ANA and/or SMA are almost exclusively requisites for the diagnosis of AIH-1 356141530. In typical cases of AIH-1, these autoantibodies are detected in significant titers (≥1;80 in adults and ≥1:40 in children) in almost half of Caucasians patients with AIH-1, while ANA alone are detected in 15% and SMA alone in 35% 53035.

The most frequent and conventional method for ANA detection is the indirect immunofluorescence (IIF) assay on cryostat sections of rodent tissues and HEp-2 cells slides (Figures 1 and 2) 14153536. Different patterns of fluorescence are found by this assay due to the large variability of target-autoantigens in the nuclei of HEp-2 cells that have been recognized 353637383940. Actually, ANA are found to be directed against single or double stranded DNA, tRNA, SSA-Ro, snRNPs, laminins A and C, cyclin A or histones 353637383940. Most commonly, a homogenous (34–58%) or speckled (21–34%) immunofluoerescence pattern is demonstrable 1415353637. So far however, neither a liver-specific nuclear antigen nor a liver-disease-specific ANA has been identified. For this reason, subtyping of ANA in cases of AIH-1 seems to have limited clinical implication and diagnostic relevance in routine clinical practice 61415353637383940.

SMA are detected by IIF on rodent liver and kidney, due to staining of vessel walls (Fig. 3), and stomach due to staining of the muscle layer (Fig. 4). SMA are directed against structures of the cytoskeleton such as actin, troponin, vimentin and tropomyosin. In AIH-1, SMA are predominantly directed against F-actin 41. The latter seems to be diagnostically more relevant in pediatric patients where SMA may be the only marker of AIH-1 even in titers as low as of 1:40 1415. Czaja et al 41 have suggested that antibodies to actin are associated with a younger age of disease onset, the presence of HLA-A1-B8-DR3 haplotype and a greater frequency of treatment failure, death from liver disease and earlier requirement for transplantation than actin-antibody negative AIH-1 patients.

ANA and/or SMA – usually in low titers – may occur in patients with chronic viral hepatitis B or C, but in most of these cases SMA lack F-actin specificity 14154243. From the clinical point of view, interferon-alpha administration is generally safe in most cases of viral hepatitis with ANA and/or SMA, although occasionally may provoke mild self-limited autoimmune disorders compared to viral hepatitis patients without ANA or SMA autoantibodies 444546. During immunosuppressive treatment, disappearance of ANA and/or SMA is observed in the majority of patients with AIH-1 47. However, autoantibody status is unable to predict immediate outcome after cessation of corticosteroid administration. Additionally, neither autoantibody titers at first diagnosis nor autoantibody behaviour in the time course of the disease are prognostic markers for AIH-1 141547. These findings indicate that ANA and SMA are not involved in the pathogenesis of AIH-1 and furthermore, their determination is more of diagnostic than prognostic value 5141547.

These autoantibodies are directed against cytoplasmic constituents of neutrophil granulocytes and monocytes. Classically, they are detected by IIF using ethanol-fixed granulocytes as substrate 48. Using the above method, two major subtypes can be distinguished. ANCA showing a diffuse or granular cytoplasmic staining (c-ANCA) and ANCA characterized by a perinuclear-staining (p-ANCA). Both c-ANCA and p-ANCA are valuable diagnostic and prognostic markers in systemic vasculitides in particular Wegener's granulomatosis and microscopic polyangiitis, respectively 4849. Proteinase-3 has been identified as the major target-autoantigen of c-ANCA in cases with Wegener's granulomatosis, while myeloperoxidase is the documented autoantigen of p-ANCA in most patients with microscopic polyangiitis 4849. Since then, ANCA (in most cases of p-ANCA type) were detected in a high prevalence in other inflammatory disorders of unknown aetiology such as, inflammatory bowel disease (more frequently in ulcerative colitis than in Crohn's disease) 5051 and primary sclerosing cholangitis (PSC), a liver disease that is frequently associated with ulcerative colitis 5152.

Several recent studies however, have also documented the presence of high titers of p-ANCA in the sera of patients with AIH-1 (Fig. 5; prevalence range 40–96%) 535455565758 and to a much lesser extent in PBC patients (prevalence range 0–39%) 545859. Occasionally, high titers of c-ANCA can be detected in AIH-1 (Dalekos GN, 2002, unpublished observations). In contrast, p-ANCA have not been detected in serum samples from patients with AIH-2 57. Low ANCA titers are detected infrequently in patients with alcoholic or chronic viral liver diseases 545760. However, a recent large study in 516 patients with hepatitis C virus (HCV) infection revealed the presence of ANCA in as high as 55.6% of patients 61. Interestingly these investigators have shown that all HCV positive sera with ANCA had c-ANCA pattern on IIF and contained proteinase-3 specificity 61. The clinical relevance of this finding remains to be determined. In patients with AIH-1, PBC or PSC the detection of ANCA appears to be associated with a more severe disease course or the presence of cirrhosis 5462. The latter suggestion however, was not confirmed by more recent studies 5354.

To determine the antigenic specificities of ANCA, antigen-specific enzyme linked immunosorbent assays (ELISAs) and Western blotting followed by immunodetection can be performed 1463. Using these techniques it became obvious that in AIH-1 the target-autoantigens recognized are multiple including cathepsin G, catalase, alpha-enolase, lactoferrin, actin and high mobility group (HMG) non-histone chromosomal proteins HMG1 and HMG2 1454565862636465. However, the determination of antigenic specificities of ANCA seems to have limited clinical relevance in patients with AIH-1 14155456.

In conclusion, the detection of ANCA may be a useful additional marker in searching for AIH-1, in particular in ANA/SMA/anti-LKM-1 negative cases of AIH. With the exception of a recent paper by Wu et al 61 the detection of ANCA is rather rare in chronic viral hepatitis 14545760. The latter may prove useful in the differential diagnosis between patients with AIH and those with viral hepatitis who tested positive for ANA or SMA. Furthermore, since ANCA appear to be relatively rare in PBC 5459, these autoantibodies may prove useful for distinguishing between genuine cases of AIH and cases of PBC with features overlapping with those of AIH 6. However, due to the lack of specificity for the diagnosis of AIH and to its obscure role – if any – in AIH, their routine determination is not recommended 1415.

The asialoglycoprotein receptor (ASGP-R) is a liver-specific glycoprotein of the cell membrane. Its main function is the internalization of asialoglycoproteins by binding a terminal galactose residue to coated pits. Anti-ASGP-R autoantibodies are detected in 88% of patients with AIH (both types) 6667. However, these autoantibodies are also found in some patients with PBC, chronic viral hepatitis B and C and alcoholic liver disease although at lower frequency and lower titers 14156667.

The ASGP-R is preferentially expressed on the surface of periportal liver cells where piecemeal necrosis is found as a marker of severe inflammatory activity in patients with AIH 68. This finding may suggest a possible immunopathogenetic involvement of anti-ASGP-R autoantibodies in AIH 69. The general presumption is that target of potentially tissue-damaging autoreactions in AIH must be liver-specific and available to the immune system in vivo (e.g. expression on the surface of hepatocytes). So far, ASGP-R is the only target-autoantigen that has been positively identified and fulfils these criteria 6869. Additional support to this emerged from the determinations of anti-ASGPR autoantibodies in consecutive AIH patients. The levels of anti-ASGP-R autoantibodies vary according to the inflammatory activity of the disease. In addition, anti-ASGP-R antibody titers decreased significantly in response to immunosuppression, while they reappear when the disease has relapsed 6670. These autoantibodies may be diagnostically helpful when other autoantibodies are not detected and AIH is suspected. However, due to the belief that anti-ASGPR antibody represents a general marker of liver autoimmunity and the limitations in its detection (requires chemically purified ASGP-R, which is not yet widely available), its routine use is not generally recommended.

The anti-SLA autoantibodies were described for the first time in 1987 20. They cannot be detected by IIF on common substrate. A competitive ELISA or a radioimmunoassay usually detects these autoantibodies 203271. SLA is found in 100000 g supernatant of liver homogenate and represent a cytosolic protein which is neither organ nor species specific 72. However, the highest concentrations are found in liver and kidney tissues. The anti-SLA autoantibodies are detected in patients with AIH alone or in combination with SMA and/or ANA 30313273. As noted above, similarities in the clinical profile between patients with AIH-1 (ANA and/or SMA positive) and AIH patients with anti-SLA alone in addition with an approximately 30% seropositivity overlap between anti-SLA and SMA and/or ANA suggest that anti-SLA is rather an additional important marker for the diagnosis of AIH-1, than a marker of a third type of AIH 614303132.

A scientific group from Tuebingen, Germany described for the first time the anti-LP autoantibodies in 1981 21. The LP antigen was predominantly detected in the S100 supernatant of liver and pancreas homogenates, indicating that this antigen was a soluble protein. Until recently, anti-LP and anti-SLA autoantibodies were thought to be different 202122. However, Wies et al report 74 provides convincing evidence and confirms previous suggestions that anti-SLA and anti-LP are one and the same autoantibody (anti-SLA/LP). In addition, the same study demonstrated that the identified target-autoantigen of anti-SLA/LP autoantibodies (a 35–50 kDa protein) was neither cytokeratins 8 or 18 71 nor glutathione-S-transferase isoenzyme 75. The results from two independent groups 7677 were similar with those found by Wies et al 74. After screening of cDNA expression libraries they identified a previously unknown amino acid sequence, which presumably encodes a UGA-suppressor tRNA-associated protein, as the targen-autoantigen of anti-SLA/LP autoantibodies 7677. The UGA-suppressor serine tRNA-protein complex is likely to be involved in cotranslational selenocysteine incorporation in human cells 78. It was then obvious that the identification of SLA/LP autoantigen would allow the establishment of a reliable, universally available diagnostic test for AIH but also it would provoke the investigation in the area of autoimmune liver diseases.

Regarding disease specificity, anti-SLA/LP autoantibodies have not been detected in patients with AIH-2, PBC, PSC, chronic viral hepatitis, alcoholic liver disease and non-hepatic autoimmune diseases by standardized ELISAs using reference autoantibody or recombinant antigen 20327379. Ballot et al 32 also showed that these autoantibodies are different from anti-LC1. For these reasons, anti-SLA/LP has been considered as a valuable and specific diagnostic marker of AIH 31327374767779. However, a recent study from the United Kingdom 80 has shown that anti-SLA/LP autoantibodies can also be detected in AIH-2 and in children with PSC. These investigators used eukaryotically expressed tRNP ((Ser) Sec)/SLA as target in a radioligand assay (RLA) which is well known as a more sensitive test than ELISAs and immunoblot due to its ability to identify antibodies directed to conformational epitopes 818283. Their novel findings need confirmation from other research groups and particularly to address whether anti-SLA/LP reactivity is also present in adult PSC. Recent data confirmed the previous finding that patients with anti-SLA/LP display a more severe course of AIH 798084. The latter suggest that anti-SLA/LP may be linked to the pathogenesis of the autoimmune process although the exact function and its role in autoimmunity are so far unclear 1415. From the clinical point of view however, this autoantibody may be helpful in an attempt to reduce the group of cryptogenic hepatitis and/or cirrhosis.

Three types of anti-LKM autoantibodies have been identified 3514151830637285. The LKM type 1 autoantibody (anti-LKM-1) is the characteristic serologic marker for the diagnosis of AIH-2 5186372. These autoantibodies were first described by Rizzetto et al 86, using the IIF method on rodent liver and kidney sections. The characteristic features of anti-LKM-1 autoantibodies are the diffuse staining of cytoplasm of the entire liver lobule and the exclusive staining of the P3 portion of the proximal renal tubules (Fig. 6) 18. Due to this staining pattern of kidney sections anti-LKM-1 can be easily distinguished from AMA, which stain proximal and distal renal tubules (Fig. 7). Western blots with hepatic and renal microsomes revealed a protein band at 50 kDa 51415637287.

The major target-autoantigen of anti-LKM-1 autoantibodies in AIH-2 has been identified as the cytochrome P450 2D6 (CYP2D6) 878889. It has been shown that anti-LKM-1 autoantibodies inhibit the enzymatic activity of CYP2D6 in vitro, but not in vivo 90. Epitope mapping experiments of CYP2D6 autoantigen have defined at least four different linear epitopes 9192. The most immunodominant epitopes of CYP2D6 were amino acids 257–269 and 321–351, which are recognized in about 70% and 50% of AIH-2 cases, respectively 9192. Two infrequent epitopes consisting of amino acids 373–389 and 410–429 are also recognized by anti-LKM-1 in some cases 92. Recently, Klein et al 93 and Kerkar et al 94 reported another immunodominant epitope of CYP2D6 (amino acids 193–212) recognized in about 70% and 93 % of AIH-2 patients, respectively. However, due to failure of inhibition of CYP2D6 enzymatic activity using epitope specific antibodies and since the absorption of the above linear epitopes was unable to absorb inhibitory anti-CYP2D6 autoantibodies, the existence of additional conformational epitopes on CYP2D6 autoantigen has been postulated 95.

It is noteworthy to state here that depending on the geographic origin, 0–7% of patients with chronic hepatitis C – irrespective of the HCV genotype – develop anti-LKM-1 autoantibodies 61443639697. Recently, two studies have shown a higher prevalence of anti-LKM autoantibodies (up to 10%) in a small number of children or adult patients with HCV infection 9899. As stated for AIH-2, CYP2D6 is the major target autoantigen recognized by anti-LKM-1 autoantibodies in HCV patients 1415818283889293949596. However, we and others have failed to document CYP2D6 as the major target autoantigen of anti-LKM antibodies in HCV/anti-LKM positive sera 9899. In addition, recently Miyakawa et al 100 identified CYP2E1 and CYP3A4 as target autoantigens of anti-LKM autoantibodies in two patients with anti-LKM-positive chronic hepatitis C. Taking together, these findings may further indicate the heterogeneous autoimmune reactions that might take place in anti-LKM positive patients with chronic hepatitis C.

The antigenic sites on CYP2D6 autoantigen recognized by anti-LKM-1 autoantibodies are different in AIH-2 and HCV/anti-LKM-1 positive cases 92939495101102103104. For example, the major linear epitope of 257–269 amino acids, as well as the newly reported peptide of 193–212 amino acids are recognized in 70–93% of AIH-2 patients but only in 18–50% of HCV/anti-LKM-1 positive patients 839394101. Additional support to the presence of conformation-dependent anti-LKM-1 autoantibodies in HCV/anti-LKM-1 positive serum samples has emerged from previous studies 99102105. In the latter studies only about 30% of HCV/anti-LKM-1 positive sera reacted with 50 kDa component using Western blot assays, while additional bands at 59 kDa, 70 kDa and 80 kDa were detected 99102105. However, even taking into account the above additional bands, no more than 45% of all sera tested reactive by Western blot. In contrast, a significant proportion of the previous negative sera tested positive for anti-LKM-1 using a specific competitive ELISA, while denaturation of the antigens prior to perform the ELISA resulted in complete loss of the signal 105.

Recently, the development of a more sensitive and specific assay for the detection of anti-LKM-1 autoantibodies was achieved 1415. This novel assay is a quantitative RLA based on immunoprecipitation using ³⁵S-methionine-labelled CYP2D6 antigen obtained by in vitro transcription and translation synthesis 81828399104. Using this assay it was shown that the anti-LKM-1 titers do not differ significantly between AIH-2 and HCV/anti-LKM-1 positive patients 818283. The presence of anti-LKM-1 in some patients with HCV infection led to the proposal for a further division of AIH-2 into AIH-2a (younger, predominantly female patients without evidence of HCV infection) and AIH-2b (older, predominantly male patients with HCV infection) 13106. Nowadays however, after the marked improvements in the reliability and availability of tests for HCV detection such a subdivision of AIH-2 appears unreasonable and tends to be deleted. Actually, HCV/anti-LKM-1 positive patients represent cases of "true" HCV infection with autoimmune features 6107.

From the clinical point of view, screening for anti-LKM autoantibodies is recommended before the initiation of interferon-alpha therapy in HCV patients and if found positive a careful monitoring appears reasonable because occasionally interferon-alpha may unmask, or provoke autoimmune hepatic reactions and even "true"AIH 643104108109110. Dalekos et al 104 studied antibody titers and performed epitope mapping of LKM-1-positive sera from patients with chronic hepatitis C. Interestingly, a patient with a high LKM-1 titer and autoantibodies directed against an epitope of amino acids 257–269, which are preferentially recognized by patients with AIH-2, showed exacerbation of the disease under interferon-alpha treatment. In contrast to other patients with HCV infection, this patient further recognized a rarely detected epitope on the C-terminal third of the protein. These results suggest that determination and monitoring of CYP2D6 autoantibody titers by both IIF and the RLA in combination with epitope mapping of CYP2D6 in HCV/anti-LKM positive patients before the initiation of interferon-alpha treatment, might be helpful in an attempt to identify those patients at risk of developing undesired autoimmune reactions 104.

The mechanism(s) of the development and the pathogenic role of anti-LKM-1 autoantibodies in hepatocellular injury are still unclear. It has been suggested that viral infections by herpes simplex virus (HSV) and related viruses may trigger the autoantibody formation through molecular mimicry in at least some individuals with AIH-2 91. Manns et al 91 tested 26 LKM-positive sera using Western blot with partial sequences of recombinant CYP2D6. Eleven sera recognized a short minimal epitope of eight amino acids with the sequence DPAQPPRD. Twelve other clones recognized a larger epitope containing this eight-amino acid core sequence. The search of electronic databases revealed an interesting match of the minimal epitope with the primary structure of the immediate early protein IE 175 of HSV-1 now known as infected cell protein 4 (ICP4) of HSV-1 (Fig. 8). Sequence identity was present for the PAQPPR sequence. This hypothesis was further supported by a case study in a pair of identical twins 111. In this study, one sister suffered from AIH-2 but the other one was healthy. Interestingly, only the sister suffering from AIH-2 was HSV positive, and her serum recognized the viral protein ICP4 in lysates of HSV-infected cells 111. So far however, overall evidence for mimicry as a driving force of AIH is not convincing.

Besides molecular mimicry, chemical modification of self-proteins and/or immunological cross-reactivity to homologous autoantigens may also provide potential triggers for autoimmune responses. The latter has been suggested by Choudhuri et al 112 who have shown that in AIH-2 patients the linear epitope 321–351 of CYP2D6 cross reacts with amino acids 33–51 of carboxypeptidase-H (the target autoantigen of islet cell autoantibodies in insulin dependent diabetes mellitus), as well as, with amino acids 307–325 of 21-hydroxylase (major target autoantigen in Addison's disease). These findings possibly indicate the presence of a common motif of CYP2D6, carboxypeptidase-H and 21-hydroxylase, which may contribute through a cross reactive immune response to the development of multiple endocrinopathies in the course of AIH-2. Additional support to this hypothesis emerged from two recent studies by Kerkar et al 94 and Bogdanos et al 113. In the first study the authors were able to show the similarity and cross-reactivity between the immunodominant epitope 193–212 of CYP2D6 and homologues of two unrelated viruses (HCV 2977–2996 and CMV 121–140) 94. In the second study the researchers investigated whether the immunodominant epitope 252–271 of CYP2D6 in anti-LKM-1 positive AIH-2 and homologues from the NS5B and E1 proteins of the HCV polyprotein and the ICP4 of HSV-1 are targets of humoral immune response in anti-LKM-1 positive and anti-LKM-1 negative HCV infected patients and furthermore whether this response is cross-reactive 113. The hypothesis of molecular mimicry and cross-reactivity in LKM-1 production has not been addressed experimentally. The authors for the first time gave experimental support to the notion that molecular similarity between CYP2D6, HCV and HSV can result in LKM-1 production via a cross-reactive response in genetically susceptible individuals (interestingly only the HCV positive/LKM-1 positive patients with viral/self cross-reactivity possessed the HLA B51 allotype) 113. Taking together, the above studies suggest that multiple exposure to viruses mimicking self may represent an important pathway to the development of autoimmunity 94113.

Two possible mechanisms have been proposed for the involvement of anti-LKM-1 autoantibodies in the pathogenesis of liver injury. The first appears to be a direct binding of these autoantibodies to hepatocytes, leading to lysis of liver cells, while the second is associated with an anti-LKM-1 induction of activating liver-infiltrating T lymphocytes, which indicates the combination of B and T cell activity in the autoimmune process involved 114115116117. A prerequisite for both anti-LKM-1 production and the activation of pathogenetic mechanisms involved in liver injury, is the expression of CYP2D6 on the surface of the patients' hepatocytes. Under this context, Ma et al 118 showed that key residues of a major CYP2D6 epitope (316–327) are exposed on the surface of the molecule and may represent key targets for anti-CYP2D6 production. In addition, recent data provides convincing evidence that anti-LKM-1 autoantibodies recognize CYP2D6 exposed on the plasma membrane of hepatocytes from either AIH-2 or HCV/anti-LKM-1 positive patients 114115 suggesting a pathogenetic role for these autoantibodies in hepatic tissue damage either in AIH-2 or in some cases of HCV/anti-LKM-1 positive patients 104109110115.

So far, anti-LKM type 2 autoantibodies (anti-LKM-2) have been detected only in some cases of drug-induced hepatitis caused by tienilic acid 1463. The target autoantigen of anti-LKM-2 has been documented as the CYP2C9 85. A proposed mechanism for the induction of anti-LKM-2 could be the binding of an active metabolite of the drug to the CYP2C9 protein, which then becomes antigenic 14637285.

Anti-LKM type 3 autoantibodies (anti-LKM-3) alone or in combination with anti-LKM-1 are also detected in about 5–10% of patients with AIH-2 16119. In contrast to anti-LKM-1 and anti-LKM-2 autoantibodies, which on immunofluorescence stain liver and kidney tissues only, with anti-LKM-3 additional fluorescence signals may be present with tissue from the pancreas, adrenal gland, thyroid, and stomach. Family 1 of UDP-glycuronosyltransferases (UGT1) is the main target autoantigen of anti-LKM-3 autoantibodies (molecular weight of 55 kDa) 119120. These autoantibodies were first described in about 13% of patients with chronic hepatitis D, but not in patients with chronic hepatitis B or C 121. However, three recent reports have shown the presence of anti-LKM-3 autoantibodies in some patients with HCV infection 99122123. These findings may further support the heterogeneous phenomenon of the HCV-induced autoimmunity.

In 1988 a second autoantibody marker of AIH-2 was recognized 19. This autoantibody was found to react to a liver cytosolic protein. The autoantibody is organ specific but not species specific and was therefore called anti-LC1 19. The anti-LC1 autoantibodies are characterized by a cytoplasmic staining of the periportal hepatocytes when the IIF assay is used for their detection. The hepatocellular layer around the central veins is not stained 19124. These findings indicate that the target autoantigen of anti-LC1 autoantibodies is not uniformly distributed in rodent liver tissues. They are detected in about 30% of patients with AIH-2 19124 and in approximately 50% of all anti-LKM-1 positive cases 125. It is noteworthy that the anti-LC1 autoantibodies proved to be the only serological marker in 10% of patients with AIH 19.

The detection of anti-LC1 autoantibodies by IIF is obscured due to the anti-LKM-1 pattern that frequently found in most of the anti-LC1 positive sera. For these reasons other techniques such as, the ouchterlony double diffusion, immunoblot or counter-immunoelectrophoresis are required for their detection 19124125126. By immunoblotting, anti-LC1 positive serum samples recognize a liver specific cytosolic protein of 58–62 kDa 124125126. Recently the molecular target of anti-LC1 was identified as the formiminotransferase cyclodeaminase (FTCD) 127, which is a polymeric bifunctional enzyme involved in folate metabolism. However, another group demonstrated the arginninosuccinate lyase (ASL) as the target autoantigen of a weak precipitin line detected by the ouchterlony double diffusion assay in patients with autoimmune or viral hepatitis 128.

Anti-LC1 autoantibodies have been proposed as a more specific marker of AIH-2 than anti-LKM-1 autoantibodies, since in the original reports their presence was never associated with HCV infection 19124. However, a recent study by Lenzi et al 125 confirmed the above aspect only in the pediatric subset of their patients, while a substantial proportion of the adults with anti-LC1 autoantibodies had also markers of HCV infection. The significance of the association between anti-LC1 autoantibodies and HCV infection remains uncertain and has to be established 106129. In contrast to what has been found for anti-LKM-1, the titers of anti-LC1 autoantibodies appear to parallel with disease activity 130. The latter may indicate a possible involvement of anti-LC1 in the pathogenesis of AIH-2. However, the clinical significance of anti-LC1 is not yet completely defined.