- Email: [email protected]
Purification and characterization of natural antibodies that recognize a human brain lectin
Journal of Neuroimmunology ELSEVIER
Journal of Neuroimmunology 57 (1995) 9-15
Purification and characterization of natural antibodies that recognize a human brain lectin Didier Lutomski a, Michel Caron alb,Philippe Bourin ‘, Christine Lefebure b, Dominique Bladier a,b, Raymonde Joubert-Caron a,* a Laboratoire de Biochimie et Technologie des Prot&nes, Universite' Paris-Nord, UFR “SMBH L6onard de Vinci”, 74 rue Marcel Cachin, 93012 Bobigny Cedex, France b Laboratoire Central de Biochimie, H6pitaf Avicenne, Bobigny, France ’ Laboratoire d’lmmunologie Cellulaire, Centre de Transfusion Sanguine des Armies, H6pital Percy. Clamart, France
Received 10 June 1994; revised 1 September 1994; accepted 23 September 1994
Abstract
We have recently identified oligoclonal IgG antibodies that are related to a human brain lectin (HBL14) from serum and cerebrospinal fluid of patients with neurological disorders. They were termed lectin-like IgG (L-IgG) (Joubert-Caron et al., 1994a,b). In this paper, the occurrence of antibodies reactive both towards HBL14 and L-IgG was investigated. Binding of antibodies to HBL14 was demonstrated by solid-phase ELISA and chromatography on immobilized HBL14. Fab fragments of these antibodies were also shown to bind to HBL14. The specificity of the antibodies towards HBL14 was studied using a panel of different antigens. Our data show that individual sera from healthy people as well as a pool of immunoglobulins from 80 blood donors contain an IgG autoreactivity to HBL14, while no IgM autoreactivity was detected. Anti-HBL14 antibodies from sera were purified using affinity chromatography on immobilized HBL14. Affinity chromatography further allowed us to demonstrate that the binding of anti-HBL14 antibodies was mediated through their Fab fragments. A higher amount of anti-HBL14 antibodies was purified using a L-IgG-depleted fraction of sera. The binding of anti-HBL14 antibodies to L-IgG was confirmed by ELISA. Finally, anti-HBL14 antibodies were found to be polyreactive. These results indicate the occurrence of a novel class of natural antibodies reactive towards a human brain lectin and suggest that these antibodies may participate in immunoregulatory mechanisms probably though idiotypic/anti-idiotypic interaction. Keywords:
Natural autoantibodies;
Anti-idiotype
antibodies;
Brain lectin; Immunoregulation
1. Introduction An endogenous lectin from human brain (HBL14), belonging to the well-conserved 1CkDa lectin family showing a specificity for /?-galactoside structures (for review see Caron et al., 1990; Harrison, 19911, has been purified (Bladier et al., 1989) and fully sequenced (Bladier et al., 1991). A recombinant protein (rIML1) (Couraud et al., 1989) identical to HBL14 and the homologous molecule found in electric organ of Torpedo electricus (Teichberg et al., 1975) have been shown to possess immunoregulatory properties suppressing
* Corresponding 48 38 77 77
author. Phone (+33-l)
0165-5728/95/$09.50
48 38 77 54; Fax (+33-l)
0 1995 Elsevier Science B.V. All rights
SSDIO165-5728(94)00152-9
reserved
experimental autoimmune myasthenia gravis (Levi et al., 1983) or experimental autoimmune encephalomyelitis (Offner et al., 1990). However, mechanisms involved in immune responses mediated by 14kDa lectins remain ill-defined. Recently, oligoclonal P-galactoside-binding y-immunoglobulins (IgG), antigenically related to HBL14, have been detected in serum and CSF from patients with neurological disorders, and purified from serum of healthy people using a specific anti-HBL14 antibody as immunosorbent (Joubert-Caron et al., 1994a,b). They were termed lectin-like IgG (L-IgG) and were found to react with cellular antigenic determinants expressed in mouse brain after surgical injury (Joubert-Caron et al., 1994b). On the basis of (i) the reactivity of the L-IgG Fab fragments to a rabbit anti-HBL14 antibody, in-
10
D. Lutomski et al./Journal of Neuroimmunology57 (1995) 9-15
hibitable by HBL14 and (ii) a specificity for P-galactosides close to that of the lectin, it was suggested that L-IgG may be anti-idiotypic antibodies (Ab2) mimicking the human brain lectin (Joubert-Caron et al., 1994a,b). The presence of antibodies (Abl) directed against HBL14 or related molecules, which may have generated L-IgG (Ab2) was thus speculated. We have therefore investigated the presence of antibodies reacting towards HBL14 in normal sera and sera from patients. In the present study, antibodies reacting with HBL14 were identified in the sera of normal individuals by ELISA, and were purified by immunoaffinity chromatography on immobilized HBL14, from a pool of healthy blood donors. The specificity of the purified antibodies was evaluated by using different antigens and their immunoreactivity against L-IgG was studied.
2. Materials and methods 2.1. Serum samples and antigens Sera from healthy blood donors were obtained from the blood bank at the Centre de Transfusion Sanguine des ArmCes (CTSA). HBL14 was prepared according to Avellana-Adalid et al. (1990). L-IgG and L-IgG-depleted fraction were prepared as reported by Joubert-Caron et al. (1994a). Actin (calf muscle), cytochrome c (bovine heart), myelin basic protein (bovine brain) and thyroglobulin (porcine type II) were purchased from Sigma, myoglobin (bovine heart type II) was obtained from Calbiochem (San Diego, CA), and laminin was a gift from J. Botty from Dr. M. Aubery’s laboratory, INSERM U 180 (Paris, France).
M citrate buffer (pH 5.2) and 10 ~1 H,O,). The reaction was stopped with 100 ~1 of 0.5 M H,SO, for 15 min. Absorbance was measured at 492 nm and the results were expressed as optical density COD). To ascertain that the interaction was not due to a lectinsugar recognition, control experiments were performed in presence of 0.1 M lactose in the incubation buffer. 2.3. Human sera and immunoglobulin preparations IgG were purified by thiophilic adsorption (Lihme and Heegaard, 1991; Joubert-Caron et al., 1994a) from a pool of sera from healthy blood donors according to Hutchens (1992). Briefly 40 ml of a pool of human sera were supplemented for 1 h at 4°C with 10% (w/v> solid ammonium sulfate, then centrifuged at 10 000 X g for 10 min at 4°C. The supernatant was passed at a rate of 0.5 ml/min through 40 ml of thiophilic adsorbent (T-gel) (7 cm x 2.5 cm) equilibrated with 50 mM Tris containing 10% (w/v) ammonium sulphate at 15°C. After washing with the same buffer until the OD,,, reached zero, the adsorbed IgG were eluted with 0.1 M NaBO,, 4 H,O/HCl (pH 8.5). The IgG fraction was biotinylated (Biot-IgG) by incubation with NHS-Biotin (Sigma, La Verpillikre, France) at 1 mg/ml in DMSO used at a final ratio of 10 pg/mg of IgG, for 4 h at 20°C. IgG-depleted serum and Biot-IgG were dialysed against TBS (50 mM Tris/CH,COOH pH 7.4, 150 mM NaCl, 0.02% (w/v) NaN,) and stored at 4°C until used. Albumin, IgG, IgA and IgM concentrations were determined in the pass-through and in the eluted fractions, using laser immunonephelometry (BNA, Behring).
2.2. ELISA for anti-HBL14 antibodies
2.4. Isolation of anti-HBLl4 ity chromatography
antibodies by immunoaffin-
Microtiter plates were coated with 100 ~1 of purified HBL14 (Avellana-Adalid et al., 19901, at a concentration of 1 pg/ml in incubation buffer (0.4 M NaCl, 0.1 M K,HPO, /KH,PO, pH 7.4, EDTA 1 mM, 0.01% BSA (w/v)) for 2 h at 37°C and overnight at 4°C. After washing with T-PBS (10 mM K,HPO,/KH,PO, pH 7.4, 0.05% (v/v) Tween 20>, the plates were blocked with 300 ~1 of 3% (w/v> BSA in T-PBS, for 2 h at 37”C, then incubated with 100 ~1 of 1:400 to 1:1600 dilutions of human serum samples in triplicate, for 1 h at 37°C. After washing, positive wells were detected with either rabbit anti-human IgG or anti-IgM conjugated to horse radish peroxidase (antiHRP IgG or anti-HRP IgM) (Dakopatts, Denmark) used at 1:5000, by 1 h incubation at 37°C. Enzymatic activity was revealed using 100 ~1 orthophenylene diamine chloride (OPD) as substrate (30 mg/15 ml of 0.1
HBL14 purified by affinity chromatography was coupled to divinylsulfone (DVS)-activated agarose according to Cornillot et al., (1992) and used as immunosorbent. Biot-IgG were allowed to interact with immobilized HBL14 in Tris/CH,COOH pH 7.7, 0.45 M NaCl) containing 0.1 M lactose (Tris-Lac). In these conditions the interaction of HBL14 with saccharidic structures was abolished. 3 ml of immunosorbent were first equilibrated with Tris-Lac and 100-120 mg of Biot-IgG were loaded onto the column at room temperature. The gel was then washed with Tris-Lac and with Tris buffer until no protein could be detected in the effluent. The bound IgG were eluted with PBS (0.05 M Na,HPO,/NaOH pH 11.5, 0.15 M NaCl) at 4°C. Eluted proteins were neutralized to pH 7.7 and stored at 4°C until assayed. Purity of isolated proteins was confirmed by SDS-PAGE.
D. Lutomski et al. /Journal
2.5. Interaction of Fab fragments with HBLlI-immunosorbent
of Neuroimmunology 57 (1995) 9-15
of biotinylated IgG
Fab fragments were prepared by 2% (w/w) papain digestion of the Biot-IgG fraction in presence of 2 mM EDTA and 15 mM cysteine (Raychaudhuri et al., 198.5) for 150 min at 37°C under gentle agitation. Iodoacetamide (0.1 M) was used to terminate digestion. The digest was then dialysed twice against 2 1 of 50 mM Tris buffer for 4 h. The remaining intact Biot-IgG and papain were removed from the mixture using their thiophilic properties (Lutomsky et al., 1994). Briefly, 37 ml (1.8 mg/ml) of the mixture was supplemented with 10% ammonium sulfate and passed through a column containing 6 ml of T-gel (data not shown). The effluent containing both Fab and Fc fragments were then dialysed against PBS (pH 8) and loaded onto an immobilized Protein A column (2.5 X 2 cm) (Pierce, Oud Beijerland, Netherlands). Fab fragments were recovered in the pass-through, and dialysed 4 h against TBS. The separation of the fragments was checked using anti-IgG light chains antibody (Dakopatts, Denmark). and the purity of the Fab fragments was established by SDSPAGE (Lutomsky et al., 1994). Fab fragments (2.25 mg) were loaded onto immobilized HBL14 equilibrated in Tris buffer. The column was carefully washed and the retained Fab fragments were sequentially eluted with Tris-Lac and PBS pH 11.5, respectively. Fab fragments were neutralized as described for IgG.
0.25
,
0.1
_I
0.05
0
/ l/400
l/800
1/l 600
DILUTION
Fig. 1. Binding of antibodies from sera of a healthy individual (0 ) and immunoglobulins from a pool of 80 blood donors (0) to HBL14coated onto microtiter plates. Results are expressed as optical density units at 492 nm COD). Reaction was monitored by HRP-labelled rabbit anti-human IgG.
2.6. Microtiter plate assay The amount of Biot-IgG and Biot-Fab fragments in each eluted fraction from the immunosorbent was determined with streptavidin-peroxidase (Strep-HRP, Sigma) as reported by Caron et al. (1993). 2.7. Specificity of anti-HBL14 antibodies In order to assess their specificity, the biotinylated anti-HBL14 antibodies at various concentrations were allowed to react in triplicate with different antigens immobilized at 5 pug/ml (50 ~1) on polystyrene plates. After incubation and washings, Strep-HRP was added, allowed to incubate 1 h at 37°C and the enzyme activity was measured as described for ELISA.
3. Results 3.1. Screening for antibodies autoreacting towards HBLl4
Individual sera from 14 healthy subjects, and a pool of immunoglobulins (Ig) from 80 blood donors were analysed separately for the presence of anti-HBL14 antibodies, on immobilized HBL14 by ELISA. The
antibody binding was revealed with either anti-p or anti-y conjugates. The results, expressed as OD,,,, are illustrated in Fig. 1 and show that antibodies were detected in all sera. The OD,,, values obtained with 1:400 dilution ranged from 0.08 to 0.21, varying from one donor to another. The values observed with the pool of Ig COD,,, = 0.17) were within this range. Antibodies in all the serum samples reacting towards HBL14 belonged to the IgG class. No significant binding was observed with serum IgM. The presence of 0.1 M lactose in incubation buffer did not modify the interaction. 3.2. Purification of IgG and isolation of antibodies reacting with HBLl4
To purify antibodies reacting towards HBL14, a two-step procedure was used. A pool of serum was made using four sera and was separated into 40-ml aliquots. Firstly, from 40 ml of the pool, 300 mg of Ig were isolated on thiophilic support. The column yielded 7.5 mg of Ig per ml of gel and 65% of the serum Ig were recovered. Laser immunonephelometry enabled to determine that albumin, IgA and IgM were found in the pass-through exclusively while IgG were found in
D. Lutomski et al. /Journal
12
of Neuroimmunology 57 (1995) 9-15
the eluted fraction. The latter were biotinylated as described in Materials and methods. Secondly, 100 mg of Biot-IgG were passed through the immunosorbent, extensively washed, and then eluted with the basic buffer. 0.1% of the loaded IgG were retained after a single passage representing 87 pg protein. To evaluate the effect of L-IgG on the interaction between immobilized HBL14 and auto-antibodies, a fraction of IgG depleted in L-IgG were prepared according to Joubert-Caron et al. (1994a). After biotinylation, the depleted fraction was loaded onto the immunosorbent instead of the total Biot-IgG. A relatively higher amount of anti-HBL14 antibodies (152 pg) was eluted. 3.3. Interaction
of Fab fragments
with HBLl4
im-
munosorbent
Among Fab fragments retained on HBL14, two populations were distinguished. The first population (45%) was eluted with 0.1 M lactose, thus indicating that the carbohydrate recognition region was a potential site of binding for anti-HBL14 Fab fragments. The second population (55%) eluted with PBS (pH 11.5) most probably bound other regions of the brain lectin molecule. 3.4. Evaluation of specificity of anti-HBLl4
antibodies
by enzyme immunoassay
The polyreactivity of anti-HBL14 antibodies was assessed by ELISA. Plates coated with different antigens were incubated with different concentrations of antibodies isolated on HBL14 immunosorbent, and binding to immobilized antigens was measured by using Strep-HRP. Whether with homologous antigen (HBL14) or with GIgG, binding was found to occur to the same extent (Fig. 2). Anti-HBL14 antibodies also recognized various heterologous antigens, although at varying levels. The mean reactivity of anti-HBL14 antibodies with heterologous antigens relative to that with HBL14 and L-IgG ranged between 67 and 172%. For some of the antigens, the binding was weaker, e.g. thyroglobulin, 67%; cytochrome c, 73%, and laminin, 74%; while it was stronger for actin (172%), and about similar for myelin basic protein (MBP) (95%). The difference in autoreactivity against HBL14, as detected by ELISA, existing between the total IgG population and purified anti-HBL14 antibodies is represented in Fig. 3. Affinity-purified anti-HBL14 antibodies bound to HBL14 significantly stronger than the total IgG preparation. 4. Discussion We have previously identified antibodies (L-IgG) mimicking a human brain lectin, in sera of healthy
I
I
8
20
10
I
33
I
I
50
AUTOANTBODYCONCENTRATlON ti/ml)
Fig. 2. Binding of immuno-purified anti-HBL14 antibodies with different antigens measured by solid-phase ELISA. HBL14 (m ) and L-IgG (0) as well as different unrelated antigens, actin (01, cytochrome c (-1, laminin (+I, myoglobin (A ), MBP (I), and thyroglobulin (0) were coated at 5 pg/ml (50 ~1) onto microtiter plates and bound anti-HBL14 antibodies were detected by HRP-labelled rabbit anti-human IgG.
subjects as well as in sera and CSF of patients with neurological disorders (Joubert-Caron et al., 1994a,b). The L-IgG were purified using an anti-HBL14 anti-
_;
0.2
d 0.1
I
10
20 IgG CONCENTRATION
I
33
I
I
I
50 @g/ml)
Fig. 3. Binding of the total IgG preparation (0) and of affinity-isolated anti-HBL14 antibodies (Dl to HBL14 as measured by solidphase ELISA. HBL14-coated plates (5 pg/ml, 50 ~1) were incubated with isolated anti-HBL14 antibodies or with the whole IgG population at different concentrations. Reaction was monitored by HRP-labelled rabbit anti-human IgG.
D. Lutomski et al. /Journal
of Neuroimmunology 57 (1995) 9-1.5
body as an immunosorbent (Joubert-Caron et al., 1994a). Our working hypothesis is that the L-IgG could be specific for anti-HBL14 autoantibodies and would participate in the regulation of expression of antiHBL14 antibodies under physiological conditions. This hypothesis might explain the immunoregulatory properties reported for this family of lectins following their injection into the circulation (Levi et al., 1983; Offner et al., 1990). Our previous findings thus prompted us to investigate the presence of antibodies reacting both with HBL14 and L-IgG. The results obtained in the present study showed that reactivity of IgG, but not of IgM with the 1CkDa human brain lectin, was detected in whole serum from healthy individuals as well as in a pool of Ig from 80 blood donors, suggesting that the anti-HBL14 antibodies belong to natural autoantibodies (NAA) (Avrameas et al., 1981; Avrameas, 1991,1993). The reactivity of serum IgG towards immobilized HBL14 detected by ELISA was not mediated by the lectin properties of HBL14 as the presence of lactose in the incubation buffer did not affect the interaction. This result is consistent with the loss of the carbohydrate-binding activity of HBL14 when it was adsorbed on microtiter plate (Caron et al., 1994). Anti-HBL14 antibodies were isolated from the rest of the serum IgG by immunoaffinity chromatography on immobilized HBL14. To prevent an eventual interaction between the lactosamine structures carried by Fc fragments of IgG (Hamako et al., 1993) and the ligand HBL14, affinity chromatography was performed in the presence of a competitive sugar, and a high molarity of sodium chloride. These conditions efficiently abolished the interaction of HBL14 with glycoproteins such as asialofetuin (data not shown), and were preferred to a deglycosylation which can affect the IgG structure. However, the carbohydrate-binding region of HBL14 which is highly conserved among this family of proteins (Abbott and Feizi, 1991a; Bladier et al., 1991) may be a binding site for autoantibodies, as suggested by the lactosyl-specific interaction of a population of Fab fragments with immobilized HBL14. In this study, it is likely that only antibodies reacting with other regions of the lectin were detected by ELISA and adsorbed on immobilized HBLl4 since lactose was present in the loading buffer of chromatography experiments. In order to examine whether the autoreactivity of IgG to HBL14 was affected by the simultaneous presence of L-IgG, chromatographies on immobilized HBL14 were carried out by using an IgG fraction depleted in L-IgG (Joubert-Caron et al., 1994a). The amount of autoantibodies reacting towards HBL14 in the depleted-L-IgG fraction was higher than in the whole IgG, suggesting that a part of anti-HBL14 antibodies (Abl) may be bound by L-IgG (Ab2) in the total
13
IgG fraction. The binding of L-IgG to anti-HBL14 antibodies was further confirmed by solid phase ELISA. Taken together, these findings argue in favor of our initial hypothesis. Anti-HBL14 antibodies are polyspecific, like most NAA previously described (Avrameas et al., 1981; Guilbert et al., 1983; Matsiota et al., 1988; Abu-Sharda and Shoenfeld, 1993; Bememan et al., 1993; Hurez et al., 1993a,bl. As HBL14 and other members of the 1CkDa lectin family are not glycosylated (Bladier et al., 1991; Harrison, 19911, the polyreactivity cannot be attributed to recognition of common carbohydrate determinants. Anti-HBL14 antibodies display a significant amount of reactivity with actin and to a lesser extent with MBP. An antigenic relationship between 14-kDa lectin and the encephalitogenic domain of MBP that has been reported previously (Abbott et al., 1989) might explain the contribution of the lectin to the prevention of encephalomyelitis (Offner et al., 1990; Abbott and Feizi, 1991b). The cross-reactivity of antibovine heart lectin monoclonal antibodies with MBP has been shown to be related to the recognition of the tetrapeptide sequence WGAE common to MBP and bovine lectin (Abbott et al., 1989). However, HBLl4 contains a different tetrapeptide (WGTE) in the same position, and no cross-reactivity was found between MBP and HBL14 using rabbit polyclonal antibodies to HBL14 (Caron, 1993). Thus, it is unlikely that an actual antigenic cross-reactivity with a well-defined epitope contributes to the reaction between anti-HBL14 NAA and MBP observed in this study. The function of NAA in healthy individuals is not yet clear. Natural antibodies to peripheral nerve myelin may occur without clinical disease-associated neuropathy in healthy people (Vrethem et al., 1991). The distinction between natural and disease-associated autoantibodies is subtle. It is not clear whether the structure, fine specificity and genetics of autoantibodies in healthy individuals differ from those of autoantibodies from patients with autoimmune disease. The development of pathological autoimmunity may thus result from abnormal expansion of natural unmutated, possibly autoantigen-driver clones. Significant modifications in the titers of NAA have been observed in the serum of patients with autoimmune as well as non-autoimmune diseases. For example, titers of NAA against the nerve growth factor were found to be increased in sera of patients with autoimmune diseases (Dicou et al., 1993), and antibodies to Purkinje cells were detected in serum of patients with autoimmune neuropathy (Nemni et al., 1993). The elevated titers of anti-HBL14 antibodies that we observed in several patients with different neurological diseases are intriguing (in preparation). Further investigations are necessary to understand the role of these antibodies in CNS immunological insults.
14
D. Lutomski et al./Journal
of Neuroimmunology 57 (1995) 9-15
In conclusion, the findings presented here indicate that serum from healthy individuals contains polyspecific natural antibodies reacting with a highly conserved 1CkDa lectin from human brain and oligoclonal IgG mimicking this lectin (Joubert-Caron et al., 1994a,b). The contribution of intravenously administered 1CkDa lectin, in protection from myasthenia gravis or encephalomyelitis has been reported previously (Levi et al., 1983; Offner et al., 1990). Taking into account the immunoregulatory role of NAA (Avrameas, 1991; Amital-Teplizki et al., 1993; Kazatchkine et al., 1994), and the immunomodulatory properties of the 16kDa lectin family @.evi et al., 1983; Offner et al., 1990), we suggest that the observed protective effect might be due to a modification of an equilibrium within the idiotypic/anti-idiotypic network.
Acknowledgements This study was supported by grants to the Laboratoire de Biochimie et Technologie des ProtCines from the Direction de la Recherche et des Etudes Doctorales (DRED), from the Delegation a la Recherche Clinique de 1’Assistance Publique, and from Fondation de France. The authors thank M. Buatois for laser immunonephelometry assays, and S.V. Kaveri, INSERM U 28 for useful suggestions.
References Abbott, M.W. and Feizi, T. (1991a) Soluble 14 kDa /I-galactosidespecific bovine lectin. Evidence from mutagenesis and proteolysis that almost the complete polypeptide chain is necessary for integrity of the carbohydrate recognition domain. J. Biol. Chem. 266, 5552-5557. Abbott, M.W. and Feizi, T. (1991b) Antigenic cross-reaction between P-galactoside binding lectin and myelin basic protein. J. Neuroimmunol. 31. 179. Abbott, M.W., Mellor, A., Edwards, Y. and Feizi, T. (1989) Soluble/3-galactoside binding lectin. cDNA cloning reveals the complete amino acid sequence and an antigenic relationship with the major encephalitogenic domain of myelin basic protein. Biochem. J. 259, 283-290. Abu-Sharda, M. and Shoenfeld, Y. (1993) What are natural autoantibodies? In: Y. Schoenfeld and D.A. Isenberg (Eds.), Natural Antibodies. Their Physiological Role and Regulatory Significance. CRC Press, Boca Raton, Ann Arbor, London, Tokyo, pp. 16-35. Amital-Teplizki, H., Lymbery, Y., Tomer, Y. and Schoenfeld, Y. (1993) Natural antibodies: self recognition and physiological autoimmunity. In: Y. Schoenfeld and D.A. Isenberg (Eds.), Natural Antibodies. Their Physiological Role and Regulatory Significance. CRC Press, Boca Raton, Ann Arbor, London, Tokyo, pp. 175-204. Avellana-Adalid, V., Joubert, R., Bladier, D. and Caron, M. (1990) Biotinylated derivative of a human brain lectin: synthesis and use
in affinoblotting for endogenous ligand studies. Anal. Biochem. 190, 26-31. Avrameas, S. (1991) Natural autoantibodies: from ‘horror autotoxicus’ to ‘gnothi seauton’. Immunol. Today 12, 154-159. Avrameas, S. (1993) Natural antibodies: self recognition and physiological autoimmunity. In: Y. Schoenfeld and D.A. Isenberg (Eds.), Natural Antibodies. Their Physiological Role and Regulatory Significance. CRC Press, Boca Raton, Ann Arbor, London, Tokyo, pp. l-14. Avrameas, S., Guilbert, B. and Dighiero, G. (1981) Natural antibodies against tubulin, actin, myosin, myoglobin, thyroglobulin, fetuin, albumin and transferrin are present in normal sera and monoclonal immunoglobulins from multiple myeloma and Waldenstrom’s macroglobulinemia may express similar antibody specificities. Ann. Immunol. (Inst. Pasteur) 132C, 231-236. Berneman, A., Guilbert, B., Eschrich, S. and Avrameas, S. (1993) IgG auto- and polyreactivities of normal human sera. Mol. Immunol. 30, 1499-1510. Bladier, D., Joubert, R., Avellana-Adalid, V., Kemeny, J.L., Doinel, C., Amouroux, J. and Caron, M. (1989) Purification and characterization of a P-galactoside-binding lectin from human brain. Arch. Biochem. Biophys. 269, 433-439. Bladier, D., Le Caer, J.-P., Joubert, R., Caron, M. and Rossier, J. (1991) P-Galactoside soluble lectin from human brain: complete aminoacid sequence. Neurochem. Int. 18, 275-281. Caron, M. (1993) Enzyme immunoassay (EIA) for human endogenous @-galactoside-binding lectin. In: H.-J. Gabius and S. Gabius (Eds.), Lectins and Glycobiology. Springer-Verlag, Berlin, Heidelberg, pp. 191-195. Caron, M., Bladier, D. and Joubert, R. (1990) Soluble galactoside-binding vertebrate lectins: a protein family with common properties. Int. J. Biochem. 22, 1379-1385. Caron, M., Joubert-Caron, R., Cartier, J.-R., Chadli, A. and Bladier, D. (1993) Study of lectin-ganglioside interactions by high-performance liquid affinity chromatography. J. Chromatogr. 646, 327333. Caron, M., Chadli, A., Bladier, D. and Joubert-Caron, R. (1994) Lectins in human brain: some characteristics of the 14 kDa /3-galactoside-binding lectin. In: J. Fischer, TC Bog-Hansen and E. Van Driesche (Eds.), Lectins, Biology, Chemistry, Clinical Chemistry. Textop, Copenhagen, Vol. 10 (in press). Cornillot, J.-D., Caron, M., Joubert-Caron, R. and Bladier, D. (1992) Use of an immobilized human endogenous lectin for the purification of complementary ligands. Int. J. Biochem. 24, 1585-1589. Couraud, P.-O., Casenti-Borocz, D., Bringman, T.S., Griffith, J., McGrogan, M. and Nedwin, G.E. (1989) Molecular cloning, characterization and expression of a human 14 kDa lectin. J. Biol. Chem. 251, 1310-1316. Dicou, E., Hurez, D. and Nerriere, V. (1993) Natural autoantibodies against the nerve growth factor in autoimmune diseases. J. Neuroimmunol. 47, 159-168. Guilbert, B., Dighiero, G. and Avrameas, S. (1983) Naturally occurring antibodies against nine common antigens in human sera. I. Detection, isolation and characterization. J. Immunol. 128, 27792787. Hamako, J., Matsui, T., Ozeki Y., Mizuochi, T. and Titani, K. (1993) Comparative studies of asparagine-linked sugar chains of immunoglobuhn G from eleven mammalian species. Comp. Biochem. Physiol. 106B, 949-954. Harrison, L. (1991) Soluble p-galactoside-binding lectins in vertebrates. In: T.C. Bog-Hansen (Ed.), Lectin Reviews. Sigma Chemical Company, St. Louis, MO, Vol. 1, pp. 17-39. Hurez, V., Dietrich, G., Kaveri, S.V. and Kazatchkine, M.D. (1993a) Polyreactivity is a property of natural and disease-associated human autoantibodies. Stand. J. Immunol. 38, 190-196. Hurez, V., Kaveri, S.V. and Krzatchkine, M.D. (1993b) Expression
D. Lutomski et al. /Journal
of Neuroimmunology 57 (1995) 9-15
and control of the natural autoreactive IgG repertoire in normal serum. Eur. J. immunol. 23, 783-789. Hutchens (1992) Thiophilic adsorption chromatography. In: A. Kenney and S. Fowell (Eds.), Methods in Molecular Biology. Humana Press, Totowa, Vol. 11, pp. 1-15. Joubert-Caron, R., Caron, M., Bochet, P., Chadli, A., Delaporte, P. Schuller, E. and Bladier, D. (1994a) Oligoclonal P-galactoside-binding immunoglobulins antigenically related to 14 kDa lectin in human serum and cerebrospinal fluid: purification and characterization. Int. J. Biochem. 26, 813-823. Joubert-Caron, R., Chadli, A., Lutomski, D., Quinonero, J., Belin, C., Bladier, D. and Caron, M. (1994b) Molecular resemblance between a 14 kDa lectin from human brain (HBL14) and some oligoclonal P-galactoside-binding y immunoglobulins: A finding which might explain the immunoregulatory properties of the 14 kDa lectins. In: J. Fischer, T.C Bog-Hansen and E. Van Driesche (Eds.), Lectins, Biology, Chemistry, Clinical Chemistry. Textop, Copenhagen, Vol. 10 (in press). Kazatchkine, M.D., Dietrich G., Hurez, V., Ronda, N., Bellon, B., Rossi, F. and Kaveri, S.V. (1994) V region-mediated selection of autoreactive repertoires by intravenous immunoglobulins (i.v. Ig). Immunol. Rev. 139, 79-107 Levi, G., Tarrab-Hazdai, R. and Teichberg, V.I. (1983) Prevention and therapy with electrolectin of experimental autoimmune myasthenia gravis in rabbits. Eur. J. Immunol. 13, 500-507. Lihme, A. and Heegaard, P.M.M. (1991) Thiophilic adsorption chromatography: the separation of serum proteins. Anal. Biochem. 192, 64-69. Lutomsky, D., Joubert-Caron, R., Bourin, P., Bladier, D. and Caron,
15
M. (1994) Use of thiophilic adsorption in the purification of biotinylated Fab fragments. J. Chromatogr. (in press). Matsiota, P., Blancher, A., Doyon, B., Guilbert, B., Clanet, M., Kouvelas, E.D. and Avrameas, S. (1988) Comparative study of natural antibodies in the serum and cerebrospinal fluid of normal individuals and patients with multiple sclerosis and other neurological diseases. Ann. Inst. Pasteur/Immunol. 139, 99-108. Nemni, R., Camerlingo, M., Fazio, R., Casto, L., Quattrini, A., Mamoli, D., Lorenzetti, I., Canal, N. and Mamoli, A. (1993) Serum antibodies to Purkinje cells and dorsal root ganglia neurons in sensory neuronopathy without malignancy. Ann. Neurol. 34, 848-854. Offner, M., Celnik, B., Bringman, T.S., Casentini-Borocz, D., Nedwin, G.E. and Vanderbark, A.A. (1990) Recombinant human P-galactoside-binding lectin suppresses clinical and histological signs of experimental autoimmune encephalomyelitis. J. Neuroimmunol. 28, 177-184. Raychaudhuri, G., McCool, D. and Painter, R.H. (1985) Human IgGl and its Fc fragment bind with different affinities to the Fc receptors on the human U 937, HL-60 and ML-1 cell lines. Mol. Immunol. 22, 1009-1019. Teichberg, V.I., Silman, I., Beitsch, D.D. and Resheff, G. (1975) A P-galactoside binding protein from electric organ tissue of Electrophorus electricus. Proc. Natl. Acad. Sci. USA 72, 1383-1387. Vrethem, M., Skogh, T., Berlin, G., Holmgren, H. and Ernerudh, J. (1991) Antibodies to peripheral nerve myelin may occur without clinical neuropathy in healthy persons. J. Neuroimmunol. 32, 219-222.
Journal of Neuroimmunology 57 (1995) 9-15
Purification and characterization of natural antibodies that recognize a human brain lectin Didier Lutomski a, Michel Caron alb,Philippe Bourin ‘, Christine Lefebure b, Dominique Bladier a,b, Raymonde Joubert-Caron a,* a Laboratoire de Biochimie et Technologie des Prot&nes, Universite' Paris-Nord, UFR “SMBH L6onard de Vinci”, 74 rue Marcel Cachin, 93012 Bobigny Cedex, France b Laboratoire Central de Biochimie, H6pitaf Avicenne, Bobigny, France ’ Laboratoire d’lmmunologie Cellulaire, Centre de Transfusion Sanguine des Armies, H6pital Percy. Clamart, France
Received 10 June 1994; revised 1 September 1994; accepted 23 September 1994
Abstract
We have recently identified oligoclonal IgG antibodies that are related to a human brain lectin (HBL14) from serum and cerebrospinal fluid of patients with neurological disorders. They were termed lectin-like IgG (L-IgG) (Joubert-Caron et al., 1994a,b). In this paper, the occurrence of antibodies reactive both towards HBL14 and L-IgG was investigated. Binding of antibodies to HBL14 was demonstrated by solid-phase ELISA and chromatography on immobilized HBL14. Fab fragments of these antibodies were also shown to bind to HBL14. The specificity of the antibodies towards HBL14 was studied using a panel of different antigens. Our data show that individual sera from healthy people as well as a pool of immunoglobulins from 80 blood donors contain an IgG autoreactivity to HBL14, while no IgM autoreactivity was detected. Anti-HBL14 antibodies from sera were purified using affinity chromatography on immobilized HBL14. Affinity chromatography further allowed us to demonstrate that the binding of anti-HBL14 antibodies was mediated through their Fab fragments. A higher amount of anti-HBL14 antibodies was purified using a L-IgG-depleted fraction of sera. The binding of anti-HBL14 antibodies to L-IgG was confirmed by ELISA. Finally, anti-HBL14 antibodies were found to be polyreactive. These results indicate the occurrence of a novel class of natural antibodies reactive towards a human brain lectin and suggest that these antibodies may participate in immunoregulatory mechanisms probably though idiotypic/anti-idiotypic interaction. Keywords:
Natural autoantibodies;
Anti-idiotype
antibodies;
Brain lectin; Immunoregulation
1. Introduction An endogenous lectin from human brain (HBL14), belonging to the well-conserved 1CkDa lectin family showing a specificity for /?-galactoside structures (for review see Caron et al., 1990; Harrison, 19911, has been purified (Bladier et al., 1989) and fully sequenced (Bladier et al., 1991). A recombinant protein (rIML1) (Couraud et al., 1989) identical to HBL14 and the homologous molecule found in electric organ of Torpedo electricus (Teichberg et al., 1975) have been shown to possess immunoregulatory properties suppressing
* Corresponding 48 38 77 77
author. Phone (+33-l)
0165-5728/95/$09.50
48 38 77 54; Fax (+33-l)
0 1995 Elsevier Science B.V. All rights
SSDIO165-5728(94)00152-9
reserved
experimental autoimmune myasthenia gravis (Levi et al., 1983) or experimental autoimmune encephalomyelitis (Offner et al., 1990). However, mechanisms involved in immune responses mediated by 14kDa lectins remain ill-defined. Recently, oligoclonal P-galactoside-binding y-immunoglobulins (IgG), antigenically related to HBL14, have been detected in serum and CSF from patients with neurological disorders, and purified from serum of healthy people using a specific anti-HBL14 antibody as immunosorbent (Joubert-Caron et al., 1994a,b). They were termed lectin-like IgG (L-IgG) and were found to react with cellular antigenic determinants expressed in mouse brain after surgical injury (Joubert-Caron et al., 1994b). On the basis of (i) the reactivity of the L-IgG Fab fragments to a rabbit anti-HBL14 antibody, in-
10
D. Lutomski et al./Journal of Neuroimmunology57 (1995) 9-15
hibitable by HBL14 and (ii) a specificity for P-galactosides close to that of the lectin, it was suggested that L-IgG may be anti-idiotypic antibodies (Ab2) mimicking the human brain lectin (Joubert-Caron et al., 1994a,b). The presence of antibodies (Abl) directed against HBL14 or related molecules, which may have generated L-IgG (Ab2) was thus speculated. We have therefore investigated the presence of antibodies reacting towards HBL14 in normal sera and sera from patients. In the present study, antibodies reacting with HBL14 were identified in the sera of normal individuals by ELISA, and were purified by immunoaffinity chromatography on immobilized HBL14, from a pool of healthy blood donors. The specificity of the purified antibodies was evaluated by using different antigens and their immunoreactivity against L-IgG was studied.
2. Materials and methods 2.1. Serum samples and antigens Sera from healthy blood donors were obtained from the blood bank at the Centre de Transfusion Sanguine des ArmCes (CTSA). HBL14 was prepared according to Avellana-Adalid et al. (1990). L-IgG and L-IgG-depleted fraction were prepared as reported by Joubert-Caron et al. (1994a). Actin (calf muscle), cytochrome c (bovine heart), myelin basic protein (bovine brain) and thyroglobulin (porcine type II) were purchased from Sigma, myoglobin (bovine heart type II) was obtained from Calbiochem (San Diego, CA), and laminin was a gift from J. Botty from Dr. M. Aubery’s laboratory, INSERM U 180 (Paris, France).
M citrate buffer (pH 5.2) and 10 ~1 H,O,). The reaction was stopped with 100 ~1 of 0.5 M H,SO, for 15 min. Absorbance was measured at 492 nm and the results were expressed as optical density COD). To ascertain that the interaction was not due to a lectinsugar recognition, control experiments were performed in presence of 0.1 M lactose in the incubation buffer. 2.3. Human sera and immunoglobulin preparations IgG were purified by thiophilic adsorption (Lihme and Heegaard, 1991; Joubert-Caron et al., 1994a) from a pool of sera from healthy blood donors according to Hutchens (1992). Briefly 40 ml of a pool of human sera were supplemented for 1 h at 4°C with 10% (w/v> solid ammonium sulfate, then centrifuged at 10 000 X g for 10 min at 4°C. The supernatant was passed at a rate of 0.5 ml/min through 40 ml of thiophilic adsorbent (T-gel) (7 cm x 2.5 cm) equilibrated with 50 mM Tris containing 10% (w/v) ammonium sulphate at 15°C. After washing with the same buffer until the OD,,, reached zero, the adsorbed IgG were eluted with 0.1 M NaBO,, 4 H,O/HCl (pH 8.5). The IgG fraction was biotinylated (Biot-IgG) by incubation with NHS-Biotin (Sigma, La Verpillikre, France) at 1 mg/ml in DMSO used at a final ratio of 10 pg/mg of IgG, for 4 h at 20°C. IgG-depleted serum and Biot-IgG were dialysed against TBS (50 mM Tris/CH,COOH pH 7.4, 150 mM NaCl, 0.02% (w/v) NaN,) and stored at 4°C until used. Albumin, IgG, IgA and IgM concentrations were determined in the pass-through and in the eluted fractions, using laser immunonephelometry (BNA, Behring).
2.2. ELISA for anti-HBL14 antibodies
2.4. Isolation of anti-HBLl4 ity chromatography
antibodies by immunoaffin-
Microtiter plates were coated with 100 ~1 of purified HBL14 (Avellana-Adalid et al., 19901, at a concentration of 1 pg/ml in incubation buffer (0.4 M NaCl, 0.1 M K,HPO, /KH,PO, pH 7.4, EDTA 1 mM, 0.01% BSA (w/v)) for 2 h at 37°C and overnight at 4°C. After washing with T-PBS (10 mM K,HPO,/KH,PO, pH 7.4, 0.05% (v/v) Tween 20>, the plates were blocked with 300 ~1 of 3% (w/v> BSA in T-PBS, for 2 h at 37”C, then incubated with 100 ~1 of 1:400 to 1:1600 dilutions of human serum samples in triplicate, for 1 h at 37°C. After washing, positive wells were detected with either rabbit anti-human IgG or anti-IgM conjugated to horse radish peroxidase (antiHRP IgG or anti-HRP IgM) (Dakopatts, Denmark) used at 1:5000, by 1 h incubation at 37°C. Enzymatic activity was revealed using 100 ~1 orthophenylene diamine chloride (OPD) as substrate (30 mg/15 ml of 0.1
HBL14 purified by affinity chromatography was coupled to divinylsulfone (DVS)-activated agarose according to Cornillot et al., (1992) and used as immunosorbent. Biot-IgG were allowed to interact with immobilized HBL14 in Tris/CH,COOH pH 7.7, 0.45 M NaCl) containing 0.1 M lactose (Tris-Lac). In these conditions the interaction of HBL14 with saccharidic structures was abolished. 3 ml of immunosorbent were first equilibrated with Tris-Lac and 100-120 mg of Biot-IgG were loaded onto the column at room temperature. The gel was then washed with Tris-Lac and with Tris buffer until no protein could be detected in the effluent. The bound IgG were eluted with PBS (0.05 M Na,HPO,/NaOH pH 11.5, 0.15 M NaCl) at 4°C. Eluted proteins were neutralized to pH 7.7 and stored at 4°C until assayed. Purity of isolated proteins was confirmed by SDS-PAGE.
D. Lutomski et al. /Journal
2.5. Interaction of Fab fragments with HBLlI-immunosorbent
of Neuroimmunology 57 (1995) 9-15
of biotinylated IgG
Fab fragments were prepared by 2% (w/w) papain digestion of the Biot-IgG fraction in presence of 2 mM EDTA and 15 mM cysteine (Raychaudhuri et al., 198.5) for 150 min at 37°C under gentle agitation. Iodoacetamide (0.1 M) was used to terminate digestion. The digest was then dialysed twice against 2 1 of 50 mM Tris buffer for 4 h. The remaining intact Biot-IgG and papain were removed from the mixture using their thiophilic properties (Lutomsky et al., 1994). Briefly, 37 ml (1.8 mg/ml) of the mixture was supplemented with 10% ammonium sulfate and passed through a column containing 6 ml of T-gel (data not shown). The effluent containing both Fab and Fc fragments were then dialysed against PBS (pH 8) and loaded onto an immobilized Protein A column (2.5 X 2 cm) (Pierce, Oud Beijerland, Netherlands). Fab fragments were recovered in the pass-through, and dialysed 4 h against TBS. The separation of the fragments was checked using anti-IgG light chains antibody (Dakopatts, Denmark). and the purity of the Fab fragments was established by SDSPAGE (Lutomsky et al., 1994). Fab fragments (2.25 mg) were loaded onto immobilized HBL14 equilibrated in Tris buffer. The column was carefully washed and the retained Fab fragments were sequentially eluted with Tris-Lac and PBS pH 11.5, respectively. Fab fragments were neutralized as described for IgG.
0.25
,
0.1
_I
0.05
0
/ l/400
l/800
1/l 600
DILUTION
Fig. 1. Binding of antibodies from sera of a healthy individual (0 ) and immunoglobulins from a pool of 80 blood donors (0) to HBL14coated onto microtiter plates. Results are expressed as optical density units at 492 nm COD). Reaction was monitored by HRP-labelled rabbit anti-human IgG.
2.6. Microtiter plate assay The amount of Biot-IgG and Biot-Fab fragments in each eluted fraction from the immunosorbent was determined with streptavidin-peroxidase (Strep-HRP, Sigma) as reported by Caron et al. (1993). 2.7. Specificity of anti-HBL14 antibodies In order to assess their specificity, the biotinylated anti-HBL14 antibodies at various concentrations were allowed to react in triplicate with different antigens immobilized at 5 pug/ml (50 ~1) on polystyrene plates. After incubation and washings, Strep-HRP was added, allowed to incubate 1 h at 37°C and the enzyme activity was measured as described for ELISA.
3. Results 3.1. Screening for antibodies autoreacting towards HBLl4
Individual sera from 14 healthy subjects, and a pool of immunoglobulins (Ig) from 80 blood donors were analysed separately for the presence of anti-HBL14 antibodies, on immobilized HBL14 by ELISA. The
antibody binding was revealed with either anti-p or anti-y conjugates. The results, expressed as OD,,,, are illustrated in Fig. 1 and show that antibodies were detected in all sera. The OD,,, values obtained with 1:400 dilution ranged from 0.08 to 0.21, varying from one donor to another. The values observed with the pool of Ig COD,,, = 0.17) were within this range. Antibodies in all the serum samples reacting towards HBL14 belonged to the IgG class. No significant binding was observed with serum IgM. The presence of 0.1 M lactose in incubation buffer did not modify the interaction. 3.2. Purification of IgG and isolation of antibodies reacting with HBLl4
To purify antibodies reacting towards HBL14, a two-step procedure was used. A pool of serum was made using four sera and was separated into 40-ml aliquots. Firstly, from 40 ml of the pool, 300 mg of Ig were isolated on thiophilic support. The column yielded 7.5 mg of Ig per ml of gel and 65% of the serum Ig were recovered. Laser immunonephelometry enabled to determine that albumin, IgA and IgM were found in the pass-through exclusively while IgG were found in
D. Lutomski et al. /Journal
12
of Neuroimmunology 57 (1995) 9-15
the eluted fraction. The latter were biotinylated as described in Materials and methods. Secondly, 100 mg of Biot-IgG were passed through the immunosorbent, extensively washed, and then eluted with the basic buffer. 0.1% of the loaded IgG were retained after a single passage representing 87 pg protein. To evaluate the effect of L-IgG on the interaction between immobilized HBL14 and auto-antibodies, a fraction of IgG depleted in L-IgG were prepared according to Joubert-Caron et al. (1994a). After biotinylation, the depleted fraction was loaded onto the immunosorbent instead of the total Biot-IgG. A relatively higher amount of anti-HBL14 antibodies (152 pg) was eluted. 3.3. Interaction
of Fab fragments
with HBLl4
im-
munosorbent
Among Fab fragments retained on HBL14, two populations were distinguished. The first population (45%) was eluted with 0.1 M lactose, thus indicating that the carbohydrate recognition region was a potential site of binding for anti-HBL14 Fab fragments. The second population (55%) eluted with PBS (pH 11.5) most probably bound other regions of the brain lectin molecule. 3.4. Evaluation of specificity of anti-HBLl4
antibodies
by enzyme immunoassay
The polyreactivity of anti-HBL14 antibodies was assessed by ELISA. Plates coated with different antigens were incubated with different concentrations of antibodies isolated on HBL14 immunosorbent, and binding to immobilized antigens was measured by using Strep-HRP. Whether with homologous antigen (HBL14) or with GIgG, binding was found to occur to the same extent (Fig. 2). Anti-HBL14 antibodies also recognized various heterologous antigens, although at varying levels. The mean reactivity of anti-HBL14 antibodies with heterologous antigens relative to that with HBL14 and L-IgG ranged between 67 and 172%. For some of the antigens, the binding was weaker, e.g. thyroglobulin, 67%; cytochrome c, 73%, and laminin, 74%; while it was stronger for actin (172%), and about similar for myelin basic protein (MBP) (95%). The difference in autoreactivity against HBL14, as detected by ELISA, existing between the total IgG population and purified anti-HBL14 antibodies is represented in Fig. 3. Affinity-purified anti-HBL14 antibodies bound to HBL14 significantly stronger than the total IgG preparation. 4. Discussion We have previously identified antibodies (L-IgG) mimicking a human brain lectin, in sera of healthy
I
I
8
20
10
I
33
I
I
50
AUTOANTBODYCONCENTRATlON ti/ml)
Fig. 2. Binding of immuno-purified anti-HBL14 antibodies with different antigens measured by solid-phase ELISA. HBL14 (m ) and L-IgG (0) as well as different unrelated antigens, actin (01, cytochrome c (-1, laminin (+I, myoglobin (A ), MBP (I), and thyroglobulin (0) were coated at 5 pg/ml (50 ~1) onto microtiter plates and bound anti-HBL14 antibodies were detected by HRP-labelled rabbit anti-human IgG.
subjects as well as in sera and CSF of patients with neurological disorders (Joubert-Caron et al., 1994a,b). The L-IgG were purified using an anti-HBL14 anti-
_;
0.2
d 0.1
I
10
20 IgG CONCENTRATION
I
33
I
I
I
50 @g/ml)
Fig. 3. Binding of the total IgG preparation (0) and of affinity-isolated anti-HBL14 antibodies (Dl to HBL14 as measured by solidphase ELISA. HBL14-coated plates (5 pg/ml, 50 ~1) were incubated with isolated anti-HBL14 antibodies or with the whole IgG population at different concentrations. Reaction was monitored by HRP-labelled rabbit anti-human IgG.
D. Lutomski et al. /Journal
of Neuroimmunology 57 (1995) 9-1.5
body as an immunosorbent (Joubert-Caron et al., 1994a). Our working hypothesis is that the L-IgG could be specific for anti-HBL14 autoantibodies and would participate in the regulation of expression of antiHBL14 antibodies under physiological conditions. This hypothesis might explain the immunoregulatory properties reported for this family of lectins following their injection into the circulation (Levi et al., 1983; Offner et al., 1990). Our previous findings thus prompted us to investigate the presence of antibodies reacting both with HBL14 and L-IgG. The results obtained in the present study showed that reactivity of IgG, but not of IgM with the 1CkDa human brain lectin, was detected in whole serum from healthy individuals as well as in a pool of Ig from 80 blood donors, suggesting that the anti-HBL14 antibodies belong to natural autoantibodies (NAA) (Avrameas et al., 1981; Avrameas, 1991,1993). The reactivity of serum IgG towards immobilized HBL14 detected by ELISA was not mediated by the lectin properties of HBL14 as the presence of lactose in the incubation buffer did not affect the interaction. This result is consistent with the loss of the carbohydrate-binding activity of HBL14 when it was adsorbed on microtiter plate (Caron et al., 1994). Anti-HBL14 antibodies were isolated from the rest of the serum IgG by immunoaffinity chromatography on immobilized HBL14. To prevent an eventual interaction between the lactosamine structures carried by Fc fragments of IgG (Hamako et al., 1993) and the ligand HBL14, affinity chromatography was performed in the presence of a competitive sugar, and a high molarity of sodium chloride. These conditions efficiently abolished the interaction of HBL14 with glycoproteins such as asialofetuin (data not shown), and were preferred to a deglycosylation which can affect the IgG structure. However, the carbohydrate-binding region of HBL14 which is highly conserved among this family of proteins (Abbott and Feizi, 1991a; Bladier et al., 1991) may be a binding site for autoantibodies, as suggested by the lactosyl-specific interaction of a population of Fab fragments with immobilized HBL14. In this study, it is likely that only antibodies reacting with other regions of the lectin were detected by ELISA and adsorbed on immobilized HBLl4 since lactose was present in the loading buffer of chromatography experiments. In order to examine whether the autoreactivity of IgG to HBL14 was affected by the simultaneous presence of L-IgG, chromatographies on immobilized HBL14 were carried out by using an IgG fraction depleted in L-IgG (Joubert-Caron et al., 1994a). The amount of autoantibodies reacting towards HBL14 in the depleted-L-IgG fraction was higher than in the whole IgG, suggesting that a part of anti-HBL14 antibodies (Abl) may be bound by L-IgG (Ab2) in the total
13
IgG fraction. The binding of L-IgG to anti-HBL14 antibodies was further confirmed by solid phase ELISA. Taken together, these findings argue in favor of our initial hypothesis. Anti-HBL14 antibodies are polyspecific, like most NAA previously described (Avrameas et al., 1981; Guilbert et al., 1983; Matsiota et al., 1988; Abu-Sharda and Shoenfeld, 1993; Bememan et al., 1993; Hurez et al., 1993a,bl. As HBL14 and other members of the 1CkDa lectin family are not glycosylated (Bladier et al., 1991; Harrison, 19911, the polyreactivity cannot be attributed to recognition of common carbohydrate determinants. Anti-HBL14 antibodies display a significant amount of reactivity with actin and to a lesser extent with MBP. An antigenic relationship between 14-kDa lectin and the encephalitogenic domain of MBP that has been reported previously (Abbott et al., 1989) might explain the contribution of the lectin to the prevention of encephalomyelitis (Offner et al., 1990; Abbott and Feizi, 1991b). The cross-reactivity of antibovine heart lectin monoclonal antibodies with MBP has been shown to be related to the recognition of the tetrapeptide sequence WGAE common to MBP and bovine lectin (Abbott et al., 1989). However, HBLl4 contains a different tetrapeptide (WGTE) in the same position, and no cross-reactivity was found between MBP and HBL14 using rabbit polyclonal antibodies to HBL14 (Caron, 1993). Thus, it is unlikely that an actual antigenic cross-reactivity with a well-defined epitope contributes to the reaction between anti-HBL14 NAA and MBP observed in this study. The function of NAA in healthy individuals is not yet clear. Natural antibodies to peripheral nerve myelin may occur without clinical disease-associated neuropathy in healthy people (Vrethem et al., 1991). The distinction between natural and disease-associated autoantibodies is subtle. It is not clear whether the structure, fine specificity and genetics of autoantibodies in healthy individuals differ from those of autoantibodies from patients with autoimmune disease. The development of pathological autoimmunity may thus result from abnormal expansion of natural unmutated, possibly autoantigen-driver clones. Significant modifications in the titers of NAA have been observed in the serum of patients with autoimmune as well as non-autoimmune diseases. For example, titers of NAA against the nerve growth factor were found to be increased in sera of patients with autoimmune diseases (Dicou et al., 1993), and antibodies to Purkinje cells were detected in serum of patients with autoimmune neuropathy (Nemni et al., 1993). The elevated titers of anti-HBL14 antibodies that we observed in several patients with different neurological diseases are intriguing (in preparation). Further investigations are necessary to understand the role of these antibodies in CNS immunological insults.
14
D. Lutomski et al./Journal
of Neuroimmunology 57 (1995) 9-15
In conclusion, the findings presented here indicate that serum from healthy individuals contains polyspecific natural antibodies reacting with a highly conserved 1CkDa lectin from human brain and oligoclonal IgG mimicking this lectin (Joubert-Caron et al., 1994a,b). The contribution of intravenously administered 1CkDa lectin, in protection from myasthenia gravis or encephalomyelitis has been reported previously (Levi et al., 1983; Offner et al., 1990). Taking into account the immunoregulatory role of NAA (Avrameas, 1991; Amital-Teplizki et al., 1993; Kazatchkine et al., 1994), and the immunomodulatory properties of the 16kDa lectin family @.evi et al., 1983; Offner et al., 1990), we suggest that the observed protective effect might be due to a modification of an equilibrium within the idiotypic/anti-idiotypic network.
Acknowledgements This study was supported by grants to the Laboratoire de Biochimie et Technologie des ProtCines from the Direction de la Recherche et des Etudes Doctorales (DRED), from the Delegation a la Recherche Clinique de 1’Assistance Publique, and from Fondation de France. The authors thank M. Buatois for laser immunonephelometry assays, and S.V. Kaveri, INSERM U 28 for useful suggestions.
References Abbott, M.W. and Feizi, T. (1991a) Soluble 14 kDa /I-galactosidespecific bovine lectin. Evidence from mutagenesis and proteolysis that almost the complete polypeptide chain is necessary for integrity of the carbohydrate recognition domain. J. Biol. Chem. 266, 5552-5557. Abbott, M.W. and Feizi, T. (1991b) Antigenic cross-reaction between P-galactoside binding lectin and myelin basic protein. J. Neuroimmunol. 31. 179. Abbott, M.W., Mellor, A., Edwards, Y. and Feizi, T. (1989) Soluble/3-galactoside binding lectin. cDNA cloning reveals the complete amino acid sequence and an antigenic relationship with the major encephalitogenic domain of myelin basic protein. Biochem. J. 259, 283-290. Abu-Sharda, M. and Shoenfeld, Y. (1993) What are natural autoantibodies? In: Y. Schoenfeld and D.A. Isenberg (Eds.), Natural Antibodies. Their Physiological Role and Regulatory Significance. CRC Press, Boca Raton, Ann Arbor, London, Tokyo, pp. 16-35. Amital-Teplizki, H., Lymbery, Y., Tomer, Y. and Schoenfeld, Y. (1993) Natural antibodies: self recognition and physiological autoimmunity. In: Y. Schoenfeld and D.A. Isenberg (Eds.), Natural Antibodies. Their Physiological Role and Regulatory Significance. CRC Press, Boca Raton, Ann Arbor, London, Tokyo, pp. 175-204. Avellana-Adalid, V., Joubert, R., Bladier, D. and Caron, M. (1990) Biotinylated derivative of a human brain lectin: synthesis and use
in affinoblotting for endogenous ligand studies. Anal. Biochem. 190, 26-31. Avrameas, S. (1991) Natural autoantibodies: from ‘horror autotoxicus’ to ‘gnothi seauton’. Immunol. Today 12, 154-159. Avrameas, S. (1993) Natural antibodies: self recognition and physiological autoimmunity. In: Y. Schoenfeld and D.A. Isenberg (Eds.), Natural Antibodies. Their Physiological Role and Regulatory Significance. CRC Press, Boca Raton, Ann Arbor, London, Tokyo, pp. l-14. Avrameas, S., Guilbert, B. and Dighiero, G. (1981) Natural antibodies against tubulin, actin, myosin, myoglobin, thyroglobulin, fetuin, albumin and transferrin are present in normal sera and monoclonal immunoglobulins from multiple myeloma and Waldenstrom’s macroglobulinemia may express similar antibody specificities. Ann. Immunol. (Inst. Pasteur) 132C, 231-236. Berneman, A., Guilbert, B., Eschrich, S. and Avrameas, S. (1993) IgG auto- and polyreactivities of normal human sera. Mol. Immunol. 30, 1499-1510. Bladier, D., Joubert, R., Avellana-Adalid, V., Kemeny, J.L., Doinel, C., Amouroux, J. and Caron, M. (1989) Purification and characterization of a P-galactoside-binding lectin from human brain. Arch. Biochem. Biophys. 269, 433-439. Bladier, D., Le Caer, J.-P., Joubert, R., Caron, M. and Rossier, J. (1991) P-Galactoside soluble lectin from human brain: complete aminoacid sequence. Neurochem. Int. 18, 275-281. Caron, M. (1993) Enzyme immunoassay (EIA) for human endogenous @-galactoside-binding lectin. In: H.-J. Gabius and S. Gabius (Eds.), Lectins and Glycobiology. Springer-Verlag, Berlin, Heidelberg, pp. 191-195. Caron, M., Bladier, D. and Joubert, R. (1990) Soluble galactoside-binding vertebrate lectins: a protein family with common properties. Int. J. Biochem. 22, 1379-1385. Caron, M., Joubert-Caron, R., Cartier, J.-R., Chadli, A. and Bladier, D. (1993) Study of lectin-ganglioside interactions by high-performance liquid affinity chromatography. J. Chromatogr. 646, 327333. Caron, M., Chadli, A., Bladier, D. and Joubert-Caron, R. (1994) Lectins in human brain: some characteristics of the 14 kDa /3-galactoside-binding lectin. In: J. Fischer, TC Bog-Hansen and E. Van Driesche (Eds.), Lectins, Biology, Chemistry, Clinical Chemistry. Textop, Copenhagen, Vol. 10 (in press). Cornillot, J.-D., Caron, M., Joubert-Caron, R. and Bladier, D. (1992) Use of an immobilized human endogenous lectin for the purification of complementary ligands. Int. J. Biochem. 24, 1585-1589. Couraud, P.-O., Casenti-Borocz, D., Bringman, T.S., Griffith, J., McGrogan, M. and Nedwin, G.E. (1989) Molecular cloning, characterization and expression of a human 14 kDa lectin. J. Biol. Chem. 251, 1310-1316. Dicou, E., Hurez, D. and Nerriere, V. (1993) Natural autoantibodies against the nerve growth factor in autoimmune diseases. J. Neuroimmunol. 47, 159-168. Guilbert, B., Dighiero, G. and Avrameas, S. (1983) Naturally occurring antibodies against nine common antigens in human sera. I. Detection, isolation and characterization. J. Immunol. 128, 27792787. Hamako, J., Matsui, T., Ozeki Y., Mizuochi, T. and Titani, K. (1993) Comparative studies of asparagine-linked sugar chains of immunoglobuhn G from eleven mammalian species. Comp. Biochem. Physiol. 106B, 949-954. Harrison, L. (1991) Soluble p-galactoside-binding lectins in vertebrates. In: T.C. Bog-Hansen (Ed.), Lectin Reviews. Sigma Chemical Company, St. Louis, MO, Vol. 1, pp. 17-39. Hurez, V., Dietrich, G., Kaveri, S.V. and Kazatchkine, M.D. (1993a) Polyreactivity is a property of natural and disease-associated human autoantibodies. Stand. J. Immunol. 38, 190-196. Hurez, V., Kaveri, S.V. and Krzatchkine, M.D. (1993b) Expression
D. Lutomski et al. /Journal
of Neuroimmunology 57 (1995) 9-15
and control of the natural autoreactive IgG repertoire in normal serum. Eur. J. immunol. 23, 783-789. Hutchens (1992) Thiophilic adsorption chromatography. In: A. Kenney and S. Fowell (Eds.), Methods in Molecular Biology. Humana Press, Totowa, Vol. 11, pp. 1-15. Joubert-Caron, R., Caron, M., Bochet, P., Chadli, A., Delaporte, P. Schuller, E. and Bladier, D. (1994a) Oligoclonal P-galactoside-binding immunoglobulins antigenically related to 14 kDa lectin in human serum and cerebrospinal fluid: purification and characterization. Int. J. Biochem. 26, 813-823. Joubert-Caron, R., Chadli, A., Lutomski, D., Quinonero, J., Belin, C., Bladier, D. and Caron, M. (1994b) Molecular resemblance between a 14 kDa lectin from human brain (HBL14) and some oligoclonal P-galactoside-binding y immunoglobulins: A finding which might explain the immunoregulatory properties of the 14 kDa lectins. In: J. Fischer, T.C Bog-Hansen and E. Van Driesche (Eds.), Lectins, Biology, Chemistry, Clinical Chemistry. Textop, Copenhagen, Vol. 10 (in press). Kazatchkine, M.D., Dietrich G., Hurez, V., Ronda, N., Bellon, B., Rossi, F. and Kaveri, S.V. (1994) V region-mediated selection of autoreactive repertoires by intravenous immunoglobulins (i.v. Ig). Immunol. Rev. 139, 79-107 Levi, G., Tarrab-Hazdai, R. and Teichberg, V.I. (1983) Prevention and therapy with electrolectin of experimental autoimmune myasthenia gravis in rabbits. Eur. J. Immunol. 13, 500-507. Lihme, A. and Heegaard, P.M.M. (1991) Thiophilic adsorption chromatography: the separation of serum proteins. Anal. Biochem. 192, 64-69. Lutomsky, D., Joubert-Caron, R., Bourin, P., Bladier, D. and Caron,
15
M. (1994) Use of thiophilic adsorption in the purification of biotinylated Fab fragments. J. Chromatogr. (in press). Matsiota, P., Blancher, A., Doyon, B., Guilbert, B., Clanet, M., Kouvelas, E.D. and Avrameas, S. (1988) Comparative study of natural antibodies in the serum and cerebrospinal fluid of normal individuals and patients with multiple sclerosis and other neurological diseases. Ann. Inst. Pasteur/Immunol. 139, 99-108. Nemni, R., Camerlingo, M., Fazio, R., Casto, L., Quattrini, A., Mamoli, D., Lorenzetti, I., Canal, N. and Mamoli, A. (1993) Serum antibodies to Purkinje cells and dorsal root ganglia neurons in sensory neuronopathy without malignancy. Ann. Neurol. 34, 848-854. Offner, M., Celnik, B., Bringman, T.S., Casentini-Borocz, D., Nedwin, G.E. and Vanderbark, A.A. (1990) Recombinant human P-galactoside-binding lectin suppresses clinical and histological signs of experimental autoimmune encephalomyelitis. J. Neuroimmunol. 28, 177-184. Raychaudhuri, G., McCool, D. and Painter, R.H. (1985) Human IgGl and its Fc fragment bind with different affinities to the Fc receptors on the human U 937, HL-60 and ML-1 cell lines. Mol. Immunol. 22, 1009-1019. Teichberg, V.I., Silman, I., Beitsch, D.D. and Resheff, G. (1975) A P-galactoside binding protein from electric organ tissue of Electrophorus electricus. Proc. Natl. Acad. Sci. USA 72, 1383-1387. Vrethem, M., Skogh, T., Berlin, G., Holmgren, H. and Ernerudh, J. (1991) Antibodies to peripheral nerve myelin may occur without clinical neuropathy in healthy persons. J. Neuroimmunol. 32, 219-222.