Plasma microRNA Signature as Predictive Marker of Clinical Response to Therapy During Multiple Sclerosis

Abstract

Objective: Despite the availability of effective therapies for Multiple Sclerosis (MS), the unpredictable nature of disease progression and the variability in individual treatment outcomes call for reliable biomarkers. This pilot study aims to investigate the potential of plasma circulating microRNAs (miRNAs) as predictive biomarkers for clinical responses to dimethyl fumarate (DMF), a widely used oral treatment for MS.

Methods: Peripheral blood samples were collected from nineteen treatment-naïve people with relapsing-remitting MS (pwRRMS) before and after 3, 6, 12, and 24 months of DMF administration, as well as from nineteen healthy individuals. MiRNAs were quantified by RT-qPCR after plasma RNA extraction, and peripheral blood immune cells were analyzed by flow cytometry. Pathway enrichment and protein-protein interaction analyses were performed to identify the biological processes and molecular networks associated with mRNAs targeted by the specific DMF-modulated miRNAs.

Results: We identified a DMF-modulated miRNA signature with significant changes occurring at early treatment stages. Notably, specific miRNAs were correlated with both clinical and immunological outcomes upon DMF treatment, including lymphocyte count reduction (let-7b-5p and miR-223-3p) and disease progression over 2 years (miR-223-3p, miR-23a-3p, miR-23b-3p, miR-27a-3p, and miR-27b-3p), suggesting their potential as predictive biomarkers for treatment response. Moreover, the validated mRNA targets of DMF-modulated miRNAs were enriched for IL-6 signaling and NRF2-dependent antioxidant pathways, highlighting the potential molecular mechanisms underpinning DMF efficacy.

Interpretation: This small exploratory study underscores the potential of plasma circulating miRNAs as candidate biomarkers for predicting therapeutic outcomes in MS and it calls for validation in larger studies, which may enhance our understanding of disease pathophysiology and offer a promising tool for personalized treatment strategies.

Keywords: circulating microRNA; dimethyl fumarate (DMF); multiple sclerosis.

FIGURE 1

Schematic outline summarizing the study flow. (A) Schematic experimental design summarizing healthy controls and treatment‐naïve pwRRMS enrollment and (B) schematic overview of quantitative analyses of the blood samples performed in parallel is reported. PBMCs, peripheral blood mononuclear cells.

FIGURE 2

DMF‐dependent modulation of the plasma circulating miRNome. (A) Whisker plots showing blood circulating miRNA levels expressed as relative quantities (Log2) from healthy controls and pwRRMS before (T0) and after 3 (T1), 6 (T2), 12 (T3), and 24 (T4) months of DMF treatment, selected for being significantly different (t‐test, p < 0.05) in at least two comparisons; asterisks are relative to comparison with the RRMS‐T0 group. (B) Longitudinal behavior of blood circulating miRNAs selected for being significantly different (t‐test, p < 0.05) in at least two comparisons among different time points; graphs report the fold change versus RRMS‐T0. (C) Hierarchical clustering of healthy controls and pwRRMS before (T0) and after 3 (T1), 6 (T2), 12 (T3), and 24 (T4) months of DMF treatment based on differentially expressed miRNAs, by an ANOVA testing (p < 0.05). *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005.

FIGURE 3

The levels of plasma circulating miRNAs at disease onset correlate with those of soluble factors upon DMF treatment. Heatmap reporting the correlation index (Spearman r values, red if positive, blue if negative) of differentially expressed miRNAs and soluble factors in pwRRMS at disease onset (T0) and upon DMF‐treatment.

FIGURE 4

Association of plasma circulating let‐7b‐5p and miR‐223‐3p with DMF‐dependent lymphopenia. (A) Whisker plots showing the number (per microliter) of peripheral blood immune cells from healthy controls and pwRRMS before (T0) and after 3 (T1), 6 (T2), 12 (T3), and 24 (T4) months of DMF treatment, selected for being significantly different (t‐test, p < 0.05) in at least two comparisons; asterisks are relative to comparison with the RRMS‐T0 group. *p ≤ 0.05; **p ≤ 0.005; and ***p ≤ 0.0005. (B) Scatter plots showing the correlation between plasma circulating let‐7b‐5p (upper panels) and miR‐223‐3p (lower panels) levels at T0 (before starting DMF treatment) and the number (cells/mm³) of peripheral blood cell subpopulations at later time points upon DMF treatment in pwRRMS; r and p values are also reported.

FIGURE 5

Identification of a specific plasma circulating miRNA signature in the prediction of response to DMF. (A) Scatter plots showing plasma circulating miRNAs positively correlated with EDSS; r and p values are also reported. (B, C) Whisker plots showing the reported blood circulating miRNAs expressed as relative quantities (Log2) (B) or fold changes (upon DMF treatment) (C), as quantified in pwRRMS stratified on the basis of EDSS change (decreased or unchanged/increased) clinically reported upon DMF treatment. *p ≤ 0.05; **p ≤ 0.005. (D) Venn diagram showing the overlap of the sub‐lists of miRNAs identified in the reported different analyses.

FIGURE 6

Functional Annotation Clustering and Protein–Protein Interaction network of miRNA‐targeted mRNAs. (A) Gene Ontology (GO) enrichment analysis (DAVID) of mRNA targets identified for miR‐223‐3p, miR‐23a‐3p, miR‐27b‐3p, miR‐27a‐3p, and miR‐23b‐3p. The top 5 enriched annotation clusters are displayed based on their enrichment scores. GO terms are grouped into clusters, each annotated with relevant biological processes. (B) Protein–Protein Interaction (PPI) network analysis using STRING and Cytoscape of the mRNA targets of the above‐mentioned miRNAs. IL‐6 (the major hub based on the betweenness centrality score) is highlighted in red.