Seronegative Lambert-Eaton myasthenic syndrome

Patients and methods.

The study was based on a large cohort of LEMS patients seen by a Japanese neurologist between 1992 and 1999. LEMS was diagnosed from clinical findings and electrophysiologic evidence of impaired neuromuscular transmission. One hundred and ten patients were followed, and all available information was reviewed. Clinical features were checked at least twice, at the initial diagnosis and during the follow-up period (4 months to 7 years); these included sex, age at onset, malignancy, duration of LEMS symptoms at the time of tumor detection, neurologic findings, and electrophysiological findings. All 110 patients were tested for anti-P/Q-type VGCC antibodies by an immunoprecipitation method that used ¹²⁵I-ω-conotoxin MVIIC and human cerebellar extracts, as described previously.4,5⇓ Patients with titers less than three standard deviations above the mean for the normal controls were classified as seronegative.

IgG was purified from serum samples that were obtained from two seropositive LEMS patients, two seronegative patients, and three healthy controls. We injected 10 mg of purified IgG IV into BKTO strain mice. Twenty-four to 48 hours after injection, miniature end-plate potentials (MEPP) and end-plate potentials (EPP) were recorded from the diaphragm by the conventional method. For each animal, EPPs or MEPPs were recorded from 10 or 15 different end-plates. Student’s t-test was used to compare the means of data sets that approximated a Gaussian distribution.

Results.

All the patients were Japanese (73 men and 37 women). Age at onset ranged from 17 to 80 years (mean 62 years). Carcinoma was detected in 69%, of whom 61% had SCLC. Neurologic symptoms preceded the diagnosis of malignancy in 84% of the patients. Predominant neurologic findings were lower limb weakness in 97%, hyporeflexia/areflexia in 85%, upper limb weakness in 80%, autonomic dysfunction in 37%, and blepharoptosis in 28%. Cerebellar signs were detected in 9%, all of whom had SCLC. The most common autonomic dysfunction symptom was dry mouth (table 1). Serum P/Q-type VGCC antibody titers were measured repeatedly, and 17 (15%) of the 110 LEMS patients were judged seronegative. Of these seronegative patients, nine were men and eight were women. The mean age at onset was 63 years. Two (12%) of the 17 seronegative patients had SCLC, whereas 70% of the seropositive group had it. A marked difference was observed (p < 0.001) between seronegative and seropositive groups. The frequency of neurologic findings was similar for both seronegative and seropositive patients (table 2).

In the passive transfer experiments, the MEPP amplitudes of the seropositive and seronegative patients with LEMS did not differ from those of the controls. Quantal contents of the EPPs, however, were equally decreased in both the seropositive and seronegative LEMS patients compared with the controls (table 3).

Discussion.

We know of only two reports on the epidemiology of LEMS.2,6⇓ O’Neil et al.2 described the clinical features of 50 consecutive British patients with LEMS. The analysis indicated that a patient with LEMS has a 62% risk of having an underlying SCLC. Thereafter, it was reported that 28 LEMS patients in West Virginia were retrospectively analyzed, and SCLC was found in 43%. A trend over time for a lower frequency of associated cancer was noted.6 In our study, carcinoma was detected in 69% of the LEMS patients, of whom 61% had SCLC. This observation of SCLC incidence is equivalent to that reported by O’Neill et al. Other than SCLC, leukemia or malignant lymphoma was an associated malignancy, indicative that LEMS is accompanied by lymphoproliferative disorders. After analyzing the clinical findings, we found no significant difference between the patients analyzed by O’Neill et al. and our LEMS patients, except for cranial nerve sign and autonomic dysfunction (see table 1). But cranial nerve signs were usually mild and often transient. Our LEMS patients had a lower incidence of autonomic dysfunction, but some who had an abnormal Saxon test (a quantitative test for salivary secretion) did not complain of dry mouth; i.e., they were asymptomatic. Therefore, we find no clinical difference between our LEMS patients and those reported by O’Neill et al. It is noteworthy that 9% of our LEMS patients also had cerebellar ataxia (not commented on by O’Neill), and all of these patients had SCLC and high anti-P/Q-type VGCC antibody titers. Therefore, cerebellar ataxia appears to be less common in seronegative LEMS.

We reviewed patients with seronegative myasthenia gravis (MG) to investigate the pathogenesis of seronegative LEMS. Previous studies found 10% to 20% of patients with MG had no detectable antibodies to acetylcholine receptor (AChR).7 The clinical features of the seronegative patients did not differ from seropositive patients, except for the indication that thymomas are less common in the seronegative patients.8 Furthermore, the passive transfer study results suggest that seronegative MG is an autoantibody-mediated disorder of neuromuscular transmission.9 The target antigens and mechanism of action in such seronegative patients are still unknown. Recently, it was reported that 70% of the AChR-Ab-seronegative but none of the AChR-Ab-seropositive MG patients had serum auto-antibodies against the muscle-specific receptor tyrosine kinase, which mediates agrin-induced clustering of AChR during synapse formation and is also expressed at the mature neuromuscular junction.10

No anti-P/Q-type VGCC antibodies were detected in 15% of our LEMS patients. Neurologic findings were similar for the seropositive and seronegative groups. The incidence of SCLC was lower for seronegative than seropositive LEMS patients. This agrees with the findings for seronegative MG as to the low incidence of thymic pathology. Autoimmune rather than paraneoplastic factors may have a major role in both seronegative MG and seronegative LEMS. In our passive transfer study, the quantal contents of EPP were decreased in mice injected with IgG from seropositive or seronegative LEMS patients, although further studies using more seronegative IgG samples will be required. These findings are a strong indication that seronegative LEMS is an antibody-mediated autoimmune disorder. There are several possible explanations for seronegative LEMS patients: (1) The immunoprecipitation assay done with ¹²⁵I-ω-conotoxin MVIIC does not detect antibodies directed specifically against the toxin-binding site, and therefore patients with antibodies to that site would appear to be seronegative. In a previous study, we investigated ω-conotoxin MVIIC-blocking antibodies in LEMS patients, but none of them had either negative-binding or positive-blocking antibodies. We speculate that patients who have only these blocking antibodies are extremely rare; (2) Antibodies bind to P/Q-type VGCC, but they may be directed at epitopes not present in the soluble VGCC extract or may have too low an affinity for their detection by the soluble assay system. Steroid therapy may decrease antibody titer, but it was used for only one of our seronegative patients; and (3) Antibodies that bind to molecules other than P/Q-type VGCC could be responsible for impaired neuromuscular transmission. Studies of cultured cell lines are needed to identify the exact targets of seronegative LEMS.