JOURNAL OF NEUROLOGY, NEUROSURGERY AND PSYCHIATRY

Mariotto S, Ferrari S, Gastaldi M, Franciotta D, et. al.

J Neurol Neurosurg Psychiatry. 2019 Apr 5

DOI: 10.1136/jnnp-2018-320287

Introduction

The identification of myelin oligodendrocyte glycoprotein antibodies (MOG- Ab) in sera of subjects with inflammatory central nervous system (CNS) conditions has improved the differential diagnosis between multiple sclerosis (MS) and anti- body-mediated demyelinating disorders.

However, the heterogeneous clinical course of MOG-Ab-positive cases complicates outcome prediction. Cerebrospinal fluid (CSF) and serum neurofilament light chain (NfL) levels reliably predict disease activity in patients with clinically isolated syndrome and MS. 

We assessed serum and CSF NfL concentration in patients with MOG-Ab-associated conditions according to clinical/paraclinical characteristics to investigate NfL as a  iomarker of disease severity.

Materials and methods

Between 2015 and 2018, we included 38 consecutive MOG-Ab-positive patients tested at our institution and 38 age-matched unaffected controls. Clinical data at sampling and at last follow-up, MRI and CSF findings were collected in standardised case report forms. Brain/spinal cord MRIs were obtained within 2 months from blood drawn for  OG-Ab/NfL analysis and assessed by referring neurologists blinded to assay results.

Blood samples were collected, centrifuged and stored at −80°C. Presence of serum AQP4-Ab was analysed using a commercial cell-based assay (Euroimmun, Lübeck,  Germany). Two independent investigators analysed MOG-Ab using recombinant live cell-based immunofluorescence assay with HEK293A cells transfected with full-length MOG (human MOG alpha-1 EGFP fusion protein), as previously described.

Investigators blinded to clinical data performed NfL analysis in all sera and  when available, in CSF samples. Sera/CSF of MOG-Ab-positive cases were obtained during the index attack, with the exception of two patients asymptomatic at sampling. Measurement of NfL concentration was analysed in duplicates usingSIMOA Nf-light kit with SR-X immunoassay analyser (Quanterix, Boston, Massachusetts, USA) which runs ultrasensitive paramagnetic bead-based enzyme-linked immunosorbent assays.

Statistical analysis

Continuous/categorical variables were reported as median (range) and percentages, respectively. Mann-Whitney U test and Wilcoxon signed-rank test were used for pairwise comparison of NfL levels within and between groups, as appropriate according to matching. Correlations were analysed computing the Spearman coefficient. Receiver operating characteristic (ROC) analysis was performed to assess the association between NfL concentration and outcomes of interest, as appropriate. Statistical significance was set at alpha <0.05, two-tailed. Analyses were performed using SPSS Statistics V.21 (IBM, USA).

Results

Characteristics of MOG-Ab-positive cases

Detailed demographic, clinical and radiological data of the analysed cohort are reported in table 1. All MOG-Ab-positive cases were negative for AQP4-Ab. The attack at sampling (index attack) was considered ‘severe’ when neurological examination showed an increase >2 points of the crude score of any Expanded Disability Status Scale (EDSS) functional system (compared with a neurological examination performed prior to the index attack in case this was not the first event). The clinical course was classified as (1) monophasic, when only one clinical acute/subacute event occurred during the entire follow-up, and (2) relapsing, in patients with one or more relapses  defined according to McDonald criteria. Radiological activity was defined as the presence of gadolinium-enhancing lesions or new or unequivocally enlarging T2 lesions on
brain/spinal cord MRI in comparison with a previous scan. Recovery was considered complete if neurological examination was normal and no symptoms were reported
(EDSS score 0 or equal to baseline value), absent if no improvement was observed (EDSS score at last follow-up ≥EDSS score at nadir), partial in all the other cases. The
final diagnosis was defined in each patient at the last follow-up according to latest diagnostic criteria.

NFL Analysis

In both patients and controls we found a significant positive correlation between age at sampling and serum NfL levels (ρ=0.497; p<0.001). Serum NfL concentration was higher in MOG-Ab positive patients (median 10.7 pg/mL, range 2.1–101.5) than in age-matched controls (median 5.9, range 1.3–13.8); p=0.005. In the MOG-Ab positive group, females had higher serum NfL levels compared with males (11.8 [4.0–101.5] vs 7.3 [2.1– 42.9], p=0.016). Moreover, NfL levels were higher in subjects with a severe attack (median 15.3 pg/mL, range 4–101.5) compared with those with a mild-to-moderate event (median 7.3, range 2.1–23.0; p=0.002), independent of age and sex (data not shown). ROC analysis revealed a moderate accuracy of serum NfL concentration for predicting attack severity (area under the curve=0.79; 95% CI=0.64
to 0.94; p=0.003). NfL cut-off value of 11.55 pg/mL provided 71.4% sensitivity and 79.2% specificity for a severe attack.

Median NfL concentration in the nine analysed CSF samples was 357.8 pg/mL (range 165.7–2281), and significantly correlated to that in serum (ρ=0.72, p=0.006). Finally, serum NfL levels correlate with EDSS score at sampling (ρ=0.47, p=0.003) and tend to correlate with EDSS score at last follow-up (ρ=0.29, p=0.07). No significant  ssociations were found between NfL levels, type of clinical presentation, relapse rate, final diagnosis, radiological findings and MOG-Ab titres.

Discussion

We herein used a highly sensitive assay to detect NfL levels in serum/CSF of patients with MOG-Ab-related conditions, in comparison with age-matched healthy controls. We observed that: (1) NfL levels are increased in older patients, likely reflecting age-dependent neuronal degeneration; (2) serum NfL levels are significantly increased in MOG-Ab-positive cases compared with controls, suggesting that significant axonal damage could occur in this condition; (3) serum and CSF NfL values are significantly correlated in MOG-Ab-associated disorders, providing evidence that blood sampling can be reliably used to measure NfL concentration in this condition; (4) serum NfL  correlates with attack severity and might predict long-term outcome in MOG-Ab-related disorders. Until now, only MOG-Ab titre, which is usually higher on relapse than in remission, has been proposed as a possible biomarker of disease activity in this condition. However, MOG-Ab titre seems not reliable in predicting the clinical course on an individual-patient basis, and regular monitoring of MOG-Ab level  is not clearly recommended. Our data confirm and extend previous observations obtained using a different technique in a smaller cohort and support the broader use of NfL as an accessible and repeatable biomarker of tissue damage. These findings are of particular relevance in MOG-Ab-related conditions where it is essential to improve the prediction of short and long-term prognosis. Several factors might explain the lack of association between the risk of relapse and NfL levels in our cohort, including the relatively short follow-up and the preferential correspondence between NfL levels and severity of the index clinical event.

On the other hand, the lack of association between NfL values and MRI measures of disease activity could be explained by the partial lack of association between clinical presentation and the extension of MRI involvement in our cohort. This observation is in line with previous reports andmight indicate once again that NfL levels reflect the extent of axonal damage occurring at disease presentation rather than overall disease activity. Limitations of our study include the wide range of follow-up duration across cases, the small sample size and the lack of regular NfL monitoring in single cases over time. However, our observation that serum NfL levels correlate to CSF ones and to the severity of the clinical event in MOG-Ab-positive patients expands the possible use of NfL as a promising and reliable biomarker.

Future studies in larger cohorts of symptomatic treatment-naïve cases with assess ent of NfL levels over time are warranted to evaluate intraindividual changes, to stratify patients according to relapse risk/worsening of disability, and to validate the use of NfL as a biomarker to predict disease course and guide therapeutic interventions in clinical practice.