distinguished two regions recognized by T cells from MG patients, that are restricted by different Class II (DR) antigens. 1. Newsom-Davis. J. & Vincent, A. ( 1982) ...
629th MEETING. LONDON
637
The role o f T helper cells is likely t o be crucial in the production of the autoantibody to AChRs in MG. Knowledge of the sequence of the human AChR subunits provides the basis for studies aimed at defining the T-cell epitopes in this disease, and their HLA Class I1 restriction. Preliminary studies using recombinant human a-subunit appear to have distinguished two regions recognized by T cells from MG patients, that are restricted by different Class I I (DR) antigens.
5 . Drachman. D. B.. Adams, K. N., Stanley. E. F. & Pestronk, A . ( 1 980) J. Neirrd. Neirrosiirg. /'.sychicit. 43. 6 0 1-6 10 6. Lennon, V. A,. McCormick. 11. J.. Lambert. E. H., Greisrnann. G. E. & Atassi. M. Z. ( 1985) /'roc,. Not/. Auld. . k i . U.S.A. 8 2 , 8805-8809 7. Compston. 11. A. S., Vincent. A,. Newsom-Davis. J. & Batchelor. J. R. ( 1980)Hruiri 103, 579-60 I 8. Newsom-Davis, J.. Harcourt, G., Sommer, N.. Beeson, D., Willcox, N., Vincent, A . & Kothbard. J. (1989) J . Aictoimmicnil;v in the press 9. Tzartos. S.J. & Lindstrom, J. M. (1980) /'roc. Nut/. Acad. Sci. 11.7ss-759 10. Hohlfeld, K..l'oyka, K. V.. Tzartos, S. J., Carson, W. & ContiTronconi. B. M. ( 1 9 8 7 ) /'roc. N u d Acad. Sci. U.S.A. 84,
1. Newsom-Davis. J. & Vincent, A. ( 1982) in C'lbiiccil A.spec~tsof Immirnolop (Lachmann, P. J. & Peters, D. K., eds), 4th edn., pp. 1 0 1 I - 1068. Blackwell Scientific Puhlications, Oxlord 2 , Patrick. J . & Lindstrom. J . M. ( 1 9 7 3 ) Scirricr 180, 8 7 1-87? 3. Vincent. A . & Newsom-Davis, J. ( 1980) J. Niwro/. Neiirosirrg.
5379-5383 1 1 . Beeson, D.. Brydson, M., Wood, H.. Vincent, A. & NewsomDavis. J. ( 1989) Hiocheni. Soc. Truris. 1I , 2 19-220 12. Harcourt, G. C., Sommer. N.. Rothbard. J., Willcox, H. N. A. & Newsom-Davis. J. i 1988) J . ( ' h i . I r i i ~ e t r .82, 1295- I300
Psychiut. 43, 5 7 7-5 89 4. Engel. A. G. (1980) J. Neicrol. Neiirosiirg Psvcliiui. 43, 571-589
Received 2 February I989
~'0rlc~llrsiorl.s
Naturally occurring autoantibodies to cholesterol in humans CARL R. ALVING, GLENN M. SWARTZ. J K and NABILA M. WASSEF Departmerit of Membrane Biochemistry, Walter Reed Ariny lristitirtr of'Kesearch, Washington. DC' 20.07-5/00, U.S.A. Rabbit antisera that caused flocculation and complement fixation with cholesterol were first described more than sixty years ago (Sachs & Klopstock. 1925). The antisera were obtained by injection of suspensions o f cholesterol with heterologous serum and serum proteins. and the ability to induce such antisera was subsequently confirmed by numerous laboratories (reviewed by Landsteincr, 1 Y45 ). Several laboratories also confirmed specificity for the phenomenon and showed that apparent specific immune reactions to cholesterol and related sterols, such as ergosterol and cholesterol derivatives, could be raised in rabbits lacking such activities. However, the colloidal nature of the cholesterol-containing antigens used for immunization and testing caused unique deviations from normally observed serological patterns with serum from immunized animals and this raised questions among some investigators about the existence of true anti-cholesterol antibodies (Wadsworth et ul., 1935; reviewed by Landsteiner, 1045). In recent years there has been renewed interest in the possibility of interactions of cholesterol with complement and the immune system. Cholesterol-dependent complement activation and complement-induced damage t o liposomes loaded with cholesterol was observed with normal human sera (Alving et ul., 1977). Complement activation occurred via the classical pathway, and the ability t o induce complement activation was transferred by an antibody-like factor from human serum to guinea-pig serum lacking activity. Numerous other laboratories have also documented activation of the classical and alternative pathways o f human and animal complement by cholesterol (Pang et NI., 1979; Hasselbacher & Hahn, 1980; Hammerschmidt et al., 198 1 ; Vogt et NI., 1985; Niculescu ct d.. 19x5; Seifert & Kazatchkine, 1987). Abbreviations used: LLISA, enzyme-linkcd immunosorbent assay; FBS. fetal bovine serum; PHS. phosphate-buffered saline.
Vol. I7
Despite the many reports of cholesterol-induced complement activation by normal human serum, true immunological reactions of cholesterol with antibodies in human serum have not yet been described. The theoretical possibility that such antibodies could exist has been confirmed with modern techniques by the demonstration that cholesterol is highly immunogenic in mice, and murine monoclonal antibodies that react with crystalline cholesterol are easily obtained (Swartz et al., 1988). In the present study, we have now found that most of the human sera that we have tested contain naturally occurring IgG and IgM autoantibodies that react with crystalline cholesterol. A few sera also contain I g A or IgE antibodies to cholesterol. The antibodies are readily detected by an enzyme-linked immunosorbent assay (ELISA)that uses crystalline cholesterol as an antigen. The detailed general method for measuring antibodies to cholesterol by ELISA was described by Swartz et al. (1088). Crystalline cholesterol that was used as a test antigen was recrystallized three times from hot ethanol. It had a melting temperature of 148°C and was > 9 0 % pure by thin-layer chromatography. It lacked detectable oxidation products. including peroxides, although a slight formation of peroxides did gradually occur with time. Ethanolic solutions of cholesterol (1 pg) were dried on the bottoms of microtitre wells, and the wells were blocked with 10"/0 fetal bovine serum in phosphate-buffered saline (FBS-PBS). Human test serum (50 pI of a l/lOO dilution in FBS-PBS) was added, followed by appropriate goat anti-( human immunoglobulin) (anti-Ig), anti-lgM, anti-lgG, anti-lgA o r anti-lgE serum conjugated to alkaline phosphatase (Kirkegaard and Perry Laboratories, Gaithersburg, MD, U S A . ). The fluorescent reaction (detected at 405 nm) was produced by adding substrate (p-nitrophenyl phosphate). The data are expressed as values obtained after subtraction of A,,,5 values from control assays run in the absence of cholesterol. The human sera that were tested for antibodies t o cholesterol were collected from several sources, and consisted either of preimmunization bleedings from 7 1 normal male volunteers (mean age mid-20s) participating in a research vaccine trial (group A ) (also used by Alving et al., 1977).normal male and female untreated controls participating in a double-blind
BIOCHEMICAL SOCIETY TRANSACTIONS
638 clinical drug trial (group B), or sera from males (aged 35-65) obtained immediately before elective cardiac catherization (group C). Most of the human sera tested contained detectable antibodies to cholesterol. As shown in Fig. 1, among 189 sera examined from groups €3 and C. the range of ELISA activity was quite broad, but only five sera gave a corrected A,,,, of less than 0.2, and none less than 0.1. When 7 1 normal sera (group A ) that were examined in a previous study (Alving el al., 1977) were re-examined by ELISA methods, nearly every serum manifested varied levels of IgG and IgM antibodies to cholesterol (Fig. 2). Separate experiments revealed that several sera also had apparent low levels o f IgA antibodies, and one serum may even have had a low level of IgE antibodies to cholesterol. As shown in Fig. 2, there was an excellent correlation between relative levels of IgG and IgM. If an arbitrary absorbance value of 0.1 is used as a criterion of positive results, only three sera apparently lacked both IgG and IgM activities. The distribution of absorbance values for IgC activities among the sera from group A is shown in Fig. 3. The mean of the absorbance
0.000-c 0
i
20
40
60
80
100
120
140 160
180 200
Serum sample (arranged by decreasing order) Fig. I . Antibodies to cholesterol in human sera Antibodies to cholesterol in I89 human sera (groups B and C) were analysed by ELISA using goat anti-(human Ig). The serum samples are arranged in descending order of ELISA activity. Background absorbance ( A controls were run in the absence of cholesterol and subtracted from the data shown. The control absorbance values that were subtracted were invariably < 0.07 (mean control A,,,, was 0.05).
.,,)
1.2
.. .
t
oJ* - 0
, 0.2
4
0.4
0.6
0.8
1.0
IgG ( A , , , , )
Fig. 2. Correlation of IgG and IgM ELISA vuluesfor each of the 71 individual human serum samples from group A
30T
1.
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8 0.9
1.0
IgG (A,,,,)
Fig. 3. Distribution atiu1y.si.s of’lgG aiitihodies t o cholesterol The IgG ELISA values are taken from the experiment shown in Fig. 2. values was 0.238 f 0.182 (sD 1. and three sera had absorbance values more than two standard deviations above the mean. When immune sera to cholesterol were originally describedin Germany in the 1920s and 1930s. they elicited a considerable degree o f interest (Landsteiner. 1945). Although the immunological techniques for working with lipid antigens (mainly flocculation and complement fixation) were well established at that time and were even relatively sophisticated. the biophysical complexities of the antigens caused complicated effects that prevented complete acceptance of the concept that true antibodies to cholesterol and other lipids could be produced. In recent years. research based on the development of liposomes. solid-phase immunoassays and hybridoma techniques has confirmed that bona fide antibodies t o numerous lipids can be produced. At least partly because of these new techniques there has been a resurgence of interest in lipids as antigens (Alving. I977 and 1986). One of the striking observations that has emerged from modern studies in lipid immunology is that naturally occurring antibodies (and autoantibodies) t o lipids, including glycolipids (Alving & Richards, 1Y 77; Richards Bi Alving, 1980), phosphospholipids (Alving, 1977. 1983. I984 and 1986) and lipid A from endotoxin (Mattsby-Baltzer Bi Alving, 1984) are widely distributed in normal human and animal sera. We have now found that IgG and IgM autoantibodies to cholesterol are present in nearly all normal adult human sera, and a few sera may even contain low levels of IgA or IgE autoantibodies to cholesterol. Although adverse consequences of autoantibodics t o cholesterol have not yet been observed directly in humans. we have previously proposed that complement activation by cholesterol might be a factor in the pathogenesis of atherosclerosis (Alving et al., 1977). As mentioned earlier, complement activation induced by cholesterol has been confirmed recently by many laboratories. The concept that cholesteroldependent complement activation might amplify ischaemic injury was supported by a report suggesting evidence of cholesterol-dependent complement activation by an immunoglobulin in the plasma of a patient with sevcre ulcerating atherosclerosis (Hammerschmidt el al., 1 98 1 ). Complement complexes (CSb-Y), as well as IgG. IgA, IgM, Clq, C3c and C4, have also been found localized in human atherosclerotic aortic wall (Niculescu et ul., 1985). We now suggest that the almost universal presence of autoantibodies to cholesterol, and cholesterol-dependent complement activation by naturally occurring autoantibodies, might play a role in the pathogenesis of vascular diseases, including atherosclerosis. The degree of such ii role, if any, will be investigated in the futurc by looking for positive or negative 1989
639
629th MEETING, LONDON correlations o f antibody titres and cholesterol levels, and correlations o f antibody titrcs with vascular diseases. A s yet the ubiquitous presence of naturally occurring antibodies t o cholesterol has prevented the drawing of immediate conclusions in these areas, and correlations will require statistical analyses o f populations of individuals at risk.
(revised edition), Harvard University Press, Cambridge (Reprinted in I962 by Dover Publications, New York) Mattsby-Baltzer, I. & Alving, C. R. (1984) Rev. Infect. Dis. 6 , 553-557 Niculescu, F., Rus, H.. Cristea. A. & Vlaicu. R. ( 1 985) Immunol. Leu. 10, 109- I 14 Pang, A. S. I).. Katz, A. & Minta. J. 0 . ( 1 979) J. Immunol. 123, 1 1 17-1 122 Richards, R. L. & Alving, C. R. ( 1980) in ('ell .SiirJic.r Glycolipids Alving,C. R . ( 1977)in TheAririgc,,r.s(Sela,M.,ed.), Vol. 4 . p ~ 1-73. . (Sweeley, C. C.. ed.), ACS Symposium 128. pp. 461-473, Academic Press, New York American Chemical Society. Washington, DC Alving, C. R. ( 1983) in Liposome lxrrrrs (Bangham. A. D., ed.). pp. Sachs,H.& Klopstock,A.( 1925) Hiochem. Z. 159,491-501 269-276. Academic Press. London Seifert. P. S. & Kazatchkine, M. I>. ( 1 987) Mol. Immiinol. 24, Alving, C. R.( 1 984j Riochem. So(. Trums. 12.342-344 1303- 1308 Alving. C. R. ( 1986) ('hem. I'h?.;. /,rpid\ 40, 303-3 I 4 Alving, C. R. & Richards, R. I>. (1977) / m m i ~ n ~ d ~ ~ ~ m14, i . s / r ySwartz, Jr, G. M., Gentry, M. K., Amende.- L. .M,. BlanchetteMackie, E. J. & Alving, C. R. ( 1988) /'roc. Nut/. Acrid. Sci. U.S.A. 383-389 85, 1902- 1906 Alving, C. R., Richards. R. L. & (iuirguis. A. A. ( 1977)J. Immioiol. Vogt, W., von Zabern, I.. Damerau. B.. Hesse. D.. Liihmann. B. & 118,342-347 Nolte. R. ( 1985) Mol. Immirnol. 22. 10 1 - I06 Hammerschmidt. D. E.. Greenberg. C. S . . Yamada. 0.. Craddock. Wadsworth, A.. Maltaner. E. & Maltaner, F. [ 1935)J. Immunol. 29, P. R. & Jacobs. H. S. ( 1 98 I ) J . I h . ('/in. Med. 98. 68-77 135-149 Hasselbacher, P. & Hahn, J. ( I 980) AtI~erosc7erosis37. 239-245 Landsteincr. K. ( 1945) T/w .Spcci/i'c~iry of Sero/ogic.til Kuic,tioris Received 3 I January I989
Lipoprotein receptors and atherosclerosis G E R T M . KOSTNER Itistitirte (fMeclicul Riochemistn. lJtiiver.sity of Grriz, A-8010 Gruz, Aiistriri Sirminun
T h e whole lipoprotein spectrum of human plasma may be divided into atherosclerotic and anti-atherosclerotic lipoproteins. To the first class belong apolipoprotein (apo)B and some apoE-containing lipoproteins of the very-low-density (VLDL). intermediate-density ( I D L ) and low-density ( L D L ) lipoprotein fractions. Anti-atherosclerotic lipoproteins are apoA-containing high-density lipoproteins (HDL). Circulating plasma lipoproteins are catabolized mainly by specific cell surface receptors ( R )which react with apoB and a p o E (B/E-R). for apoE ( E - R ) o r for apoA (HDL-R). Whereas the B/E-R and E-R are responsible for the cellular uptake of lipoproteins and their lipid load by various organs. HDL-R are thought to promote lipid (cholesterol) efflux. There is an additional class of lipoprotein rcceptors, the s o called scavenger-R which are responsible for the removal of altered o r degraded lipoproteins from the circulation. Under normal physiological conditions. the conccrtcd action of these receptors warrants efficient lipoprotein turnover and direction into target organs. Derangements o f this system, however, may lead to the deposition and accumulation of atherogenic lipids, notably free cholesterol ( F C ) and cholesteryl esters ( C E ) in arterial tissue causing atherosclerosis and cardiac death.
lritrodiictioti T h e main function of apolipoproteins is the transport of lipids in plasma in a soluble form. In addition, they activate lipolytic enzymes and target lipoproteins into various organs. There are two main sources of circulating lipoproteins: the intestine and the liver. Intestinal lipoproteins transport Abbreviations used: apoA. -B. -E. apolipoproteins A, B and E. respectively; VLIIL, very-low-density lipoprotein; IIIL, intermediate-density lipoprotein; LDL, low-density lipoprotein; HDL, high-density lipoprotein; FC, free cholesterol; CE, cholesteryl ester; TG. triacylglycerol; PL, phospholipid; -AC. acetylation; -Ma. maleylation; -Ox, oxidation; MPM. mouse peritoneal macrophages.
Vol. 17
mainly dietary lipids, notably triacylglycerols (TG),whereas the liver secretes lipoproteins containing endogenously produced lipids (rcviewed by Kostner, 1983).Although the rationale for the secretion of apoB-containing lipoproteins might be the export o f TG from organs, nascent lipoproteins bear a relatively high amount o f cholesterol in addition to phospholipids (PL). TG and PL are hydrolysed in plasma by specific lipases, leaving 'remnant' particles which are enriched in C E . T h e latter derives from FC by the action of lecithin: cholesterol acyltransferase (LCAT ). T h e CE-rich lipoproteins. also considered to be most atherogenic, are IDL. LDL as well as Lp(a) (an apoB-containing lipoprotein o f very high athcrogenicity). We speculated that Lp(a) mainly serves the purpose of secreting cholesterol from the liver independently from TG (Kostner, 1988). 7he R/E- u t d E-receptors. Table 1 lists the receptors specific for different apolipoproteins and also informs on their distribution in various organs. T h e best characterized receptor without doubt is the B/E-R (Brown &L Goldstein, 19741, which may be found on cells of any tissue. T h e liver, however, seems t o be the key organ which regulates plasma LDL and cholesterol concentration via this receptor. B/E-R are also highly abundant on cells of the adrenal gland. T h e B/E-R has a high affinity ( K c , 1 0 - x - l O - " mol/l) for apoBcontaining lipoproteins, but in addition an approx. 100-fold higher affinity for apoE-containing ones. T h c liver has an additional receptor which is specific for apoE-containing lipoproteins. but does not bind apoB. These two receptors promotc the catabolism of the majority of circulating atherogenic lipoproteins. We have studied the specificity o f the B/E-R and E-R on Hep-G2 cells in culture with respect to binding of various lipoprotein-density classes (Krempler et al., 1987).A representative Competition experiment is shown in Fig. 1. Similar experiments were carried out by incubating cells with Iz51labelled chylomicron remnants (data not shown). From this study. we concluded that chylomicron remnants and large VLDL remnants are primarily bound by the E-R, whereas small VLDL, IDL and LDL arc bound to the B/E-R. HDLcontaining apoE (HDL-E)are bound t o both receptors yet to the E-R with much higher affinity. In further experiments. we also studied the regulation of these two receptors: incubation of Hep-G2 cells in LDL-
