Multifocal motor neuropathy

Multifocal motor neuropathy (MMN) is a slowly progressive disorder of peripheral nerves that usually presents with asymmetric distal weakness in the upper extremities.^1-3 Electrodiagnostic studies show focal blockade of impulse conduction along motor axons.^1-3 Serum IgM antibodies to GM1 ganglioside are common in MMN, reported in 30 to 63% of patients.^2,4-7 The target of serum antibodies, if any, in the remaining patients has been unclear. We previously found that IgM in MMN patient sera binds to GM1 as a component in a lipid mixture (with galactocerebroside and cholesterol sulfate [GGC]) more commonly than to GM1 alone.⁸ In this study, we examined the frequency of binding of serum IgM to pure GM1 ganglioside that was attached covalently to ELISA plates containing secondary amino groups(Co-GM1). We examined sera from patients with MMN and compared the results with those in ALS and other control groups.

Methods. Serum samples. We tested sera from 27 consecutive patients with MMN^1,2 who had been examined by one of the authors (A.P.). As control subjects, we tested 215 other sera. These included 117 patients examined at the Neuromuscular Center at Washington University, St. Louis, who met accepted criteria for: chronic inflammatory demyelinating polyneuropathy (CIDP) (22 patients), demyelinating Guillain-Barré syndrome (GBS) (22 patients), ALS (clinically definite by World Federation of Neurology criteria) (22 patients), asymmetric distal lower motor neuron syndromes without conduction block⁹(10 patients), systemic immune disorders without neurologic involvement (22 patients), and idiopathic sensory-motor polyneuropathies (19 patients). We also evaluated serum from patients with ALS included in a treatment trial of ciliary neurotrophic factor (67 patients) and Chinese patients with acute immune neuropathies (41 patients) (including subgroups with axonal [29 patients] and demyelinating [14 patients] findings on electrophysiologic studies), who were tested for IgG anti-GM1 antibodies as part of one of our prior studies.¹⁰

ELISA antibody assay. Sera were assayed for IgM binding to purified glycolipids using conventional ELISA plates as previously described^8,9,11 or Nunc CovaLink NH microwell plates (Nunc, Roskilde, Denmark). In conventional ELISA plates(Immulon 2, Dynatech, Chantilly, VA), individual lipid antigens, including GM1 ganglioside (0.15 µg) and GD1a ganglioside (0.15 µg) (Sigma), in 50 µl of methanol were added to wells and evaporated to dryness. For the GGC lipid mixture (GM1 ganglioside, 0.15 µg; galactocerebroside, 1.5µg; cholesterol, 1.5 µg), individual lipids were dissolved in methanol, mixed in appropriate proportions, evaporated to dryness, reconstituted in phosphate-buffered saline (PBS), pH 7.2 (0.01 M with 0.15 M NaCl; 100 µL), placed in wells, incubated overnight at 4 °C, and washed three times with PBS-0.05% Tween 20. For plates containing secondary amino groups that covalently bind polysaccharide groups such as those on glycolipids, GM1 ganglioside (0.15 µg) and GD1a ganglioside (0.15µg) were dissolved in 100 µL of 1 N-(3-dimethylaminopropyl)-N-ethyl-carbodiimide (Sigma), incubated in wells overnight at 4 °C, and washed three times with PBS-0.05% Tween 20.

Remaining binding sites in all ELISA wells were blocked with 1% human serum albumin in PBS (100 µL) for 4 hours at room temperature. Plates were then washed three times with 1% bovine serum albumin (BSA) in PBS. Subsequent steps were performed at 4 °C. Between steps, washing (five times) was performed using PBS with 1% BSA without detergent. All sera were tested in duplicate. Serum was tested by adding dilutions (1:10³ to 1:10⁶ in PBS with 1% BSA) to wells overnight. The binding of IgM was measured using 4-hour exposure to goat anti-human IgM linked to horseradish peroxidase (Organon Teknika-Cappel, West Chester, PA) in PBS with 1% BSA(1:20,000). Color was developed with 100 µL substrate buffer (0.1 M citrate buffer, pH 4.5 with 0.004% H₂O₂ and 0.1 o-phenylenediamine) for 30 minutes. Optical density (OD) was determined at 450 nm. A serum antibody with a titer of x was detectable(>0.05 OD units over control subjects) up to a dilution of at least 1/x. The titer of selective serum IgM binding to GM1 ganglioside was then calculated by subtracting the level of IgM binding to GD1a in the serum. Based on initial control studies, we designated high titers of selective IgM anti-GM1 binding as those greater than 1,800.

Statistics. Antibody titers are presented to two significant figures. The significance of differences between groups was calculated using the Mann-Whitney rank sum test.

Results. Sera from 85% of MMN patients (23/27) had high titers of selective IgM binding to GM1 ganglioside (>1,800) when measured using Co-GM1 ELISA methods (table). This is significantly greater (p < 0.001) than the percentages of the same MMN sera found to have selective IgM binding with testing against GM1 (37%; 10/27) or the GGC lipid mixture (52%; 14/27) on standard ELISA plates. Nine of 27 sera had high IgM anti-GM1 titers using Co-GM1 methods (range, 1,900 to 13,000) but were negative using the other two types of testing. In addition to being detected more frequently with Co-GM1 ELISA methods, the titers of IgM versus GM1 in MMN patients were generally higher (p< 0.001), averaging 31,000 ± 15,000, compared with 7,200 ± 4,400 with GM1 antibodies measured using standard plates and 3,600 ± 1,300 with the GGC antigen.

IgM anti-GM1 antibodies were also present in 30% of patients (3/10) with slowly progressive lower motor neuron syndromes with no conduction block but a clinical phenotype similar to MMN. One other group, Chinese patients with acute immune neuropathies, also frequently (35%) had high titers of selective IgM binding to GM1 ganglioside. IgM anti-GM1 titers were high in a similar proportion of the subgroups with predominantly axonal changes (38%; 11/29) and with demyelination (29%; 4/14). Titers in the group of Chinese patients with acute immune neuropathies were lower than those in MMN, averaging 2,100 ± 520. Only 2% of sera (1/43) had a titer above 10,000 compared with 44% of MMN patients (12/27; p < 0.001). No serum from US patients with GBS, CIPD, ALS, sensory-motor polyneuropathy, or systemic autoimmune disorders had high titers (>1,800) of selective IgM binding to GM1 ganglioside.

In a final experiment, we tested the effects of covalently linking antigen to ELISA plates on the measurement of another antiglycolipid antibody. Covalent binding of sulfatide to ELISA plates eliminated, rather than increased, specific IgM binding to that antigen. Sera with selective IgM anti-sulfatide titers of 5,400, 5,600, 7,800, and 2,800 detected using conventional ELISA plates showed no selective IgM binding to sulfatide covalently linked to plates.

Discussion. The Co-GM1 ELISA method of testing for selective high-titer IgM binding to GM1 ganglioside was positive in 85% of our patients with MMN (23/27). This is a significantly greater (p < 0.001) prevalence of serum autoantibodies in MMN patients than was detected using other methods. We found that 33% of MMN patient sera (9/27) were negative using the other two assay methods but had high titers of selective IgM binding to GM1 when tested using ELISA with Co-GM1. One patient (no. 1) had high titers of IgM binding to the GM1-containing lipid mixture but not to Co-GM1 or to pure GM1 on standard plates, bringing the overall prevalence of disease-specific positive serum testing in MMN to 89%. The increased frequency of anti-GM1 antibodies detected using Co-GM1 methods seems likely to be related to the configuration of the GM1 ganglioside bound covalently to amine groups in the ELISA well. Further studies are required to determine whether the antigenic epitope(s) is all related to the carbohydrate moiety, or as suggested in our previous studies,^8,12 the lipid environment of the GM1 carbohydrates also plays a role in antibody binding.

The 89% frequency of occurrence of selective serum IgM antibodies to Co-GM1 or the GGC lipid mixture in MMN, at titers with specificity for the disease, is much higher than that found in studies from other laboratories, in which the prevalence has ranged from 30 to 63%.^5-7 The wide disparity among laboratories in the prevalence of serum IgM antibodies emphasizes the need for clinical validation of testing methods. Laboratories without clinical validation of their anti-GM1 methods, or detecting a low prevalence of anti-GM1 antibodies in MMN, should not offer the test as a commercial service, because it is then not useful in the clinical evaluation of patients with motor neuropathy.

Using Co-GM1 as antigen, the ELISA titers of IgM anti-GM1 antibodies in MMN were often quite high, averaging above 30,000 and with 12 of 27 patient sera above 10,000. MMN patients tested by standard ELISA methods averaged 23% of the titers found using Co-GM1. The Chinese acute immune neuropathies, the only clinically different group with any sera having titers of IgM versus GM1 above 1,800, averaged about 7% of the MMN level at 2,100 ± 520. IgM anti-GM1 antibodies were even present in Chinese patients (29%; 4/14) with demyelinating GBS-like disorders. This is in contrast (p = 0.017) to the GBS patients in the United States in whom IgM anti-GM1 antibodies were not detected (0/22). This disparity suggests that GBS-like syndromes may have varying underlying immune pathogenesis, or at least predisposing events, in different geographic locations.

Our Co-GM1 method is specific, in addition to being sensitive, for the diagnosis of MMN. This adds to the distinction between MMN and CIDP. The presence of serum anti-GM1 antibodies suggests that a patient with a predominantly demyelinating neuropathy has MMN, which rarely responds to prednisone, rather than CIDP, which often does. The limited spectrum of occurrence of serum IgM anti-GM1 antibodies in other disorders suggests that these antibodies are a marker of immune motor neuropathies. IgM anti-GM1 antibodies were present in 35% of patients with acute motor neuropathies in the Chinese population, whether demyelinating or axonal, and in 30% of patients with slowly progressive lower motor neuron disorders without conduction block. In contrast, IgM anti-GM1 antibodies were never found in 80 sera from patients with another motor disorder, ALS, which is probably not immune mediated.

The high sensitivity (89%) and specificity of serum IgM anti-GM1 antibodies detected using Co-GM1 and GGC methods makes them useful in the workup of motor neuropathy syndromes. Our results are comparable with those for anti-acetylcholine receptor antibodies, which also have approximately 90% sensitivity and high specificity for their associated disorder, myasthenia gravis.¹³ In general, the combination of antibody testing and electrophysiologic evaluation in myasthenia gravis and MMN adds certainty to the diagnosis. Electrophysiologic testing in search of motor conduction block during the workup of MMN is often time consuming and, in some cases with severely affected nerves or no clear findings of demyelination, difficult to interpret. Anti-GM1 antibody testing, with the high sensitivity and specificity of the Co-GM1 and GGC methods, can clarify the diagnosis, reduce the time and expense of electrophysiologic evaluation, and suggest which therapeutic interventions are appropriate.^2,4 This is important because immune disorders are often treated with modalities that are either very expensive or associated with significant side effects.^2,13 Negative antibody testing early in the course of patients with motor neuron disorders but no, or only equivocal, upper motor neuron signs should be useful in reducing the suspicion of MMN, especially when electrodiagnostic testing fails to detect conduction block or other suggestions of demyelination. High-titer IgM anti-GM1 antibodies detected by Co-GM1 or GGC methods suggest a diagnosis of MMN, or a related motor neuropathy, and reduce the likelihood of other types of immune neuropathies or ALS.