Breaking self-tolerance toward cytochrome P4502E1 (CYP2E1) in chronic hepatitis C: Possible role for molecular mimicry

Research Article

Breaking self-tolerance toward cytochrome P4502E1 (CYP2E1) in chronic hepatitis C: Possible role for molecular mimicry Salvatore Suttia, Matteo Vidalia, Cristina Mombelloa, Massimo Sartoric, Magnus Ingelman-Sundbergb, Emanuele Albanoa,* a

Department of Medical Sciences and Interdisciplinary Research Centre for Autoimmune Diseases (IRCAD), University ‘‘Amedeo Avogadro” of East Piedmont Novara, Italy; bSection of Pharmacogenetics, Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden; cGastroenterology Unit, Ospedale Maggiore della Carità, Novara, Italy

Background & aims: Circulating auto-antibodies targeting conformational antigens on cytochrome P4502E1 (CYP2E1) are detectable in patients with chronic hepatitis C (CHC) and are associated with more severe necro-inflammation. This study investigated the antigen specificity and the possible origin of these auto-antibodies. Methods: CYP2E1 site-directed mutagenesis and molecular simulation were used to characterize the epitope specificity of CHC-associated anti-CYP2E1 auto-antibodies. Results: Immunoprecipitation experiments using differently mutated human CYP2E1s revealed that conformational antiCYP2E1 antibodies targeted two epitopes located on the molecule surface in an area between Lys324-Glu346 at J–K’’ helices overlapping. Such epitopes were not recognized by the sera targeting linear CYP2E1 antigens. The CYP2E1324–346 peptide showed good homology with two sequences (NS5b438–449 and NS5b456–465) within the NS5b protein of hepatitis C virus (HCV). Consistently, conformational anti-CYP2E1 IgG bind to GST-conjugated NS5b438–449 and NS5b456–465 more efficiently than those recognizing CYP2E1 linear antigens. Competition experiments confirmed the cross-reactivity of conformational anti-CYP2E1 IgG with both NS5b438–449 and NS5b456–465. Moreover, mice immunized with GST-conjugated NS5b438–449 or NS5b456–465 peptides developed antibodies recognizing human CYP2E1. Conclusions: In CHC patients cross-reactivity between CYP2E1 and specific sequences in HCV-NS5b protein can promote the development of auto-antibodies targeting conformational epitopes on the CYP2E1 surface that might contribute to hepatic injury. Ó 2010 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Keywords: Cytochrome P450; Autoimmunity; HCV infection; Liver fibrosis. Received 21 January 2010; received in revised form 9 March 2010; accepted 22 March 2010; available online 31 May 2010 * Corresponding author. Address: Department of Medical Sciences, University ‘‘Amedeo Avogadro” of East Piedmont, Via Solaroli 17, 28100 Novara, Italy. Tel.: +39 0321 660642; fax: +39 0321 620421. E-mail address: [email protected] (E. Albano). Abbreviations: CYP, cytochrome P450; CHC, chronic hepatitis C; HCV, hepatitis C virus; LKM-1, Liver Kidney Microsome type-1.

Introduction Cytochrome P450s (CYPs) are a broad family of enzymes involved in the bio-transformation of a variety of endogenous and exogenous compounds [1]. Although CYP isoenzymes are widely distributed in many tissues, autoimmune reactions involving specific CYPs are features of autoimmune adrenal insufficiencies [2] and of autoimmune responses in some liver diseases [3–6]. In particular, auto-antibodies against CYP2C9 and CYP1A2 can be found in tienilic acid- or dihydralazine-induced hepatitis, respectively [7,8], while auto-antibodies against CYP2E1 are detectable in hepatitis caused by halothane, halogenated hydrocarbons, and isoniazid as well as in alcoholic liver disease [9– 11]. Anti-CYP1A1 auto-antibodies are also a characteristic of the hepatitis associated with type-1 polyendocrine syndrome [12]. Finally, both humoral and cellular immune responses targeting CYP2D6 are a hallmark of type-2 autoimmune hepatitis [3–6]. The implication of anti-CYP autoimmunity in causing liver damage has received experimental support by a recent study showing that mouse infection with CYP2D6-expressing Ad5 adenovirus leads to the production of anti-CYP2D6 IgG that cause immune-mediated liver injury by recognizing the same epitopes targeted by human auto-antibodies [13]. The breaking of self-tolerance toward CYPs is not unique to idiopathic or drug-induced hepatitis as Liver Kidney Microsome type-1 (LKM-1) auto-antibodies directed against CYP2D6 are detectable in 1–5% of the patients with chronic hepatitis C (CHC) [14,15]. Furthermore, antibodies against CYP2A6, CYP3A4, and CYP2E1 have been also observed in CHC patients [16,17]. In this latter respect, we have recently reported that about 40% of CHC patients have increased titers of IgG recognizing human CYP2E1 in immuno-enzymatic assays [18]. CHC-associated antiCYP2E1 IgG do not cross-react with CYP2D6 and their presence is independent from virus genotypes or alcohol consumption [18]. CHC-associated anti-CYP2E1 auto-antibodies appear to be heterogeneous in their antigen specificity targeting both conformational and linear epitopes [18,19]. In our hands, the prevalent presence of conformational anti-CYP2E1 auto-antibodies, rather than overall anti-CYP2E1 auto-reactivity, is associated with a 11- and 4-fold increased risk of diffuse necro-inflammation and fibrosis, respectively [19]. Furthermore, in patients undergoing liver transplantation for end-stage HCV infection, high titres of

Journal of Hepatology 2010 vol. 53 j 431–438

Research Article conformational anti-CYP2E1 IgG before transplant as well as their post-transplant persistence are independent risk factors for severe post-transplant recurrence of hepatitis C [20]. In this study, we sought to investigate the epitope specificity, as well as the possible origins of conformational anti-CYP2E1 auto-antibodies associated with HCV infection.

ferent sera (1:20 dilution in RIPA buffer) and then 2 h at 4 °C with 50 ll of protein A-Sepharose CL4B beads (50% w/v suspension in PBS) (Amersham Biosciences, Amersham, UK). The immunocomplexes bound to protein A-Sepharose beads were recovered by centrifugation, washed three times with 1.5 ml PBS, solubilized in 40 ll of SDS buffer pH 6.8 and used for SDS/PAGE electrophoresis [22]. The membranes were probed with a monoclonal mouse IgG toward the His-6-tag tail (Amersham Biosciences, Amersham, UK) and horseradish peroxidase conjugated anti-mouse immunoglobulins using Western Lightning Chemiluminescence Reagent Plus. The band intensities were measured by videodensitometry and expressed as percent of wild-type CYP2E1.

Materials and methods Sequence analysis and molecular computer simulation Patient recruitment For this study, we used 30 sera with high titres of anti-CYP2E1 auto-antibodies obtained from un-treated patients with chronic hepatitis C (CHC) referring to the Ospedale Maggiore della Carità in Novara. All the patients were negative for hepatitis B surface antigen and HIV infection. The HCV genotype was determined using the INNO-Lipa HCV kit (Innogenetics, Zwijndrecht, Belgium). Genotype 1 (1a or 1b) was detected in 13 patients (44%), genotype 2 (2a or 2a/c) in 15 patients (50%) and genotype 3 in 2 patients (6%). The presence of anti-CYP2E1 IgG was determined by an in-house enzyme-linked solid phase immunoassay (ELISA) using recombinant human CYP2E1 as antigen [18]. The predominant targeting of conformational (n = 20 sera) or linear CYP2E1 (n = 10 sera) antigens was preliminary assessed by the capacity of the different sera to recognize folded and unfolded human CYP2E1 in, respectively, immunoprecipitation and Western blotting assays, as previously reported [18,19]. All subjects gave informed consent to the analysis and the study was planned according to the guide-lines of the local ethical committee. Site-directed mutagenesis of CYP2E1 and prokaryote expression of recombinant CYPs Human CYP2E1 was cloned into the expression plasmid pCWori+ between the restriction sites NdeI and HindIII. Additional bases encoding 6 C-terminal histidines were added and the nucleotides encoding the first 18 amino acids were removed in order to optimize the expression [21]. The mutant CYP2E1s were generated by using the QuickChange XL Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacture instructions. The forward and reverse DNA primers used for each mutation are reported in Table 1. The XL-1 Blue supercompetent Escherichia coli strain was transfected with the different plasmids by heat shock and selected on LB-agar plates containing ampicillin. The correct sequence of the inserts was confirmed by automated DNA sequencing. Wild-type and mutated CYP2E1 variants were expressed and purified as previously reported [22]. Reaction of human sera with mutated CYP2E1s The recognition of wild-type and mutated CYP2E1s by CHC sera was estimated in immunoprecipitation experiments. Briefly, CYP2E1 (10 pmol) solubilized in RIPA buffer was incubated over-night at 4 °C in an orbital shaker with 5 ll of the dif-

CYP2E1 and HCV-NS5b sequence alignment was performed using the ClustalW2 program (http://www.ebi.ac.uk/tools/clustalw2/index.html), while the SWISSMODEL server (http://www.expasy.ch/swissmod/SWISS-MODEL/.html) was used for protein modeling. The molecular model of CYP2E1 was based on the CYP2C5 crystal structure and potential motifs for amino acid exchanges were selected by computer simulation in surface a helices and b sheets as previously reported [22]. Route mean square (RMS) errors and atomic distances were calculated using DeepView-SwissPdbViewer software (http://www.expasy.org/spdbv/). CYP2E1 graphical representation was made using Rastop (version 2.0.3) by Ph. Valadon (La Jolla, USA). Cross-reactivity assays with HCV proteins NS5b438–449 and NS5b456–465 nucleotide sequences from genotype 1b HCV were generated by DNA-free PCR using D-nucleotide triphosphate mixture (0.5 mM), Dream Taq polymerase (0.5 U/ll) (Fermentas International Inc., Burlington, Canada) and the following primers (0.5 lM): NS5B1b438–449 forward: GCTAGCGCGCATGGTGGAGGTGGCTCCCAACTTGAGAAAGCC CTAGATTGTCAG NS5B1b438–449 reverse: CGATGCTAGCCCCGTAGATCTGACAATCTAGGGCTTTCTC NS5B1b456–465 forward: GCTAGCGCGCATGGTGGAGGTGGCTCCCCACTTGACCTACCT CAG NS5B1b456–465 reverse: CGATGCTAGCTCGTTCAATGATCTGAGGTAGGTCAAGTGG The amplicons were separated by electrophoresis and the correct sequence was confirmed by DNA sequencing. The sequences were inserted in a modified pGex4T-1 cloning vector by using BsshII and NheI restriction enzymes and T4 DNA ligase (Invitrogen, Carlsbad, CA, USA). DH5’a E. coli strains transfected with the different plasmids were selected on LB-agar plates containing ampicillin (50 lg/ml). Single colonies were expanded for a 16 h culture in LB medium plus ampicillin and then further grown 5 h at 30 °C with isopropyl b-D-1-thiogalactopyranoside (1 mM). Bacterial extracts were incubated 1 h at 4 °C with Glutathione–Agarose (Sigma–Aldrich, Milan, Italy) and GST-conjugated peptides were purified by column chromatography using 10 mM reduced glutathione in Tris– HCl 50 mM pH 9.5. The cross-reactivity of CHC sera with NS5B438–449 and NS5B456–465 peptides was evaluated by ELISA coating polystyrene Maxi-Sorp ELISA plates with 2 lg

Table 1. Mutations inserted in CYP2E1 molecule and DNA primers used for in situ mutagenesis reactions. Mutations

Primers

K243 AAA ? A243 GCA + E244 GAG ? A244 GCG + E248 GAA ? A248 GCA + K251 AAG ? A251 GCG K324 AAG ? A324 GCG

GCAGCGTATGTGTCTGCAAGGGTGGCGGAGCACCATCAATC GATTGATGGTGCTCCGCCACCCTTGCAGACACATACGCTGC CCCTGAGATCGAAGAGGCGCTCCATGAAGAAATTGACAGGG CCCTGTCAATTTCTTCATGGAGCGCCTCTTCGATCTCAGGG GCTCCATGAAGAAATTGACGCGGTGATTGGGCCAAGCCG CGGCTTGGCCCAATCACCGCGTCAATTTCTTCATGGAGC GCCGAATCCCTGCCATCGCGGATAGGCAAGAGATGC GCATCTCTTGCCTATCCGCGATGGCAGGGATTCGGC CCCTGCCATCAAGGATGCGCAAGAGATGCCCTACATGG CCATGTAGGGCATCTCTTGCGCATCCTTGATGGCAGGG CCATCAAGGATAGGCAAGCGATGCCCTACATGGATGC GCATCCATGTAGGGCATCGCTTGCCTATCCTTGATGG CCAGAACACTTCCTGAATGAAAATGGAGCGTTCAAGTACAGTGAC GTCACTGTACTTGAACGCTCCATTTTCATTCAGGAAGTGTTCTGG CCTGAATGAAAATGGAAAGGCCAAGTACAGTGACTATTTCAAGCC GGCTTGAAATAGTCACTGTACTTGGCCTTTCCATTTTCATTCAGG

R331 AGG ? A331 GCG K342 AAA ? A342 GCG R344 AGG ? A344 GCG E346 GAG ? A346 GCG K420 AAG ? A420 GCG F421 TTC ? A421 GCC

432

Journal of Hepatology 2010 vol. 53 j 431–438

JOURNAL OF HEPATOLOGY of GST-NS5b peptides in bicarbonate buffer, pH 9.6. The results were expressed as optical density (O.D.) at 495 nm after subtracting the background reactivity with the carrier protein. For competition assays, human sera were incubated overnight on a rotor wheel with either 0.1 nmol/ml of CYP2E1 or 4 lg/ml of, respectively, GST-NS5b438–449, GST-NS5b456–465 or the carrier protein. Immunocomplexes were separated by 8 min centrifugation at 13,000g and the supernatants were tested for the remaining reactivity towards the two NS5b-derived peptides or CYP2E1. The results were expressed as percent of non-pre-adsorbed sera. Immunization experiments Male FVB mice (25 g body weight) were immunized by sub-cutaneous injections with 100 lg of GST-conjugated NS5b438–449 and NS5b456–465 peptides or carrier GST in complete Freund’s adjuvant and re-busted after 3 weeks with the same antigens in incomplete Freund’s adjuvant. The mice were sacrificed after 1 week from re-busting. The sera from three mice were pooled and used for Western blotting and immunoprecipitation experiments as described above. Data analysis and statistical calculation SPSS statistical software v.12.0 (SPSS Inc., Chicago, Illinois, USA) was used for statistical analysis. Differences between groups were estimated by non-parametric Mann–Whitney or Chi-square test. Normality distribution was preliminarily assessed by the Kolmogorov–Smirnov and the Shapiro–Wilk tests and corrections were performed by logarithmic transformation.

Results Characterization of the conformational epitopes targeted by antiCYP2E1 associated with CHC To characterize the epitope specificity of CHC-associated conformational anti-CYP2E1 auto-antibodies we took advantage of a previously described approach for the identification of potential antigenic motifs based on the substitution of specific amino acids selected by computer simulation on the surface of the CYP2E1 molecule [22]. Site-directed mutagenesis was used to produce seven mutated CYP2E1 in which alanine replaced the charged residues of Lys324, Arg331, Lys342, Glu346, Lys420, and Phe421 in the J, J0 helices and the loop connecting K’’ and L helices of CYP2E1 C-terminal portion. A further mutated form containing four alanine substitutions at Lys243, Glu244, Glu248, and Lys251 in the G helix was also used. These amino acids were selected because of identified conformational antigenic areas on the CYP2E1 surface and because previously they have been proven not to affect the correct CYP2E1 folding [22]. The capacity of 20 CHC sera with predominant reactivity against conformational anti-CYP2E1 antigens to recognize the different CYP2E1 variants was assayed by immunoprecipitation assays. Fig. 1 shows, when compared to the wild-type CYP2E1, that the different sera immunoprecipitated mutated CYP2E1s with either increased or decreased efficiency, likely reflecting variations in antibody affinity in relation to the mutation-induced changes in the antigen charges. By taking into account variations of more than 50%, we observed that mutations of Lys324 and Arg331, alone or in combination, modified CYP2E1 recognition by 17 sera (85%). CYP2E1 immunoprecipitation by the three remaining sera was instead influenced by the Lys342 and Glu346 substitutions (Fig. 1). The change of these amino acids together with the substitution of the neighboring Arg344 effectively modified CYP2E1 recognition by fourteen other sera. Altogether, 17 sera (85%) were responsive to Lys342, Arg344, and Glu346 substitutions. The Lys420 and Phe421 mutations were less effective and involved only six sera (30%) (Fig. 1). Conversely, the combined substitution of Lys243, Glu244, Glu248, and

WT

K - 324 R- 331 K-342 R-344 E -346 K- 420 F- 421 4 helixG

P 26 P 41 P 46 Serum

WT

P 50

100

Lys324 305

Arg331 128

Lys342 99

Arg344 79

Glu346 98

Lys420 130

Phe- 4-helix 421 G 99 92

P9

100

62

181

141

105

72

70

73

P 20

100

357

72

140

78

70

49

69

120

P 26

100

72

249

91

99

167

142

7

137

P 37

100

165

127

100

170

95

127

91

78

P 30

100

398

57

118

19

93

119

37

69

P 36

100

326

66

193

80

260

113

62

65

P 41

100

95

186

68

221

234

116

132

128

121

P 45

100

141

336

231

165

100

85

103

88

P 49

100

168

162

181

53

76

123

100

57

P 33

100

53

32

64

371

464

48

43

80

P 22

100

31

112

78

28

100

115

57

142

P 47

100

21

83

79

25

61

62

35

127

P 43

100

43

50

70

178

55

99

91

93

P 29

100

27

22

46

51

86

100

75

69

P 46

100

14

43

21

90

28

30

52

135 108

P 52

100

23

38

24

89

99

84

91

P 44

100

131

127

171

106

150

124

131

136

P 48

100

90

128

191

103

155

144

100

101

P 42

100

126

123

26

90

343

71

52

68

WT

K - 324 R- 331 K-342 R-344 E -346 K- 420 F- 421 4 helixG

P8 P 35 Fig. 1. Identification of possible antigenic areas targeted by CHC-associated conformational anti-CYP2E1 IgG. The sera of 20 patients with reactivity towards conformational CYP2E1 epitopes were tested in immunoprecipitation experiments for the capacity to recognize His-6-tagged human CYP2E1 in which alanine substituted selected charged amino acids on the molecular surface. The results are expressed as percent of the recovery of modified CYP2E1s as compared to the wild-type protein, included as a reference in each blot. The black squares indicate a reduction in antibody recognition >50% and the grey squares an increase >50%. The different sera are identified in ‘‘italic”. The blots on the top are shown as examples of the immunoprecipitation experiments. The bottom bands represent the Western blot reactivity against the different mutated CYP2E1s of two randomly selected CHC sera recognizing linear CYP2E1 epitopes.

Lys251 in the G helix did not appreciably influence CYP2E1 immunoprecipitation by all the sera tested (Fig. 1). None of the mutated amino acids affected CYP2E1 recognition by 10 sera mainly targeting linear CYP2E1 antigens (Fig. 1). Computer simulation revealed that the amino acids influencing CYP2E1 immunoprecipitation were located in close proximity and identified an antigenic area about 1.9–2.0 nm wide at the overlapping of J, J0 and K’’ helices (Fig. 2). Moreover, by considering the above combinations in the responses to the different mutations we identified two different potential epitopes including the sequences Lys324-Arg331 and Lys342-Glu346, respectively, (Fig. 2). Cross reactivity of conformational anti-CYP2E1 auto-antibodies with HCV proteins Molecular mimicry with viral proteins has been proposed as a possible cause for the development of anti-CYP auto-antibodies

Journal of Hepatology 2010 vol. 53 j 431–438

433

Research Article lowered CYP2E1 recognition by conformational anti-CYP2E1 sera (Fig. 5B).

Arg-344 Lys-342 Phe-421 Glu-346 Arg-331 COOH NH2

Lys-324

Lys-420

Fig. 2. Localization of the conformational antigenic area identified by the site-directed mutagenesis experiments on the simulated three dimensional structure of CYP2E1. The J0 and K00 helices are shown in light grey, while mutated amino acids that effectively influence antibody binding are shown in dark grey. Dotted lines indicate putative conformational epitopes comprising aa324–334 and aa335–346, respectively.

in patients with CHC [14,15]. The alignment of the CYP2E1324–346 peptide with the HCV poly-protein demonstrated significant homology (load score 21–26) between the CYP2E1324–334 and CYP2E1335–346 sequences and those corresponding to two sequences in RNA-dependent-DNA polymerase HCV-NS5b protein (NS5b438–449 and NS5b456–465, respectively) (Fig. 3). These homologies were well conserved among the different HCV genotypes except for HCV4a (Fig. 3). Similarities in the tri-dimensional structure (RMSE = 0.062 nm) were also evident between CYP2E1324–346 and the molecular model of HCV1b-NS5b438–465 sequence available from RCSB Protein Data Bank (http:// wwwrcsb.org/pdb/home/home.do) (Fig. 3). Furthermore, when located on the CYP2E1 structure the amino acids homologous or synonymous with those in the HCV1b-NS5b438–465 sequence were mostly present on the molecule surface in the area encompassing the two conformational epitopes (Fig. 4A and B). These observations and the notion that the CHC patients carrying conformational anti-CYP2E1 IgG were infected by HCV genotype 1 (n = 9; 45%) and genotype 2 (n = 11; 55%) prompted us to verify whether viral peptides might cross-react with conformational anti-CYP2E1 auto-antibodies. As the HCV-NS5b438–449 and -NS5b456–465 sequences were located in the inner part of the molecule, the two peptides were cloned as GST conjugated in the pGex4T-1 vector and expressed in E. coli. ELISA assays using purified GST-NS5b438–449 and GST-NS5b456–465 demonstrated that the sera with predominant reactivity against conformational anti-CYP2E1 epitopes bind to the two NS5b-derived peptides more efficiently than the sera of CYP2E1 linear antigens or CYP2E1-negative sera (Fig. 4C and D). Cross reactivity of conformational anti-CYP2E1 sera was further verified by competition experiments. Fig. 5A shows that pre-absorption with CYP2E1 (0.1 nmol/ml) reduced by about 80% the IgG binding to the two NS5b-derived peptides. In a similar way, the pre-adsorption with GST-NS5b438–449 and GST-NS5b456–465 (4 lg/ml) also

434

Antibodies raised against HCV-NS5b-derived peptides cross-react with human CYP2E1 Further evidence for the possible role of molecular mimicry in the development of anti-CYP2E1 auto-reactivity was obtained by immunizing mice with either GST-NS5b438–449, GST-NS5b456–465, or the carrier GST. Fig. 5C, shows that the sera obtained from mice immunized with the two viral peptides contained IgG recognizing human CYP2E1 in both immunoprecipitation and Western blotting experiments. Conversely, no cross-reactivity with CYP2E1 was observed using sera from GST-immunized mice (Fig. 5C).

Discussion Previous studies have shown that conformational anti-CYP2E1 auto-antibodies detected in newly diagnosed CHC patients as well as in subjects with post-transplant recurrence of hepatitis C are associated with an increased severity of necro-inflammatory injury [18–20]. At present, the mechanisms by which conformational anti-CYP2E1 auto-antibodies participate in hepatocyte damage in CHC are poorly understood. In order to propose a pathogenic role for anti-CYP antibodies it is necessary that the specific CYP localizes on the cell plasma membranes and the epitopes recognized are accessible to binding by the antibody [5,6]. By using a panel of CYP2E1 variants containing single substitutions in selected amino acids located on the molecular surface, we show that conformational anti-CYP2E1 IgG mainly targets two distinct epitopes in an area comprised between Lys324 and Glu346 at the juxtaposition of the J and J’ helices. The close proximity of these epitopes with the loop between K’’ and L helices (Fig. 2) can explain why Lys420 and Phe421 mutations interfere with CYP2E1 recognition by some CHC sera. The Lys324-Glu346 antigenic structure partially overlaps with the epitope recognized by conformational anti-CYP2E1 auto-antibodies associated with halothane-induced hepatitis [22]. However, in contrast to the latter, CHC-related anti-CYP2E1 auto-antibodies display more variable responses to CYP2E1 mutations. This likely reflects a heterogeneous polyclonal response characterized by antibodies showing either increased or decreased affinity in relation to the modifications of the antigen charges. Moreover, the sera from CHC patients do not target an epitope in the G helix (Lys243 and Lys251) that is recognized by anti-CYP2E1 IgG from halothane hepatitis patients [22]. Drug metabolizing CYPs display a high degree of sequence homology [1]. In this context, the J, K, and L helices might represent an important antigenic area in CYPs. Indeed, these helices are binding sites of anti-CYP2C9 and anti-CYP3A4 conformational antibodies associated with tienilic acid-induced autoimmune hepatitis and hypersensitivity reactions to aromatic anti-convulsing drugs, respectively [7,8]. Moreover, sequence alignment demonstrates that the CYP2E1324–346 antigen is close to the peptides CYP2D6316–327 and CYP2D6321–379 that are among the targets of anti-CYP2D6 LKM-1 antibodies [23,24]. It is well established that a fraction of hepatocyte CYP2E1 is located in the outer layer of the plasma membrane where they are transported via the Golgi secretory vesicles [25]. Structural simulations based on CYP2E1 interaction

Journal of Hepatology 2010 vol. 53 j 431–438

JOURNAL OF HEPATOLOGY

R344

CYP2E1335-346

NS5b456-465

I 341

P460

R465

P339 I 462

P456

L439

G334

L325

D444

P335 D330

G449

CYP2E1324-334

NS5b438-449

Fig. 3. Alignment of the CYP2E1324–346 sequence with a portion (aa 421–480) of NS5b from different HCV genotypes. Amino acid alignment was performed using the ClustalW2 program as reported in the methods section. Homologous and synonymous amino acids are shown in dark grey and light grey, respectively. The values on the right represent sequence homology load scores. The molecular structure of the CYP2E1324–346 and HCV1b NS5b421–480 fragments are shown at the bottom. The continuous and dotted circles show the structures of the corresponding sequences in CYP2E1- and NS5b-derived peptides. The SWISS-MODEL server was used for protein modeling.

with membrane lipids indicate that the area comprising J–L helices is located on the outer portion of the molecule and is well accessible to the antibody binding site [22,26]. This is consistent with previous observations about the recognition of plasma membrane CYP2E1 by the auto-antibodies present in CHC sera [18] as well as by those associated with halothane-induced hepatitis [27]. Thus, as in the case of LKM-1 antibodies [28], the binding of conformational anti-CYP2E1 IgG to epitopes on J–L helices has the capability to trigger antibody-mediated cytotoxicity. Nonetheless, we cannot exclude that these auto-antibodies might be the hallmark of T-lymphocyte-mediated responses, as some overlap has been reported in CYP2D6 epitopes recognized by B- and T-lymphocytes from type-2 autoimmune hepatitis patients [29,30]. So far little is known about the mechanisms leading to the development of conformational anti-CYP2E1 IgG in HCV infected subjects. Although alcohol abuse has been shown to stimulate the

breaking of self-tolerance toward CYP2E1 [11], in our hands alcohol intake by CHC patients does not influence anti-CYP2E1 IgG prevalence [18]. Furthermore, the fact that anti-CYP selfreactivity in CHC is highly specific to individual CYPs [15–17] and no cross-reactivity occurs between anti-CYP2E1 antibodies and CYP2D6 [18], this rules out the possibility that CYP release from damaged hepatocytes might trigger the breaking of selftolerance. Current research on the mechanisms by which HCV infection promotes the onset of autoimmune reactions have focused on the peculiar ability of HCV to stimulate B-lymphocyte activation as well as on molecular mimicry [3–6,15]. Molecular mimicry defines the possibility that auto-reactive T-lymphocytes can be activated by cross-reacting with microorganism antigens that share similar linear or conformational structures [31]. By now, mimicry mechanisms involving viral proteins are recognized to play a key role in the pathogenesis of several human autoimmune diseases [31]. In this contest, mimicry between

Journal of Hepatology 2010 vol. 53 j 431–438

435

Research Article

K342

R331

COOH K324

R342 K342

NH2

NS5b 438-449

1.0 0.8 0.6

a, b b

0.4 0.2

IgG reactivity (OD490nm)

IgG reactivity (OD490nm)

E346

NS5b456-465

0.6 0.5

a, b

0.4 0.3

b

0.2 0.1 0

0 Conf Lin CYP2E1CYP2E1+

Conf Lin CYP2E1+

CYP2E1-

Fig. 4. Localization of the amino acids shared with HCV-NS5b438–465 sequence on the CYP2E1 surface (A and B). NS5b438–465 homologous and synonymous amino acids are shown in dark grey and light grey, respectively. The arrows indicate the amino acids that have been mutated for epitope identification. (Panels C and D) Anti-CYP2E1 auto-antibodies from CHC patients recognize antigens in HCV-NS5b-derived peptides. The binding to GST-conjugated NS5b438–449 and NS5b456–465 peptides was evaluated in ELISA using 20 sera with predominant IgG against conformational (CYP2E1 + Conf) and 10 sera with IgG against predominant linear (CYP2E1 + Lin) CYP2E1 epitopes. As a reference, 5 sera negative for anti-CYP2E1 IgG (CYP2E1-) were also used. The values are expressed as optical densities and boxes include the values within 25th and 75th percentile, while the horizontal bars represent the medians. Eighty percent of the values are comprised between the extremities of the vertical bars (10th–90th percentile). Statistical significance: (a) p <0.05 vs sera with linear anti-CYP2E1 IgG (95% CI 0.04–0.42 and 0.01–0.025); (b) p <0.001 vs anti-CYP2E1-negative sera (95% CI 0.11–0.37).

CYP2D6 and HCV proteins, including E1, NS3, NS5a, and NS5b proteins, accounts for the development of LKM-1 auto-antibodies [32–34] and auto-reactive T-lymphocytes directed to antigens in CYP2D6, CYP2A6, and CYP2A7 [14,35]. Moreover, cross-reactivity 436

between CYP2D6 and Cytomegalovirus and herpes simplex virus proteins has also been observed [33,36,37]. Here, we show that the CYP2E1324–346 sequence, encompassing the conformational antigenic area recognized by CHC-associated conformational

Journal of Hepatology 2010 vol. 53 j 431–438

JOURNAL OF HEPATOLOGY Targets

80 60 40

CYP2E1

NS5b438-449

Target CYP2E1

50

NS5b456-465

20 0

Percent reduction

Percent reduction

100

40 30

IP

20

Antiserum against:

10 0

Competitor CYP2E1

WB

a a

NS5b438-449

NS5b 456-465

GST carrier

8-4 43

b S5 -N T GS

49

6-4 45

b S5 -N T GS

65

T GS

E1 P2 Y C

Competitors Fig. 5. Cross reactivity between CYP2E1 and NS5b438–449 and NS5b456–465 peptides. (A and B) The conformational anti-CYP2E1 CHC sera (n = 20) were pre-adsorbed with either CYP2E1 (0.1 nmol/ml) (A) or GST-conjugated NS5b438–449 and NS5b456–465 peptides (B) and evaluated for their binding to the two NS5b-derived peptides and human CYP2E1 respectively. The values are expressed as percent of the reactivity of the same non-preabsorbed sera. Statistical significance: p <0.001 vs sera pre-adsorbed with the carrier GST (95% CI 31.4–15.4 and 17.7–38.2). (C) Sera from mice immunized with GST-conjugated NS5b438–449 and NS5b456–465 peptides cross-react with human CYP2E1. Western blot (WB) or immunoprecipitation (IP) were performed using sera from SVB mice immunized with either GST-NS5b438–449, GST-NS5b456–465 or the carrier GST protein and recombinant His-6-taged human CYP2E1 as antigen. Mouse sera were used at 1:100 and 1:50 dilutions for Western blotting and for immunoprecipitation, respectively. A polyclonal rabbit-anti-CYP2E1 serum (dilution 1:5000) was used as a reference. In immunoprecipitation experiments CYP2E1 was revealed with a monoclonal mouse IgG toward the His-6-tag tail.

anti-CYP2E1 antibodies, has good homology with two short peptides (NS5b438–449 and NS5b456–465) well conserved in NS5b RNA-dependent-DNA polymerase of most of HCV genotypes. NS5b438–465 homologous/synonymous amino acids are also located on the CYP2E1 surface structure in the area encompassing the two conformational epitopes. Accordingly, recombinant GST-conjugated NS5b438–449 and NS5b456–465 peptides are preferentially recognized by CHC sera containing conformational antiCYP2E1 IgG. Furthermore, cross-reactivity between CYP2E1 and the two peptides is supported by competition experiments and by the development of antibodies targeting human CYP2E1 following mouse immunization with either GST-NS5b438–449 and GST-NS5b456–465. The fact that two distinct NS5b-derived peptides are recognized by conformational anti-CYP2E1 autoantibodies and promote anti-CYP2E1 IgG reactivity further supports the hypothesis that the CYP2E1324–346 antigenic area might contain two distinct epitopes (Fig. 2) driving polyclonal antibody responses. Moreover, NS5b438–449 and NS5b456–465 sequences are well conserved in most of the HCV genotypes, thus explaining why the detection of anti-CYP2E1 conformational IgG in CHC patients is independent from the genotype of the infecting viruses [18,19]. Surprisingly, CHC sera efficiently bind to the NS5b438–449 peptide that in the molecular model shows the least structural similarity with the homologous CYP2E1324–334 sequence (Fig. 3). This suggests the possibility that, when not constrained in the whole molecule, NS5b438–449 might acquire an a-helix conformation similar to that of the CYP2E1324–334 peptide. Altogether these data point to the possibility that B-cell molecular mimicry with HCV-NS5b might cause the breaking of self-tolerance towards specific conformational structures in CYP2E1. Interestingly, cross-reactivity between the CYP2D6245–288 sequence and HCV-NS3 (serine-protease and helicase) or HCV-NS5a (phospho-protein) has been proposed as a cause for CHC-associated conformational LKM-1 antibodies [34]. Nonetheless, we cannot exclude that molecular mimicry with other viral proteins might also promote anti-CYP2E1 autoreactivity, as demonstrated for anti-LKM-1 auto-antibodies [32,36,37]. In conclusion, we show that conformational anti-CYP2E1 auto-antibodies associated with chronic hepatitis C target specific epitopes on the molecule surface. Furthermore, the crossreactivity of these auto-antibodies with distinct structures in

HCV-NS5b suggests that molecular mimicry might be responsible for the breaking of self-tolerance towards CYP2E1 during HCV infection. Conflict of Interest The authors who have taken part in this study declared that they do not have anything to disclose regarding funding or conflict of interest with respect to this manuscript. Acknowledgments This study was supported by grants from the University of East Piedmont and the Regional Government of Piedmont (Progetto Ricerca Sanitaria Finalizzata 2007). References [1] Coon MJ. Cytochrome P450: nature’s most versatile biological catalyst. Annu Rev Pharmacol Toxicol 2005;45:1–25. [2] Betterle C. Autoimmune adrenal insufficiency and autoimmune polyendocrine syndromes: autoantibodies, autoantigens, and their applicability in diagnosis and disease prediction. Endocr Rev 2002;23:327–354. [3] Strassburg CP, Manns MP. Autoantibodies and autoantigens in autoimmune hepatitis. Semin Liver Dis 2002;22:339–352. [4] Vergani D, Mieli-Vergani G. The impact of autoimmunity on hepatocytes. Sem Liver Dis 2007;27:140–151. [5] Bogdanos DP, Dalekos GN. Enzymes as target antigens of liver-specific autoimmunity: the case of cytochrome P450s. Curr Med Chem 2008;15:2285–2292. [6] Bogdanos DP, Mieli-Vergani G, Vergani D. Autoantibodies and their antigens in autoimmune hepatitis. Sem Liver Dis 2009;29:241–253. [7] Lecoeur S, Andre C, Beaune PH. Tienilic acid-induced autoimmune hepatitis: anti-liver and anti-kidney microsomal type 2 autoantibodies recognize a three-site conformational epitope on cytochrome P4502C9. Mol Pharmacol 1996;50:326–333. [8] Leeder JS, Gaedigk A, Lu X, et al. Epitope mapping studies with human anticytochrome P450 3A antibodies. Mol Pharmacol 1996;49:234–243. [9] Bourdi M, Chen W, Peter RM, et al. Human cytochrome P450 2E1 is a major autoantigen associated with halothane hepatitis. Chem Res Toxicol 1996;9:1159–1166. [10] Hoet P, Graf ML, Bourdi M, et al. Epidemic of liver disease caused by hydrochlorofluorocarbons used as ozone-sparing substitutes of chlorofluoro carbons. Lancet 1997;350:556–559. [11] Vidali M, Stewart SF, Rolla R, et al. Genetic and epigenetic factors in autoimmune reactions toward cytochrome P4502E1 in alcoholic liver disease. Hepatol 2003;37:277–285.

Journal of Hepatology 2010 vol. 53 j 431–438

437

Research Article [12] Lankisch TO, Jaeckel E, Strassburg CP. The autoimmune polyendocrinopathy–candidiasis–ectodermal dystrophy or autoimmune polyglandular syndrome type 1. Semin Liver Dis. 2009:307–314. [13] Holdener M, Hintermann E, Bayer M, et al. Breaking tolerance to the natural human liver autoantigen cytochrome P450 2D6 by virus infection. J Exp Med 2008;205:1409–1422. [14] Strassburg CP, Vogel A, Manns MP. Autoimmunity and hepatitis C. Autoimmun Rev 2003;2:322–331. [15] Ferri S, Muratori L, Lenzi M, et al. HCV and autoimmunity. Curr Pharmacol Des 2008;14:1678–1685. [16] Dalekos GN, Obermayer-Straub P, Bartels M, et al. Cytochrome P450 2A6: a new hepatic autoantigen in patients with chronic hepatitis C virus infection. J Hepatol 2003;39:800–806. [17] Miyakawa H, Kitazawa E, Kikuchi K, et al. Immunoreactivity to various human cytochrome P450 proteins of sera from patients with autoimmune hepatitis, chronic hepatitis B, and chronic hepatitis C. Autoimmunity 2001;33:23–32. [18] Vidali M, Occhino G, Ivaldi A, et al. Detection of auto-antibodies against cytochrome P4502E1 (CYP2E1) in chronic hepatitis C. J Hepatol 2007;46: 605–612. [19] Sutti S, Vidali M, Occhino G, et al. Autoantibodies against cytochrome P4502E1 (CYP2E1) conformational antigens in patients with chronic hepatitis C. J Hepatol 2008;48 (suppl. 1):S65. [20] Rigamonti C, Vidali M, Donato MF, et al. Serum autoantibodies against cytochrome P450 2E1 (CYP2E1) predict severity of necroinflammation of recurrent hepatitis C. Am J Transplant 2009;9:601–609. [21] Gillam EM, Guengerich FP. Maximizing expression of the eukaryotic cytochrome P450 in Escherichia coli. Methods Enzymol 1996;272:3–14. [22] Vidali M, Hidestrand E, Eliasson E, et al. Use of molecular simulation for mapping conformational CYP2E1 epitopes. J Biol Chem 2004;279: 50949–50955. [23] Ma Y, Thomas MG, Okamoto M, et al. Key residues of a major cytochrome P4502D6 epitope are located on the surface of the molecule. J Immunol 2002;169:277–285. [24] Sugimura T, Obermayer-Straub P, Kayser A, et al. A major CYP2D6 autoepitope in autoimmune hepatitis type 2 and chronic hepatitis C is a three-dimensional structure homologous to other cytochrome P450 autoantigens. Autoimmunity 2002;35:501–513.

438

[25] Neve EPA, Ingelman-Sundberg M. Intracellular transport and localization of microsomal cytochrome P450. Anal Bioanal Chem 2008;392: 1075–1084. [26] Neve EPA, Hidestrand M, Ingelman-Sundberg M. Identification of sequences responsible for intracellular targeting and membrane binding of rat CYP2E1 in yeast. Biochemistry 2003;42:14566–14575. [27] Eliasson E, Kenna JG. Cytochrome P450 2E1 is a cell surface autoantigen in halothane hepatitis. Mol Pharmacol 1996;50:573–582. [28] Muratori L, Parola M, Ripalti A, et al. Liver/kidney microsomal antibody type 1 targets CYP2D6 on hepatocyte plasma membrane. Gut 2000;46:553–561. [29] Ma Y, Bogdanos DP, Hussain MJ, et al. Polyclonal T-cell responses to cytochrome P450IID6 are associated with disease activity in autoimmune hepatitis type 2. Gastroenterol 2006;130:868–882. [30] Longhi MS, Hussain MJ, Bogdanos DP, et al. Cytochrome P450IID6-specific CD8 T cell immune responses mirror disease activity in autoimmune hepatitis type 2. Hepatology 2007;46:472–484. [31] Münz C, Lünemann JD, Taegue Getts M, Miller SD. Antiviral immune responses: triggers of or triggerd by autoimmunity. Nat Rev Immunol 2009;9:246–258. [32] Manns MP, Griffin KJ, Sullivan KF, et al. LKM-1 autoantibodies recognize a short linear sequence in P450IID6, a cytochrome P-450 monoxygenase. J Clin Invest 1991;88:1370–1378. [33] Kerkar N, Choudhuri K, Ma Y, et al. Cytochrome P4502D6193–212: a new immunodominant epitope and target of virus/self cross-reactivity in liver kidney microsomal autoantibody type 1-positive liver disease. J Immunol 2003;170:1481–1489. [34] Marceau G, Lapierre P, Beland K, et al. LKM1 autoantibodies in chronic hepatitis C infection: a case of molecular mimicry? Hepatology 2005;42:675–682. [35] Kammer AR, van der Burg SH, Grabscheid B, et al. Molecular mimicry of human cytochrome P450 by hepatitis C virus at the level of cytotoxic T cell recognition. J Exp Med 1999;190:169–176. [36] Yamamoto AM, Cresteil D, Homberg JC, Alvarez F. Characterization of antiliver-kidney microsome antibody (anti-LKM1) from hepatitis C viruspositive and negative sera. Gastroenterology 1993;104:1762–1767. [37] Bogdanos DP, Lenzi M, Okamoto M, et al. Multiple viral/self immunological cross-reactivity in liver kidney microsomal antibody positive hepatitis c virus infected patients is associated with the possession of HLA B51. Int J Immunopathol Pharmacol 2004;17:83–92.

Journal of Hepatology 2010 vol. 53 j 431–438