Serum markers which reflect MS disease activity could help personalize MS therapeutics. Longitudinal samples from patients undergoing autologous haematopoetic stem cell transplantation (HSCT) for aggressive MS represent a valuable cohort to search for such biomarkers, as these patients had very active disease prior to treatment followed by durable supression of inflammatory disease activity after treatment.
To investigate changes in candidate serum proteins in patients with active MS compared to controls as well as before and after HSCT in relation to clinical and MRI outcomes.
97 proteins of interest were identified including established markers of inflammation and neurodegeneration. Levels were quantified using an in-house antibody colocalization microarray in 24 MS patients with aggressive relapsing MS at baseline compared to 10 controls. Pre-post HSCT changes were analyzed over 10 timepoints pre and up to 36 months post HSCT. We used principal componant analysis for data reduction prior to correlation as well as mixed effects models of individual proteins to compare changes in levels to clinical and MRI covariates of interest (age, EDSS score, relapses, sustained progression, lesional and volumetric MRI measures).
Levels of 19 proteins differed between MS patients at baseline and controls and 17 proteins differed comparing baseline and 12-months post HSCT (simple t tests, p<0.1); we focused on the levels of these proteins for subsequent analyses. 7 proteins were identified in both comparisons including amphiregulin, cathepsin, CRP, GRO, HAI-1 and leptin, which may indicate normalization post HSCT. 8/24 patients developed sustained EDSS progression in the absence of ongoing relapses post HSCT; using mixed effects models, of the 17 candidate proteins, the longitudinal trajectory of CRP levels differed in patients who developed sustained progression compared to those who did not (B=-0.003, p=0.045). Component analysis was used to summarize clusters of proteins into a single value based on internal correlation/discordance. At baseline, one cluster of proteins (CRP, KLK14, PAI-1, IGFBP-7, PDGF) correlated with preceding rapid progression from diagnosis to EDSS 6 (p=0.011, r=0.80) and EDSS worsening in the preceding 24 months (p=0.047, r=0.46). A different cluster of proteins (HAI-1, amhiregulin, FAS, capthespsin B, e-cadherin, GFAP) correlated with the pretreatment rate of brain atrophy. Comparing pre-post changes, one cluster correlated with rate of brain atrophy in the first year post HSCT (p=0.024, r=0.059).
This exploratory analysis of longitudinal serum biomarkers changes pre and post HSCT provides hypothesis generating observations worthy of future investigation.