The Lipoic Acid Containing Components of the 2-Oxoacid

Spontaneous occurrence of chronic non-suppurative destructive cholangitis and ... in MRL/lpr mice: Possible animal model for primary biliary cirrhosis...
0 downloads 0 Views 4MB Size
Download PDF
Chem. Res. Toxicol. 1996,8, 736-746

736

The Lipoic Acid Containing Components of the 2-Oxoacid Dehydrogenase Complexes Mimic Trifluoroacetylated Proteins and Are Autoantigens Associated with Halothane Hepatitis Nora Frey,’ Urs Christen,? Paul Jeno,* Stephen J. Yeaman,$ Yoshiharu Shimomura,”J. Gerald Kenna,l A. Jay Gandolfi? Leo Ranek,* and Josef Gut*$+ Department of Pharmacology, Biocenter of the University, CH-4056 Basel, Switzerland, Department of Biochemistry, Biocenter of the University, CH-4056 Basel, Switzerland, Department of Biochemistry and Genetics, Medical School, University of Newcastle upon Tyne, Newcastle upon Tyne NE2 4HH, U.K., Department of Bioscience, Nagoya Institute of Technology, Gokiso, Showa-Ku, Nagoya, Japan, St. Mary’s Hospital Medical School, Imperial College of Science, Technology and Medicine, Norfolk Place, London W2 IPG, U.K., Department of Anesthesiology, University of Arizona Health Science Center, Tucson, Arizona 85724, and Medical Department A, Rzjgshospitalet, DK-2100 Copenhagen, Denmark Received January 30, 1995@

Anti-CF3CO antibodies, monospecific toward trifluoroacetylated proteins (CFsCO-proteins), which are elicited in experimental animals and humans exposed to the anesthetic agent halothane, cross-react with a n unknown protein of approximately 52 kDa, constitutively expressed in tissues of experimental animals and humans not previously exposed to the agent. Using anti-CF3CO antibody, the protein(s) of 52 kDa could be immunoprecipitated from solubilized rat heart homogenate. Two-dimensional gel electrophoretic analysis revealed the presence of distinct major (Pl, P2) and minor (P3, P4, P5) protein components with apparent molecular masses of 52 kDa. From each of the components P1 and P2, the amino acid sequences of three peptides were determined and found to exhibit 100% identity with the corresponding amino acid sequences of the E2 subunit of the rat 2-oxoglutarate dehydrogenase complex (OGDC). Additionally to the E2 subunit of OGDC, anti-CF3CO antibody also recognized on immunoblots the purified E2 subunit of the branched chain 2-oxoacid dehydrogenase complex (BCOADC) and protein X, a constituent of the pyruvate dehydrogenase complex (PDC), in a G(RS)-lipoicacid, manner sensitive to competition by P-(trifluoroacety1)-L-lysine (CF~CO-L~S), (lipoyl-Lys). Furthermore, a discrete population of autoantibodies and NG-(G(RS)-lipoyl)-~-lysine was identified in sera of patients with halothane hepatitis which could not discriminate between the lipoylated target epitope present on the E2 subunit of OGDC and epitopes on CF3CORSA, used as model for CFaCO-proteins. These data suggest that the autoantigenicity of these proteins in halothane hepatitis is based on the molecular mimicry of CF3CO-Lys by lipoic acid, the prosthetic group common to protein X and the E2 subunits of OGDC and BCOADC.

Introduction Trifluoroacetyl adducts to proteins arise through covalent modification of target proteins by the acyl halide intermediate CF3COC1,which is elicited upon oxidative, cytochrome P450-dependent metabolism of halothanel and other, structurally closely related compounds of the pentahaloethane-type (1-5). With either halothane or the chlorofluorocarbon substitute 2,2-dichloro-l,l,l-trifluoroethane (HCFC 123)as substrates, the predominant adduct formed in the process has been identified by 19FNMR as NB-(trifluoroacety1)-L-lysine (CF3CO-Lys (3)). *Author to whom correspondence should be addressed, at his present address: Ciba-Geigy AG, Drug Metabolism and Exploratory Toxicology, K-136.1.19, CH-4002 Basel, Switzerland. Phone: (+41)61 696 36 45; FAX: (+41) 61 696 62 12. Department of Pharmacology, Biocenter of the University, Basel. t Department of Biochemistry, Biocenter of the University, Basel. 5 University of Newcastle upon Tyne. ‘I Nagoya Institute of Technology. Imperial College of Science, Technology and Medicine. # University of Arizona Health Science Center. * Rijgshospitalet. Abstract published in Advance ACS Abstracts, June 1, 1995. +

@

Trifluoroacetylated proteins (CF3CO-protein~)~ are detected in both the livers of rodents (4, 6-9) and liver biopsies of humans exposed to halothane (10, 11). The formation of CF3CO-proteinsis not restricted to the liver, however; CFsCO-proteins have immunochemically been detected in the kidneys ( 4 ) ,the heart (121, and the testes (13)of rats and in the olfactory epithelium and lungs of mice (9)exposed to the agent. Moreover, in the rat liver, CF3CO-proteins have been located not only in hepatocytes but also in Kupffer cells (141, which are liver Abbreviations used halothane, 2-bromo-2-chloro-l,l,l-trifluoroethane; HCFC 123, 2,2-dichloro-l,l,l-trifluoroethane; CF3CO-RSA, trifluoroacetylated rabbit serum albumin; CFsCO-protein, trifluoroacetylated protein; CF&O-Lys, W-(trifluoroacety1)-L-lysine;Lys(Ac), hrs-acetyl-L-lysine;lipoyl-Lys, hrs-(G(RS)-lipoyl)-~-lysine, PBC, primary biliary cirrhosis; ECL, enhanced chemiluminescence; HRP, horseradish peroxidase; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; PDC, pyruvate dehydrogenase complex; OGDC, 2-0x0glutarate dehydrogenase complex; BCOADC, branched chain 2-oxoacid dehydrogenase complex; PBS, phosphate-buffered saline. The term “CF3CO-protein” refers to a trifluoroacetylated protein with no reference made to the identity or function of that particular protein, except where indicated. The term “anti-CFsCO antibody” refers to the monospecific antibody obtained from an anti-CFaCO antiserum through affinity purification on a CF3CO-Lys affinity matrix (26).

0 1995 American Chemical Society

Lipoylated Autoantigens in Halothane Hepatitis resident macrophages competent for antigen presentation

(15). In rare cases (i.e., in between 1in 37000 and 1in 3700 of individuals exposed to halothane on one or more occasions) a severe form of hepatic damage, coined halothane hepatitis (16, 171, develops. Halothane hepatitis is thought to have an immunological basis. m i c t e d individuals have serum antibodies directed toward distinct liver microsomal CF3CO-proteins ( I 7); based on their reactivity with patients' sera, several of these CF3CO-proteins have been isolated from livers of halothaneexposed rats and were identified by amino acid sequencing and/or molecular cloning as protein disulfide isomerase (18))microsomal carboxylase(191,calreticulin (201,ERp72 (211, GRP 78 (221, and ERp99 (23). Current evidence indicates that all halothane-exposed human individuals produce CF3CO-proteins (10,24-26). However, it is not understood why, in contrast to the few susceptible individuals afnicted with the disease, the vast majority of individuals appears to immunologically tolerate CF3CO-proteins. One of several putative mechanisms that could lead to this apparent unresponsiveness (i.e., tolerance) toward CF3CO-proteinsmight be the occurrence in the repertoire of immunological self of structures which mimic CF3CO-proteinsvery closely. Conceptually, such pre-existing self-determinants might promote the development of immunological tolerance toward CF3COproteins in the repertoire of T- andor B-lymphocytes through mechanisms such as clonal deletion, induction of clonal anergy, andor induction of clonal ignorance (27-33). In the search for such self-determinants, we have detected proteins of 52 and 64 kDa, constitutively expressed in the liver of humans as well as in several tissues of rats not previously exposed to halothane, which carry epitopes cross-reactive with CF3CO-proteins when probed for with a monospecific anti-CFsCO antibody (26). The protein of 64 kDa was identified as the dihydrolipoamide acetyltransferase component (E2 subunit) of the mitochondrial pyruvate dehydrogenase complex (PDC) (34). Competitive inhibition by 6(RS)-lipoic acid3 and CF3CO-Lys of the recognition of purified E2 subunit of PDC and of CF3CO-proteins by anti-CF3CO antibody suggested that the lipoyl-binding domains of the E2 subunit, to which lipoic acid is covalently bound as a prosthetic group, were involved in the process of molecular mimicry of CF3CO-proteins (34). In fact, the high degree of structural relatedness of CF3CO-Lys and 6(RS)lipoic acid was identified as the structural basis of molecular mimicry of CFsCO-proteins by the E2 subunit of PDC (35). As one consequence of molecular mimicry, the E2 subunit of human PDC was identified, by the use of autoantibodies present in patients' sera, as an autoantigen associated with halothane hepatitis (35). In this report, we have isolated, based on its high abundancy in this tissue, one of the previously elusive, constitutively expressed protein(s) of 52 kDa from the heart of rats and identified it as the E2 subunit of the 2-oxoglutarate dehydrogenase complex (OGDC). Moreover, we tested the hypothesis if, as a consequence of molecular mimicry of CF3CO-Lys by 6(RS)-lipoicacid (34, 35), the E2 subunits of OGDC and of the branched chain 2-oxoacid dehydrogenase complex (BCOADC),as well as G(RS)-lipoicacid and lipoyl-Lys were used in their oxidized form; for purposes of general discussion, the term lipoic acid was used without indicating the stereochemical assignment at position 6.

Chem. Res. Toxicol., Vol. 8, No. 5, 1995 737 protein X , a constitutent of PDC, were autoantigens associated with halothane hepatitis.

Experimental Procedures Materials. N6-(Trifluoroacetyl)-L-lysine(CF3CO-Lys) and NG-acetyl-L-lysine (Lys(Ac)) were obtained from Senn Chemicals (Dielsdorf, Switzerland). NG-(G(RS)-lipoyl)-~-lysine (lipoyl-Lys) was custom synthesized by and obtained from Peninsula Laboratories (Belmont, CA). 6(RS)-Lipoic acid (oxidized form), L-lysine, trifluoroacetic acid, aprotinin, leupeptin, pepstatin, soybean trypsin-chymotrypsin inhibitor, phenylmethanesulfonyl fluoride, and taurocholic acid were all obtained from Sigma (St. Louis, MO). BioLyte 5/7, BioLyte 3/10, goat antirabbit IgG (HSL) horseradish peroxidase (HRP) conjugate, goat anti-human IgG (H+L) HRP-conjugate, M i - G e l 102, and MiGel Hz were purchased from Bio-Rad Laboratories (Richmond, CAI. The enhanced chemiluminescence (ECL) detection system was obtained from Amersham International (Amersham, U. IC). Trifluoroacetylated rabbit serum albumin (CF3CO-RSA) was synthesized as described (26). The monospecific anti-CFjCO antibody was obtained as described from a polyclonal rabbit anti-CFjCO-RSA antiserum through afinity purification on a M i - G e l 102 amino-terminal agarose column, to which CFjCOLys had been coupled (26). Aliquots of the final preparation (0.1 mg of IgG/mL) were stored at -80 "C and thawed only once. Human Sera and Purified Complexes. Sera from halothane-exposed healthy control individuals and sera from patients with the clinical diagnosis of halothane hepatitis were kindly provided by three of the authors: A.J.G. (collection 11, J.G.K. (collection 2), and L.R. (collection 3). Pyruvate dehydrogenase complex (PDC) and 2-oxoglutarate dehydrogenase complex (OGDC) were both isolated from human heart at biopsy as previously described (36);purified branched chain 2-oxoacid dehydrogenase complex (BCOADC) was isolated from rat liver as described (37). Gel Electrophoresis and Immunoblotting. Protein samples were mixed with equal volumes of dissociation buffer to give a final concentration of 12 mM Tris-HC1 (pH 6.81, 8% (w/v) SDS, 10% (v/v) glycerol, and 40 mM dithiothreitol and were heated to 95 "C for 10 min. SDS-PAGE (38)was performed at room temperature with a Mini Protean I1 cell (Bio-Rad). For samples of PDC, OGDC, and BCOADC a 4.5%stacking gel and a 8.5%separating gel were used, and electophoresis was for 100 min at 100 V. Otherwise, a 4.5%stacking and a 7.5%separating gel was used at 200 V for 40 min. Proteins were transferred electrophoretically (39) to nitrocellulose at 100 V for 1 h i n 25 mM Tris-HC1 buffer, containing 192 mM glycine and 20% (v/v) methanol. After transfer, the nitrocellulose was stained with amido black, destained, and blocked for 2 h at room temperature with phosphate-buffered saline (PBS; comprising 10 mM NazHPO4, 3 mM KH~POI,137 mM NaC1, pH 7.41, containing 2% (w/v) dry milk powder and 0.02% (w/v) Thimerosal (subsequently termed the blocking solution). The nitrocellulose was cut into strips and either incubated with anti-CFjCO antibody (diluted 1:200) or human serum (diluted as indicated) in blocking solution (500pL) for 18 h at room temperature, washed (4 times, 5 min each) with blocking solution, and incubated for 2 h with goat anti-rabbit horseradish peroxidase-conjugated second antibody or, where appropriate, with goat anti-human horseradish peroxidase-conjugated second antibody (diluted 1:2000) at room temperature. Strips were washed (4times, 5 min each) with blocking solution and with PBS, pH 7.4 (4 times, 5 min each and once for 20 min). Visualization was by the ECL detection system. Where indicated, incubation of polypeptides, blotted onto nitrocellulose, with the anti-CF&O antibody was done in presence of effectors. When human serum was used, Tween20 (0.05% (v/v)) was included in all washing steps with blocking solution. Antibody Exchange Immunochemistry. Antibody exchange experiments were performed according to Hammarback and Vallee (40). Samples of OGDC (6 pg/cm gel width) were subjected to SDS-PAGE and electrophoretically transferred to

738 Chem. Res. Toxicol., Vol. 8, No. 5, 1995 nitrocellulose as described above. After blocking for 2 h, the area of the blot containing the E2 subunit of OGDC was cut out and incubated with the serum of patient 34 (serum collection 1) in blocking solution for 16 h at a dilution of 1:15. After washing with blocking solution (4 times, 5 min each), the nitrocellulose strip containing a complex between the E2 subunit of OGDC and the immobilized, affinity-purified autoantibodies was cut in disks of 2.5 x 2.5 mm. Four of these nitrocellulose disks were coincubated with one target disk of 2.5 x 5 mm containing CF3CO-RSA (0.9 pg of proteidcm slot width) in blocking solution for 16 h. Effectors were included where indicated. After washing with blocking solution (4 times, 5 min each), the strips were incubated for 2 h a t room temperature with goat anti-human HRP-conjugated second antibody. Strips were washed with blocking solution (4 times, 5 min each) and with PBS, pH 7.4 (4 times, 5 min each and once for 20 min). No Tween 20 was used in any step of the transfer experiment. The detection by enhanced chemiluminescence was followed by scanning densitometry (Molecular Dynamics 300A, operated with Molecular Dynamics Image Quant v3.0 software) as described (35, 41 1. Immunoaffinity Purification of the 52 kDa Protein(s) and Two-Dimensional Gel Electrophoresis. Rat heart total homogenate (10 mg/mL) was solubilized at 4 "C in a buffer of 50 mM Tris-HC1 (pH 7.4) containing 0.5 M NaCI, 0.5 mM phenylmethanesulfonyl fluoride, 60 mg/mL soybean trypsinchymotrypsin inhibitor, 20 mgimL aprotinin, 0.7 mg/mL leupeptin, 0.7 mgimL pepstatin, and 10 mg/mL taurocholic acid for 90 min on a n end-over-end shaker. After centrifugation (105000g, 1 h), the pellet was discarded and the supernatant was recirculated (10 mL/h) at 4 "C for 16 h over a n anti-CF&O antibody imunoaffinity column (3 mL bed volume), preequilibrated with 50 mM Tris-HC1 (pH 7.4) containing 0.5 M NaCl and 10 mg/mL taurocholic acid. The immunoaffinity matrix had been prepared by coupling anti-CF3CO antibody to M i - G e l Hz (Bio-Rad, 1.2 mg of IgG coupledml of gel). After the sample application, the immunoaffinity column was washed (15 bed volumes) with 50 mM Tris-HC1 (pH 7.41, containing 0.5 M NaCl and 10 mg/mL taurocholic acid. To displace the bound proteids), one bed volume of the same buffer containing 100 mM CF3CO-Lys was applied, and after incubation for 16 h on a n end-over-end shaker and centrifugation (5 min, 800g), the supernatant containing the displaced protein(s) was collected. The immunoaffinity-purified protein(s), comprising a mixture of proteins of 64 and 52 kDa, was then subjected to twodimensional gel electrophoresis. Isoelectric focusing was performed in the first dimension and SDS-PAGE in the second dimension according to the method of O'Farrell (42). Isoelectric focusing was carried out in a 2 x 105 mm rod gel with 4% (w/v) acrylamide containing 2% (v/v) carrier ampholytes (1.6% (viv) BioLyte 5/7 and 0.4% (v/v) BioLyte 3/10) to establish a pH gradient ranging from 3.5 to 9.5. The immunoaffinity purified fraction of taurocholatesolubilized rat heart homogenate was precipitated with 25% (wi v) trichloroacetic acid to remove CF3CO-Lys. The pellet was directly resuspended in first dimension sample buffer (9.5 M urea, 2% (v/v) NP 40, 5% (v/v) P-mercaptoethanol, 1.6% (v/v) BioLyte 517, and 0.4% (vlv) BioLyte 3/10) and was incubated for 10 min at room temperature prior to application onto the gel tube; the sample was overlaid with a buffer containing 9 M urea, 0.8% (viv) BioLyte 5/7, 0.4% (viv) BioLyte 3/10, and 0.05% (v/v) bromophenol blue. Isoelectric focusing was typically for 16 h at 400 V followed by 1h at 800 V using 10 mM H$04 (pH 1.5) a s anode buffer and 20 mM NaOH (pH 11.5) as cathode buffer. After the first dimension, the rod gel was removed, soaked in SDS-PAGE sample buffer (12 mM Tris-HC1, (pH 6.8) containing 8% (w/v) SDS, 10% (v/v) glycerol, and 40 mM dithiothreitol) for 10 min at room temperature, and then transferred onto a polyacrylamide gel. SDS-PAGE was performed according to Laemmli (38)using a 4.5% stacking and a 8.5% separating gel. The proteins were then either electrophoretically transferred onto nitrocellulose to perform immunoblotting studies or stained within the gel with 0.1% (wiv)

Frey et al. Coomassie blue in 1%(v/v) acetic acid and 40% (v/v) methanol for 10 min and destained in 1%(v/v) acetic acid and 40% (v/v) methanol until protein spots were visible. The areas of the gel containing the protein spots of interest were excised and stored at -80 "C until further analysis.

In Situ Digestion and Amino Acid Sequence Analysis. The gel pieces were further destained with two washes (200 p L each) of 40% (v/v) 1-propanol for 5 min followed by two extractions (200 pL each) of 0.2 M NH4HC03/50% (v/v) acetonitrile for 5 min ( 4 3 ) . The gel pieces were then dried until all NH4HC03 and acetonitrile had evaporated. The proteins were then reduced at 60 "C for 45 min i n 50 pL of 125 mM Tris-HC1, pH 8.0, containing 0.1% (w/v) SDS, 10 mM DTT,and 1 mM EDTA, followed by alkylation in 50 mM iodoacetamide at room temperature for 15 min i n the dark. The digestion of proteins within the gel pieces was then started by adding 0.5 pg of endoproteinase LysC (Wako Chemicals, Neuss, Germany) and continued at 37 "C for 18 h. The supernatant was carefully collected, and the gel pieces were first extracted twice with 50 pL of 0.1 M Tris-HC1 (pH 8.0) for 30 min each at room temperature followed by two more extractions with 50 pL of 80% (v/v) acetonitrile containing 0.05% (viv) trifluoroacetic acid for 30 min at room temperature. The supernatants of the organic extracts were taken to dryness and then dissolved in the combined supernatants of the preceding aqueous extracts. An equal volume of 1 M guanidine hydrochloride was added, and incubation was for 30 min a t room temperature. After centrifugation at lOOOOg for 10 min, the supernatant was subjected to separation by RP-HPLC as described (34, 4 3 ) ; peptidic fragments were collected and stored a t -80 "C until amino acid sequence analysis ( 4 3 ) .

Results Identificationof the E2 Subunit of OGDC. Monospecific anti-CFsCOantibodies, directed against CF3COproteins which are elicited in tissues of experimental animals and humans exposed to the anesthetic agent halothane, were shown to recognize cross-reactive proteins of 64 and 52 kDa in several tissues of rats and in the liver of humans not previously exposed to the drug (14). Previously, by the use of the anti-CF3C0 antibody as an immunoaffinity matrix, the protein of 64 kDa was purified in a single step from solubilized rat heart microsomal and mitochondrial fractions and identified as the E2 subunit of PDC (34). Under those conditions, the protein of 52 kDa could not be immunoprecipitated to any considerable degree and its identity remained unknown. However, when instead of solubilized microsomal and mitochondrial fractions of heart tissue, as used in those earlier experiments (341,the starting material was rat heart total homogenate, solubilized with taurocholate, it was possible to co-precipitate the protein of 52 kDa with the E2 subunit of PDC (Figure lA, section 2) by the use of the anti-CFsCO antibody as an immunoaffinity matrix. Analytical two-dimensional gel electrophoresis with isoelectricfocusingin the first dimension followed by SDS-PAGE in the second dimension was performed and revealed a separation of the proteids) of 52 kDa into two major (arbitrarily designated P1, P2) and three minor (P3, P4, P5) protein components (Figure lA, section 1)which were all recognized by anti-CFsCO antibody on immunoblots (Figure lB, section 1) in a manner sensitive to competition by CF&O-Lys (Figure lC, section 1)but not to competition by L-lysine (data not shown). In order to establish the identity of each of the discrete protein components, cross-reactive with anti-CF3CO antibody, preparative two-dimensional gel electrophoresis

Chem. Res. Toxicol., Vol. 8, No. 5, 1995 739

Lipoylated Autoantigens in Halothane Hepatitis I

A

1

P1: peptide 1 OGDC-E2 (rat) OGDC-E2 (human) peptide 2 OGDC-E2 (rat)

VEGGTPLFTLR 123 _ _ _ - - - _ _ -133 _134 ----------- 144 AKPAEAPATAHK 142

peptide 3

B B

. .t

P2: peptide 1 OGDC-E2 (rat) OGDC-E2 (human)

_ - _ _ _ - _ _153 ____ AXAFALQEQPWNAV

VEGGTPLFTL 123 134

peptide 2 OGDC-E2 (rat)

154

peptide 3 OGDC-E2 (rat) OGDC-E2 (human)

431 442

__________

----------

132 143

AAPEAPAAPPPPVAP

_______________

168

AVEDP

_____

-----

435 446

Figure 2. Amino acid sequence comparison. Proteolytic frag-

C

!+)I I f I

I

I I

I 1 I f

Figure 1. Immunoaffinity purification and two-dimensional gel electrophoresis of proteins of 52 kDa cross-reactive with antiCF3CO antibody. Rat heart total homogenate was solubilized and subjected to immunoafinity chromatography on a matrix containing anti-CF3CO antibody coupled to Affi-Gel Hz as described in the Experimental Procedures. (A) Isoelectric focusing (in a pH gradient established by the use of BioLyte 3/10 and BioLyte 518 ampholytes (see Experimental Procedures) followed by SDS-PAGE was used to analyze copurified proteins (100ng of proteidtube) by two-dimensional gel electrophoresis (section 1);one-dimensional analysis of the proteins (100ng of proteidcm slot width) by SDS-PAGE was performed simultaneously (section 2). Proteins were stained by amido black after their electrophoretic transfer to a nitrocellulose membrane. The orientation of the pH gradient from acidic (+) to basic (-) is indicated; arrowheads indicate the relative migration positions of the E2 subunit of PDC (64 kDa, upper) and the proteins of 52 kDa (lower). The corresponding immunoblots were developed with anti-CF&O antibody in the absence (B) or the presence (C) of 100 mM CF&O-Lys.

was performed. Protein components corresponding to P1 and P2 were excised and digested in situ by endoproteinase LysC, and three randomly selected peptides of each protein component were subjected to amino acid microsequence analysis (43). The obtained amino acid sequences of the three peptides of each P1 (Figure 2A) and P2 (Figure 2B) revealed 100% identity with the corresponding deduced amino acid sequences of the rat E2 subunit of OGDC (44), with the exception of peptide 3 of P1 which revealed a minor deviation. In the case of peptide 1of P1 and peptides 1and 3 of P2,100%sequence identity with the corresponding deduced amino acid

ments of the protein components (see also Figure 1)P1 (A) and P2 (B) were separated by RP-HPLC. From randomly selected fragments, the sequences corresponding to the peptides 1, 2, and 3 were obtained. The sequences were compared with the corresponding deduced amino acid sequences of the rat (44)and human (45)E2 subunit of OGDC. Identical residues are indicated by (-1; 6) indicates a gap introduced for maximal alignment, and X indicates a n ambiguous residue.

sequences of the human E2 subunit of OGDC (45) was also found. At this point, it is not clear why the E2 subunit of OGDC is separated into two discrete protein components upon two-dimensionalgel electrophoresis and if these two components might represent distinct isoforms of the E2 subunit of OGDC. One should note, however, that, in the same experiments, the E2 subunit of PDC (which copurified in the course of the immunoafinity purification step with the E2 subunit of OGDC) and the additional proteins of 52 kDa (i.e., the protein components P3, P4, and P5 (Figure lA, section l)), whose identities remain unknown at present, separated into three protein components also (Figure lA,lB, sections 1).An apparent separation of human E2 subunit of PDC into three protein components was noted before (46). In addition, using a human placental cDNA clone coding for the E2 subunit of PDC, Southern blot analysis of human genomic DNA revealed the presence of several bands (47), suggesting the occurrence of one or multiple genes coding for forms of E2 subunit proteins; similarly, several RNA transcripts have been identified by Northern blot analysis which may reflect the transcription of one or more genes related to the gene(s) coding for the E2 subunit of PDC (47,48). These data raise the possibility of the expression of distinct forms of the E2 subunit of PDC; a similar scenario might lead to the expression of putative isoforms of the E2 subunit of OGDC. The identity of P3, P4, and P5 (Figure lA, section 1)remains unknown, primarily because of problems with their recovery. Based on their apparent molecular mass of 52 kDa and their immunochemical properties (i.e., recognition by anti-CFsCO antibody (Figure lB, sections 1 and 2) in a manner sensitive to competition by CF3CO-Lys (Figure lC,sections 1 and 2)), all of which coincide (see below) with properties of protein X, a constituent of PDC, and the E2 subunit of BCOADC, it is tempting to speculate that P3, P4, and P5 are related to these proteins; accordingly, experiments aimed at their identification are in progress.

740 Chem. Res. Toxicol., Vol. 8, No. 5, 1995

Frey et al.

t\b

A

:\b

B

:\b

C

.116

El

. 97

E1B -

E3 E2 -

E2 Ela

?

-

66

-

45

- 29 lmmunoblot

lmmunoblot

PDC

OGDC

lmmunoblot

BCOADC

Figure 3. Competitive immunoblotting analysis of the binding of anti-CFsCO antibody to protein X (A) and E2 subunit proteins of OGDC (B) and BCOADC (C). Immunoblots of purified complexes of (A) human PDC (1pg of proteidcm slot width), of (B) human OGDC (1pg of proteidcm slot width), and of (C) rat BCOADC (2 pg of proteidcm slot width) were incubated with anti-CF&O antibody in the absence (Control) and the presence of CF3CO-Lys (100 mM), G(RS)-lipoic acid (50 mM), lipoyl-Lys (40 mM), and L-lysine (100 mM). Protein staining was by amido black, and immunoblots were developed with the ECL detection system. The migration distances in SDS-PAGE of proteins of known molecular mass are indicated.

Recognition of the E2 Subunits of OGDC, BCOADC, and of Protein X by Anti-CF&O Antibody. We have recently shown that the monospecific anti-CF3CO antibody cannot discriminate between CF3CO-Lys, the predominant adduct present on CF3COproteins, and lipoic acid, the prosthetic group of the E2 subunit of PDC (34, 35). The E2 subunit of OGDC is structurally and functionally closely related to the E2 subunit of PDC and the E2 subunit of BCOADC, as well as to protein X, a constituent of PDC. A striking similarity among these subunit proteins is the presence of one (OGDC, BCOADC, protein X) or two (PDC) lipoyl binding domains to which the prosthetic group lipoic acid is covalently attached at the NG-amine group of a critical lysine residue (49,50). Therefore, anti-CF3CO antibody should also recognize, besides the E2 subunit of OGDC, the E2 subunit of BCOADC and protein X. In fact, on immunoblots, anti-CFsCO antibody strongly reacted with human protein X (Figure 3A) present as a constituent of PDC when isolated as a whole complex from human heart tissue (36). Similarly, anti-CF3CO antibody did recognize the E2 subunit of rat BCOADC (Figure 3C). The recognition of these two target proteins, as well as that of the E2 subunits of OGDC (Figure 3B) and PDC (Figure 3A), by anti-CFsCO antibody was sensitive to competition by CF&O-Lys, 6(RS)-lipoicacid, and lipoyl-Lys, but not (with the exception of, for as of yet unknown reasons, the E2 subunit of BCOADC (Figure 3C)) to competition by L-lysine, indicating that the recognition of the lipoylated target proteins by anti-CF&O antibody was highly specific and targeted toward the lipoic acid moiety of these antigenic targets. Note that reactivity of anti-CF3CO antibody was exclusively with the lipoylated target proteins (i.e., the corresponding E2 subunits and protein X) of the tested complexes; no reactivity with the nonlipoylated subunit components of the complexes (i.e., Ela, Elp, E3) was evident. For unknown reasons, a staining a t around 66 kDa, which correlated with neither protein staining nor the presence and absence of inhibitors, was noted on the immunoblots of the rat BCOADC (Figure 3C). Recognition of Protein X and of the E2 Subunits of OGDC and BCOADC by Sera of Patients with Halothane Hepatitis. There is ample evidence indicat-

ing that sera from patients with halothane hepatitis contain antibodies which recognize discrete CF3COproteins, as well as CF3CO-RSA, to varying extents when tested by immunoblotting and/or ELISA-based analysis (51-53), with some sera not recognizing CF3CO-proteins on immunoblots, but in ELISA-based analysis only (52). Thus, if the concept of immunochemical mimicry of CF3CO-Lys by 6(RS)-lipoicacid is valid, those sera of patients with halothane hepatitis which recognize CF3CO-proteins should also recognize lipoylated proteins; that is, patients' sera might contain (a) discrete population(s) of autoantibodies which, analogous to the monospecific anti-CF3CO antibody, cannot discriminate between CF3COproteins and lipoylated proteins. In fact, autoantibodies in patients' sera were recently identified (35)which could not discriminate the immunochemically recognized determinants on CF3CO-proteins (Le., CF3CO-Lys) and the E2 subunit of PDC (i.e., lipoic acid), respectively, based on the very close structural relatedness of CF3CO-Lys and 6(RS)-lipoicacid. Here we show that, from a pool of 35 patients' sera available to us for analysis, 17 sera recognized CF3CO-RSA as a model antigen (Table 1). From the subset of these latter sera, 13 sera recognized human protein X (shown for one serum in Figure 4A,lane 2), 13 sera recognized the E2 subunit of human OGDC (shown for one serum in Figure 4B, lane 2), and 9 sera recognized the E2 subunit of rat BCOADC (shown for one serum in Figure 4C, lane 2) as antigenic targets on immunoblots. Note that, in keeping with our earlier reports (34, 35), the E2 subunit of PDC was also recognized (Figure 4A, lane 2) by 13 of the 17 patients' sera. The titers of autoantibodies in the individual sera, directed toward protein X and the E2 subunits of OGDC and BCOADC, ranged from 1 5 0 up to at least 1:1600 when tested on immunoblots (not shown). Except for some sera (4 sera of 16 sera tested in this regard) which also recognized (Figure 5A) the E3 subunit of PDC (or a proteolytic fragment of the E2 subunit of PDC which appears to comigrate with the E3 subunit): the reactivity of patients' sera was usually restricted to the lipoylated subunit proteins (Figure 4A-C, lanes 2) and rarely extended to the non-lipoylated subunit proteins (i.e., Ela, S. J. Yeaman, unpublished observation.

Chem. Res. Toxicol., Vol. 8, No. 5, 1995 741

Lipoylated Autoantigens in Halothane Hepatitis A

1

2

3

B

4

1

3

C

4

1

2

3

4

- - lmmunoblot

lmmunoblot

I

2

PDC

lmmunoblot

OGDC

BCOADC

Figure 4. Recognition of protein X and E2 subunit proteins by sera of patients with halothane hepatitis. Immunoblots of purified complexes of (A) human PDC (1p g of proteidcm slot width), of (B) human OGDC (1pg of proteidcm slot width), and of (C) rat BCOADC (2 pg of proteidcm slot width), respectively, were incubated (lanes 2) with (A) serum of patient 8 (sera collection 2; 1:200 dilution), with (B) serum of patient 4 (sera collection 2; 1:1600 dilution), and with (C) serum of patient 2 (sera collection 2; 1:400). Corresponding immunoblots were incubated with t h e serum (1:400 dilution) of a halothane-exposed but healthy control individual (lanes 3) or with the serum (1:400 dilution) of a n unexposed individual (lanes 4). Protein staining (lanes 1)was by amido black, and immunoblots were develoDed with the ECL detection svstem. The migration distances in SDS-PAGE of proteins of known molecular mass are indicated. Table 1. Recognition of Antigenic Targets b y Sera of Patients with Halothane H e p a t i t i e antigenic targets

E2 subunits of patients’ sera

CF3CO-RSA

protein X

OGDC

BCOADC

collection 1 collection 2 collection 3 control serab control seraC

7 (18) 7 (12)

5 5

5 5

4 2

3(5) - (4)b

3

3

3

- (4Y

-

-

-

a The sera of patients with halothane hepatitis and corresponding control sera were tested on immunoblots for reactivity toward the indicated antigens as described in the Experimental Procedures. From the total number of sera of patients in each collection (figures in parentheses) available for analysis, only the subset of those sera that recognized CF3CO-RSA on immunoblots in a manner sensitive to competition by CF3CO-Lys and G(RS)-lipoic acid but not to competition by blysine was further analyzed for reactivity with the antigenic targets a s indicated. Sera collection 1was from A. J. Gandolfi‘s laboratory; sera collection 2 was from J. G. Kenna’s laboratory; sera collection 3 was from L. Ranek‘s laboratory. Figures indicate the number of patients’ sera reacting with the respective target antigen. -, No reactivity detectable. Sera of halothane-exposed, but healthy control individuals. Sera from unexposed control individuals.

Elp, E3) present in the native complexes (Figure 4A-C, lanes 1) used in these experiments. When analyzed individually, six sera of the subset of 17 patients’ sera which reacted with CF3CO-RSA (Table 1)recognized the E2 subunits of human OGDC and PDC and rat BCOADC, and human protein X. Four sera recognized the E2 subunits of human OGDC and PDC, and human protein X. Two sera recognized the E2 subunits of human PDC and of rat BCOADC, and human protein X. Two sera recognized the E2 subunit of human OGDC only; one serum recognized the E2 subunits of human PDC and of rat BCOADC only, and one serum recognized none of these antigenic targets to any noticeable degree. None of the lipoylated or non-lipoylated antigenic targets was recognized by sera from halothane-exposed, but healthy individuals (Figure 4A-C, lanes 3) or by sera from unexposed human individuals (Figure 4A-C, lanes 4). Additional experiments were performed to further characterize patients’ sera with respect to recognition of lipoylated target antigens, which appeared to be complex. In a first set of experiments, we randomly selected patients’ sera for antibody exchange experiments. As demonstrated with the serum of patient 34 (sera collec-

A

n

E? -

protein X D

E3 E2

w

IC

-- *

Q

PI

i i

t

Figure 5. Competitive immunoblotting analysis of the recognition of antigenic targets by sera of patients with halothane hepatitis. (A) The serum of patient 11(sera collection l),at a dilution of 1:1500, was reacted on immunoblots with the purified complex of human PDC (1pg of proteidcm slot width) in the absence (Control) and in the presence of CF3CO-Lys (50 mM) and 6(RS)-lipoicacid (50 mM). (B) The serum of patient 4 (sera collection 21, at a dilution of 1:800, was reacted on immunoblots with the purified complex of human OGDC (1pg of proteidcm slot width) in the absence (Control) and in the presence of CF3CO-Lys (25 mM) and 6(RS)-lipoic acid (25 mM). Immunoblots were developed with the ECL detection system. Note t h a t in the immunoblotting system used the apparent molecular masses of the E2 subunits of human PDC (A) and human OGDC (B), respectively, are 64 and 52 kDa while t h a t of the E 3 subunit is about 55 kDa for both complexes.

tion l), a discrete population of autoantibodies could be selectively affinity-purified on immunoblot-immobilized E2 subunit of human OGDC and spontaneously exchanged.from the source E2 subunit of human OGDC to the target antigen CF3CO-RSA (Figure 6). The recognition of CF3CO-RSAby this discrete population of patients’ autoantibodies was sensitive to competition by CF3COLys (50 mM), 6(RS)-lipoic acid (25 mM), and lipoyl-Lys (20 mM), but not to competition by L-lysine (50 mM). These experiments confirmed the occurrence in patients’ sera of a discrete population of autoantibodies whose properties were identical to those of the monospecific CF3CO antibody. Neither binding to lipoylated target proteins nor transfer of antibodies to CF3CO-RSA was observed with sera from both halothane-exposed and unexposed healthy control individuals (not shown). Note that, with some patients’ sera, transfer of autoantibodies

742 Chem. Res. Toxicol., Vol. 8, No. 5, 1995

I

Frey et al. subunits of OGDC, BCOADC, and PDC) by these latter patients’ autoantibodies remain to be determined.

$

a

Discussion

1

Figure 6. Antibody exchange immunochemistry. The serum of patient 34 (sera collection 11, at a dilution of 1:15, was incubated on immunoblots with purified human OGDC (6 pgl cm slot width) as shown in Figure 4B, lane 2. Four disks (2.5 x 2.5 mm) containing the E2 subunit of OGDC and bound autoantibody as the sole source for human autoantibody were then cut from the immunoblots and coincubated with one target disk (2.5 x 5 mm) stemming from immunoblots that contained CF3CO-RSA (0.9pg of proteidcm slot width) as the sole target antigen in the absence (Control) and the presence of CF3COLys (50 mM), 6(RS)-lipoic acid (25 mM), lipoyl-Lys (20 mM), and L-lysine (50 mM). Blots were developed with the ECL detection system, the films obtained were subjected to scanning densitometry as described (4), and relative signals, corrected for background signal of each individual disk, were plotted; ( 0 ) indicates that no signal above background was detectable. Inset: corresponding immunoblots. Arrowhead indicates the relative migration position of CF3CO-RSA.

from the source E2 subunit of OGDC to the target CF3CO-RSA was below the detection limit of the experimental approach employed here. In a second set of experiments, the recognition of E2 subunits by patients’ sera on immunoblots was performed in the presence of the competitors CF3CO-Lys (50 mM) and 6(RS)-lipoicacid (50 mM). With some patients’ sera, as demonstated with the serum of patient 11 (sera collection 1)and the E2 subunit of PDC as target antigen, the recognition of the lipoylated antigens was slightly affected (about 40%reduced as measured by densitometric scanning (data not shown)) in the presence of CF3CO-Lys (Figure 5A) and almost completely abolished in presence of G(RS)-lipoicacid (Figure 5A), suggesting the presence in these sera of a considerable proportion of autoantibodies specific for the lipoyl motif of the E2 subunit protein. The recognition by the serum of patient 11 of the E3 subunit of PDC, a non-lipoylated target antigen the recognition of which was noted with some patients’ sera, was not affected by CF3CO-Lys and 6(RS)lipoic acid (Figure 5A). In contrast, as noted in previous experiments (35, 54) and demonstrated here with the serum of patient 4 (sera collection 2) and the E2 subunit of OGDC as target antigen, some of the patients’ sera recognized lipoylated antigens in a manner not or only marginally sensitive to competition by CF3CO-Lys (25 mM) and G(RS)-lipoic acid (25 mM (Figure 5B)). These data suggest the occurrence in these sera of higher proportion(s) of autoantibodies recognizing epitopes on native, lipoylated target proteins other than the central lipoic acid motif. The sites of recognition (i.e., epitopes) of the lipoylated antigens (i.e., protein X and the E2

Anti-CFSCO antibodies, monospecific toward CF3COproteins, recognize previously unidentified proteins of 52 kDa, constitutively expressed in several tissues of the rat (26) and in the liver of human individuals (26) not exposed to the anesthetic agent halothane. Here, we have identified the E2 subunit of OGDC as one component of the protein(s) of 52 kDa. Furthermore, we demonstrate that the E2 subunits of OGDC and of BCOADC, as well as protein X, a component of PDC, mimic CF3CO-proteins and that they are autoantigens associated with halothane hepatitis. These conclusions are based on three major observations. First, the E2 subunit of OGDC was immunoaffinity-purified as one of several protein components of the proteins of 52 kDa from solubilized rat heart homogenate by the use of the antiCF3CO antibody as an affinity matrix. The identity of the E2 subunit of OGDC was established based on 100% sequence homology between several internal peptides and corresponding regions of the deduced amino acid sequences of rat (44)and human (45)liver E2 subunit of OGDC. Second, the E2 subunit of rat and human OGDC and the E2 subunit of rat BCOADC and protein X are all recognized by anti-CF&O antibody. The recognition of these antigens on immunoblots by anti-CF&O antibody is abolished in the presence of CF&O-Lys, 6(RS)lipoic acid, and lipoyl-Lys, but not in the presence of L-lysine, indicating that the presence of the prosthetic group lipoic acid in the two E2 subunit proteins and in protein X, is responsible for their recognition by anti-CF3CO antibody. This is in keeping with earlier data obtained in this laboratory (35)showing that the presence of lipoic acid in Llip, the lipoylated form of the recombinantly expressed (55) inner lipoyl domain of the E2 subunit of human PDC, was sufficient but necessary for the recognition of Llip by anti-CF&O antibody. Finally, the E2 subunits of OGDC and BCOADC, as well as protein X are recognized as autoantigens in their native, non-trifluoroacetylated form by human autoantibodies present in sera of patients with halothane hepatitis, but not in sera of halothane-exposed or unexposed healthy control individuals. As demonstrated with the serum of patient 34 (sera collection 1) by antibody exchange, the former sera can contain (a) discrete population(s) of autoantibodies which recognized, after prior affinity purification on the E2 subunit of the human OGDC, the model antigen CF3CO-RSA in a manner sensitive to competition by CF&O-Lys, G(RS)-lipoicacid, and lipoylLys, but not to competition by L-lysine. Previous data from our laboratory indicated that CF3CO-Lys and lipoic acid, respectively, are the minimal but sufficient structural entities present on CF3CO-proteins and the E2 subunit of PDC (or subdomains thereof)which immunochemically, by the use of anti-CF&O antibody, were not discernible and constituted the basis for molecular mimicry of CF3CO-proteins by the E2 subunit of PDC (34,35). The inability of antibodies to discriminate between CF3CO-proteins and lipoylated antigens is not a unique property of the experimentally elicited anti-CF3CO antibody. Thus, in patients with halothane hepatitis, but not in halothane-exposed healthy control individuals or in unexposed individuals, (a) discrete population(s) of autoantibodies which cannot discriminate between CF3-

Lipoylated Autoantigens in Halothane Hepatitis CO-Lys and lipoic acid is (are) a regular part of the immune response occurring in the disease state. Consequently, because they all carry covalently bound lipoic acid as a prosthetic group, protein X and the E2 subunits of OGDC and BCOADC, along with the E2 subunit of PDC (351, are autoantigens associated with halothane hepatitis. The immunological significance of molecular mimicry of CF3CO-Lys by lipoic acid and its relation to halothane metabolism-dependent formation of CFsCO-proteins is not yet clear. However, protein X and the E2 subunits of OGDC and BCOADC (as well as that of PDC) are constitutively expressed as part of self throughout the body; it might therefore be valid to assume that, except in rare cases (see below), individuals are immunologically tolerant toward these proteins. One might speculate that individuals, all of whom produce CF3CO-proteins after exposure to halothane, could profit from this situation because immunological tolerance, established toward the lipoic acid motif, might coincidentially also provide tolerance toward CF~CO-L~S, the major haptenic structure present in CF3CO-proteins. If so, the occurrence in patients’ sera of a discrete population of autoantibodies directed against lipoic acid, the motif common to protein X and the E2 subunits of OGDC and BCOADC, would indicate that immunological tolerance toward lipoic acid is not developed in individuals who prove susceptible to halothane hepatitis. The mechanisms of breakdown of presumed tolerance toward the lipoic acid motif remain obscure at present. By the use of anti-CF3CO antibody as a probe, we have identified (41 a fraction of patients (-70%) with halothane hepatitis who exhibit, in the liver, low levels of immunodetectable protein(s) of 52 kDa, one component of which we have identified as the E2 subunit of OGDC (this report), and of the E2 subunit of PDC (34, 351, while a remaining two of the,seven patients examined exhibited normal levels of the E2 subunit of PDC and the protein(s) of 52 kDa, despite the fact that these two patients suffered from a very severe form of hepatitis (41). I t is currently not known if this apparent low expression of the E2 subunit of PDC and the protein(s) of 52 kD is a contributing factor in the etiology of halothane hepatitis; it is also not yet known if these irregularities are acquired and whether they are tissue specific or constitutive throughout the body. The importance of the levels of expressed self-antigen in establishing tolerance to self was demonstrated in the B- and T-cell repertoires of strains of mice which express different concentrations of transgene-encoded hen egg lysozyme (56). T-Cells were shown t u be tolerant in all strains of mice in which lysozyme was expressed, irrespective of the transgene concentration, while B-cell tolerance did not develop below critical serum lysozyme concentrations and could only be restored by induction of elevated expression levels of the transgene. Our data suggest that an “endogenous hapten”-carrier system (comprised within various lipoyl domains) molecularly mimics a hapten-carrier system (comprised within CF3CO-proteins) that is generated in situ upon exposure of individuals to the exogenous agent halothane. One might speculate that, in healthy control individuals, where protein X and the E2 subunits of OGDC, BCOADC, and PDC are expressed a t “normal levels” (41), tolerance in the B-cell and the T-cell repertoire might be established, while in patients who lack adequate expression of these proteins, such tolerance might be poorly developed. A sudden challenge of the latter individuals by a high

Chem. Res. Toxicol., Vol. 8, No. 5, 1995 743 density of CF3CO-proteins, comprising the molecular mimic of lipoic acid, namely, CF~CO-L~S, could lead to the production of the autoantibodies which are not able to discriminate between CF3CO-Lys and lipoic acid. Alternatively, in susceptible individuals, CF3CO-Lysrelated epitopes in high density might provide additional signals to overcome limits of T-cell tolerance to selfproteins when presented in nonphysiological concentrations (57). In addition, a very recent report indicated the occurrence of a repertoire of natural antibodies reacting with lipoic acid in nonimmunized BALB/c mice (58). Although not yet demonstrated, natural anti-lipoic acid autoantibodies might also occur in susceptible humans, and the transient generation of molecular mimics of lipoic acid, namely, CF3CO-Lys,might lead to an enhancement of the production of anti-lipoic acid antibodies. The immune response in patients with halothane hepatitis is complex, however. Patients’ sera contain discrete populations of antibodies with specificities toward epitopes comprising not the lipoic acid motif (or its structural mimic CF3CO-Lys (see below))and/or parts of the carrier molecules. Thus, we detected in patients’ sera autoantibodies which recognized protein X and the E2 subunits of OGDC and BCOADC in a manner not, or only partially, sensitive to competition by even high concentrations of CF3CO-Lys or G(RS)-lipoic acid. These data suggest the recognition by these autoantibodies of epitopes not comprising the lipoic acid motif, of conformational epitopes which may or may not include the lipoic acid motif, or of epitopes accessible after spreading of the immune response (59)to cryptic determinants of protein X and the various E2 subunit proteins. Parallel to our findings, recent data indicated that patients’ sera contain discrete populations of autoantibodies toward the native, non-trifluoroacetylated form of human microsomal carboxylase (601, the native form of protein disulfide isomerase (181, and the native form of a neoantigen of 58 kDa with high amino acid sequence homology to but not the activity of phosphatidylinositol-specific phospholipase C (61). These antigenic targets had previously been isolated, based on their recognition by patients’ sera, in their corresponding trifluoroacetylated form (i.e., as CF3CO-proteins) as prominent neoantigens from liver microsomes of halothane-exposed rats (18,19, 61). The collective evidence from these data suggested the occurrence in patients’ sera of (a) discrete population(s) of autoantibodies which recognized the haptenic CF3CO-Lys motif, epitopes which may not comprise the CF3CO-Lys motif, conformational epitopes which may or may not include the CF3CO-Lys motif, or epitopes accessible after spreading of the immune response to cryptic determinants of the corresponding carrier molecule. Intra- and/ or intermolecular determinant spreading, i.e., the progressive recognition by autoantibodies and/or autoreactive T-cells of epitopes other than the initially immunogenic epitope in the course of an ongoing immune reaction, has been demonstrated in murine models of T-cell-mediated autoimmune diseases such as murine experimental allergic encephalomyelitis (62) or in the development of diabetes in non-obese diabetic (NOD)-mice (63). Intramolecular determinant spreading in the course of halothane hepatitis is indirectly suggested as a mechanism that leads to the observed diverse repertoire of autoantibodies directed toward epitopes carrying either the lipoic acid or the CF3CO-Lys motifs as well as toward chemically unmodified epitopes of the corresponding carrier molecules. The fact that some patients’ sera also

744 Chem. Res. Toxicol., Vol. 8, No. 5, 1995

recognize the E3 subunit protein (Figure 5A) as an autoantigen might be suggestivefor intermolecular spreading as well. Protein X and the E2 subunits of OGDC, BCOADC, and PDC, have been identified as autoantigens (47,6468) in primary biliary cirrhosis, an autoimmune cholestatic liver disease characterized by progressive obliteration of intrahepatic bile ducts. The lipoylated domains of these autoantigens are immunodominant epitopes in that sera of human individuals afflicted with primary biliary cirrhosis preferentially recognize such domains (69-71). A dominant role of lipoic acid (which may be mimicked by octanoic acid) in the binding of human sera was demonstrated with the Escherichia coli E2 subunit of PDC as antigen (611, and recently, indirect evidence suggested that the lipoic acid moiety might specifically be recognized by (a) discrete population(s) of antibodies present in patients’ sera because such sera tend to exhibit a higher affinity toward the lipoylated form (Llip) than toward the un-lipoylated form (Ulip) of recombinantly expressed inner lipoyl domain of the human E2 subunit of the PDC (72). In the light of molecular mimicry of CF3CO-Lys by lipoic acid, it is interesting to note that, in one study (51), antisera of 17%of patients afflicted with PBC cross-reacted with CFSCO-RSA; however, such cross-reactivities were not restricted to sera of patients with primary biliary cirrhosis. We have also found recognition on immunoblots of CF&O-RSA by one out of ten sera of patients with primary biliary cirrho~is.~ The notion that protein X and the E2 subunits OGDC, BCOADC, and PDC are autoantigens associated with both conditions clearly raises the possibility of a hitherto unknown cross-sensitization of patients for halothane hepatitis and primary biliary cirrhosis. This might be of immediate clinical significance because patients with primary biliary cirrhosis, whether phenotypically overt or not, might be at greater risk for an aggravating immune reaction when exposed to halothane. Any pathogenic role of patients’ autoantibodies directed against protein X and the E2 subunits of OGDC, BCOADC, and PDC (and, because of molecular mimicry, toward CFSCO-proteins also) in the etiology of halothane hepatitis remains to be established. Moreover, the exact role of protein X, and the E2 subunits of OGDC, BCOADC, and PDC, in the etiology of halothane hepatitis remains to be demonstrated. In addition, although associated with halothane hepatitis, neither the autoantibodies nor the autoantigens characterized here can be considered as specific markers of the disease because of their association with other forms of autoimmune disease. Also, it remains to be elucidated whether the occurrence of a repertoire of self motifs (i.e., lipoic acid) that mimic trifluoroacetylated motifs (Le., CF3CO-Lys) might help to establish immunological tolerance of individuals toward CFsCO-proteins and whether aberrant expression of such self motifs might render individuals susceptible to immunological reactions toward offending CF3COproteins. In order to answer these questions, further experimentation is clearly needed. Regardless, molecular mimicry of drug metabolite-modified protein adducts by epitopes expressed on endogenous self-proteins could be of general relevance to immune-mediated adverse drug reactions.

Acknowledgment. This work was supported by the Swiss National Science Foundation (Grant 31-252.91) U. Christen and J. Gut, unpublished observation.

Frey et al.

and the Roche Research Foundation. J.G. was the recipient of Research Career Development Awards 3-018.0.87 and 3130-009067.87/2 from the Swiss National Science Foundation.

References (1) Kenna, J. G., Neuberger, J., and Williams, R. (1987) Identification

by immunoblotting of three halothane-induced liver microsomal polypeptide antigens recognized by antibodies in sera from patients with halothane-associated hepatitis. J . Pharmacol. Exp. Ther. 242, 733-740. (2) Kenna, J. G., Satoh, H., Christ, D. D., and Pohl, L. R. (1988) Metabolic basis for a drug hypersensitivity: antibodies in sera from patients with halothane hepatitis recognize liver neoantigens that contain the trifluoroacetyl group derived from halothane. J . Pharmacol. Exp. Ther. 245, 1103-1109. (3) Harris, J. W., Pohl, L. R., Martin, J. L., a n d h d e r s , M. W. (1991) Tissue acylation by the chlorofluorocarbon substitute 2,2-chlorol,l,l-trifluorcethane.Proc. Natl. Acad. Sci. U.SA.88,1407-1410. (4) Huwyler, J., Aeschlimann, D., Christen, U., and Gut, J. (1992) The kidney as a novel target tissue for protein adduct formation associated with metabolism of halothane and the candidate chlorofluorocarbon replacement 2,2,-dichloro-l,l,l-trifluoroethane. Eur. J . Biochem. 207, 229-238. ( 5 ) Harris, J. W., Jones, J. P., Martin, J. L., LaRosa, A. C., Olson, M. J., Pohl, L. R., and Anders, M. W. (1992) Pentahaloethanebased chlorofluorocarbon substitutes and halothane: correlation of in vivo hepatic protein trifluoroacetylation and urinary trifluoroacetic acid excretion with calculated enthalpies of activation. Chem. Res. Toxicol. 5, 720-725. (6) Satoh, H., Fukuda, Y., Anderson, D. K., Ferrans, V. J., Gillette, J. R., and Pohl., L. R. (1985) Immunological studies on the mechanism of halothane-induced hepatotoxicity: Immunohistochemical evidence for trifluoroacetylated hepatocytes. J . Pharmacol. Exp. Ther. 233, 857-862. (7) Vergani, D., Mieli-Vergani, G., Alberti, A,, Neuberger, J., Davies, A. L. W. F., and Williams, R. (1980) Antibodies to the surface of halothane altered rabbit hepatocytes in patients with halothane associated hepatitis. N . Engl. J. Med. 303, 66-71. ( 8 ) Hubbard, A. K., Roth, T. P., Schuman, S., and Gandolfi, A. J. (1989) Localization of halothane-induced antigen in situ by specific anti-halothane metabolite antibodies. Clin. Exp. Immunol. 76,422-427. (9) Heijink, E., De Matteis, F., Gibbs, A. H., Davies, A., and White, I. N. H. (1993) Metabolic activation of halothane to neoantigens inC57B1/10 mice: immunochemical studies. Eur. J. Pharmacol. 248, 15-25. (10) Kenna, J. G., Neuberger, J., and Williams, R. (1988) Evidence for expression in human liver of halothane-induced neoantigens recognized by antibodies in sera from patients with halothane hepatitis. Hepatology 8, 1635-1641. (11) Pohl, L. R., Kenna, J. G., Satoh, H., Christ, D. D., and Martin, J. L. (1989) Neoantigens associated with halothane hepatitis. Drug Metub. Rev. 20, 203-217. (12) Huwyler, J.,and Gut, J.(1992) Exposure to the chlorofluorocarbon substitute 2,2-dichloro-l,l,l-trifluoroethane and the anesthetic agent halothane is associated with the transient protein adduct formation in the heart. Biochem. Biophys. Res. Commun. 184, 1344-1349. (13) Kenna, J . G., Martin, J. L., Satoh, H., and Pohl, L. R. (1990) Factors affecting the expression of trifluoroacetylated liver microsomal protein neoantigens in rats treated with halothane. Drug Metab. Dispos. 18, 788-793. (14) Christen, U., Burgin, M., and Gut, J. (1991) Halothane metabolism: Kupffer cells carry and partially process trifluoroacetylated protein adducts. Biochem. Biophys. Res. Commun. 175,256-262. (15) Rubinstein, D., Roska, A. K., and Lipsky, P. E. (1987) Antigen presentation by liver sinusoidal lining cells after antigen presentation in vivo. J . Immunol. 138, 1377-1382. (16) National Halothane Study (1966) Summary of the National Halothane Study: possible association between halothane anesthesia and postoperative hepatic necrosis. J . Am. Med. Assoc. 197, 121-134. (17) Pohl, L. R., Satoh, H., Christ, D. D., and Kenna, J . G. (1988) The immunologic and metabolic basis of drug hypersensitivities. Annu. Reu. Pharmacol. Toxicol. 28, 367-387. (18) Martin, J. L., Kenna, J. G., Martin, B. M., Thomassen, D., Reed, G. F., and Pohl, L. R. (1993) Halothane hepatitis patients have serum antibodies that react with protein disulfide isomerase. Hepatology 18, 858-863. (19) Satoh, H., Martin, B. M., Schulik, A. H., Christ, D. D., Kenna, J . G., and Pohl, L. R. (1989) Human anti-endoplasmic reticulum

Lipoylated Autoantigens in Halothane Hepatitis antibodies in sera from patients with halothane-induced hepatitis are directed against a trifluoroacetylated carboxylesterase. Proc. Natl. Acad. Sci. U.S.A. 86, 322-326. (20) Butler, L. E., Thomassen, D., Martin, J. L., Martin, B. M., Kenna, J. G., and Pohl, L. R. (1992)The calcium-binding protein calreticulin is covalently modified in rat liver by a reactive metabolite of the inhalation anesthetic halothane. Chem. Res. Toxicol. 5, 406-410. (21) Pumford, N. R., Martin, B. M., Thomassen, D., Burris, J. A., Kenna, J. G., Martin, J . L., and Pohl, L. R. (1993) Serum antibodies from halothane hepatitis patients react with the endoplasmic reticulum protein ERp72. Chem. Res. Toxicol. 6, 609-615. (22) Davila, J. C., Martin, B. M., and Pohl, L. R. (1992) Patients with halothane hepatitis have serum antibodies directed against glucose regulated stress protein GRP78IBiP. Toxicologist 12,255. (23) Thomassen, D., Martin, B. M., Martin, J. L., Pumford, N. R., and Pohl, L. R. (1989) The role of a stress protein in the development of a drug-induced allergic response. Eur. J. Pharmacol. 183,11381139. (24) Satoh, H., Gillette, J. R., Davies, H. W., Schulick, R. D., and Pohl, L. R. (1985) Immunochemical evidence of trifluoroacetylated cytochrome P450 in the liver of halothane treated rats. Mol. Pharmacol. 28,468-474. (25) Neuberger, J. M., Kenna, J . G., and Williams, R. (1987)Halothane hepatitis: attempt to develop an animal model. Int. J . Immunopharmacol. 9, 123-132. (26) Christen, U.,Biirgin, M., and Gut, J. (1991)Halothane Metabolism: Immunochemical evidence for molecular mimicry of trifluoroacetylated liver protein adducts by constitutive polypeptides. Mol. Pharmacol. 40, 390-400. (27) Kappler, J. W., Roehm, N., and Marrack, P. C. (1987) T-cell tolerance by clonal elimination in the thymus. Cell 49,273-280. (28) Kisielow, P., Bliithmann, H., Staerz, U.D., Steinmetz, M., and von Boehmer, H. (1988) Tolerance in T-cell receptor transgenic mice involves deletion of nonmature CD4+8&thymocytes. Nature 333,742-746. (29) Kappler, J. W., Staerz, U.,White, J., and Marrack, P. C. (1988) Self-tolerance eliminates T cells specific for Mls-modified products of the major histocompatibility complex. Nature 332, 35-40. (30) Ramsdell, F., and Folkwes, B. J. (1990) Clonal deletion versus clonal anergy: The role of the thymus in inducing self tolerance. Science 248, 1342-1348. (31) Schwartz R. H. (1990)A cell culture model for T lymphocyte clonal anergy. Science 248, 1349-1356. (32) Sprent, J., Gao, E.-K., and Webb, S. R. (1990)T cell reactivity to MHC molecules: Immunity versus tolerance. Science 248, 13571363. (33) Ohashi, P. S., Oehen, S., Buerki, K., Pircher, H., Ohashi, C. T., Odermatt, B., Malissen, B., Zinkernagel, R. M., and Hengartner, H. (1991)Ablation of “tolerance” and induction of diabetes by virus infection in viral antigen transgenic mice. Cell 65, 305-317. (34) Christen, U.,Jeno, P., and Gut , J. (1993) Halothane metabolism: The dihydrolipoamide acetyltransferase subunit of the pyruvate dehydrogenase complex molecularly mimics trifluroacetyl-protein adducts. Biochemistry 32, 1492- 1499. (35) Christen, U., Quinn, J., Yeaman, S. J., Kenna, J . G., Clarke, J. B., Gandolfi, A. J., and Gut, J. (1994) Identification of the dihydrolipoamide acetyltransferase subunit of the human pyruvate dehydrogenase complex as an autoantigen in halothane hepatitis. Molecular mimicry of trifluoroacetyl-lysine by lipoic acid. Eur. J . Biochem. 223, 1035-1047. (36) Palmer, J. M., Bassendine, M. F., James, 0. F. W., and Yeaman, S. J. (1993) Human pyruvate dehydrogenase complex as an autoantigen in primary biliary cirrhosis. Clin. Sci. 85, 289-293. (37) Shimomura, Y., Nanaumi, N., Suzuki, M., Popov, K. M., and Harris, R. A. (1990) Purification and partial characterization of branched chain a-ketoacid dehydrogenase kinase from rat liver and rat heart. Arch. Biochem. Biophys. 283, 283-299. (38) Laemmli, U. K. (1970) Cleavage of structural proteins during assembly at the head of bacteriophage Td. Nature 227, 680-685. (39) Towbin, H., Staehelin, T., and Gordon, J. (1979)Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. U.S.A. 76, 4350-4354. (40) Hammarback, J. A., and Vallee, R. B. (1990)Antibody exchange immunochemistry. J . Biol. Chem. 265, 12763-12766. (41) Gut, J., Christen, U., Huwyler, J., Burgin, M., and Kenna, J . G. (1992) Molecular mimicry of trifluoroacetylated human liver protein adducts by constitutive proteins and immunochemical evidence for its impairment in halothane hepatitis. Eur. J . Biochem. 210, 569-576. (42) O’Farrell, P. H. (1975) High resolution two-dimensional electrophoresis of proteins. J . Biol. Chem. 250, 4007-4021.

Chem. Res. Toxicol., Vol. 8, No. 5, 1995 746 (43) Jeno, P., Mini, T., Moes, S., Hintermann, E., and Horst, M. (1995) Internal sequences from proteins digested in polyacrylamide gels. Anal. Biochem. 224, 75-82. (44) Nakano, K , Matuda, S., Yamanaka, T., Tsubouchi, H., Nakagawa, S., Titani, K., Ohta, S., and Miyata, T. (1991) Purification and molecular cloning of succinyltransferase of the rat a-ketoglutarate dehydrogenase complex. J . Biol. Chem. 266, 19013-19017. (45) Nakano, K., Matuda, S., Sakamoto, T., Takase, C., Nakagawa, S., Ohta, S., Ariyama, T., Inazawa, J., Abe, T., and Miyata, T. (1993) Human dihydrolipoamide succinyltransferase: cDNA cloning and localization on chromosome 14q24.2-q24.3. Biochim. Biophys. Acta 1216, 360-368. (46) Wicking, C. A., Scholem, R. D., Hunt, S. M., and Brown, G. K. (1986)Immunochemical analysis of normal and mutant forms of human pyruvate dehydrogenase. Biochem. J. 239, 89-96. (47) Coppel, R. L., McNeilage, L. J., Surh, C. D., Van de Water, J., Spithill, T. W., Wittingham, S., and Gershwin, M. E. (1988) Primary structure of the human M2 mitochondrial autoantigen of primary biliary cirrhosis: Dihydrolipoamide acetyltransferase. Proc. Natl. Acad. Sci. U.S.A. 85, 7317-7321. (48) Thekkumkara, T. J., Ho, L., Wexler, I. D., Pons, G., Liu, T.-Ch., and Patel, M. S. (1988) Nucleotide sequence of a cDNA for the dihydrolipoamide acetyltransferase component of human pyruvate dehydrogenase complex. FEBS Lett. 240, 45-48. (49) Yeaman, S. J. (1986) The mammalian 2-oxoacid dehydrogenases: a complex family. Trends Biochem. Sci. 11, 293-296. (50) Perham, R. N. (1991) Domains, motifs, and linkers in 2-oxo acid dehydrogenase multienzyme complexes: A paradigm in the design of a multifunctional protein. Biochemistry 30, 8501-8512. (51) Martin, J . L., Kenna, J . G., and Pohl, L. R. (1990)Antibody assays for the detection of patients sensitized to halothane. Anesth. Analg. (N.Y.) 70, 154-159. (52) Kenna, J . G., Scatchard, K., van Pelt, F. N. A. M., and Knight, T. L. (1992)Antibodies to halothane-induced liver antigens in sera from patients with halothane hepatitis: evidence for a conformation-dependent antigen. Br. J . Clin. Pharmacol. 34, 154P. (53) Hastings, K. L., Thomas, C., Hubbard, A. K., and Gandolfi, A. J . (1991) Screening for antibodies associated with halothane hepatitis. Br. J. Anaesth. 67, 722-728. (54) Gut, J., Christen, U.,and Huwyler, J . (1993) Mechanisms of halothane toxicity: Novel insights. Pharmacol. “her. 58, 133155. (55) Quinn, J., Diamond, A. G., Masters, A. K., Brookfield, D. E., Wallis, N. G., and Yeaman, S. J. (1993)Expression and lipoylation in Escherichia coli of the inner lipoyl domain of the E2 component of the human pyruvate dehydrogenase complex. Biochem. J . 298, 81-85. (56) Adelstein, S., Pritchard-Briscoe, H., Anderson, T. A., Crosbie, J., Gammon, G., Loblay, R. H., Rasten, A,, and Goodnow, C. C. (1991) Induction of self-tolerance in T cells but not B cells of transgenic mice expressing little self antigen. Science 251, 1223- 1225. (57) Schild, H., Rotschke, O., Kalbacher, H., and Rammensee, H.-G. (1990) Limit of T cell tolerance to self proteins by peptide presentation. Science 247, 1587-1589. (58) Legastelois, S., Thomas, V., Quash, G., MQtais,M.-P., Tebib, J., Moreira, A., & Monier, J.-C. (1994)Naturally occurring antibodies reacting with lipoic acid: screening method, characterization and biochemical interest. J . Immunol. Methods 171, 111-119. (59) Lehmann, P. V., Sercarz, E. E., Forsthuber, T., Dayan, C. M., and Gammon, G. (1993) Determinant spreading and the dynamics of the autoimmune T-cell repertoire. Immunol. Today 14,203-208. (60) Smitt, N. C. M., Kenna, J. G., Harrison, D. J., Tew, D., and Wolf, C. R. (1993) Autoantibodies to hepatic microsomal carboxylesterase in halothane hepatitis. Lancet 342, 963-964. (61) Martin, J . L., Reed, G. F., and Pohl, L. R. (1993)Association of anti-58 kDa endoplasmic reticulum antibodies with halothane hepatitis. Biochem. Pharmacol. 46, 1247-1250. (62) Lehmann, P. V., Forsthuber, T., Miller, A,, and Sercaz, E. E. (1992)Spreading of T-cell autoimmunity to cryptic determinants of an autoantigen. Nature 358, 155-157. (63) Kaufman, D. L., Clare-Salzler, M., Tian, J., Forsthuber, T., Ting, G. S. P., Robinson, P., Atkinson, M. A., Sercarz, E. E., Tobin, A. J., and Lehman, P. V. (1993)Spontaneous loss of T-cell tolerance to glutamic acid decarboxylase in murine insulin-dependent diabetes. Nature 366, 69-72. (64) Van de Water, J., Fregeau, D., Davis, P., Ansari, A,, Danner, D. J., Leung, P., Coppel, R., and Gershwin, M. E.(1988)Autoantibodies of primary biliary cirrhosis recognize dihydrolipoamide acetyltransferase and inhibit enzyme function. J. Immunol. 141, 2321-2324. (65) Fussey, S. P. M., Guest, J. R., James, 0. F. W., Bassendine, M. F., and Yeaman, S. J. (1988) Identification and analysis of the major M2 autoantigens in primary biliary cirrhosis. Proc. Natl. Acad. Sci. U.S.A. 85, 8654-8658.

Frey et al.

746 Chem. Res. Toxicol.,Vol. 8, No. 5, 1995 (66) Fussey, S. P. M., Bassendine, M. F., James, 0. F. W., and Yeaman, S. J. (1990) Characterization of the reactivity of autoantibodies

in primary biliary cirrhosis. FEBS Lett. 246,49-53. (67) Fregeau, D. R., Prindiville, T., Coppel, R. L., Kaplan, M., Dickson, E. R., and Gershwin, M. E. (1990) Inhibition of a-ketoglautarate dehydrogenase activity by a distinct population of autoantibodies recognizing dihydrolipoamide succinyltransferase in primary biliary cirrhosis. Hepatology 11, 975-981. (68) Surh, C. D., Roche, T. E., Danner, D. J.,Ansari, A., Coppel, R. L., Prindiville, T., Dickson, E. R., and Gershwin, M. E. (1989) Antimitochondrial autoantibodies in primary biliary cirrhosis recognize cross-reactive epitopes on protein X and dihydrolipoamide acetyltransferase of pyruvate dehydrogenase complex. Hepatology 10,127-133. (69) Leung, P. S., Iwayama, T., Coppel, R. L., & Gershwin, M. E. (1990) Site-directed mutagenesis of lysine within the immunodominant autoepitope of PDC-E2. Hepatology 12,1321-1328.

(70) Van de Water, J., Gershwin, M. E., Leung, P., Ansari, A,, and Coppel, R. L. (1988) The autoepitope of the 74-kD mitochondrial autoantigen of primary biliary cirrhosis corresponds to the function2 site of dihydr&poamide acetyltransferase. J . Exp. Med. 167,1791-1799. (71) Fussey, S. P. M., Mi, S. T., Guest, J . R., James, 0. F. W., Bassendine, M. F., and Yeaman, S. J. (1990)Reactivity of primary biliary cirrhosis sera with Escherichia coli dihydrolipoamide acetyltransferase (E2p): Characterization of the main immunogenic region. Proc. Natl. Acad. Sci. U.S.A. 87,3987-3991. (72) Quinn, J., Diamond, A. G., Palmer, J. M., Bassendine, M. F., James, 0. F. W., and Yeaman, S. J. (1993) Lipoylated and unlipoylated domains of human PDC-E2 as autoantigens in primary biliary cirrhosis: Significance of lipoate attachment. Hepatology 18,1384-1391.

TX9500150