Vitamin D supplementation and serum neurofilament light chain in interferon‐beta‐1b‐treated MS patients

Abstract Objectives Serum neurofilament light chain (sNfL) is a promising biomarker of MS activity, progression, and treatment response. The aim of the present study was to address whether sNfL concentrations are affected by supplementation of vitamin D and correlate with disease activity in interferon‐beta‐1b (IFNb‐1b)‐treated Finnish MS. Materials and Methods Serum samples were available of 32 participants of the Finnish vitamin D randomized controlled trial (17 vitamin D/15 placebo). Serum 25 hydroxyvitamin D was measured using radioimmunoassay and sNfL using single‐molecule array (Simoa). Correlation of sNfL with brain magnetic resonance imaging (MRI) activity, burden of disease (BOD, mm3), and disability was assessed at the study baseline and at 52 weeks. Results Serum NfL concentrations were similar in the patients randomized to high‐dose vitamin D and placebo at the study baseline and at month 12 follow‐up (p‐value). Concentrations of sNfL were higher in patients with Gadolinium‐enhancing lesions in brain MRI: median (95% CI) sNfL was 14.84 (9.9–42.5) pg/ml and 11.39 (8.9–13.2) pg/ml in patients without Gd+ lesions (p = .0144) and correlated with enhancing lesion volume (Pearson r = .36, p = .037) at the study baseline but not at week 52. Serum NfL did not correlate with the MRI BOD or disability measured by expanded disability status scale and 25‐foot walk test. Conclusion In this small cohort of clinically stable IFN‐treated Finnish MS patients, sNfL levels were similarly low in patients supplemented with high‐dose vitamin D or placebo. Subclinical disease activity in MRI was associated with higher sNfL levels.


| INTRODUC TI ON
Neurofilaments are components of the neuronal cytoskeleton released in cerebrospinal fluid (CSF) in association with central nervous system damage. Neurofilament light chain (NfL) is promising as a biomarker in MS activity, progression, and treatment response (Teunissen & Khalil, 2012). The concentrations of NfL in serum (sNfL) and CSF among MS patients are highly correlated (Novakova, Zetterberg, & Sundström, 2017). Vitamin D can also be considered for a serum biomarker of MS. A number of observational studies have shown that low serum levels of 25-hydroxyvitamin D (25[ OH]D) are associated with clinical relapses or inflammatory disease activity detected in brain magnetic resonance imaging (MRI) in either treatment naïve MS patients or patients treated with immunomodulatory drugs (Ferre, Clarelli, & Aferruzza, 2018;Loken-Amsrud, Holmoy, & Bakke, 2012;Soilu-Hänninen, 2005). Among patients with MS treated with interferon-beta-1b, low 25(OH)D levels early in the disease course have been shown to be a strong risk factor also for long-term disease progression (Ascherio et al., 2014). However, conclusive evidence of the benefit of vitamin D supplementation in MS is still lacking, although it is supported by secondary MRI endpoints from several randomized controlled trials (RCTs; Hupperts, Smolders, & Vieth, 2019;Soilu-Hänninen, Åivo, & Lindström, 2012).
Two previous studies have investigated the effect of vitamin D supplementation on NfL in serum (sNfL) among participants of RCTs with high-dose vitamin D or placebo supplementation in MS (Holmøy, Røsjø, & Zetterberg, 2019;Smolders, Mimpen, & Oechtering, 2020). Neither of these studies supported and effect of vitamin D on sNfL, but among the Dutch participants of the SOLARIUM study, higher week 48 NfL levels showed a trend for a higher risk of combined unique active lesions in brain MRI (Smolders et al., 2020). Among Swedish MS patients, an inverse association between serum 25(OH)D and CSF-NfL levels has been shown (Sandberg, Biström, & Salzer, 2016).
We have previously performed an RCT with weekly administration of either 20,000 IU of cholecalciferol or identical placebo capsules in 66 Finnish MS patients (Soilu-Hänninen et al., 2012).
In this paper, we measured serum NfL from serum samples available from 32 patients of the RCT at the study baseline and of N = 25 patients (13 vitamin D/12 placebo) at baseline and 52-week follow-up. Serum NfL concentrations were measured in duplicate with single-molecule array (Simoa). Correlation of serum NfL with the number and volume of Gadolinium (Gd)-enhancing lesions on brain MRI and with MRI burden of disease and disability were assessed. The aim of the present study was to address whether vitamin D supplementation is associated with changes in circulating NfL concentrations in FNB-1-b-treated Finnish MS patients and whether serum NfL concentrations correlate with disability and MRI markers of disease burden and activity.

| Study design
The design and clinical and MRI results of the Finnish vitamin D study (cholecalciferol as an add-on treatment to subcutaneously administered interferon b-1b [IFNB] for the treatment of MS, NCT01339676) have been published previously (Soilu-Hänninen et al., 2012). Briefly, 66 RRMS patients aged 18-55 years with IFNb-1b use for at least 1 months and no neutralizing antibodies to IFNb were detected and baseline 25(OH)D levels below 85 nmol/L, were randomized to either 20,000 IU of cholecalciferol (Swiss-Caps), or identical placebo capsules, for 52 weeks.
All participants gave written informed consent. The study was approved by the ethics committee of Turku University and Turku University Hospital and the National Agency of Medicine, Helsinki, Finland.

| Vitamin D and NfL measurements
Serum samples for the 25(OH) D analysis were collected at the study baseline and months 6 and 12 visits, freshly frozen, and stored at −70°C until analyses. Serum 25(OH)D was measured using DiaSorin (Stillwater) 125 I RIA Kit. The sensitivity of this assay is 4.0 nmol/L, and intra-assay coefficient of variations (CVs) were <10%. Serum NfL concentrations were measured in duplicate according to the manufacturer's instructions using the NfL advantage kit for Simoa (Catalog number 103186; Quanterix) (Kampman, Steffensen, Mellgren, & Jorgensen, 2012;Rissin, Kan, & Campbell, 2010). Dynamic range of the assay is 0.174-1800 pg/ ml, and median intra-assay CV was 3.5% (interquartile range 1.7%-6.3%, min-max range 0.02%-23%). The Simoa assays were performed in the biomarker laboratory of the Institute of Clinical Medicine-Neurology, University of Eastern Finland, Kuopio, Finland.

| MRI analyses
Brain MRI using a 1.5 Tesla scanner and dual echo T2/PD and postcontrast T1-weighted sequences covering the whole brain in transverse imaging plane with 3 mm slice thickness was performed at the study baseline and at 52 weeks and centrally analyzed at the Neuroimaging Research Unit, Vita-Salute University, Milan, Italy, as described previously (Soilu-Hänninen et al., 2012). Analyses in-

| Statistical analyses
Statistical analyses were conducted with GraphPad Prism (GraphPad Software) and SPSS (SPSS Inc., version 20.0). Descriptive statistics are provided as mean and standard deviation for normally distributed data and median (95% confidence interval) for skewed variables. Because of non-normal distribution of the sNfL levels, natural logarithm transformation was used to create normally distributed data. Unpaired t test was used to compare differences in the logarithmic sNfL levels between the treatment arms at the study baseline and at week 52 and between patients with or without Gadoliniumenhancing lesions. Correlation of log-transformed sNfL levels to other variables was tested using Pearson's correlation. p-values <.05 were considered significant.  Table 1. Age, body mass index, disease duration, or smoking status did not significantly correlate with the serum NfL (not shown). Disability during the study was measured by the expanded disability status scale (EDSS) and 25-foot walk test as described earlier (Soilu-Hänninen et al., 2012). In the original sample, there was not significant difference in the clinical endpoints at week 52, but there was a trend for less increase in the MRI BOD and significantly less Gd-enhancing lesions in the vitamin D arm than in the placebo arm at week 52. The distribution of the 52-week clinical and MRI variables between the original sample, and the subset included in the current study are shown in the Table S1.     Figure S1).

| RE SULTS
At week 52, the enhancing lesions were smaller, ranging from 48 to 52 mm 3 and did not significantly correlate with the sNfL concentration (not shown). No correlation was seen between MRI BOD and sNfL either at the study baseline (r = −.12, p = .48) or at month 12 (r = .08, p = .67).

F I G U R E 2
Serum neurofilament light chain (Nfl) levels (pg/ml) in the patients with Gd+ levels and without Gd+ lesions in brain MRI at the study baseline. Bars show median and 95% confidence interval. Significance was tested with unpaired t test of logtransformed sNfL values the SOLARIUM study, the patients were also using IFN, but in the Norwegian vitamin D RCT, about half of the patients were not receiving disease-modifying therapies Kampman et al., 2012;Smolders et al., 2020). In the latter study, the subgroup of patients not treated with disease-modifying therapies showed a 30% decrease in sNfL during high-dose vitamin D supplementation. Thence, it is possible that in untreated MS patients, there is an effect with vitamin D on NfL that is masked by the IFN treatment.
In the study by Holmøy et al from Norway ( All the patients included in our study were using IFNb-1b and had to have a positive MxA test as a marker of IFN response to be included in the RCT (Soilu-Hänninen et al., 2012). The treatment with interferon has been shown to lead to a drop in serum NfL in a previous study (Varhaug, Barro, & Bjornevik, 2017). Patients experiencing Gd-enhancing lesions during IFN therapy had higher serum NfL levels than patients without enhancing lesions in the study by Varhaug et al. (2017). Our results are in line with this, and we also saw a correlation of sNfL concentrations with enhancing lesion volume at the study baseline. We did not have serum samples prior to IFN therapy from our patients nor control patients samples included. The serum NfL levels in our patients without enhancing lesions were low and in the same range as reported by Holmøy et al. (2019) from the Norwegian vitamin D study, and somewhat lower than in the study by Smolders et al. (2020). In a previous study from the biomarker laboratory of the University of Eastern Finland, patients without neurological diseases but with primary psychiatric disorders approximately 15 years older the MS patients in our study had similar NfL concentrations as the MS patients without Gd-enhancing lesions of our study (Katisko, Cajanus, and Jääskeläinen (2020).
In our cohort of relatively mildly disabled MS patients with short disease durations and no secondary progressive patients included, there was no correlation with sNfL and disability. In a previous Finnish study including older MS patients with longer disease duration and also patients with progressive MS, a correlation of sNfL with disability was found (Högl, Rissanen, & Barro, 2018). These results support sNfL being a marker of also axonal degeneration at least in patients with more advanced MS disease. The MRI T2 lesion volume that we measured from our patients can be considered a cumulative measure of total lesion formation in MS (Chitnis, Gonzales, & Healy, 2018).
We did not, however, find any correlation of sNf and the MRI burden of disease in our patients. This may have been driven by the absence of a statistically significant sample size, but may be also explained by the serum samples being collected approximately 7 years after disease onset and 3 years after the initiation of the IFN therapy. In a 10- year follow-up of MS patients followed from the disease onset, the strongest correlation between the sNfL and long-term lesion volume accumulation was observed between NfL samples collected during the first year after disease onset (Chitnis et al., 2018).
The major limitation of our study is the small sample size, since serum samples of only half of the patients included in the parental Finnish vitamin D RCT could be retrieved for the analyses. The strengths are the randomized controlled setting of the study, the MRIs performed with stringent imaging protocol and centrally analyzed in a reference center, and all the patients having an IFN response verified with an MxA test.

AUTH O R CO NTR I B UTI O N
KH involved in study concept and data collection, statistical analyses, interpretation of data, and drafting of the manuscript; OJ involved in Nf analyses, statistical advice, and critical review of the manuscript, S-K H involved in NL analyses and critical review of the manuscript; M S-H involved in study concept and supervision, data collection, drafting, and critical review of the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data available upon request from the corresponding author upon reasonable request.