrophages in the biliary microenvironment in disease states such as PSC^200,201. IL-8 (REFS^202,203), probably derived largely from activated cholangiocytes, is another protein consistently found upregulated in the bile of patients with cholestatic liver disease^204–206. Although it serves as a potent chemoattractant for monocytes, IL-8 secretion could also be involved in shaping the strong neutrophil signature observed in biliary disease states such as PSC^200,207. Numerous proteins found in bile probably reflect neutrophil activation in the biliary microenvironment (for example, S100A8, S100A9, S100A12 and MMP9), albeit more so in PSC than other biliary diseases^200,207. Additional cytokines secreted by cholangiocytes (such as IL-6) are probably involved in autocrine signalling²⁰⁸, reinforcing the cholangiocyte–immune system crosstalk, promoting cholangiocyte proliferation and, over time, facilitating cholangiocarcinoma development during chronic biliary inflammation^209,210.

Future directions

Much remains to be understood regarding the mechanisms of cholangiocyte pathobiology. Future advances will need to both exploit new technical advances and address fundamental gaps in our knowledge. The field of cholangiocyte biology has been held back by the absence of a cholangiocyte-specific promoter to develop and examine the phenotypes of cholangiocyte gene modifications using Cre recombinase technology. One strategy to gain cell specificity for cholangiocyte Cre expression could use liver-specific promoters in transgenic animals to drive expression of a flippase (FLP) recombinase in hepatoblasts during embryonic development. The FLP recombinase would be expressed in the descendants of the hepatoblasts, including both hepatocytes and cholangiocytes. A second construct expressing a Cre recombinase driven by cytokeratin 19 (which is expressed only in cholangiocytes within the liver) and held in check by a FLP recognition target (FRT)-flanked stop codon would then be introduced into the animals. The FLP recombinase would excise the stop codon, permitting expression of Cre only in cholangiocytes. Although conceptually attractive, we are not aware of anyone establishing this mouse to date, and other approaches would be welcome and are encouraged.

Two validated and now established technical advances include the development of patient-derived induced pluripotent stem cells (iPSCs), which can be differentiated into cholangiocytes, and the generation of 3D organoids. Several groups have now reported on iPSCs differentiated into cholangiocytes and cholangiocyte organoids with considerable success and insight^211–223. For example, organoids have been used to identify a clonogenic subpopulation of mouse cholangiocytes and unique surface markers for this proliferative cell population²¹³. In addition, iPSC-derived cholangiocytes have been used to regenerate the extrahepatic bile duct in the mouse²²⁴. These new tools will allow further mechanistic studies to improve our understanding of the mechanisms of human cholangiocyte biology and generate more functional in vitro models.

A myriad of pertinent questions persist regarding cholangiocyte pathobiology (BOX 2). This list is certainly not exclusive but rather highlights many relevant knowledge gaps that can be addressed with current methodologies. We hope these questions will help guide future research priorities and emphasize to funding agencies the importance of these questions to human health. We look forward to the future answers to these and other questions.

Box 2 |. Key questions.

• How is cholangiocyte regeneration regulated?

• What cholangiocyte cell death processes occur and predominate?

• Does elimination of reactive ductular cells reverse liver injury?

• How do cholangiocytes interact with innate and adaptive immune cells?

• How do cholangiocytes interact with the intestinal and biliary microbiome?

• How do cholangiocytes modify and engage the biomatrix of the biliary system?

• What is the effect of changes in cholangiocyte-mediated alterations of bile duct permeability?

• What is the role of cholangiocytes in drug and xenobiotic metabolism and transport?

Conclusions

Despite comprising only ~5% of the cells in the liver, cholangiocytes are essential for health. Considerable information now exists regarding their highly complex and regulated transport functions and contributions to bile composition and flow. The role of cholangiocytes in liver regeneration has also become more sharply defined. Current scientific investigation is focused on their role in liver immunobiology, inflammation and fibrosis, a line of enquiry that is important in unravelling the pathogenic mechanisms causing the cholangiopathies. We encourage more work on these processes, which will hopefully result in better therapy for these diseases.

Key points.

• Cholangiocytes are epithelial cells lining the intrahepatic and extrahepatic bile ducts; they are heterogeneous in size and function and contribute to bile composition and flow by solute transport processes.

• Cholangiocytes contribute to liver regeneration, especially when hepatocyte regeneration is compromised, as is often the case in human chronic liver diseases.

• Cholangiocytes can become activated and participate in inflammation by secreting chemokines and cytokines and can also directly modulate the biology of myofibroblasts, the cell type responsible for collagen deposition within the liver.

• Cholangiocytes can become senescent and participate in the senescence-associated secretory phenotype, a cell fate also characterized by cytokine generation and release.

• Cholangiocytes participate in hepatic immunobiology, particularly by expressing Toll-like receptors (TLRs), contributing to immunoglobulin A (IgA) biology, and by cellular crosstalk with the innate and adaptive immune system.

• Cholangiocytes are damaged in a variety of human liver diseases termed the cholangiopathies, which are in need of optimized therapies and represent a current unmet need in clinical medicine.