Individual Restriction Of Fine Specificity Variability In Anti-GM1 IgG Antibodies Associated With Guillain-Barré Syndrome

Elevated titers of serum antibodies against GM1 ganglioside are associated with a variety of autoimmune neuropathies. Much evidence indicates these autoantibodies play a primary role in the disease processes, but the mechanism for their appearance is unclear. We studied the fine specificity of anti-GM1 antibodies of the IgG isotype present in sera from patients with Guillain-Barré syndrome (GBS), using thin-layer chromatogram-immunostaining of GM1, asialo-GM1 (GA1), GD1b and GM1-derivatives with small modifications on the oligosaccharide moiety. We were able to distinguish populations of antibodies with different fine specificity. Remarkably, individual patients presented only one or two of them, and different patients had different populations. This restriction in the variability of antibody populations suggests that the appearance of the anti-GM1 antibodies is a random process involving restricted populations of lymphocytes. With the origin of disease-associated anti-GM1 antibodies as a context, this finding could provide explanation for the “host susceptibility factor” observed in GBS following enteritis with GM1 oligosaccharide-carrying strains of Campylobacter jejuni.

Scientific RepoRts | 6:19901 | DOI: 10.1038/srep19901 these results suggest that the "binding site drift" mechanism could also be contributing to the induction of anti-GM1 antibodies of the IgG isotype.

GBS patients' sera display different anti-GM1 IgG antibody populations. Thirty GBS sera having
anti-GM1 IgG antibodies were selected for this study. Specificity of patient antibodies was assessed by thin-layer chromatography (TLC)-immunostaining and soluble antigen-binding inhibition assay (SABIA). A full summary of serum antibody cross-reactivities and clinical features of GBS patients is shown in Table 1. Antibodies that recognize GM1 can have four different fine specificities, depending if they cross-react or not with two structurally related glycolipids: GA1, desialylated form of GM1; and GD1b, a GM1 molecule with an additional sialic acid residue 7,13 . TLC-immunostaining patterns of patient sera were variable. Four representative cases are shown in Fig. 1. Almost half (13) of the sera stained only GM1 (Fig. 1B), whereas the rest also showed cross-reactivity with GA1 (Fig. 1C), GD1b (Fig. 1D) or with both glycolipids (Fig. 1E).

Fine specificity variability of anti-GM1 IgG antibody populations is restricted within each individual GBS patient.
In all GBS patients, preincubation of sera with soluble GM1 inhibited the binding of anti-GM1 IgG antibodies to TLC-adsorbed GM1 but also to GA1 and GD1b (results not shown), indicating that cross-reacting anti-GM1 antibodies are involved in the staining of GA1 and GD1b. It is clear that sera showing reactivity only with GM1 contained only one antibody population defined by fine specificity (GM1-specific), but sera having cross-reacting antibodies can have more than one population. From twelve sera showing cross-reactivity with both GA1 and GD1b, six contained only one population ~ binding to all three glycolipids ( Fig. 2A) was inhibited by preincubation with either GA1 (Fig. 2B) or GD1b (Fig. 2C). In the other six sera, binding to GM1 was not completely inhibited by GA1 (Fig. 2E) or by GD1b (Fig. 2F) indicating that, in addition to cross-reacting antibodies, the sera contained also the GM1-specific population.
The remaining sera showed only one type of cross-reactivity: three of them cross-reacted only with GA1 and two only with GD1b (see Fig. 1C,D). In all sera reacting with GA1, binding to GM1 was completely inhibited by soluble GA1, indicating only one population of antibodies (result not shown). In contrast, both sera cross-reacting exclusively with GD1b contained also a GM1 specific population (results not shown).
Although four different populations of anti-GM1 antibodies can be clearly distinguished according to their cross-reactivity with GA1 and GD1b, some additional heterogeneity was observed within these populations. The six sera containing only the population that cross-reacted with GA1/GD1b (Fig. 3A) presented different staining patterns ( Fig. 3B): from a serum showing similar cross-reactivity for both glycolipids, to a serum preferentially cross-reacting with one of them.

Anti-GM1 specific IgG antibodies vary their structural requirements between different GBS patients.
To study the antibody population specific for GM1 in more detail, chemically modified GM1 molecules were used as antigen (Fig. 4A). As exemplified in Fig. 4B, the chemical modification of certain functional groups in the GM1 molecule reduced partially or completely the binding of patient antibodies. Binding to GM1-derivatives was inhibited by preincubation of the sera with soluble GM1, indicating that the same antibodies are involved in the binding to both, the derivatives and the unmodified GM1 (results not shown). Different immunoreactivity patterns with the derivatives were found. Although some patients showed similar results, the patterns of reactivity with the derivatives were quite variable among the different sera (Fig. 4C).

Discussion
When sera from GBS patients were analyzed by TLC-immunostaining using GM1, GA1 and GD1b gangliosides as antigen, an interesting observation was done: immunostaining pattern was quite different among patients. A further characterization of the antibodies showed a remarkable result: patients had a restricted variability in antibody populations defined by fine specificity. From all different antibody populations we were able to distinguish, individual patients presented only one or two of them, and different patients had different populations. The meaning of this restriction can be analyzed by considering the structure of the recognized antigen. The oligosaccharide moiety of GA1 is included in the structure of GM1, and that of GM1 in GD1b. Based on this fact, the different cross-reactivity of anti-GM1 antibody populations can be explained by the recognition of different areas of GM1 oligosaccharide by the antibodies 7 . A similar interpretation can be done if we analyze the reactivity  of GM1-specific antibodies (Fig. 4C) in the context of the GM1 three-dimensional structure (shown in Fig. 4A). Areas differentially recognized account for different antibody binding site structures and, consequently, different B-lymphocyte clones. Therefore, restriction of antibody population variability would indicate that one or few B-lymphocyte clones are involved in each patient´s immune response. This idea implies that the resultant antibodies produced will be monoclonal or oligoclonal, a fact that is known to occur in autoimmune diseases [14][15][16] . This has been recently described for neuropathy-associated anti-GM1 antibodies through immunoglobulin light chain usage detection 17 . In addition, by affinity purification and further isoelectric focusing of anti-GM1 IgG antibodies from three GBS patients, Townson et al. depicted an oligoclonal type of response 18 .
Most of the patients studied here had a preceding diarrhea, an indication that the "molecular mimicry" mechanism was involved in the generation of antibodies. On the other hand, immunization of rabbits with GM1 in a proper adjuvant induces a classic polyclonal antibody response, including isotype changes and presence of different anti-GM1 antibody populations 13,19 . Consequently, the induction of a classical immune response without population restriction would be expected from the "molecular mimicry" mechanism. At this point, a question emerges: why did this not occur? One possible answer is provided by the "binding site drift" hypothesis 11 . This hypothesis was developed to explain the origin of disease-associated anti-GM1 IgM antibodies present in patients with neuropathies 7 . It is based in three facts: i. GM1 is a self-antigen and consequently B-cell clones recognizing GM1 with high affinity should not be present in normal individuals; ii. IgM antibodies that recognize GM1 with low affinity and a defined fine specificity are part of the normal human repertoire of anti-bacterial antibodies; and iii. Disease-associated IgM antibodies have higher affinity for GM1, and show restricted variability in fine specificity. The hypothesis proposes that patient B-cell clones originate from normally occurring ones (Fig. 5). B-lymphocytes producing normal anti-GM1 antibodies spontaneously mutate their V genes, thus modifying their binding sites. Some of these mutations increase the binding affinity for GM1, and the new B-lymphocytes can now be stimulated by self or foreign GM1. During the process, fine specificity can change and various potential paths can be followed, generating antibody populations with distinct fine specificity. Each lymphocyte follows one of these paths at random ("drift") and if only one or a few lymphocytes are involved, a restricted pattern of populations will be generated. Before or during the infection, in those diarrheal patients where normally occurring B cell clones undergo the "drift" process, antibodies with restricted fine specificity will be induced. If the Reactivity of a patient serum with chemically modified GM1. GM1-derivatives were immunostained with a patient serum having specific anti-GM1 antibodies. One plate was stained with orcinol reagent for chemical visualization of the derivatives. (C) Thirteen patient sera having reactivity only with GM1 (not GA1 or GD1b) were used for TLC-immunostaining of GM1-derivatives. Reactivity within each serum was expressed as full, reduced or no reactivity compared to the corresponding anti-GM1 reactivity.

Figure 5. Generation of Guillain-Barré syndrome-associated anti-GM1 IgG antibodies with individual
restriction of fine specificity variability in the context of "binding site drift" and "molecular mimicry" mechanisms. B cells producing normally occurring anti-GM1 antibodies ("normal reactivity") can undergo spontaneous mutations of V genes, randomly re-shaping their binding sites ("binding site drift"). This can lead to eventual increase in binding affinity for GM1 and also to various potential paths generating antibody populations with distinct fine specificity (represented as shaded areas in the different "types"). If only a few lymphocytes get involved, a restricted pattern of fine specificity is produced. During an infection with GM1 oligosaccharide-carrying bacteria ("molecular mimicry"), those diarrheal patients that experienced the "drift" process will get anti-GM1 antibodies with restricted fine specificity induced, and Guillain-Barré syndrome will process occurs in only a few patients, this could explain why only a minority of patients infected with GM1 oligosaccharide-carrying strains of C. jejuni develop GBS 20 .
In summary, the emergence of restricted patterns for anti-GM1 antibody populations through events described by the "binding site drift" hypothesis can account for the puzzling "host susceptibility factor" frequently observed in close association with the "molecular mimicry" mechanism in GBS 12 .

Methods
Patients. Sera from 30 GBS patients carrying anti-GM1 IgG antibodies were collected at Dokkyo Medical University, Tochigi, Japan, with prior approval from its Ethics Committee. Written informed consent was obtained from every patient. Serum samples taken during the first three weeks after the disease onset, before immune treatment, were stored at -80 °C until use. Sera were analyzed at the Argentinean laboratory. For this purpose, small volume of sera were lyophilized and transported by courier service. Previous experiments done with human and rabbit sera indicated that this treatment does not modify antibody activity (titer, affinity and fine specificity) of anti-ganglioside antibodies. All procedures were approved by the Ethics Committee of CIQUIBIC-CONICET. Criteria for inclusion were a positive spot for GM1 or GM1 and GA1 / GD1b in thin-layer chromatography (TLC)-immunostaining at 1/200 dilution, and no spot for other gangliosides. Twenty-four patients (80%) had diarrhea episodes preceding neurological symptoms. All experiments were performed in accordance with Ethical Guidelines on Research Involving Human Subjects 21 .
Glycolipids. GM1, GD1a, GD1b and GT1b were obtained from human brain. Folch upper phase was purified by DEAE-chromatography 22 , and HPLC on Iatrobeads silica gel column 23 . GA1 was prepared by acid hydrolysis of cow brain gangliosides 24 . GM1 derivatives. D1: Specific modification of C6 of GM1 terminal galactose was done by enzymatic treatment (galactose oxidase) and oxime formation. Briefly, 2.5 mg of GM1 was dissolved in 1 ml of 50 mM sodium phosphate, pH 7.0, containing 1% Triton X-100 and incubated for 24 h. with 15 U of galactose oxidase (Sigma, St. Louis, MO) at 37 °C. The aldehyde product was desalted by a Sep-Pak C18 cartridge (Millipore Corp., Milford, MA) and purified by HPLC 23 . For oxime formation, the purified GM1-aldehyde was dried and dissolved in 1 ml of saturated hydroxylamine hydrochloride in pyridine. After 24 h at RT, 2 ml of water was added and the resulting oxime of GM1 was purified by a Sep-Pak C18 cartridge. D2: De-N-acetylation of GM1 sialic acid was done by mild alkaline hydrolysis in aqueous 90% N-butanol 25 . D3: The glycerol chain of GM1 sialic acid was oxidized to the 7-and 8-aldehyde forms by mild periodation and then reduced to the corresponding truncated primary alcohols as described by Spiegel et al. 26 .
All GM1 derivatives were further purified by HPLC.
Soluble antigen-binding inhibition assay. Inhibition of antibody binding to plate bound ganglioside antigen was accomplished by incubating the sera with 0.1 mM GM1, GA1 or GD1b for 60 min before adding them to the plates.