Thai Karine, Rebillard Rose-Marie, Klement Wendy, Ayoub Othmane, Tastet Olivier, Ben Ahmed Skander, Zierfuss Bettina, Grasmuck Camille, Tea Fiona, Bourbonniere Lyne, Margarido Clara, Hoornaert Chloé Juliette, Carrier Francis, Gowing Elizabeth, Dubé Mathieu, Zandee Stephanie, Girard Marc, Duquette Pierre, Lahav Boaz, Arbour Nathalie, Larochelle Catherine, Prat Alexandre
Department of Neuroscience, Université de Montréal, Montréal, Canada.
Centre de Recherche du Centre Hospitalier de l'Université de Montréal, Montréal, Canada; and.
Neurol Neuroimmunol Neuroinflamm. 2025 Sep;12(5):e200426. doi: 10.1212/NXI.0000000000200426. Epub 2025 Jul 3.
Multiple sclerosis (MS) is an immune-mediated demyelinating disease of the CNS characterized by a heterogeneous disease trajectory, highlighting the need for biomarkers to predict disease activity. Current disease-monitoring tools primarily reflect existing disease damage rather than impending activity. Peripheral blood mononuclear cells (PBMCs) are an ideal source of potential biomarkers due to their accessibility and their known role in MS pathology. Among PBMCs, myeloid cells are key players in MS pathogenic processes, yet they have not been as extensively studied than lymphocytes. The objective of our study was to identify indicators of MS disease activity through immune profiling.
We analyzed PBMCs using high-dimensional flow cytometry with a panel focusing on myeloid cells. We performed unsupervised clustering analyses to define a comprehensive immune landscape at a single-cell resolution. Supervised machine learning methods were used to extract immune features indicative of MS activity.
We analyzed PBMCs from 135 individuals with MS with retrospective longitudinal follow-up and 44 healthy controls (HCs). Among the individuals with MS, 53 were untreated and were compared with HCs. Using an elastic-net model, 20 immune features were identified as contributors to the classification of MS and HCs (receiver operating characteristic-AUC 0.8881). To explore associations between immune features and disease activity, we focused on individuals with relapsing-remitting MS (n = 106). We identified a subpopulation of classical monocytes (CMs) with high expression of human leukocyte antigen - DR isotype (HLA-DR) and positive for CD141 (HLA-DRCD141) as a predictor of impending relapses over 2 years (hazard ratio [HR] 2.8, 95% CI 1.6-4.7) and disability worsening in patients with higher relapse activity. HLA-DRCD141 CMs could be retrieved by manual gating using 9 parameters and were similarly indicative of 2-year relapse risk (HR 1.9, 95% CI 1.3-2.8), highlighting its potential as a practical, translational approach. Compared with the widely studied biomarker serum neurofilament light chain reflecting acute activity, HLA-DRCD141 CMs provided a stronger prognostic value for impending relapse risk, suggesting different kinetics related to the underlying pathology.
Our findings suggest that the frequency of HLA-DRCD141 CMs could serve as a valuable predictor of disease activity complementary to current clinical tools to guide evidence-based treatment decisions.
多发性硬化症（MS）是一种中枢神经系统的免疫介导性脱髓鞘疾病，其疾病进程具有异质性，这凸显了对预测疾病活动的生物标志物的需求。当前的疾病监测工具主要反映已有的疾病损伤，而非即将发生的疾病活动。外周血单个核细胞（PBMC）因其易于获取以及在MS病理过程中已知的作用，是潜在生物标志物的理想来源。在PBMC中，髓样细胞是MS致病过程中的关键参与者，但与淋巴细胞相比，它们尚未得到广泛研究。我们研究的目的是通过免疫谱分析确定MS疾病活动的指标。
我们使用聚焦于髓样细胞的高维流式细胞术分析PBMC。我们进行无监督聚类分析，以单细胞分辨率定义全面的免疫格局。使用监督式机器学习方法提取指示MS活动的免疫特征。
我们分析了135例有回顾性纵向随访的MS患者和44例健康对照（HC）的PBMC。在MS患者中，53例未接受治疗，并与HC进行比较。使用弹性网络模型，确定了20个免疫特征作为MS和HC分类的贡献因素（受试者操作特征曲线下面积[AUC]为0.8881）。为了探索免疫特征与疾病活动之间的关联，我们重点关注复发缓解型MS患者（n = 106）。我们确定了一个人类白细胞抗原-DR同种型（HLA-DR）高表达且CD141呈阳性的经典单核细胞（CM）亚群（HLA-DRCD1 | 41），作为未来2年即将复发的预测指标（风险比[HR]为2.8，95%置信区间[CI]为1.6 - 4.7），并且在复发活动较高的患者中残疾情况会恶化。HLA-DRCD141 CM可以通过使用9个参数的手动设门来获取，并且同样指示2年复发风险（HR为1.9，95% CI为1.3 - 2.8），突出了其作为一种实用的转化方法的潜力。与广泛研究的反映急性活动的生物标志物血清神经丝轻链相比，HLA-DRCD141 CM对即将复发风险具有更强的预后价值，表明与潜在病理相关的不同动力学。
我们的研究结果表明，HLA-DRCD141 CM的频率可以作为疾病活动的有价值预测指标，补充当前的临床工具，以指导基于证据的治疗决策。
