If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Laboratory Medicine, Institute of Biomedicine, University of Gothenburg, Gothenburg, SwedenDepartment of Clinical Chemistry, Sahlgrenska University Hospital, Gothenburg 413 85, Sweden
Multiple sclerosis (MS) is an immune-mediated demyelinating disorder of the central nervous system. The glycosphingolipid sulfatide, a lipid particularly enriched in the myelin sheath, has been shown to be involved the maintenance of this specific membrane structure. Sulfatide in cerebrospinal fluid (CSF) may reflect demyelination, a dominating feature of MS. We investigated the diagnostic utility of CSF sulfatide isoform levels to separate different courses or phenotypes of MS disease.
Material and methods
This was a mono-center, cross-sectional study of relapsing-remitting MS (RRMS) (n = 45) and progressive MS (PMS) (n = 42) patients (consisting of primary PMS (n = 17) and secondary PMS (n = 25)) and healthy controls (n = 19). In total, 20 sulfatide isoforms were measured in CSF by liquid chromatography-mass spectrometry.
Results
CSF total sulfatide concentrations, as well as CSF sulfatide isoform distribution, did not differ across the study groups, and their levels were independent of disease course/phenotype, disease duration, time to conversion to secondary PMS, age, and disability in MS patients.
Conclusion
CSF sulfatide isoforms lack diagnostic and prognostic utility as a biomarker for progressive MS.
). Typical hallmarks of MS pathology are multiple lesions characterized by inflammatory infiltrates, demyelination and axonal loss, followed by astrogliosis. Over time, deficient remyelination results in impaired nerve conduction with subsequent persistent loss of myelin contributing to the accumulation of symptoms and disability often seen in MS (
). Most frequently, the disease starts with a relapsing-remitting course (RRMS), characterized by transient periods of neurological symptoms followed years later by a secondary progressive course (SPMS) with slowly increasing disability with or without superimposed relapses (
). In 10–15% of patients the disease is progressive from onset and designated primary progressive MS (PPMS). Although several attempts have been made to separate these disease phenotypes on genetic, biochemical and neuroradiological grounds (
) still the distinction between them is essentially based on their clinical features.
Sulfatide, a highly multifunctional glycosphingolipid particularly enriched in the myelin sheath, has been shown to be involved in oligodendrocyte differentiation, maintenance of the myelin sheath as well as formation of the axon-glial junctions at the paranodal region (
). Thus, alterations in sulfatide metabolism are suggested to contribute to myelin deterioration as well as synaptic dysfunction, neurological decline and inflammation observed in different conditions associated with myelin pathology. Sulfatide has been proposed as a target for the autoimmune attack in MS and altered levels of sulfatide and anti-sulfatide antibodies in serum and CSF in this disease has been reported previously (
), we demonstrated that levels of total CSF sulfatide and the specific isoforms C24:1, C26:1 and C26:1-OH isoforms were significantly increased and C23:0-OH was significantly decreased in PMS as compared with newly diagnosed RRMS patients and normal age-matched controls. In multivariate discriminant analysis the CSF sulfatide isoform pattern in PMS patients was distinct and non-overlapping with that of RRMS patients and control individuals. However, sulfatide levels did not correlate with tested CSF biomarkers (neurofilament light (NF-L), glial fibrillary acidic protein (GFAP)), magnetic resonance imaging (MRI including T2 or contrast-enhancing T1 lesions, brain volume) nor clinical parameters (disability measured with Expanded Disability Status Scale (EDSS)) (
). In the present study, we aimed to confirm and further explore the sulfatide isoform pattern in CSF in relation to disease course and progressive phenotypes, introducing a larger cohort with RRMS, SPMS and PPMS patients. The overall aim was to investigate the diagnostic utility of sulfatide isoforms in the conversion of RRMS to the progressive phase of MS.
2. Patients and methods
2.1 Study participants
We consecutively included 87 patients with MS, 45 RRMS and 42 PMS patients (17 PPMS and 25 SPMS) at the Department of Neurology, Sahlgrenska University Hospital, Gothenburg, Sweden, all fulfilling the 2017 revised McDonald criteria (
) . Inclusion criteria: RRMS patients were age-matched to PMS. Exclusion criteria: any other neurological disease. No randomization was performed to allocate subjects in the study. No sample calculation was performed since we included a larger patient cohort compared to previous study (
). The clinical data were extracted from the patients’ medical records and from the Swedish MS registry (www.neuroreg.se). All MS patients were examined by a trained neurologist and their neurological disability was scored according to the EDSS (
) at the time of sampling. Healthy donors (HD) selected for this study were age-matched and none had any neurological symptoms or a history of neurological disease. MRI scans of brain and cervical columns were performed with 1.5 or 3.0 Tesla machine, using a standard protocol for MS, which included T1-weighted sequence following a dose of intravenous gadolinium (Gd) contrast, T2-weighted sequence, and fluid-attenuated inversion recovery sequence according to Swedish MS guidelines (
Guidelines for the use of magnetic resonance imaging in diagnosing and monitoring the treatment of multiple sclerosis: recommendations of the Swedish multiple sclerosis association and the Swedish neuroradiological society.
). Number of contrast-enhancing T1 lesions were extracted from the patients’ medical records. The characteristics of patients and HD are shown in Table 1. This observational study was not pre-registered.
Table 1Demographic and clinical profile of MS patients and HD.
RRMS
PMS
PPMS
SPMS
HD
n = 45
n = 42
n = 17
n = 25
n = 19
Age, mean (range)
46 (26–58)
52 (36–68)
52 (36–68)
53 (43–67)
55 (27–71)
Gender (M//F)
13//32
19//23
7//10
12//13
10//9
Currently treated with DMT
17
5
0
5
NA
EDSS, mean (range)
2.4 (0–7)
4.9 (2–8.5)
4.2 (2–6.5)
5.4 (2.5–8.5)
NA
MRI enhancing lesions, mean (range)
1.3 (0–15)
0.26 (0–3)
0.47 (0–3)
0.12 (0–1)
NA
Relapse within 3 months
13
0
0
0
NA
Relapse within 3 weeks
10
0
0
0
NA
Disease duration [years], mean (range)
8.3 (0.1–31.0)
13.7 (0.4–31.3)
5.5 (0.4–13.5)
19.2 (2.1–31.3)
NA
Conversion from RRMS to SPMS [years], mean (range)
The study was approved by the Regional Ethics Committee of Gothenburg (T487–14) and carried out in accordance with the code of ethics of the World Medical Association (declaration of Helsinki). All patients and HD provided written-informed consent.
2.3 Cerebrospinal fluid sampling
The spinal tap and the CSF samples were handled according to the consensus protocol of the BioMS-EU network for CSF biomarker research in MS (
). CSF was collected in polypropylene tubes, gently mixed by inverting the tube, and centrifuged at room temperature at 2000 x g for 10 min. The CSF samples were aliquoted (0,5 mL) and stored at -80 °C pending analyses.
2.4 Quantification of CSF sulfatide
The quantification of sulfatide isoforms was performed as previously described (
). No blinding of the samples was performed, however, the sample preparation and analysis of sulfatides were randomized and all samples were simultaneously analyzed. Briefly, sulfatide were automatically extracted from 100 µL CSF using the robot-assisted BUME method (
). R-values refer to spearman correlation coefficients. P-values were calculated using the two-tailed Mann-Whitney U test. FDR p < 0.05 are considered statistically significant. Normality was assessed using the Shapiro-Wilks test. All variables were deemed non-normally distributed hence non-parametric tests were used throughout.
2.6 Multivariate discriminant analysis
LDA is a so-called supervised dimensionality reduction technique. A supervised learning algorithm is given access to information on which samples belong to which group as opposed to for example principal components analysis (PCA) which only is concerned about overall variation in the data. Dimensionality reduction reduces many variables into few variables (normally two) which may be visualized. The LDA algorithm reduces the dimensions into the number of pre-defined groups minus 1. The LDA algorithm's objective function consists of maximizing the distance between the groups while minimizing the distance within the groups.
3. Results
3.1 The global isoform pattern of sulfatide in CSF does not distinguish PMS patients from RRMS patients nor from age-matched controls
Our previous study demonstrated altered levels of total CSF sulfatide and the specific isoforms C24:1, C26:1, C26:1-OH and C23:0-OH isoforms in PMS as compared with newly diagnosed RRMS patients and age-matched HD (
). In the present study, we aimed to investigate whether these results could be reproduced in a larger patient cohort including subgroups of PMS patients (PPMS and SPMS). The levels of sulfatide isoforms were analyzed as previously by LC‐MS/MS and we identified 20 different sulfatide isoforms in CSF that were compared between the different patient groups and HD. However, the levels of neither total CSF sulfatide nor the specific sulfatide isoforms described previously (
) were significantly altered in PMS patients as compared with RRMS patients and HD (Fig. 1A, Suppl Table 1). This was further confirmed by constructing receiver-operating curves (AUC < 0,6 Suppl Fig. 1). Similar results were obtained after subdividing of the PMS patients into PPMS and SPMS (Fig. 1B). Analysis of the remaining CSF sulfatide isoforms showed no significant difference neither comparing PMS, RRMS patients and HD (Suppl Fig. 2, Suppl Table 1), nor after subdividing of the PMS patients into PPMS and SPMS (data not shown).
Fig. 1CSF abundance of total sulfatide and specific isoforms (ST C24:1, C26:1, C26:1-OH, C23:0-OH) in progressive MS (PMS) patients, relapsing-remitting MS (RRMS) patients and healthy donors (HD). A. Levels in the combined PMS group were compared to RRMS and HD. B. PMS subgroups PPMS and SPMS were compared to RRMS and HD. Y-axis shows scaled data (mean scaled to 0 and standard deviation scaled to 1). See also Suppl Table 1 for group mean of sulfatide isoform concentrations (nmol/L) and standard deviations.
Fig. 2Linear discriminant analysis of global sulfatide isoform pattern in CSF from healthy donors (HD), relapsing-remitting MS (RRMS) and progressive MS (PMS) patients. This analysis shows no distinct separation between the groups, indicating lack of diagnostic utility of any of the CSF sulfatide isoforms.
To obtain a global comparison of various sulfatide isoforms in the different patient groups and HD, LDA was performed. The result is displayed in a two-dimensional plot in Fig. 2. Although slight group differences were visible, LDA of the sulfatide isoform pattern in CSF was unable to clearly differentiate PMS from RRMS patients and HD. Similar results were obtained after subdividing of the PMS patients into PPMS and SPMS (data not shown). Disease modifying therapies did not influence the findings above (statistical analysis performed after excluding patients under treatment, n = 22 RRMS and SPMS patients see Table 1, data not shown).
3.2 No correlation of CSF sulfatide isoform levels with age, disease duration, duration of progressive course, number of MRI lesions or EDSS score in MS patients
We performed correlation analysis between CSF-sulfatide levels (total and selected isoforms) and demographic and clinical characteristics. No relationships were found between CSF sulfatide concentrations and disease duration in this cohort of MS patients (also when subdivided in RRMS, PPMS and SPMS, data not shown). Suppl Figure 3 shows selected data for total sulfatide and the C24:1, C26:1, C26:1-OH and C23:0-OH sulfatide isoforms divided in RRMS and PMS. Similarly, no significant findings were observed when correlating sulfatide concentrations to EDSS scores (Suppl Table 2). Furthermore, sulfatide levels did not increase with age and did not correlate with the number of contrast-enhancing lesions on MRI (data not shown). Regression analysis was also performed to investigate whether any sulfatide isoforms showed association with duration of progressive phase (i.e., time since conversion from RRMS to SPMS or disease duration for PPMS). No significant associations could be found (Suppl Table 3), except for 16:0 and 18:1 sulfatide in SPMS patients (FDR adjusted p-values 0.0016 and 0.01 respectively). However, no correlation of these two sulfatide isoforms was found for other disease parameters of MS.
4. Discussion
In the present study, we aimed to reproduce and confirm previous findings (
) of CSF sulfatide specific isoforms to differentiate PMS from RRMS patients in a larger cohort, and further to explore the sulfatide isoform pattern in CSF in relation to the progressive phenotypes, duration, and severity of PMS. However, we neither showed a gradient with increasing or decreasing levels of sulfatide isoforms correlating to disease phenotype nor with disease severity and disease duration. Furthermore, the sulfatide isoform concentrations were not related to the duration of PMS. Thus, the present study did not confirm the diagnostic utility of CSF sulfatide isoforms in separating MS phenotypes.
There might be several possible explanations for the discrepancy between the results of our two studies. First, the present RRMS and HD cohorts were age-matched to PMS patients and therefore older, and the RRMS patients had longer disease duration than the RRMS cohort of the previous study (
). The mean age of RRMS patients in the previous as compared to the present study cohort was 35 and 53 years, respectively, which affected the disease duration (mean 2.3 vs 13.8 years). Second, we included a higher number of PMS patients in the present study with a broader range of disability. Third, in contrast to the previous study (
), RRMS patients had lower inflammatory activity and PMS patients had higher inflammatory activity in the present study. In fact, most RRMS patients of the previous study recently had their clinical onset of MS and their diagnostic MRI frequently showed contrast enhancement. Mean number of contrast-enhancing lesions in RRMS patients of the previous as compared with the present study was 2.8 (range 0–12) and 1.3 (range 0–15), respectively, and the proportion of patients with relapse was considerably higher (22/29 vs 13/45, respectively). However, PMS patients had slightly higher activity on MRI as compared with the previous study, with mean number of contrast-enhancing lesions of 0.26 (range 0–3) vs 0.08 (range 0–1). Neither PMS patients of the present, nor the previous study had any signs of relapse. Fourth, the size of the patient material could be another confounding factor because limited study populations have a greater risk of findings that are non-generalizable.
The method of choice might also influence the results when comparing different studies. Thin-layer chromatography with immunodetection was the first method with sufficient sensitivity to detect sulfatide in CSF, which was used in the study by Haghighi et al. (
) reporting increased CSF sulfatide in MS patients. This technique shows rather poor reproducibility between runs and thus requires simultaneous analysis of the entire cohort, including the controls, which is hard to achieve due to automation difficulties. LC-MS/MS is now the golden standard for CSF sulfatide quantification and our recently published high-throughput method shows good sensitivity and reproducibility, using in house semi-synthesized sulfatide with C19:0 fatty acid as internal standard (
). The sulfatide concentrations obtained in healthy individuals were close to the levels observed for the HD samples in the previous MS study (100 ± 30 nmol/L [mean+/- SD, n = 19], ages 27–71 years vs 84 ± 23 nmol/L [mean+/- SD, n = 16], ages 24–50 years (
The regression analysis performed to investigate whether any CSF sulfatide isoforms showed association with duration of progressive phase indicated significant associations for the 16:0 and 18:1 sulfatide isoforms (FDR adjusted p-values 0.0016 and 0.01 respectively for SPMS). Neither did these specific sulfatide isoforms differ in the previous study comparing PMS to RRMS and HD, nor associated with relevant biomarkers or EDSS scores (
). The explained variance of the 16:0 and 18:1 sulfatide isoforms and disease duration was further found to be low (adjusted R2 0.038 and 0.01 respectively). Thus, although a statistically significant relationship may exist, the clinical importance of this relationship appears limited.
In one previous study, sulfatide isoform pattern (four selected isoforms) in serum of RRMS patients was analyzed (
). The authors found that the serum levels of C18:0 and C24:1 sulfatide correlated with disease status in RRMS patients. These findings were followed up more recently in a study that analyzed sulfatide isoforms in isolated extracellular vesicles (EVs) from plasma of MS patients (
). EVs are membrane nanoparticles of different size believed to be involved in intercellular communication by transferring proteins, lipids and RNA between cells (
). The EV composition is hypothesized to mirror the releasing cell type and thereby reflect the biological system in health and disease. Moyano et al. found a significant increase of C16:0 sulfatide specifically in small EVs from plasma of patients with MS (mainly RRMS patients) and further propose C16:0 sulfatide in small EVs to represent a unique indicator of MS pathophysiology (
). Most likely, EVs in body fluids such as plasma or CSF maintain the biological features of the releasing cells and thereby include potential disease markers. An attractive hypothesis is that EVs isolated from CSF could reveal differences in biomarker content (including sulfatide) that is not observed in analyses of total CSF. Future studies are needed to resolve these questions.
Based on our data in the present report, we conclude that CSF sulfatide isoforms lack diagnostic utility in separating multiple sclerosis patients from normal individuals as well as in differentiating progressive from relapsing-remitting multiple sclerosis. The rationale for sampling and patient material selection as well as the quantification method and preanalytical conditions are possible confounding factors for the contradictory results of CSF sulfatide as a biomarker in MS.
Founding
The study was supported by grants from the Swedish State Support for Clinical Research (Grant No. ALFGBG-722081), Regional FoU grant Västra Götalandsregionen (Grant No. 260 101), NEURO Sweden, NEURO Gothenburg, Edith Jacobsons Foundation to JL, by a grant from the Swedish Research Council (Grant No. 2017–01821) to SC and by grant from NEURO Gothenburg, Edith Jacobsons Foundation, the Gothenburg Foundation for Neurological Research (ISNF) to LN.
CRediT authorship contribution statement
Lenka Novakova: Methodology, Conceptualization, Writing – original draft, Writing – review & editing. Marcus Henricsson: Formal analysis, Data curation, Writing – review & editing. Elias Björnson: Formal analysis, Writing – review & editing. Markus Axelsson: Writing – review & editing. Jan Borén: Writing – review & editing. Igal Rosenstein: Writing – review & editing. Jan Lycke: Methodology, Conceptualization, Writing – original draft, Writing – review & editing. Susanna L. Cardell: Methodology, Conceptualization, Writing – original draft, Writing – review & editing. Maria Blomqvist: Methodology, Conceptualization, Writing – original draft, Writing – review & editing.
Declaration of Competing Interest
JL reported receiving honoraria for lectures and for advisory boards from Biogen, Novartis, Merck, Alexion, Roche, Sanofi, and BMS; and unconditional grants from Biogen and Novartis; and is serving on the editorial board of the Acta Neurologica Scandinavica outside the submitted work. LN has received honoraria for lectures from Biogen, Novartis, Teva and Merck, and for advisory boards from Merck, Janssen and Sanofi. MS, EB, MA, JB, IR, SC and MB report no conflicts of interest.
Acknowledgments
We gratefully acknowledge patients and healthy donors for providing their precious CSF samples.
Guidelines for the use of magnetic resonance imaging in diagnosing and monitoring the treatment of multiple sclerosis: recommendations of the Swedish multiple sclerosis association and the Swedish neuroradiological society.
00209-2&id=gr1.jpg){kind=link}
00209-2&id=gr2.jpg){kind=link}