The Association of Axonal Damage Biomarkers and Osteopontin at Diagnosis Could Be Useful in Newly Diagnosed MS Patients

Abstract

(1) Background: Multiple sclerosis (MS) is a biologically highly heterogeneous disease and has poor predictability at diagnosis. Moreover, robust data indicate that early disease activity strongly correlates with future disability. Therefore, there is a need for strong and reliable biomarkers from diagnosis to characterize and identify patients who require highly effective disease-modifying treatments (DMTs). Several biomarkers are promising, particularly neurofilament light chains (NFLs), but the relevance of others is less consolidated. (2) Methods: We evaluated a panel of axonal damage and inflammatory biomarkers in cerebrospinal fluid (CSF) and matched serum obtained from a cohort of 60 newly diagnosed MS patients. Disability at diagnosis, negative prognostic factors, and the initial DMT prescribed were carefully recorded. (3) Results: We observed correlations between different axonal biomarkers: CSF and serum NFL versus CSF total tau; and between the inflammatory marker osteopontin (OPN) and axonal biomarkers CSF p-Tau, CSF total tau, and serum NFL. CSF and serum NFL and total tau, as well as CSF OPN, positively correlated with EDSS at diagnosis. Moreover, CSF and serum NFL levels were increased in patients with gadolinium-enhancing lesions (p = 0.01 and p = 0.04, respectively) and in those treated with highly effective DMT (p = 0.049). Furthermore, CSF OPN and both CSF and serum NFL levels significantly differentiated patients based on EDSS, with a combined ROC AUC of 0.88. We calculated and internally validated biomarker (in particular serum NFL) thresholds that significantly identified patients with higher disability. Finally, CSF OPN levels and dissemination in the spinal cord were significant predictors of EDSS at diagnosis. (4) Conclusions: These preliminary exploratory data confirm the pathological interconnection between inflammation and axonal damage from early disease stages, contributing to early disability. Follow-up data, such as longitudinal disability scores, repeated serum measurements, a healthy control group, and external validation of our results, are needed. We suggest that combining several fluid biomarkers may improve the clinical characterization of patients.

Keywords: biomarker; multiple sclerosis; neurodegeneration; neurofilament; osteopontin; tau.

Figure 1

Box plots comparing biomarker levels (expressed in Log10) in patients with multiple sclerosis (MS) stratified by expanded disability status scale (EDSS) scores at diagnosis using a cut-off of 2.5. Each subplot illustrates the distribution of a specific biomarker in cerebrospinal fluid (CSF) or serum: (a) CSF osteopontin (OPN) levels are significantly elevated in patients with EDSS > 2.5 compared to those with EDSS ≤ 2.5 (p < 0.05). (b) CSF neurofilament light chain (NFL) levels are also significantly higher in the EDSS > 2.5 group (p < 0.05). (c) Serum NFL concentrations show a similar pattern, with significantly increased levels in the EDSS > 2.5 group (p < 0.05). (d) Phosphorylated tau (p-Tau) levels in CSF do not significantly differ between the two groups (ns = not significant). (e) Serum OPN levels also show no significant difference between the EDSS categories. (f) Total tau (t-Tau) levels in CSF are not significantly different between groups. Statistical significance: p  <  0.05 is indicated by an asterisk (*); ns = not significant.

Figure 2

ROC curves computed for each biomarker based on disability evaluated with the EDSS (using a cut-off of 2.5). Significant results (p < 0.05) with an AUC between 0.7 and 0.8 were obtained for cerebrospinal fluid and serum neurofilament light chain and cerebrospinal fluid osteopontin.

Figure 3

Multivariate ROC curves computed for combined biomarkers based on disability evaluated with the EDSS using a cut-off of 2.5. Both the models were statistically significant (p < 0.05). On the left, the model with cerebrospinal fluid osteopontin and neurofilament light has an AUC of 0.878; on the right is the same model after adding serum neurofilament light chains, with an AUC of 0.885.