Mechanisms of Action of Anti-GM 1 and Anti-GQ 1b Ganglioside Antibodies in Guillain-Barré Syndrome

Transcription

1 S144 Mechanisms of Action of Anti-GM 1 and Anti-GQ 1b Ganglioside Antibodies in Guillain-Barré Syndrome Hugh J. Willison, Graham O Hanlon, Gary Paterson, Colin P. O Leary, Jean Veitch, George Wilson, Mark Roberts, Teresa Tang, and Angela Vincent University Department of Neurology, Institute of Neurological Sciences, Southern General Hospital, Glasgow, and Department of Clinical Neurology, University of Oxford, Institute of Molecular Medicine, John Radcliffe Hospital, Oxford, United Kingdom Anti-GM 1 and anti-gq 1b ganglioside antibodies are found in association with acute and chronic peripheral neuropathies, including Guillain-Barré syndrome. They are believed to arise as a result of molecular mimicry with immunogenic microbial polysaccharides. Although anti-ganglioside antibodies are suspected to play a causal role in neuropathy pathogenesis, the details of this have yet to be proven. The approach in this laboratory to solving this issue has been to generate anti-gm 1 and anti-gq 1b monoclonal antibodies from peripheral blood lymphocytes of affected patients and to study their immunolocalization in peripheral nerve and their electrophysiologic effects in animal models in which peripheral nerve sites are exposed to anti-ganglioside antibodies. These data show that anti-ganglioside antibody reactive epitopes are widely distributed in peripheral nerve and can cause electrophysiologic abnormalities in a variety of model systems; thus, these data support the view that anti-ganglioside antibody reactive epitopes may directly contribute to neuropathy pathogenesis. Gangliosides are a class of ú100 structurally distinct membrane glycolipids that are highly enriched in the nervous system. They are composed of a ceramide tail inserted in the lipid bilayer and an oligosaccharide moiety containing sialic acid oriented extracellularly [1]. Gangliosides modulate a wide variety of neuronal functions and also act as bacterial toxin receptors and targets for autoantibodies [2]. Extensive studies on IgM paraproteins present in the sera of patients with chronic peripheral neuropathies led to the discovery that gangliosides and other glycolipids were important antigens for neuropathyassociated autoantibodies. As a result of studies on chronic neuropathies, antibodies specific for a range of structurally different gangliosides were sought and first observed almost a decade ago in the sera of patients with acute postinfectious neuropathy termed Guillain-Barré syndrome (GBS) [3]. Since then, there have been many publications on the relationship between anti-ganglioside antibodies and acute or chronic neuropathies; see [4 9] for detailed reviews on different aspects of this subject. There is strong, combined clinical and serologic evidence that anti-ganglioside antibodies are significantly associated with GBS. Anti-GM 1 antibodies are associated with a proportion of predominantly motor, demyelinating and axonal forms of GBS [10 13], in addition to being found in chronic motor neuropathies [14 15]. Anti-GQ 1b antibodies are found in ú90% of patients with the regional GBS variant called Miller Fisher syndrome (MFS) [16 18], to which a chronic counterpart also exists [19] (reviewed in [20]). What is less clear is the extent to which anti-ganglioside antibodies contribute to the pathophysiology of GBS and the mechanism(s) by which they do so. This field of research is contentious, with good reason, because it is clouded by a number of complex and often contradictory threads of opinion and data. Issues of rele- vance will be outlined and illustrated by our data and data from other laboratories, with particular focus on the fine specificity of anti-gm 1 and anti-gq 1b antibodies, the distribution of and access to reactive antigens in peripheral nerve, and pathophysi- ologic evidence from animal studies on the causal role of anti- GM 1 and anti-gq 1b antibodies. Presented: Workshop on the Development of Guillain-Barré Syndrome following Campylobacter Infection, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland, August Clinical specimens were studied with informed consent, according to local guidelines. Grant support: Wellcome Trust, Scottish Motor Neurone Disease Association, and Guillain-Barré Syndrome Support Group of Great Britain (to H.J.W.); Medical Research Council and Muscular Dystrophy Group/Myasthenia Gravis Association (to A.V.). Reprints or correspondence: Dr. Hugh J. Willison, University Department of Neurology, Southern General Hospital, Glasgow, G51 4TF, UK. The Journal of Infectious Diseases 1997;176(Suppl 2):S by The University of Chicago. All rights reserved /97/76S2 0012$02.00 Anti-GM 1 and Anti-GQ 1 b Antibodies Bind a Wide Range of Target Epitopes in Peripheral Nerve Anti-GM 1 and anti-gq 1b ganglioside antibodies are usually referred to in terms of the reactive ganglioside species that were first identified. The resulting terminology should be viewed with caution since anti-gm 1 and anti-gq 1b antibodies invariably share reactivity to greater or lesser extents with other neural glycoconjugates, which could be important pathogenic targets. Well-recognized patterns of shared anti-gm 1 IgG and IgM antibody reactivity include terminal Gal(b1-3)GalNAccontaining glycolipids, such as asialo-gm 1 (GA1) and GD 1b

2 JID 1997;176 (Suppl 2) Action of Ganglioside Antibodies in GBS S145 [21]; internal epitopes common to GM 1 and GM 2 [22]; and 1), and DRG cell bodies are CT positive. Thus, the physical a wide array of Gal(b1-3)GalNAc-binding peripheral nerve distribution of GM 1 clearly extends beyond that which would glycoproteins [23]. Similarly, anti-gq 1b antibodies detected in be predicted by the clinical phenotype associated with antipersons with MFS often react with GT 1a and, to a lesser extent, GM 1 antibodies, which predominantly involves motor rather with other di- or polysialylated gangliosides, including GD 1b, than sensory nerves. GD 3, and GT 1b [24 25]. They may also react with sialylated One explanation for this is that not all GM 1 is available for glycoproteins, as present, for example, on red blood cells [20]. anti-gm 1 antibody binding in the normal neural microenvironment. There may also be as yet unidentified peripheral nerve target Alternatively, anti-gm 1 antibody binding to GM 1 is equal antigens for these antibodies. throughout, but the pathophysiologic effects of such binding These shared reactivities are established by ELISA and thinlayer differ between sites. The former explanation would account chromatography results on purified or chromatographi- for some of the diversity of immunostaining seen with different cally separated gangliosides that have been isolated from their human anti-gm 1 IgM and IgG antibodies and antisera. For physiologic milieu and assayed at 4 C. They may correlate example, in unfixed rodent and human peripheral nerve frozensection poorly with antibody behavior and binding characteristics in and teased-fiber studies, we have found that human biologic membranes at body temperature. Using both purified monoclonal anti-gm 1 antibodies with apparently similar specificities monoclonal antibodies (MAbs) and polyclonal antisera, we and and avidities for GM 1 selectively stain GM 1 -containing others have shown that both the presence of accessory lipids membranes in unique, antibody-specific patterns [27]. The antibody and the assay temperature have important influences on antibody Sm1 (see figure 1) predominantly stains myelin in transand binding. For example, we have demonstrated that sera verse sections of rat peripheral nerve and stains neuronal cell with high titers of anti-gq 1b antibodies have a poorer reactivity bodies and associated axons in the DRG. In contrast, another to liposomally incorporated GQ 1b (phosphatidylcholine:choles- pair of anti-gm 1 antibodies (Bo1/3, not shown) stain a different terol:gq 1b at 4:5:1) and may have no reactivity at 37 C compared population of DRG neurons and fail to stain any other periphmonoclonal with that at 4 C [25]. We made similar observations with eral nerve structures. Another anti-gm 1 antibody (Br1) stains anti-gm 1 antibodies [26]. When considering the peripheral nerve myelin, axons, and extracellular matrix mate- mechanism of action of these antibodies, it is clearly important rial. In blocking experiments using excess, unlabeled CT, some to consider these variables. patterns of anti-gm 1 antibody staining can be completely blocked by CT, whereas others cannot. This indicates that both GM 1 and unidentified cross-reactive neural ligands can be The Distribution of and Access to Anti-GM 1 and Antibound by anti-gm 1 antibodies. GQ 1b Antibody Reactive Antigens in Peripheral Nerve In addition to the fine specificity of anti-gm 1 antibodies, the The presence of shared reactivities between anti-gm 1 or antibody class (IgM or IgG) is an important variable in allowing anti-gq 1b antibodies and multiple ganglioside and glycoconju- binding to GM 1, presumably because IgG is more penetrant gate species implies that many peripheral nerve antigens may than IgM, being smaller and of higher affinity. For example, be available for antibody binding with subsequent pathophysio- antisera from GBS patients that contain IgG anti-gm 1 antibodies logic effects. This is borne out by the diversity of antibody readily bind CT-positive nodes of Ranvier in teased fibers patterns seen in immunocytochemical studies on normal pe- [11], whereas we and others have great difficulty showing antiripheral nerve and complicates their interpretation, even when GM 1 IgM antibody binding to nodes in teased fibers or intact using monoclonal reagents in favor of whole human or animal preparations. antisera. The immunocytochemical data relating to anti-gq 1b anti- Most immunocytochemical studies have focused on anti- body binding are less extensive than that for GM 1 because GM 1 antibodies. With one exception [27 28], no GBS patient monospecific GQ 1b ligands are scarce. One study has shown derived human MAbs are available; thus, immunocytochemical GQ 1b to be enriched in human oculomotor nerves (a major site data relating to this have been extrapolated from studies using of clinical pathology in MFS) compared with peripheral nerve anti-gm 1 IgM antibodies associated with multifocal motor neu- and has also shown binding of a GQ 1b mouse MAb to human ropathy [29 32], anti-gm 1 IgG-containing sera from GBS patients oculomotor nerve nodes of Ranvier [24]. Our immunocyto- [11], and cholera toxin B subunit (CT), a high affinity chemical studies have used a human monoclonal IgM antibody GM 1 -specific ligand [33]. (Ha1, derived from a patient with a chronic ataxic neuropathy When defined by cholera toxin binding, GM 1 is very widely superficially resembling MFS) of broader specificity than that distributed in the human and rodent peripheral nervous system. to GQ 1b alone in that the Ha1 antibody also binds to other It is present at motor end plates and in terminal motor nerves, NeuAc(a2-8)NeuAc-linked gangliosides, including GD 1b, in compact myelin and axonal membranes, at nodes of Ranvier GT 1a,GT 1b, and GD 3 [34]. Many of these gangliosides have in peripheral nerve, and in the dorsal and ventral spinal roots. specific patterns of immunolocalization in the nervous system. In dorsal root ganglia (DRG), myelinated dorsal root fibers Since some MFS-associated antibodies also cross-react with appear to react more strongly than do ventral root fibers (figure these disialylated gangliosides, cautious extrapolation is possi-

3 S146 Willison et al. JID 1997;176 (Suppl 2) Figure 1. Immunofluorescence staining of rodent neural tissues with cholera toxin (CT, A and B) and anti-gm 1 IgM monoclonal antibody Sm1 (panels C and D). A, In dorsal root ganglion (DRG; area enclosed by broken line represents extent of neuronal cell bodies) CT stained dorsal root fibers (drf) more strongly than ventral root fibers (vr) and also stained subpopulation of DRG neuronal cell bodies (bar Å 100 mm). B, CT-stained nodes of Ranvier in teased fibers (bar Å 25 mm). C, In DRG, Sm1 stained subpopulation of CT-positive neurons and proximal axons (arrowhead; bar Å 50 mm). D, In transverse sections of sciatic nerve, Sm1 stained compact myelin in myelinated fiber profiles (bar Å 50 mm). ble. Using Ha1, we have found a very wide distribution of considered when setting up and interpreting models of antibody-mediated reactive tissue antigens in rodent and human peripheral nerve, neuropathy, as described below. including motor end plate regions, DRG, and peripheral and cranial (oculomotor) nerve myelin or axons (or both). We have also observed strong immunoreactivity with anti-gm 1 antibod- Pathophysiologic Evidence from Human and Animal ies in transverse sections of human oculomotor nerve in addi- Studies on the Causal Role of Anti-GM 1 and Anti-GQ 1b tion to peripheral nerve. This indicates that the GM 1 -GQ 1b Antibodies peripheral-cranial dichotomy (based on clinical features of the diseases associated with the relevant antibodies) is unlikely to The pathogenicity of anti-ganglioside antibodies in animal be accounted for by selective distribution of peripheral motor models has been addressed, using a variety of techniques, nerve antigen alone. mainly in rodents and rabbits. Immunization of rabbits with Antibody access to cryptic sites, be they protected by the gangliosides themselves can induce neuropathy, including GM 1 blood nerve barrier or more subtle microenvironmental con- and GD 1b [30, 36], but has not been demonstrated for GQ 1b. straints, such as the antigen density and the adjacent molecular However, this approach has not been particularly successful milieu, would appear to be an important factor in controlling because gangliosides are poor immunogens in their own right, the pathogenic potential of anti-ganglioside antibodies. It is and the fine specificity of the immune response obtained in likely on theoretical and experimental grounds that the blood- this manner may differ from that in the natural disease. Immuninerve barrier must be breached to allow antibody full access zation with bacterial lipopolysaccharides or their parent organisms to the nerve microenvironment [35]. These issues need to be that contain reactive epitopes shared with GM 1 or GQ 1b

4 JID 1997;176 (Suppl 2) Action of Ganglioside Antibodies in GBS S147 is more likely to be successful, particularly if done via the least in part, through action at the motor nerve terminals. This natural route of exposure (e.g., oral or respiratory routes). does not contradict the certainty, from clinical and pathologic Our approach to pathophysiology has been 2-fold: either studies, that other neural sites, such as motor nerve trunks, to expose isolated organ preparations to anti-gm 1 and GQ 1b muscle spindles, and afferent tracts, are also pathologically antibodies over a several-hour period or to passively transfer affected in MFS. antibodies to rodents by intraperitoneal injection over several The findings from our studies and those of others on the site days or weeks and then perform ex vivo electrophysiologic and mechanism of action of anti-gm 1 antibodies are less clear assessment. We have focused on the mouse hemidiaphragm (table 1). Conduction block has been demonstrated following preparation for several reasons. First, our immunocytochemical intraneural injection of sera containing anti-gm 1 in some studies studies using human MAbs have indicated that both anti-gm 1 [40, 41] but was not seen in another detailed study despite and anti-gq 1b (Ha1) antibody reactive epitopes are present in the presence of injected anti-gm 1 antibody bound nodes of the distal motor nerve and motor end plate. Second, this site Ranvier 6 days following intraneural injection [42]. In vitro lies outside the blood-nerve barrier and thus might be readily studies have also yielded some positive results. Arasaki et al. amenable to antibody penetration, compared with a main nerve [43] applied antisera containing anti-gm 1 to an isolated sciatic trunk. Third, human electrophysiologic evidence suggests that nerve segment and demonstrated acute conduction block. In the terminal motor nerve is a clinically important site of pathophysiology. another study using a voltage clamp technique on isolated sin- Fourth, botulinum toxins, which exert their para- gle myelinated rat nerve fibers, Takigawa et al. [44] induced lytic effects through disruption of nerve terminal exocytosis, a complement-independent increase in the K / current and, in do so by gaining access to the motor nerve terminal via ganglio- the presence of complement, observed a decrease in the Na / side-containing ectoacceptors, suggesting that anti-ganglioside current and a progressive increase in the nonspecific leakage antibodies may also be able to exert effects at this site [37]. current. Using the mouse hemidiaphragm model in conjunction In the mouse hemidiaphragm organ bath preparation, sera with anti-gm 1 positive antisera from chronic motor neuropacontaining anti-gq 1b antibody and IgG from patients with thy patients, we observed an increase in the stimulus threshold MFS (dialyzed in physiologic buffer and diluted 1:2) caused required to elicit end plate potentials and a progressive decline an acute rise in spontaneous neurotransmitter release over in end plate potential amplitude [45]. These data, taken in 30 min and then exerted a complete blocking effect on spontaneous conjunction with immunocytochemical data on the neural dis- and nerve-evoked neurotransmitter release over 2 tribution of GM 1, suggest that distal motor nerves, nodes of 3 h [38]. Using a perfused macro patch-clamp electrode, Ranvier, and their associated ion channels may be important Buchwald et al. [39] showed that a fast and fully reversible targets for anti-gm 1 antibodies. presynaptic blocking factor is present in MFS IgG. In passive Using GBS sera with high-titer anti-gm 1 antibodies, we have transfer experiments followed by ex vivo analysis comparing not obtained clear patterns of abnormalities. For example, in normal IgM with Ha1 IgM (5 mg of IgM/day for 10 days a recent series of blinded studies on 4 anti-gm 1 positive and injected intraperitoneally into mice producing human IgM 4 anti-gm 1 negative GBS sera and a range of control sera, levels ú1 mg/ml of mouse serum and anti-ganglioside antibody we observed a substantial reduction in 2 of the 8 GBS sera in titers ú1/1000 in Ha1-treated animals), a significant end plate potential amplitude (to õ5 mv in 30 min) and block fall in end plate potential amplitude (23.4 vs. 8.8 mv) and (in 7 of 9 and 6 of 10 fibers, respectively) but no effect in rise in stimulus threshold to evoke nerve transmission (1.3 control sera. One of the GBS sera was anti-gm 1 positive; the vs. 4.3 V) was observed [34]. second was anti-gm 1 negative but had a very high IgG anti The mechanism by which these effects take place is currently asialo-gm 1 titer of ú1/20,000. Thus, not all anti-gm 1 antibody unknown. Spontaneous and high potassium-evoked (80 mm positive samples exerted an effect in this system, and 1 K / ) acetylcholine release from the nerve growth factor stimu- anti-gm 1 negative sample did exert an effect. The samples lated rat pheochromocytoma PC12 cell line is unaffected by were not controlled for complement or cytokine levels, which MFS sera or Ha1 (O Leary CP, unpublished results). This is may be important cofactors. Apart from these results, electrophysiologic despite the fact that this cell line expresses GQ 1b and other studies on anti-gm 1 antibody postive GBS sera polysialylated gangliosides in its nerve growth factor differen- have not yet been reported, and further work is required to tiated state, reacts strongly with the MAb Ha1, and is sensitive establish the relevance of these findings. to botulinum toxin. These data suggest that anti-gq 1b antibodies may not exert a direct, pharmacologic effect on neurotransmitter release but additionally require complement or other Conclusions serum factors. Alternatively, mammalian neuromuscular junc- Gangliosides regulate many diverse physiologic processes tions containing anti-gq 1b antibody reactive epitopes may [46]. They are asymmetrically located in the plasma membrane have a particular vulnerability to the effects of these antisera. and anchored in the lipid bilayer by their ceramide tail, These data collectively suggest that anti-gq 1b antibodies exposing their sialylated oligosaccharide core extracellularly, may exert their pathophysiologic effects on motor nerves, at a site readily accessible to antibody binding. GM 1 and GQ 1b

5 S148 Willison et al. JID 1997;176 (Suppl 2) Table 1. Name of first author, year of publication, and reference of electrophysiologic studies using defined anti-gm1 and anti-gq1b antisera. Study Physiologic system Antibody Mode of delivery Summary of results Willison, 1996 [34] Isolated mouse GQ 1b, IgM (human) Passive transfer, then Decrease in EPP amplitude; hemidiaphragm organ bath increased stimulus threshold Roberts, 1994 [38] Isolated mouse GQ 1b, IgG (human) Organ bath Increase in MEPP frequency, hemidiaphragm then block of quantal release Buchwald, 1995 [39] Isolated mouse GQ 1b, IgG (human) Via macro-patch-clamp Reversible block of quantal hemidiaphragm electrode release Santoro, 1992 [40] Rat sciatic nerve (serial GM 1, IgM (human) Intraneural injection Conduction block and temporal in vivo study) dispersion Uncini, 1993 [41] Rat sciatic nerve (serial GM 1, IgM (human) Intraneural injection Conduction block and temporal in vivo study) dispersion Harvey, 1995 [42] Rat sciatic nerve (serial GM 1, IgG and IgM Intraneural injection No effects in vivo study) (human) Arasaki, 1993 [43] Isolated, desheathed rat GM 1, IgM (human) Organ bath Acute conduction block sciatic nerve GM 1 (rabbit) Takigawa, 1995 [44] Isolated, single rat GM 1 (rabbit) Via microelectrode Increased K / current and myelinated fiber decreased Na / current Roberts, 1995 [45] Isolated mouse GM 1, IgM (human) Organ bath and passive Decrease in EPP amplitude and hemidiaphragm transfer conduction block NOTE. EPP Å end plate potential; MEPP Å miniature EPP. are highly enriched in neural membranes, particularly at synapassociated 10. Yuki N, Yoshino H, Sato S, Miyatake T. Acute axonal polyneuropathy with anti-gm1 antibodies following Campylobacter enteritis. tic and nodal regions. They interact with growth factor recep- Neurology 1990;40: tors, ion channels, and cell recognition and adhesion molecules, 11. Gregson NA, Jones D, Thomas PK, Willison HJ. Acute motor neuropathy and they have widespread effects on signal transduction sys- with antibodies to GM1 ganglioside. Neurology 1991;238: tems. It seems highly reasonable that antibodies directed 12. van den Berg LH, Marrink J, de Jager AEJ, et al. Anti-GM1 antibodies in against gangliosides, having gained access to the relevant neu- patients with Guillain-Barré syndrome. J Neurol Neurosurg Psychiatry 1992;55:8 11. ral membranes, might exert pathophysiologic effects either di- 13. Nobile-Orazio E, Carpo M, Meucci N, et al. Guillain Barré syndrome rectly or indirectly via complement activation. Evidence is associated with high titers of anti-gm1 antibodies. J Neurol Sci 1992; emerging that this is so, although the precise mechanisms by 109: which this takes place remain to be clarified. 14. Pestronk A, Cornblath DR, Ilyas AA, et al. A treatable multifocal neuropa- thy with antibodies to GM1 ganglioside. Ann Neurol 1988;24: Pestronk A. Invited review: motor neuropathies, motor neuron disorders and anti-glycolipid antibodies. Muscle Nerve 1991;14: References 16. Chiba A, Kusonoki S, Shimizu T, Kanazawa I. Serum IgG antibody to ganglioside GQ1b is a possible marker of Miller Fisher syndrome. Ann 1. Yu RK, Saito M. Structure and localization of gangliosides. In: Margolis Neurol 1992;31: RU, Margolis RK, eds. Neurobiology of glycoconjugates. New York: 17. Yuki N, Sato S, Tsuji S, Ohsawa T, Miyatake T. Frequent presence of Plenum Press, 1989:1 34. anti-gq1b antibody in Fisher s syndrome. Neurology 1993;43: Tettamanti G, Riboni L. Gangliosides and the modulation of function of 18. Willison HJ, Veitch J, Paterson G, Kennedy PGE. Miller Fisher syndrome neural cells. Adv Lipid Res 1993;25: is associated with serum antibodies to GQ1b ganglioside. J Neurol 3. Ilyas AA, Willison HJ, Quarles RH, et al. Serum antibodies to gangliosides Neurosurg Psychiatry 1993;56: in Guillain Barré syndrome. Ann Neurol 1988;23: Ilyas AA, Quarles RH, Dalakas MC, Fishman PH, Brady RO. Monoclonal 4. Quarles RH, Ilyas AA, Willison HJ. Antibodies to glycolipids in demyelin- IgM in a patient with paraproteinemic polyneuropathy binds to ganglioating diseases of the human peripheral nervous system. Chem Phys sides containing disialosyl groups. Ann Neurol 1985;18: Lipids 1986;42: Herron B, Willison HJ, Veitch J, Roelcke D, Illis LS, Boulton FE. Mono- 5. Latov N. Antibodies to glycoconjugates in neurological disease. Clin As- clonal IgM cold agglutinins with anti-pr 1d specificity in a patient with pects Autoimmun 1990;4: peripheral neuropathy. Vox Sang 1994;67: Willison HJ. Antiglycolipid antibodies in peripheral neuropathy: fact or 21. Baba H, Daune GC, Ilyas AA, et al. Anti-GM1 ganglioside antibodies fiction? J Neurol Neurosurg Psychiatry 1994;57: with differing fine specificities in patients with multifocal motor neurop- 7. Hughes RAC, Rees JH. Guillain-Barré syndrome. Curr Opin Neurol 1994; athy. J Neuroimmunol 1989;25: : Ilyas AA, Willison HJ, Dalakas MC, Whittaker JN, Quarles RH. Identifi- 8. Hartung HP, Pollard JD, Harvey GK, Toyka KV. Immunopathogenesis cation and characterization of gangliosides reacting with IgM paraproand treatment of the Guillain-Barré syndrome. Muscle Nerve 1995;18: teins in three patients with neuropathy associated with biclonal gammo pathy. J Neurochem 1988;51: O Leary C, Willison HJ. Immunological investigation. Curr Opin Neurol 1995;8: Thomas FP, Lee AM, Romas SN, Latov N. Monoclonal IgMs with anti- Gal(b1-3) GalNAc activity in lower motor neuron disease; identification

6 JID 1997;176 (Suppl 2) Action of Ganglioside Antibodies in GBS S149 of glycoprotein antigens in neural tissue and cross-reactivity with serum 35. Pollard JD, Westland KW, Harvey GK, et al. Activated T-cells of nonimmunoglobulins. J Neuroimmunol 1989;23: neural specificity open the blood-nerve barrier to circulating antibody. 24. Chiba A, Kusonoki S, Obata H, Machinami R, Kanazawa I. Serum anti- Ann Neurol 1995;37: GQ1b IgG antibody is associated with ophthalmoplegia in Miller Fisher 36. Kusunoki S, Shimizu J, Chiba A, Ugawa Y, Hitoshi S, Kanazawa I. syndrome and Guillain Barré syndrome. Neurology 1993;13: Experimental sensory neuropathy induced by sensitization with ganglio- 25. Willison HJ, Veitch J. Immunoglobulin subclass distribution and binding side GD1b. Ann Neurol 1996;39: characteristics of anti-gq1b antibodies in Miller Fisher syndrome. J 37. Kitamura M, Iwamori M, Nagai Y. Interaction between Clostridium botuli- Neuroimmunol 1994;50: num neurotoxin and gangliosides. Biochem Biophys Acta 1980;628: 26. Willison HJ, Paterson G, Kennedy PGE, Veitch J. Cloning of human anti GM1 antibodies from motor neuropathy patients. Ann Neurol 1994;35: 38. Roberts M, Willison HJ, Vincent A, Newsom-Davis J. Serum factor in the Miller Fisher variant of Guillain-Barré syndrome blocks neurotrans- 27. O Hanlon GM, Paterson GJ, Wilson G, Doyle D, McHardie P, Willison HJ. mitter release. Lancet 1994;343: Anti-GM1 ganglioside antibodies cloned from autoimmune neuropathy 39. Buchwald B, Weishaupt A, Toyka KV, Dudel J. Immunoglobulin G from patients show diverse binding patterns in the rodent nervous system. J a patient with Miller-Fisher syndrome rapidly and reversibly depresses Neuropathol Exp Neurol 1996;55: evoked quantal release at the neuromuscular junction of mice. Neurosci- 28. Paterson G, Wilson G, Kennedy PGE, Willison HJ. Analysis of anti-gm1 ence Lett 1995;201: antibodies cloned from motor neuropathy patients demonstrates diverse 40. Santoro M, Uncini A, Corbo M, et al. Experimental conduction block variable region gene usage with extensive somatic mutation. J Immunol induced by serum from a patient with anti-gm1 antibodies. Ann Neurol 1995;155: ;31: Schluep M, Steck AJ. Immunostaining of motor nerve terminals by IgM 41. Uncini A, Santoro M, Corbo M, Lugaresi A, Latov N. Conduction abnor- M protein with activity against gangliosides GM1 and GD1b from a malities induced by sera of patients with multifocal motor neuropathy patient with motor neuron disease. Neurology 1988;38: and anti-gm1 antibodies. Muscle Nerve 1993;16: Thomas FP, Trojaborg W, Nagy C, et al. Experimental autoimmune neu- 42. Harvey GK, Toyka KV, Zielasek J, Kiefer R, Simonis C, Hartung HP. ropathy with anti-gm1 antibodies and immunoglobulin deposits at Failure of anti-gm1 IgG or IgM to induce conduction block following nodes of Ranvier. Acta Neuropathol 1991;82: intraneural transfer. Muscle Nerve 1995;18: Corbo M, Quattrini A, Lugaresi A, Santoro M, Latov N, Hays AP. Patterns 43. Arasaki K, Kusonoki S, Kudo N, Kanazawa I. Acute conduction block in of reactivity of human anti-gm1 antibodies with spinal cord and motor neurons. Ann Neurol 1992;32: vitro following exposure to anti-ganglioside sera. Muscle Nerve 1993; 32. Corbo M, Quattrini A, Latov N, Hays AP. Localization of GM1 and 16: Gal(b1-3)GalNAc antigenic determinants in peripheral nerve. Neurol- 44. Takigawa T, Yasuda H, Kikkawa R, Shigeta Y, Saida T, Kitasato H. ogy 1993;43: Antibodies against GM(1) ganglioside affect K / and Na / currents 33. Ganser AL, Kirschner DA, Willinger M. Ganglioside localization on my- in isolated rat myelinated nerve-fibers. Ann Neurol 1995;37:436 elinated nerve fibres by cholera toxin binding. J Neurocytol 1983;12: Roberts M, Willison HJ, Vincent A, Newsom-Davis J. Multifocal motor 34. Willison HJ, O Hanlon GM, Paterson GJ, et al. A somatically mutated neuropathy human sera block distal motor nerve conduction in mice. human anti-ganglioside antibody that induces experimental neuropathy Ann Neurol 1995;38: in mice is encoded by the variable region heavy chain gene, V1 18. J 46. Merrill AH, Hannun YA, Bell RM. Introduction: sphingolipids and their Clin Invest 1996;97: metabolites in cell regulation. Adv Lipid Res 1993;25:1 21.