Baseline miR-10a, miR-146a, and miR-155 expression and subsequent clinical and radiological outcomes in multiple sclerosis

Abstract

Multiple sclerosis (MS) is a chronic, immune-mediated central nervous system disorder with a multifactorial etiology, driven by genetic–environmental interplay. MicroRNAs (miRNAs), small non-coding RNAs, act as key post-transcriptional regulators within epigenetic mechanisms, integrating these determinants into gene expression networks. Increasing evidence implicates miRNA dysregulation as an important contributor to MS pathogenesis and progression. Between 2015 and 2018, expression levels of miR-10a, miR-146a, and miR-155 were obtained from whole blood samples of patients with MS using RT-PCR and evaluated to investigate associations between baseline miRNA expression patterns and subsequent clinical and radiological outcomes. Based on these data, the medical records of 54 patients who remained under follow-up (2015–2023) were reviewed retrospectively. Following expression analysis, the number of relapses, annualized relapse rate, changes in Expanded Disability Status Scale (EDSS) score, radiological activity, and achievement of No Evidence of Disease Activity (NEDA-3) status were assessed over a mean follow-up period of approximately 5 years. A total of 54 patients (40 female and 14 male) were included. The mean follow-up period was 62.44 ± 21.14 months (14–81 months). The median age at diagnosis was 31 years (16–53 years). Differences in miR-10a expression levels were found in 37 patients (68.5%), miR-146a in 36 (66.7%), and miR-155 in 41 (75.9%). miR-10a dysregulation was significantly associated with the final EDSS score (P = .007) and the presence of cervical spinal cord lesions at baseline (P = .040). miR-146a and miR-155 dysregulation were significantly associated with the development of new cervical spinal cord lesions (P = .020 and P = .025, respectively). miR-155 also showed a nominal association with overall radiological activity (P = .049). Exploratory ROC analysis showed limited discriminatory performance of miR-155 for radiological activity (AUC = 0.655, 95% CI: 0.513–0.779; P = .0387). In this exploratory retrospective study, baseline expression patterns of miR-10a, miR-146a, and miR-155 were associated with selected subsequent clinical and radiological outcomes in patients with MS. Associations were observed between miR-10a and disability measures, between miR-146a/miR-155 and new cervical spinal cord lesions, and between miR-155 and radiological activity. These findings should be considered hypothesis-generating and require confirmation in larger prospective studies with longitudinal miRNA assessment and independent validation.

1. Introduction

Multiple sclerosis (MS) is a chronic, immune-mediated disorder of the central nervous system (CNS) in which inflammatory, demyelinating, and neurodegenerative processes intersect to drive disease activity and progression. The condition predominantly manifests in young adulthood, and the onset of irreversible disability at this life stage carries substantial individual and societal consequences. The relapsing–remitting phenotype, the most common clinical presentation, is characterized by pronounced inflammatory activity and frequent relapses in the early course. Over subsequent years, the frequency and severity of these relapses typically diminish, with neurodegenerative mechanisms emerging as the principal drivers of clinical deterioration. Contemporary evidence suggests that inflammation and neurodegeneration coexist from disease onset, with their relative contributions varying between individuals.^[1,2]

The etiology of the disease is multifactorial, arising from complex interactions between genetic susceptibility and environmental exposures, including cigarette smoking, vitamin D deficiency, obesity during early life, and prior infection with Epstein–Barr virus. MicroRNAs function as key epigenetic regulators mediating these interactions. These small (~19–24 nucleotides in length), endogenous, and evolutionarily conserved non-coding RNAs play a crucial role in the post-transcriptional regulation of gene expression.^[3] MicroRNA (miRNAs) modulate key processes within both the innate and adaptive immune systems, thereby influencing immunological responses relevant to disease pathophysiology. Through such regulatory effects, they have the potential to alter disease characteristics and progression patterns. Epigenetic mechanisms, including miRNA-mediated regulation, may be actively involved not only in disease initiation but also in determining prognosis. This epigenetic influence may account for the marked heterogeneity in disease course observed among patients with similar demographic characteristics, disease duration, and histories of immunomodulatory therapy. Moreover, these mechanisms may contribute to the neurodegenerative phase of the disease, potentially rendering it more prominent, more frequent, or more rapidly progressive.

In current research on MS, 2 miRNAs have been identified as molecules of particular interest: microRNA 155 (miR-155) and microRNA 146a (miR-146a).^[4] Available evidence identifies miR-155 as an “inflamma-miR,” given its role as a potent promoter of inflammatory processes. Functionally and phenotypically, miR-155 acts as a critical mediator in the pathogenesis of autoimmune diseases, primarily through its influence on myeloid cell polarization. In contrast, miR-146a functions as a key regulator in the negative feedback control of inflammatory signaling pathways.^[5] One study investigating miRNAs involved in the innate immune cell response to microbial infection identified miR-132, miR-155, and miR-146a, which exhibited marked upregulation upon bacterial lipopolysaccharide stimulation.^[6] In addition to these, microRNA 10a (miR-10a) serves as an important post-transcriptional regulator modulating inflammatory responses. The miR-10 gene family has been noted for its genomic location within HOX clusters, which function as critical developmental regulators, and its high degree of evolutionary conservation across species.^[7]

Despite significant advances, the molecular mechanisms driving MS remain incompletely elucidated, and the contribution of miRNAs to disease pathogenesis is an area of active investigation. Building on our previous work characterizing miRNA expression profiles in whole blood and stool samples from patients with relapsing–remitting MS (RRMS), the present study aims to delineate the clinical and radiological associations of miR-155, miR-146a, and miR-10a with disease activity, lesion distribution, and subsequent disability progression.^[8,9] Elucidating these relationships may provide further insights into the associations between miRNA expression patterns and the subsequent clinical and radiological course of MS.

2. Materials and methods

2.1. Study design, data collection, and outcomes

This exploratory, retrospective observational study evaluated the long-term clinical and radiological outcomes of the original patient cohort in whom whole-blood expression levels of miR-10a, miR-146a, and miR-155 had been assessed between 2015 and 2018 as part of our previous research project (OUAP(T)2015/23). The present study relates these previously obtained miRNA expression data to subsequent longitudinal clinical and radiological outcomes. The study protocol was approved by the local ethics committee (approval no. 2023-17/18), and all participants waived consent for this study in accordance with the written informed consent form obtained in the previous study. Patient records were retrospectively reviewed for individuals initially selected based on the 2010 revised McDonald criteria. During follow-up, diagnoses were confirmed or newly established in accordance with the 2017 revision of the McDonald criteria. Detailed information on patient recruitment and study design is provided in Figure 1.

Figure 1.

The number of patients included in the current study after follow-up, their diagnoses, and the study design. CIS = clinically isolated syndrome, miRNA = microRNA, MS = multiple sclerosis, RIS = radiologically isolated syndrome, RRMS = relapsing-remitting multiple sclerosis.

Participants were eligible if they had attended follow-up at our center for at least 12 months after the initial expression analysis; had complete clinical records, including number of relapses, Expanded Disability Status Scale (EDSS) scores, and immunomodulatory therapy details; had available follow-up cranial and cervical magnetic resonance imaging (MRI) scans; and met the 2017 revised McDonald criteria for the diagnosis of MS. Patients lost to follow-up or with incomplete data for any of the specified variables were excluded.

After application of these criteria, 54 patients were included in the analysis. Data collected comprised demographic characteristics, number of relapses, annualized relapse rates, immunomodulatory treatments, EDSS scores, and cranial and cervical MRI scans, with thoracic 3T MRI scans included when available. In addition, expression data for miR-10a, miR-146a, and miR-155 from our previous study were incorporated.^[8] Disability progression was defined as an increase in EDSS score of 1.5 points for patients with a baseline EDSS score of 0, 1.0 point for those with a baseline EDSS score <5.5, or 0.5 points for those with a baseline EDSS score ≥5.5, confirmed over at least 6 months.

Radiological analyses were performed by evaluating all imaging studies obtained up to the date of the last visit, provided that cranial and cervical MRI images were obtained at least 12 months after the initial imaging procedure. Radiological activity was defined as new or enlarging T2-hyperintense lesions and/or new contrast-enhancing lesions (CELs) on available follow-up MRI. Evaluated parameters included the number and anatomical localization of newly developed T2-hyperintense demyelinating lesions (cortical/juxtacortical, periventricular, infratentorial, cervical, and thoracic), enlargement of preexisting lesions, and the presence of CELs. MRI examinations were obtained as part of routine clinical follow-up, and clinical and radiological data were retrospectively reviewed by 2 neurologists. Detailed imaging protocol parameters and formal inter-rater reliability data were not uniformly available.

2.2. Expression analysis

Whole-blood expression data for miR-10a, miR-146a, and miR-155 were obtained from our previous research project (OUAP(T)2015/23). Expression data for 2 of these miRNAs were reported previously.^[8] Total RNA was extracted from whole-blood samples using TRIzol Reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol, eluted in 50 μL of preheated Elution Solution, and quantified with a UV–Vis spectrophotometer/NanoDrop (Beckman Coulter). RNA integrity was verified by agarose gel electrophoresis, and samples were stored at–80°C until reverse transcription. cDNA synthesis was performed using the miRCURY LNA™ Reverse Transcription Kit (QIAGEN, Germany) according to the manufacturer’s instructions, using a Bio-Rad T100 Thermal Cycler (Bio-Rad), and cDNA was diluted 1:60 in RNase-free water immediately before qRT–PCR (real-time polymerase chain reaction). Reactions were performed on a LightCycler 480 II system (Roche, Germany) using miRNA-specific LNA™ PCR primer sets for hsa-miR-10a (YP00204778), hsa-miR-146a-5p (YP00204688), and hsa-miR-155 (YP00204308), with U6 snRNA (RNU6, YP00203907) as the endogenous control, and the miRCURY SYBR® Green Master Mix (2×) (QIAGEN, Germany). Data were analyzed using the GeneGlobe Data Analysis Center (QIAGEN; http://www.qiagen.com/geneglobe). Relative miRNA expression was quantified using the comparative 2^−ΔΔCt method, with RNU6 serving as the endogenous reference for normalization. The mean ΔCt value of the healthy control group from the original study was used as the calibrator in the ΔΔCt calculation. Fold-regulation values were derived from the fold-change values according to the GeneGlobe analysis algorithm. A fold-regulation cutoff of 2 was used as the threshold for expression classification.

2.3. Statistical analysis

The data were examined using the Shapiro–Wilk test to determine whether the data presented a normal distribution. The results are presented as mean ± standard deviation, median (minimum-maximum), or frequency and percentage. Kruskal–Wallis and Mann–Whitney U tests were used for non-normally distributed data. The Bonferroni correction was applied to post hoc pairwise comparisons following significant overall Kruskal–Wallis tests. Because this was an exploratory candidate miRNA study evaluating 3 prespecified miRNAs rather than a high-dimensional screening analysis, no additional global multiple-testing correction across different miRNAs was applied. Accordingly, statistically significant findings were interpreted as exploratory and hypothesis-generating. Categorical variables were compared between groups using Pearson chi-square, Fisher exact, and Fisher-Freeman-Halton tests. Pearson correlation coefficient was calculated to determine the relationship between the variables. Statistical significance was set at α = 0.05. Statistical analyses were performed using IBM SPSS ver.28.0 (IBM Corp. Released 2021. IBM SPSS Statistics for Windows, Version 28.0. Armonk: IBM Corp.).

An exploratory receiver operating characteristic (ROC) curve analysis was performed for baseline miR-155 expression in relation to radiological activity. The area under the curve (AUC), 95% confidence interval, sensitivity, specificity, and optimal cutoff value were reported. ROC analyses were performed using MedCalc Statistical Software version 22.006 (MedCalc Software Ltd., Ostend, Belgium). Because of the limited sample size and the limited number of outcome events, multivariable adjustment was not considered statistically appropriate. Accordingly, all reported associations are based on unadjusted exploratory analyses.

3. Results

3.1. Demographic data and clinical findings

A total of 54 patients (40 female and 14 male) were included in the study. The mean follow-up period was 62.44 ± 21.14 months (14–81 months). The median age at diagnosis was 31 years (16–53 years). The mean number of relapses before and after sampling was 3.18 ± 2.36 and 3.83 ± 2.62, respectively. The annual relapse rate was 1.00 ± 1.20 and 0.37 ± 0.22, respectively. The mean EDSS score was 3.03 ± 1.46 before sample analysis and 3.27 ± 1.88 at the final visit. Disability progression was detected in 42.6% (n = 23) of patients. Half of the patients (n = 27) experienced a new MS relapse. The proportion of patients who achieved No Evidence of Disease Activity-3 (NEDA-3) status – defined as the absence of new relapse, disability progression, and radiological activity – was 20.4% (n = 11). Immune modulatory treatments used during expression analysis and at the last visit were recorded as follows: interferon (n = 18; 5), glatiramer acetate (n = 11; 5), teriflunomide (n = 5; 9), dimethyl fumarate (n = 1; 4), and fingolimod (n = 10; 16). In addition, ocrelizumab (n = 7), cladribine (n = 7), and alemtuzumab (n = 1) treatments were initiated. Overall, treatment was changed in 34 of 46 patients (73.9%) during follow-up. New treatment was initiated in 7 of 8 patients who were previously diagnosed with clinically isolated syndrome or radiologically isolated syndrome.

3.2. Radiological findings

A total of 213 MRI examinations, taken within 3 months prior to the expression analysis period and up to the date of the final visit, were evaluated. The most recent imaging was performed within 1 year prior to the last visit. Based on cranial MRI, the median number of imaging studies was 4 (2–10). Among the patients, 63.0% (n = 34) had cortical/juxtacortical lesions, 98.1% (n = 53) had periventricular lesions, 42.6% (n = 23) had infratentorial lesions, and 77.8% (n = 42) had demyelinating lesions in the cervical segment. Of the 10 thoracic MRIs evaluated, lesions were detected in 2 patients.

During the follow-up period, new demyelinating lesions developed in 11.1% (n = 6) of cortical/juxtacortical, 27.8% (n = 15) of periventricular, 7.4% (n = 4) of infratentorial, and 17.0% (n = 9) of cervical spinal cord regions, as well as in the thoracic spinal cord in 2 patients. No patients exhibited enlargement of existing lesions without the development of new lesions.

When all imaging studies were evaluated, the proportion of patients with lesions showing contrast enhancement was 20.4% (n = 11), and when new T2 hyperintense demyelinating lesions/enlargement of existing lesions were included, the total radiological activity rate was 44.4% (n = 24). Given the high rate of baseline periventricular lesions (98.1%) and the small number of thoracic MRIs obtained both at baseline and during follow-up (n = 10), these variables were not statistically analyzed.

3.3. Expression analyses

The mean values for miR-146a, miR-10a, and miR-155 were as follows: Fold Change (FC) 1.22 ± 2.77, 14.31 ± 72.27, and 0.97 ± 2.96; Fold Regulation (FR) −5.06 ± 10.31, −37.63 ± 97.92, and −4.72 ± 5.61, respectively; and Log2 2^–∆∆CT values of −3.07 ± 2.64, −3.70 ± 6.43, and −3.76 ± 2.60, respectively.

Differences in expression levels were observed for miR-146a in 36 patients (66.7%), miR-10a in 37 patients (68.5%), and miR-155 in 41 patients (75.9%). When grouped by expression level differences, an increase in EDSS score was observed in the group with downregulated miR-10a, while the baseline EDSS score was higher in the group with upregulated miR-146a (Table 1). No statistically significant associations were found with other clinical characteristics.

Table 1.

Clinicoradiological characteristics according to miRNA fold-regulation groups.

Clinicoradiological features			Baseline miRNA status
			miR-10a			miR-146a			miR-155
			Upregulation n = 10 (%18.5)	Downregulation n = 37 (%68.5)	No differential expression n = 7 (%13)	Upregulation n = 8 (%14.8)	Downregulation n = 36(%66.7)	No differential expression n = 10(%18.5)	Upregulation n = 5(%9.3)	Downregulation n = 41(%75.9)	No differential expression n = 8(%14.8)
Diagnosis age (yr; <30, >30)			0.063			0.330			0.520
Sex			0.899			0.729			0.050
Number of relapses		Baseline	0.113			0.484			0.633
		Final	0.387			0.564			0.191
		Total	0.309			0.881			0.779
ARR		Baseline	0.906			0.314			0.418
		Final	0.230			0.262			0.236
		Total	0.864			0.497			0.489
EDSS		Baseline	0.315			0.044 (↑)			0.421
		Final	0.294			0.266			0.384
		Increase during follow-up	0.045 (↓)			0.802			1.000
Radiological features (MRI)	Baseline	Cortical/ Juxtacortical	0.755			0.241			0.710
		Infratentorial	0.837			0.084			0.250
		Cervical	0.114			0.231			0.323
	Follow-up after analysis	Cortical/ Juxtacortical	0.821			0.072			0.437
		Periventricular	0.722			0.484			0.251
		Infratentorial	0.545			0.407			1.000
		Cervical	0.248			0.413			0.832
		CEL	0.590			0.525			0.306
		New/Enlarging T2 Lesion	0.594			0.398			0.256

miRNA expression status was classified according to fold-regulation values: ≥+2 (upregulated), ≤−2 (downregulated), and −2 < fold regulation <+2 (no differential expression). Bold P values are significant (<.05).

↓: Downregulated, ↑: Upregulated.

ARR = annualized relapse rate, CEL = contrast-enhancing lesion(s), EDSS = Expanded Disability Status Scale, miRNA = microRNA, MRI = magnetic resonance imaging.

When each patient’s expression analyses were evaluated according to clinical and radiological characteristics, miR-10a dysregulation was associated with the final EDSS score (P = .007) and the presence of cervical segment lesions at baseline (P = .040), while dysregulation of miR-146a and miR-155 was associated with the development of new cervical lesions (P = .020 and P = .025, respectively; Fig. 2). No significant differences were found for other parameters (Table 2).

Figure 2.

Associations between baseline miRNA expression and radiological outcomes during a mean follow-up of 5 years. (A) Dysregulation of miR-146a and miR-155 was associated with the development of new cervical demyelinating lesions. (B) Dysregulation of miR-155 was associated with overall radiological activity, whereas miR-10a and miR-146a showed no significant association. miR-10a = microRNA 10a, miR-146a = microRNA 146a, miR-155 = microRNA 155, miRNA = microRNA, MRI = magnetic resonance imaging.

Table 2.

Clinical and radiological characteristics according to baseline miRNA dysregulation.

Clinicoradiological features			Baseline miRNA expression*
			miR-10a	miR-146a	miR-155
Diagnosis age			0.648	0.769	0.461
Sex			0.992	0.502	0.465
Number of relapses		Baseline	0.806	0.135	0.236
		Final	0.058	0.693	0.270
		Total	0.789	0.288	0.637
ARR		Baseline	0.626	0.179	0.525
		Final	0.253	0.687	0.471
		Total	0.451	0.450	0.998
EDSS		Baseline	0.052	0.527	0.293
		Final	0.007	0.837	0.999
		Increase during follow-up	0.373	0.368	0.345
Radiological Features (MRI)	Baseline	Cortical/Juxtacortical	0.734	0.929	0.872
		Infratentorial	0.827	0.077	0.128
		Cervical	0.040	0.323	0.080
	Follow-up after analysis	Cortical/Juxtacortical	0.841	0.103	0.216
		Periventricular	0.954	0.692	0.306
		Infratentorial	0.644	0.083	0.200
		Cervical	0.227	0.020	0.025
		CEL	0.715	0.217	0.555
		New/Enlarging T2 Lesion	0.632	0.107	0.052

ARR = annualized relapse rate, CEL = contrast-enhancing lesion(s), EDSS = Expanded Disability Status Scale, miR-10a = microRNA 10a, miR-146a = microRNA 146a, miR-155 = microRNA 155, miRNA = microRNA, MRI = magnetic resonance imaging.

^*Log2 2^–∆∆CT, Bold P values are significant (<.05).

Regarding longitudinal outcomes, miR-155 dysregulation showed a nominal association with radiological activity (P = .049; Fig. 2). However, no significant differences were observed for any of the 3 miRNAs with respect to new relapses, disability progression, or NEDA-3 status (Table 3).

Table 3.

Baseline miRNA expression according to clinical and radiological outcomes.

Clinical and radiological outcome		miR-10a		miR-146a		miR-155
		Log2 2–∆∆CT	P value	Log2 2–∆∆CT	P value	Log2 2–∆∆CT	P value
New relapse	Yes	−3.30 ± 2.71	.697	−3.00 ± 8.26	.655	−4.16 ± 2.51	.359
	No	−2.86 ± 2.60		−4.40 ± 3.89		−3.36 ± 2.68
EDSS progression	Yes	−3.46 ± 2.90	.368	−3.40 ± 4.67	.373	−4.24 ± 2.71	.345
	No	−2.96 ± 2.42		−5.33 ± 2.87		−3.63 ± 2.39
Radiological activity	Yes	−3.72 ± 2.55	.107	−3.34 ± 8.45	.632	−4.63 ± 2.02	.049
	No	−2.56 ± 2.64		−3.99 ± 4.33		−3.06 ± 2.83
No evidence of disease activity (NEDA-3)	Yes	−2.97 ± 2.32	.838	−5.59 ± 3.40	.237	−3.37 ± 2.64	.569
	No	−3.10 ± 2.74		−3.22 ± 6.95		−3.86 ± 2.61

Baseline miRNA expression (Log2 2^–∆∆CT), presented as mean ± SD. Bold P values indicate statistical significance (P < .05).

EDSS = Expanded Disability Status Scale, miR-10a = microRNA 10a, miR-146a = microRNA 146a, miR-155 = microRNA 155, miRNA = microRNA, MRI = magnetic resonance imaging, NEDA-3 = no evidence of disease activity-3.

An exploratory ROC analysis was performed for baseline miR-155 expression in relation to radiological activity. miR-155 showed limited discriminatory performance (AUC = 0.655, 95% CI: 0.513–0.779; optimal cutoff ≤−5.31; sensitivity 50.0%; specificity 83.3%; P = .0387; 24 events among 54 patients; Fig. 3).

Figure 3.

Exploratory ROC analysis of baseline miR-155 expression for radiological activity during follow-up. The analysis included 54 patients, of whom 24 had radiological activity and 30 had no radiological activity. The AUC was 0.655 (95% CI: 0.513–0.779; P = .0387). The optimal cutoff value was ≤−5.31, with a sensitivity of 50.0% and specificity of 83.33%. AUC = area under the curve, CI = confidence interval, miR-155 = microRNA 155, ROC = receiver operating characteristic.

4. Discussion

miRNAs, which are key epigenetic regulators, are thought to play a role in the development and clinical course of multiple sclerosis, including both disease activation and progression. miRNAs are generated from hairpin-shaped precursors of double-stranded RNAs (dsRNAs), and hundreds of highly conserved genomic regions encoding miRNAs have been identified in the human genome.^[10] In recent years, increasing interest in miRNA biology has led to a growing number of studies focusing on multiple sclerosis. Investigations in human samples and experimental animal models have identified numerous miRNAs, some of which are upregulated, whereas others are downregulated.^[11-13]

In vitro studies have shown that miR-155 regulates neurosteroid synthesis–related proteins specific to MS lesions in the central nervous system, induces Th1 and Th2 development, promotes macrophage-mediated release and myelin phagocytosis by decreasing CD47 expression, and increases the growth of Th1 and Th17 cells.^[14] Furthermore, miR-155 may promote cytokine production in the brain in response to IL-1β and IFN-γ.^[15] miR-146a is highly expressed in active MS lesions, experimental autoimmune encephalomyelitis, and other autoimmune diseases. NF-κB–stimulated miR-146a forms an essential negative feedback loop that regulates innate immune responses in monocytes. It also increases IL-17 production and promotes astrocyte-associated neuroinflammatory responses by targeting IL-6 and COX-2. Therefore, similar to its role in rheumatoid arthritis and psoriasis, it participates in the reactivation and pro-inflammatory steps in MS.^[16,17]

However, miR-146a may contribute to potentiating NF-κB activation and sustaining the inflammatory response in human astroglial cells.^[18] Data on miR-10a are limited in the literature. It inhibits the IL-6R signaling pathway, acts as one of the senescence-associated factors with an essential role in the biological processes of aging, is linked to aggressive phenotypes of ovarian cancer and the growth of pancreatic cancer, and facilitates the development of osteoarthritis.^[19-23] Inflammation has been reported to exert an anti-inflammatory effect only in sepsis, while inducing inflammatory responses in the hippocampal neurons of epileptic rats.^[24,25]

The heterogeneous clinical course of MS and the unresolved aspects of its pathophysiology contribute to the complexity of long-term disease outcomes. While some patients experience relatively limited disability accumulation, others develop substantial inflammatory activity or sustained disability worsening over time. The factors underlying these differences warrant further investigation.

The present study represents a patient cohort with age and sex distributions consistent with the general MS population. The follow-up period allowed assessment of subsequent clinical and radiological disease activity and sustained disability changes.

Consistent with our earlier findings, baseline miR-155 showed a nominal association with radiological activity rather than clinical relapses during follow-up, while baseline miR-146a expression status was associated with baseline EDSS. Of particular note, downregulation of miR-10a was associated with increased EDSS scores, a more critical clinical parameter of long-term outcome. Individual-level analyses showed a significant association between final EDSS scores and baseline miR-10a expression. Although data on this miRNA remain limited, an experimental study reported higher miR-10a expression in the medulla oblongata and spinal cord in a mouse model.^[26] To date, only 1 study has documented reduced miR-10a expression in MS patients.^[27] In our study, miR-10a expression was associated with cervical segment lesions of the spinal cord at baseline. During the disease course, however, the development of new cervical lesions was associated with baseline miR-155 and miR-146a expression patterns. These findings suggest that baseline miR-10a expression patterns may be associated with cervical spinal cord involvement at baseline, but were not associated with subsequent lesion localization or overall radiological activity during follow-up.

We evaluated associations between baseline miRNA expression patterns and subsequent clinical and radiological outcomes, which constituted the central focus of the study. No significant associations were observed between the 3 miRNAs and new relapses, disability progression, or NEDA-3 status. miR-155 expression patterns were associated with the development of new cervical spinal cord lesions and showed a nominal association with overall radiological activity. Given the reported involvement of miR-155 in inflammatory signaling and its proposed therapeutic relevance in MS, these associations warrant further investigation.^[28] These findings should be interpreted cautiously because of the retrospective design, small sample size, treatment changes during follow-up, and absence of longitudinal miRNA measurements. In particular, treatment changes during follow-up may have influenced relapse occurrence, radiological activity, disability progression, and NEDA-3 status. The observed associations are, therefore, hypothesis-generating and do not establish a causal or prognostic role for these miRNAs. Larger prospective studies incorporating serial miRNA measurements and validation cohorts are needed to determine whether these associations are reproducible and clinically relevant.

5. Limitations

This study related previously obtained baseline miRNA expression data to subsequent clinical and radiological outcomes during follow-up. Because miRNA expression was not reassessed at the final visit or during follow-up, the study cannot determine whether miRNA levels changed in parallel with radiological activity or disability progression. Additional limitations include the incomplete availability of some demographic and clinical variables due to the retrospective study design. Some potentially relevant demographic and clinical factors, such as smoking status, body mass index, lifestyle-related characteristics, and vitamin D supplementation, were not consistently documented throughout follow-up. The relatively small sample size may also have reduced statistical power and contributed to wide confidence intervals in intergroup comparisons. Furthermore, treatment exposure represents an important potential confounder, as disease-modifying therapies were changed during follow-up. Adjustment for treatment exposure was not feasible because of the limited sample size and the heterogeneity of treatment regimens.

Our results provide preliminary insights; however, prospective studies incorporating serial miRNA measurements and larger, more homogeneous patient groups, particularly with respect to disease-modifying treatment exposure, are required to validate these findings and better account for potential confounding factors.

Acknowledgments

We are grateful to Prof Dr Güven Özkaya for his assistance with the statistical analysis of this study.

Author contributions

Conceptualization: Furkan Sarıdaş.

Data curation: Furkan Sarıdaş, Havva Tezcan Ünlü.

Formal analysis: Furkan Sarıdaş, Havva Tezcan Ünlü.

Investigation: Furkan Sarıdaş, Havva Tezcan Ünlü, Özlem Taşkapılıoğlu.

Methodology: Furkan Sarıdaş, Havva Tezcan Ünlü, Gülşah Çeçener.

Project administration: Furkan Sarıdaş.

Resources: Furkan Sarıdaş, Havva Tezcan Ünlü, Gülşah Çeçener, Özlem Taşkapılıoğlu, Ömer Faruk Turan.

Supervision: Gülşah Çeçener, Özlem Taşkapılıoğlu, Ömer Faruk Turan.

Validation: Havva Tezcan Ünlü, Gülşah Çeçener.

Visualization: Furkan Sarıdaş, Havva Tezcan Ünlü.

Writing – original draft: Furkan Sarıdaş, Havva Tezcan Ünlü.

Writing – review & editing: Gülşah Çeçener, Özlem Taşkapılıoğlu, Ömer Faruk Turan.