Biomarkers of multiple sclerosis disease course in the blood and in the gut

Multiple sclerosis (MS) is a chronic neurological disorder with a high world-wide prevalence. The disease is characterized by substantial clinical heterogeneity. There is a general lack of prognostic biomarkers, and despite many years of intense research, tangible therapeutic targets remain scarce. The first goal of this thesis was to investigate the prognostic potential of blood-based candidate biomarker glial fibrillary acidic protein (GFAP) in MS, as cerebrospinal fluid (CSF) GFAP has previously been linked with disability worsening. The second goal was to contribute to a better understanding of whether and how the gut microbiota and clinical characteristics of MS are linked, as many studies – mainly in animal models – implicate the gut microbiota in MS pathology, with the exact role of the gut microbiota in people with MS remaining unclear.
In chapter 3, we assessed the potential of plasma glial fibrillary acidic protein (GFAP) compared to plasma neurofilament light chain (NfL) as prognostic biomarker for disability worsening in a mixed cohort of relapsing-remitting (RR) and primary progressive (PP)MS. We showed that both plasma GFAP and -NfL relate to disability worsening over 4-5 years based on ROC curve analysis, but only NfL was significantly related to worsening in a survival analysis.
In chapter 4, we assessed the gut microbiota composition in relation to worsening. We found a link between the presumably dysbiotic, inflammation-associated Bacteroides 2 (Bact2) enterotype at baseline and disability worsening over 4-5 years in patients with (RR)MS. In this chapter we also showed that summary metrics of gut microbiota composition (i.e. enterotypes and diversity metrics) were relatively stable over three months' time.
Results from the many cross-sectional case-control studies on the MS gut microbiota composition are quite divergent, which might be due to several reasons, such as incomplete confounder control and the use of relative microbiome profiling (RMP) approaches. In previous research1,2, quantitative microbiome profiling (QMP) has been shown to provide advantages over RMP to link host characteristics to the gut microbiota composition. In chapter 5, using a large cross-sectional case control design, we identified stool moisture content (a proxy for transit time) as the only major covariate of gut microbiota composition with a larger effect size than MS diagnosis. We confirmed that the Bact2 enterotype is increased in MS versus controls3. Again, we emphasized the importance of implementing gut transit time as covariate by showing that several taxon differences found in previous studies may merely reflect differences in stool moisture content between patients and controls. The implementation of QMP proved to be relevant as well, as fecal bacterial cell loads were significantly lower in the MS group than in the control group.
In people with the Prevotella enterotype, we found lower disability levels and lower levels of serum GFAP, compared to people with a Bact2 and Ruminococcus enterotype. Also, the short-chain fatty acid (SCFA) producer Faecalibacterium was decreased in PPMS versus controls, and showed an inverse correlation with GFAP in the PPMS cohort. These results point towards a potential beneficial role of SCFA producers in MS, as has been suggested in previous research4.
Longitudinal studies are required to determine whether gut microbiota metrics are stable in time and whether disease-modifying treatment (DMT) influences gut microbiome composition. With clinical follow-up over six months and repeated fecal sample collection in a recently diagnosed cohort (chapter 6), we were able to investigate the gut microbiota composition close to diagnosis, and the effect of starting a DMT. We confirmed that enterotypes are relatively stable over time, but that specific genera tend to vary a lot between the different time points, consistent with previous reports in healthy individuals. We did not identify clear patterns of gut microbiome changes after the start of a DMT, but this remains to be further investigated in larger cohorts.
Also, longitudinal set-ups can help elucidate whether specific disease events are associated with gut microbiota alterations and whether these alterations seem secondary to disease events or rather the other way around. In a third longitudinal study set-up in chapter 7, we investigated the dynamics of the gut microbiota during and in the six months after a relapse. During a relapse (i.e. within six weeks after relapse onset), we found no differences in beta diversity metrics, alpha diversity, enterotypes or bacterial cell loads in comparison with a clinically stable cohort. However, from the relapse timepoint to the subsequent timepoints there was a decrease in bacterial cell loads and richness, and an increase in fecal moisture content. This was independent of corticosteroid treatment. Consistent with this finding, the Bact2 prevalence was significantly higher at six weeks after a relapse in comparison with a remission cohort. This suggests a link between inflammatory disease activity in MS and the gut microbiota. These findings advocate for gut microbiota changes secondary to relapse activity, rather than the other way around, but this remains to be confirmed in larger studies with repeated fecal sampling before relapse.