Primary sclerosing cholangitis (PSC) is a chronic cholestatic liver disease of unknown etiology but lymphocytic portal tract infiltration is suggestive of an immune-mediated basis for this disease. Associations with inflammatory bowel disease (IBD) especially ulcerative colitis (UC), and with particular autoimmune diseases, as well as the genetic associations further suggest PSC may be an immune-mediated disease. The immunogenetics of PSC have been the subject of active research and several HLA and non-HLA associated genes have been implicated in the development of the disease. Lymphocytes derived from the inflamed gut may enter the liver via the enterohepatic circulation to cause hepatic disease. PSC may be triggered in genetically susceptible individuals by infections or toxins entering the portal circulation through a permeable colon and hence evoking an abnormal immune response.

Primary sclerosing cholangitis (PSC) is a chronic disease of the intra and/or extrahepatic bile ducts. It is characterized by a concentric obliterative fibrosis that leads to bile duct strictures (Figure 1). In many, this in turn progresses to biliary cirrhosis and hepatic failure. Approximately one third of patients will develop cholangiocarcinoma[1]. PSC is frequently associated with inflammatory bowel disease (IBD) usually ulcerative colitis (UC) and those with Crohn’s have disease predominantly affecting the colon. Approximately three quarters of the Northern European population with PSC have concomitant IBD particularly extensive UC[2]. 4.0%-7.5% of patients with UC have PSC[3].

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Cholangiogram showing beading and dilatation of the intra and extra hepatic bile ducts- the diagnostic features of primary sclerosing cholangitis (PSC).

The term “secondary sclerosing cholangitis (SSC)” is used for a disease with similar clinical features to PSC but where a direct causative agent for the pathological process is known. Such agents include choledocholithiasis with intraductal stones, surgical damage to bile ducts, ischaemia from hepatic artery occlusion, infections, and chemical agents such as drugs. Table 1 comprises a full list of possible causes of SSC with a section also showing the conditions which can mimic sclerosing cholangitis on cholangiography. There is little good data on the natural history of SSC and very little information regarding the immunological processes occurring during the progression of SSC is known although liver biopsies often show similar changes to those of PSC with ductopenia and patchy inflammation. The remainder of this chapter will concentrate on the etiopathogenesis of PSC.

The etiology and pathogenesis of PSC remain very poorly understood. The insidious onset of the disease makes the identification of an aetiological agent very unlikely. As the disease is associated with autoantibodies and HLA haplotypes as well as being closely related to IBD it would appear to be immune mediated. An autoimmune mediated destructive process is also suggested by lymphocytic infiltration into areas of portal damage.

PSC is not however a classical autoimmune disease, as it occurs with a 2:1 male predominance compared with the female predominance found in classical autoimmune diseases such as primary biliary cirrhosis (PBC) and autoimmune hepatitis (AIH). Moreover PSC does not have the characteristic response to immunosuppressive treatment as seen in classical autoimmune disease (Table 2).

Circumstantial evidence that PSC may be immune mediated comes from the independent association of PSC with a number of autoimmune diseases. 119 patients with PSC were studied by Saarinen et al[4]. Each PSC patient with IBD was matched to an IBD patient without PSC; 24% of the PSC patients had one or more autoimmune disorders outside the liver and colon compared with only 9% in the IBD group without PSC. Nine patients in the PSC group had 2 or more autoimmune diseases compared with only 2 in the IBD group. Diabetes mellitus and thyroid diseases were the most common in both groups. It is noteworthy that associated autoimmune disease did not seem to influence the outcome or clinical presentation of PSC[4].

Simultaneous or sequential occurrence of PSC and AIH has been described in both adult and pediatric populations[5]. The reported prevalence of this overlap syndrome is variable from 8%-54% and depends on the age of the study population, the type of scoring system used for diagnosis and the completeness of the analysed data.

In general, sclerosing cholangitis in children is characterized by more pronounced autoimmune features with a clinical overlap with AIH. This condition “autoimmune sclerosing cholangitis in childhood” has been addressed elsewhere in this issue (see Miele Verghani).

Atypical anti-neutrophil cytoplasmic antibodies (ANCA) are present in the serum of up to 88% patients with PSC (33%-88%)[5]. They are however not specific for PSC and are found in UC (60%-87%), and AIH (50%-96%)[6]. These ANCA are distinct from perinuclear-staining antineutrophil cytoplasmic antibody (p-ANCA) found in microscopic polyangiitis and cytoplasmic-staining antineutrophil cytoplasmic antibody c-ANCA in Wegener’s granulomatosis.

Immunoblotting showed reactivity in 92% of IBD or hepatobiliary disease patients with an atypical p-ANCA to a myeloid specific nuclear protein with a molecular mass of 50 kDa[7]. The target antigen in PSC for these atypical ANCA is probably a neutrophil nuclear envelope protein. viz tubulin-beta isotype 5[8]. Terjung and colleagues have suggested that the term p-ANNA is therefore more appropriate as the recognised antigen is not cytoplasmic but originating in the nuclear membrane[7].

The importance of these autoantibodies in the develo-pment of PSC is unknown. Titres of ANCA correlate with disease activity in the systemic vasculitides, whereas in contrast there is a poor correlation between ANCA and clinical parameters in PSC[9–11]. Titres of ANCA remain unchanged after a transplant in PSC and after a colectomy in UC. Current evidence suggests that they are unlikely to play a role in the pathogenesis of PSC.

A high proportion of non-specific autoantibodies in addition to p-ANNA are found in patients with PSC (Table 3). They are of unclear relevance and unhelpful in diagnosis. These include anti nuclear antibodies (20%-67%), antimitochondrial antibodies (< 10%) and antithyroperoxidase antibodies (7%-16%)[5]. Anti-cardiolipin antibodies were found in 66% of PSC patients compared to 4% controls by Angulo but no resultant associations with thrombotic disease were demonstrated[12].

Significantly more PSC patients have autoantibodies to surface antigens expressed on biliary epithelial cells (BEC) than patients with PBC, AIH or normal controls. These induce increased expression of CD44 on the BEC and increased production of IL-6 by BEC[13]. Anti-BEC autoantibodies may be both IgM and IgG. IL-6 induces BEC proliferation in vitro and suppresses BEC apoptosis, and it is increased in the bile in cholangitis and in the serum in cholangiocarcinoma. Persistent IL-6 production may be in part, responsible for the bile duct changes seen in PSC.

Antibodies to the baker’s yeast, Saccharomyces cerevisiae (ASCA) have been reported in IBD especially active Crohn’s disease. ASCA are not autoantibodies but there does seem to be some genetic predisposition to their presence. ASCA has also been seen in autoimmune liver disease including PSC but no conclusions can be drawn from their presence[14].

PSC is not attributable to one gene locus and is a non-Mendelian (complex) disorder. A number of associations have been made with HLA haplotypes as well as a number of other genes. There is controversy as to whether there is a primary susceptibility allele but PSC is probably acquired through inheriting a combination of genetic polymorphisms that act together to cause susceptibility to disease. The genetics of PSC is still the subject of active research.

There is a T cell predominant portal infiltrate in PSC although the relative proportions and importance of the CD4 and CD8 cells are not known. CD4 cells are seen more commonly in the portal tracts and CD8 cells predominate in areas of interface hepatitis[41]. The cell infiltrate may change as the disease progresses. These cells are functional and are likely to be involved in the pathogenesis of disease. In the peripheral circulation there does appear to be a fall in CD8 cells as the disease progresses. This only occurs late in disease so is unlikely to be significant in disease pathogenesis[41–44].

Bo and colleagues showed that cell proliferation and function of liver derived T lymphocytes is impaired in PSC patients compared with liver derived T cells obtained from normal controls or patients with other autoimmune liver diseases[45]. They believe this is due to exposure to high levels of TNF in vivo and this exposure may be chronic.

Previously relatively high levels of TNF-α have been seen in T cell lines from liver biopsies in patients with different stages of PSC while decreased levels were observed in PBC patients. Therefore increased levels of TNF-α are present in PSC patients whether the disease be early or late stage[4546].

The association between PSC and IBD led to Vierling’s hypothesis that colonic bacteria enter the portal circulation through a leaky mucosa in IBD thereby causing PSC (Figure 2)[55].

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Vierling’s hypothesis of the pathogenesis of primary sclerosing cholangitis[72].

Bacterial antigens may act as molecular mimics in genetically susceptible people and cause an immune reaction responsible for initiating PSC. The bacteria are able to get through gut walls made permeable by colitis or in theory by any infective episode of acute infective or inflammatory colitis. Chemokines and cytokines are then released from Kupffer cells in the liver attracting macrophages/monocytes, lymphocytes, activated neutrophils and fibroblasts to the portal tracts. Vierling further suggested that the concentric fibrosis resulting could cause atrophy of the BEC secondary to ischaemia. The bile duct loss would lead to progressive cholestasis, further fibrosis and secondary biliary cirrhosis. This does not explain however why there are fewer PSC patients with Crohn’s colitis as compared with UC and why there can be an associated stricturing of the pancreatic duct.

Portal bacteremia has been described in UC patients undergoing colectomy[56]. A study looking at explanted livers showed higher bacterial positivity rates in bile and bile ducts in PSC patients compared with PBC patients, and α-haemolytic streptococci accounted for 46% of the bacterial strains found. Bile duct cannulation at endoscopic retrograde cholangiopancreatography (ERCP) could have accounted for this bacterial presence[57]. The study went on therefore to compare patients with PSC who had undergone ERCP to those who had not, in order to evaluate the potential role of these bacteria in the etiopathogenesis of PSC. Positive cultures were obtained from 3 of the naive PSC patients and from 6 of the PSC patients with previous ERCP. α-haemolytic streptococci were again the commonest bacteria seen. As most naive PSC patients were found to have negative bacterial cultures this bacteria is unlikely to play a primary role in etiopathogenesis but may be involved in disease progression[58].

Recent molecular studies have shown an increased prevalence of H pylori and other non-gastric Helicobacter species in cholestatic liver diseases compared with healthy controls and noncholestatic liver disease. In PSC positivity was significantly but weakly associated with UC[59].

Ponsioen et al have suggested an association between PSC and previous Chlamydia infection after the finding of an increase in seroprevalence of Chlamydia anti-lipopolysaccharide (LPS) antibodies in PSC patients, although no Chlamydia antibodies were found in liver tissue and thus the significance is unclear[60].

Among animal models, none has yet have been developed showing all the features of PSC, although a rat model in which there is small bowel bacterial overgrowth has shown hepatic injury somewhat similar to that seen in human PSC[6162].

Abnormal accumulation of lipopolysaccharide (a bacterial endotoxin), presumably derived from portal blood, in the biliary epithelium has been shown in a rat model with a self-filling blind intestinal loop, and therefore may be involved in the pathogenesis of bile duct injury associated with intestinal injury[63]. Are these studies very persuasive?

PSC is strongly linked to IBD but it also runs a course independent from the bowel disease illustrated by the fact that the disease can develop many years after colectomy. Grant et al hypothesized that T lymphocytes generated in the gut during active inflammation persist as long-lived memory cells and undergo enterohepatic circulation and can then trigger an inflammatory response in the liver when activated by an appropriate stimulus. The nature of the stimulus remains unclear; possibilities include hepatic expression of the original priming antigen or possibly mediation solely by the aberrant expression of gut specific adhesion molecules and chemokines[64].

There is overlapping expression of many molecules between the gut and liver including the two potential addressins vascular adhesion protein-1 (VAP-1) and mucosal addressin cell adhesion molecule-1 (MAdCAM-1). VAP-1 expression on liver endothelium is normally far stronger than that seen on mucosal vessels. In IBD gut expression is greatly increased, suggesting that lymphocytes from the liver may be able to enter the inflamed gut using VAP-1. MAdCAM-1 endothelial expression was thought to be restricted to the gut but has been recently seen on portal endothelium in inflammatory liver disease (including PSC) associated with IBD. Mucosal lymphocytes express α₄β₇, which allows adhesion to hepatic MAdCAM-1 suggesting it may play a role in lymphocyte recruitment[65–67]. They propose that memory lymphocyte cells recirculate between liver and gut using either/both MAdCAM-1 and VAP-1[65].

The chemokine CCL21 activates lymphocyte adhesion to MAdCAM-1 dependent on α₄β_7. CCL21, thought to only exist in secondary lymphoid tissue is upregulated in portal associated lymphoid tissue in PSC and plays an important role in recruiting lymphocytes. Expression of the gut-associated chemokine CCL25 (thymus-expressed chemokine (TECK)) has also been shown in PSC liver sinusoidal endothelium but was absent in liver in AIH/PBC. A significant population of CCR9+ mucosal lymphocytes (capable of binding CCL25) has been detected infiltrating PSC liver tissue compared with controls and matched peripheral blood, thus supporting the hypothesis of a T cell enterohepatic recirculation. CCR9 lymphocytes co-express the gut homing integrin α₄β₇. Therefore CCL25 recruits CCR9+ lymphocytes to the liver in PSC by triggering adhesion to MAdCAM-1[6667]. MAdCAM-1 and CCL25 are upregulated to the liver in inflammatory liver diseases whereas previously they were thought to be restricted to the gut. Conversely VAP-1, normally expressed in the liver, is up regulated in the gut in IBD[68].

However this does not explain why PSC is associated more with UC than Crohn’s disease, as it would be predicted that just as many memory T cells are produced in Crohn’s disease as in UC.

Immune mechanisms play an important role in the pathogenesis of PSC, although it is remains unclear whether it is a classical autoimmune disease (Tables 2 and 6). There are strong MHC genetic associations including HLA molecules and the MIC molecules. HLA haplotypes however do not account for all the genetic susceptibility in the development of PSC and there is uncertainty about the importance of genes outside this region. Bacterial antigens may act as molecular mimics in hosts who are genetically susceptible and therefore cause an immune reaction leading to PSC initiation. Lymphocytes may move from the inflamed gut in IBD via the enterohepatic circulation and cause inflammation of the liver when activated by a specific stimulus such as bacterially derived antigens.