MG indicates myasthenia gravis.
Elena Cortés-Vicente, Eduard Gallardo, María Ángeles Martínez, Jordi Díaz-Manera, Luis Querol, Ricard Rojas-García, Isabel Illa. Clinical Characteristics of Patients With Double-Seronegative Myasthenia Gravis and Antibodies to Cortactin. JAMA Neurol. 2016;73(9):1099–1104. doi:10.1001/jamaneurol.2016.2032
Double-seronegative myasthenia gravis (dSNMG) includes patients with myasthenia gravis (MG) without detectable antibodies to the nicotinic acetylcholine receptor (AChR) or to muscle-specific tyrosine kinase (MuSK). The lack of a biomarker hinders the diagnosis and clinical management in these patients. Cortactin, a protein acting downstream from agrin/low-density lipoprotein receptor–related protein 4 (LRP4)/MuSK, has been described as an antigen in dSNMG.
To describe the frequency and clinical features of patients with dSNMG who have cortactin antibodies.
Design, Setting, and Participants
A retrospective cross-sectional study was conducted at Hospital de la Santa Creu i Sant Pau, an institutional practice referral center in Barcelona, Spain, between May 1, 2015, and November 30, 2015. We included 250 patients with a definitive diagnosis of MG with available serum samples at the time of diagnosis. Descriptive and comparative data analyses were performed.
Cortactin antibodies were measured by enzyme-linked immunosorbent assay and Western blot; AChR, MuSK, and anti–striated muscle antibodies were detected using a standard method; and LRP4 antibodies were tested using a cell-based assay.
Main Outcomes and Measures
The primary outcome was the frequency of patients with dSNMG who have cortactin antibodies. Secondary outcomes were demographic, clinical, neurophysiological, and laboratory data.
Of 250 patients (mean [SD] age at onset, 49.7 [21.2] years; 56% female), 38 (15.2%) had dSNMG, 201 (80.4%) had MG with AChR antibodies, and 11 (4.4%) had MG with MuSK antibodies. Cortactin antibodies were identified in 28 patients with MG: 9 of 38 (23.7%) who had dSNMG, 19 of 201 (9.5%) who had MG with AChR antibodies (significantly lower than those with dSNMG: 9.5% vs 23.7%; P = .02), and 0 of 11 who had MG with MuSK antibodies; 0 of 29 controls had cortactin antibodies. At onset, among the 9 patients with dSNMG and cortactin antibodies, 6 had ocular MG and 3 had Myasthenia Gravis Foundation of America clinical classification IIA. Two patients with ocular MG developed generalized MG. The group with dSNMG and cortactin antibodies, compared with those who had MG with AChR antibodies, more frequently had mild forms at onset (100.0% vs 62.7%; P = .03), had fewer bulbar signs at maximal worsening (0% vs 41.3%; P = .01), and were younger at onset (median [interquartile range], 34.9 [9.5] vs 53.9 [38.5] years; P = .03); the group with dSNMG and cortactin antibodies also more frequently had ocular MG at onset than those with MG and AChR antibodies, although the difference was not statistically significant (66.7% vs 40.8%; P = .17). Of 17 patients with ocular dSNMG, 4 (23.5%) had antibodies to cortactin.
Conclusions and Relevance
In this study, patients with cortactin antibodies and dSNMG had an ocular or mild generalized phenotype of MG. Including the detection of cortactin antibodies in the routine diagnosis of dSNMG may be helpful in ocular MG.
Myasthenia gravis (MG) is an autoimmune antibody-mediated disease of the neuromuscular junction characterized by fatigable muscle weakness. Autoantibodies against the nicotinic acetylcholine receptor (AChR) are detectable in up to 80% of cases of generalized MG and are present in only 50% of patients with ocular-restricted MG.1- 5 Antibodies to muscle-specific tyrosine kinase (MuSK) are detected in around 5% of patients with generalized MG. In these patients, the disorder may be clinically different from MG with AChR antibodies, being characterized by early facial, bulbar, neck, and respiratory weakness.6- 10
In around 15% of patients with generalized MG and in up to 50% of patients with ocular forms of the disease, no antibodies against AChR and MuSK are detected. These cases are identified as double-seronegative MG (dSNMG). The clinical presentation of this subgroup of MG is similar to MG with AChR antibodies in terms of muscle weakness distribution, disease severity, and response to immunotherapy and plasma exchange, suggesting that it is an immune-mediated disorder of the neuromuscular junction. The lack of a detectable pathogenic antibody or an autoimmune biomarker in dSNMG has hindered the diagnosis and clinical management in these patients, especially in ocular forms of the disease.11- 13
Double-seronegative MG encompasses a heterogeneous group of patients with antibodies to different neuromuscular junction proteins. Antibodies against low-density lipoprotein receptor–related protein 4 (LRP4) has been described in subgroups of patients with dSNMG (2%-27%).14- 16 Low-affinity IgG1 AChR antibodies binding clustered AChRs have been found in 38% of patients with dSNMG in a cellular assay using human embryonic kidney cells.17,18 The finding of these autoantibodies, regardless of their pathogenic role in vivo, indicates an ongoing immune process and supports the initiation of immunotherapy. However, a subgroup of patients still has unidentified antibodies to an antigen that is as yet unknown.
Our group recently described a new target antigen, cortactin, in 19.7% of patients with dSNMG.19 Cortactin was found using a human protein array approach. It was considered a good candidate because it is concentrated at the neuromuscular junction and acts downstream from agrin/LRP4/MuSK, promoting AChR clustering.20,21 Simultaneously, cortactin antibodies were discovered using mass spectrometry analysis and were found by enzyme-linked immunosorbent assay in 20% of patients with polymyositis.22 Overall, these antibodies were reported in around 10% of patients with other autoimmune diseases.19,22
The aim of this study is to describe the results of testing the initial serum sample (the serum sample obtained at the first visit) for cortactin antibodies in our series of patients with a definitive diagnosis of MG. We report the MG phenotype of patients with these autoantibodies and discuss its usefulness in the clinical setting.
Question What are the frequency and clinical features of patients who have double-seronegative myasthenia gravis (dSNMG) with cortactin antibodies?
Findings In this cross-sectional study of 250 patients, cortactin antibodies were identified in 9 of 38 patients with dSNMG (23.7%). These patients had an ocular or mild generalized phenotype of myasthenia gravis.
Meaning Including the detection of cortactin antibodies may be helpful in the routine diagnosis of dSNMG.
This was a retrospective cross-sectional study conducted between May 1, 2015, and November 30, 2015. We selected patients from the neuromuscular unit at Hospital de la Santa Creu i Sant Pau, an institutional practice referral center Barcelona, Spain, who fulfilled the criteria for definitive MG and had available serum samples at the time of diagnosis. Serum samples from healthy controls were also included in the cortactin antibody assays.
Myasthenia gravis was diagnosed by a consultant neurologist based on compatible clinical features together with 1 or more of the following criteria: (1) positive results on an AChR or MuSK antibody assay; (2) electrophysiological study findings compatible with a postsynaptic neuromuscular junction disorder (repetitive stimulation, single-fiber electromyography, or both); and (3) a response to cholinesterase inhibitors. In dSNMG, the diagnosis was always confirmed by abnormal findings on a neurophysiological study.
Written informed consent was obtained from all patients. The study was approved by the ethics committee at Hospital de la Santa Creu i Sant Pau.
The main clinical features of the disease were reviewed: weakness distribution and severity according to the Myasthenia Gravis Foundation of America clinical classification23; presence of thymoma by thoracic computed tomography or thymus pathology in patients undergoing thymectomy; and treatment required (no treatment or cholinesterase inhibitors, and immunosuppressive therapy). We also recorded demographic data (sex, date of birth, age at onset) as well as positive results on assays for AChR, MuSK, cortactin, LRP4, and anti–striated muscle antibodies. Testing for low-affinity anti-AChR antibodies was not available.
Antibodies against AChR and MuSK were tested according to the manufacturer’s instructions by radioimmunoassay in serum samples collected from all patients at diagnosis. Cortactin antibodies were measured using the in-house enzyme-linked immunosorbent assay, and the results were confirmed by Western blot using a purified recombinant protein cortactin (OriGene) as previously described.19 A positive control and a negative control were prediluted from 200 to 6.25 standard units and were used as calibrators. A sample was considered positive when the value was at least 20 standard units. Values from 12.5 to 20 standard units were considered moderately positive and were always analyzed by Western blot to confirm positivity. Serum samples were tested for antibodies against the nonrelated recombinant protein TIF1γ (OriGene) to confirm the specificity of the assay. Antibodies against LRP4 were tested using a transfected human embryonic kidney 293 cell–based test as previously described.19 Anti–striated muscle antibodies were detected using standard immunofluorescence on skeletal muscle (primate) slides (Inova Diagnostics Inc) following the manufacturer’s instructions.
A descriptive data analysis was performed. The frequencies of symptoms are reported as percentages. Demographic characteristics are reported as medians and interquartile ranges. Differences in baseline characteristics between patient subgroups were evaluated using Fisher exact test when comparing categorical variables and Mann-Whitney U test when comparing quantitative variables. Data analysis was carried out using Stata for Windows version 13.0 statistical software (StataCorp LP).
We studied 250 patients (mean [SD] age at onset, 49.7 [21.2] years; 56% female) who fulfilled the inclusion criteria, including 38 (15.2%) with dSNMG, 201 (80.4%) with MG and AChR antibodies, and 11 (4.4%) with MG and MuSK antibodies (Figure). Serum samples from 29 healthy individuals were analyzed as controls.
We identified a total of 28 patients with MG who had cortactin antibodies. Patients with dSNMG had cortactin antibodies with a frequency of 23.7% (9 of 38 patients). No LRP4 and anti–striated muscle antibodies were detectable in this group of patients. No TIF1γ antibodies were detected. Cortactin antibodies were detected in 19 of 201 patients who had MG with AChR antibodies, a significantly lower rate than in patients with dSNMG (9.5% vs 23.7%; P = .02). Of 11 patients who had MG with MuSK antibodies, none had cortactin antibodies (Figure). No healthy controls had antibodies against cortactin (P = .004). Of 17 patients with ocular dSNMG, 4 (23.5%) had antibodies to cortactin.
Patients with dSNMG and cortactin antibodies (n = 9) had a median age at onset of 34.9 years (interquartile range, 9.5 years), and 77.8% were female. Of these patients, 6 presented with restricted ocular forms and 3 presented with mild generalized forms (Myasthenia Gravis Foundation of America clinical classification IIA) at onset. During the follow-up, 2 of the 6 patients with ocular MG developed a generalized disease. No patients presented with bulbar signs; severe forms and admission at the intensive care unit; or thymoma. Immunosuppressive therapy was required in 55.6% of the patients.
Patients with dSNMG and cortactin antibodies, compared with patients who had MG with AChR antibodies, more frequently had mild forms (I and IIA) at onset (100.0% vs 62.7%; P = .03), less frequently had bulbar signs at maximal worsening (0% vs 41.3%; P = .01), and were significantly younger at onset (median [interquartile range], 34.9 [9.5] vs 53.9 [38.5] years; P = .03). They also more frequently had restricted ocular forms at onset, although this difference was not statistically significant (66.7% vs 40.8%; P = .17). Although not statistically significant, the group of patients with dSNMG and cortactin antibodies, compared with those who had MG with AChR antibodies, had lower rates of immunosuppressive therapy required (55.6% vs 77.1%; P = .22), generalized disease development (33.3% vs 53.7%; P = .42), and presence of thymoma (0% vs 14.4%; P = .62) (Table 1). When adjusting for age and comparing patients younger than 50 years in both groups, patients with dSNMG and cortactin antibodies, compared with those who had MG and AChR antibodies, had a higher frequency of ocular MG (at onset: 75.0% vs 30.4%; P = .02; at maximal worsening: 50.0% vs 13.0%; P = .02) and fewer bulbar symptoms at maximal worsening (0% vs 34.8%; P = .05) (Table 2).
Among patients with dSNMG, the subgroup with cortactin antibodies compared with the subgroup without them were younger (median [interquartile range], 34.9 [29.7-39.2] vs 49.4 [35.2-61.7] years; P = .03). Although not statistically significant, the subgroup with cortactin antibodies had milder forms at onset (9 of 9 patients [100%] vs 22 of 29 patients [75.9%]; P = .16) and less bulbar involvement (0 of 9 patients vs 10 of 29 patients [34.5%]; P = .08). No differences were found in treatment required (5 of 9 patients [55.6%] vs 18 of 29 patients [62.1%]; P > .99).
When comparing the subgroup of patients with MG who had both AChR and cortactin antibodies with patients who had MG and AChR antibodies alone, we did not find differences in sex (female: 11 of 19 patients [57.9%] vs 94 of 182 patients [51.6%]; P = .64), age at onset (median [interquartile range], 50.6 [22.6-73.2] vs 54.1 [31.2-69.3] years; P = .75), restricted ocular forms (6 of 19 patients [31.6%] vs 76 of 182 patients [41.8%]; P = .47), mild forms (I and IIA) (10 of 19 patients [52.6%] vs 116 of 182 patients [63.7%]; P = .46), bulbar symptoms at maximal worsening (7 of 19 patients [36.8%] vs 76 of 182 patients [41.8%]; P = .81), presence of thymoma (2 of 19 patients [10.5%] vs 27 of 182 patients [14.8%]; P > .99), or treatment required (14 of 19 patients [73.7%] vs 141 of 182 patients [77.5%]; P = .78).
Cortactin antibodies were detected in a subgroup of patients with dSNMG. These patients were younger than those who had MG with AChR antibodies and those who had dSNMG without cortactin antibodies, and they had a predominance of ocular or mild generalized clinical forms and no bulbar signs. The MuSK, LRP4, and anti–striated muscle antibodies were not associated with cortactin antibodies. Some patients with MG and AChR antibodies had cortactin antibodies, but at a significantly lower rate than those with dSNMG (P = .02). No healthy controls had cortactin antibodies.
Double-seronegative MG has classically been considered similar to MG with AChR antibodies in terms of clinical features and response to treatment. Classifying patients with MG into subgroups according to serological criteria has implications in diagnosis, prognosis, and therapy.1 The finding of MuSK antibodies among AChR-seronegative patients with MG defined a subgroup with predominant involvement of cranial and bulbar muscles, with an excellent and maintained response to prednisone and other immune therapies, including rituximab.9 Among patients with dSNMG, those with LRP4 antibodies have a female preponderance and predominating ocular or mild generalized MG.1 These antibodies have a low incidence, and they could coexist with MuSK antibodies. The clinical characteristics of patients with low-affinity antibodies to clustered AChR have recently been described.24 Patients have early onset and milder disease with predominantly isolated ocular symptoms and a low generalization rate, but 25% present with mild bulbar symptoms. A commercial test to measure these antibodies is not yet available, so these subgroups can be identified in only a few centers. Our results confirm that patients with dSNMG and cortactin antibodies present with predominantly ocular or mild clinical forms of the disease, and interestingly, we did not find bulbar signs initially or during follow-up. These findings suggest that cortactin antibodies may imply good prognosis in patients with dSNMG.
It is difficult to diagnose the ocular forms of MG at an immunological level because, in contrast with the generalized forms, up to 50% of patients with ocular MG are seronegative for AChR antibodies. In our series, as up to 24% of patients with ocular dSNMG had antibodies to cortactin, the detection of these antibodies constituted a useful biomarker especially in this group of patients. A biomarker of disease in ocular MG is especially relevant because the differential diagnosis of ptosis and diplopia includes several diseases, the pharmacological test may be confusing, and results on electrophysiological studies may be borderline or normal in some patients.
Cortactin antibodies were found in 9.5% of patients with MG and AChR antibodies, but the clinical features of this subgroup of patients were similar to those found in patients with MG and only AChR antibodies. This suggests that AChR antibodies dominate when found in coexistence with cortactin antibodies. In this context, the good prognostic value of cortactin antibodies would not be applicable. We did not test the serum samples of these patients for low-affinity IgG1 AChR antibodies, but we would expect to find them at similar rates to those found for high-affinity AChR antibodies. We found that MuSK, LRP4, and anti–striated muscle antibodies were not associated with cortactin antibodies in our series.
Cortactin is a postsynaptic neuromuscular junction intracellular protein that acts downstream from agrin/LRP4/MuSK, promoting clustering of AChR. Because of this important role in neuromuscular transmission, cortactin has been considered as a potential antigen in the pathogenesis of MG. While the pathogenicity of antibodies to AChR was long ago proven and the pathogenic role has also been shown for MuSK antibodies,5 it remains to be determined how antibodies to cortactin are generated and whether they are pathogenic. The fact that most patients had no other autoantibodies suggests that the cortactin antibodies did not arise as a consequence of epitope spreading or neuromuscular junction damage. However, the presence of cortactin antibodies has been described in other neuromuscular autoimmune disorders19,24 and also in around 5% of healthy controls.19 This suggests that they may not be involved in the pathogenesis of MG, but they should be considered a good biomarker in dSNMG.
In this study, we confirm that cortactin antibodies are significantly prevalent in dSNMG, a clinical entity in which the absence of a diagnostic marker may hinder the clinical management of patients. For this reason, although its pathogenic role has not yet been proven, the finding of cortactin antibodies in patients with dSNMG is a useful clinical tool and supports the use of immunosuppressive therapies.
The main limitation of the study is that although we tested a wide sample of serum samples from patients with MG, the group of patients with dSNMG and cortactin antibodies is small. In a larger sample of patients with dSNMG, we might be able to reach statistical significance in terms of sex, treatment required, generalized disease development, or presence of thymoma, as we found clear differences in these variables between both groups.
We consider that cortactin antibodies are a biomarker in a significant proportion of patients with dSNMG, especially those with ocular dSNMG. In addition, the presence of cortactin antibodies supports the diagnosis of autoimmune MG and, if clinically required, treatment with immunosuppressive therapies.
Corresponding Author: Isabel Illa, MD, PhD, Neuromuscular Diseases Unit, Department of Neurology, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Sant Antoni Maria Claret 167, Barcelona 08025, Spain (firstname.lastname@example.org).
Accepted for Publication: May 5, 2016.
Published Online: July 5, 2016. doi:10.1001/jamaneurol.2016.2032.
Author Contributions: Dr Illa had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Cortés-Vicente, Illa.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Cortés-Vicente, Illa.
Critical revision of the manuscript for important intellectual content: Gallardo, Martínez, Díaz-Manera, Querol, Rojas-García, Illa.
Statistical analysis: Cortés-Vicente, Illa.
Obtained funding: Illa.
Administrative, technical, or material support: Martínez.
Study supervision: Gallardo, Rojas-García, Illa.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by grant FIS 13/0937 from Instituto de Salud Carlos III Ministry of Economy and Innovation (principal investigator, Dr Illa). Dr Querol is supported by grant JR13/00014 from Instituto de Salud Carlos III.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Cándido Juárez, PhD, Department of Immunology, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain, provided support and helpful advice; he received no compensation.