Molecular mimicry: sensitization of Lewis rats with Campylobacter jejuni lipopolysaccharides induces formation of antibody toward GD3 ganglioside

Abstract

Recently we have reported cases of demyelinating inflammatory neuropathy showing elevated titers of anti-GD3 antibodies, which occurs rarely in Guillain-Barré syndrome. To examine the correlation between the anti-GD3 antibody titer and Campylobacter jejuni infection, we sensitized female Lewis rats with lipopolysaccharides (LPSs) from serotype HS19 of C. jejuni and examined changes in nerve conduction velocity and nerve conduction block (P/D ratio). After 16 weeks of sensitization, animals revealed decreases of nerve conduction velocity and conduction block (P/D ratio) and high titer of anti-GD3 antibodies. These anti-GD3 antibodies also blocked transmission in neuromuscular junctions of spinal cord-muscle cells cocultures. The GD3 epitope was verified to be located on the Schwann cell surface and nodes of Ranvier in rat sciatic nerve. To determine the target epitope for GD3 antibodies in causing nerve dysfunction, the LPS fraction containing the GD3 epitope was purified from the total LPS by using an anti-GD3 monoclonal antibody-immobilized affinity column. Subsequently, chemical analysis of the oligosaccharide portion was performed and confirmed the presence of a GD3-like epitope as having the following tetrasaccharide structure: NeuAcalpha2-8NeuAc2-3Galbeta1-4Hep. Our data thus support the possibility of a contribution of GD3 mimicry as a potential pathogenic mechanism of peripheral nerve dysfunction.

Fig. 1

Animals were subjected to diagnostic electromyography measurement in the tail nerve without surgical operation. A: Nerve conduction waves were recorded from stimulations of the distal and proximal electrode positionings on the tail. B: Latency_{proximal-distal} was determined by subtracting proximal from distal latency of nerve conduction waves. NCV and nerve conduction block (P/D ratio) were calculated by the following formulas, respectively: NCV (m/sec) = 0.03/latency_{proximal-distal}; P/D ratio = proximal to distal amplitude ratio among nerve conduction waves. If P/D ratio is less than 1.0, nerve conduction blocks may occur in the segment between electrode positionings.

Fig. 2

Time course of Ab titers after sensitization. Average IgG (solid columns) and IgM (open columns) Ab titers against GM1, GD3, LPS, or KLH in the sera of LPS-treated animals (N = 8).

Fig. 3

Nerve dysfunction and demyelination. A: Time course of changes in NCV and P/D ratio. Solid columns, LPS-treated animals (N = 8); open columns, vehicle-treated animals (N = 8); stippled columns, nontreated animals (N = 4). Bars are standard deviations, and statistical differences are determined by Student's t-test (*P < 0.05, **P < 0.01). B: Luxol fast blue staining in the transverse sections of rat sciatic nerve. a: Arrows show areas of demyelination (LPS-1). b: Control animals. Scale bar = 10 μm.

Fig. 4

NCV (A) and P/D ratio (B) are plotted against anti-GD3 (1), -GM1 (2), or -LPS (3) Ab titers in LPS-treated rats at 16 weeks after immunization. Circles correspond to each of the animals; in particular, the three arrowed animals (LPS-1–3) are characterized by high titers of anti-GM1 and -GD3 Abs with severity of nerve conduction dysfunctions as described in Results. r Value is the correlation coefficient between Ab titer and nerve conduction (NCV or P/D ratio).

Fig. 5

Effect of immunoabsorption on Ab-binding and NMJ-blocking activities. A: TLC overlay with IgG GSL Abs of rat serum (LPS-1–3 in Fig. 4B), serum immunoabsorbed with GM1 (LPS-1′–3′), and serum immunoabsorbed with GM1 and GD3 (LPS-1″–3″). Lanes 1, GM1; 2, GM2; 3, GD1a; 4, GD1b; 5, GT1b; 6, GQ1b; 7, GD3; 8, GA1. B: Time course of inhibitory effect on muscle action potential frequencies in spinal cord–muscle cocultured cells. Arrow shows addition of rat serum, 10 μl of LPS-1–3 (a); serum immunoabsorbed with GM1 only, 10 μl of LPS-1′–3′ (b); and serum immunoabsorbed with GM1 and GD3, 10 μl of LPS-1″–3″ (c).

Fig. 6

Localization of GD3 epitope in rat sciatic nerve. Teased fibers of rat sciatic nerve immunofluoresence stained by anti-GD3 mAb R24 (A), TRITC-peanut agglutinin (B), or anti-GD3 mAb R24 and TRITC-peanut agglutinin (C). Cryosections of rat sciatic nerve immunofluoresence stained by anti-GD3 mAb R24 (D), anti-NF (E), or anti-GD3 mAb R24 and anti-human neurofilament Ab (F). Scale bar = 10 μm.

Fig. 7

Chemical analysis of the epitopic structure of HS19-derived LPS recognized by anti-GD3 mAb R24. A: The fraction (100 ng) eluted with n-propanol:H₂O:triethylamine (75:20:5, v/v/v; solvent 3) by silica gel column chromatography was developed on an HPTLC plate and examined by TLC overlay using anti-GD3 mAb R24. B: After purification by anti-GD3 mAb R24-immobilized affinity chromatography, the oligosaccharides of LPS were converted to ABOE derivatives and analyzed by MALDI-TOF mass spectrometry. The arrowed peak at m/z 1,186.1 is assigned to protonated molecular ion [M + H]⁺. The peaks at m/z 1,229.0 and 1,208.1 are sodium adduct ions [M + 2Na−2H]⁺ or [M + Na]⁺, respectively. The peaks at m/z 916.3 and 625.6 correspond to loss of one or two NeuAc residues from the sodium adduct ions [M + Na-NeuAc]⁺ and [M + Na-2NeuAc]⁺, respectively. C: The proposed carbohydrate structure of the epitope recognized by anti-GD3 mAb R24.