autoantibodies of primary biliary cirrhosis recognize ...

0022-1767/88/1417-2321$02.00/0 THEJOURNAL OF IMMUNOLOGY Copyright 0 1988 by The American Association of Immunologists

Vol. 141.

2321-2324. No. 7.October 1. 1988 Printed In U . S . A .

AUTOANTIBODIES OFPRIMARYBILIARYCIRRHOSISRECOGNIZE DIHYDROLIPOAMIDE ACETYLTRANSFERASEAND INHIBITENZYME FUNCTION' JUDY VAN DE WATER,* DAVID FREGEAU,* PAUL DAVIS,* AFTAB ANSARI,' DEAN DANNER,' PATRICK LEUNG,* ROSS COPPEL,' AND M. ERIC GERSHWIN2* From the +Division of Rheumatology andClinical Immunology. University of Calqornia at Davis. Davis, CA 95616; 'Department of Pathology, Emory University Atlanta, GA 30322; 'Departmentsof Pediatrics and Biochemistry,Emory and Eliza Hall Institutef o r Medical Research, Royal Melbourne Hospital, University, Atlanta.GA 30322; and 'Walter Victoria, 3050. Australia

Autoantibodies against mitochondria occur in the sera of patients with primary biliary cirrhosis (PBC) with characteristic reactivity to an inner membrane protein of approximately 74 kDa. To precisely define theseautoantigens, we recently cloned andsequenced a rat liver cDNA (pRMIT)that encodes for all of the epitopes recognized by Ig to the 74-kDa autoantigen. In the present study we have used this recombinant probe as a tool, in addition to purified enzymes, to demonstrate by immunoblotting that the 74-kDa mitochondrial autoantigen is dihydrolipoamide acetyltransferase (EC 2.3.1.12).the core protein of the pyruvate dehydrogenase complex. Furthermore, and of particular interest, inhibition of pyruvate dehydrogenase enzyme activity was demonstrated after incubation with sera from patients with PBC but not from normal volunteers or patients with chronic active hepatitis. Such inhibition was abrogated by absorption of the PBC sera with an expressing subclone ofpRMIT, designated pRMIT-603. Identification of dihydrolipoamide acetyltransferase a s the target of autoimmunity in PBC provides a reagent that can be used to determine mechanisms by which this molecule is recognized. It will allow study of whether autoimmune reactivity, at the humoral or T cell level, is the basis for the pathogenesis of PBC. Additionally, such data present evidence of functional inhibition of a critical metabolicenzyme.Dihydrolipoamide acetyltransferase is well-known to mitochondrial biochemistry and, similar to identified autoantigens in other autoimmune diseases, is highly conserved in evolution.

of diseases (1, 2). Such reactivity has been demonstrated using standard immunofluorescent assays, ELISA, and Complement fixation techniques (3, 4). Recently, however, therehas been considerable effort to determinethe biochemical nature of the mitochondrial molecules being recognized and these studies have clearly demonstrated that mitochondrial Ag can be distinguished on the basis of m.w. (5,6). For example, sera from patients withPBC3 react predominantly with a mitochondrial inner membrane protein of approximately 74 kDa (7). To identify the 74-kDa mitochondrial protein recognizedby PBC sera, we examined reactivity to a well characterized mitochondrial inner membraneprotein from the a-ketoacid dehydrogenase pathway, thePDH complex. PDH contains three catalytic components: E l , the decarboxylase (Ela 46kDa and E l @36kDa); E2, the lipoamide acetyltransferase (74kDa); and E3, the dihydrolipoamide dehydrogenase (55 kDa). The dihydrolipoamide acetyltransferase component (E2) contains a lipoic acid co-factor that is covalently bound to a lysine residue and is involved in the transfer of an acyl group to coenzyme A. Based on well-known enzyme m.w., we concentrated on PDH-E2 as a possible target of anti-mitochondrial antibodies. The justification for this hypothesis was based on the partially homologous sequence of our previously reported recombinant mitochondrial autoantigen with a clone of Escherichia coli PDH.E2 (8). The identification of the 74 kDa as lipoamide acetyltransferase by immunoblotting and sequence homology allowed us to definitively address the interaction between the autoantibodies of PBC and PDH.E2. We report herein that PBC autoantibodies specifically recognize PDH-E2 and more importantly, these antibodies inhibit enzyme functional activity.

Autoantibodies against mitochondria have been documented to occur in the seraof patients with wide a variety Received for publication January 27. 1988. Accepted for publication June 16. 1988. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ' This work was supported by funds received by United States Public Health Service Grant DK 39588 and from the Australian National Health of the and Medical Research Council. M.E.G. is aseniorinvestigator Arthrltis Foundation of America. Address correspondenceand reprint requests to Dr. M. Eric Gershwin, Division of Rheumatology/Allergy and Clinical Immunology. TB 192. University of California at Davis. Davis, CA 95616.

MATERIALS AND METHODS

Sera. Sera were collected from 30 patients having a well established clinical and laboratory diagnosis of PBC. In addition, sera from 15 patients with CAH and 20 healthy volunteers were used in this study. Sera were obtained fromthe clinics of the University of California a t Davis and the Walter and EHza Hall Instltute for Medical Research, Melbourne, Australia. Preparation of affinity purified antisera. Purified recombinant protein was covalently linked to cyanogen bromide Sepharose CL48 gel as previously described (9. IO). Six PBC sera were diluted in 3 ml of PBS/O.O5% Tween and incubated overnight with the coupled gel. The column was washed with100 bed vol of PBS/O.O5% Tween Abbreviations used in this paper: PBC. primary biliary cirrhosis: CAH. chronic active hepatitis:PDH. pyruvate dehydrogenase.

232 1

2322

AUTOANTIGENS OF BILIARY PRIMARY

a t a flow rate of 4 bed vol/h. Bound antibodies were elutedand then specificity confirmed as previously described (9). Isolation of mitochondrial recombinant polypeptlde. A clone of pRMIT containing thecDNA that encodes forthe 74-kDa autoantigen (9)was incubated overnight a t 37°C in Luria-Bertani broth containing 10 pdml ampicillin. A total of 18 h later it was diluted for log phase growth and was induced with 10 mM isopropyl-thiogalactosidase for 4 h. The E. coli preparation was then harvested and the recombinant fused polypeptide purified to homogeneity as described (9).The recombinant polypeptide encoded by pRMIT-603 was purified as above for useIn absorption studies (1 1). A s a further absorption control. the purified recombinant polypeptide produced by pRFA (the liver-specific F Ag was used throughout (12). Mitochondrlal enzyme complexes. A crude preparation of the PDH complex (13-15) was purchased from Sigma. Inc. (St. Louis. MO) and the dihydrolipoamide acetyltransferase further purified to homogeneity byHPLC using a n ion exchangeresin.Thepeaks resulting from HPLC fractionation were run on a n SDS-PAGE gel and probed with PBC sera at 1/1000 or with control sera at 1/100 (9).Due to the presenceof breakdown products or contaminants in commercial PDH. a preparation of branched-chain ketoacid dehydrogenase purified from bovine liver (16-18) that contained some PDH.E2 was also reacted with PBC sera at 1/1000. The enzyme activity and thepurification of this preparation haspreviously been described in detail (1 6). Immunoblottlng of enzymes. PAGE was performed on 1.5-mmthick slab gels in 0.1% SDS. using a 4.75% stackinggel and a 10% resolvinggel. Samples were diluted withTris-HC1,pH 6.8. containing 4% SDS. 20% glycerol. and 5%2-ME (sample buffer) andboiled for 5 min. Approximately 10 pg protein were loaded in each gel lane and proteins separated and blotted (9). The sera used for probing were from patients with PBC (1/1000 and 1/10.000). CAH (l/lOO). or normal volunteers (l/lOO). In addition. the strips were probed with sera from six patients with PBC sera that hadbeen affinity purified against the recombinant fused polypeptide expressed bypRMIT. After three IO-min washes withblotto, "'I-sheep anti-human Ig (0.07 Ci/ml) or '"I-goat anti-mouse ig (0.06 Ci/ml) was added and allowed to incubatefor 1 hour.All steps were performed a s described (9). Absorption of PBC sera. Serafrom six PBC patients were diluted 1/100. 1/1000. and 1/10.000 in PBS with 1% BSA. These diluted samples were absorbed with 100 pg PDH complex overnight at 4°C and used to probe blots of a n expresslng clone of pRMIT (9. 11). Sera from these same patientswere also absorbed with 100 pg of purified pRMIT induced fusion polypeptide and probed against PDH. A s a further control, all sera were absorbed against anotherrecombinant autoantigen recently cloned and sequenced from our laboratory, the rat F liver-specific Ag (pRFA)(12).Finally. sera absorbed against the PDH complex were used to probe the recombinant autoantigen in a n ELlSA assay to quantitate the absorptlon. Briefly. 2 $dm1 of Ag was used to coat microtiter plates overnight. After blocking. absorbed or unabsorbed sera were incubated in parallel for 1 h a t room temperature. After washing. the plates were incubated with goat antihuman peroxidase conjugate diluted 1/4000 (Tago, Burlingame. CA) for 30 min and reactivity visualized with 40.0 mM 2.2'azinobis(3ethyl benzthiazoline sulfonic acid) in 0.05 M citric acid buffer pH 4.2 containing 0.5 M H202. Sera from PBC patients and five CAH controls were diluted and absorbedwlth 50 pgml of the clone pRMIT-603 overnight a t 4°C. Llpoamlde acetyltransferase Lnhlbttion. To assay for inhibition of pyruvate dehydrogenase activity, sera from 10 patients with PBC and seven controls (five patients with CAH and two normal volunteers) were used. NADH production was quantitated spectrophotometrically over time at 340 nm. Thereaction mixture contained 50 pM potassium phosphate (pH 8.0): 2.5 pM NAD': 0.2 pM thiamin pyrophosphate: 0.13 pM coenzyme A: 0.32 pM dithiothreitol. 1 pM MgC1'. and 2 pM sodium pyruvate, pH 7.4 (19). PDH (Sigma. St. Louis. MO) with a n enzyme activity of 2.2 U/ml was diluted 1/10 in 1 % BSA-PBS and 10pl pre-incubated for 10 min a t room temperature with 90 pl 1% BSA-PES as a positive control. 90 pl of CAH serum diluted lo-' to in 1 % BSA-PBS a s a negative serum control. and 90 pl ofPBC sera diluted IO-' to lo-' in 1 % BSA-PBS. The 100 pl enzyme mixture was then added to 900 pl of the reaction mixture that had been equilibrated to 30°C. The increase in absorbance was measured over 120 s and the ratet = 20 s compared for each group. The assay was standardizedto a rate of t = 20 s of 0.065 to 0.075 with 1% BSA + PBS before incubation with the test sera. RESULTS

Immunoblotting. A total of 29 of 30 sera from patients with PBC reacted with theHPLC purified EC 2.2.3.12 74-

CIRRHOSIS

kDa core protein of the PDH complex, or dihydrolipoamide acetyltransferase by immunoblotting (TableI). In contrast, serafrom 15 patients with chronic active hepatitis and 20normal volunteers hadno detectable reactivity to the enzyme at a 1/100dilution or 10-fold the concentration used with sera from patients with PBC. Absorption. Further experiments wereperformed to confirm that dihydrolipoamide acetyltransferase is the major 74-kDa PBC autoantigen and that thisis encoded by the cDNA clone we had previously isolated. PBC sera and sera affinity purified against the recombinant fused polypeptidereacted with EC 2.3.1.12when probed against a preparation of branched-chain ketoacid dehydrogenasewhichcontained dihydrolipoamideacetyltransferase (Fig. 1). Absorption of 10 PBC sera with the purified recombinantautoantigen encoded by pRMIT eliminatedreactivityagainst HPLC purifieddihydrolipoamide acetyltransferase (Fig. 2):reactivity was unchanged whenPBC sera wasabsorbed against the control pRFA. In addition. absorption of PBC sera with the proteins of the PDH complex removed reactivity against the TABLE I Reactlultu of sera wlth dlhudrollpoamlde acetultransferase sera

Dliutlon

Dlhydrollpoamlde Acetyltransferase

PEC Affinity serab

1/1000

29/30"

CAH

]/loo

174 k h l

616

0115

Normals 1/100 0120 a p < 0.001 compared to CAH and normal sera. Student's t-test. bAffinitypurified antisera prepared as described (9).

A B C

D

74 -

52-

Flgure 1 . Purified branched-chain ketoaciddehydrogenase (RCKD)enzymes at 1 ppjlane are probed with sera from patients with PRC at I / 1.000 (lanes A and B). Note reactivity to the 74-kDa dihydrolipoamide is also noted. Further work acetyltransferase. Reactivity to 52 RCKD (E2). is in progress to confirm PRC sera antibodyactivity to these RCKD enzymes. In contrast. [lanr C). sera from a patient with CAH. even when probed at 1/100. had no reactivity. In lane D. sera from a patient with PRC that had been affinity purified against the recombinant fused polypeptide were usedto probe the BCKD complex. Note the reactivity to only dihydrolipoamideacetyltransferase.

2323

AUTOANTIGENS OF PRIMARY BILIARY CIRRHOSIS TABLE 111 Inhlbltlon ofPDHfunctlon by PBC autoantlbodles

A B

Patlent

?6

Inhlbltlon"

% lnhlblllon after

Absorption with DRM1T-603b

18.8 3.2 29.1 12.5 23.5 4.7

94.5' 62.5 92.8 29.2 100 54.0 90.2 78.2 48.5 54.7

3 6 7 24 71 78 88 90 91 96

pRMlT ELISA OD

0.769' 0.666 0.747 0.590 0.806 0.490 0.715 0.823 0.686 0.705

0

4.7 14.1 11.0

*

Flgure 2. In lane A HPLC purifled dlhydrollpoamlde acetyltransferase was run at1 @/lane and probed wlth sera from a patlent wlth PBC at 1/ 10,000. Note the 74-kDa reactlvlty. In lane B the same sera. absorbed wlth the cloned and purified recomblnant fused polypeptlde. Is nonreactlve. The mlnor bands In lane A are breakdown products; they are absorbed outas seen In lane B. TABLE II ELISA ofautoantlbodfes agalnst recornblnantfused polypeptlde encoded bu nRMITafter absorntlonwfthPDH"

" The rate t = 20 s for the controls (n = 7) was 0.065 0.005. or 0% lnhlbitlon; thlsIs ldentlcal to1% BSA-PBS controls. bThe % lnhlbltlon was unchanged after absorptlon wlth the control clone pRFA. c p = 0.019 when % Inhlbltlon Is compared to ELlSA value. Student's t-test.

UI 0

N

II

c

PBC Serum Dllullonb

Unabsorbed ODC

1/1000 1/10.000 1I1 00.000

0.791 f 0.262 0.441 f 0.200 0.169 f 0.058

OD after Absorptlon wlth Purlfled PDH rnmolrxd r""

*

0.174 0.026" 0.074 f 0.022' 0.022 f 0.020'

"The mean OD for control seras Is 0.252 (1/1000); 0.098 (1/10.000) and 0.008 (1/100.000).These background values are unchanged by absorptlon. bn=6. OD 492 nm. mean f SD.

The mean OD for PBC sera is unchanged when the controlrecomblnant F alloantlgen Is used ( 12). 'p < 0.01. Student's t-test, compared to unabsorbed sera.

recombinant autoantigen. Similar results were obtained in ELISA. The mean absorbanceof unabsorbed PBC sera at 1/1000 was 0.791 k 0.262 (mean& SD) when assayed against therecombinant autoantigen (Table 11). After absorption with the PDH complex, it decreased to background valuesof 0.174 & 0.026. Enzyme inhibitton. Pre-incubation of PDH with PBC sera inhibited enzyme activity in 10/10 patients tested. The degree of enzyme inhibitionby PBC sera ranged from 30 to 100% and wasdirectly proportional to the patient ELISA OD value, p = 0.019 (Table 111). The average enzyme rate (t = 20 s) for patients with CAH and normal volunteers were similar, and the data combined were 0.065 f 0.005. This value was identical to the enzyme activity when the reaction mixture was incubated without sera in thepresence of 1% BSA in PBS. In contrast, the enzyme rate was significantly reduced to 0.019 f 0.106 when incubated in the presence of sera from patients with PBC at a 1/100 dilution ( p s 0.01). Values for individual inhibition are shownin Table 111. The ratet = 20 s values were compared as percent of control values to allow for variation between assays. A dilution curve of one PBC assay patient with an average antibody titer demonstrated that the most effective dilution for complete inhibition ofPDH activity was1/100,although inhibitory activity was readily detected at dilutions up to 1/1000 (Fig. 3). Further, PBC sera absorbed with the recombinant polypeptide encoded by pRMIT-603,but not

w I-

4:

K

0.01

0.00

I...**" I 102

I

1b 3

1o 4

1o 5

1b 6

DILUTION

Flgure 3. Serial 10-fold dllutlons. of one representatlve PBC patlent serum and one CAH serum, were Incubated wlth PDH and the enzyme activity rate 5 = 20 s measured. The most effectlve serum dilution for complete Inhlbltlon of PDH actlvlty was seen at a dllutlon of 1/100. although Inhlbltlon was observed at 1/1.000. Little change was noted wlth dllutlon of control serum.

if absorbed with the control clone, lost the ability to inhibit PDH activity. Absorption of the control sera with pRMIT-603 or pRFA had no effect on the enzyme activity (Table 111). DISCUSSION

Previously we reported the identification of a cDNA encoding a portion of the 74 kDa autoantigen including epitopes recognized by Ig from patients with PBC. This mitochondrial autoantigen was hitherto believed to be part of or at least closely associated with the ATPase complex of the inner membranes of mitochondria (6). However, at thetime of publication of pRMIT, gene bank computer analysis failed to produce the identity of the cloned protein (9.20). A s described herein, this Ag is now identified as dihydrolipoamide acetyltransferase which is the core protein that binds together the PDH complex (8.21). PDH occupies a criticalposition in energy metabolism and is comprised of multiple copies of three types of enzymatic components, the decarboxylase (El),which contains two subunits m.w. 46.000 and 36.000. the dihydrolipoamide acetyltransferase(E2) with a m.w. of 74.000 and thelipoamide dehydrogenase (E3)with a m.w. of 55.000. PDH.E2 contains a lipoyl-lysine binding site. During catalysis, the substrate is carried in thioester

2324

AUTOANTIGENS OF PRIMARY BILIARY CIRRHOSIS

2. Berg, P. A.,R. Klein, and J. J. Lindenborn-Fotinos. 1986. Antimllinkage by the lipoyl-lysine swinging arms of E2 that tochondrialantibodies in primary biliary cirrhosis. J. Hepatol. transfer the substrate between the E 1and E3 active sites. 2: 123. The lipoyl-lysine residue protrudes between the El and 3. Kenna. J. G.. J. Neuberger. E. Davies, A. L. W. F. Eddleston, a n d R. Williams. 1984. A slmple enzyme-linked immunosorbent assay E3 subunits from a n inner portion of the E2core. Current for detection of anti-mitochondrial antibodies. J. Irnrnunol. Methods evidence indicates that mammalian PDH.E2 contains one 73:401. 4. Kaplan, M. M.. J. V. Gandolfo, and E. G . Quaroni. 1984.An enzymelipoyl-lysine residue, although a second group capableof linked immunosorbent assay (ELISA)for detecting antlmitochondrlal undergoing slow acetylation is also present on the E2 antibody. Hepatology 4:727. polypeptide (22).Escherichia coli PDH-E2 contains three 5. Lindenborn-Fotinos, J.. H. Baum. and P. A. Berg. 1985. Mitochondrial antibodies in primary biliary cirrhosis: species and nonspecies lipoic acid residues in three repeating highly conserved specific determinants of M2 antigen. Hepatology 5 ~ 7 6 3 . domains (8). 6. Baum, H., and C. Palmer. 1985. The PBC specific antigen. Mol. Aspects Med. 8:201. Recently our laboratoryidentified the subclone pRMIT7. Miyachi. K.,S. Watanabe. Yamashiki,T. Hiwatashi, andF. Ichida. 603 that contains the immunodominant epitope recog1984. Precipitating antimitochondrial antibodies in Japanese panized by PBC sera (1 1). The sequence IETDKATIGF, tients with primary biliary cirrhosis. Am. J. Gastroenterol. 79~704. 8. Packman, L. C., G . Hale, and R. N. Perham. 1984. Repeating funcwhich is contained in this epitope, is homologous to the tional domainsin thepyruvate dehydrogenasemultienzyme complex bovine heart PDH-E2 sequence (8)and also has signifiof Escherichia coli. EMBO J. 3:1315. cant homology to E . coli PDH.E2. Thus, this region is 9. Gershwin. M.E.,I. R. Mackay, A. Sturgess, andR. L. Coppel. 1987. Identification and specificity of a cDNA encoding the 74 kD mitohighly conserved in naturewhich is consistent with what chondrial antigen recognized in primary biliary cirrhosis. J. Irnrnuhas been reported for other autoantigens. Based on the nol. 138:3525. reactivity ofPBC sera to PDH.E2, it was reasoned that 10. Crewther. P. E., A. E. Bianco, G. V. Brown, R.L. Coppel, H.D. Stahl. D. J. Kemp, and R. F. Anders. 1986. Affinity purification of these autoantibodies could inhibit the function of PDH, human antibodies directed against cloned antigens of Plasmodium possibly by binding tothe lipoyl-lysine residue of PDH.E2. falciparum. J. Irnrnunol. Methods 86:257. 11 Van d e Water, J., M. E. Gershwin, P. b u n g . , A. Ansari, and R. L. In fact, peptide studies have shown that PBC sera recogCoppel. 1988. The autoepltope of the 74 kD mitochondrial autoannize the region surrounding this lipoic acid binding site tigen of primary biliary cirrhosis. J. Exp. Med. in press. (11). In our study, pre-incubation of PBC sera with PDH 12 Gershwin. M. E.,R. L. Coppel, E. Bearer, M. G . Peterson, A. Sturgess. and I. R. Mackay. 1987. Molecular cloning of the liver specific inhibitedenzymefunction in10/10patients tested. rat F antigen. J. irnrnunol. 139:3828. Moreover, absorption of these same sera with pRMIT-603 1 3 Stanley, C. J.. and R. N. Perham. 1980. Purification of 2-oxo acid dehydrogenase multienzyme complexes from ox heart by a new resulted in a loss of inhibitory activity. The data dismethod. Blochern. J. i91:147. cussed herein indicatethat not only do PBC autoantibod- 14. Matuda, S., T.Shirahama. T. Saheki. S. Miura. and M. Mori. 1983. ies bind to a critical enzyme involved in cell energy mePurification and immunochemical studies of pyruvate dehydrogenase complex from rat heart, and cell-free synthesis of lipoamide tabolism, but these antibodies also inhibit the function dehydrogenase, a component of the complex. Blochlrn. Bfophys. of this enzyme. This raises some interesting questions Acta 741:86. concerning the functionof PBC autoantibodies in disease 15. Hodgson, L. A..0. G . De Marcucci, and J. G . Lindsay. 1986. Lipoic acid is the site of substrate-dependent acetylation of component X pathogenesis and why such reactivity is directed against ox heart pyruvate dehydrogenasemultienzyme complex. J. Blochem. a cytoplasmic enzyme. 158:595. There areseveral possible mechanisms by which inner 16. Heffelfinger, C., E. T.Sewell, and D. J. Danner. 1983.Identification of specific subunits of highly purified bovine liver branched-chain mitochondrial proteins may become the target of an auketoacid dehydrogenase. Blochernfstry 22:5519. toimmuneresponse,includingabnormalitiesin the 17. Goodwin. G . W.. M. J., Kuntz, R. Paxton, and R. A. Harris. 1987. Enzymatic determination of the branched-chain a-keto acids. Anal. structure or location of these proteins in biliary epitheBlochern. 162536. lium. Structural changes in the portion of the gene en- 18. Danner. D. J., E. T. Sewell, a n d L.J. Elsas. 1982.Cloflbrlc acid and coding the leader sequencemay cause aberranttargeting phenylpyruvic acid as biochemical probes for studying soluble bovine liver branched chain ketoacid dehydrogenase. J. Blol. Chern. of the enzyme to the cell membrane rather than the 257:659. mitochondrial membrane with consequent immune pres19. Pettit. F. H.. and L. J. Reed. 1982. Pyruvate dehydrogenase complex from bovine kidney and heart. Methods Enzyrnol. 89:376. entation. A recent report has shown evidence of the cellM. E.,R. L. Coppel, and I. R. Mackay. 1987. Primary surface localization of a n Ag cross-reactive with AMA 20. Gershwin. biliary cirrhosis and mitochondrial autoantigens-insights from mo(23). The specific tissue damage seen requires that the lecular biology. Hepatology 8:147. re-targeting be confined to biliary epithelium and this 21. Kreze. G . B., and H. Ronft. 1981. Pyruvate dehydrogenase complex from Baker’s yeast. 2. Molecular structure, dissociation. and impllwould imply allelic forms of the gene, only one of which cations for the origin of mitochondria. Eur. J. Blochern. I19:581. has been mutated. There is evidence from studies on 22. Bradford, A. P.. S . Howell, A. Aitken. L. A. James, and S. J. Yeaman. 1987. Primary structure around the lipoate-attachment influenza infection that internal proteins may stimulate site on the E2 component of bovine heart pyruvate dehydrogenase a cytotoxic response via presentation of processed protein complex. Blochern. J. 245:919. in association with class I molecules (24-26). Altered or 23. Ghadiminejad 1.. and H. Baum. 1987. Evidence of the cell-surface localization of antigen cross-reacting with the mitochondrial antiabnormal enzymesmay be degraded at anincreased rate bodies of primary biliary cirrhosis. Hepatology 8 ~ 7 4 3 . leading to increased presentation at the surface. These 24. Townsend, A. R.. F. M. Gordon, and J. Davey. 1985. Cytotoxic T cells recognize fragments of the influenza nucleoprotein. Cell hypotheses may now be tested with the reagents we 42r457. possess. Alternatively, the antibodyresponse may be 25. Gotch, F., J. Rothbard. K. Howland, A. Townsend, a n d A. McMichael. 1987. Cytotoxic T lymphocytes recognize a fragment of secondary to pre-existing cellular damage and enzyme influenza virus matrix protein in associatlon with HLA-A2. Nature release. Interestingly, the lipoamide or lipoyl conjugates 6:881. of these enzymes are powerful adjuvants (27, 28) and 26. Townsend. A. R., J. Bastin, K. Gould, and G . G . Brownlee. 1986. Cytotoxic T lymphocytes recognize influenza hemagglutinin that may provide a basis for theincreased antibody response lacks a signal sequence. Nature 324:575. of PBC patients to these proteins. 27. Ohmori. H., T.Yamuchi, and I. Yamamoto. 1986. Augmentation of REFERENCES 1. Walker, J. G.. D. Doniach. I. Roitt, a n d S. Sherlock. 1965. Serological tests in diagnosis of primary biliary cirrhosis. Lancet 1:827.

the antibody response by lipoic acid in mice. I. Analysis of the mode of action in a n in vitro culture. J a p . J. Pharrnacol. 42135. 28. Ghadiminejad I., and H.Baum. 1986. Augmentation of the antibody responses by lipoic acid in mice. 11. Restoration of the antibody response in immunosuppressive mice. Jap. J . Pharrnacol. 4 2 2 7 5 .