Abstract
Cerebrospinal fluid (CSF) concentrations of neurofilament chains are promising biomarkers of axonal and neuronal damage, and several studies have shown that the CSF levels of neurofilaments are increased in patients with multiple sclerosis (MS). Neurofilaments are composed of light, intermediate, and heavy chains, of which mainly the heavy and light chains have been studied in MS. Both neurofilament light and heavy chain concentrations are increased and correlate with each other in CSF of MS patients, although some studies indicate that increases in neurofilament heavy chain concentrations are somewhat less pronounced than are light chain increases. 1 Furthermore, whereas neurofilament heavy chain concentrations correlate with clinical disability scores, the prognostic value of neurofilament light chain increases for predicting future worsening may be somewhat higher. 2
Neurofilament heavy chains are found in different phosphorylated isoforms. Several studies have investigated the concentrations of these isoforms in CSF in MS. In a recent study, CSF concentrations of the heavy chain isoform recognized by the SMI35 antibody was found to increase with age, and after age correction, neurofilament concentrations were found to be higher in all disease stages in MS compared to controls. Relapses were associated with higher CSF concentrations compared with stable disease. Finally, phosphorylated neurofilament levels correlated with Expanded Disability Status Scale (EDSS) scores in patients with relapsing-remitting multiple sclerosis (RRMS), and the correlation was most prominent in RRMS during relapse. 1 CSF neurofilament light chain levels also seem to have prognostic value as conversion to secondary-progressive MS and severe disability as assessed by the MS severity score was more likely in cases with high neurofilament light chain levels early in the disease when these patients were examined at follow-up visits after a mean of 14 years (range, 8–20 years) after disease onset. 3
Several studies have indicated that neurofilament levels are also responsive to treatment. 1 Treatment with fingolimod decreased neurofilament light chain levels compared with placebo, and reductions in neurofilament levels correlated with an improvement in relapse and magnetic resonance imaging (MRI) outcomes, 4 and in patients treated with natalizumab, those experiencing a relapse showed higher neurofilament levels compared with those in remission. 5 In this issue of Multiple Sclerosis Journal, Petzold et al. 6 report that patients with high CSF phosphorylated neurofilament heavy (NfH-SMI35) concentrations had accelerated brain atrophy compared to patients with lower concentrations at follow-up after 15 years. Brain atrophy is generally associated with increasing disability, although progression in EDSS was minimal in this study and changes did not reach statistical significance. The findings are of interest as the disease course was relapsing-remitting in seven, secondary progressive in six, and primary progressive in two of the patients, particularly as we desperately need prognostic biomarkers in progressive MS.
In patients with progressive MS, natalizumab treatment decreased CSF neurofilament light chain levels and was accompanied by increases in magnetization transfer ratio in both cortical gray and normal-appearing white matter and EDSS improvement. 7
It could be argued that CSF neurofilament may be ready for use in clinical trials as a secondary outcome measure or as part of a composite primary outcome measures in clinical trials of MS, although best documented in RRMS. Also neurofilament as baseline prognostic biomarkers could be used for enrichment of patient cohorts selected for clinical trials or as a stratification variable if they predict both short-term and long-term prognosis. Biomarkers that can predict disease worsening are greatly needed, not least for patients with progressive MS. In RRMS, gadolinium-enhancing lesions or new T2 lesions on MRI are the most sensitive markers of active (inflammatory) disease and correlate with clinical relapses. This correlation has enabled the conduct of short-time placebo-controlled proof-of-concept phase II trials in RRMS using MRI activity as an outcome measure. In contrast, in progressive MS, there are no MRI biomarkers with predictive properties as to the worsening of disease over longer periods, which can be used in short-term phase II trials of new disease-modifying drugs, and thereby could be used to select which drugs should be tested in very costly phase III trials in patients with progressive MS.
Overall, neurofilament chains are promising prognostic biomarkers, but most studies have been small and the findings should be corroborated in larger trials or observational studies. In addition, while neurofilaments have shown encouraging results in groups of patients, it remains to be shown that they have sufficient predictive value to be useful at the individual level. Finally, whereas repeated lumbar punctures for CSF analysis may be acceptable for patients participating in a clinical trial, it will not be suitable for use in daily clinical practice. Hence, it is encouraging that studies have shown that neurofilament levels in blood could be a valuable biomarker. Serum neurofilament light chain concentrations in early RRMS are increased and correlate with CSF levels and with MRI measures of demyelination and disease severity, suggesting that serum studies of neurofilament concentrations may find use as a surrogate outcome in future clinical trials as well as a routine biomarker of treatment response. 8