Pathogenesis of primary biliary cholangitis (primary biliary cirrhosis)

INTRODUCTION — Primary biliary cholangitis (PBC; previously referred to as primary biliary cirrhosis) is characterized by a T-lymphocyte-mediated attack on small intralobular bile ducts (picture 1A-B). A continuous assault on the bile duct epithelial cells leads to their gradual destruction and eventual disappearance (picture 2). The sustained loss of intralobular bile ducts causes the signs and symptoms of cholestasis, and eventually results in cirrhosis and liver failure [1-3]. (See "Clinical manifestations, diagnosis, and prognosis of primary biliary cholangitis".)

This topic will review the pathogenesis of PBC. The diagnosis and management of PBC are discussed separately. (See "Clinical manifestations, diagnosis, and prognosis of primary biliary cholangitis" and "Overview of the management of primary biliary cholangitis".)

CLUES ABOUT ETIOLOGY BASED ON THE EPIDEMIOLOGY OF PBC — The precise cause of PBC is unknown but, as with other autoimmune diseases, is related to genetic susceptibility and environmental factors [4-6]. A number of environmental causes have been implicated, including several bacteria, viruses, toxins, and drugs, which are described below [7-11]. Some of the most compelling evidence for an environmental factor has been derived from epidemiologic studies, which have demonstrated geographic clustering, clustering of cases across time, and seasonal variation in the diagnosis of PBC [9,10].

As examples, the incidence of PBC has been estimated to be 200 to 251 per million in South Wales and Northeast England, less than 25 per million in Canada and Australia, and almost zero in Sub-Saharan Africa and India [12-16]. A limitation of these studies is that the varying incidence rates may in part be due to genetic differences in the populations and to differences in study methodology. However, that an environmental factor may be involved was strongly implicated in a study from a well-defined geographical population cohort in Northern England [17] and around toxic waste sites in New York City [8]. Analysis of cases demonstrated unequivocal clustering in well-demarcated geographic regions.

The unanswered question in PBC is what causes or triggers the T-cell attack on small bile duct epithelial cells? Available data suggest that PBC is an autoimmune disease [18,19]. Like other better characterized autoimmune diseases, there appear to be at least two distinct requirements for PBC to develop: genetic susceptibility; and a triggering event that initiates the autoimmune attack on bile duct cells.

The prevalence of PBC in families with one affected member is estimated to be 100-fold greater than that in the general population [20,21]. However, the disorder is not inherited in any simple recessive or dominant pattern. Familial occurrences of the disease have included sisters [22], brothers [23], brothers and sisters [24], and parent and child [25]. In addition, unaffected family members are more likely than controls to have impaired T-cell regulation [26] and increased numbers of circulating autoantibodies [23]. However, there is not a significant increase in antimitochondrial antibody, the serologic marker of PBC, in healthy family members if purified recombinant mitochondrial autoantigens are used in the assay [27]. (See 'Antimitochondrial antibodies' below.)

Genetic susceptibility — There have been extensive genome-wide association studies (GWAS) in patients with PBC, and a number of associations have been documented [28]. However, while such associations are of considerable interest, they have not yet been translated into useful clinical testing [29-31]. There is a weak association between PBC and haplotype HLA-DR8 [32-34] and the HLA-DPB1 gene in some populations [35,36]. Accumulated data suggest that there may be an inherited abnormality of immune regulation, perhaps an inability to suppress an inflammatory attack on small bile ducts once it is initiated [32,35,37]. One study found an association of PBC with a variant in the CTLA4 gene, which encodes a co-inhibitory immunoreceptor involved in self-tolerance [38]. Another report found a significant association with genetic variants at the HLA class II, IL12A, and IL12RB2 loci (which have a role in interleukin-12 immunoregulatory signaling) [31]. The importance of IL12 has been underscored in animal models and in a case report of a child with congenital IL12 deficiency who developed PBC [39]. Other susceptibility loci that have been identified include SPIB (which encodes for a transcription factor involved in B-cell receptor signaling and T-cell lineage decisions), IRF5 (which encodes for interferon regulatory factor 5), TNPO3 (which encodes for transportin 3), and MMEL1 (which encodes for a metallo-endopeptidase) [40,41]. Interestingly, IRF5 and TNPO3 have also been implicated in the pathogenesis of systemic lupus erythematosus, systemic sclerosis and Sjogren syndrome, suggesting that there is a genetic overlap in PBC and these disorders. The concordance of PBC in identical twins is approximately 50 percent, somewhat higher than the concordance of autoimmunity in identical twins with lupus and rheumatoid arthritis [42,43].

Why the disease primarily affects females is unclear. A plausible theory relates to immunological tolerance, a component of which depends upon genes located on the X chromosome [44]. One study suggested that females with PBC were more likely to have monosomy of the X chromosome (a marker reflecting haploinsufficiency for specific X-linked genes) compared with healthy controls or those with hepatitis C [45].

Finally, there have been suggestions for an epigenetic role in the pathogenesis of PBC, (ie, methylation), and therefore modification of DNA [46].

MOLECULAR MIMICRY — Molecular mimicry refers to a phenomenon where antigens evoking an immune response have enough similarity to endogenous proteins to incite an autoimmune reaction. Such a mechanism has been widely proposed to initiate autoimmunity in patients with primary biliary cholangitis (PBC). Several candidate inciting agents (including bacteria, viruses and various chemicals) have been proposed, but none has been proven definitively.

Infections — An infectious etiology for triggering PBC has been suspected, although a specific etiologic agent has not been unequivocally identified. Several agents have been implicated; some have been disproven while others remain topics of investigation.

It is possible that infection and the presence of antimitochondrial antibodies (AMA) (see 'Antimitochondrial antibodies' below) may be pathogenetically linked. A study comparing peripheral blood mononuclear cells from patients with PBC with controls found that patients with PBC had a much higher number of inducible IgM-producing B-cells and that each B-cell produced a greater quantity of IgM protein compared with controls [47]. Induction of the B-cells was achieved by stimulation with unmethylated CpG DNA motifs (non-proteinaceous repeated sequences found frequently in bacteria). The authors hypothesized that patients with PBC are hyperresponsive with respect to IgM production and that the heightened IgM production may reflect how various microorganisms interact with the immune system in patients with PBC. Interestingly, there is a significant increase in AMA-specific plasmablasts in the blood of patients with PBC [48].

Retroviruses — Data from one laboratory suggested that infection with a retrovirus may be associated with PBC in some patients. However, these data have not been replicated [49] and have largely been disregarded.

Biochemical improvement from combination antiretroviral treatment was observed in an older open-label pilot study (also from the same group), a finding that has never been confirmed [50].

Propionibacterium acnes — A possible etiologic role for Propionibacterium acnes (P. acnes) was suggested in a study in which sequences of the organism were detected by PCR-amplifying granulomas from patients with PBC [51]. The significance of this observation remains uncertain but may relate to the appearance of granulomas in some patients with PBC.

Chlamydia pneumoniae — A pilot study reported that antigens to Chlamydia pneumoniae were detectable more often in the livers of patients with PBC compared with controls with other causes of liver disease, suggesting a possible etiologic role [52]. However, these findings were disputed in subsequent report using more detailed immunohistochemical analysis [53]. Furthermore, no benefit from treatment with tetracycline was observed in a pilot open-label study [54].

Escherichia coli — Antibodies reacting against the mitochondrial human pyruvate dehydrogenase complex crossreact with the E. coli pyruvate dehydrogenase complex implicating E. coli infection (particularly urinary tract infection) in the pathogenesis of PBC [55-59]. A problem with this theory is that antibodies against the E. coli are often in lower titers than against the human complex, and the E. coli antibodies are more frequent in patients with advanced disease (rather than being present in patients with early-stage disease, which would be more supportive of a role as an inciting agent). However, subsequent studies using different recombinant constructs of E. coli have suggested that infection with E. coli may be an early inciting event [60].

N. aromaticivorans — Novosphingobium aromaticivorans (N. aromaticivorans) is a ubiquitous gram-negative organism that metabolizes organic compounds and estrogens. Two of its proteins share homology with the E2 component of pyruvate dehydrogenase (PDC-E2), the major epitope recognized by antimitochondrial antibodies. A role for the bacteria in the pathogenesis of PBC was suggested in a pilot study in which 100 percent of 77 PBC patients (who were anti-PDC-E2 positive) had high antibody titers against the organism compared to none of 195 controls [61]. Reactivity to the organism was also found in another study [62]. These observations provide increasing evidence that exposure to N. aromaticivorans is associated with the development of PBC, the clinical implications of which are uncertain [63].

Lactobacillus — A case report described a patient who developed PBC after receiving intramuscular vaccination with a vaccine containing several lactobacillus species to treat recurrent vaginosis [64]. The patient had AMA-M2 antimitochondrial antibodies that reacted against the major PBC-specific mitochondrial auto-epitope (ie, pyruvate dehydrogenase complex E2) but reacted even more strongly against Lactobacillus delbrueckii. The authors speculated that the lactobacillus vaccination caused the development of PBC through molecular mimicry.

CHEMICAL EXPOSURE — Certain compounds (particularly halogenated hydrocarbons) are capable of inducing antibodies that have an affinity for the human pyruvate dehydrogenase complex (sometimes even having greater affinity for native mitochondrial antigens than for the halogenated hydrocarbon) [65,66]. In a rabbit model, one compound (bromohexanoate ester) induced high titers of antimitochondrial antibodies that were similar to those seen in humans with PBC [67]. However, the animals did not develop liver lesions similar to PBC. Such lesions did develop 18 months after immunization in a guinea pig model [68]. Additional data have led to the induction of histologic lesions that resemble PBC in mice immunized with a common component of cosmetics, 2-octynoic acid [69]. Residence near a toxic waste site was associated with PBC in one epidemiologic study [8].

ANTIMITOCHONDRIAL ANTIBODIES — Antimitochondrial antibodies are the serologic hallmark of PBC (see "Clinical manifestations, diagnosis, and prognosis of primary biliary cholangitis").

There are four principal autoantigens that are targets for antimitochondrial antibodies (AMA), which have collectively been referred to as the "M2" subtype of mitochondrial autoantigens: the E2 subunits of the pyruvate dehydrogenase complex, the branched chain 2-oxo-acid dehydrogenase complex, the ketoglutaric acid dehydrogenase complex, and the dihydrolipoamide dehydrogenase-binding protein [3]. Each of these autoantigens participates in oxidative phosphorylation (a process occurring in the inner mitochondrial membrane by which ATP is formed) and have a great deal of homology.

A major advance in our understanding of PBC occurred with the identification and cloning of these antigens [70]. Most AMA react against the dihydrolipoamide acetyltransferase component (E2 subunit) of pyruvate dehydrogenase (PDC-E2) [71-74]. PBC is the only disease in which there are B- and T-cells that are autoreactive against PDC-E2 (image 1). Thus far, all of the mitochondrial autoantigens screened have been targets of only the M2 antimitochondrial antibodies. Other antimitochondrial antibodies previously described, anti-M4, anti-M8, and anti-M9, are most likely artifacts of the methods used to detect them [75,76]. Human antimitochondrial antibodies inhibit the enzymatic activity of their target enzyme complexes in vitro [72].

Antimitochondrial antibodies are found in 95 percent of patients with PBC when using the most sensitive detection techniques [77], and have a specificity of 98 percent for the disease [71]. However, their role in pathogenesis is unclear as illustrated by the following observations:

●AMA titers vary greatly (more than 20-fold) among patients but tend to be stable over time in an individual patient; antibody titers do not correlate with disease severity or rate of progression [78,79].

●There is no apparent difference in the clinical spectrum or course of patients with PBC who are AMA-positive or AMA-negative [80].

●Antimitochondrial antibodies raised in animals immunized with recombinant human pyruvate dehydrogenase do not cause bile duct damage or any recognizable disease [81].

●Patients with either AMA-negative or AMA-positive PBC have a similar response to treatment with ursodeoxycholic acid or liver transplantation [82].

T-CELL MITOCHONDRIAL RESPONSE — Although the role of antimitochondrial antibodies in the pathogenesis of primary biliary cholangitis (PBC) is uncertain, accumulating data suggest a direct role for T-lymphocytes. CD4+ and CD8+ cells are present in high concentration in the portal triads of patients with PBC, often surrounding and infiltrating necrotic bile ducts [83-87]. Autoreactive T-lymphocytes from PBC patients are specific for PDC-E2 [88-91]. There is a 100- to 150-fold increase in precursor frequencies of PDC-E2-specific T-cells isolated from the liver and hilar lymph nodes compared to peripheral blood mononuclear cells in patients with PBC [84]. PDC-E2, normally found in the inner mitochondrial membrane of all cells, is aberrantly expressed in the luminal surface of bile duct epithelial cells only in patients with PBC [85].

The immunodominant epitope recognized by MHC class II restricted CD4+ T-cells in patients with PBC has been identified as amino acids 163 to 176 of PDC-E2 [90,91]. The immunodominant epitope recognized by MHC Class I restricted CD8+ T-cells has also been identified and is similar (amino acids 159 to 167 of PDC-E2) [87]. This peptide induces proliferation of specific MHC-class I restricted CD8+ cytotoxic T-lymphocytes from most HLA-A*0201 patients with PBC [92] and the frequency of cytotoxic T-lymphocytes recognizing this epitope is increased 10-fold in liver compared to blood in patients with PBC [87]. An alanine substitution at position 5 of this epitope significantly reduced peptide-specific cytotoxicity and interferon gamma production by cytotoxic CD8+ T-lymphocytes derived from peripheral blood monocytes of patients with PBC [93].

These data suggest that bile duct epithelial cells expressing this epitope are targeted by CD8+ T-lymphocytes in patients with PBC. Why bile duct epithelial cells express this epitope is still unknown. One possibility is that IgA is complexed to PDC-E2 in the serum of patients with PBC and this antigen-antibody complex is then transcytosed from the basal to apical side of the bile duct epithelial cell en route to its secretion in bile [85]. Serum IgA has been found at mitochondrial surfaces and immune complexes containing PDC-E2, and IgA have been found in bile duct lumens in patients with PBC [85,94].

In addition to the T-lymphocyte mediated destruction of small bile ducts, secondary damage to hepatocytes may occur from the accumulation in the liver of increased concentrations of potentially toxic substances, such as bile acids, which are normally secreted into bile (algorithm 1). The naturally occurring bile acids (cholic acid, chenodeoxycholic acid, and deoxycholic acid) are all detergents and can dissolve cell membranes if present in a sufficiently high concentration [95-97]; such toxic concentrations are often reached in cholestasis (figure 1) [98].

WHY TISSUE INJURY IN PBC IS CONFINED TO THE LIVER — A paradox of PBC is that the target mitochondrial antigens are present in every cell in the body while the immune attack appears to be highly specific for biliary epithelium. In addition, PBC may reoccur following liver transplantation. Since the donor liver contains donor MHC genes, this observation suggests that immune mechanisms, other than adaptive immunity, may perpetuate disease. Proposed mechanisms include innate immune responses [99]. One theory is that there may be qualitative differences in the processing of apoptotic bile-duct cells that account for the specificity, but the precise mechanisms involved are incompletely understood [100-102].

Another hypothesis proposes that any factor (whether it be an infection or a chemical) that damages bile duct epithelial cells may unmask an antigen in these cells that shares some epitopic similarities with pyruvate dehydrogenase complex E2 [103,104].

A molecule that shares some antigenic determinants with the E2 subunit of pyruvate dehydrogenase but is distinct from it has been identified on the luminal surface of biliary epithelial cells of patients with PBC early in the course of disease (image 1) [105,106]. The presence of this antigen, found only in bile duct epithelial cells of patients with PBC, could explain why cell injury in PBC is limited to biliary epithelial cells.

Expression of this autoantigen on the luminal surface of biliary epithelial cells may also provoke an antibody mediated attack by IgA antibodies, the antibody present in bile [105]. In addition, this E2-like antigen, together with the appropriate class II major histocompatibility complex antigens, and another molecule required for antigen presentation, BB1/B7, could be the target of activated CD8+ lymphocytes. All of these molecules are found in and around damaged bile ducts in PBC [106]. The E2 component appears in damaged bile ducts before the HLA class II antigens and BB1/B7 supporting a sequence whereby the bile duct epithelial cell injury precedes the T-cell mediated immune response. The histology of PBC is also consistent with T-lymphocyte mediated cytotoxicity as is the fact that the E2 antigens stimulate interleukin-2 production by cloned T-cells isolated from liver tissue [103].

ROLE OF RETAINED BILE ACIDS — The "foamy" degeneration of hepatocytes in primary biliary cholangitis (PBC) has been attributed to the noxious effects of bile acids (picture 3) [95]. Cholestasis per se causes increased expression of HLA class I antigens on hepatocytes, thereby rendering them better targets for activated T-lymphocytes. In this regard, there is increased evidence that the biliary system requires a bicarbonate umbrella, and that disruption of this umbrella makes biliary cells more prone to damage. In patients with PBC, emerging data suggest that the bicarbonate umbrella is defective. Whether the umbrella dysfunction is primary or secondary to disease is unknown [107].

Reversal of these pathogenetic events may explain the efficacy of ursodeoxycholic acid, a dihydroxy bile acid, which has been approved by the FDA for the treatment of PBC. Studies in both humans and experimental animals suggest that treatment with ursodeoxycholic acid reduces expression of HLA class I antigens by hepatocytes and also lessens the toxicity of retained naturally occurring bile acids such as cholic acid and chenodeoxycholic acid [96,97]. (See "Overview of the management of primary biliary cholangitis".)

FETAL MICROCHIMERISM — The persistence of allogeneic fetal cells in the maternal system has been termed fetal microchimerism. These cells have been postulated to have a role in the development of maternal diseases such as scleroderma [108], providing the rationale for studies of microchimerism in patients with PBC. Several studies on microchimerism in PBC have produced discordant results [109-112]. Thus, at the present time, the role of microchimerism is unproven [113].

CLINICAL PROGRESSION — The pathogenesis of biliary cirrhosis itself is straightforward after the immunologic injury and is similar to that which occurs in a variety of chronic cholestatic diseases, such as bile duct strictures and biliary atresia in children. Chronic impairment of bile flow leads to portal and parenchymal inflammation, liver cell necrosis, scarring, and eventually to cirrhosis and liver failure. The clinical progression, however, requires years, and at present there are no clear prognostic factors that predict the rate of liver failure in a given patient. Furthermore, it is likely that there are various immunological mechanisms at play during the progression of disease (ie, there may be different pathways involved in the inflammatory response, the destruction of bile ducts, and the development of fibrosis). (See "Clinical manifestations, diagnosis, and prognosis of primary biliary cholangitis".)

SUMMARY

●Background – Primary biliary cholangitis is characterized by a T-lymphocyte-mediated attack on small intralobular bile ducts (picture 1A-B). A continuous assault on the bile duct epithelial cells leads to their gradual destruction and eventual disappearance (picture 2). (See 'Introduction' above.)

●Pathogenesis – The pathogenesis of primary biliary cholangitis is incompletely understood but appears to involve genetic susceptibility and environmental factors. It is likely that more than one environmental trigger may elicit bile duct epithelial injury in a genetically susceptible individual. This will lead to both adaptive and innate immune responses that lead to portal inflammation and bile duct epithelial damage. (See 'Clues about etiology based on the epidemiology of PBC' above.)

●Clinical progression – Once immune-mediated bile duct injury has been established, the disease progresses due to chronic cholestasis and secondary inflammation and scarring. (See 'Clinical progression' above.)