1166

ARE ANTIMITOCHONDRIAL ANTIBODIESIN PRIMARY BILIARY CIRRHOSISINDUCED BY R(ROUGH)-MUTANTS

OF ENTEROBACTERIACEAE?

ROMAN STEMEROWICZ1BERND MÖLLER1ARNE RODLOFF2

MARINA FREUDENBERG4

UWE HOPF1CHRISTEL WITTENBRINK1RICHARD REINHARDT3

CHRIS GALANOS4

Medizinische Klinik, Universitätsklinikum Charlottenburg1 andInstitut für Medizinische Mikrobiologie und Infektionsimmunologie,2

Freie Universität Berlin; Max-Planck-Institut für MolekulareGenetik, Abt Wittmann,3 Berlin; and Max-Planck-Institut für

Immunbiologie, Freiburg,4 Federal Republic of Germany

Summary Antimitochondrial antibody (AMA)-positive serum samples from 45 patients

with primary biliary cirrhosis (PBC) and AMA-negativeserum samples from patients with chronic liver diseases,systemic lupus erythematosus, or rheumatoid arthritis werestudied by an immunoblot technique with mitochondriafrom bovine heart and pig kidney and with several strains ofgram-negative bacteria as antigens after separation bysodiumdodecylsulphate polyacrylamide gel electrophoresis.Serum from patients with PBC recognised up to threemitochondrial antigens with molecular weights of 70 kD, 50kD, and 42 kD. Equivalent patterns were found with bandsat 70-80 kD and 50-52 kD with Enterobacteriaceae as

antigens. AMA-reactive polypeptides were localised in theribosomal and membrane fractions from Enterobacteriaceaebut differed from known outer membrane proteins.Conversely, PBC-specific mitochondrial antigens at 70 kDand 50 kD were recognised by rabbit antisera againstSalmonella minnesota Rb and Rc mutants but not by antiseraagainst wild-type Enterobacteriaceae. Absorptionexperiments and two-dimensional immunoblotting studiesconfirmed that mitochondria and gram-negative bacteriashare identical PBC-specific determinants. It seems that

PBC-specific antigens are expressed in gram-negativebacteria and that these antigens may be immunogenic inmutants with defective polysaccharide synthesis. The datasupport the hypothesis of a bacterial aetiology for PBC.

Introduction

THE aetiology and pathogenesis of primary biliarycirrhosis (PBC) are unknown. Some properties of PBC,such as the presence of antimitochondrial antibodies

(AMA),l.2 antibodies directed against other subcellular

structures,3 circulating immune complexes,4 high levels of

IgM with monomeric IgM,s familial occurrence of thedisease6 and association with other immune disorders,’suggest that immunological processes are involved in itspathogenesis. AMA were found in up to 100% of patientswith PBC by means of sensitive techniques.2,g In the westernblotting technique, AMA recognise up to five distinctmitochondrial polypeptides of varying molecular weightdepending on the species used as a source of antigen.9-16Polypeptides reacting with AMA have also beendemonstrated in mitochondrial preparations from yeastsand higher plants.12,17Two studies have shown reactivity of serum samples from

PBC patients with some polypeptides from EscherichiaColi.10,12 Extensive absorption procedures indicated thatAMA might cause this reaction.10 In our study samples ofserum from patients with PBC and from patients with otherdiseases were tested by immunoblotting with various

gram-negative bacteria, including representativeEnterobacteriaceae with smooth(S) or rough(R) forms of thelipopolysaccharide molecule, and with subcellular bacterialstructures. In addition, we studied the reactions withmitochondrial polypeptides of antisera raised in rabbitsagainst various bacterial strains.

Patients and Methods

We studied 45 patients with PBC of histological stages 1-IV,18 40patients with chronic inflammatory liver diseases but no AMA, 10patients with systemic lupus erythematosus, 10 patients withrheumatoid arthritis, and 30 healthy controls (table i). All patientswith liver disease had been followed for more than a year, and atleast one liver biopsy had been carried out for each.

In routine clinical studies AMA, antinuclear antibodies, andsmooth muscle antibodies were detected by an immuno-fluorescence technique on cryostat sections from mouse liver andkidney. Liver membrane antibodies were determined byimmunofluorescence with isolated rabbit hepatocytes19 and

immunoblotting.2oMitochondria from bovine heart and pig kidney were prepared as

described by Martinez and McCauley21 and stored in small volumesat - 80oC.The bacterial strains used as antigens in the study are

summarised in table 11. E coli wild strains and E coli R(rough)mutants were isolated from patients; the Salmonella minnesotamutants were obtained from the Max-Planck-Institut, Freiburg;Neisseria meningitidis, Halobacteriurn marismortui, and SulfolobaGsacidocaldarius were donated by the Max-Planck - Institut, Berlin. Allother bacterial strains used were reference strains from appropriateinstitutions (American Type Culture Collection, USA; DeutscheSammlung fur Mikroorganismen, Gottingen, FRG).

Published methods were used for the isolation of E coli wholemembranes22 and outer membranes23 and the preparation ofbacterial ribosomes and ribosomal subunits from E coli andArchaebacteria.24

TABLE I-CLINICAL AND SEROLOGICAL DATA OF PATIENTS

Sex

CAH = chronic active hepatitis; SLE = systemic lupus erythematosus; ANA = antinuclear antibodies; LMA= liver membrane antibodies; SMA = smoothmuscle antibodies.

1167

TABLE II-POLYPEPTIDE ANTIGENS OF VARIOUS

GRAM-NEGATIVE BACTERIA REACTING IN IMMUNOBLOT STUDIES

WITH AMA-POSITIVE SAMPLES FROM PATIENTS WITH PBC

I

*Archaebacteria.LPS = lipopolysaccharide; S = smooth form: R = rough form.

Antisera against bacteria were obtained after subcutaneousimmunisation of rabbits with heat-killed bacteria in incompleteFreund’s adjuvant.2s .

Sodium dodecylsulphate/polyacrylamide gel electrophoresis wascarried out26 with 10-15% gradient gels. The gels were run in twosteps (at 200 V and a constant current of 80 mA for 1 h, then at 500 Vand 120 mA) in a "tris"-glycine buffer (pH 83). Molecular weightstandards were included in each run. Gels were either blotted orstained with Coomassie brilliant blue G 250. After electrophoresis,proteins were transferred from the gels to nitrocellulose sheets.2’ Aside portion of the blot was stained with Coomassie blue, and theother portion was soaked overnight at 45°C in phosphate-bufferedsaline containing 2% (weight/volume) bovine serum albumin and2% gelatin. Subsequently, the sheets were incubated with serumsamples diluted in phosphate-buffered saline (patient samples1/100; rabbit antisera 1/250) containing 2% (by volume) fetal calfserum. Immunodetection was carried out with horseradish-

peroxidase-conjugated mouse anti-human or goat anti-rabbit

F(ab), (Dianova, Hamburg, FRG) and 0-diaminobenzidine.Two-dimensional gel electrophoresis was carried out as

described by O’Farrel.2g After isoelectric focusing in the pH range35-70, the mitochondrial polypeptides were separated in gradientgels according to their molecular weight, stained or transferred tonitrocellulose as described above, and tested either with

representative patient serum samples or with rabbit antisera.1 ml samples of AMA-positive serum from patients and rabbit

antisera were incubated with 50 mg lyophilised bacteria or bovinemitochondrial proteins for 30 min at 37°C and stored at 4°C

overnight. Absorbed samples were centrifuged at 105 g for 30 min.The procedure was repeated three times. Absorbed and non-absorbed samples at the same dilutions were screened byimmunoblotting with either bacterial or mitochondrial proteinfractions as antigens.

Results

Immunoblot Analysis with Mitochondrial Antigens

Samples from 45 patients with PBC reacted with up tothree mitochondrial polypeptides (molecular weights 70 kD,50 kD, and 42 kD) in mitochondrial preparations from

TABLE III-REACTIVITY OF PBC PATIENTS’ SERUM SAMPLES

WITH POLYPEPTIDES FROM BOVINE HEART MITOCHONDRIA BY

IMMUNOBLOTTING

bovine heart and pig kidney (fig 1). There were fourdifferent patterns of AMA reaction in patients with PBC(table III). The most common pattern involved staining ofboth 70 kD and 50 kD proteins (23 of 45 patients). Allsamples tested from PBC patients recognised at least one ofthese two polypeptides. None of the samples from patientswith other liver diseases, systemic lupus erythematosus, orrheumatoid arthritis, or healthy controls was reactive withPBC-specific mitochondrial polypeptides.

Immunoblot Studies with Gram-negative Bacteria andBacterial Protein Sub fractions

16 representative serum samples from patients with PBCwere tested with several gram-negative bacterial strains asantigens (table II). The PBC-specific reaction patterns wereobserved with all the bacteria used except for two

Archaebacteria strains. All PBC samples tested reacted withEnterobacteriaceae and showed identical patterns, withmain polypeptide bands at molecular weight 70-80 kD and50-52 kD. Only one protein was stained (52 kD) withEnterobacter cloacae as antigen. Bacteroides fragilis andAcinetobacter calcoaceticus, two gram-negative specieswhich are not Enterobacteriaceae, showed PBC-specificbands at 60 kD and 95-100 kD, or an additional band at40 kD. Samples from patients with other liver diseases,systemic lupus erythematosus, or rheumatoid arthritis, andhealthy controls showed no reaction in the molecular weightrange 40-100 kD. 5 of the 10 samples from patients withalcoholic cirrhosis showed a band at 5 kD, the region specificfor lipid A.

Further immunoblotting studies with membrane andribosomal preparations as well as with ribosomal subunitsfrom E coli were carried out to localise the PBC-specifictarget antigens in bacterial cells. The AMA-reactive 52 kDpolypeptide was detectable in both whole membrane andouter membrane fractions from wild-type and R-mutantbacteria (fig 2), in the whole ribosomal fraction, in the crude70 S ribosomal fraction, and in 30 S ribosomal subunits, butwas absent from 50 S ribosomal subunits. Reactivity withthe 70-80 kD polypeptide was seen with all bacterialmembrane fractions and with whole E coli ribosomes, butnot with more highly purified ribosomal 30 S and 50 Ssubunits.

Reactivity of Rabbit Antisera

Antisera raised in rabbits against all the bacterial strainsused in the study were tested by immunoblotting withmitochondria from bovine heart and pig kidney as antigensas well as by immunofluorescence on mouse kidney sections.Antisera against S minnesota R345 (Rb mutant) or S

minnesota R5 (Rc mutant) recognised polypeptides ofmolecular weight 70 kD and 50 kD, corresponding to themitochondrial binding patterns obtained with serum from

1168

Fig 1-Immunoblot patterns of rabbit antiserum against Sminnesota R5 and AMA-positive serum from patient with PBCtested with mitochondria from bovine heart (B) and pig kidney(S) as antigens.

PBC patients (fig 1). Rabbit antisera against wild-formEnterobacteriaceae did not react with mitochondria.Antisera against S minnesota R345 and R5 showed a

mitochondrial fluorescence pattern on mouse kidneysections, but antisera against wild-type bacteria were

negative.

Two-dimensional Immunoblot Studies

After two-dimensional gel electrophoresis, immuno-blotting was carried out to confirm the identity of antigensreacting with AMA from patients and with rabbit antibodiesagainst sahnonella R-mutants. Between pH 4-5 and 5-8,AMA from patients with PBC recognised up to nine

separate bovine mitochondrial polypeptides in three groupsat 70 kD, 50 kD, and 42 kD. The AMA binding patternswere very similar to those formed by rabbit antibodiesagainst S minnesota R5. Two polypeptides at 70 kD and twoof five polypeptides at 50 kD were recognised both by PBCpatients’ serum and by rabbit antisera against S minnesotaR5; polypeptides at 42 kD reacted only with patients’samples.

Absorption Studies

Absorption studies were carried out to determine whetherserum from patients with PBC and rabbit antisera againstbacteria recognise crossreactive mitochondrial and bacterialpolypeptides. When serum samples from PBC patients withAMA reactivity were absorbed with mitochondria, they losttheir reactivity with the mitochondrial 70 kD and 50 kDpolypeptides and with the bacterial polypeptides at 75-80kD and 50-52 kD. When AMA-positive samples frompatients were absorbed with S minnesota R5, their reactivitywith bacterial polypeptides was abolished, whereas

reactivity with mitochondrial 70 kD and 50 kD polypeptideswas only diminished. Absorption of rabbit antiserum

against S minnesota R5 with mitochondria abolished thereaction with both 70 kD and 50 kD mitochondrial

Fig 2-Immunoblot of serum sample from a patient with PBC(AMA pattern type III in table III) tested with E coli whole bacteriaand subfractions as antigens.

Lane 1 = E coli 06 K2:H 1 whole bacteria; lane 2 = whole membranepreparation E coh Kl:H7; lane 3=outer membrane preparation E colaKl :H7; lane 4 = whole membrane preparation E coli 0114:H4; lane 5 = wholemembrane preparation E coli 0114:H9; lane 6=ribosomes E coli; lane7 = ribosomal 30 S subunits: lane 8 = ribosomal 50 S subunits.

polypeptides and their bacterial counterparts at 75-80 kDand 50-52 kD.

Discussion

AMA-positive serum samples from 45 patients with PBCrecognised up to three polypeptides of molecular weight 70kD, 50 kD, and 42 kD in our immunoblot studies withmitochondria from bovine heart and pig kidney as antigensources. Other studies had similar findings with three orfour PBC-specific mitochondrial polypeptides,9,15 whichhave been attributed to the M2 complex located on the innermitochondrial membrane. 1029Our immunoblot studies with various gram-negative

bacteria as antigens showed that PBC serum samples reactwith bacterial polypeptides of molecular weight 70-100 kDand 50-52 kD which are present in nearly all gram-negativebacteria tested but are absent from Archaebacteria. Sincethese reactions with bacterial polypeptides are stronglyassociated with PBC, the molecular weight range of bacterialpolypeptides corresponds to PBC-specific mitochondrialpolypeptides, and the antibody activity can be absorbed bymitochondria, there is evidence that the bacterial

polypeptides are recognised by AMA. Reactivity of PBCserum with polypeptides from E coli was initially describedby Lindenborn-Fotinos et apo and confirmed by Frazer etall2 by means of the immunoblot technique.Our immunoblot studies with purified subcellular

fractions from E coli show that at least one of the AMAreactive bacterial polypeptides is located in ribosomes. Theband at 52 kD was found with 30 S ribosomal subunits as the

antigenic fraction but not with 50 S ribosomal subunits.Therefore, we suggest that the 50-52 kD polypeptide is aribosomal component. This conclusion accords with the

findings of Wittmann et al 30 defming the molecular weightsof ribosomal proteins from E coli in the range 10-55 kD, andis supported by the consistent reaction of AMA with variousbacterial strains of a polypeptide of about 50 kD molecularweight. Such consistency is characteristic of ribosomal

1169

proteins (M. Achtman, personal communication). E colipolypeptides of molecular weight 75-80 kD might representmembrane proteins or proteins associated with ribosomes.Variation in reactions at 70-100 kD in various gram-

negative bacteria is consistent with the phylogeneticallydetermined wide variety of membrane proteins. Thepresence of both proteins in crude ribosomal and membranefractions might be attributable to contamination of thesepreparations.Immunisation of rabbits with the R345 and R5 mutants of

S minnesota induced antibodies which reacted withmitochondria in immunoblot studies and on tissue sectionsfrom mouse kidneys in a pattern that closely resembled thatestablished for serum from PBC patients. Two-dimensionalimmunoblotting and absorption experiments withmitochondria supported our assumption of partial identityof these mitochondrial and bacterial polypeptidesrecognised by experimental antisera.

While immunisation with R-mutants results in antibodyproduction against core oligosaccharides and proteincomponents of the cell wall, immunisation with wild-formbacteria primarily induces formation of antibodies againstpolysaccharide chains of lipopolysaccharide molecules. TheR-mutants differ from wild forms mainly in defective

synthesis of the saccharide portion of lipopolysaccharide.This results in a different molecular organisation of the cellmembrane. In consequence, R-mutants might possessseveral different characteristics, especially those related toimmunogenic properties,3’ sensitivity to detergents or

hydrophobic antibiotics,32 and hepatic clearance.3334Since the structure and metabolism of mitochondria are

similar to those of aerobic bacteria, it has been suggested thatmitochondria have evolved from free-living prokaryotestaken up by amoeboid cells with subsequent integration ofmitochondrial metabolism.35 This endosymbiont theoryimplies that the outer mitochondrial membrane is derivedfrom the protoeukaryote and that the inner mitochondrialmembrane, including the cristae, represents the plasmamembrane of the integrated bacterium. This hypothesis hasfound additional support from studies showing thatbacterial and mitochondrial ribosomal RNA moleculesshare several regions of preserved primary and secondarystructures.36 The presence of polypeptides in bacterialribosomes and membrane fractions which react with

PBC-specific AMA, as shown in this study, may beinterpreted as evidence supporting this theory on the basis ofa structural analysis of protein antigens. It is conceivablethat PBC-specific AMA might react with mitochondrialpolypeptides located on the inner membranes andribosomes of mitochondria. This hypothesis accordswith the finding that there is no antigenic relation betweenthe two dominant PBC-specific mitochondrial poly-peptides. 15,16The phylogenetic conservation and wide prevalence of

such bacterial and mitochondrial polypeptides imply theexistence of immune tolerance to these antigens in man.Such immune tolerance may be terminated when

polypeptides are presented in a strong immunogenic form orwhen a defect in the immune system occurs, possiblybecause of a genetic predisposition. We propose the

hypothesis that the AMA in patients with PBC are inducedby R-mutants of gram-negative bacteria which may bepresent in the gut. Although AMA probably have nopathogenetic importance in PBC, our fmdings suggest abacterial aetiology of the disease.

We plan to test our hypothesis by determining whetherPBC patients carry abnormally large numbers of intestinalR-mutants. Since R-mutants will probably be found inhealthy subjects also, it will be necessary to characterise theR-mutants’ immunogenic properties towards the AMA-reactive polypeptides. It would then be clear whether thecharacteristics of the R-mutants predispose patients to PBCor whether a defect in the immune reaction to bacterial

antigens predominates. We plan to include urine samples inour studies because of the reported association betweenurinary tract infection and PBC.3’ It is interesting thatOlling and colleagues311 found a larger proportion of E coliR-mutants with greater sensitivity to bacterial antibodies inthe stools of children with symptomless urinary tract

infection than in those of healthy children; the R-mutantswere also found in the urine of the children with infection.

This work was supported by the Forschungsschwerpunkt "Hepatitis" ofthe Freie Universitat Berlin. We thank Dr M. Achtman and Dr W.Strittmatter (Max-Planck-Institut fur Molekulare Genetik, West Berlin) forNeisseria meningmdis G 2019 strain and for helpful suggestions; Dr L. Beutin(Robert-Koch-Institut, West Berlin) for donation of membrane fractionsfrom E coli; and Mrs K. Langmach for technical help.

Correspondence should be addressed to U. H., Medizinische Klinik undPoliklinik, Universitätsklmikum Charlottenburg, Spandauer Damm 130,1000 Berlin 19, Federal Republic of Germany.

REFERENCES

1. Walker JG, Doniach D, Roitt IM, Sherlock S. Serological tests in the diagnosis ofprimary cirrhosis. Lancet 1965; i: 827-31.

2. Sherlock S, Scheuer PJ. The presentation and diagnosis of 100 patients with primarybiliary cirrhosis. N Engl J Med 1973; 289: 674-78.

3. Doniach D, Roitt IM, Walker JG, Sherlock S. Tissue antibodies m primary biliarycirrhosis, active chronic (lupoid) hepatitis, cryptogenic cirrhosis and other liverdiseases and their clinical implications. Clin Exp Immunol 1966, 1: 237-62.

4. Thomas HC, DeVilliers D, Potter BJ, et al. Immune complexes in acute chronic liverdiseases. Clin Exp Immunol 1978; 31: 150-57.

5. Fakunle YM, Aranguibel F, DeVilliers D, Thomas HC, Sherlock S. Monomeric (7S)IgM in chronic liver disease. Clin Exp Immunol 1979; 38: 204-10

6. Galbraith RM, Smith M. Mackenzie RM, Tee DE, Doruach D, Williams R. Highprevalence of seroimmunologic abnormalities in relatives of patients with activechronic hepatitis or primary biliary cirrhosis. N Engl J Med 1974; 290: 63-69.

7. Culp KS, Fleming CR, Duffy J. Autoimmune associations in primary biliary cirrhosis.Mayo Clin Proc 1982; 57: 365-70

8. Kaplan MM, Gandolfo JV, Quaroni EG. An enzyme-linked immunosorbent assay(ELISA) for detecting antimitochondrial antibody. Hepatology 1984; 4: 727-30.

9. Baum H, Palmer C. The PBC-specific antigen. Mol Aspects Med 1985; 8: 201-34.10. Lindenborn-Fotinos J, Baum H, Berg PA Mitochondrial antibodies in primary

biliary cirrhosis species and nonspecies specific determinants of M2 antigenHepatology 1985; 5: 763-69.

11. Mendel-Hartvig I, Nelson BD, Loof L, Totterman TH. Primary biliary cirrhosis:further biochemical and immunological characterization of mitochondrial antigens.Clin Exp Immunol 1985; 62: 371-79

12. Frazer IH, Mackay IR, Jordan TW, Whittingham S, Marzuki S. Reactivity ofanti-mitochondrial autoantibodies m primary biliary cirrhosis; definition of twonovel mitochondrial peptide autoantigens. J Immunol 1985; 135: 1739-45.

13. Alderuccio F, Toh BH, Barnett AJ, Pedersen JS. Identification and characterization ofmitochondrial autoantigens in progressive systemic sclerosis: identity with the72 000 dalton autoantigen m primary biliary cirrhosis. J Immunol 1986; 137:1855-59.

14. Ishii H, Saifuku K, Namihisa T. Reactivities and clinical relevance ofantimitochondrial antibodies to four mitochondrial inner membrane proteins insera of patients with primary biliary cirrhosis. Hepatology 1987; 7: 134-36.

15. Manns M, Gerken G, Kyriatsoulis A, Trautwein C, Reske K, Meyer zumBuschenfelde KH. Two different subtypes of antimitochondrial antibodies areassociated with primary biliary cirrhosis. identification and characterization byradioimmunoassay and immunoblotting. Hepatology 1987; 7: 893-99.

16. Gerschwin ME, Mackay IR, Sturgess A, Coppel RL. Identification and specificity of acDNA encoding the 70 kD mitochondrial antigen recognized in primary biliarycirrhosis. J Immunol 1987; 138: 3525-31.

17. Uzoegwu PN, Baum H, Williamson J. Detection of an antigen to primary biliarycirrhosis in wild type and petite mutant Saccharomyces cerevisiae. Cell Biol Int Rep1984; 8: 987-92.

18. Scheuer PJ. Primary biliary cirrhosis. In: Liver biopsy interpretation, 3rd ed. LondonBaillière Tindall, 1980: 47-56

19. Hopf U, Meyer zum Buschenfelde KH, Amold W. Detection of a liver membraneautoantibody in HBsAg-negative chronic active hepatitis. N Engl J Med 1976; 294:

1170

20. Jahn HU, Hopf U, Moller B, Stemerowicz R, Klein R, Berg PA. Liver membraneautoantibodies (LMA) in chronic hepatitis of the autoimmune type:characterization of the target antigen. J Hepatol 1987; 5: 535.

21. Martinez P, McCauley R. Studies on the flavins in rat liver mitochondrial outermembranes. Biochim Biophys Acta 1977; 497: 437-46.

22. Achtman M, Schwuchow S, Helmuth R, Morelli G, Manning PA. Cell-cell-interaction in conjugating Escherichia coli: con mutants and stabilization of matingaggregates. Mol Gen Genet 1978; 164: 171-83.

23. Manning PA, Beutin L, Achtman M. Outer membrane of Escherichia coli: propertiesof the F sex factor traT protein which is involved in surface exclusion. J Bacteriol1980; 142: 275-95.

24. Grote M, Dijk J, Reinhardt R. Ribosomal and DNA binding proteins of thethermoacidophilic archaebacterium Sulfolobus acidocaldarius. Biochim BiophysActa 1986; 873: 405-13.

25. Freudenberg MA, Bog-Hansen TC, Back U, Galanos C. Interaction of

lipopolysaccharides with plasma high-density liproprotein in rats. Infect Immum1980; 28: 373-80.

26. Laemmli UK. Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature 1970; 227: 680-85.

27. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins frompolyacrylamide gels to nitrocellulose sheets: procedure and some applications. ProcNatl Acad Sci USA 1979; 76: 4350-56.

28. O’Farrel PH. High resolution two-dimensional electrophoresis of proteins. J BiolChem 1975; 250: 4007-21.

29. Berg PA, Doniach D, Roitt IM, Cooper HM. Mitochondnal antibodies in primary

biliary cirrhosis IV. Significance of the membrane structure for the complementfixing antigen. Immunology 1969; 17: 281-87.

30. Wittmann HG, Littlechild JA, Wittman-Liebold B. Structure of ribosomal proteinsIn: Chambliss G, Craven GR, Davies J, Davis K, Kahan L, Nomura M (eds).Ribosomes—structure, function and genetics. Baltimore: University Park Press,1980: 51-88.

31. Galanos C, Freudenberg MA, John K, Freudenberg N, Luderitz O. Bacterialendotoxins: chemical and biological properties. In: Keppler D, Popper H, BianchiC, Reutter W. Mechanism of hepatocyte injury and death. Lancaster: MTP Press,1983: 203-13.

32. Lugtenberg B, Van Alphen L. Molecular architecture and functioning of the outermembrane of Escherichia coli and other gram-negative bacteria. Biochim BiophysActa 1983; 737: 51-115.

33. Hopf U, Ramadori G, Moller B, Galanos C. Hepatocellular clearance function ofbacterial lipopolysaccharides and free lipid A in mice with endotoxic shock Am JEmergency Med 1984; 2: 13-19.

34. Freudenberg MA, Freudenberg N, Galanos C. Time course of cellular distribution ofendotoxin in liver, lungs and kidneys of rats. Br J Exp Pathol 1982; 63: 56-64

35. Margulis L. Origin of eukaryotic cells. New Haven: Yale University Press, 1970.36. Künzel H, Kochel HG Evolution of rRNA and origin of mitochondria. Nature 1981,

293: 751-55.37. Boroughs AK, Rosenstein IJ, Epstein O, et al. Bactenuna and primary biliary

cirrhosis. Git 1984; 25: 133-37.38. Oiling S, Hanson LA, Holmgren J, et al. The bactericidal effect of normal human

serum on E coli strains from normals and from patients with urinary tract infectionsInfection I 1973; 1: 24-28.

EFFECT OF ACUTE ADMINISTRATION OFINSULIN-LIKE GROWTH FACTOR I IN

PATIENTS WITH LARON-TYPE DWARFISM

Z. LARONB. ERSTER

B. KLINGERS. ANIN

Institute of Pediatric and Adolescent Endocrinology, BeilinsonMedical Center, Sackler Faculty of Medicine, Tel Aviv University,

Israel

Summary Biosynthetic insulin-type growth factor I(IGF-I) was given as an intravenous bolus

of 75 µg/kg to 9 patients with Laron-type dwarfism. Rapidonset of hypoglycaemia (-45% of basal level) was

associated with a reduction in plasma insulin (-55% ofbasal level); thus, the lack of growth hormone receptors inthis condition does not apply to IGF-I receptors andpost-receptor pathways. Long-term treatment maytherefore be beneficial in Laron-type dwarfism.

INTRODUCTION

IN 19661 and 1968, we described a new hereditarysyndrome which clinically (dwarfism, obesity, and

micropenis), and in terms of metabolic responses

(hypoglycaemia and high free fatty acids), was

indistinguishable from isolated growth hormone deficiency.The characteristic difference between the two diseases wasthe high level of circulating growth hormone (GH) in thenew syndrome. Subsequent investigations showed that, inthese patients, the GH was similar in immunologicaP andbiological4 behaviour to that of the hormone in healthysubjects. In these patients, the inactivity of circulatinghormone, as well as that of exogenous GH, is due to a lack ofGH receptors,5 so that they are unable to produce insulin-like growth factor (IGF-I, somatomedin-C).6Recombinant DNA technology has made possible the

TABLE I-CLINICAL DATA FOR STUDY PATIENTS

I I I I -

I

CA = chronological age; BA = bone age; A = adult.

TABLE II-BLOOD GLUCOSE RESPONSE TO INTRAVENOUS IGF-I

BOLUS

Glucose in mmol/1.

biosynthesis of IGF-I; we report the acute response ofpatients with Laron-type dwarfism to the administration ofrecombinant (r) IGF-I.

SUBJECTS AND METHODS

5 male and 4 female patients with Laron-type dwarfism, aged 11to 33 years of age, were studied (see table i). IGF-i was synthesisedby recombinant DNA technology’ (Fujisawa Pharmaceutical

Company, Osaka) and had an identical aminoacid sequence to thatof natural IGF-I. The purity of the preparation (lot 144770 K) was99% as determined by reverse phase high pressure liquidchromatography. The content of IGF-I was 97 9%, and that ofmet59(0)-IGF-i was 2-1%.