Exploring blood biomarkers for disease progression and treatment response in multiple sclerosis

This thesis focuses on developing blood biomarkers for multiple sclerosis (MS) disease progression (part I), evaluating neurofilament light chain (NfL) as a treatment monitoring marker (part II), and advancing clinical implementation of NfL, contactin-1 (CNTN1), and glial fibrillary acidic protein (GFAP) (part III). In part I, a natalizumab-treated MS cohort (n=84) was divided into progressors and non-progressors to identify biomarkers linked to disease progression. Serum samples were collected at baseline and at 3, 12, 24 months, and last follow-up (median 5 years). Chapter 2 studied NfL, a biomarker for neuro-axonal damage. NfL levels decreased significantly within 3 months after natalizumab initiation, then remained stable and did not differ between progressors and non-progressors, suggesting it reflects relapse-related injury rather than chronic progression. Chapter 4 assessed biomarker associations with brain MRI volumetrics using Sequence Adaptive Multimodal SEGmentation (SAMSEG) to handle different MRI scanners. Yearly changes in brain, thalamic, and ventricular volumes did not differ between groups. However, NfL increases were linked to brain and thalamic atrophy. Only NfL at 1 year predicted these volume changes, likely due to residual inflammation and pseudo-atrophy early in treatment. Chapter 3 focused on CNTN1, involved in axo-glial interactions disrupted in MS. MS patients had lower CNTN1 levels than healthy controls (n=222). In the natalizumab cohort, CNTN1 decreased after treatment, with baseline and 12-month levels lower in progressors. Baseline CNTN1 predicted progression, but longitudinal CNTN1 did not correlate with MRI volumetrics, except lower CNTN1 at 3 months predicted ventricular enlargement, possibly reflecting axo-glial changes. Chapter 5 studied GFAP, a CNS astrocyte cytoskeleton protein. GFAP reached nadir within 3 months, faster than NfL’s 12-month nadir, suggesting quicker response and less sensitivity to residual activity. Like NfL, GFAP showed no direct link to progression but predicted ventricular enlargement at baseline and 1 year, indicating roles in active inflammation and chronic processes. GFAP may supplement NfL and CNTN1 depending on timing and MRI region. Chapter 6 used a proteomic immunoassay measuring 1472 serum proteins to find novel biomarkers linked to progression and brain atrophy. Twenty proteins differed consistently between groups; seven remained significant after multiple testing correction. Pathway analysis highlighted metabolism—especially fat digestion and sphingolipid pathways—pointing to lipid metabolism’s role in MS progression, possibly related to oligodendrocyte/myelin energy metabolism. Part II evaluated NfL as a treatment monitoring tool. Chapter 7 studied 42 JC-virus positive patients switching from natalizumab to ocrelizumab due to PML risk. Disease activity was absent in 83% of direct switchers and 50% of indirect switchers. NfL remained stable but transiently increased with PML or MS activity and decreased with ocrelizumab. NfL lacks specificity to replace MRI for differentiating PML from MS activity. Chapter 8 described a B-cell tailored ocrelizumab dosing protocol during the COVID-19 pandemic. Using 10 B-cells/μL for re-dosing, median intervals extended to 34 weeks. Disease recurrence was low (1.9%), though limited follow-up restricts conclusions. Group NfL stayed stable, but individual NfL sometimes failed to detect MRI activity. Part III, chapter 9 addressed pre-analytical factors affecting NfL, CNTN1, and GFAP measurements. Seven experiments tested sample handling in 5 controls and 5 MS patients. NfL and GFAP were measured via Simoa; CNTN1 via Luminex. Biomarkers were generally stable. Recommendations included centrifugation within 24 hours for NfL and limits on freeze-thaw cycles (CNTN1 max 2; NfL/GFAP max 4). Finally, chapter 10 evaluated serum NfL’s clinical impact via 166 questionnaires on 157 patients. NfL influenced clinical decisions in 19.3% of cases and increased clinician certainty. Lower NfL correlated with reduced expectations of MRI activity. NfL was most valued for new symptoms and differential diagnosis. Findings highlight NfL’s limitations and the need for more sensitive, specific MS biomarkers.