Auroimmunify. 1992, Vol. 14. pp. 105-1 10 Reprints available directly from the publisher Photocopying permitted by license only

0 1992 Harwood Academic Publishers GmbH Printed in the United States

A NEW PLAYER IN THE ANTIPHOSPHOLIPID SYNDROME: THE p2 GLYCOPROTEIN I COFACTOR

GUIDO VALESINI* and YEHUDA SHOENFELDT V l i n i c a Medica I , Fondazione A. Cesalpino, Universita di Roma “La Sapienza” ?Research Unit of Autoimmune Diseases, Deparrement of

Medicine B , Sheba Medical Center, Tel Hashomer

(Received April 21,1992; in final form August 20,1992)

The study of antiphospholipid (aPL) antibodies has been greatly developed in recent years and conclusive evidence now exists Concerning the correlation between aPL and clinical signs such as thrombosis, throm- bocytopenia, abortion, and fetal loss.

Several hypotheses have been put forward concerning the pathogenic mechanism of aPL, but none has received final confirmation from experimental data.

Many studies have been devoted to characterizing the antigens recognized by the different aPL autoanti- bodies and to a cofactor involved in the binding of autoantibodies and phospholipids; this cofactor has been identified as an apolipoprotein, the 82 glicoprotein I (WGPI) or APO-H.

Direct evidence now exists which suggests that both the a G P 1 and the phospholipid comprise the epitope to which aPL are directed. On the other hand anti-bGPI antibodies have been identified in sera of patients suffering from SLE and primary Antiphospholipid Syndrome.

AGPI is normally present in human plasmaherum and possesses numerous inhibitory functions in multiple coagulation pathways. The amino acid sequence of bGP1 has been identified and found to consist of five repeating units that belong to the complement control protein (CCP) superfamily.

This development of knowledge related to aPL has followed three steps respectively: I . the standardization of the techniques of detection; 2. identification of the clinical signs related to the autoantibodies; and finally 3. the discovery of a new player, the AGPI cofactor.

KEY WORDS: Antiphospholipid, glycoprotein, APO-H.

THE ANTIPHOSPHOLIPID SYNDROME (APLS)

The study of antiphospholipid (aPL) antibodies has been greatly developed in recent years since specific and sensitive tests became available.

The anticardiolipin (aCL) assay as originally described by Harris et a/.’ and subsequently modified by Loizou et aL2 has played a central role in increas- ing our knowledge in this field. A large number of sera, from patients suffering from autoimmune and infectious diseases have been screened and conclusive evidence now exists concerning the correlation between aPL autoantibodies detectable in Systemic Lupus Erythematosus (SLE) and related conditions and clinical signs such as thrombosis, thrombocyto- penia, abortion, fetal loss, etc3.

The old phenomenon of Lupus Anticoagulant (LAC) has also been regarded in a different light as a method of detection of aPL antibodies using coagu- lation assay4. The close association of serological abnormalities detectable in SLE and related to the aPL antibodies and the above mentioned clinical signs, has led to the definition of the APLS (Table 1)5*6.

Correspondence to: Prof G. Valesini, Clinica Medica 1. Universiti di Roma “La Sapienza”. Policlinico Umbeno I. 00161 Roma ( I tah) .

aCL and LAC have been considered as playing a pathogenic role in the induction of the clinical mani- festations reported in Table 1. Several hypotheses have been put forward concerning this pathogenic mechanism (Table 2), but none has received final con- firmation from experimental data’.

Table 1 Characteristics of the antiphospholipid syndrome

I . Antiphospholipid (aPL) antibody detectable as anticardiolipin (aCL), Lupus Anticoagulant (LAC), and VDRL.

2. Venous and arterial thrombosis, thrombocytopenia, recurrent abortion or fetal loss, hemolytic anemia, leg ulcers, arterial occlusions, livedo reticularis. and transverse myelitis.

Presence of circulating aPL antibody and at least two of the aforementioned signs and symptoms.

Table 2

-e f fec t on endothelial prostacyclin production (?) -inhibition of thrombomodulin-mediated protein C activation (?) -inhibition of fibrinolytic activity (?) -inhibition of antithrombin 111 activity (?) -functional effect on platelets and/or erythrocytes (?)

Mechanisms proposed to explain aPL pathogenicity

I05

Aut

oim

mun

ity D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y O

saka

Uni

vers

ity o

n 12

/01/

14Fo

r pe

rson

al u

se o

nly.

106 G. VALESINI AND Y. SCHOENFELD

aCL was shown to play a pathogenic role in two recently developed experimental models; in one, poly- clonal aCL, purified from a patient with APLS, and mouse monoclonal aCL were passively transfused into the tail vein of ICR and BALB/C mice. The trans- fusion was followed by the symptoms of APLS. The experimentally induced APLS was characterized by thrombocytopenia, coagulation abnormalities, and increased resorption rate of fetuses (the equivalent of recurrent fetal loss in humans)'.

In another experiment, the same results were observed after active immunization with aCL in the footpads'.

The increase of knowledge concerning aPL in recent years raises two intriguing questions about: 1. the identity of the antigens recognised by the different aPL autoantibodies (aCL, LAC, etc.); 2. the identity of aPL autoantibodies in both autoimmunity and infectious diseases.

THE aCL COFACTOR

More information is now available on these two points; in fact, McNeil et al." have recently perfor- med experiments leading to the following con- clusions:

aPL, purified from plasma (or serum) by sequential phospholipid affinity, and cation-exchange chrom- atography, exhibit typical binding i n aCL ELISA but do not possess LAC activity in phospholipid- dependent clotting tests. This strongly suggest that aCL and LAC represent distinct antibody subgroups' ' * '* ;

purified aPL bind to the phospholipid antigen only when assayed in the presence of normal human plasma, human serum or bovine serum, suggesting the existence of a plasma/serum cofactor. The cofactor has been identified as an apolipoprotein: APO-H of pz glicoprotein I (p2 GPI)'3.

Galli et al. published similar results leading to the conclusion that aCL antibodies are directed not to car- diolipin, but to the plasma protein cofactor14.

Data from different laboratories have subsequently demonstrated that the pZ GPI found in sera or added as purified material, enhances binding of aCL to CL in a dose dependent fashion. However, the antigen recog- nised by these autoantibodies seems to be a confor- mational epitope resulting from the interaction between p2 GPI and (anionic) ph~sphol ipids '~*l~ '" .

In Galli's view, the autoantibodies recognize an epitope on the p2 GPI and this apolipoprotein binds to the negatively charged ph~spholipid'~.

This hypothesis could explain the cross-reaction of aCL with all the anionic phospholipids, even though

the spatial headgroups geometry differs markedly between these molecules'".

More direct evidence now exists which suggest that both the p2 GPI and the phospholipid comprise the epitope to which aCL antibodies are directed. In fact, it is well known that pt GPI also binds to heparin and DNA, but aCL antibodies were found not to bind to a heparin-& GPI complex, or to pz GPI covalently linked to sepharo~e '~.

On the other hand, Tincani et al." have recently reported the presence of anti pz GPI antibodies in sera of patients suffering from SLE and primary APS. These authors have employed affinity purified IgG from patient sera in a solid phase radioimmunoassay. Cross-inhibition experiments suggest the association of 2 different antibody populations (antiphospholipid and anti p2 GPI).

Shapiro el al." reported similar results using affin- ity purified aCL on a solid phase ELISA; they pointed out the experimental evidence that aCL react with neo-antigen on pz GPI which becomes exposed when this protein forms a complex with anionic phospho- lipid, and that this neo-antigen is similar or identical to that exposed when GPI is absorbed into the sur- face of a microtiter well'".

FACTS ABOUT pz GPI (TABLE 3)

The p:! GPI is normally present in human plasma/serum at a concentration of 200 pg/ml. Pre- viously, it was reported" that the amount of pz GPI in plasma is genetically determined in such a way that concentrations are controlled by a pair of autosomal codominant alleles, BgN and BgD. Individuals homo- zygous for BgN have p2 GPI concentrations higher than 160pglml (range 160-300): this is the N (normal) phenotype.

Individuals homozygous for BgD have p2 GPI con- centrations lower than 50 pg/ml (phenotype D, decreased) and those heterozygous have an amount of pz GPI comprised between 50-140 pg/ml (phenotype I, intermediate).

Cleve has reported that the concentration of p2 GPI in serum is also influenced to some extent by non gen- etic factors such as sex, age, and various pathologic conditions; it is frequently decreased in liver cir- rhosis2Is''. The decrease is highly correlated with pseudocolinesterase activity, this is presumably a reflection of the fact that both pz GPI and pseudo- colinesterase are produced in the liverz3.

The Pt GPI concentration increases with age: in fact the level in newborn infants is significantly lower than in the adult".

p2 GPI concentrations are increased in smokers" and in patients suffering from diabetes mellitus with retinopathy". I t is not yet clear how this finding fits in

Aut

oim

mun

ity D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y O

saka

Uni

vers

ity o

n 12

/01/

14Fo

r pe

rson

al u

se o

nly.

ANTIPHOSPHOLIPID COFACTOR I07

Table 3 Facts about &glycoprotein I (APO-H). From: CIin ExpRheurnotol 10: 1-5. 1992 (with permission)

1. aCL and LA most probably play a pathogenic role in the induction of the various clinical manifestations (e.g. thromboembolism. recurrent fetal loss) of the anti-phospholipid syndrome.

2. &GPI present in serum or added as purified material. enhances antibody binding to CL.

3. aCL binds to a - G P I complexed with CL on solid phase.

4. &GPI is required for prolongation of the dRVVT clotting time by LA positive plasmas.

5. aCL purified from syphilis sera do not require /?&PI for binding, in this way differing from “autoimmune” aCL.

6. &GPI is a glycoprotein rich in cystein and proline, with a molecular weight of 50 KD.

7. A-GPI displays interspecies homology in sequence despite serological differences in the binding of aCL.

8. &GPI is an inhibitor of the intrinsic blood coagulation pathway, ADP-induced platelet aggregation and prothrombinase activity of activated platelets. It binds to washed platelets.

9. A-GPI binds to negatively charged phospholipid micelles.

10. The normal concentrations of &GPI in serum is around 200 pg/ml.

I I . &-GPI displays higher levels in primary and SLE-associated APLS and no correlation with aCL titers or clinical findings.

12. Anti-&-GPI can be detected in primary and SLE-associated APLS without any correlation with aCL titers or clinical findings.

13. Heterologous &GPI alone or combined with CL is immunogenic in mice and rabbits and can induce the production of both anti-&GPI and aPL antibodies.

14. The binding of &GPI to CL seems to induce changes in phospholipid conformation from the bilayer to the hexagonal form.

with the vascular damage observed in these Table 4 Relationship between GPI and coagulation

conditions. The p2 GPI: The p 2 GPI possesses numerous inhibitory func-

lions in multiple coagulation pathways. One of the most promising aspects of finding p2 GPI as a com- ponent of the aCL antigen is the possibility that aCL -inhibits adenosine diphosphated mediated platelet antibodies interfere with the function of PZ GPI in vivo, thereby conferring a prothrombotic diathesis.

-inhibits anionic macromolecular intrinsic pathway activation”; -acts as an antiprothrombinase in v i f d ;

~ ~ ~ ~ ~ ~ ~ s ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ; i n ”‘;

RELATIONSHIP BETWEEN p2 GPI AND COAGULATION

As reported in Table 4, p2 GPI may interfere with the coagulation pathway at various level^^"^^. The p2 GPI may be regarded, indeed, as a physiologic anticoagu- lant. However, previous studies have reported that the rarely identified Bg’ hymozygotes with undetectable p2 GPI levels do not suffer thrombotic com- p~ica t ions~~.

Since patients with aPL antibodies suffer from thrombosis and aPL antibodies react with p2 GPI, the relationship between serum concentrations of p2 GPI and the levels of aCL autoantibodies have been investigated.

GPI in patients with serum aCL antibodies. McNeil and co- workers”. employing an RIA with iodinated p2 GPI,

Two groups have reported an increase of

found a group of patients with aCL antibodies and very high p2 GPI levels. There was no correlation between PZ GPI levels and the clinical events such as thromboses.

The elevated PZ GPI level may be genetically deter- mined and such a predisposition may represent a pre- requisite to develop antiphospholipid antibodies. On the other hand, the increase in pZ GPI levels may be secondary to the presence of antibodies and represent a compensatory phenomenon.

Tincani and co-workerst9 employed a Single Radial Immuno-Diffusion (SRID) and found a similarly high percentage of high producers of p2 GPI among patients with primary antiphospholipid syndrome (PAPS). In this study as well, pz GPI values did not correlate with clinical signs or laboratory parameters (thrombosis, fetal loss, thrombocytopenia).

Aut

oim

mun

ity D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y O

saka

Uni

vers

ity o

n 12

/01/

14Fo

r pe

rson

al u

se o

nly.

I08 G. VALESINI AND Y. SCHOENFELD

BIOCHEMISTRY OF pz GPI

The p2 GPI, first described by Schultze et a/.” as a perchloric acid soluble glycoprotein of human serum, is a protein of approximately 50 KDa.

The physicochemical and immunochemical proper- ties of this glycoprotein in rat and human are very simi~a?’”~.

The amino acid sequence of /?2 GPI was originally determined by protein-sequencing techniques by Losier et al.” who demonstrated that it is 326 residues in length. The sequence consists of five repeating units (60 amino acid residues), that belong to the complement control protein (CCP) ~upe r fami ly~~ and is generally designed as short consensus repeats (SCRs).

These repeating motifs are presented in proteins involved in the regulation of complement activation and in a growing number of non-complement proteins such as interleukin-2 receptors and leucocyte adhesion molecu~es~~.

A partial cDNA sequence for rat p2 GPI has been published in abbreviated format as well as the com- plete nucleotide and deduced amino acid sequence of human p2 GPI. The major portion of the rat cDNA- derived amino acid sequence corresponds to aa resi- dues 52-326 of the human sequence. The two sequences are 82.5% identical over this region3’.

The site of synthesis of pz GPI was investigated by Northem-blot analysis and experimental evidence indicates that hepatocytes are a major site of biosynth- esis”. The p2 GPI has been reported as behaving as an acute phase protein3*.

The pz GPI as an apolipoprotein binds lipids; the phospholipids bound to p2 GPI undergo structural modification”. In fact, it has been recently demon- strated that /?z GPI is capable of inducing CL to undergo a bilayer to hexagonal transition”. The result of this interaction could be the “de novo” expression of a conformational epitope recognized by a specific autoantibody (Figure 1).

Conflicting results have been reported from immun- ization experiments: Gharavi and co-workers injecting purified human pz GPI into rabbits and mice were able to detect an antiphospholipid and anti p2 GPI activity in the immunized animals. The immunization with CL alone failed to induce either aPL or anti Bz G P P . In Gharavi’s experiments, immunization of mice with purified mouse /3z GPI induced the anti p2 GPI anti- body, but the antisera lacked any aPL activity4’.

On the other hand, Rauch, immunizing mice with cardiolipin mixed with p2 GPI, produced aCL anti- bodies, while mice immunized with CL or /$ GPI alone did not produce aCL antibodies, although they did produce anti p2 GPI antibodies4’.

These two sets of experiments led to the apparently conflicting conclusion that: 1. aPL are induced by the

Figure 1 The CL (a) binds to GPI (b) and undergoes some conformational modifications. This binding leads to the expression of a new epitope (c) recognized by specific autoantibodies (d) present in autoimmune sera.

binding of foreign p2 GPI to endogenous PL “in V ~ V O ” ~ ~ ; 2. The interaction of CL with pz GPI renders CL immunogenic.

In another set of experiments we observed that rab- bits and mice immunized with human p2 GPI synthe- size antibodies reactive with GPI alone and p2 GPI complexed with CL42.

In the evaluation of the above mentioned exper- iments one must take into account the possible inter- ference of endogenous /?* GPI, or the amount of such glycoprotein present in bovine serum or contaminat- ing the preparations of purified albumin generally employed in blocking and/or diluting and/or washing buffers“’.

This latter aspect is of great practical relevance, in fact it may condition the results of the experiments because of what has been called “p2 dependence”

Aut

oim

mun

ity D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y O

saka

Uni

vers

ity o

n 12

/01/

14Fo

r pe

rson

al u

se o

nly.

ANTIPHOSPHOLIPID COFACTOR 109

phenomenon. CabraI4’ has recently reported that the p2 dependence is effective during the blocking step of the ELISA procedures more than during the dilution and washing procedures.

CONCLUSIONS

The field of aPL antibodies has greatly enlarged and increased during the last 10 years. Such an expansion followed three steps: 1 . standardization of the tech- niques of aPL detection; 2. identification of clinical signs and symptoms related to the aPL antibodies; and 3. discovery of a new player, the pZ GPI cofactor.

The next step would be to go deeper inside the pathophysiology of aPL and anti p2 GPI antibodies.

References

I . Harris EN, Boey ML, Mackworth-Young CG. Gharavi AE, Patel BM, Loizou S, Hughes GRV. Anticardiolipin antibodies: detection by radioimmunoassay and association with throm- bosis in Systemic Lupus Erythematosus. Lancet 1983; ii: 121 1-1214

2. Loizou S , McCrea JD. Rudge AC, Reynolds R, Boyle CC. Harris EN. Measurement of anticardiolipin antibodies by an enzyme-linked immunosorbent assay (ELISA): standardization and quantitation of results. Clin E.rp Immunol 1985; 62:

3. Alarcon Segovia D. Delete M, Oria CV. Sanchez-Guerrero J, Gomez-Pacheco L, Cabiedes J. Fernandez L, Ponce De Leon S. Antiphospholipid antibodies and the antiphospholipid syn- drome in Systemic Lupus Erythematosus. A prospective analy- sis of 500 consecutive patients. Medicine 1989; 68: 353-365

4. Derksen RHWN, Kater L. Lupus Anticoagulant: revival of an old phenomenon. CIin Exp Rheumatol1985; 3: 349-357

5 . Sammaritano LR, Gharavi AE, Lockshim MD. Antiphospholi- pid antibody syndrome: immunologic and clinical aspects. Semin Arthritis Rheum 1990; 20: 81-96

6. Hughes GRV. Asherson RA, Khamashta MA. Antiphospholipid

738-745

syndrome: linking many specialties. Annals Rheum D;s 1989; 48: 355-356

7. Meroni PL, Tincani A, Harris EN, Valesini G. Hughes GRV,

8

9.

10.

I I .

12.

13.

Balestrieri G. The pathophysiology of anti-phospholipid anti- bodies. Clin Exp Rheumatol 1989; 71s-3: 81-84 Blank M, Cohen J. Toder V, Shoenfeld Y. Induction of anti- phospholipid syndrome in naive mice with mouse lupus mono- clonal and human polyclonal anti-cardiolipin antibodies. Proc Natl Acad Sci USA 1991; 88: 3069-3073 Bakimer R, Fishman P, Blank M, Sredni B, Djaldelty M, Shoenfeld Y. Induction of Primary Antiphospholipid Syndrome in mice by immunization with a human monoclonal anticardiol- ipin antibody (H-3). JClin Invest 1992; 89: 1558-1563 McNeil HP. Hunt JE, Krilis SA. New aspects of anticardiolipin antibodies. Clin E.rp Rheumatol 1990; 8: 525-527 McNeil HP. Chesterman CN. Krilis SA. Anticardiolipin anti- bodies and lupus anticoagulants comprise separate antibody subgroups with different phospholipids binding characteristics. Erirish Journ of Haemarol 1989; 73: 506-5 13 McNeil HP. Krilis SA. Chesterman CN. Purification of anti- phospholipid antibodies using a new affinity method. Throm- bosis Research 1988; 52: 64 1-648 McNeil HP. Richard JS. Chesterman CN, Krilis SA. Antiphos- pholipid antibodies are directed against a complex antigen that

includes a lipid-binding inhibitor of coagulation: /?z-giyco- protein-I (apolipoprotein-H). Proc Natl Acad Sci USA 1990; 87: 4120-4124

14. Galli M, Comfurios P. Maassen C, Hemker HC, De Baets MH. van Breda-Vriesman PJC, Barbui T, Zwaal RFA, Bevers EM. Anticardiolipin antibodies (ACA) directed not to cardiolipin but to a plasma protein cofactor. Lancet 1990; 335: 154447

15. Harris NE. Pierangeli S. What is the “true’’ antigen for antipho- spholipid antibodies? Lancer 1990; 330, 1505

16. Harris NE, Pierangeli S , Barquinero J. Ordi-Ros J. Anticardio- lipin antibodies and binding of anionic phospholipids and serum protein. Lancer 1990; 336: 505-506

17. McNeil HP. Specificity of antiphospholipid antibodies. Anti- phospholipid antibodynupus Anticoagulant workshop. NIH Bethesda. September 25, 1991

18. Shapiro SS, Thiagarajan P. Essex D, Love T. Studies on the specificity of anticardiolipin antibodies. Antiphospholipid antibodynupus Anticoagulant workshop. NIH Bethesda. Sep- tember 25, 1991

19. Tincani A. /?z GPI seric levels and antibodies against /?* GPI in patients with primary and secondary anti-phospholipid syn- drome. International workshop on B2-GPI and antiphospholipid antibodies. Milano. October 25-26. 1991

20. Atkin J, Rundle AT. Serum /?z-glycoprotein-I phenotype fre- quencies in an English population. Humangenerik 1974; 21:

21. Cleve H. Genetic studies on the deficiency of /?2-glycoprotein-I of human serum. Humangenetik 1968; 5: 294-304

22. Cleve H. Rittner C. Further family studies on the genetic con- trol of /??-glycoprotein-I concentration in human serum. Humangenerik 1969; 7: 93-97

23. Propert DN. The relation of sex, smoking status, birth rank, and parental age to A-glycoprotein-I levels and phenotypes in a sample of Australian Caucasian adults. Human Genet 1978; 43:

24. Schousboe I . /??-glycoprotein I: a plasma inhibitor of the con- tact activation of the intrinsic blood coagulation pathway. Blood 1985; 66: 1086-1091

25. Nimpf J, Bevers EM, Bomans PHH, Till U. Wurm H. Kostner GM, Zwaal RFA. Protrombinase activity of human platelets is inhibited by /?!-glycoprotein-I. Eiochim et Eiophys Acta 1986;

26. Canfield WM, Kisiel W. Evidence of normal functional levels of activated protein C inhibitor in combined vector V/VIII deficiency disease. J Clin Invest 1982; 70: 12661272

27. Nimpf J. Wurm H, Kostner G. /?2-glycoprotein-I (apo-H) inhibits the release reaction of human platelets during ADP- induced aggregation. Atherosclerosis 1987; 63: 109-1 14

28. Nimpf J, Bevers EM. Bomans PHH, Ti1 U, Wurm H, Kostner GM, Zwaal RFA. Protrombinase activity of human platelets is inhibited by /??-glycoprotein-I. Eiochim et Eiophys Acra 1986;

29. Schousboe I. Binding of A-glycoprotein-I to platelets: effects of adenylate cyclase activity. Thrombosis Research 1980 19:

30. Hoeg JM, Segal P, Gregg RE, Chang YS, Lindgren FT, Adam- son GL, Frank M. Brikman C, Brewer HB Jr. Characterization of plasma lipids and lipoproteins in patients with &Glyco- protein I (Apolipoprotein H) deficiency. Artherosclerosis 1985; 55: 25-34

3 I . McNeil HP. Anticardiolipin antibody-p,GPI interaction. Inter- national workshop on /?:GPI and antiphospholipid antibodies. Milano, October 25-26, 1991

32. Shultze HE. Heide K. Haupt M. Uber ein bisher unbekanutes niedermolekulares /??-globulin des humanserums. Narurwissen- shid?en I96 1 ; 48: 7 19-72 I

33. Polz E, Wurm H. Kostner GM. Investigations on /??-glyco- protein-l in the rat; isolation from serum and demonstration in lipoprotein density fractions. / / i t J Eiochem 1980; 1 I: 265-270

8 1-84

28 1-288

884: 142-149

884: 142-149

225-237

Aut

oim

mun

ity D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y O

saka

Uni

vers

ity o

n 12

/01/

14Fo

r pe

rson

al u

se o

nly.

I10 G . VALESINI AND Y. SCHOENFELD

34. Wurm H. &-glycoprotein-I (Apoprotein H) interactions with phospholipid vesicles. l n r J Biochem 1984; 16: 5 11-515

35. Lozier 1, Takashashi N, Putnam FW. Complete amino acid sequence of human plasma A-glycoprotein-I. Proc Narl Acad Sci USA 1984: 81: 364044

36. Reid KBM. Day AJ. Structure-function relationships of the complement components. lmrnunology Today 1989; 10:

37. Steinkasserer A, Estaller C, Weiss EH. Sim RB, Day AJ. Com- plete nucleotide and deduced amino acid sequence of human &-glycoprotein-I. Biochem J 1991; 277: 387-391

38. Bevers EN. New insight on the mode of action of lupus anti- coagulant. International workshop of AGPI and antiphospholi- pid antibodies. Milano, October 25-26. 1991

39. Janoff AS. Relationship of lipid and antibody: phospholipid three dimensional structure determines antigenicity. Antiphos-

177-180

pholipid antibodynupus Anticoagulant workshop. NIH Bethesda. September 25, 1991

40. Gharavi AE, Sammaritano LR. Induction of antiphospholipid antibodies by immunization with a lipid binding protein, /3? glycoprotein. Antiphospholipid antibodynupus Anticoagulant workshop. NIH Bethesda. September 25, 1991

41. Rauch J. Immunology of anti-phospholipid antibodies. Anti- phospholipid antibodykupus Anticoagulant workshop. NIH Bethesda. September 25. 1991

42. Valesini G. Induction of anti-aGPI by active immunization. International workshop on AGPI and antiphospholipid anti- bodies. Milano. October 25-26, 199 1

43. Cabral R. AGPI dependence of antiphospholipid antibodies. International workshop on AGPI and antiphospholipid anti- bodies. Milano, October 25-26, 199 I

Aut

oim

mun

ity D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y O

saka

Uni

vers

ity o

n 12

/01/

14Fo

r pe

rson

al u

se o

nly.