Background

Quantification of the activity and progression of multiple sclerosis (MS) is an important tool for research on MS as well as its clinical treatment. Currently, disease activity and progression assessments rely on qualitative clinical evaluations or the acquisition of radiographic data such as magnetic resonance imaging (MRI).

Objectives

Quantifying MS disease activity (DA) and progression (DP) instead through the use of blood biomarkers would provide a significant reduction in barriers to such testing (e.g. monetary cost, time and specialized personnel requirements, invasiveness, operational difficulty, etc.). The use of an ensemble of proteins representing various biological pathways involved in MS pathophysiology would also provide useful insights into this complex and heterogeneous disease.

Methods

We investigated proteomic biomarkers associated with different levels of MS DA and DP using 205 blood serum samples from 88 patients (University Hospital Basel), extracting protein levels using Proximity Extension Assays (PEA) from Olink^TM. We then conducted a focused statistical analysis on 21 proteins that were selected for a custom MS assay panel development project based on their association with endpoints in previous studies. We corrected these protein levels using clinical data, including: age, sex, disease duration, age of the bio-banked sample, and medication status. We then compared protein levels to five different radiographic and clinical endpoints.

– Primary Endpoint: Gadolinium (Gd) enhanced lesion count

– Secondary Endpoints: T2 lesion volume, Expanded Disability Status Scale (EDSS) score, Clinically Defined Relapse Status, and Annualized Relapse Rate (ARR)

Results

In this report, we examine the univariate performance of selected proteins on the prediction of all five endpoints, comparing it to that of several multivariate machine learning techniques. We draw distinctions between the highest performing models for each endpoint and draw connections to the underlying biology governing MS activity and progression.

Conclusions

We found significant improvements in predictive power from the use of multivariate models in comparison to even the highest performing univariate techniques. The samples analyzed in this study will be re-assayed for validation purposes alongside additional cohorts in the forthcoming 21-plex custom MS proteomic assay panel.

Background

Multiple Sclerosis (MS) is a complex and heterogeneous disease. Investigating the biological pathways and cell types involved in MS pathophysiology as represented by protein biomarker expression can help inform the development of tools to monitor disease activity, disease progression, identify early evidence of relapse, and monitor treatment response.

Objectives

To develop a blood based multiplex proteomic assay that associates with clinical and radiographic endpoints in patients with MS. These endpoints include the presence of gadolinium-enhanced (Gd+) lesions, Annualized Relapse Rate (ARR) and clinically defined relapse status (active versus stable).

Methods

Serum samples (n=690 in total) from multiple deeply-phenotyped cohorts (ACP, CLIMB and EPIC) were tested in immunoassays for the measurement of 1196 proteins using Proximity Extension Assays (PEA) from Olink^TM and for 215 proteins using xMAP^TM immunoassays from Myriad RBM, Inc. (RBM). Associated radiographic and clinical endpoints at the time of the blood draw were correlated with the protein levels. Twenty-one proteins were selected for inclusion in a custom assay based on their performance in univariate and multivariate statistical models, and replication across independent cohorts. Biological pathway modeling and network analysis were performed to ensure comprehensive representation of MS neurophysiology. Area under the curve (AUC) was selected as the key metric for model performance evaluation.

Results

Multivariate statistical ensembles restricted to the expression levels of the biomarkers selected for the custom assay achieved AUC performance of 0.827 for classification of the presence of Gd+ lesions, 0.802 for classification of clinically defined relapse status, and 0.930 for the classification of patients with Low ARR (≤0.2 relapses) vs High ARR (≥1.0 relapses). A multivariate model utilizing shifts in biomarker expression in longitudinally paired samples achieved the highest observed performance of 0.950 for classification of Gd+ lesion presence. In each case, the multivariate models significantly outperformed (p-value <0.05) the AUC of the highest performing univariate biomarker.

Conclusions

Multivariate models restricted to the 21 selected proteins effectively classified several radiographic and clinical endpoints with stronger performance than any single biomarker. A 21-plex custom assay panel is being developed for further investigation and validation using additional cohorts.

Background

Multiple sclerosis (MS) is a multifaceted disease with an intricate pathophysiology that lies at the intersection of autoimmunity, inflammation, redox imbalance, demyelination, and neurodegeneration. The varying interplay of distinct and converging mechanistic profiles in MS is believed to contribute to the heterogeneity observed in disease course and outcomes, clinical presentation, and therapeutic response. With this high degree of undeciphered molecular complexity, identifying biomolecular markers that are reproducible as well as specific to even the major subclasses of MS has been problematic. These difficulties have hindered the clinical translation of biomarkers and their use to aid in disease assessment and treatment strategies for individual MS patients.

Objectives

We posit that a biocentric framework can be leveraged to augment the prognostic capacity of MS biomarkers. By coupling machine learning findings with a computational illustration of their dynamical interactions within disease-perturbed networks, we hope to extract biomarkers that convey the full spectrum of MS pathophysiology—from disease activity to worsening and progression. Furthermore, by honing in on biomarkers that reflect the true underlying biology, we may be able to expand on the standard list of MS clinical and surrogate endpoints, further guiding patient stratification and informing targeted therapeutic selection and drug repurposing.

Methods

Using a panel of 21 protein serum biomarkers that were pre-selected per radiographic and clinical endpoints, we ran spatial expression correlation to extract a proteomic signature specific to MS organs and cell types. We modeled the functional connectivity of these proteins and then performed unsupervised clustering and network centrality analysis to identify motifs of interconnected proteins. Network motifs were later annotated using significantly enriched gene ontology terms and pathways in order to contextualize their mechanism-of-action with respect to MS.

Results

Topological mapping of protein functional interactions uncovered 10 major pathological profiles in MS: (1) myelin integrity; (2) lipid metabolism; (3) immune modulation; (4) inflammation; (5) cerebrovascular function; (6) cell-cell communication; (7) cellular energetics; (8) synaptic dynamics; (9) neuroaxonal integrity; and (10) gut microbiota. Their shifting degree of involvement was captured to characterize the different paths to disease activity and progression. A rich repertoire of immune, glial, and neuronal cells was implicated as being critical in orchestrating the synergistically evolving crosstalk between these various disease mechanisms.

Conclusions

Through a complementary integration of data analytics and systems biology, we were able to shift the focus from that of single proteins to disease processes, identifying clinical biomarkers that appear to fully recapitulate hallmarks of MS.