Three clinicopathological observations support the hypothesis that physiological blockade of conduction at the level of the axon, in addition to motor axonal degeneration, contributes significantly to clinical weakness in AMAN. First, many patients with AMAN recover quite quickly, with or without IVIg treatment, and their time course of recovery is incompatible with nerve fiber degeneration and regeneration (Ho et al., 1997a; Kuwabara et al., 1998). Second, the pathological studies in AMAN and its animal model indicate that despite severe clinical weakness, axonal degeneration is not universal in early cases, and such degeneration affects only a proportion of motor fibers even in late cases (Griffin et al., 1996c; Susuki et al., 2003). Third, the issue of mismatch between clinical and pathological findings and axonal conduction failure was examined by serial electrophysiolog-ical studies in a set of patients with GBS. In this important study, Kuwabara and colleagues (1998) examined the elec-trophysiological features of a group of GBS cases with IgG anti-GM1 antibodies and classified them into axonal or demyelinating variants on the basis of initial electrodiagnos-tic examination. Subsequent nerve conduction studies showed that all cases classified as AIDP and some axonal cases had rapid improvement in distal compound muscle action potential amplitudes. Further, cases that were initially classified as AIDP had rapid resolution of conduction slowing and block without the appearance of slow components that indicate remyelination. These findings raise the possibility that rapid recovery of CMAP amplitudes and motor conduction slowing without features of remyelination is due to a reversible conduction failure/block at the level of axon. The recovery of CMAP amplitudes in cases initially classi fied as axonal could be due to dysfunction at the level of node of Ranvier or motor nerve terminal. With this in mind, experimental evidence for physiological effects of anti-gan-glioside antibodies on the nodes of Raniver and motor nerve terminal is presented next.
There are several reasons to consider the node of Ranvier as a candidate site for the pathophysiological effects of anti-ganglioside antibodies. First, most complex gangliosides are concentrated at the nodes of Ranvier. Second, in myelinated fibers axonal target antigens are particularly exposed at this site. Third, ion channels are clustered at nodes of Ranvier, and disruption of their function can lead to conduction abnormalities and finally, structural integrity of the node of Ranvier is critical for nerve fiber conduction. Therefore, experimental studies, mostly with anti-GM1 antibodies, have focused on node of Ranvier dysfunction in both animal and in vitro models. Nevertheless, discrepant findings have been reported by different investigators.
Initially, two separate studies showed that intraneural injections, in rats, of human anti-GM1 antibodies caused acute conduction block (Santoro et al., 1992; Uncini et al., 1993), but this finding was not confirmed by another group that used immunoglobulin fractions with IgG or IgM anti-GM1 reactivity purified from patients with GBS or multifocal motor neuropathy (Harvey et al., 1995). The basis for these discrepant findings has not been resolved, but it could reflect factors that were not strictly matched: (1) anti-GM1 antibody properties such as fine specificity and/or affinity, (2) whole serum versus purified immunoglobulins and anti-GM1 titers, and (3) source and activity of the exogenous complement. This issue of anti-ganglioside antibody-mediated conduction failure is rekindled by a recent report indicating that, in a rabbit model of AMAN associated with anti-GM1 antibodies, conduction abnormalities were noted in myelinated fibers in the spinal roots (Susuki et al., 2003). To explore this issue further, we have recently examined the effects of an anti-GD1a-reactive monoclonal antibody injected intraneurally in rat sciatic nerves. The serial nerve conduction studies in this model indicate development of conduction failure that reverses over time. Notably, exogenous complement was required in this model to induce conduction abnormalities (J. Spies, K. Sheikh, and colleagues, unpublished findings). Overall, these findings support the concept that anti-ganglio-side antibodies with certain specificities can cause conduction failure in myelinated nerve fibers.
Different in vitro studies with anti-GM1 antibodies also report discrepant findings. The first such study examined the effects of human and rabbit sera with anti-GM1 reactivity on desheathed rat sciatic nerve preparations and found mild reduction of CMAP amplitudes (Arasaki et al., 1993). An extremely provocative finding by Takigawa and colleagues (1995) provided strong support to the concept of anti-GM1 antibody-mediated nodal ion channel dysfunction. This study, by using a voltage clamp technique on isolated single myelinated rat nerve fibers, found that rabbit sera with anti-GM1 reactivity caused increased potassium current, as elicited by step depolarization, in a complement-independent manner, whereas these antibodies in the presence of active complement caused decreased sodium currents and eventually blocked the channels irreversibly. These investigators reproduced their findings with GBS sera containing anti-GM1 antibodies (Takigawa et al., 2000). The hypothesis of sodium channel dysfunction was further supported by a report by Weber et al. (2000) showing that rabbit sera with IgG anti-GM1 reactivity caused reversible blockade of the voltage-gated sodium channels in a neural cell line. It was postulated that complexes of anti-GM1 antibodies and complement products block the ion-conducting pore of the channel directly. In contrast, two other studies in vitro that used human or rabbit sera with anti-GM1 reactivity did not observe acute conduction abnormalities (Hirota et al., 1997; Paparounas et al., 1999). The basis of these discrepant results is not resolved and could reflect differences in the antibody properties or methodologies used by different investigators.
One possible mechanism of sodium channel dysfunction in the Takigawa and Weber studies could be that anti-GM1 antibodies bound directly to cross-reactive oligosaccharides on highly sialylated and glycosylated sodium channels led to abnormalities in ion channel conduction. We examined this issue in the spinal roots of dystrophic mice that contained large amyelinated fibers with randomly distributed patches of sodium channels in the absence of nodes of Ranvier. This preparation allowed us to ask whether sodium channels have an obligatory co-localization with GM1-binding ligands (Sheikh et al., 1999). Our results showed that GM1-binding ligands did not co-localize with sodium channels in this situation, suggesting that anti-GM1 antibody-mediated nodal dysfunction is likely to be indirect and not due to direct targeting of sodium channels. However, this issue needs further resolution at the level of GBS-derived anti-GM1 antibodies and sodium channel isotype(s) concentrated at nodes of Ranvier in human peripheral nerves.
The motor nerve terminal is a susceptible site for anti-ganglioside antibody-mediated injury because it is enriched in gangliosides such as GM1, GD1a, and GQ1b, and it lacks a blood-nerve barrier. We have previously reported that this site degenerates in AMAN (Ho et al., 1997b), and an elec-trophysiological study suggests that this site is affected in some cases of FS (Uncini and Lugaresi, 1999). Experimental pathophysiological studies have examined both complement-dependent and complement-independent effects of anti-ganglioside antibodies on this site in phrenic nerve hemidiaphragm preparations.
a. Complement-Dependent Pathophysiological Effects In a series of studies, Hugh Willison's group in Glasgow (Willison et al.,1996; Goodyear et al.,1999; Plomp et al., 1999) extensively investigated the effects of both human and experimental antibodies, with specificities similar to FS, on motor nerve terminal in phrenic nerve hemidiaphragm preparations both in vitro and ex vivo. They have demonstrated that anti-GQ1b reactive antibodies bind to neuro-muscular junctions and induce massive quantal release of acetylcholine from nerve terminals, eventually blocking the neuromuscular transmission on a presynaptic basis. This effect is dependent on the activation of complement, probably through the alternate pathway, but neither requires activation of the classical pathway nor the formation of membrane attack complex. These anti-ganglioside antibody-mediated changes in motor nerve terminal physiology closely resemble the effects of the paralytic neurotoxin, a-latrotoxin. Subsequent studies by this group showed that in this model, there is antibody and complement deposition accompanied by morphological disruption of the nerve terminal involving the loss of major cytoskeletal components, including neurofilaments (O'Hanlon et al., 2001). Notably, both calcium depletion and calpain inhibition protected the cytoskeleton from degradation, supporting the hypothesis that calcium ingress and activation of calcium-dependent proteases, calpains, are involved in degradation of the motor nerve terminal cytoskeleton (O'Hanlon et al., 2003). It appears that a-latrotoxin like-effects at NMJ are not restricted to antibodies with GQ1b reactivity, but can also be seen with antibodies with GD1a-reactivity (H. Willison, personal communication), thus supporting the notion that motor nerve terminal is a target site for anti-ganglioside antibody-mediated injury in AMAN.
b. Complement-Independent Immunopharmacological effects of antibodies In a parallel set of studies in a perfused macro-patch clamp electrode model, Buchwald and colleagues demonstrated blockade of evoked acetylcholine release and reduced amplitude of postsynaptic potentials, suggestive of combined presynaptic and postsynaptic blockade, after application of purified IgG from patients with GBS and FS directly to the nerve terminal and a concentration-dependent presynaptic blockade with a monoclonal antibody with GQ1b-reactivity (Buchwald et al. 1995, 1998, 2002). These effects were complement-independent and reversible, suggesting a direct immunophar-macological effect of the antibodies at the neuromuscular junction. We have examined the effects of IgG monoclonal antibodies with GD1a, GD1b, and GM1 reactivity in this model. Our results showed depressed evoked quantal release to a significant yet different extent with different mAbs, whereas the amplitude of postsynaptic currents was not significantly affected. The blockade was reversible or partially reversible after washout with these mAbs
(B. Buchwald K. Sheikh, unpublished observations). It is possible that anti-ganglioside antibody-mediated blockade of presynaptic evoked quantal release is due to interference with presynaptic calcium channels or proteins of the vesicle release machinery. Gangliosides are expressed on the outer leaflet of the plasma membrane and anti-ganglioside antibodies could indeed form an antibody-antigen complex that alters the function of the presynaptic channel proteins. Further, these results indicate that anti-ganglioside antibodies with different specificities cause distinct pathophy-siological effects, supporting the hypothesis that anti-ganglioside antibody specificity contributes to selective nerve fiber dysfunction and heterogeneity of clinical manifestations in immune neuropathies.
Was this article helpful?