Although Virchow first described arteriosclerosis as an inflammatory disorder in 1885 it has taken a long time to become recognised in modern medicine as an active inflammatory disease rather than simply a problem of lipid storage in the vessel wall. In the following we present analogies and potential common pathogenetic mechanisms in the pathobiology of arteriosclerosis and PSC based on recent findings derived from animal models for both diseases 78910. Even when it is taken into account that the human body may have a limited repertoire of reactions to noxious and/or infectious stimuli, the similarities between PSC and arteriosclerosis in regard to their pathobiology are stunning and deserve further exploration.

All known critical molecular and cellular events underlying the pathogenesis of arteriosclerosis are apparently linked to alterations in lipid homeostasis 3. As such, the induction of endothelial vascular cell adhesion molecules (Table 1), currently believed to represent the major initiators in arteriosclerosis, is related to the accumulation of oxidized LDL (oxLDL) and phospholipids on the endothelial surface (Figure 1) 3. As such, oxidized lipoprotein particles (e.g. oxidized phospholipids and short-chain aldehydes arising from lipoprotein oxidation) can induce endothelial VCAM-1 expression via TNF-α- or IL1-β-induced nuclear factor-κB activation 4.

Since the liver is the major organ responsible for the production and degradation of apoB-100-containing lipoproteins, it is most likely that these molecules may also be engaged in the pathogeneses of liver diseases. Moreover, the fact that bile secretion is the most effective means of cholesterol elimination from the human body leads to the suggestion that altered biliary lipid secretion and/or abnormal lipid oxidation leading to oxidative stress may also be critically involved in the initiation and/or perpetuation of PSC. In parallel to arteriosclerosis, enhanced cholangiocellular VCAM-1 expression was recently demonstrated in multidrug resistance gene (Mdr2/Abcb4) knockout mice (Mdr2^-/-), a well characterized mouse model that closely mirrors the macroscopic and microscopic pathology of PSC 811. Potentially toxic bile acids synthesized from cholesterol may represent first-line candidates for the induction of cholangitis in Mdr2^-/- mice 12. Bile acids are normally packed into mixed micelles together with phospholipids and cholesterol to protect cholangiocytes from potentially toxic bile acids, which may otherwise cause necrosis or apoptosis of cholangiocytes 6. Absence of phospholipids as a consequence of Mdr2 defects not only results in unopposed bile acid toxicity but also in cholesterol supersaturated bile, which could facilitate oxidation. The critical role of biliary lipotoxicity in the sense of deranged and consequently toxic bile composition in the development of cholangitis was also underlined by the enhanced cholestatic phenotype in bile acid-fed Mdr2^-/- mice 13. Despite the apparent discrepancy between decreased biliary cholesterol concentration in Mdr2^-/- mice and increased serum cholesterol levels in arteriosclerosis, the common denominator may be abnormal cholesterol oxidation in both conditions. Alternatively, abnormal metabolic products of phosphatidylcholines (e.g. lysoPC) could also be excreted into bile in Mdr2^-/- mice or reach the cholangiocytes also from the perbiliary plexus or lymph vessels instead from the luminal side. Moreover, bile acids may be able to induce a reactive phenotype of cholangiocytes characterized by the production of several proinflammatory and profibrogenic cytokines and chemokines as well as their corresponding receptors (summarized in Table 1 and shown in Figure 2) 6. Conversely, some bile acids, their receptors, and synthetic agonists were recently shown to influence the degree of arteriosclerosis in several animal models 1415161718. This fits well with the rapidly progressing insights into the transcriptional regulation by nuclear receptors (e.g. FXR, LXR) of bile acid synthesis/transport on the one hand and lipid metabolism/transport on the other 171920. Finally, toxic bile acids can activate endothelial cells 21. Taken together, exposure to abnormal luminal lipid content and composition, ultimately resulting in lipid oxidation, may be critically involved in both conditions.

Endothelial and bile duct epithelial cells may not only be the victims/targets but also the culprits/effectors that actively participate in inflammatory processes in arteriosclerosis and PSC, respectively. The endothelium represents the first-line contact, where the accumulation of oxLDL induces the expression of adhesion molecules (e.g., VCAM-1) and proinflammatory cytokines/chemokines (e.g., MCP-1, see also Table 1) critically involved in the initiation and perpetuation of arteriosclerosis, making them potential therapeutic targets 4. Most interestingly, the cytokine/chemokine/receptor armamentarium of reactive endothelial cells and reactive cholangiocytes appears to be largely comparable, as summarized in Table 16.

Endothelial binding and penetration of monocytes and T lymphocytes in nascent atheromas by adhesion molecules is pivotal in the development of arteriosclerosis 4. Moreover, induction of endothelial-derived adhesion molecules reduces barrier function via alterations of the delicately regulated tight junction system, allowing oxidized lipids and marginalized/attached inflammatory cells to migrate into subendothelial regions of the vessel 22. The MCP-1/CCR2 interaction together with the action of M-CSF represents a critical step for migrated monocytes to differentiate into activated macrophages/foam cells, which themselves are able to produce numerous cytokines, reactive oxygen species (ROS), and growth factors (Table 1). Penetrating T lymphocytes further perpetuate this inflammatory response in a complex orchestrated interplay of cytokines and tissue factors (Table 1, Figure 1).

In analogy to arteriosclerosis, hydrophobic bile acids can induce MCP-1 and TNF-α expression to increase tight junction permeability of cholangiocytes in vitro 2324. In line with these findings, increased tight junction permeability was also demonstrated in Mdr2^-/- mice and human cholangiopathies, a mechanism which could permit regurgitation of bile (rich in bile acids, phospholipids, cholesterol, lipoprotein X, CD66a) from the bile duct lumen into the portal tract (Figure 2) 825. Leaking of potentially chemotactic/reactive cholephiles together with enhanced expression and secretion of adhesion molecules may then induce the migration of neutrophils and T lymphocytes from the peribiliary plexus out into the subepithelial space and into the bile duct epithelium, leading to cholangitis (Figure 2). Despite the fact that cytotoxic T cells were shown to induce cholangiocyte apoptosis in humans 26, their role in PSC is still a matter of debate since these cells are virtually absent in precirrhotic PSC stages. The hypothesis becomes questionable that cholangiocytes are the primary target of the immune system in PSC. The role of portal and/or penetrated neutrophils (probably leading to ROS formation and destruction of the basement membrane) in PSC also deserves more detailed studies because these cells may be critically involved in the activation of periductal myofibroblasts by aldehydic end-products of lipid peroxidation.

Differentiated monocytes/macrophages and T cells that find their way out through the leaky endothelial barrier exert their pro-atherogenic effects by the secretion of cytokines and growth factors driving smooth muscle cell (SMC) migration and proliferation as well as extracellular matrix (ECM) deposition (Table 1). The development of fibrous plaques results from proliferating SMCs, cholesterol ester and ECM components. This complex interplay is mediated by extensive crosstalk between T-cells and macrophages (CD40/CD40L interactions). Increased homocystein and angiotensin II levels in the vessel wall were also demonstrated to be important mediators of SMC proliferation as well as their activation. In addition, molecules such as osteopontin, CD44 and TGF-β were demonstrated to be critically involved in SMC activation. Reduction of arteriosclerosis in corresponding atherogenic diet-fed knock-out mice 272829 further underlines the importance of these factors for the development of arteriosclerosis.

A number of these central mediators of arteriosclerosis were also demonstrated to originate from activated cholangiocytes in cholangiopathies 303132 (Table 1 and Figure 2). Cholangiopathies are frequently accompanied by ductular proliferation/reaction, which is a potential stimulus for biliary fibrosis. The question whether ductular proliferates induce/stimulate proliferation and activation of periductal myofibroblasts and how this might happen (e.g., via stimulation with cytokines such as TGF-β or PDGF, brain-derived peptides such as NGF and BDNF, epithelial-mesenchymal transformation) is critical and still to be resolved.

Little is known about the potential importance of the extracellular matrix composition in the pathogenesis of both entities and how modifications (e.g. binding of oxLDL/phospholipids to proteoglycans, cleavage and signalling through proteinases, collagenases, gelatinases, resulting advanced glycosylated endproducts or transglutamination of collagen fibres) could influence the perpetuation of both diseases. In arteriosclerosis, diminished stability of the matrix scaffold is driven by MMPs, collagenases, and the recently identified ADAMTSs (a disintegrin-like metalloprotease with thrombospondin type-1 motifs) 33. These enzymes could either play a role in the resolution of vessel stenoses or also destabilize plaques 3435. In parallel, these enzymes could also be engaged in the chronic persistent wound-healing process in PSC. It is attractive to speculate that chronic activation of MMPs could further increase duct permeability. In addition, persistent activation of ADAMTSs could further increase the inflammatory response through their thrombospondin-1 motive (TSP-1) on TGF-β 3637 or in the sense of a self-perpetuating process by generating chemotactic/immunogenic ECM products.

The exact role of MDR3 variants (the human orthologue of rodent Mdr2) in the pathogenesis of PSC is still open to question. Two studies (with a total number of 80 PSC patients) found no differences in the number of genetic MDR3 variants between healthy individuals and PSC patients 3839. The findings of these two genetic studies are consistent with a normal biliary bile acid/PC ratio (indicative of normal MDR3 function) previously reported in 45 PSC patients 40. However, the number of PSC patients tested so far is still to small to definitely answer the question whether MDR3 defects may play a pathogenetic role and reductions in PC excretion may be transient and independent of genetic factors. Moreover, since PSC may represent a complex/polygenetic disease, a major role of MDR3 mutations in "unselected" PSC patients would be surprising. Rather, MDR3 genetic variations could play a role in selected PSC subgroups (e.g., small duct PSC, young women suffering from PSC, PSC patients with gallstones). Such "unusual" PSC patients could have been missed in previous studies despite the fact that cholelithiasis can be part of its clinical spectrum 41. Although the role of MDR3 variants in the pathogenesis of PSC is still controversial, Mdr2 knockout mice may still represent a worthwhile cholangiopathy model. The Mdr2 knockout mouse model is the first and so far only non-surgical animal model clearly showing the macroscopic (bile duct strictures and dilatations of the large bile ducts) and microscopic features (onion skin-type like pericholangitis and periductal fibrosis) of sclerosing cholangitis in humans 842. An additional advantage of this model system is the high reproducibility and low variance in the cholestatic phenotype of these animals making them also an ideal model for therapeutic interventional studies 43. Irrespective of the underlying cause of sclerosing cholangitis, the Mdr2^-/- model is an appropriate animal model for sclerosing cholangitis since it reliably reproduces the final common pathways of bile duct injury and (peri)biliary fibrosis. However, Mdr2^-/- mice do not show some of the (pheno)typical features observed in human PSC such as the association with cholangiocarcinoma and inflammatory bowel disease. finally, the Mdr2^-/- model may also have general validity as model for cholangiopathies resulting from bile toxicity with a deranged biliary phospholipid/bile salt secretion.

Despite the stunning similarities between PSC and arteriosclerosis in regard to their pathobiology as outlined above, there are also obvious differences between both entities such as the age of onset, different risk factors, and different co-factors (outlined in Table 1) which deserve attention. The exact definition and discrimination of similarities and differences between PSC and artiosclerosis should further help to understand the pathogenesis of both diseases.

(i)

(ii)

(iii)

(iv)