Zheng Fuze, Lin Yawen, Qiu Liangliang, Zheng Ying, Zeng Minghui, Lin Xiaodan, He Qifang, Lin Yuhua, Chen Long, Lin Xin, Chen Xinyue, Lin Lin, Wang Lili, He Junjie, Lin Feng, Yang Kang, Wang Ning, Lin Minting, Lian Sheng, Wang Zhiqiang
Department of Neurology and Institute of Neurology of First Affiliated Hospital, Institute of Neuroscience, and Fujian Key Laboratory of Molecular Neurology, Fujian Medical University, Fuzhou 350005, China.
College of Computer and Data Science, Fuzhou University, and Fujian Key Laboratory of Network Computing and Intelligent Information Processing, Fuzhou 350108, China.
Brain. 2025 Feb 3;148(2):613-625. doi: 10.1093/brain/awae309.
Facioscapulohumeral muscular dystrophy type 1 (FSHD1) patients exhibit marked variability in both age at onset (AAO) and disease severity. Early onset FSHD1 patients are at an increased risk of severe weakness, and early onset has been tentatively linked to the length of D4Z4 repeat units (RUs) and methylation levels. The present study explored potential relationships among genetic characteristics, AAO and disease severity in FSHD1. This retrospective and observational cohort study was conducted at the Fujian Neuromedical Centre (FNMC) in China. Genetically confirmed participants with FSHD1 recruited from 2001 to 2023 underwent distal D4Z4 methylation assessment. Disease severity was assessed by FSHD clinical score, age-corrected clinical severity score (ACSS) and onset age of lower extremity involvement. Mediation analyses were used to explore relationships among genetic characteristics, AAO and disease severity. Finally, machine learning was employed to explore AAO prediction in FSHD1. A total of 874 participants (including 804 symptomatic patients and 70 asymptomatic carriers) were included. Multivariate Cox regression analyses indicated that male gender, low DUZ4 RUs, low CpG6 methylation levels, non-mosaic mutation and de novo mutation were independently associated with early onset in FSHD1. Early onset patients (AAO < 10 years) had both a significantly higher proportion and an earlier median onset age of lower extremity involvement compared to the typical adolescent onset (10 ≤ AAO < 20 years), typical adult onset (20 ≤ AAO < 30 years) and late onset (AAO ≥ 30 years) subgroups. AAO was negatively correlated with both clinical score and ACSS. We found that AAO exerted mediation effects, accounting for 12.2% of the total effect of D4Z4 RUs and CpG6 methylation levels on ACSS and 38.6% of the total effect of D4Z4 RUs and CpG6 methylation levels on onset age of lower extremity involvement. A random forest model that incorporated variables including gender, age at examination, inheritance pattern, mosaic mutation, D4Z4 RUs and D4Z4 methylation levels (at CpG3, CpG6 and CpG10 loci) performed well for AAO prediction. The predicted AAO (pAAO) was negatively correlated with ACSS (Spearman's ρ = -0.692). Our study revealed independent contributions from D4Z4 RUs, D4Z4 methylation levels, mosaic mutation and inheritance pattern on AAO variation in FSHD1. AAO mediates effects of D4Z4 RUs and methylation levels on disease severity. The pAAO values from our random forest model informatively reflect disease severity, offering insights that can support efficacious patient management.
1型面肩肱型肌营养不良症（FSHD1）患者在发病年龄（AAO）和疾病严重程度方面均表现出显著差异。早发型FSHD1患者出现严重肌无力的风险增加，且早发初步被认为与D4Z4重复序列（RUs）的长度和甲基化水平有关。本研究探讨了FSHD1患者的遗传特征、发病年龄和疾病严重程度之间的潜在关系。这项回顾性观察队列研究在中国福建神经医学中心（FNMC）进行。对2001年至2023年招募的经基因确诊的FSHD1参与者进行了远端D4Z4甲基化评估。通过FSHD临床评分、年龄校正临床严重程度评分（ACSS）和下肢受累发病年龄评估疾病严重程度。采用中介分析来探讨遗传特征、发病年龄和疾病严重程度之间的关系。最后，运用机器学习来探索FSHD1患者发病年龄的预测。共纳入874名参与者（包括804名有症状患者和70名无症状携带者）。多变量Cox回归分析表明，男性、低DUZ4 RUs、低CpG6甲基化水平、非嵌合突变和新发突变与FSHD1的早发独立相关。与典型青少年发病（10≤AAO<20岁）、典型成人发病（20≤AAO<30岁）和晚发（AAO≥30岁）亚组相比，早发患者（AAO<10岁）下肢受累的比例显著更高，且下肢受累的中位发病年龄更早。发病年龄与临床评分和ACSS均呈负相关。我们发现发病年龄发挥中介作用，占D4Z4 RUs和CpG6甲基化水平对ACSS总效应的12.2%，以及D4Z4 RUs和CpG6甲基化水平对下肢受累发病年龄总效应的38.6%。一个纳入性别、检查时年龄、遗传模式、嵌合突变、D4Z4 RUs和D4Z4甲基化水平（在CpG3、CpG6和CpG10位点）等变量的随机森林模型在发病年龄预测方面表现良好。预测的发病年龄（pAAO）与ACSS呈负相关（Spearman相关系数ρ=-0.692）。我们的研究揭示了D4Z4 RUs、D4Z4甲基化水平、嵌合突变和遗传模式对FSHD1患者发病年龄变异的独立贡献。发病年龄介导了D4Z4 RUs和甲基化水平对疾病严重程度的影响。我们随机森林模型得出的pAAO值能有效反映疾病严重程度，为支持有效的患者管理提供了见解。
