Primary biliary cirrhosis: What do autoantibodies tell us?

Abstract

Primary biliary cirrhosis (PBC) is a chronic, progressive, cholestatic, organ-specific autoimmune disease of unknown etiology. It predominantly affects middle-aged women, and is characterized by autoimmune-mediated destruction of small- and medium-size intrahepatic bile ducts, portal inflammation and progressive scarring, which without proper treatment can ultimately lead to fibrosis and hepatic failure. Serum autoantibodies are crucial tools for differential diagnosis of PBC. While it is currently accepted that antimitochondrial antibodies are the most important serological markers of PBC, during the last five decades more than sixty autoantibodies have been explored in these patients, some of which had previously been thought to be specific for other autoimmune diseases.

Key Words: Primary biliary cirrhosis; Autoimmune disease; Autoantibody; Anti-mitochondrial antibody; Anti-gp210 antibody; Anti-sp100 antibody; Anti-centromere antibodies

INTRODUCTION

Primary biliary cirrhosis (PBC) is a progressive autoimmune liver disease characterized by infiltration of lymphocytes in portal tracts, progressive destruction of intrahepatic small bile ducts and the presence of serum antimitochondrial antibodies (AMA)[1,2]. As is the case for the majority of autoimmune diseases, PBC affects predominantly women. Recent investigations have suggested that PBC, sometimes asymptomatic, is not a rare disease. During the last several years advanced biochemical assays, further delineation of specific liver histological findings, more effective serum autoantibody detection methods and improved diagnostic abilities have led to higher prevalence estimates worldwide[3-5]. Currently it is believed that PBC is likely to be triggered by a combination of environmental factors including infection in a genetically susceptible individual. This hypothesis is supported by the high concordance rate of PBC among first-degree relatives and in homozygous twins (approximately 60%)[6,7]. Specific immunologic damage to biliary epithelium and a mechanism of tissue destruction in PBC has been elucidated[8,9]. In addition, epitopes of T cells and B cells targeting mitochondrial autoantigens have been identified[10-12]. Furthermore, a number of autoantibodies previously thought to be specific markers for another autoimmune disease have been detected in patients with PBC.

Disease progression and clinical manifestations in PBC varies. The fact that a variety of autoantibodies have been detected in PBC suggests the disease has a complicated pathogenesis. In this review, the properties of these autoantibodies and their autoantigen characteristics, as well as their pathogenetic and clinical significance were discussed.

AMA

The presence of AMA is pathognomonic for PBC[13], and it is generally accepted that AMA can be detected in serum years before the advent of any clinical manifestation or biochemical abnormality[14-16]. AMA were first described in 1958[17] in sera from patients with chronic liver disorders and then detected by Walker et al[18] in 1965 using an immunofluorescence test. In the past 40 years an enormous number of experimental studies have focused on AMA, and numerous rewarding discoveries have been made. There are nine subtypes of AMA, four of which have been involved in PBC, including anti-M2, anti-M4, anti-M8 and anti-M9. It has been demonstrated that the autoantigens recognized by anti-M2 are located in the inner membranes of mitochondria, whereas those recognized by anti-M4, anti-M8 and anti-M9 are located in the outer mitochondrial membranes. Anti-M9 can be detected in both anti-M2-positive and -negative PBC patients, while anti-M4 is only positive in the presence of anti-M2. All four of these AMA subtypes are relatively specific for the diagnosis of PBC.

Anti-M2

M2 has been found to contain five antigenic determinants, with molecular weights of 70 kDa (a), 56 kDa (b), 51 kDa (c) 45 kDa (d) and 36 kDa (e), all of which were identified subsequently as members of the 2-oxoacid dehydrogenase complex of enzymes within the mitochondrial respiratory chain, including the E2 subunit of the pyruvate dehydrogenase complex (PDC-E2), the E2 subunit of the branched-chain 2-oxoacid dehydrogenase complex, the E2 subunit of the 2-oxoglutarate dehydrogenase complex, E1t alfa subunits of PDC and E3 binding protein (protein X)[19,20]. The exact molecular weight of the M2 band differs among laboratories according to mitochondria species being used and specifics of techniques for antigen preparation and detection. In patients with PBC, approximately 90%-95% of serum samples react against PDC-E2, making this the most important autoantigen in the disease. Anti-M2 is the most important subtype used in routine diagnostic tests for PBC. Its level in affected sera is high and it also exists in other body fluids such as saliva and bile[21-23]. As AMA is considered to be the hallmark of PBC, a positive test is potentially diagnostic, or at least indicative that the individual is at increased risk for future development of PBC[15].

Anti-M4

The anti-M4 antibody was originally detected in patients with chronic cholestatic liver disease (mixed form) associated with two different types of complement-fixing AMA[24]. M4 is a single antigen with molecular weight of 52 kDa. It can be detected by a complement fixation test but not immunoblotting. Unlike M2, the M4 antigen is trypsin-insensitive and its band at sucrose densities is 1.08 to 1.14. Anti-M4 is found predominantly in patients with histological features of chronic active hepatitis and PBC. Recent studies have identified the major proteins in the M4 fraction which is related to the PDC-E1 subunits and sulphite oxidase[25,26].

Anti-M8

The M8 antigen is also trypsin-sensitive with a band at sucrose densities of 1.16 to 1.24. Anti-M8 has been found only in coexistence with anti-M2, the presence of anti-M8 indicates progressive disease activity. On the other hand, not all anti-M2-positive patients have anti-M8. Like M4, the M8 antigen also locates in the outer mitochondrial membranes[27].

Anti-M9

Anti-M9 antibody was accidentally found when testing anti-M2-positive sera against trypsinized submitochondrial particles from rat liver shown to be devoid of anti-M2[28]. Anti-M9 antibody is detected predominantly in patients with asymptomatic and early PBC, and it also can be positive in anti-M2-negative PBC patients. Unlike anti-M4 and anti-M8, which seem to reflect disease activity, anti-M9 antibody occurs early in PBC. Patients with only anti-M9 have all the typical biochemical features found in classic anti-M2-positive patients, but seem to have slower disease progression and benign outcome, whereas patients having complement-fixing antibodies against anti-M2, anti-M4, and anti-M8 seem to have more active disease and worse outcome[29-31], though this finding wasn’t supported by a blinded study on Dutch PBC patients conducted by Vleggaar et al[32].

ROLE OF AMA IN PBC PATHOGENESIS

Although AMA serve as highly sensitive markers for the diagnosis of PBC, autoantibodies against various mitochondrial enzymes can frequently be detected in patients with other diseases, such as primary Sjögren’s syndrome (pSS), scleroderma, autoimmune hepatitis[33,34] and some infectious diseases like tuberculosis and viral hepatitis[35-38]. It is very interesting that the prevalence of AMA in first-degree relatives of PBC probands is as high as 13.1%, whereas in gender, age, race, and residence-matched controls the prevalence is only 1%, suggesting that environmental risks and genetic determinants are likely implicated in the etiology of PBC[7].

As no clinical correlation can be found, and animal models with serum AMA do not consistently have PBC-like liver lesions, the exact role played by AMA in the immunopathology and pathogenesis of PBC remains elusive. However, current data indicate that the destruction of biliary cells is mediated by liver-infiltrating autoreactive T cells specific for the dominant PDC-E2 autoantigen[39]. The dominant epitopes of autoreactive T and B cells have been identified. The CD4⁺ T cell epitope appears to localize to peptides 163-176, the CD8⁺ T cell epitope appears to localize to peptides 159-167, while the B cell epitope appears to localize to peptides 167-186[39-43]. Furthermore, the most prominent immune features of autoreactive CD4⁺ and CD8⁺ T cells can be detected in peripheral blood from patients with PBC. The disease-related AMA-specific CD8⁺ T cells are enriched up to about 10-fold and the CD4⁺ T cells are enriched up to more than 100-fold in liver compared to peripheral blood samples[42,44]. Presently the data suggest that B and T cells in PBC patients respond simultaneously to the same autoantigen, and that both are involved in the pathogenesis of PBC.

Study of stored sera of well-characterized PBC patients followed for 7-28 years indicate that AMA levels are not associated with disease severity and progression. Most studies except the one conducted by Poupon et al[45] did not support that AMA levels could be affected by treatments during the observation period[45-47]. In fact, low levels of AMA persisted for up to 11 years following liver transplantation[47]. AMA are non-organ- and non-species-specific, and contain IgA, IgG and IgM subclasses. Data from PBC patients demonstrate that the presence of AMA IgA in sera or saliva might be associated with disease progression[23] and some studies suggested that greater concentrations of AMA IgA in biliary and mucosal secretions, constant transcytosis, would render the exposed cells more susceptible to apoptosis resulting in subsequent bile duct damage[48], while others proposed the hypothesis that AMA IgA can be transported to the vascular side of the bile duct cell where it can induce apoptosis by reacting with PDC-E2-like molecules located on the luminal surface cell membrane[49]. Many studies have demonstrated that the different AMA IgG subclasses have different clinical significance. PBC patients positive for IgG3 AMA had histologically more advanced disease and were more frequently cirrhotic than those who were negative. Furthermore, there was a positive correlation between AMA IgG3 titers and Mayo risk scores: this subclass is associated with poor prognosis, possibly reflecting the peculiar ability of this isotype to engage mediators of immunological damage[50].

Currently it is believed that a positive AMA titer is virtually pathognomonic of current PBC or risk for future development of the disorder, although the mechanisms leading to the generation of AMA have not been elucidated. Several possible mechanisms have been suggested regarding the generation of AMA, such as oxidative damage, molecular mimicry and changed biliary epithelial cell (BEC) apoptosis[51,52]. The fact that high levels of AMA can be detected in patients with acute liver failure supports the hypothesis that oxidative stress-induced liver damage may lead to induction of AMA[53]. But it is also surprisingly true that the AMA in these patients disappear rapidly, suggesting the pathogenesis of PBC is multifactorial. It has been demonstrated that molecular mimicry between bacterial or viral antigens and mitochondrial antigens can trigger the generation of AMA in PBC[54,55]. Modification of the inner lipoyl domain of E2 with halide or ethyl halide results in increased reactivity of AMA from PBC patients, suggesting that xenobiotics might make cellular components antigenic[56].

There is growing evidence showing that the onset of PBC may be the result of inefficient removal of apoptotic cells. It is of interest to note that a recent report proposed that PDC-E2 in patients with PBC is released without caspase cleavage from apoptotic BEC, supported by the fact that glutathionylation of the lysine lipoic acid moiety on the PDC-E2 is sometimes, though not commonly, decreased by serum AMA via Bcl-2[57]. Other studies show that apoptotic cells are phagocytosed by BECs, a function mediated by anti-CD16, and so consequently act as an exogenous source of autoantigens in cholangiocytes[9,58]. Defects in the elimination of apoptotic cells can lead to secondary necrosis accompanied by subsequent release of intracellular components, which might explain the generation of autoantibodies against intracellular antigens like AMA[59].

Further studies in the field of AMA in PBC have led to speculation about the existence of an AMA-negative PBC subgroup. It is not clear whether there is indeed such a subgroup, having distinct features, or if this is an artifact due to technical limitations of current AMA detection methods leading to false-negative results in some PBC patients[60]. Present data indicate that there is no discernable difference between AMA-positive and -negative PBC in terms of clinical manifestations, liver biochemistry and histopathology findings, disease course, as well as response to treatment[61-63]. As more sensitive and specific serologic tests are applied, many patients initially believed to be AMA-negative are subsequently found to be AMA-positive[64,65]. These findings cast doubt on the existence of a true AMA-negative PBC subgroup.

ANTINUCLEAR ANTIBODIES IN PBC

Antinuclear dot antibodies (SP100, PML, NDP52 and SP140)

PBC patients often have autoantibodies with nuclear dot (ND) stain patterns in the indirect immunofluorescence (IIF) assay. The major antigens associated with ND are as follows: sp100 proteins, which are transcription-activating proteins autoantigenic primarily in patients with PBC and occasionally in rheumatic disorders[66,67]; promyelocytic leukemia (PML) protein, a transformation and cell-growth suppressing protein aberrantly expressed in PML cells that was discovered in studies of the development of acute PML; NDP52, a protein of the myosin VI binding partners which was previously shown to contribute to innate immunity[68,69]; and sp140 proteins, which are identified as autoantigenic proteins in PBC recently. Sp100 and PML were discovered in the context of leukemic transformation and as autoantigens in PBC[70]. They are reported to be co-autoimmunogenic, often in patients with PBC[71]. The sp100 antigen was described by Szostecki et al[66] as a peptide of 480 amino acids with an aberrant electrophoretic mobility to 100 kDa, and a calculated molecular weight of 53 kDa. It was subsequently characterized by complementary DNA cloning, and the deduced amino acid sequence was found to contain sequence similarities with HIV-1 nef proteins[72]. The prevalence of anti-sp100 antibodies in PBC is about 25%, and it appears to be highly specific for a diagnosis of PBC, but only when other diseases can be excluded and the typical clinical context is present[73,74]. The presence of anti-sp100 antibodies serves as a serologic marker of PBC, which could be useful in clinics to confirm the diagnosis, especially in AMA-negative PBC patients[75,76]. Recent data indicate that as reports of AMA-negative PBC decrease due to development of more sensitive and specific serologic tests for serum AMA, anti-sp100 antibodies appear to be more common in AMA-positive PBC patients than in those who are AMA-negative[77,78]. Also, anti-sp100 antibodies are increasingly found to be present in many clinical conditions, such as systemic lupus erythematosus (SLE) and pSS. It is of interest to note that among female patients with urinary tract infections but no liver disease, 80% of the AMA-positive, but none of the AMA-negative patients were also positive for anti-sp100 antibodies. It is also well established that among PBC patients, about 74% of patients with urinary tract infections were positive for anti-sp100, whereas the positivity was only 4.8% in PBC patients without urinary tract infections[79]. Given the high specificity of anti-sp100 as an immunoserological hallmark of PBC, these findings support the hypothesis that some infections such as Escherichia coli are involved in the induction of PBC-specific autoimmunity.

PML protein was discovered in cells of patients with acute PML as a protein fused with the retinoic acid receptor-a (RAR)[80,81]. PML protein functions as a nuclear hormone receptor transcriptional coactivator[82]. Subsequently it was shown to form ND patterns when tested by immunofluorescence microscopy with serum anti-PML antibodies from PBC patients. Anti-PML antibodies often coexist with anti-sp100 antibodies in individuals with PBC[71], and are present in about 19% of PBC patients[83]. Current study indicates that anti-PML antibodies are highly specific for PBC even when autoantibodies against mitochondrial antigens are not found[84].

Anti-sp140 antibodies were recently identified for the first time in patients with PBC by Granito et al[85]. They are present in about 15% of PBC patients and are highly specific for PBC. Anti-sp140 antibodies coexist with anti-sp100 and anti-PML antibodies. No association was found between anti-sp140 and any particular clinical feature of PBC.

Antinuclear pore antibodies (gp210 and p62)

In addition to AMA, a number of nuclear antigens have been recognized as targets of antinuclear antibodies (ANA) in patients with PBC, including several components of the nuclear pore complex (NPC), such as the gp210 and p62 proteins. These antibodies have a nuclear periphery fluorescence pattern in the IIF assay, as first reported by Ruffatti et al[86] in 1985. Several reports revealed that the frequency of PBC-specific nuclear envelope antibodies ranged from 16% to 30%[76,87], and that the frequency increased greatly when fluorescent-labeled specific antiserum of the IgG subclass was applied[88,89]. In 1990 a study by Lassoued et al[90] showed that autoantibodies from patients with PBC having a punctate fluorescence pattern in IIF react with a protein of molecular mass approximately 200 kDa, which was identified as the NPC membrane protein gp210[91]. Gp210 is an integral glycoprotein of the nuclear pore consisting of three main domains: a large glycosylated luminal domain, a single hydrophobic transmembrane segment and a short cytoplasmic tail. Gp210 is recognized by antibodies in approximately 25% of patients with PBC[92]. The gp210 epitope recognized by most of the autoantibodies is a 15 amino acid stretch in the cytoplasmic, carboxyl-terminal domain of the protein. In the ANA category, these anti-gp210 antibodies are particularly significant since they are highly specific for PBC[93,94]. In addition, several reports link the presence of anti-gp210 antibodies in PBC patients with disease severity and poor prognosis. Since the presence of anti-gp210 antibodies correlates with an unfavorable disease course and more rapid progression, it is useful for monitoring the effect of ursodeoxycholic acid and for the early identification of patients at high risk for end-stage hepatic failure, and so may potentially become an important prognostic marker in PBC patients[95,96]. Findings to date clearly indicate that anti-gp210 antibodies having the best predictive value regarding progression to end-stage hepatic failure. The proposed mechanism for this predictive role is based on the following hypothesis that the breakdown of immunological tolerance to mitochondrial antigens such as PDC-E2 is not enough for the progression to hepatic failure, whereas the breakdown of immunological tolerance to nuclear antigens such as gp210, in which molecular mimicry is involved as well as increased and aberrant expression of gp210 in small bile ducts, may play a crucial role[97].

A few years after the discovery of anti-gp210 antibodies in PBC, reactivity of PBC sera with a 60 kDa component of NPC was reported. Anti-p62 antibodies, which also generate a perinuclear pattern in IIF, were first described in 1987[98-100]. They occur as frequently as the anti-gp210 glycoprotein autoantibodies[101], and with a specificity for PBC of up to 97%. Anti-p62 antibodies reacting with the 60 kDa component localize to the NPC. The frequency of anti-p62 antibodies in PBC is about 30%-55%. Their presence in PBC is not associated with AMA, but is associated with disease progression. Data from a multicenter study indicated that anti-p62 complex antibodies might be related to the progressive or advanced stage of PBC[99,102], that their prevalence is higher in symptomatic patients and that they are associated with more severe disease, defined as the presence of cirrhosis or its complications. In addition, it has been reported that anti-p62-positive patients have higher levels of serum bilirubin and more marked inflammatory infiltrates on liver biopsy[87].

Antinuclear envelope antibodies (Lamin and Lamin B receptor)

The nuclear envelope is a bilayered membranous structure that can be divided into five distinct components: the inner nuclear membrane, having a distinct set of integral membrane proteins; the outer nuclear membrane; a perinuclear space, which is continuous with the lumen of the endoplasmic reticulum; the pore domains, regions where the inner nuclear membrane and outer nuclear membrane come together and fuse; and an underlying nuclear lamina, containing the nuclear lamins[103]. A smooth membrane fluorescence pattern is characteristic of the presence of antibodies to nuclear lamins in IIF using sera from PBC patients. Three subtypes of anti-lamin antibodies have been described: anti-lamin A, B and C[102,104-106]. Anti-lamin antibodies do not seem to be disease-specific as they are found in patients with several different autoimmune disorders, such as SLE, chronic fatigue syndrome, and PBC[107-110]. Anti-lamin A, B and C antibodies are detected with frequencies of 6%-8% in sera from patients with PBC. The usual scenario is to find anti-lamin A and C together, and less frequently either anti-lamin B alone or all three in the same patient[111].

Lamin B receptor (LBR) is a protein integral to the inner nuclear membrane with a nucleoplasmic, amino-terminal domain of 208 amino acids, followed by a carboxyl-terminal domain with eight putative transmembrane segments. Anti-LBR antibodies from PBC patients recognize the nucleoplasmic, amino-terminal domain but not the carboxyl-terminal domain. Anti-LBR antibodies appear to be highly specific for PBC, but their clinical significance is unclear. The prevalence of anti-LBR antibodies in PBC is approximately 2%-6%[76,102,112,113].

Anti-centromere antibodies

Anti-centromere antibodies (ACA) are important diagnostic markers of systemic sclerosis (SSc), found in about 25% of these patients[114]. In patients with CREST syndrome or limited cutaneous SSc, the positivity rises to 50%-90%. ACA in SSc are usually associated with a good prognosis, though they are not specific for SSc. ACA can be detected in patients with other rheumatic diseases including pSS, SLE and PBC (about 30%)[115-120]. It is of interest to note that several subtypes of ACA have been identified, including anti-CENP-A, anti-CENP-B, anti-CENP-C and anti-CENP-O antibodies[121]. Research during the past several years has found that prevalence of the ACA subtypes differs among various autoimmune diseases[122]. Recent studies have demonstrated that ACA positivity in patients with PBC is of significant predictive value for progression to portal hypertension[123,124].

OTHER AUTOANTIBODIES DETECTED IN PBC

Although extensive research has focused on AMA, it is of interest to note that, to date, more than sixty different autoantibodies have been found in PBC patients. Some target at nuclear or cytoplasmic molecules and cell membranes, while others react with lipid components. Some, like AMA, occur frequently and almost universally in PBC, while others, like anti-lamin and anti-LBR, are present in only a few patients. It should be noted that among these autoantibodies, some are not specific for any disease, and some are thought to be more closely related to other autoimmune diseases, such as anti-CCP which is relatively specific for rheumatoid arthritis[125,126]. Prevalence and properties of these autoantibodies in PBC are summarized in Table 1.

Table 1 Autoantibodies in primary biliary cirrhosis that are closely related to other autoimmune diseases.

No.	Autoantibody	Autoantigen properties	Prevalence in PBC (%)	Clinical associations	Ref.
ANA
1	Anti-chromatin	Chromatin	8.9-25.0	Anti-chromatin antibodies are reported to be associated with disease activity in AIH, but their roles in PBC remains to be investigated	[99,127-129]
2	Anti-dsDNA	Double-stranded deoxyribonucleic acid	17.0-22.0	Anti-dsDNA antibodies are one of important criteria for the diagnosis of SLE. Co-existence of AMA and anti-dsDNA autoantibodies can be considered the serological profile of AIH/PBC overlap syndrome	[76,127,128,130,131]
3	Anti-ssDNA	Single-stranded deoxyribonucleic acid	59.0-71.0	Anti-ssDNA antibodies can be detected in many diseases	[132,133]
4	Anti-histone	Histone	43.6	Anti-histone antibodies are generally considered to be related to drug-induced lupus, though it can be detected in many autoimmune diseases including PBC	[132,134]
5	Anti-scl-70	Topoisomerase-1	3.0-24.0	Anti-scl-70 antibody serve as a specific maker for diffuse SSc and presents in 30%-60% of subjects with diffuse SSc	[127,132]
6	Anti-Sm	Proteins of 28/29, 16, 16.5, 18, and 12, 11, 6 kDa which participate in pre-messenger RNA processing into spliced mature mRNA	7.0-34.0	Anti-Sm autoantibodies are highly specific for SLE	[76,127,132,135]
7	Anti-SSA	Intracellular ribonucleoproteins of 60 and 52 kDa that are associated with small RNAs	5.0-30.0	Anti-Ro(SS-A) and anti-La(SS-B) antibodies are more frequently seen in SS and SLE. Their presence in PBC suggests that PBC often overlaps with SS	[76,127,132,135,136]
8	Anti-SSB	An intracellular ribonucleoprotein of 47 kDa that are associated with small RNAs	2.0-21.0
9	Anti-RNP	Ribonucleoprotein	5.0	More frequently seen in SLE	[76,127]
10	Anti-Jo-1	Histidyl tRNA synthetase	26.0	Anti-Jo-1 antibodies are predominantly detected in patients with myositis	[135]
11	Anti-U1RNP	U1snRNPs that contain specific proteins of 70, 33 and 20 kDa	3.1-5.0	Anti-U1snRNP antibodies predominantly present in SLE, and can be detected in PBC patients. The clinical significance of anti-U1snRNP antibodies in PBC is unknown	[137,138]
Other liver diseases-associated autoantibodies
12	Anti-SMA	A variety of target antigens including F-actin, G-actin, myosin, tropomyosin, troponin, desmin, vimentin, keratin, etc.	8.0-25.0	Anti-SMAs present mainly in AIH-I, and can also be detected in chronic active hepatitis. The presences of anti-SMAs in PBC are potential indicators of AIH/PBC overlap syndrome	[131,139]
13	Anti-SLA	SLA and liver and pancreas antigen	2.0-3.9	Anti-SLAs are autoantibodies seen in AIH-III.The presences of SLA autoantibodies in PBC indicate secondary autoimmune hepatitis	[9,140,141]
14	Anti-LKM	Liver kidney microsomal antigen	0.7	Anti-LKM antibodies occur preferentially in AIH-II. Anti-LKM autoantibodies can be seen in 21.4% of HCV-infected PBC patients, which suggests a close association between LKM and HCV-infected PBC	[142]
15	Anti-ASGPR	Asialoglycoprotein receptor	22.0-23.0	Anti-ASGPR antibodies mainly present in AIH and PBC. The autoimmune responses against ASGPR have been implicated in the development of AIH and PBC	[143-146]
16	Anti-LCM	Liver cell membrane specific antigen	42.0	Anti-LCM antibodies are detected predominantly in patients with HBsAg-negative chronic active hepatitis, but are also found in other liver diseases such as PBC	[147-149]
17	Anti-LSP	Liver specific protein	48.5	Anti-LSP antibodies present in viral hepatitis and autoimmune liver disease, and are found to correlate with severity of periportal inflammation and piecemeal necrosis in PBC	[144,150]
18	Anti-calreticulin	Calreticulin	20.0	Anti-calreticulin antibodies present in autoimmune liver disorder and IBD. They are not specific for PBC	[151,152]
19	Anti-FH	Fumarate hydratase	19.4	Anti-FH antibodies are found to be present predominantly in AIH. It can also be detected in PBC and other liver disease. The prevalence and clinic significance of anti-FH in PBC need further study	[153]
20	Anti-PGAM-B	Phosphoglycerate mutase isozyme B	16.7	Anti-PGAM-B antibodies are found to be present in 70.0% of AIH and 16.7% of PBC. It is also present in about 10% of viral hepatitis and 3.7% of healthy control. The clinical significance of anti-FH needs further study	[154]
21	Anti-p97/VCP	P97/valosin-containing protein	12.5	Anti-p97/VCP antibodies predominantly present in PBC, and can be detected in about 9.7% of AIH. The presence of anti-p97/VCP antibodies in PBC suggests less progressive disease course and benign prognosis	[155-157]
22	Anti-GSTA1-1	Glutathione S-transferase	10.0	Anti-GST autoantibodies are detected in 16.0% of AIH and 10.0% of PBC. Patients of AIH with positive anti-GST have severe diseases and poor prognosis	[158]
23	Anti-ASL	Argininosuccinate lyase	23.0	Anti-argininosuccinate lyase is a newly identified autoantibody in liver disease and its clinical relevance remains unknown	[159]
24	Anti-calmodulin	Calmodulin	IgM 50.0	Anti-calmodulin autoantibodies neither associate with anti-SMA, ANA and AMA, nor with hyperglobulinemia. The clinic significant of anti-calmodulin is unclear	[160]
			IgA 42.9
Gastroenteropathy-associated autoantibodies
25	ASCA	Baker's yeast saccharomyces cerevisiae	24.2	ASCA serves as a serological marker of Crohn’s disease, and has also been detected in other autoimmune disorders and in 5%-6.3% of blood donors. The prevalence of ASCA in AIH is 20%-30%, in AMA-negative PBC 44%. ASCA is common in PBC patients and correlates with higher level of circulating IgA. The prevalence of ASCA in PBC may be an indirect sign of enhanced mucosal immunity, but does not necessarily indicate concomitant inflammatory bowel disease	[161-163]
26	Anti-Galectin-3	Galectin-3, a member of -galactoside-binding lectins	30.0	Anti-Galectin-3 autoantibodies are primarily associated with Crohn's disease, and correlate negatively with disease activity. The significance of anti-Galectin-3 IgG autoantibodies in patients with PBC is unknown	[164]
27	Anti-tTG	Tissue transglutaminase	10.0-26.7	Anti-tTG autoantibody is mainly found in celiac disease. The prevalence of anti-tTG in PBC varies due to different types of substrate utilized in detection	[127,165,166]
28	AGA	Gliadin	16.0-21.0	Anti-gliadin antibodies are considered as the most reliable serological markers for celiac disease. They are also frequently seen in PBC, and IgA subclass of anti-gliadin antibodies are more pronounced in patients with Scheuer's stage III-IV disease	[166,167]
Vasculitis-associated autoantibodies
29	ANCA	Antigens including proteinase 3, myeloperoxidase, bactericidal/permeability-increasing protein, lactoferrin,human leukocyte elastase, cathepsin G, lysozyme, azurocidin, etc.	2-26	ANCAs are primarily associated with systemic vasculitides such as Wegener’s granulomatosis, microscopic polyangiitis and Churg-Strauss syndrome	[76,131,168]
30	Anti-MPO	Myeloperoxidase	9.0	Predominantly in microscopic polyangiitis, necrotizing and crescentic glomerulonephritis, Churg-Strauss syndrome	[169]
31	Anti-PR3	Proteinase 3	3.0	Predominantly in Wegener’s granulomatosis, and also detectable in microscopic polyangiitis, necrotizing and crescentic glomerulonephritis	[127]
32	Anti-LF	Lactoferrin	25.0-35.7	Are detected in several autoimmune disorders, such as Crohn's disease, SLE, systemic vasculitides. They are not specific markers for PBC	[170,171]
Thrombophilia-associated autoantibodies
33	Anti-β2GPI	β(2)-glycoprotein I	IgG 2-15	Represent specific features of patients with antiphospholipid syndrome. Their presence in PBC often indicates severe disease and worse prognosis	[127,172,173]
34	Acl	Cardiolipin	IgG 27.3
35	Anti-PS	Phosphatidylserine	IgM 75
36	Anti-PT	Prothrombin	IgG 7
37	Anti-PE	Phosphatydilethanolamine	IgG 5
Diabetes mellitus-associated autoantibodies
38	Anti-GAD	Glutamic acid decarboxylase	5.5	Anti-GAD occurs preferentially in the patients with type 1 diabetes. Clinical significance of Anti-GAD in PBC is unclear	[174]
39	Anti-SOX13	Transcriptional factor SOX13	18.0	SOX13 was initially identified in type 1 diabetes. The present of anti-SOX13 in PBC may merely indicate an immune response to products of damage to parenchymal tissue	[175]
Autoimmune thyroid diseases-associated autoantibodies
40	Anti-TG	Thyroglobulin	54.5	Anti-TG, anti-TPO and anti-TR are markers of autoimmune thyroid diseases. Their significances in PBC are unknown	[176]
41	Anti-TPO	Thyroid peroxidase	45.5
42	Anti-TR	TSH receptor	9.1
Others autoantibodies
43	Anti-CCP	Cyclic citrullinated peptide	2.7-4.0	Anti-CCP antibodies are highly specific for RA with sensitivity of 60%-70%. Presence of anti-CCP antibodies in PBC patients suggests RA overlap	[125,126,177]
44	Anti-ClpP	Microbial caseinolytic proteases P	30-47	ClpP is highly conserved among bacteria. Anti-ClpP in PBC suggests infection factors and molecular mimicry involved in the pathogenesis	[178,179]
45	Anti-β-subunit of bacterial RNA polymerase	β-subunit of bacterial RNA-polymerase	32.8	These autoantibodies in PBC, suggest bacterial triggers of PBC	[180]
46	Anti-EPO	Eosinophil peroxidase	52.5	PBC patients with positive anti-EPO antibodies have less peripheral eosinophils	[181]
47	Anti-p53	Nuclear protein of 53 kDa that regulates cell proliferation and apoptosis	8.0	Anti-p53 autoantibodies are commonly seen in malignancies and organ-specific autoimmune diseases such as type 1 diabetes, thyroid diseases, PBC and AIH	[182]
48	Anti-acetylcholine receptor	Nicotinic acetylcholine receptor	58.8-74.0	Anti-acetylcholine receptor antibodies are primarily associated with myasthenia gravis, though PBC patients with positive anti-acetylcholine receptor antibodies do not have clinical symptoms of myasthenia	[169,183,184]
49	Anti-CAII	Carbonic anhydrase II	18-31	Anti-CAII antibody is likely a nonspecific marker of autoimmunity. It has been detected in a variety of autoimmune diseases, including Graves’ disease, type 1 diabetes, SS, SLE, AIH and PBC. In cases of PBC, no significant correlation has been found between anti-CAII antibody and AMA	[185-189]
50	Anti-α enolase	α-enolase	28.6	Anti-α-enolase antibodies present in a variety of inflammatory and autoimmune disorders, such as SLE, IBD, RA and AIH, and are not likely to be specific markers for any disease. They might be involved in destruction of biliary epithelium and are associated with hepatic failure	[190-195]
51	Anti-HSP	Heat shock proteins	45.7	Enhanced biliary expression of heat shock protein is found in PBC. Anti-HSPs are common in PBC, and are related to titers of AMA. They might cross-react with the main mitochondrial antigens in PBC	[196-199]
52	Anti-FKBP12	FK506 binding protein 12	44.4	The significance of anti-FKBP12 antibodies in PBC is unclear	[200]

SLA: Soluble liver antigen; PBC: Primary biliary cirrhosis; AMA: Antimitochondrial antibodies; SLE: Systemic lupus erythematosus; SSc: Systemic sclerosis; ANA: Antinuclear antibodies.

CONCLUSION

The presence of serum autoantibodies is characteristic of PBC, and is useful in the clinical diagnostic process in combination with histology and imaging studies. Numerous autoantibodies are found in sera from patients with PBC. This suggests that the development of PBC is a multi-factorial process. With growing numbers of clinical studies of autoimmune diseases and extensive application of more sensitive testing methods for antibodies, it has gradually been realized that the association between an individual autoantibody and autoimmune disease is not as specific as previously thought. AMA is very sensitive and anti-gp210 and anti-sp100 are highly specific for PBC. Other antibodies found in PBC, such as ACA, ASCA, ANCA and anti-sm, could also be found in other autoimmune diseases[131,161,162,168]. Although some autoantibodies are believed to be associated with the pathogenesis of PBC, these associations are likely to be extremely complicated and surely exert complex effects in many different ways. It is hard to understand these delicate associations based on our current knowledge of PBC, and further advanced studies are required to elucidate the pathogenesis of this autoimmune disease.