Inflammatory multiple-sclerosis plaques generate characteristic metabolic profiles in cerebrospinal fluid

Abstract

Background: Multiple sclerosis (MS), an inflammatory disease of the central nervous system, manifests itself in numerous forms and stages. A number of brain metabolic alterations have been reported for MS patients vs. control subjects. However, metabolite profiles of cerebrospinal fluid (CSF) are not consistent among the published MS studies, most probably due to variations in the patient cohorts studied. We undertook the first investigation of highly homogeneous MS patient cohorts to determine characteristic effects of inflammatory MS plaques on the CSF metabolome, including only patients with clinically isolated syndrome (CIS) with or without inflammatory brain plaques, and controls.

Methodology/principal findings: CSF obtained by lumbar puncture was analyzed by proton magnetic resonance spectroscopy. 27 metabolites were quantified. Differences between groups of control subjects (n = 10), CIS patients with (n = 21) and without (n = 12) inflammatory plaques were evaluated by univariate statistics and principal component analysis (PCA). Seven metabolites showed statistically significant inter-group differences (p<0.05). Interestingly, a significant increase in beta-hydroxyisobutyrate (BHIB) was detected in CIS with vs. without active plaques, but not when comparing either CIS group with control subjects. Moreover, a significant correlation was found, for the first time, between CSF lactate concentration and the number of inflammatory MS brain plaques. In contrast, fructose concentrations were equally enhanced in CIS with or without active plaques. PCA based on all 27 metabolites yielded group-specific clusters.

Conclusions/significance: CSF metabolic profiles suggest a close link between MS plaque activity in CIS patients on the one hand and organic-acid metabolism on the other. Our detection of increased BHIB levels points to a hitherto unsuspected role for this compound in MS with active plaques, and serves as a basis for further investigation. The metabolic effects described in our study are crucial elements in the explanation of biochemical mechanisms involved in specific MS manifestations.

Figure 1. Selected CSF metabolite concentrations [mM] in patients with or without inflammatory brain MS plaques (abbreviated ‘inflamm’ and ‘non-infl’, respectively), and controls.

Data represent means±s.e.m.

Figure 2. Percentile plots for selected CSF metabolite concentrations [mM] in patients with or without inflammatory brain MS plaques (abbreviated ‘inflamm’ and ‘non-inflamm’, respectively), and controls.

These plots show the percentage of data less than or equal to an observed value. Marked deviations from normal distribution are readily recognized by major deviations from a straight line. The percentile plots for fructose illustrate the presence of CSF with particularly increased concentrations in CIS patients, independently of the presence of inflammatory plaques.

Figure 3. Scattergrams for lactate and acetate CSF concentrations [mM] in patients with or without inflammatory brain MS plaques (abbreviated ‘inflamm’ and ‘non-inflamm’, respectively), and controls.

Marked deviations from normal distribution are readily recognized for lactate (inflamm) since data are clustered at low values (around 2.2 mM), but also for acetate (inflamm) where data are clustered at high values (around 0.16 mM).

Figure 4. Lactate concentration [mM] in CSF of CIS patients with inflammatory plaques vs. number of inflammatory plaques, with linear fit.

Figure 5. Two-dimensional projection of a three-dimensional CSF metabolite plot, based on the first three principal components obtained by PCA.

All control subjects (green dots) are clustered within the broken ellipse that also contains one data point from each of the groups with and without inflammatory plaques (red and black dots, respectively).

Figure 6. Two-dimensional projection of a three-dimensional CSF metabolite plot, based on the first three principal components obtained by PCA (projection angle changed relative to Figure 5).

All CIS patients with inflammatory plaques (red dots), except for two, are clustered within the broken ellipse that also contains four data points from patients without inflammatory plaques (black dots).