The etiology of the Guillain-Barré syndrome (GBS) still remains elusive. Recent years have witnessed important advances in the delineation of the mechanisms that may operate to produce nerve damage. Evidence gathered from cell biology, immunology, and immunopathology studies in patients with GBS and animals with experimental autoimmune neuritis (EAN) indicate that GBS results from aberrant immune responses against components of peripheral nerve. Autoreactive T lymphocytes specific for the myelin antigens P0 and P2 and circulating antibodies to these antigens and various glycoproteins and glycolipids have been identified but their pathogenic role remains unclear. The multiplicity of these factors and the involvement of several antigen nonspecific proinflammatory mechanisms suggest that a complex interaction of immune pathways results in nerve damage. Data on disturbed humoral immunity with particular emphasis on glycolipid antibodies and on activation of autoreactive T lymphocytes and macrophages will be reviewed. Possible mechanisms underlying initiation of peripheral nerve-directed immune responses will be discussed with particular emphasis on the recently highlighted association with Campylobacter jejuni infection.

In the second part of our review the role of antecedent infections in the pathogenesis of GBS is discussed. The association with Campylobacter jejuni (C. jejuni) is highlighted and the concept of molecular mimicry, i.e., sharing of epitopes between microbes and peripheral nerve, explained. Alternative mechanisms to relate an infection with the immune-mediated neuropathy are elaborated. Current therapies of the GBS include plasma exchange, high-dose intravenous immunoglobulins, and supportive treatment directed to secondary complications. Published therapeutic trials are reviewed and future approaches are outlined. Principles of general care are also summarized.

Recent studies from our laboratory and by other investigators have shown that autoreactive CD4+ cells specific for peripheral nerve P2 protein have a powerful effect on blood-nerve barrier permeability. In this study we injected CD4+ T cells reactive to a nonneural antigen (ovalbumin) systemically and achieved their accumulation in the tibial nerve of Lewis rats by previous intraneural injection of ovalbumin. Selected rats were given systemic demyelinating antibody (antigalactocerebroside) to provide an indicator of changes in the permeability of the blood-nerve barrier, and the animals were monitored by sequential neurophysiological studies and histology. Circulating ovalbumin-specific T cells accumulated at sites of intraneural ovalbumin injection without inducing demyelination in control animals. In rats with circulating galactocerebroside antibodies, local conduction block and demyelination were seen in the region of T-cell accumulation. Electron microscopy demonstrated dissolution of some tight junctions between endothelial cells in areas of T-cell accumulation, and T cells traversing the endothelium between endothelial cells and through their cytoplasm. Endothelial cell damage was evident in these areas. This study demonstrates breakdown of the blood-nerve barrier by activated T cells, even of nonneural specificity, allowing the development of focal conduction block and demyelination in the presence of circulating antimyelin antibodies.

In order to confirm the reported pathogenicity of human antibodies to monosialoganglioside GM1, immunoglobulin fractions with high anti-GM1 IgG or IgM titers were prepared from patients with Guillain-Barré syndrome and multifocal motor neuropathy respectively. These fractions were injected intraneurally into rat tibial nerves with fresh human complement. Neither the anti-GM1 IgG nor the anti-GM1 IgM fraction induced significant focal conduction block or slowing compared to a pooled fraction prepared from 5 normal individuals. In contrast, rabbit experimental allergic neuritis serum included as a positive control was highly active. Transverse sections of injected nerve failed to show evidence of demyelination. Staining for human immunoglobulin in cryostat sections showed the presence of injected anti-GM1 antibody bound to nodes of Ranvier up to 6 days following intraneural transfer. These data fail to confirm previous reports of conduction block from intraneural transfer of anti-GM1 serum and suggest that such electrophysiological effects may be the result of factors other than or in addition to anti-GM1 antibodies.