Journal of Neuroimmunology, 36 (1992) 209-215 © 1992 Elsevier Science Publishers B.V. All rights reserved 0165-5728/92/$05.00
209
JNI 02119
Determination of the affinity of monoclonal human IgM for myelin-associated glycoprotein and sulfated glucuronic paragloboside J.C. Brouet a X. Mariette a, A. Chevalier a and B. Hauttecoeur b Laboratory of Immunology and Immunopathology, U.108 INSERM, H6pital Saint-Louis, Paris, France, and ~ Laboratoire de Chimie Organique, Institut Pasteur, 75015 Paris, France (Received 8 July 1991) (Revised received 16 September 1991) (Accepted 17 September 1991)
Key words: Monoclonal IgM; Myelin-associated glycoprotein; Sulfated glucuronic paragloboside; Affinity constant
Summary We determined the association constant of eight monoclonal IgM with two of their targets, i.e. the myelin-associated glycoprotein (MAG) or the sulfated glucuronic paragloboside (SGPG). All IgM had a 10- to 100-fold higher affinity for MAG than for SGPG. The affinity of the different IgM for MAG ranged from 1.3 x 10 -6 to 7 x 10 -9 tool/liter. The Scatchard plots for MAG were curvilinear, half of the sites being of high or low affinity. In contrast, the plots were linear in the assay using SGPG. No obvious correlations were found between the fine specificity of these IgM for the glucuronyl sulfate epitope and their affinity, although most IgM with a high affinity reacted exclusively with SGPG derivatives retaining a sulfate group. There was no parallelism between the severity of the neuropathy and the affinity of the IgM for MAG or SGPG.
Introduction Monoclonal IgM reactive with the myelin-associated glycoprotein (MAG), two nerve glycolipids, the sulfated glucuronic paragloboside (SGPG) and its higher homolog (SGLPG) as well as with small molecular weight nerve glycoproteins have been implicated in the pathogenesis of some hu-
Correspondence to: J.C. Brouet, M.D., Laboratory of Immunology and Immunopathology, U.108 INSERM, H6pital Saint-Louis, 1, avenue Claude Vellefaux, 75475 Paris Cedex 10, France.
man peripheral neuropathies (reviewed in Steck et al., 1987). Several pieces of evidence support this possibility. First, deposits of monoclonal IgM are found around myelin sheaths along with terminal components of complement (Monaco et al., 1990). Second, intraneural injection of anti-MAG IgM led to demyelination (Hays et al., 1987; Sergott et al., 1988). Third, passive transfer of anti-MAG IgM in primates triggered some nerve alterations (Dancea et al., 1989). Although most patients suffer from a sensorimotor polyneuropathy, the severity of the disease or the predominance of sensory or motor symptoms is quite variable from patient to patient.
210
These differences in the course of the neurologic disease remain unexplained. It appears unlikely that they are related to the serum level of the monoclonal IgM or to its target on MAG or SGPG since all IgM studied react with a glucuronyl sulfate epitope of the latter molecule; differences in the fine specificity of anti-MAG IgM, however, do exist (Ilyias et al., 1990). We determine here the intrinsic affinity of eight monoclonal IgM for MAG and SGPG and search for correlations with the severity of the neuropathy. Material and methods
Patients and sera The main clinical data and course of the neuropathy were reviewed from the clinical charts of the patients by a physician who was unaware of the laboratory data. The severity of neuropathy was graded according to the following disability scale: 0 = normal; 1 = clinical or electrophysiological signs (or both) of neuropathy without symptoms of neuropathy (subclinical neuropathy); 2 = mild motor or sensory symptoms (or both) with or without mild functional impairment; 3 = moderately disabled by motor and sensory symptoms including ataxia; 4 = requiring assistance in eating, dressing, or using a walking aid; 5 = not ambulatory. Follow-up ranged between 5 and 20 years (mean 10 years). Serum monoclonal IgM were isolated from patients' sera and purified as previously reported (Mihaesco et al., 1989). The concentration of IgM in samples used for affinity constant measurement was determined by a classical sandwich enzyme-linked immunosorbent assay (ELISA) as reported (Grillot-Courvalin et al., 1986). Purification of MAG, SGPG and SGPG derivath, es MAG was purified from human brain white matter according to the procedure described by Quarles and Pasnak (1977). Purified MAG gave a single 100 kDa band by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) analysis. Its concentration was measured by Coomassie blue staining uging albumin as standard.
SGPG was prepared from human sciatic nerve according to Ariga et al. (1987) with slight modifications. The acidic glycolipid was obtained by elution from DEAE-Sephadex A25 (Pharmacia) column with 0.5 M potassium acetate in methanol and desalting by extensive dialysis against water at 4°C; lipids contaminants were removed by elution from a silica gel 60 (Merck, Darmstadt, Germany) column with c h l o r o f o r m / m e t h a n o l 4: 1. SGPG was eluted with c h l o r o f o r m / m e t h a n o l / water 55 : 45 : 10 as a single component as shown by high performance thin-layer chromatography (HPTLC). The 3-sulfomethyl-glucuronyl paragloboside (SMeGPG) was prepared from SGPG according to the method of Handa and Nakamura (1984) with addition of methyl iodide to the acidic glycolipid in dimethyl sulfoxide at room temperature. S M e G P G was eluted with 0.08 M ammonium acetate in methanol from a DEAE-Sephadex chromatography column followed by SepPak C18 Cartridge (Waters Associates) chromatography to remove salts. The methyl ester of glucuronyl parag[oboside (MeGPG) was prepared by mild acid methanolysis (0.05 HC1 in methanol for 5 h at room temperature) of S M e G P G and purified by DEAE-Sephadex column chromatography. The glucuronyl paragloboside (GPG) was obtained by saponification (0.2 M N a O H in methanol 4 h at 37°C) of MeGPG. The salts were removed by chromatography on Sep-Pak C18 Cartridge. The 3-sulfoglucosyl paragloboside (SGlcPG) was prepared from S M e G P G by addition of sodium borohydride (10 m g / m l ) in methanol at room temperature overnight. The excess of sodium borohydride was destroyed by addition of acetic acid (2 M) and the mixture was applied on Sep-Pak C18 Cartridge to obtain the reduced derivative (Chou et al., 1986). Purity of SGPG derivatives was checked in each case by thin-layer chromatography as described (Hauttecoeur et al., 1990).
Determination of the affinity constant The determination of the dissociating constant ( K D) of antigen-antibody equilibria in solution was done according to Friguet et al. (1985). In
211
this assay, the proportion of antibody which remains unsaturated at equilibrium is measured by an indirect ELISA. A n t i - M A G binding by E L I S A was determined as previously reported (Shy et al., 1984), with some modifications; in brief, glutaraldehyde precoated plates were incubated with 1 /xg of M A G and saturated with L-lysine. For S G P G reactivity, an E L I S A was performed exactly as described by McGinnis et al. (1988), wells were coated with 20 ng SGPG. In both assays, monoclonal IgM were added to the wells for 1 h at room temperature followed by six washes in Tween 0.05% phosphate-buffered saline (PBS); /3-galactosidase conjugated goat anti-human /xchain antibodies were added for 3 h at 37°C. After six washings, the reaction was revealed by addition of substrate for 1 h at 37°C. For the affinity measurement, various concentrations of M A G (1 × 10 -9 to 5 × 10 -6 M) or S G P G (5 × 10 - 9 t o 5 >( 10 - 6 M) in 0.050 ml PBS-Tween were mixed with a constant amount of monoclonal IgM in 0.050 ml PBS-Tween for 18 h at room temperature in microtiter plates precoated with 3% bovine serum albumin (BSA). The concentration used for each IgM in both M A G and S G P G ELISAs was deduced from a preliminary
25
E L I S A calibration to ensure that the absorbance was linear in the range of concentrations explored. According to these results, the concentration of IgM ranged from 2.2 × 10 - 7 to 4.4 X 10 - l ° M. After the incubation period, the concentration of free antibody was measured by the indirect E L I S A described above.
Results Dissociation constant o f monoclonal IgM Representative Scatchard plots for different IgM in the M A G or S G P G assays are shown in Figs. 1 and 2. In the former assay, the IgM featured high and low affinity binding sites. The value for each K D is shown in Table 1. For technical reasons, the curves were not reliable in three instances at low concentration of antigen and only low affinity K o values could be determined. The reasons why these IgM exhibited two kinds of binding sites are manyfold and will be discussed in the next section; however, it should be noted that the IgM preparations we used were homogenous since in seven cases sequencing of the heavy and light
25-
25-
7
7 =[
? o
o
~2
o
©
"-2
>
~.5
1.5
1
0.5
0.5
0,5
~ i
05
1
V
05
1
V
015
"'L"-
"IB-
r~ 1
Fig. 1. Scatchard plots of the binding of MAG to IgM of patient R.O.G. (A), IgM of patient B.L.A. (B) and IgM of patient F.U.E. (C). The free antibody concentration at equilibrium was measured by ELISA. u is the fraction of bound antibody and a the concentration of free antigen at equilibrium.
212 TABLE 1
TABLE 2
DISSOCIATION CONSTANT (MOL/LITER) O F E I G H T M O N O C L O N A L IgM lgM
Clinical
H i g h K D ~'
FOR
MAG
Low K D b
score R.O.G. L.O.C. B.L.A.
4 3 3
7×10 7×10 6x10
~ s s
3 ×10 2.6×10 3 ×10
7 7 7
A.Z.I. F.U.E.
4 4
5 x l l l -~ 4×10 s
3 2
×10 ×10
7 7
P.E.C.
2
_ c)
2
×l(I
Y.S.E.
6
5
_ c)
1.5× 10 ~
D.E.P.
5
_ c)
1.3 × 10 ~'
" K D value at low antigen concentrations. i, KD value at high antigen concentrations. Not determined (see text).
chains was performed which revealed a single N-terminal amino acid sequence (Brouet et al., 1989). In contrast, in the SGPG assay, Scatchard plots were linear, all sites having a similar affinity for the glycolipid. K D values are given in Table 2. In three cases, the K D value was too low to be determined with confidence within the technical constraints of the assay.
\
7 i
5-
DISSOCIATION CONSTANT (MOL/LITER) OF F I V E M O N O C L O N A L IgM
lgM
KD
R.O.G.
6 . 6 x 10 s
L.O.C, B.L.A.
3 × 10 t' 2 . 7 × 10 -~'
F.U.E, Y.S.E.
1.2× 10 ~ 1.7× 10 4
FOR SGPG
Fine specificity of monoclonal IgM The reactivity of the IgM for SGPG and some of its chemical derivatives was determined as suggested by Ilyias et al. (1990) in order to see whether there were some correlations between the affinity of the IgM and its fine specificity for distinct epitopes of the glucuronyl sulfate determinant. Results are shown in Table 3. All IgM bound best to the SGPG and gave various patterns of reactivity with the four derivatives studied.
Clinical correlates There were no obvious correlations between the severity of the neuropathy and the affinity of the monoclonal IgM for M A G or SGPG (see
7
~
42 \
\x
\
Fig. 2, Scatchard plots of the binding of SGPG to |gM of patient R . O . G , ( A ) a n d I g M of patient B . L A , (B). The free antibody concentration at equilibrium was measured by E L I S A . u a n d u/a as in Fig. 1.
213 TABLE 3 REACTIVITY OF M O N O C L O N A L IgM WITH SGPG AND ITS CHEMICAL DERIVATIVES IgM
SGPG
GPG
SMeGPG
SGIcGPG
MeGPG
R.O.G. B.L.A. A.Z.I. P.E.C. D.E.P. F.U.E. L.O.C. Y.S.E.
+++ +++ +++ +++ +++ +++ +++ +++
(+) +
++ ++ + + + + ++ ++
+ (+) (+) + (+) (+) (+)
(+) -
+ + + Maximal binding (100%). + + Optical density between 50 and 80% binding to SGPG. + Optical density between 30 and 50% binding to SGPG. ( + ) Optical density between 10 and 30% binding to SGPG. - Optical density less than 10% maximal binding.
Table 1). In particular the two IgM with the lowest affinities were isolated from patients with highly disabling neuropathy. However, one of these two patients was considerably improved by short courses of plasmapheresis over a 10-year period. Patients R.O.G., A.Z.I. and F.U.E. partially improved initially under plasmapheresis and chemotherapy but further progression of the neuropathy was not influenced by any treatment. Patients L.O.C., B.L.A., D.E.P. and P.E.C. were not or slightly responsive to various therapeutic regimens.
Discussion The possible involvement of monoclonal IgM in the pathogenesis of human peripheral neuropathies stimulated during recent years a number of studies aiming to characterize their fine antibody activity to M A G and glycolipids (Ilyias et al., 1990), their antigenic specificities as well as their structural features (Brouet et al., 1989). In contrast to other monoclonal IgM having a defined antibody activity, such as rheumatoid factors and cold agglutinins, anti-MAG lgM express mainly private idiotopes (Saito et al., 1983; Kahn, 1985; Schmitt et al., 1987; Brouet et al., 1989) and seem to use pairing of different variable regions of their heavy and light chains to build their combining sites (Brouet et al., 1989). The latter
finding was rather unexpected in view of the narrow specificity of these IgM for epitopes carried by the glucuronyl sulfate determinant of SGPG. It may indicate that the human variable gene repertoire for this determinant is rather large. Finally, present structural data on light chains from anti-MAG IgM support the possibility that the expansion of anti-MAG synthesizing B cells may be antigen driven (Mihaesco et al., 1989). In view of the latter possibility and taking into account the natural course of the neuropathy which is quite variable from patient to patient, we wished to determine the affinity constant of these monoclonal IgM for both M A G and SGPG. All IgM had a 10- to 100-fold higher affinity for M A G than for SGPG. This holds true for IgM having either a relatively high or low affinity for MAG. Similarly, the IgM having the highest affinity for M A G also featured the best affinity for SGPG. Therefore the determinant recognized by these IgM was either more accessible or repetitive on M A G than on SGPG. Of note, whereas the Scatchard plots for S G P G were linear with homogenous binding sites, the same analysis on M A G yielded curvilinear plots with two types of binding sites with relatively low or high affinity. Whereas it is difficult to rule out trivial explanations such as aggregation of M A G molecules at high concentrations leading to decreased receptor binding, the possibility of different functional sites in a single IgM molecule is not without precedent. For instance, monoclonal IgM with anti-IgG antibody activity could bind only five IgG molecules which behave as a monovalent antigen (Stone and Metzger, 1968a). Therefore it seems likely that when one site of the IgM subunit is saturated, the effective binding constant of the second one may be considerably smaller than the intrinsic affinity constant (Stone and Metzger, 1968b). This hypothesis fits well with our data which showed that half of the sites were either of high or low affinity. Although structural data on anti-MAG IgM are still scarce, we should like to outline that two IgM (F.U.E. and D.E.P.) with VK~v light chains exhibiting a high n u m b e r of amino acid substitutions (Mihaesco et al., 1989) had either a relatively high (4 x 10 -8 for IgM F.U.E.) or low (_< 5 x 10 -6 for IgM D.E.P.) affinity constant for
214
MAG. Therefore, at least in these cases, the rate of mutation does not seem to result in an increased affinity. Further studies should more thoroughly address this point which is of critical importance to gain insight into the pathogenesis of the disease. Taken together, the affinity of different IgM for MAG ranged from 1 . 3 × 1 0 ~ to 7 × 1 0 '~ mol/liter and for SGPG from 1.7 x 10 -4 to 6.6 x 10 s tool/liter. These values are higher than that reported for monoclonal IgM with anti-IgG specificity (Stone and Metzger, 1968b). Most of them are obviously lower than that of post-immunization antibodies. The affinity of these IgM is in the same order of magnitude as that of the so-called natural polyreactive antibodies found in normal sera; however, these IgM to MAG or SGPG are devoid of notable polyreactivity (Aissa-Feunira et al., 1991). We do not find obvious correlations between the affinity of these IgM and their fine antibody activity. Definitive conclusions must, however, await the study of a larger group of patients in view of the diversity of the epitopes carried by the glucuronyl sulfate determinant. Likewise, there were no correlations between the severity of the neuropathy as assessed by a disability scale and the affinity of the various IgM. However, one must take into account that in this small series of patients some selection bias may exist (for instance purification of the IgM implied that the patients were under plasmapheresis) and that the time elapsed between the onset of symptoms and diagnosis or treatment differed widely from patient to patient. In addition, the possibility exists that other molecules such as nerve low molecular weight glycoproteins, which we have not studied here, may be involved in the pathogenesis of the neuropathy.
References Aissa-Feunira, F.B., Benamar-Elgaaied, A., Bouguerra, A., Schmitt, C., Brouet, J.C. and Dellagi, K. (1991) A human monoclonal IgG reactive with a private idiotype of a monoclonal IgM with autoantibody activity against myelin associated glycoprotein. Eur. J. Immunol. 4, 1065-1068. Ariga, T., Kohriyama, T., Freddo, L., Latov, N., Saito, M., Kon, K., Ando, S., Suzuki, M., Hemling, M.E., Rinehart,
Jr., K.L., Kusunoki, S. and Yu, R.K. (1987) Characterization of sulfated glucuronic acid containing glycolipids rcacting with lgM M-proteins in patients with neuropathy. J. Biol. Chem. 262, 848 853. Brouet, J.C., Dellagi, K., Gendron, M.C., Chevalier, A., Schmin, C. and Mihaesco, E. (1989) Expression of a public idiotype by human monoclonal IgM directed to myelin-associated glycoprotein and characterization of the variability subgroup of their heavy and light chains. J. Exp. Med. 170, 1551-1558. Chou, D.K.H., Ilyas, A.A., Evans, J.E., Costello, C., Quarles, R.H. and Jungalwala, F.B. (1986) Structure of sulfated glueuronyl glycolipids in the nervous system reacting with HNK-I antibody and some IgM paraproteins in neuropathy. J. Biol. Chem. 261, 11717-11725. Dancea, S., Dellagi, K., Renaud, F., Mahouy, C., Hauw, J.J. and Brouet, J.C. (1989) Effect of passive transfer of human anti-myelin-associated glycoprotein IgM in marmoset. Autoimmunity 3, 29-37. Friguet, B., Chaffotte, A.F., Djavadi-Ohaniance, L. and Goldberg, M.E. (1985) Measurements of the true affinity constant in solution of antigen-antibody complexes by enzyme-linked immunosorbent assay. J. Immunol. Methods 77, 305-319. Grillot-Courvalin, C., Labaume, S. and Brouet, J.C. (1986) Differentiation by interleukin 2 of a subpopulation of human B cells. Scand. J. Immunol. 23, 679-684. Handa, S. and Nakamura, K. (1984) Modification of sialic acid carboxyl group of ganglioside. J. Biochem. 95. 1323-1329. Hauttecceur, B., Schmitt, C., Dubois, C., Danon, F. and Brouet, J.C. (1990) Reactivity of human monoclonal IgM with nerve glycosphingolipids. Clin. Exp. Immunol. 80, 181 185. Hays, A.P., Latov, N., Takatsu, M. and Sherman, W.H. (1987) Experimental demyelination of nerve induced by serum of patients with neuropathy and an anti-MAG lgM M-protein. Neurology 37, 242-256. Ilyias, A.A., Chou, D.K.H., Jungalwala, F.B., Costello, C. and Quarles, R.H. (1990) Variability in the structural requirements for binding of human monoclonal anti-myelin-associated glycoprotein immunoglobulin M antibodies and HNK-I to sphingoglycolipid antigens. J. Neurochem. 55, 594-601. Kahn, S.N. (t985) Human IgM monoclonal antibodies with restricted antigenic specificity for myelin express unrelated idiotypes. J. Neurol. Sci. 69, 161-170. McGinnis, S., Kohriyama, T., Yu, R.K., Pesce, M.A. and Latov, N. (1988) Antibodies to sulfated glucuronic acid containing glycosphingolipids in neuropathy associated with anti-MAG antibodies and in normal subjects. J. Neuroimmunol. 17, 119-126. Mihaesco, E., Ayadi, H., Congy, N., Gendron, M.C., Roy, J.P., Heyermann, H., Frangione, B. and Brouet, J.C. (1989) Multiple mutations in the variable region of the K light chains of three monoelonal human IgM with anti-myelinassociated glycoprotein activity. J. Biol. Chem. 264, 2148121485. Monaco, S., Bonetti, B., Ferrari, S., Moretto, G., Nardelli, E., Tedesco, F., Mollnes, T.E., Nobile-Orazio, E., Manfredini,
215 E., Bonazzi, L. and Rizzuto, N. (1990) Complement-mediated demyelination in patients with IgM monoclonal gammopathy and polyneuropathy. New Engl. J. Med. 322, 649-652. Quarles, R.H. and Pasnak, C.F. (1977) A rapid procedure for selectively isolating the major glycoprotein from purified rat brain myelin. Biochem. J. 163, 635-637. Saito, T., Sherman, W.H. and Latov, N. (1983) Specificity of idiotypes of M-proteins that react with MAG in patients with neuropathy. J. Immunol. 130, 2496-2498. Schmitt, C., Dellagi, K., Mihaesco, E. and Brouet, J.C. (1987) Detection of cross-reactive determinants shared by human monoclonal IgM reacting with myelin-associated glycoprotein. J. Immunol. 138, 1442-1446. Sergott, R.C., Brown, M.J., Lisak, R.P. and Miller, S.L. (1988) Antibody to myelin-associated glycoprotein produces central nervous system demyelination. Neurology 38, 422-426.
Shy, M.E., Vietorisz, T., Nobile-Orazio, E. and Latov, N. (1984) Specificity of human IgM M-proteins that bind to myelin-associated glycoprotein: peptide mapping, deglycosylation, and competitive binding studies. J. Immunol. 133, 2509-2512. Steck, A.J., Murray, N., Dellagi, K., Brouet, J.C. and Seligmann, M. (1987) Peripheral neuropathy associated with monoclonal IgM autoantibody. Ann. Neurol. 22, 764-767. Stone, M.J. and Metzger, H. (1968a) Study of macromolecular interactions by equilibrium molecular sieving. J. Biol. Chem. 243, 5049-5055. Stone, M.J. and Metzger, H. (1968b) Binding properties of a Waldenstr6m macroglobulin antibody. J. Biol. Chem. 243, 5977-5984.
209
JNI 02119
Determination of the affinity of monoclonal human IgM for myelin-associated glycoprotein and sulfated glucuronic paragloboside J.C. Brouet a X. Mariette a, A. Chevalier a and B. Hauttecoeur b Laboratory of Immunology and Immunopathology, U.108 INSERM, H6pital Saint-Louis, Paris, France, and ~ Laboratoire de Chimie Organique, Institut Pasteur, 75015 Paris, France (Received 8 July 1991) (Revised received 16 September 1991) (Accepted 17 September 1991)
Key words: Monoclonal IgM; Myelin-associated glycoprotein; Sulfated glucuronic paragloboside; Affinity constant
Summary We determined the association constant of eight monoclonal IgM with two of their targets, i.e. the myelin-associated glycoprotein (MAG) or the sulfated glucuronic paragloboside (SGPG). All IgM had a 10- to 100-fold higher affinity for MAG than for SGPG. The affinity of the different IgM for MAG ranged from 1.3 x 10 -6 to 7 x 10 -9 tool/liter. The Scatchard plots for MAG were curvilinear, half of the sites being of high or low affinity. In contrast, the plots were linear in the assay using SGPG. No obvious correlations were found between the fine specificity of these IgM for the glucuronyl sulfate epitope and their affinity, although most IgM with a high affinity reacted exclusively with SGPG derivatives retaining a sulfate group. There was no parallelism between the severity of the neuropathy and the affinity of the IgM for MAG or SGPG.
Introduction Monoclonal IgM reactive with the myelin-associated glycoprotein (MAG), two nerve glycolipids, the sulfated glucuronic paragloboside (SGPG) and its higher homolog (SGLPG) as well as with small molecular weight nerve glycoproteins have been implicated in the pathogenesis of some hu-
Correspondence to: J.C. Brouet, M.D., Laboratory of Immunology and Immunopathology, U.108 INSERM, H6pital Saint-Louis, 1, avenue Claude Vellefaux, 75475 Paris Cedex 10, France.
man peripheral neuropathies (reviewed in Steck et al., 1987). Several pieces of evidence support this possibility. First, deposits of monoclonal IgM are found around myelin sheaths along with terminal components of complement (Monaco et al., 1990). Second, intraneural injection of anti-MAG IgM led to demyelination (Hays et al., 1987; Sergott et al., 1988). Third, passive transfer of anti-MAG IgM in primates triggered some nerve alterations (Dancea et al., 1989). Although most patients suffer from a sensorimotor polyneuropathy, the severity of the disease or the predominance of sensory or motor symptoms is quite variable from patient to patient.
210
These differences in the course of the neurologic disease remain unexplained. It appears unlikely that they are related to the serum level of the monoclonal IgM or to its target on MAG or SGPG since all IgM studied react with a glucuronyl sulfate epitope of the latter molecule; differences in the fine specificity of anti-MAG IgM, however, do exist (Ilyias et al., 1990). We determine here the intrinsic affinity of eight monoclonal IgM for MAG and SGPG and search for correlations with the severity of the neuropathy. Material and methods
Patients and sera The main clinical data and course of the neuropathy were reviewed from the clinical charts of the patients by a physician who was unaware of the laboratory data. The severity of neuropathy was graded according to the following disability scale: 0 = normal; 1 = clinical or electrophysiological signs (or both) of neuropathy without symptoms of neuropathy (subclinical neuropathy); 2 = mild motor or sensory symptoms (or both) with or without mild functional impairment; 3 = moderately disabled by motor and sensory symptoms including ataxia; 4 = requiring assistance in eating, dressing, or using a walking aid; 5 = not ambulatory. Follow-up ranged between 5 and 20 years (mean 10 years). Serum monoclonal IgM were isolated from patients' sera and purified as previously reported (Mihaesco et al., 1989). The concentration of IgM in samples used for affinity constant measurement was determined by a classical sandwich enzyme-linked immunosorbent assay (ELISA) as reported (Grillot-Courvalin et al., 1986). Purification of MAG, SGPG and SGPG derivath, es MAG was purified from human brain white matter according to the procedure described by Quarles and Pasnak (1977). Purified MAG gave a single 100 kDa band by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) analysis. Its concentration was measured by Coomassie blue staining uging albumin as standard.
SGPG was prepared from human sciatic nerve according to Ariga et al. (1987) with slight modifications. The acidic glycolipid was obtained by elution from DEAE-Sephadex A25 (Pharmacia) column with 0.5 M potassium acetate in methanol and desalting by extensive dialysis against water at 4°C; lipids contaminants were removed by elution from a silica gel 60 (Merck, Darmstadt, Germany) column with c h l o r o f o r m / m e t h a n o l 4: 1. SGPG was eluted with c h l o r o f o r m / m e t h a n o l / water 55 : 45 : 10 as a single component as shown by high performance thin-layer chromatography (HPTLC). The 3-sulfomethyl-glucuronyl paragloboside (SMeGPG) was prepared from SGPG according to the method of Handa and Nakamura (1984) with addition of methyl iodide to the acidic glycolipid in dimethyl sulfoxide at room temperature. S M e G P G was eluted with 0.08 M ammonium acetate in methanol from a DEAE-Sephadex chromatography column followed by SepPak C18 Cartridge (Waters Associates) chromatography to remove salts. The methyl ester of glucuronyl parag[oboside (MeGPG) was prepared by mild acid methanolysis (0.05 HC1 in methanol for 5 h at room temperature) of S M e G P G and purified by DEAE-Sephadex column chromatography. The glucuronyl paragloboside (GPG) was obtained by saponification (0.2 M N a O H in methanol 4 h at 37°C) of MeGPG. The salts were removed by chromatography on Sep-Pak C18 Cartridge. The 3-sulfoglucosyl paragloboside (SGlcPG) was prepared from S M e G P G by addition of sodium borohydride (10 m g / m l ) in methanol at room temperature overnight. The excess of sodium borohydride was destroyed by addition of acetic acid (2 M) and the mixture was applied on Sep-Pak C18 Cartridge to obtain the reduced derivative (Chou et al., 1986). Purity of SGPG derivatives was checked in each case by thin-layer chromatography as described (Hauttecoeur et al., 1990).
Determination of the affinity constant The determination of the dissociating constant ( K D) of antigen-antibody equilibria in solution was done according to Friguet et al. (1985). In
211
this assay, the proportion of antibody which remains unsaturated at equilibrium is measured by an indirect ELISA. A n t i - M A G binding by E L I S A was determined as previously reported (Shy et al., 1984), with some modifications; in brief, glutaraldehyde precoated plates were incubated with 1 /xg of M A G and saturated with L-lysine. For S G P G reactivity, an E L I S A was performed exactly as described by McGinnis et al. (1988), wells were coated with 20 ng SGPG. In both assays, monoclonal IgM were added to the wells for 1 h at room temperature followed by six washes in Tween 0.05% phosphate-buffered saline (PBS); /3-galactosidase conjugated goat anti-human /xchain antibodies were added for 3 h at 37°C. After six washings, the reaction was revealed by addition of substrate for 1 h at 37°C. For the affinity measurement, various concentrations of M A G (1 × 10 -9 to 5 × 10 -6 M) or S G P G (5 × 10 - 9 t o 5 >( 10 - 6 M) in 0.050 ml PBS-Tween were mixed with a constant amount of monoclonal IgM in 0.050 ml PBS-Tween for 18 h at room temperature in microtiter plates precoated with 3% bovine serum albumin (BSA). The concentration used for each IgM in both M A G and S G P G ELISAs was deduced from a preliminary
25
E L I S A calibration to ensure that the absorbance was linear in the range of concentrations explored. According to these results, the concentration of IgM ranged from 2.2 × 10 - 7 to 4.4 X 10 - l ° M. After the incubation period, the concentration of free antibody was measured by the indirect E L I S A described above.
Results Dissociation constant o f monoclonal IgM Representative Scatchard plots for different IgM in the M A G or S G P G assays are shown in Figs. 1 and 2. In the former assay, the IgM featured high and low affinity binding sites. The value for each K D is shown in Table 1. For technical reasons, the curves were not reliable in three instances at low concentration of antigen and only low affinity K o values could be determined. The reasons why these IgM exhibited two kinds of binding sites are manyfold and will be discussed in the next section; however, it should be noted that the IgM preparations we used were homogenous since in seven cases sequencing of the heavy and light
25-
25-
7
7 =[
? o
o
~2
o
©
"-2
>
~.5
1.5
1
0.5
0.5
0,5
~ i
05
1
V
05
1
V
015
"'L"-
"IB-
r~ 1
Fig. 1. Scatchard plots of the binding of MAG to IgM of patient R.O.G. (A), IgM of patient B.L.A. (B) and IgM of patient F.U.E. (C). The free antibody concentration at equilibrium was measured by ELISA. u is the fraction of bound antibody and a the concentration of free antigen at equilibrium.
212 TABLE 1
TABLE 2
DISSOCIATION CONSTANT (MOL/LITER) O F E I G H T M O N O C L O N A L IgM lgM
Clinical
H i g h K D ~'
FOR
MAG
Low K D b
score R.O.G. L.O.C. B.L.A.
4 3 3
7×10 7×10 6x10
~ s s
3 ×10 2.6×10 3 ×10
7 7 7
A.Z.I. F.U.E.
4 4
5 x l l l -~ 4×10 s
3 2
×10 ×10
7 7
P.E.C.
2
_ c)
2
×l(I
Y.S.E.
6
5
_ c)
1.5× 10 ~
D.E.P.
5
_ c)
1.3 × 10 ~'
" K D value at low antigen concentrations. i, KD value at high antigen concentrations. Not determined (see text).
chains was performed which revealed a single N-terminal amino acid sequence (Brouet et al., 1989). In contrast, in the SGPG assay, Scatchard plots were linear, all sites having a similar affinity for the glycolipid. K D values are given in Table 2. In three cases, the K D value was too low to be determined with confidence within the technical constraints of the assay.
\
7 i
5-
DISSOCIATION CONSTANT (MOL/LITER) OF F I V E M O N O C L O N A L IgM
lgM
KD
R.O.G.
6 . 6 x 10 s
L.O.C, B.L.A.
3 × 10 t' 2 . 7 × 10 -~'
F.U.E, Y.S.E.
1.2× 10 ~ 1.7× 10 4
FOR SGPG
Fine specificity of monoclonal IgM The reactivity of the IgM for SGPG and some of its chemical derivatives was determined as suggested by Ilyias et al. (1990) in order to see whether there were some correlations between the affinity of the IgM and its fine specificity for distinct epitopes of the glucuronyl sulfate determinant. Results are shown in Table 3. All IgM bound best to the SGPG and gave various patterns of reactivity with the four derivatives studied.
Clinical correlates There were no obvious correlations between the severity of the neuropathy and the affinity of the monoclonal IgM for M A G or SGPG (see
7
~
42 \
\x
\
Fig. 2, Scatchard plots of the binding of SGPG to |gM of patient R . O . G , ( A ) a n d I g M of patient B . L A , (B). The free antibody concentration at equilibrium was measured by E L I S A . u a n d u/a as in Fig. 1.
213 TABLE 3 REACTIVITY OF M O N O C L O N A L IgM WITH SGPG AND ITS CHEMICAL DERIVATIVES IgM
SGPG
GPG
SMeGPG
SGIcGPG
MeGPG
R.O.G. B.L.A. A.Z.I. P.E.C. D.E.P. F.U.E. L.O.C. Y.S.E.
+++ +++ +++ +++ +++ +++ +++ +++
(+) +
++ ++ + + + + ++ ++
+ (+) (+) + (+) (+) (+)
(+) -
+ + + Maximal binding (100%). + + Optical density between 50 and 80% binding to SGPG. + Optical density between 30 and 50% binding to SGPG. ( + ) Optical density between 10 and 30% binding to SGPG. - Optical density less than 10% maximal binding.
Table 1). In particular the two IgM with the lowest affinities were isolated from patients with highly disabling neuropathy. However, one of these two patients was considerably improved by short courses of plasmapheresis over a 10-year period. Patients R.O.G., A.Z.I. and F.U.E. partially improved initially under plasmapheresis and chemotherapy but further progression of the neuropathy was not influenced by any treatment. Patients L.O.C., B.L.A., D.E.P. and P.E.C. were not or slightly responsive to various therapeutic regimens.
Discussion The possible involvement of monoclonal IgM in the pathogenesis of human peripheral neuropathies stimulated during recent years a number of studies aiming to characterize their fine antibody activity to M A G and glycolipids (Ilyias et al., 1990), their antigenic specificities as well as their structural features (Brouet et al., 1989). In contrast to other monoclonal IgM having a defined antibody activity, such as rheumatoid factors and cold agglutinins, anti-MAG lgM express mainly private idiotopes (Saito et al., 1983; Kahn, 1985; Schmitt et al., 1987; Brouet et al., 1989) and seem to use pairing of different variable regions of their heavy and light chains to build their combining sites (Brouet et al., 1989). The latter
finding was rather unexpected in view of the narrow specificity of these IgM for epitopes carried by the glucuronyl sulfate determinant of SGPG. It may indicate that the human variable gene repertoire for this determinant is rather large. Finally, present structural data on light chains from anti-MAG IgM support the possibility that the expansion of anti-MAG synthesizing B cells may be antigen driven (Mihaesco et al., 1989). In view of the latter possibility and taking into account the natural course of the neuropathy which is quite variable from patient to patient, we wished to determine the affinity constant of these monoclonal IgM for both M A G and SGPG. All IgM had a 10- to 100-fold higher affinity for M A G than for SGPG. This holds true for IgM having either a relatively high or low affinity for MAG. Similarly, the IgM having the highest affinity for M A G also featured the best affinity for SGPG. Therefore the determinant recognized by these IgM was either more accessible or repetitive on M A G than on SGPG. Of note, whereas the Scatchard plots for S G P G were linear with homogenous binding sites, the same analysis on M A G yielded curvilinear plots with two types of binding sites with relatively low or high affinity. Whereas it is difficult to rule out trivial explanations such as aggregation of M A G molecules at high concentrations leading to decreased receptor binding, the possibility of different functional sites in a single IgM molecule is not without precedent. For instance, monoclonal IgM with anti-IgG antibody activity could bind only five IgG molecules which behave as a monovalent antigen (Stone and Metzger, 1968a). Therefore it seems likely that when one site of the IgM subunit is saturated, the effective binding constant of the second one may be considerably smaller than the intrinsic affinity constant (Stone and Metzger, 1968b). This hypothesis fits well with our data which showed that half of the sites were either of high or low affinity. Although structural data on anti-MAG IgM are still scarce, we should like to outline that two IgM (F.U.E. and D.E.P.) with VK~v light chains exhibiting a high n u m b e r of amino acid substitutions (Mihaesco et al., 1989) had either a relatively high (4 x 10 -8 for IgM F.U.E.) or low (_< 5 x 10 -6 for IgM D.E.P.) affinity constant for
214
MAG. Therefore, at least in these cases, the rate of mutation does not seem to result in an increased affinity. Further studies should more thoroughly address this point which is of critical importance to gain insight into the pathogenesis of the disease. Taken together, the affinity of different IgM for MAG ranged from 1 . 3 × 1 0 ~ to 7 × 1 0 '~ mol/liter and for SGPG from 1.7 x 10 -4 to 6.6 x 10 s tool/liter. These values are higher than that reported for monoclonal IgM with anti-IgG specificity (Stone and Metzger, 1968b). Most of them are obviously lower than that of post-immunization antibodies. The affinity of these IgM is in the same order of magnitude as that of the so-called natural polyreactive antibodies found in normal sera; however, these IgM to MAG or SGPG are devoid of notable polyreactivity (Aissa-Feunira et al., 1991). We do not find obvious correlations between the affinity of these IgM and their fine antibody activity. Definitive conclusions must, however, await the study of a larger group of patients in view of the diversity of the epitopes carried by the glucuronyl sulfate determinant. Likewise, there were no correlations between the severity of the neuropathy as assessed by a disability scale and the affinity of the various IgM. However, one must take into account that in this small series of patients some selection bias may exist (for instance purification of the IgM implied that the patients were under plasmapheresis) and that the time elapsed between the onset of symptoms and diagnosis or treatment differed widely from patient to patient. In addition, the possibility exists that other molecules such as nerve low molecular weight glycoproteins, which we have not studied here, may be involved in the pathogenesis of the neuropathy.
References Aissa-Feunira, F.B., Benamar-Elgaaied, A., Bouguerra, A., Schmitt, C., Brouet, J.C. and Dellagi, K. (1991) A human monoclonal IgG reactive with a private idiotype of a monoclonal IgM with autoantibody activity against myelin associated glycoprotein. Eur. J. Immunol. 4, 1065-1068. Ariga, T., Kohriyama, T., Freddo, L., Latov, N., Saito, M., Kon, K., Ando, S., Suzuki, M., Hemling, M.E., Rinehart,
Jr., K.L., Kusunoki, S. and Yu, R.K. (1987) Characterization of sulfated glucuronic acid containing glycolipids rcacting with lgM M-proteins in patients with neuropathy. J. Biol. Chem. 262, 848 853. Brouet, J.C., Dellagi, K., Gendron, M.C., Chevalier, A., Schmin, C. and Mihaesco, E. (1989) Expression of a public idiotype by human monoclonal IgM directed to myelin-associated glycoprotein and characterization of the variability subgroup of their heavy and light chains. J. Exp. Med. 170, 1551-1558. Chou, D.K.H., Ilyas, A.A., Evans, J.E., Costello, C., Quarles, R.H. and Jungalwala, F.B. (1986) Structure of sulfated glueuronyl glycolipids in the nervous system reacting with HNK-I antibody and some IgM paraproteins in neuropathy. J. Biol. Chem. 261, 11717-11725. Dancea, S., Dellagi, K., Renaud, F., Mahouy, C., Hauw, J.J. and Brouet, J.C. (1989) Effect of passive transfer of human anti-myelin-associated glycoprotein IgM in marmoset. Autoimmunity 3, 29-37. Friguet, B., Chaffotte, A.F., Djavadi-Ohaniance, L. and Goldberg, M.E. (1985) Measurements of the true affinity constant in solution of antigen-antibody complexes by enzyme-linked immunosorbent assay. J. Immunol. Methods 77, 305-319. Grillot-Courvalin, C., Labaume, S. and Brouet, J.C. (1986) Differentiation by interleukin 2 of a subpopulation of human B cells. Scand. J. Immunol. 23, 679-684. Handa, S. and Nakamura, K. (1984) Modification of sialic acid carboxyl group of ganglioside. J. Biochem. 95. 1323-1329. Hauttecceur, B., Schmitt, C., Dubois, C., Danon, F. and Brouet, J.C. (1990) Reactivity of human monoclonal IgM with nerve glycosphingolipids. Clin. Exp. Immunol. 80, 181 185. Hays, A.P., Latov, N., Takatsu, M. and Sherman, W.H. (1987) Experimental demyelination of nerve induced by serum of patients with neuropathy and an anti-MAG lgM M-protein. Neurology 37, 242-256. Ilyias, A.A., Chou, D.K.H., Jungalwala, F.B., Costello, C. and Quarles, R.H. (1990) Variability in the structural requirements for binding of human monoclonal anti-myelin-associated glycoprotein immunoglobulin M antibodies and HNK-I to sphingoglycolipid antigens. J. Neurochem. 55, 594-601. Kahn, S.N. (t985) Human IgM monoclonal antibodies with restricted antigenic specificity for myelin express unrelated idiotypes. J. Neurol. Sci. 69, 161-170. McGinnis, S., Kohriyama, T., Yu, R.K., Pesce, M.A. and Latov, N. (1988) Antibodies to sulfated glucuronic acid containing glycosphingolipids in neuropathy associated with anti-MAG antibodies and in normal subjects. J. Neuroimmunol. 17, 119-126. Mihaesco, E., Ayadi, H., Congy, N., Gendron, M.C., Roy, J.P., Heyermann, H., Frangione, B. and Brouet, J.C. (1989) Multiple mutations in the variable region of the K light chains of three monoelonal human IgM with anti-myelinassociated glycoprotein activity. J. Biol. Chem. 264, 2148121485. Monaco, S., Bonetti, B., Ferrari, S., Moretto, G., Nardelli, E., Tedesco, F., Mollnes, T.E., Nobile-Orazio, E., Manfredini,
215 E., Bonazzi, L. and Rizzuto, N. (1990) Complement-mediated demyelination in patients with IgM monoclonal gammopathy and polyneuropathy. New Engl. J. Med. 322, 649-652. Quarles, R.H. and Pasnak, C.F. (1977) A rapid procedure for selectively isolating the major glycoprotein from purified rat brain myelin. Biochem. J. 163, 635-637. Saito, T., Sherman, W.H. and Latov, N. (1983) Specificity of idiotypes of M-proteins that react with MAG in patients with neuropathy. J. Immunol. 130, 2496-2498. Schmitt, C., Dellagi, K., Mihaesco, E. and Brouet, J.C. (1987) Detection of cross-reactive determinants shared by human monoclonal IgM reacting with myelin-associated glycoprotein. J. Immunol. 138, 1442-1446. Sergott, R.C., Brown, M.J., Lisak, R.P. and Miller, S.L. (1988) Antibody to myelin-associated glycoprotein produces central nervous system demyelination. Neurology 38, 422-426.
Shy, M.E., Vietorisz, T., Nobile-Orazio, E. and Latov, N. (1984) Specificity of human IgM M-proteins that bind to myelin-associated glycoprotein: peptide mapping, deglycosylation, and competitive binding studies. J. Immunol. 133, 2509-2512. Steck, A.J., Murray, N., Dellagi, K., Brouet, J.C. and Seligmann, M. (1987) Peripheral neuropathy associated with monoclonal IgM autoantibody. Ann. Neurol. 22, 764-767. Stone, M.J. and Metzger, H. (1968a) Study of macromolecular interactions by equilibrium molecular sieving. J. Biol. Chem. 243, 5049-5055. Stone, M.J. and Metzger, H. (1968b) Binding properties of a Waldenstr6m macroglobulin antibody. J. Biol. Chem. 243, 5977-5984.
