Prevalence and clinical significance of anticardiolipin antibodies in patients with type 1 autoimmune hepatitis

Journal of Autoimmunity 24 (2005) 251e260 www.elsevier.com/locate/issn/08968411

Prevalence and clinical significance of anticardiolipin antibodies in patients with type 1 autoimmune hepatitis* Christos Liaskosa, Eirini Rigopouloub, Kalliopi Zachoua,b, Sarah Georgiadoua, Nikolaos Gatselisa, Roidoula Papamihalic, George N. Dalekosa,b,) a

Research Laboratory of Internal Medicine, Department of Medicine, Medical School, University of Thessaly, Larissa, Greece b Academic Liver Unit, Department of Medicine, Medical School, University of Thessaly, Larissa, Greece c Department of Pathology, Medical School, University of Thessaly, Larissa, Greece Received 31 October 2004; revised 28 December 2004; accepted 5 January 2005

Abstract There are few case reports on the association between autoimmune hepatitis (AIH) and anticardiolipin antibodies (anti-CLAbs) and/or antiphospholipid syndrome (APLS). We studied the anti-CLAbs prevalence in AIH and other hepatic diseases. We also investigated whether anti-CLAbs are co-factor dependent and which is their avidity since co-factor dependency or increased resistance is associated with APLS. Fifty-nine AIH patients, 228 HCV, 50 HBV, 123 with other non-viral and non-autoimmune liver disorders (nV-nALD) and 267 healthy people were investigated for anti-CLAbs and antibodies against beta-2-glycoprotein I (antib2-GPI). Resistance of IgG anti-CLAbs was evaluated using 2 M urea. IgG anti-CLAbs detected in 39% of AIH, 19.7% of HCV ( pZ0.006), 14% of HBV ( pZ0.01), 8.1% of nV-nALD ( pZ0.000) and 1.1% of healthy ( pZ0.000). IgG anti-CLAbs were associated with the presence of cirrhosis and active AIH while their resistance to urea was high. Anti-b2-GPI was detected in two AIH patients. We demonstrated a significantly higher prevalence of anti-CLAbs in patients with AIH compared to other diseases and healthy people. Anti-CLAbs were associated with AIH stage but no association was found with APLS clinical manifestations (thrombosis, pregnancy morbidity, thrombocytopenia). However, their avidity was comparable with that of APLS indicating the need for prospective studies in order to address whether anti-CLAbs in AIH may contribute to the progression of liver disease or APLS development. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Anti-cardiolipin antibodies; Anti-phospholipid antibodies; Autoimmune hepatitis; Antiphospholipid syndrome; Viral hepatitis

1. Introduction

*

This article has been presented as an abstract in the Falk Symposium on Autoimmune Liver Disease, October 12e13, 2004, Freiburg, Germany. Research Committee of the University of Thessaly and Galenica A.E. have partially supported this work. ) Corresponding author. Academic Liver Unit and Research Laboratory of Internal Medicine, Medical School, University of Thessaly, Papakiriazi 22 str., 41222 Larissa, Greece. Tel.: C30 41 0 565251; fax: C30 41 0 565250. E-mail address: [email protected] (G.N. Dalekos). 0896-8411/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.jaut.2005.01.016

Autoimmune hepatitis (AIH) is a disease of unknown aetiology, which is predominant among women and is characterised by hypergammaglobulinemia, characteristic autoantibodies, association with human leukocyte antigens DR3 or DR4 and a favourable response to immunosuppression [1e3]. Specific features of the disease are its associations with several extrahepatic immune-mediated diseases and/or syndromes [1,4e8] However, only very few case reports on the association between AIH and the presence of antiphospholipid

252

C. Liaskos et al. / Journal of Autoimmunity 24 (2005) 251e260

antibodies (anti-PLAbs) and/or antiphospholipid syndrome (APLS) have been reported so far [9e18]. APLS is an autoimmune disorder characterised by arterial or venous thrombosis and obstetric complications in association with the presence of anti-PLAbs [19e21]. Anti-PLAbs are a group of autoantibodies directed against complexes of negatively charged phospholipids and plasma proteins (co-factors) [19e23]. According to the recently established criteria, only anticardiolipin antibodies (anti-CLAbs) or lupus anticoagulant (LA), are considered as laboratory criteria for anti-PLAbs positivity in APLS [19]. It seems that cofactors such as beta-2-glycoprotein I (b2-GPI) must be present for the anti-PLAbs to bind to phospholipids in APLS patients (co-factor dependent, ‘‘autoimmune’’, or ‘‘thrombogenic’’) [19e22,24]. In several other chronic inflammatory conditions, such as ulcerative colitis or infectious diseases anti-CLAbs seem to be of nonpathogenic origin (co-factor independent or ‘‘nonthrombogenic’’) [23e29]. In this context it is rather clear that only ‘‘autoimmune’’ anti-PLAbs are correlated with APLS [19,20,23e27]. We studied the prevalence and the clinical significance of IgG and IgM isotype-specific anti-CLAbs in a consecutive series of AIH patients and in patients with other hepatic diseases, as well as in healthy people, since so far, data on anti-CLAbs with or without the APLS features is missing in AIH. In order to address whether anti-CLAbs are ‘‘pathogenic’’ (‘‘thrombogenic’’) or not we investigated their co-factor dependency and avidity (low or high resistance of anti-CLAbs binding to dissociating agents such as urea) since co-factor dependency or increased resistance of anti-CLAbs to urea is associated with the occurrence of APLS features [24,30].

2. Materials and methods 2.1. Subjects We investigated 59 patients (47 females; median age 62 years, range 6.5e78 years) with AIH (51 with AIH type 1 (AIH-1), two with AIH type 2 and six with AIH1/primary biliary cirrhosis (PBC) overlap syndrome) for the presence of IgG and IgM anti-CLAbs and for IgG and IgM antibodies against b2-GPI (anti-b2-GPIAbs). The diagnosis of AIH in the 53 patients was based on the revised descriptive criteria for diagnosis of AIH reported by the International Autoimmune Hepatitis Group [7]. These criteria include all of the following: (a) compatible liver histology (interface hepatitis with or without lobular hepatitis); (b) serum biochemistry (any abnormality in serum aminotransferases); (c) serum immunoglobulins (total serum globulin or g-globulins or IgG concentrations greater than 1.5 times the upper

normal limit); (d) serum autoantibodies (seropositivity for antinuclear antibodies (ANA), smooth muscle autoantibodies (SMA) or liver-kidney microsomal antibodies type 1 (anti-LKM-1) and seronegativity for antimitochondrial antibodies (AMA)); (e) absence of viral markers (seronegativity for markers of hepatitis A, B and C viruses); (f) absence of other aetiological factors such as alcohol consumption or use of known hepatotoxic drugs [7]. Two patients were inactive carriers of hepatitis B virus (HBV) with repeatedly undetectable serum HBV-DNA, while they had increased levels of aminotransferases, hypergammaglobulinaemia, detectable ANA and SMA in high titres and a liver biopsy compatible with AIH. In addition, 42 patients with AIH had pre-treatment scoring system above 15 (meanGSD; 17.1G0.9), while 11 patients who were under immunosuppression at the time of investigation had scoring above 17 (19.2G1.1) [7]. Patients with AIH-1/PBC overlap syndrome were considered according to the report by Chazouilleres et al. [31]. According to the latter study a patient with AIH/PBC overlap syndrome had to fulfil at least two of the three criteria for PBC (alkaline phosphatase (ALP) above twice the upper normal or g-glutamyl-transpeptidase (g-GT)>5 times upper normal; positive AMA; florid bile duct lesions on liver biopsy) as well as at least two of the following three for AIH (alaninoaminotransferase (ALT)>5 times upper normal; serum IgG above twice the upper normal or positive SMA; histologic lesions compatible with AIH) [31,32]. At the time of investigation 48 patients had not received any kind of immunosuppression in the past. Forty of them were tested at the time of diagnosis before the beginning of the treatment while eight patients had already inactive cirrhosis at diagnosis and immunosuppression was not indicated. The remaining 11 patients (Table 1) were under immunosuppression at the time of investigation (treatment duration 29.7G70 months). Six out of 11 patients were under prednizone only (15 mg/ daily; treatment duration 44.9G95.8 months) and five were under prednizone (7.5e10 mg/daily) plus azathioprine (1.5e2 mg/kg daily) (treatment duration 9.8G 6.3 months). The characteristics of AIH patients are given in detail in Table 1. According to our previous publications activity of the disease was defined by the presence of one or more of the following general symptoms: malaise, fatigue, arthralgias, jaundice, nausea, anorexia and weight loss irrespective of the presence of elevated aminotransferases [33,34]. The disease control groups consisted of 228 patients with chronic hepatitis C (HCV; 124 female; median age 59 years, range 17e83 years) including 163 patients with evidence of autoimmunity (152 seropositive for ANA and/or SMA and 11 patients with extrahepatic autoimmune manifestations but without autoantibody

C. Liaskos et al. / Journal of Autoimmunity 24 (2005) 251e260 Table 1 Characteristics of patients with autoimmune hepatitis at the time of the collection of serum samples AIH (nZ59) Disease duration (months)a Alcohol abuse (yes/no) Cirrhosis (yes/no) Other autoimmune diseases (yes/no) Antinuclear antibodies (pos/neg) Antimitochondrial antibodies (pos/neg) Smooth muscle antibodies (pos/neg) Liver-kidney-microsomal antibodies (pos/neg) Antibodies to soluble liver antigens (pos/neg) Atypical p-ANCA or ANNA (pos/neg) Atypical c-ANCA (pos/neg) Anti-ds-DNA (pos/neg) Hepatitis B surface antigen (pos/neg) Antibodies to hepatitis B surface antigen (pos/neg) Antibodies to hepatitis B core antigen (pos/neg) Hepatitis B e antigen (pos/neg) Antibodies to hepatitis B e antigen (pos/neg) Thrombosis or obstetric complicationsb (yes/no) Thrombocytopeniac (yes/no) Serum immunoglobulin G (mg/dl) Serum immunoglobulin M (mg/dl) Serum immunoglobulin A (mg/dl) AST (number of upper normal) ALT (number of upper normal) g-GT (number of upper normal) Alkaline phosphatase (ALP; normal limit 306 U/l) g-globulins (g/dl) Active disease (yes/no)d Immunosuppression (yes/no)e

41.4G56 11/48 26/33 20/39 55/4 6/53 47/12 2/57 5/54 30/29 18/41 22/37 2/57 26/33 32/27 0/59 10/49 4/55 16/43 2099G949 218G160 376G322 4.49G11.8 4.16G4.55 3.04G2.8 153G92 3.73G1.10 22/37 11/48

ANCA, anti-neutrophil cytoplasmic antibodies; ANNA, antineutrophil nuclear antibodies; AST, aspartate aminotransferase, normal upper limit 40 U/l; ALT, alaninoaminotransferase, normal upper limit 40 U/l, g-GT, g-glutamyl-transpeptidase, normal upper limit 37 U/l. All patients tested negative for antibodies against hepatitis C virus. a Duration of the disease was calculated from the date of the onset of the disease to the date of the collection of the sera. b One or more unexplained deaths of a morphologically normal foetus at or beyond the 10th week of gestation, or one or more premature births of a morphologically normal neonate at or before the 34th week of gestation due to severe pre-eclampsia or placental insufficiency, or three or more unexplained consecutive spontaneous abortions before the 10th week of gestation. c Platelet count less than 140,000/ml. d General symptoms, malaise, fatigue, arthralgias, jaundice, etc. with or without biochemical activity. e Investigation was done at presentation in 40 patients. Data are given as meanGstandard deviation.

seropositivity; 11/152 HCV patients with detectable autoantibodies had also concurrent extrahepatic autoimmune disorders), 50 patients with chronic hepatitis B (20 female; median age 57 years, range 18e71 years) and 123 patients (55 female; median age 56 years, range 20e 71) with other non-viral and non-autoimmune liver disorders (n-V and n-ALD). The latter group consisted of 24 with alcoholic liver disease, 44 with non-alcoholic steatohepatitis and 55 with undefined causes of elevated liver enzymes. The healthy controls consisted of 267 blood donors (117 female; median age 48 years, range

253

18e60 years). All donors tested negative for hepatitis B surface antigen (HBsAg), antibodies against HCV (antiHCV) and antibodies against human immunodeficiency virus (anti-HIV). Alaninoaminotransferase (ALT) values were within normal limits in all of them (26.5G 5.3 U/l). In every patient, prior history of arteriovenous thrombotic events (myocardial infraction, ischaemic stroke, pulmonary emboli, deep venous thrombosis of legs, arms or internal organs), obstetric complications and neuropsychiatric disorders (convulsions, migraine, memory loss, chorea or psychosis) were recorded. The medical history (presence or absence of symptoms at the time of diagnosis, alcohol abuse, extrahepatic manifestations, being under treatment or not) of AIH patients (Table 1) and of the disease control patients was taken into account (data not shown). Twenty AIH patients suffered from extrahepatic manifestations including Hashimoto thyroiditis, Grave’s disease, systemic lupus erythematosus (SLE), rheumatoid arthritis, etc. Histological data were available in all AIH patients included in the study, in 129 HCV, 30 HBV and 26 n-V and n-ALD. For statistical reasons, patients were divided into two groups according to the clinical evaluation and the histological data (where available): (i) those with cirrhosis and (ii) those without cirrhosis (Table 1). Serum levels of aspartate aminotransferase (AST), ALT, g-GT, ALP, g-globulins and platelets were determined using standard techniques. Similar to previous studies, patients with platelet count less than 140,000/mm3 were considered thrombocytopenic [27, 28]. Serological markers of HBV infection were determined using commercial enzyme immunoassays (Abbott GmbH, Wiesbaden-Delkenheim, Germany). Anti-HCV was determined by a third generation enzyme immunoassay (Murex Diagnostics Ltd, Central Road Temple Hill, UK). All subjects consented to participate in the study at the time of interview. The ethical committee of the University Hospital of Larissa approved the study protocol.

2.2. Detection of IgG and IgM anticardiolipin antibodies (anti-CLAbs) Coded sera were stored at
30  C until tested. Serum samples from the patients in whom we had histological data were selected on the basis of their close proximity to the liver biopsy (0e6 months). The levels of IgG and IgM isotype specific anti-CLAbs were determined by a quantitative solid phase enzyme-linked immunosorbent assay (ELISA) as described previously [27e 29,35,36]. Positive sera from patients with AIH were further tested by a commercially available ELISA (IgG and IgM anti-CLAbs, INOVA Diagnostics).

C. Liaskos et al. / Journal of Autoimmunity 24 (2005) 251e260

The specificity, reproducibility and optimal conditions of the assay have been determined in extensive preliminary experiments as described previously [27e 29,35e37]. In each assay, the between day variation of the OD values was eliminated by running serial dilutions of a positive control serum (standard curve) on each plate, as described [27e29,35e37]. Briefly, a standard curve was constructed by assaying repeatedly control sera positive for IgG anti-CLAbs in serial dilutions from 1:50 to 1: 6400. The OD value of 1:6400 dilution was arbitrarily chosen as 1 binding unit (BU). The BU values for test samples were calculated according to this curve on the plate. This was accomplished by dividing the OD of each sample tested by the OD value, which corresponded to 1 BU for that plate. Finally, the results were expressed as binding index (BI) calculated by dividing the BU of every sample by the mean BU value for the healthy control group plus four standard deviations (SD), multiplied by 100. According to this formula, a BI of 100 was defined as the cut-off point. The adoption of this stringent cut-off precludes the possibility of false positive results. According to our previous studies [26e29], a BI O200 was defined as a high positive result for anti-CLAbs, while a BI between 100 and 150 was defined as a low and a BI between 150 and 200 as a medium. All experiments were done in triplicate and all patients tested for IgG antiCLAbs in at least two subsequent samples, 6 weeks apart.

2.3. Urea resistance anti-CLAbs ELISA Urea is a chaotropic agent that interferes between antigen and antibody [38]. Therefore, ‘‘urea resistance’’ is a descriptive term indicating that the antibodies recognize the antigen in the presence of urea, because the antibody, the antigen, or the antigen-antibody interactions are resistant to urea. The ability of the IgG anti-CLAbs to recognize their antigen in the presence of urea was evaluated by ELISA, as described [30]. To set up the assay, sera with high anti-CLAbs titres tested in serial twofold dilutions to obtain a standard curve. After the incubation of sera, 50 ml of 0 M, 2 M and 4 M urea were added for 10 min at room temperature [24,30]. Thereafter, the plates were washed five times with PBS and the remaining ELISA steps were carried out as described above (Fig. 1). In order to ensure that urea treatment does not affect the binding of the antigen on the ELISA plate or does not alter significantly the cardiolipin structure, two additional experiments were done by adding urea after the overnight incubation of cardiolipin as well as after the blocking step with BSA (actually the addition of urea in these two experiments was done before the addition of the sera). It was shown that sera gave the

1400

1200

1000

800

0M UREA

OD

254

2M UREA 4M UREA

600

400

200

0 50

100

200

400

800 1600 3200

DILUTION Fig. 1. Experiments for the determination of the avidity of a high-titre IgG anti-CLAb from an APLS patient in serial dilutions (1:50 to 1:3200) after treatment with various molarities of urea. Dissociation of the antibodies from the antigen was observed and was higher with increasing molarities of urea. OD, optical density.

same reactivity in non-treated and treated with urea plates (data not shown). All IgG anti-CLAbs positive sera of AIH patients, and three from well-defined APLS patients with high titre IgG anti-CLAbs were tested in serial dilutions in duplicate series at 0 M and 2 M urea and the dilutions of untreated and dissociated complex were compared at a certain end point (fixed OD: 0.150) as described [30]. This OD value was chosen because it was above the cutoff limit of the assay (OD: 0.100) and simultaneously enabled us to evaluate the urea resistance even from sera with low reactivity in our in-house anti-CL ELISA. According to previous studies [24,30], in order to quantitate the avidity data, the following strategy was carried out. The serum dilution after urea treatment expressed as percentage of the serum dilution without treatment and corresponding to the same OD value (0.150 absorbance unit), was defined as ‘residual activity’ (RA) [24,30]. Thus, the higher the RA, the higher the reactivity of antibodies to the antigen. 2.4. Inhibition experiments The specificity of the assay for IgG anti-CLAbs, was determined in preliminary inhibition experiments, as described [27e29,35e37]. Serum samples at a dilution of 1:25 in PBS were mixed with equal volumes of cardiolipin (concentration 500 mg/ml) and doublestranded DNA (ds-DNA) (concentration 500 mg/ml). The two mixtures of each serum (final dilution 1:50)

255

C. Liaskos et al. / Journal of Autoimmunity 24 (2005) 251e260 800

2.6. Detection of other autoantibodies

700

ANA, SMA, AMA, anti-LKM and anti-neutrophil cytoplasmic antibodies were detected by indirect immunofluorescence using standard protocols and IgG antibodies to soluble liver antigens/liver pancreas (antiSLA/LP) using commercial ELISA (Euroimmun AG, Lubeck, Deutschland). IgG anti-ds-DNA antibodies were determined by an in-house quantitative ELISA as we described [29,35,37].

600

OD

500 400

no inhibitor with cardiolipin with ds-DNA

300

2.7. Statistical analysis 200

Results are expressed as meanG(SD). Data were analysed by an unpaired t-test, ManneWhitney U-test (MWU), c2 (two-by-two with Yate’s correction), Fisher’s exact test and Spearman’s rank correlation (r) where applicable. Two-sided p values less than 0.05 were considered as statistically significant. Confidence intervals (95% CI) were determined using the formula PZpG1.96( pq/n)1/2 where p is the frequency, q is 1
p and n is the number of individuals tested from each group. The Documenta Geigy Scientific Table (Ciba-Geigy Ltd, Basle, Switzerland, 1972, 7th edition) was used for 95% CI when fewer than 41 individuals were tested.

100 0

1/50 1/100 1/200 1/400 1/800 1/1600 1/3200

Dilution Fig. 2. Inhibition experiments using serial dilutions (1/50e1/3200) of a high titre IgG anti-CLAb serum from a patient with overlap autoimmune hepatitis/primary biliary cirrhosis. OD values are progressively decreased in uninhibited IgG anti-CLAbs positive serum (solid line) or after incubation with ds-DNA (broken line); OD values disappeared from the 1/50 dilution of the IgG anti-CLAbs positive serum when incubation was done with cardiolipin (dot line). OD, optical density.

3. Results were incubated for 1 h at room temperature and subsequently overnight at 4  C. Uninhibited samples containing equal volumes of serum and PBS without inhibitor were also incubated for 1 h at room temperature and subsequently overnight at 4  C [36]. Using our in-house ELISA, we determined the OD of each mixture, in serial dilutions ranging from 1:50 to 1:3200 to confirm the specificity of the assay (Fig. 2).

3.1. AIH patients A significant proportion of patients had detectable IgG and/or IgM anti-CLAbs (50.8%; 95% CI: 38e 63.6%). The respective prevalence of anti-CLAbs according to the isotype was 39% (95% CI: 26.6e 51.5%) for IgG (11 patients with low, seven with medium and five with high titres; 176.4G73 BI) and 23.7% (95% CI: 12.9e34.5%) for IgM (eight patients with low, four with medium and two with high titres; 151.9G33.6 BI) (Table 2). Nineteen out of 23 IgG antiCLAbs positive patients were consistently positive in at least two subsequent testings, 6 weeks apart for IgG antiCLAbs, including eight patients who have maintained their titres, eight with an increase and three with a decrease of the titre. All but two positive sera by the in-house ELISA (one patient’s serum from the IgG and

2.5. Detection of IgG and IgM anti-b2-GPI antibodies (anti-b2-GPIAbs) Anti-b2-GPIAbs were determined using a commercial ELISA according to the manufacturer’s instructions (QUANTA LiteÔ b2-GPI IgG and IgM, INOVA Diagnostics) [39]. Results were reported in standard IgG or IgM anti-b2-GPI units (SGU). Values greater than 20 SGU were considered positive.

Table 2 Prevalence of anticardiolipin (anti-CLAbs) and anti-b2-GPI antibodies (anti-b2-GPIAbs) in the groups of patients studied IgG anti-CLAbs AIH (%) HCV infection (%) HBV infection (%) Other diseases of the livera (%)

23/59 45/228 7/50 10/123

,

,

(39)* ** *** (19.7)* (14)** (8.1)***

IgM anti-CLAbs 14/59 48/228 6/50 2/123

*pZ0.003; **pZ0.007; ***pZ0.000; xpZ0.000; xxp!0.02; #pZ0.000; a Patients with non-viral no-autoimmune diseases of the liver.

##

x

(23.7) (21.3) (12) (1.6)x

IgG and/or IgM anti-CLAbs 30/59 77/228 13/50 11/123

xx,#,##

(50.8) (34.1)## (26)xx (9.8)#

pZ0.02 (two by two c2 after Yates’ correction).

IgG anti-b2-GPI Abs 2/59 (3.4) 7/228 (3.1) 1/50 (2) 0

256

C. Liaskos et al. / Journal of Autoimmunity 24 (2005) 251e260

one from the IgM positive group), tested also positive by the commercial ELISA kit. In contrast, only six healthy donors (2.25%; 95% CI: 0.5e4%) tested positive for IgG and/or IgM a-CL Abs; three positive for IgG and three for IgM isotype (1.12% for each; 95% CI: 0.13e2.37%). The above frequencies of anti-CLAbs were significantly higher ( pZ0.000 for each comparison) in AIH patients (50.8%, 39% and 23.7%) compared with those detected in healthy people (2.25%, 1.12% and 1.12%). Only 2/59 of AIH patients (3.4%; 95% CI: 0e4.5%) tested IgG anti-b2-GPIAbs positive (one also IgG antiCLAbs positive), but both of them without any clinical or laboratory evidence suggestive of APLS. None of the healthy people tested positive for IgG or IgM anti-b2GPI Abs ( pZ0.000). Positivity for IgG, IgM or for at least one antiCLAbs-isotype was not associated with the age, sex, alcohol abuse, disease duration and any current immunosuppressive treatment for AIH (Table 3a). Anti-CLAbs detection (frequency of positivity or subgroup of patients according to the low-medium-high titre) was not associated with reported thrombotic events, foetal losses in the past, the presence of extrahepatic autoimmune diseases (or in any particular extrahepatic disorder or in total) (Table 3a) or with the presence of thrombocytopenia (Table 3b). In addition, neither positivity nor the titre of other autoantibodies tested or the presence of markers of past HBV infection was associated with positivity for any isotype of the antiCLAbs in AIH patients (data not shown). The absence of associations in all of the above remained even when the six patients with AIH-1/primary biliary cirrhosis were excluded from the analysis (data not shown). Positivity for IgG or for at least one isotype of antiCLAbs (IgG and/or IgM) was associated with the presence of cirrhosis ( p!0.02 and p!0.10, respectively), clinically active disease ( pZ0.03 and pZ0.004, respec-

tively) (Table 3a) and biochemical activity (significantly increased mean levels of AST and ALT in anti-CLAbs positive patients compared to negative; see also Table 3b). Similar significant findings were found after the exclusion of the six patients with AIH-1/primary biliary cirrhosis overlap syndrome (data not shown). 3.2. Disease control groups Taken together, 101/401 patients (26%; 95% CI: 21.6e30.2%) with HCV, HBV and n-V and n-ALD had detectable IgG and/or IgM anti-CLAbs (statistically significant difference compared to AIH; 26% vs 50.8%; pZ0.000). In more detail, 62/401 patients (15%; 95% CI: 11.5e18.5%) tested IgG anti-CLAbs positive (statistically significant difference compared to AIH; 15% vs 39%; pZ0.000). Prevalence of IgG anti-CLAbs was significantly lower in each group of patients with HCV, HBV or n-V and nALD compared to AIH (Table 2). IgG anti-CLAbs were detected significantly more frequently in AIH patients compared with that observed in the total group of patients with chronic viral hepatitis (24/59; 39% vs 52/ 278; 18.7%, pZ0.002). Positivity for IgG, IgM and IgG and/or IgM antiCLAbs was significantly associated with the presence of cirrhosis ( pZ0.025, pZ0.004, and pZ0.004, respectively) and decreased platelets count ( pZ0.016, pZ0.011, and pZ0.004, respectively) in HCV patients. In the group of HBV patients, IgG and/or IgM anti-CLAbs were associated with lower platelet counts (167.9G 71.7!103/ml in anti-CLAbs positive; nZ13 vs 218G 59.4!103/ml in anti-CLAbs negative; nZ37; p!0.05), as well as the presence of thrombocytopenia (5/13 antiCLAbs positive patients had less than 140,000 platelets/ ml vs 3/37 anti-CLAbs negative patients; p!0.05). IgG anti-b2-GPIAbs were positive in only seven HCV

Table 3a Demographic and clinical characteristics of patients with AIH according to the presence of a positive or negative anticardiolipin test IgG

Age (years) Sex (M/F) Disease duration (months) Alcohol abuse (yes/no) Other autoimmune disease (yes/no) Thrombosis or foetal losses (yes/no) Cirrhosis (yes/no) Active disease (yes/no)a Immunosuppression (yes/no) a

IgM

IgG and/or IgM

Anti-CL pos (nZ23)

Anti-CL neg (nZ36)

Anti-CL pos (nZ14)

Anti-CL neg (nZ45)

Anti-CL pos (nZ30)

Anti-CL neg (nZ29)

53.1G18.8 5/18 31.5G50.9 3/20 6/17

56.3G17.9 7/29 47.7G59 8/28 14/22

58.3G17.3 3/11 21.9G36.4 4/10 5/9

54G18.5 9/36 47.5G60 7/38 15/30

54.8G18.5 6/24 33.6G48.9 6/24 8/22

55.3G18.1 6/23 49.5G62.5 5/24 12/17

1/22

3/33

1/13

3/42

2/28

2/27

15/8x 13/10*

11/25x 9/27*

8/6 7/7

18/27 15/30

17/13xx 17/13**

9/20xx 5/24**

2/21

9/27

3/11

8/37

4/26

7/22

General symptoms, malaise, fatigue, arthralgias, jaundiced, etc. with or without biochemical activity. Data are given as meanGSD; nZnumber of patients tested; M, male; F, female; xp!0.02; xxp!0.10; *pZ0.03; **pZ0.004 (c2 test).

257

C. Liaskos et al. / Journal of Autoimmunity 24 (2005) 251e260 Table 3b Laboratory characteristics of patients with AIH according to the presence of a positive or negative anticardiolipin test IgG

AST (U/l) ALT (U/l) g-GT (U/l) ALP (U/l) g-globulins (g/dl) Thrombocytopeniaa (no/yes)

IgM

IgG and/or IgM

Anti-CL pos (nZ23)

Anti-CL neg (nZ36)

Anti-CL pos (nZ14)

Anti-CL neg (nZ45)

Anti-CL pos (nZ30)

Anti-CL neg (nZ29)

313.8G669# 265.7G524*** 146.6G141 190.7G91.8* 3.91G1.1 15/8

94.4G261.9# 102.3G222*** 90.8G134 129G84.3* 3.62G1.12 28/8

223.9G558.8 239.7G558.5 165.2G168.3 179.1G106 4.14G1.03 10/4

166.3G449.3 143.6G301 96.1G125 144.9G86.4 3.6G1.1 33/12

253.7G593.7x 220.1G464.3*** 141G139.6 189.1G102.2*** 4G1.08** 20/10

103.7G291.7x 110.8G247.1*** 83.1G132.6 115.7G61.4*** 3.46G1.10** 23/6

Abbreviations are the same as in text. Data are given as meanGSD. nZnumber of patients tested. *pZ0.01, **p!0.10 (unpaired t-test). ***p!0.05, pZ0.006 and #pZ0.003 (ManneWhitney U-test). a Platelet count less than 140,000/mm3 (see also [27,28]).

x

patients (3%; 95% CI: 0.8e5.2%) and one female HBV patient (2%; 95% CI: 0e5.8%), who was also positive for IgG anti-CLAbs with thrombocytopenia, but without a history of thrombotic events or recurrent foetal losses in the past. All disease controls patients tested negative for IgM anti-b2-GPIAbs. 3.3. IgG anti-CLAbs reactivity in the presence of urea Twenty-three AIH patients, positive for IgG antiCLAbs were tested in serial dilutions, in duplicate series at 0 M and 2 M urea. The resistance to dissociation of the IgG anti-CLAbs from the respective antigen is illustrated by the RA (%) values (Fig. 3). RA values in AIH patients positive for IgG anti-CLAbs were 86.3G10.1 which were similar with that observed in APLS patients (Fig. 3). 100,0

90,0

RA %

80,0

70,0

60,0

50,0 AIH

APLS

DISEASE Fig. 3. The residual activity (RA%) of the sera from patients with autoimmune hepatitis (AIH; nZ23) and antiphospholipid syndrome (APLS; nZ3) after treatment with 2 M urea.

4. Discussion To the best of our knowledge, this is the first study, which showed that half of AIH patients tested positive for IgG and/or IgM anti-CLAbs. Our results are not in agreement with the findings of a previous study by de Larranaga et al. [40], who reported a positivity of anti-PLAbs in only 3% of patients with autoimmune liver diseases. These differences could be attributed to the laboratory methods used and the design of the studies [23]. So far, AIH has been associated with the presence of anti-CLAbs only in case reports [9e18]. In these cases, anti-CLAbs were associated with the presence of APLS. The present study failed to find any association between anti-CLAbs and the presence of APLS features in AIH patients. However, the detection of IgG anti-CLAbs in AIH patients was associated with more severe disease as attested by the presence of active disease and cirrhosis more frequently in patients with a positive anti-CLAb test. These findings suggest anti-CLAbs as a potentially additional marker of disease severity and activity. The latter has already been proposed for primary sclerosing cholangitis (PSC) [41]. However, anti-CLAbs were also correlated with the presence of cirrhosis in HCV, a virus which is characterised by an increased association with autoimmune responses [23,26e28]. Taken together, these results could be attributed to a consequence of the liver disease process (viral or autoimmune). However, such an explanation seems unlikely, as the disease duration did not significantly differ between anti-CLAbs positive and negative AIH patients. In AIH, neither autoantibody titres of non-organ specific autoantibodies at first diagnosis nor autoantibody behaviour in the time course of the disease are prognostic markers for the disease [1,2,7,42]. Antibodies to the liver-related antigens such as antibodies to asialoglycoprotein-receptor (anti-ASGP-R), liver cytocol type-1 antibodies (anti-LC1) and anti-SLA have had similar limitations [1e7]. Anti-ASGP-R, anti-LC1 and anti-SLA appear to correlate with disease severity and

258

C. Liaskos et al. / Journal of Autoimmunity 24 (2005) 251e260

the treatment response. In general however, autoantibodies should not be used as a tool for monitoring of treatment or to predict AIH activity and outcome [1e 7,42]. In this study, no association was found between anti-SLA and anti-CLAbs detection. Nevertheless, the use of anti-CLAbs as a biological marker of activity and severity either in AIH or PSC [41] should be confirmed by prospective studies examining a significant number of patients during a long-term follow-up period with relapses and remission. The high incidence of anti-CLAbs in AIH patients but alsodalthough in lower frequenciesdin both hepatitis C and B and other liver diseases raises the possibility of a common mechanism leading to their production in a number of different liver diseases. One attractive hypothesis could be the induction by chronic stimulus of neoantigens, by disrupting liver cell membranes [23,43,44]. The critical issue is the clinical significance of antiCLAbs, which means whether they are thrombogenic or not and whether they could be implicated in AIH pathogenesis. Anti-CLAbs have been extensively investigated in patients with viral hepatitis where they found to be one of the most common autoantibodies [23,26e28,43]. The prevalent concept is that in the majority of HBV or HCV cases, anti-CLAbs are nonpathogenic and therefore their routine determination is not justified [23,45]. However, in particular patients with special immune reactivity or with abnormal haemostatic regulation, they may exert a pro-coagulant effect and be involved in the genesis of thrombotic events. Under this context, Elefsiniotis et al. [46] found a higher prevalence of anti-CLAbs in patients with HBV-related hepatocellular carcinoma and portal vein thrombosis (PVT), compared to those without PVT. The authors suggested the possible participation of anti-CLAbs in thrombotic mechanisms and in the hypercoagulable state, which occurs in advanced liver disease. However, they do not analyse b2-GPI dependency providing no clue as to whether anti-CLAbs play a role in the initiation of these complications or merely reflect a result of tissue damage and local activation of the clotting system. We showed that anti-CLAbs in AIH patients are rather not pathogenic as demonstrated by the absence of anti-b2-GPIAbs [23,24,30]. Interestingly however, we observed that the RA% values of IgG anti-CLAbs in AIH patients were similar to those reported in APLS patients [24,30] and with those observed in three welldefined APLS patients in the present study. Elution of bound immunoglobulins with 2 M urea allows the detection of high affinity antibodies without influencing the antigen structure [24,30,47]. Therefore, urea resistance has been proven to be well associated with antibody avidity [24,30]. In a previous study, RA values equal to or higher than 30% were associated with APLS [30]. According to this article, the median RA was

statistically higher in APLS patients compared to nonAPLS with SLE [30]. In a second study, RA values of HIV-infected anti-CLAbs positive patients were significantly lower (mean value below 25%) than those with APLS concluding that HIV infection-related antiCLAbs are not thrombogenic [24]. The production of high avidity anti-CLAbs in our AIH patients may be of particular interest, raising questions on the possibility of developing APLS features during the time course of the disease. In conclusion, our report is the first to demonstrate a significantly higher prevalence of anti-CLAbs in AIH patients compared to that found in patients with other liver disorders. From the clinical point of view, antiCLAbs appear to be related to the clinical activity and the stage of AIH. These autoantibodies were not correlated with the known APLS features while they appear to be ‘‘non-pathogenic’’. However, the avidity of IgG anti-CLAbs in AIH patients was comparable with that reported for APLS patients [24,30]. Though no study examining the clinical and laboratory characteristics of APLS patients showed any association between APLS and the presence of AIH [48], our data in association with the findings from two recent studies [49,50] cannot definitely exclude the possibility of some pathogenic implication of these autoantibodies in AIH. In the first study, Shah et al. [49] observed during a follow-up period of 10 years the development of APLS in almost half of their patients who were initially positive for anti-CLAbs but without any evidence of APLS or SLE in the past. In the second study, Arbuckle et al. [50] suggested that specific autoantibodies for SLE are present many years before diagnosis. Taking into account the above studies, as well as the known propensity of AIH to be associated with extrahepatic immune-mediated syndromes, we believe that multicentre prospective studies with long-term follow-up of anti-CLAb(C)/AIH patients as well as with other autoimmune liver diseases such as PBC and PSC are needed in order to address definitely whether these autoantibodies have any clinical importance by contributing to the progression of autoimmune liver disease and the appearance of APLS or they are simply an epiphenomenon associated with the endothelial dysfunction and tissue damage.

References [1] Manns MP, Strassburg CP. Autoimmune hepatitis: clinical challenges. Gastroenterology 2001;120:1502e17. [2] Dalekos GN, Zachou K, Liaskos C, Gatselis N. Autoantibodies and defined target autoantigens in autoimmune hepatitis: an overview. Eur J Intern Med 2002;13:293e303. [3] Czaja AJ. Drug therapy in the management of type 1 autoimmune hepatitis. Drugs 1999;57:49e68.

C. Liaskos et al. / Journal of Autoimmunity 24 (2005) 251e260 [4] Obermayer-Straub P, Strassburg CP, Manns MP. Autoimmune hepatitis. J Hepatol 2000;32(Suppl 1):181e97. [5] Krawitt EL. Autoimmune hepatitis. N Engl J Med 1996;334: 897e903. [6] Van der Berg. Autoimmune hepatitis: pathogenesis, diagnosis and treatment. Scand J Gastroenterol 1998;225(Suppl):66e9. [7] Alvarez F, Berg PA, Bianchi FB, Bianchi L, Burroughs AK, Cancado EL, et al. International autoimmune hepatitis group report: review of criteria for diagnosis of autoimmune hepatitis. J Hepatol 1999;31:929e38. [8] Krawitt EL. Can you recognize autoimmune hepatitis? Postgrad Med 1998;104:145e9, 152. [9] Pathmakanthan S, Kay EW, Murray FE. Autoimmune chronic active hepatitis associated with the presence of antiphospholipid antibodies. Eur J Gastroenterol Hepatol 1998;10:155e7. [10] Gurudu MD, Mittal SK, Shaber M, Gamboa E, Michael S, Sigal LH. Autoimmune hepatitis associated with autoimmune hemolytic anemia anti anticardiolipin antibody syndrome. Dig Dis Sci 2000;45:1878e80. [11] Tanaka H, Umekawa T, Kikukawa T, Toyoda N. Autoimmune hepatitis complicated with antiphospholipid syndrome in pregnancy. Am J Reprod Immunol 2002;3:142e5. [12] Dourakis SP, Michael AE, Papanikolaou IS, Nomikou E, Thalassinou P, Hatziyannis SJ. Autoimmune hepatitis associated with the antiphospholipid syndrome. Eur J Gastroenterol Hepatol 2001;5:591e3. [13] Katayama Y, Kohriyama K, Kirizuka K, Nishizaki H, Fujii H, Tanji Y. Sjogren’s syndrome complicated with autoimmune hepatitis and antiphospholipid antibody syndrome. Intern Med 2000;1:73e6. [14] Benten D, Widjaja A, von Depka M, Flemming P, Trautwein C, Boker K, et al. Autoimmune hepatitis associated with coagulation disorder and immune thyreopathy. Z Gastroenterol 2001;10: 837e840. [15] Tauchmanova L, Rossi R, Coppola A, Luciano A, Del Viscovo L, Soriente L, et al. Antiphospholipid syndrome, adrenal failure, dilated cardiomyopathy and chronic hepatitis: an unusual manifestation of multiorgan autoimmune injury? Eur J Endocrinol 1998;139:641e5. [16] Saeki R, Kaneko S, Terasaki S, Unoura M, Ogino H, Wakabayashi T, et al. Mixed types of chronic acute hepatitis and primary biliary cirrhosis associated with the anti-phospholipid syndrome, a case report. Hepatogastroenterology 1993;40: 499e501. [17] Kesler A, Pomeraz I, Huberman M, Novis B, Kott E. Central venous thrombosis and chronic active hepatitis as part of the antiphospholipid syndrome. Postgrad Med J 1996;76:690e2. [18] Beales ILP. An acquired pseudo-Benard Soulier syndrome occurring with autoimmune chronic active hepatitis and anticardiolipin antibody. Postgrad Med J 1994;70:305e8. [19] Wilson WA, Gharavi AE, Koike T, Lockshin MD, Branch DW, Piette JC, et al. International consensus statement on preliminary classification criteria for definite antiphospholipid syndrome. Arthritis Rheum 1999;42:1309e11. [20] Levine JS, Branch W, Rauch J. The antiphospholipid syndrome. N Engl J Med 2002;346:738e46. [21] Ruiz-Irastorza G, Khamashta MA, Hughes GRV. Hughes syndrome crosses boundaries. Autoimmun Rev 2002;1:43e8. [22] Khamashta MA, Hughes GV. Antiphospholipid antibodies and antiphospholipid syndrome. Curr Opin Rheumatol 1995;7:389e94. [23] Dalekos GN, Zachou K, Liaskos C. The antiphospholipid syndrome and infection. Curr Rheumatol Rep 2001;3:277e85. [24] Petrovas C, Vlachoyiannopoulos P, Kordossis T, Moutsopoulos HM. Anti-phospholipid antibodies in HIV infection and SLE with or without antiphospholipid syndrome: comparisons of phospholipid specificity, avidity and reactivity with b2-GPI. J Autoimmun 1999;13:347e55.

259

[25] Gharavi AE, Pierangeli SS, Harris EN. New developments in viral peptides and APL induction. J Autoimmun 2000;15:227e30. [26] Dalekos GN, Mannousakis MN, Zervou E, Tsianos EV, Moutsopoulos HM. Immunologic and viral markers in antiHIV negative heroin addicts. Eur J Clin Invest 1993;23:219e25. [27] Dalekos GN, Kistis KG, Boumba DS, Voulgari P, Zervou EK, Drosos AA, et al. Increased incidence of anti-cardiolipin antibodies in patients with hepatitis C is not associated with aetiopathogenetic link to anti-phospholipid syndrome. Eur J Gastroenterol Hepatol 2000;12:67e74. [28] Zachou K, Liaskos C, Christodoulou DK, Kardasi M, Papadamou G, Gatselis N, et al. Anti-cardiolipin antibodies in patients with chronic viral hepatitis are independent of beta2glycoprotein I cofactor or features of antiphospholipid syndrome. Eur J Clin Invest 2003;33:161e8. [29] Dalekos GN, Manoussakis MN, Goussia AC, Tsianos EV, Moutsopoulos HM. Soluble interleukin-2 receptors, antineutrophil cytoplasmic antibodies and other autoantibodies in patients with ulcerative colitis. Gut 1993;34:658e64. [30] Vlachoyiannopoulos PG, Petrovas C, Tektonidou M, Krilis SA, Moutsopoulos HM. Antibodies to b2glycoprotein I: urea resistance, binding specificity and association with thrombosis. J Clin Immunol 1998;18:308e91. [31] Chazouilleres O, Wendum D, Serfaty L, Montembault S, Rosmorduc O, Poupon R. Primary biliary cirrhosis-autoimmune hepatitis overlap syndrome: clinical features and response to therapy. Hepatology 1998;28:296e301. [32] Beuers U. 2005. Hepatic overlap syndromes. J Hepatol in press. [33] Kyriakou DS, Alexandrakis MG, Zachou K, Passam F, Stathakis NE, Dalekos GN. Hemopoietic progenitor cells and bone marrow stromal cells in patients with autoimmune hepatitis type 1 and primary biliary cirrhosis. J Hepatol 2003;39:679e85. [34] Tsikrikoni A, Kyriakou DS, Rigopoulou EI, Alexandrakis MG, Zachou K, Passam F, et al. 2005. Markers of cell activation and apoptosis in bone marrow mononuclear cells of patients with autoimmune hepatitis type 1 and primary biliary cirrhosis. J Hepatol in press. [35] Manoussakis MN, Stavropoulos ED, Germanidis GS, Papasteriades CA, Garalea KL, Dontas AS, et al. Soluble interleukin-2 receptors and autoantibodies in the serum of healthy elderly individuals. Autoimmunity 1990;7:129e37. [36] Manoussakis MN, Charavi AE, Drosos AA, Kitridou RC, Moutsopoulos HM. Anticardiolipin antibodies in unselected autoimmune rheumatic disease patients. Clin Immunol Immunopathol 1987;44:297e307. [37] Manoussakis MN, Tzioufas AG, Silis MP, Pange PJ, Goudevenos J, Moutsopoulos HM. High prevalence of anticardiolipin and other autoantibodies in a healthy elderly population. Clin Exp Immunol 1987;69:557e65. [38] Kamoun PP. Denaturation of globular proteins by urea: breakdown of hydrogen or hydrophobic bonds. Trends Biochem Sci 1988;13:424e5. [39] Lewis S, Keil L, Binder W, De Bari VA. Standardized measurement of major immunoglobulin class (IgG, IgA and IgM) antibodies to b2-glycoprotein I in patients with antiphospholipid syndrome. J Clin Lab Anal 1998;12:293e7. [40] de Larranaga GF, Harris N, Pierangeli SS, Alonso BS, Schroder T, Fainboim H. Low prevalence of autoimmune antiphospholipid antibodies in hepatic diseases. Medicina 2000;60:919e22. [41] Angulo P, Peter JB, Gershwin ME, DeSotel CK, Shoenfeld Y, Ahmed AE, et al. Serum autoantibodies in patients with primary sclerosing cholangitis. J Hepatol 2000;32:182e7. [42] Czaja AJ. Behaviour and significance of autoantibodies in type 1 autoimmune hepatitis. J Hepatol 1999;30:394e401. [43] Biron C, Andreani H, Blanc P, Ramos J, Ducos J, Guigue N, et al. Prevalence of antiphospholipid antibodies in patients with

260

[44] [45]

[46]

[47]

C. Liaskos et al. / Journal of Autoimmunity 24 (2005) 251e260 chronic liver disease related to alcohol or hepatitis C virus: correlation with liver injury. J Lab Clin Med 1998;131:243e50. Gharavi AE, Pierangeli SS, Harris EN. New developments in viral peptides and APL induction. J Autoimmun 2000;15:227e30. Yuste JR, Prieto J. Anticardiolipin antibodies in chronic viral hepatitis. Do they have clinical consequences? Eur J Gastroenterol Hepatol 2003;15:717e9. Elefsiniotis I, Diamantis I, Dourakis S, Kafiri G, Pantazis K, Mavrogiannis C. Anticardiolipin antibodies in chronic hepatitis B and chronic hepatitis D infection, and hepatitis B-related hepatocellular carcinoma. Relationship with portal vein thrombosis. Eur J Gastroenterol Hepatol 2003;15:721e6. Sonnerborg A, Sallberg M. Human immunodeficiency virus type 1-infected patients with no disease progression display high

avidity antibody production to autologous V3 sequences. J Infect Dis 1995;171:509e11. [48] Cervera R, Piette JC, Font J, Khamashta MA, Shoenfeld Y, Camps MT, et al. Antiphospholipid syndrome: clinical and immunologic manifestations and patterns of disease expression in a cohort of 1,000 patients. Arthritis Rheum 2002;46: 1019e27. [49] Shah NM, Khamashta MA, Atsumi T, Hughes GR. Outcome of patients with anticardiolipin antibodies: a 10 year follow-up of 52 patients. Lupus 1998;7:3e6. [50] Arbuckle MR, Mc Clain MT, Rubertone MV, Scofield RH, Dennis G, James JA, et al. Development of autoantibodies before the clinical onset of systemic lupus erythematosus. N Engl J Med 2003;349:1526e33.