In the last decade, structural and functional synaptic alterations, collectively known as synaptopathy, have come up as a determinant pathological process contributing to the neurodegenerative damage in multiple sclerosis (MS) and its mouse model, the experimental autoimmune encephalomyelitis (EAE). Since synaptic alteration and loss are reversible, unlike loss of neurons, an early detection could permit a precocious clinical intervention with potentially better therapeutic outcomes but reliable biomarkers are not available yet.
MicroRNAs (miRs) circulating in the cerebrospinal fluids (CSF) are good candidates as possible sensitive biomarkers for MS synaptopathy-driven disease progression. They represent a new class of modulators of gene expression with stable presence in the body fluids and with a critical role in many physiological and pathological processes, especially in the central nervous system. Accordingly, it has been recently demonstrated that miR-142-3p is a crucial component in a detrimental regulatory axis of EAE/MS excitotoxic synaptic dysfunctions, by reducing the level of the glial glutamate aspartate transporter/excitatory amino acid transporter 1 (GLAST/EAAT1) protein. Moreover, miR-142-3p levels are increased in both EAE brains and CSFs of patients with relapsing-remitting MS (RRMS) and correlate with disease progression. Preliminary data also reveal that miR-142-3p is direct target of different pharmacological treatments for MS, while the action of non-pharmacological treatments, as therapeutic transcranial magnetic stimulation (TMS) to ameliorate MS spasticity, is still unknown.
Based on these considerations, a prospective and retrospective cohort study of about six years will be performed to assess whether miR-142-3p is a possible biomarker for MS synaptopathy-driven disease progression (AIM1) and for the efficacy of disease-modifying treatments (DMTs) currently used in MS therapy (AIM2a). Moreover, a genetic screening from peripheral blood will be conducted in order to identify single nucleotide polymorphisms (SNPs) in coding and/or regulating regions of miR-142-3p and GLAST/EAAT1 genes, associated with MS synaptopathy (AIM2b). Finally, a repetitive TMS stimulation protocol (iTBS) will be performed in a subgroup of screened MS patients with lower limb spasticity (interventional substudy) to evaluate the patient responsiveness to the treatment linked to the identified SNPs (AIM2c).
Given the heterogeneity and complexity of MS disease, multivariable approach will permit to dissect miR-142-3p contribution to MS course influenced by synaptopathy (AIM1).
Firstly, miR-142-3p levels in MS CSF (the day of recruitment, T0) will be correlated with other possible variables relevant to disease progression, such as:
* clinical (disease duration, estimated as the number of years from onset to the most recent assessment of disability; disability, evaluated using EDSS = Expanded Disability Status Scale; Progression Index, PI = EDSS/disease duration; change in ARR = Annualized Relapse Rate) and neuroradiological parameters (dual-echo proton density; FLAIR = fluid-attenuated inversion recovery; T2-WI = T2-weighted spin-echo images and T1-WI = pre-contrast and post-contrast T1-weighted spin-echo images after intravenous gadolinium (Gd) infusion) at T0 and once per year during a 6-year-follow-up if no relapse occurs (T12, T24, T36, T48, T60 , T72);
* levels of inflammatory and potential excitotoxic protein factors (as IL-1β, TNF and RANTES-CCL5) in the CSF (T0);
* levels of neurofilaments, beta amyloid, tau proteins and growth factors (like NGF, PDGF and BDNF) in the CSF, as possible indicators of neurodegenerative and regenerative processes occurring at the CSF withdrawal (T0).
To reduce the variable dimension, Principal Component Analysis (PCA) will be applied taking into account the contribution of miR-142-3p to disease progression as part of a complex network of molecules circulating in the CSF, and univariable and multivariable correlations will be repeated.
In multivariable analysis (based on multivariable generalized linear models, GLM), miR-142-3p levels in the CSF (or PCA components including miR-142-3p as part of the component) will be considered as the independent variable adjusting for demographical, clinical and neuroradiological values as well as different DMT treatments. A further analysis based on treatment stratifications of the patients will be attempted (AIM2a).
Lastly, the CSF levels of miR-142-3p (or PCA components including miR-142-3p) identified to associate with disease progression variables will be correlated with neurophysiological parameters, recorded by means of TMS to evaluate cortical excitability and plasticity (SICI = short interval intracortical inhibition; ICF = intracortical facilitation; LICI = long interval intracortical inhibition; PAS = Paired Associative Stimulation) in MS patients at T0. Thus, miR-142-3p circulating in the CSF will be validated as possible biomarkers of synaptopathy-driven disease progression (as single molecules or as part of a PCA component).
To identify genetic variants of miR-142-3p and GLAST/EAAT1 coding genes relevant to MS synaptopathy (AIM2b) SNPs will be analyzed at T0 and will be correlated with miR-142-3p levels in the CSF and with other possible variables relevant to disease progression as in AIM1. PCA and GLM models will be applied as in AIM1.
To evaluate treatment responsiveness in the subgroup of screened MS patients included in the interventional substudy based on a two-week protocol of iTBS for reducing lower limb spasticity, the H/M amplitude ratio of the Soleus H reflex and the Modified Ashworth Scale (MAS) will be considered before (W0) and after (W2) the stimulation protocol. Possible association between patient responsiveness to the iTBS stimulation protocol and specific SNPs will be assessed (AIM2c).
Statistical analysis will be performed using Prism GraphPad 6.0, IBM SPSS Statistics 15.0, R software and T-MEV 4.4.1. Data will be tested for normality distribution through the Kolmogorov-Smirnov and Shapiro-Wilk tests. The k-means method will be used to divide MS patients into homogeneous clusters, based on miR-142-3p levels in the CSF and other relevant parameters. Differences between two groups will be analyzed using Student's t-test, Mann-Whitney test, Fisher exact test or log-rank test, as appropriate; multiple comparisons will be performed by ANOVA followed by Tukey HSD or by Kruskal-Wallis. Pearson or nonparametric Spearman correlation coefficients will be performed to evaluate the association of miR-142-3p levels in the CSF or specific genetic variants of MIR142 and SLC1A3 (or the correspondent PCA component, see next) with continuous demographic, clinical and neuroradiological parameters (e.g Age, changes in EDSS, Number of T2 lesions, etc.). For the multiple comparisons it will be controlled the False Discovery Rate (FDR) applying the method proposed by Benjamini and Hochberg.
PCA will be applied to represent sets of potentially correlated variables (CSF levels of miR-142-3p or specific genetic variants of MIR142 and SLC1A3, inflammatory and potential excitotoxic protein factors and levels of neurofilaments, beta amyloid, tau protein and growth factors) with principal components (PC) that are linearly uncorrelated obtained using orthogonal transformation. PCs are ordered so that the first PC has the largest possible variance and only some components are selected to represent the correlated variables. As a result, the dimension of the variables is reduced.
To validate miR-142-3p as biomarker of synaptopathy-driven disease progression (measured in terms of clinical or radiological changes and TMS variables) or specific SNPs of MIR142 and SLC1A3 linked to MS synaptopathy, GLM models will be applied considering, respectively, the miR-142-3p level in the CSF (or the identified PCA components including miRs) or the genetic variants as an independent variable adjusting for demographical, clinical, neuroradiological, neurophysiological, biochemical factors and treatments.
Data will be presented as the mean (standard deviation, sd) or median (25th- 75th percentile). The significance level is established at p\<0.05.
