210

IMMUNOCHEMICAL METHODS

[2 1]

[21] I d e n t i f i c a t i o n o f H u m a n C y t o c h r o m e P 4 S 0 s as A u t o a n t i g e n s By MICHAEL P. MANNS and ERIC F. JOHNSON

Introduction Cytochrome P450 proteins are specifically recognized by human autoantibodies. Long before microsomal autoantigens were identified as members of the cytochrome P450 superfamily of proteins, microsomal autoantibodies had been discovered by indirect immunofluorescence. They became known as liver/kidney microsomal (LKM) autoantibodies owing to their reactivity with liver and kidney sections in immunofluorescence experiments.l LKM antibodies are associated with either idiopathic autoimmune or drug-induced hepatitis. For routine clinical diagnosis, microsomal autoantibodies are still detected by indirect immunofluorescence against cryostat sections from rodent liver and kidney tissue. 2 Furthermore, specific radioimmunoassays have been developed which measure the inhibitory effect of patient serum on the reaction of reference LKM antibodies with native microsomal proteins) ,4 If microsomal autoantibodies detected by immunofluorescence turn out to be of unknown antigen specificity, the patient serum is tested against a panel of purified cytochrome P450 proteins, against neonatal and adult microsomes from human liver, as well as against microsomes prepared from animals that had been treated with chemicals known to induce or suppress particular classes of cytochromes P450. In case the serum reacts with a particular purified cytochrome P450 or antigen expression is induced or suppressed by a particular agent, antigen specificity of a new microsomal autoantibody is confirmed by recognition of the expressed recombinant P450 protein. All autoantibodies against cytochrome P450 proteins that have been identified so far inhibit the function of their P450 protein in vitro. This can be used as an additional confirmatory test for the identity of the antigen. If these experiments do not clarify the antigen specificity of a new t M. Rizzetto, G. Swana, and D. Doniach, Clin. Exp. Immunol. 15, 331 (1973). 2 M. Manns, J. Hepatol. 9, 272 (1989). 3 M. Manns, K.-H. Meyer zum Biischenfelde, J. Slusarczyk, and H.-P. Dienes, Clin. Exp. lmmunol. 57, 600 (1984). 4 M. P. Mantas, K. J. Griffin, L. C. Quattrochi, M. Sacher, H. Thaler, R. H. Tukey, and E. F. Johnson, Arch. Biochem. Biophys. 280, 229 (1990).

METHODS IN ENZYMOLOGY,VOL. 206

Copyright © 1991by AcademicPress, Inc. All rights of reproductionin any form reserved.

[21]

P450 AUTOANTIGENS

211

human autoantibody, screening of a cDNA expression library becomes appropriate. Nucleic acid sequence analysis of isolated cDNA clones will clarify whether the cloned antigen is a cytochrome P450. Isolated cDNAs can then be subcloned in expression vectors suitable for the production of recombinant antigen to be used in diagnostic tests. The determination of microsomal autoantibodies against cytochrome P450 enzymes has clinical relevance for the diagnosis of inflammatory liver diseases, namely, the identification of drug-induced liver disease as well as subgroups of idiopathic autoimmune hepatitis. This has therapeutic implications for the patient. Testing for Hew Autoantibodies

If microsomal autoantibodies determined by indirect immunofluorescence and specific radioimmunoassays are of unknown antigen specificity, the following experiments are recommended for antigen identification. First, human autoantibody-positive serum is tested by immunoblotting against a panel of purified cytochrome P450 proteins. We use rabbit proteins which are often recognized by autoantibodies to human P450s of the same subfamily. Human autoantibodies in general react with conserved epitopes. Therefore, they cross-react with tissues from different species. This is also generally seen for antinuclear antibodies in rheumatological disorders, antimitochondrial antibodies in primary biliary cirrhosis, and others? We use 7.5% sodium dodecyl sulfate (SDS) gels and separate 2-10 pg of purified protein per lane. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting are performed by standard procedures. 6,7 Proteins are transferred and immobilized onto nitrocellulose filters, which are then incubated with human autoantibody-positive serum diluted 1:I00 in 2% milk powder (MP) in phosphate-bufferred saline (PBS)/Tween 20 (PBS is 15 mM sodium phosphate, 0.15 M NaCI). 7 For immunodetection we use as second antibody either peroxidase-labeled anti-human immunoglobulin G (IgG) serum and 4-chloro-l-naphthol or nSI-labeled protein A [300,000 counts/min (cpm)/ml].7 As controls, normal human serum or serum positive for other autoantibodies (i.e., antimitochondrial antibodies) are tested. If the autoantibody does not react with purified forms of rabbit cytochrome P450s the serum is tested against a 5 E. M. Tan, J. Clin. Invest. 84, 1 (1989). 6 H. Towbin, T. Staehlin, and J. Gordon, Proc. Natl. Acad. Sci. U.S.A. 76, 4350 (1979). 7 M. Manns, E. F. Johnson, K. J. Griffin, E. M. Tan, and K. F. Sullivan, J. Clin. Invest. 83, 1066 (1989).

212

IMMUNOCHEMICAL METHODS

[9.1]

panel of microsomes (2/~g/per lane) from neonatal or adult human livers and microsomes prepared from rabbits that have been treated with chemicals which induce or suppress the expression of cytochrome P450 proteins, among them TCDD (dioxin), phenobarbital, and rifampicin. If these experiments point to a particular cytochrome P450 as autoantigen, the autoantibody's specificity is confirmed by recognition of this cytochrome P450 expressed as a recombinant or by inhibition of specific drug metabolism known to be mediated by the particular cytochrome P450. The identification of cytochrome P450IA2 as a human autoantigen serves as an example. 4 First, serum from a patient with idiopathic autoimmune hepatitis was found to react with rodent liver tissue in indirect immunofluorescence and with a 50-kDa protein in human liver microsomes by immunoblotting. The serum failed to recognize human recombinant cytochrome P450IID6, the major antigen of LKMI autoantibodies. 8 In subsequent experiments it became evident that expression of the antigen was induced by TCDD in rabbits and that it reacts with purified rabbit cytochrome P450IA2. The autoantibody's specificity was confirmed by recognition of human cytochrome P450IA2 expressed from its cDNA in COS cells. Immunoblotting experiments indicated that the human serum specifically reacted with a protein in lysates from COS cells transfected with a cDNA for human cytochrome P450IA2 with a molecular weight comparable to that of the cellular microsomal protein. Lysates from COS cells transfected with a cDNA for human P450IA1 or with vector alone gave negative results. A 1:200 dilution of patient serum was tested in immunoblotting experiments against 5 /zg of human liver microsomes, 50/zg of microsomes from human derived Hep G2 cells treated with 3methylcholanthrene to induce P450IA1, and 50/zg of microsomes prepared from COS cells transfected with vector pSVL alone or pSVL carrying genes for either P450IA2 or P450IA1. LKMt autoantibodies directed against P450IID6 have been shown to inhibit bufuralol9 and sparteine 1° metabolism catalyzed by human liver microsomes, and LKM 2autoantibodies have been shown to inhibit ticrynafen (tienilic acid) 11and mephenytoin 12oxidation. The specificity of autoantibodies for cytochrome P450IA2 was further confirmed by their ability to a M. Sacher, P. Bltim¢l, H. Thaler, and M. Manns, J. Hepatol. 10, 364 (1990). 9 U. M. Zanger, H. P. Hauri, J. Loeper, J.-C. Homberg, and U. A. Meyer, Proc. Natl. Acad. Sci. U.S.A. 27, 8256 (1988). 10 M. Manns, U. Zanger, G. Gerken, K. F. Sullivan, K.-H. Meyer zum Btischenfelde, U. A. Meyer, and M. Eichelbaum, Hepatology 12, 127 (1990). i1 p. H. Beaune, P. M. Dansette, D. Mansuy, L. Kiffel, M. Finck, C. Amar, J. P. Leroux, and J. C : Homberg, Proc. Natl. Acad. Sci. U.S.A. 84, 551 (1981). t2 U. T. Meier and U. A. Meyer, Biochemistry 26, 8466 (1987).

[21]

P450 AUTOANTIGENS

213

inhibit ethoxyresorufin O-deethylation in isolated human liver microsomes. This reaction is largely catalyzed by P450IA1 and P450IA2 in rat 13,14and rabbit. 15 In addition, a monoclonal antibody developed to rat P450IA1, which also recognizes P450IA2, extensively inhibits ethoxyresorufin O-deethylation catalyzed by human liver microsomes.16 For inhibition studies 0.1, 0.5, 3.0, and 10/zl of patient or control serum were added to human liver microsomes (0.5 mg protein). Product formation was monitored fluorometrically as described previously.15 Prior to initiation of the reaction by the addition of NADPH and isocitrate/ isocitrate dehydrogenase as an NADPH regeneration system, the indicated volume of serum was incubated with the microsomes for 15 min. Whereas the autoantibody serum almost completely inhibited P450IA2mediated ethoxyresorufin O-deethylation (>90%), neither normal human serum nor autoantibodies against other cytochrome P450s such as P450IID6 (LKM~) inhibited the function of this enzyme. Autoantibodies against P450IA2 have also been reported by others in patients with dihydralazine hepatitis. ~7 Sera from the latter were able to inhibit P450IA2supported monooxygenase activity by more than 90%, namely, ethoxyresorufin and phenacetin O-deethylation. No inhibition was seen for pentoxyresorufin O-dealkylation, which is not supported by P450IA2.17 It is not known whether autoantibodies to P450IA2 from dihydralazine hepatitis patients recognize the same epitope as the autoantibody identified in our study.

Screening cDNA Expression Libraries with Human Autoantibodies Screening of a eDNA expression library with human autoantibodies becomes the appropriate experimental approach when other experiments, namely, immunoblotting against purified cytochrome P450 proteins, regulation of autoantigen expression in microsomes, recognition of recombinant cytochrome P450s, and enzyme inhibition studies, do not clarify antigen identification.

13M. D. Burke and R. T. Mayer, Drug Metab. Dispos. 3, 245 (1975). 14 F. P. Guengerich, G. A. Dannan, S. T. Wright, M. V. Margin, and L. S. Kaminsky, Biochemistry 21, 6019 (1982). 15E. F. Johnson, G. E. Schwab, and U. Muller-Eberhard, Mol. Pharmacol. 15, 708 (1979). t60. Pelkonen, M. Pasanen, H. Kuha, B. Gachalyi, M. Kairaluoma, E. A. Sotaniemi, S. S. Park, F. K. Friedman, and H. V. Gelboin, Br. J. Clin. Pharmacol. 22, 125 (1986). 17M. Bourdi, D. Larrey, J. Nataf, J. Bernuan, D. Pessayre, M. Iwasaki, F. P. Guengerich, and P. H. Beanne, J. Clin. Invest. 85, 1967 (1990).

214

[21]

IMMUNOCHEMICAL METHODS

Western Blot A

PS kDa 64-

50-

HLD HLD GTE 8.2 13.2 411

PS

[21 ]

P450 AUTOANTIGENS

215

Choice of Serum and Library From our experience, it is extremely important to thoroughly select an individual serum for screening the library. This serum should react at a high titer, that is, more than 1 : 1000 in immunoblotting, with the cellular protein. Therefore, all available sera should be tested against microsomes. Premade h-gtl 1 cDNA expression libraries can be purchased from many companies (e.g., Clontech, Palo Alto, CA; Stratagene, La Jolla, CA; Bethesda Research Laboratories, Gaithersburg, MD). The ideal approach is to prepare a cDNA library from an organ of which microsomes are available for testing antigen content by immunoblotting. Since some P450s are genetically polymorphic, it may be important to test the sera against microsomes prepared from the same liver which served for cDNA library synthesis.

Screening cDNA Expression Libraries with Human Autoantibodies cDNA clones are identified, purified, and isolated by immunodetection procedures. These methods are well standardized. For further details, see Young and Davis. TM

Immunological Identification of cDNA-Encoded Antigen After isolation of the cDNA, autoantibodies are affinity purified on fusion proteins expressed by the isolated cDNA clone. For controls, the purification procedures employ unrelated cDNA clones. Isolated cDNA clones in h phages are plated on a 150-ram plate at a concentration of 100,000 plaque-forming units (pfu). Proteins are adsorbed to nitrocellulose filters saturated with isopropyl-/3-D-thiogalactoside (IPTG) as described. 7,18

18 R. A. Young and R. W. Davis, Proc. Natl. Acad. Sci. U.S.A. 80, 1194 (1983).

FIG. 1. The 50-kDa antigen is encoded by cDNA clones HLD8.2 and HLD 13.2. Antibodies were affinity purified from nitrocellulose filters beating cDNA-encoded antigens as described. (A) Imrnunoblot of human liver microsomes (12 /zg protein) probed with whole LKM 1 antibody-positive serum (PS) at a 1:800 dilution or with affinity-purified autoantibodies isolated from hgtl 1 clones HLD8.2, HLDI3.2, or a control cDNA, GTE411. (B) Cytoplasmic immunofluorescence staining of mouse proximal renal tubular cells with autoantibodies affinity purified from antigen encoded by hgtl 1 clone HLD8.2. Magnification: x50. [Reproduced from the Journal of Clinical lnoestigation 83, 1066 (1989), 7 by copyright permission of the American Society of Clinical Investigation.]

216

IMMUNOCHEMICAL METHODS

[21]

Filters are first incubated with 2% MP in PBS/Tween and then with autoantibody-positive serum diluted 1 : 200 in 2% MP in PBS/Tween for 60 min at room temperature. After extensive washings with PBS/Tween for 60 min, nitrocellulose filters are cut into small pieces (1-2 mm diameter) and transferred to 50-ml Falcon tubes. Gloves should be used while handling the filters. Filter pieces are then washed in the Falcon tubes quickly by adding 20 ml of PBS/Tween and brief vortexing. Antibodies bound to recombinant antigen on filters are then subsequently eluted by adding 5 ml of 0.2 M glycine, pH 2.7. The filters are vortexed in the glycine buffer for 5 min. The filter pieces are then pelleted, and the supernatant is removed by Pasteur pipette into 10-ml tubes containing 75/zl of 1 M Tris buffer, pH 9.5. Additional Tris buffer may be added to reach pH 7.5. The filter bits are washed twice with PBS/Tween and may be stored at - 7 0 ° for another affinity purification procedure. After neutralization and before further analysis by immunoblotting and immunofluorescence, the eluted antibody is concentrated 25-fold, using a Minicon Macrosolute Concentrator (Amicon, Danvers, MA). Twentyfive microliters of the final eluted antibody solution is used for standard immunofluorescence testing on rodent liver and kidney cryostat sections, whereas 50 ~1 is used for immunoblotting with human liver microsomes. As demonstrated in Fig. 1, antibodies affinity purified on isolated eDNA clones HLD8.2 and HLD 13.2 react with the 50-kDa protein in human liver microsomes but not with the 64-kDa antigen recognized by the whole LKM~antibody-positive serum. In immunofluorescence, the eluted antibodies specifically recognize the cytoplasm of hepatocytes and proximal renal tubular epithelia. Antibodies purified on the unrelated GTE411 eDNA clone are unreactive. The cDNA clones are further characterized by sequence analysis if the autoantibodies purified on hgt 11-derived fusion proteins react with microsomal protein of the appropriate molecular weight as evidenced by immunoblotting and with appropriate tissue ceils in immunofluorescence studies.

Identification of Autoantigen by Sequence Comparison Sequence comparison studies will reveal whether the cloned autoantigen is a particular cytochrome P450, whether its sequence has homology with previously known cytochrome P450s, or whether the cloned antigen seems to be a new gene. Use of Recombinant Antigen as Diagnostic Reagent For immunoblotting analysis using recombinant antigen, the cDNA can be subcloned into specific expression vectors. In the case of cytochrome

[2 1]

P450 AUTOANTIGENS

217

TABLE I DIAGNOSTICSIGNIFICANCE OF RECOMBINANT LKM1 (P450IID6) ANTIGEN(LKM-R)a

CAH ANA positive ("lupoid") CAH LKM positive CAH SLA positive Acute hepatitis (A, B, C, NANB) PBC Extrahepatic cholestasis Malignancies Immunopathies Controls

N

LKM-IF (positive)

Western blot: 50 kDa (positive)

LKM-R (P450IID6) (positive)

20

0

0

0

46

46

28

40

10

0

0

0

51

0

0

0

20 10

0 0

0 0

0 0

20 20 20

0 3 0

0 0 0

0 0 0

IF, Immunofluorescence; CAH, chronic active hepatitis; ANA, antinuclear antibodies; LKM, liver kidney microsomal antibodies; SLA, soluble liver antigen antibodies; PBC, primary biliary cirrhosis; N, number of sera tested from different individual patients; LKM-IF, LKM antibodies detected by indirect immunofluroescence; Western blot; 50 kDa, LKM antibodies detected by reactivity with 50 kDa microsomal protein in Western blot; LKM-R (P450IID6), LKM1 antibodies detected by reactivity with human recombinant cytochrome P450IID6.

P450IID6, we used the fusion protein v e c t o r p A T H l l (gift of Dr. T. J. K o e r n e r , D u k e University, D u r h a m , NC). p A T H 11 contains the 5' portion of the Escherichia coli trpE gene encoding 36 k D a of anthranilate synthetase followed b y a polylinker that p o s s e s s e s an EcoRI site in the same translational f r a m e as the EcoRI site of hgt 11. After ligation and transformation into Escherichia coli R R I , colonies are selected, cultured, and p r o c e s s e d for i m m u n o b l o t analysis. 7 Owing to the low a b u n d a n c e of P450s in liver tissue, only a portion of sera showing antimicrosomal reactivity in immunofluorescence studies react with the 50-kDa protein of h u m a n liver m i c r o s o m e s in immunoblots (Table I). A diagnostic test was established using a h u m a n r e c o m b i n a n t c y t o c h r o m e P450IID6 representing 85% of the protein coding sequence o f the h u m a n P450IID6 gene 7 (Fig. 2). F o r the preparation of the r e c o m b i n a n t c y t o c h r o m e P450IID6, E. coli RR1 was t r a n s f o r m e d with the p A T H l l / H L D 8 . 2 c D N A construct. T h e synthesis of the fusion protein is induced

218

IMMUNOCHEMICALMETHODS

Serum A 123

45

[21]

Serum B 1

23

4

kDa

97.466.2-

"-F.P.

42.7-

31.0-

FIo. 2. LKM1 autoantibodies react with the 50-kDa protein of rabbit (lane 1) and human (lane 2) liver mierosomes as well as the 78-kDa recombinant antigen (F.P.) derived from the pATH11 L K M - c D N A construct (lane 3). No reaction is seen with bacterial lysates from E. coil RR1 transformed with pATH11 carrying the eDNA in the incorrect orientation (lane 4) or with vector pATHI 1 alone (lane 5). The results with two individual sera are shown (A and B). [Reproduced from the Journal of Clinical Investigation 83, 1066 (1989), 7 by copyright permission of the American Society of Clinical Investigation.]

by indoleacrylic acid (IAA) as described elsewhere. 19 Bacteria from 1.5ml cultures are pelleted for 2 rain in a Eppendorf microfuge and washed with STE medium (10 mM Tris, pH 8, 1 mM EDTA, 100 mM NaC1) before being resuspended in 3% of the original volume in STE containing lysozyme (5 mg/ml). After incubation on ice for 5 min, bacteria are frozen for 15 min at - 7 0 °, then thawed, and 1/10 volume of Triton X-100 added. After vortexing, 1/10 volume of DNase (1 mg/ml), diluted in STE with 10 mM MgC12, is added at room temperature until the lysed bacteria are no 19 W. C. Earnshaw, K. F. Sullivan, P. S. Machlin, C. A. Cooke, D. A. Kaiser, T. D. Pollard, N. F. Rothfield, and D. W. Cleveland, J. Cell Biol. 104, 817 (1987).

[2 1]

P450 AUTOANTIGENS

219

longer viscous. This takes approximately 10 min. Finally, an equal volume of 2× LSB buffer (125 mM Tris-Cl, pH 6.8, 6% SDS, 10% mercaptoethanol, 20% glycerol, and 1% bromphenol blue) is added before the solution is boiled for 2 min and pelleted. Twenty microliters of the bacterial lysate is loaded per lane for SDS-PAGE and immunoblot analysis. For control, lysates from bacteria transfected with vector alone or carrying the cDNA in the wrong orientation were tested. Such an experiment with two different patient sera is shown in Fig. 2. LKM1 autoantibodies react with human recombinant P450IID6 fusion protein at 78 kDa. Human recombinant cytochrome P450IID6 (LKM1 antigen) proved to be a specific and sensitive diagnostic reagent2° (Table I). Only LKM~positive sera, as evidenced by immunofluorescence, from patients with inflammatory liver diseases react with the P450IID6-TrpE fusion protein. Some sera giving LKM type immunofluroescence on tissue sections from patients with nonhepatic autoimmune disorders did not react with the human recombinant P450IID6. Furthermore, the test with cloned antigen proved to be more sensitive compared to immunoblotting with human liver microsomes (Table I). However, 10% of LKMt-positive sera from patients with liver diseases were found to react with a 50-kDa microsomal protein but not with recombinant P450IID6 (LKM~ antigen). One of these rare liver microsomal autoantigens has been identified as cytochrome P450IA2. 4,8 This illustrates how the application of recombinant cDNA technology can distinguish autoantigens in microsomes with similar molecular weights which are not separated by SDS-PAGE. Summary Antimicrosomal antibodies in inflammatory liver diseases all seem to be directed against members of the cytochrome P450 family of proteins. These autoantigens seem to be genetically polymorphic, the autoantibodies are inhibitory, and the autoepitopes are generally conserved among species. Anti-P450 autoantibodies share these characteristics with other autoantibodies, for example, antinuclear antibodies in sytemic lupus crythematosus.5 The identification of P450s as human autoantigens is clinically important. Diagnostic tests will be developed on the basis of cloned antigen, facilitating a better diagnosis of drug-induced and idiopathic autoimmune hepatitis. 2,4,7,H,17,2° It is unknown what triggers autoantibody production against cytochrome P450 proteins. Furthermore, their pathogenetic role 2o M. Manns, E. F. Johnson, K. J. Griffin, K.-H. Meyer zum Btischenfelde, E. M. Tan, and K. F. Sullivan, Hepatology 10, 637 (1989).

220

IMMUNOCHEMICALMETHODS

[22]

and thus their involvement in tissue destruction is unclear. In this context LKMI autoantibodies may serve as a model. Although LKM~ antibodies are inhibitory, all LKM~ antibody-positive patients tested so far are extensive metabolizers for drug metabolism mediated by P450IID6 and express this protein in their livers.l° Thus, the inhibitory LKM1 autoantibody does not sufficiently penetrate through the intact liver cell membrane to inhibit enzyme function in vivo. Presumably, tissue destruction in autoimmune hepatitis is mediated by liver-infiltrating T lymphocytes. T lymphocytes have been cloned from liver tissue that specifically proliferate in the presence of recombinant cytochrome P450IID6. 21The construction of overlapping cDNA subclones is also valuable to identify immunodominant B cell as well as relevant T cell epitopes. Acknowledgments Our own work was supported by NATO Collaboration Research Grant No. 5-2-05/RG891003 (M.M.), National Institutes of Health Grant GM 31001 (E.F.L), and the Deutsche Forschungsgemeinschaft, SFB 311, Project Al (M.M.). The authors acknowledge the support of E. M. Tan, K. J. Griffin, K, F. Sullivan, La Jolla, and K. H. Meyer zum Btischenfelde, Mainz. 21 H. L6hr, M. Manns, C. Trautwein, K.-H. Meyer zum Biischenfeldc, and B. Fleischer, Clin. Exp. Immunol. 84, 297 (1991).

1-22] A n t i p e p t i d e Antibodies in Studies of C y t o c h r o m e s P 4 5 0 I A By ROBERT J. EDWARDS, BEa_r~ARDP. MURRAY, and ALAN R. BOOBIS Introduction Antipeptide antibodies directed to specific surface regions of cytochromes P450 provide a powerful means of studying both the structure and function of the members of this superfamily of proteins. Such antibodies are suitable for many immunological techniques, including enzymelinked immunosorbent assay (ELISA), immunoblotting, immunocytochemistry, and immunoinhibition of enzyme activity. The generation of antipeptide antibodies is particularly suited to the analysis of cytochromes P450 because the primary structures of many of the members this large group of isoenzymes have been deduced by METHODS IN ENZYMOLOGY,VOL. 206

Copyright © 1991by AcademicPress, Inc. All rightsof reproductionin any form reserved.

Identification of human cytochrome P450s as autoantigens.

Antimicrosomal antibodies in inflammatory liver diseases all seem to be directed against members of the cytochrome P450 family of proteins. These auto...
3MB Sizes 0 Downloads 0 Views