Clinical, serological and genetic predictors of response to immunotherapy in anti-IgLON5 disease

Abstract

Anti-IgLON5 disease is a newly defined clinical entity characterized by a progressive course with high disability and mortality rate. While precise pathogenetic mechanisms remain unclear, features characteristic of both autoimmune and neurodegenerative diseases were reported. Data on immunotherapy are limited, and its efficacy remains controversial. In this study, we retrospectively investigated an anti-IgLON5 disease cohort with special focus on clinical, serological and genetic predictors of the immunotherapy response and long-term outcome. Patients were recruited from the GENERATE (German Network for Research on Autoimmune Encephalitis) registry. Along with clinical parameters, anti-IgLON5 immunoglobulin (Ig)G in serum and CSF, anti-IgLON5 IgG1-4, IgA and IgM in serum, neurofilament light chain and glial fibrillary acidic protein in serum as well as human leukocyte antigen-genotypes were determined. We identified 53 patients (symptom onset 63.8 ± 10.3 years, female:male 1:1.5). The most frequent initial clinical presentations were bulbar syndrome, hyperkinetic syndrome or isolated sleep disorder [at least one symptom present in 38% (20/53)]. At the time of diagnosis, the majority of patients had a generalized multi-systemic phenotype; nevertheless, 21% (11/53) still had an isolated brainstem syndrome and/or a characteristic sleep disorder only. About one third of patients [28% (15/53)] reported subacute disease onset and 51% (27/53) relapse-like exacerbations during the disease course. Inflammatory CSF changes were evident in 37% (19/51) and increased blood-CSF-barrier permeability in 46% (21/46). CSF cell count significantly decreased, while serum anti-IgLON5 IgG titre increased with disease duration. The presence of human leukocyte antigen-DRB1*10:01 [55% (24/44)] was associated with higher serum anti-IgLON5 IgG titres. Neurofilament light chain and glial fibrillary acidic protein in serum were substantially increased (71.1 ± 103.9 pg/ml and 126.7 ± 73.3 pg/ml, respectively). First-line immunotherapy of relapse-like acute-to-subacute exacerbation episodes resulted in improvement in 41% (11/27) of patients and early initiation within the first 6 weeks was a predictor for therapy response. Sixty-eight per cent (36/53) of patients were treated with long-term immunotherapy and 75% (27/36) of these experienced no further disease progression (observation period of 20.2 ± 15.4 months). Long-term immunotherapy initiation during the first year after onset and low pre-treatment neurofilament light chain were significant predictors for a better outcome. In conclusion, subacute disease onset and early inflammatory CSF changes support the primary role of autoimmune mechanisms at least at initial stages of anti-IgLON5 disease. Early immunotherapy, prior to advanced neurodegeneration, is associated with a better long-term clinical outcome. Low serum neurofilament light chain at treatment initiation may serve as a potential biomarker of the immunotherapy response.