第一部分 DEPDC5基因相关癫痫患者的基因型谱及表型谱

目的：DEPDC5基因相关癫痫抗发作药物（anti-seizure medications, ASM）疗效差异大，极少数患者不经治疗随访多年无发作，少部分患者单药即可控制发作，50%以上为药物难治。本研究旨在分析DEPDC5基因相关癫痫在多学科诊疗模式下的临床和遗传学特征及治疗转归；结合既往报道病例总结基因型谱及表型谱，阐明基因型-表型相关性，探究药物难治的危险因素，为个体化诊疗、评估预后及遗传咨询提供理论依据。

方法：收集2016年5月至2024年2月于首都儿科研究所院及北京清华长庚医院神经内科就诊的DEPDC5基因相关癫痫患者的病例资料，分析患者临床表现、脑电图（Electroencephalogram, EEG）、影像学、遗传学特征及治疗转归。以“DEPDC5”or“GATOR1”为关键词分别检索Pubmed、Web of Science及中国知网、万方数据库建库至2024年2月的文献。总结患者性别、起病年龄、癫痫表型、发育里程碑、EEG及核磁共振成像（Magnetic resonance imaging, MRI）表现、治疗转归及基因型等；利用χ²检验或Fisher确切概率法，探究不同遗传方式、变异类型及变异位点分布与癫痫起病年龄、精神运动发育、头颅MRI表现及药物治疗效果的相关性。利用多因素Logistic回归模型分析药物难治的相关危险因素。

结果：本组19例患者中，37%（7/19）≤1岁起病；多种癫痫发作类型，57.9%（11/19）为局灶性运动发作，8例为各类型癫痫综合征，其中5例家族性局灶性癫痫综合征伴可变灶、2例婴儿癫痫性痉挛综合征、1例睡眠相关过度运动性癫痫（Sleep-related hypermotor epilepsy, SHE）；37%（7/19）并存精神运动发育迟滞；15.8%（3/19）并存孤独症等共患病；94%（16/17）间歇期EEG监测到癫痫样放电，其中41.2%（7/17）为脑区性放电，41.2%（7/17）多种放电形式并存，1例（6%）双侧多脑区性放电，1例（6%）一侧性放电，余1例（6%）间歇期EEG正常；64.7%（11/17）MRI存在脑结构异常。本组共19个变异位点，13个既往未报道，其中84.2%（16/19）为截短变异，10.5%（2/19）为错义变异，5.3%（1/19）整码缺失；89.5%（17/19）遗传自亲代，10.5%（2/19）为新生变异。随访时间为6个月至7年5个月，多学科诊疗模式下癫痫无发作率为68%（13/19），其中1例（5%）病初发作稀少，未服用ASM，随访5年无发作；21.1%（4/19）单药治疗无发作，15.8%（3/19）2种以上ASM联合治疗无发作；余57.9%（11/19）为药物难治性癫痫（Drug-resistant epilepsy, DRE）。11例DRE患者中，5例经综合术前评估后行致痫灶切除，术后随访2年3月至5年8月均无发作，术后病理2例为局灶性皮质发育不良（Focal cortical dysplasia, FCD）IIa，1例FCDIIb，1例FCDIII及1例胶质细胞增多。

本组及文献报道共255例患者。约三分之一≤1岁起病；癫痫表型主要为不同

病灶起源的局灶性癫痫，SHE是最常见的癫痫综合征；37.7%存在脑结构异常；约67%为局灶性或脑区性放电；约65%为DRE。DEPDC5变异以截短变异为主（82%），其次是错义变异（13.7%）；变异多遗传自亲代（87.6%）；p.Arg243*、p.Arg843*及p.Arg1087*为最常见的变异位点。基因型-表型相关性分析表明截短变异者DRE比例（67.6%）显著高于错义变异（46.7%）（χ²=4.916，P=0.027）；与其他4个结构域相比，变异位点位于CTD者多＞1岁起病（P＜0.05）；与LassoB非结构域（40%）相比，变异位点位于LassoA者DRE比例（81.3%）更高（χ²=9.253，P=0.002）；变异位点相同，表型亦可轻重不一。≤1岁起病（OR=0.370, 95%CI: 0.171~0.798, P=0.011）及头颅MRI异常（OR=0.459, 95%CI: 0.212~0.990, P=0.047）是药物难治的独立危险因素。

结论：约三分之一DEPDC5基因相关癫痫患者≤1岁起病，以不同病灶起源的局灶性癫痫为主，间歇期EEG常见多脑区或弥漫性放电或多种放电形式并存；37.7%患者存在脑结构异常；约65%患者为DRE，经多学科综合评估行致痫灶切除术后多无发作。DEPDC5变异以截短变异为主，其次是错义变异。截短变异者DRE比例高于错义变异；与其他结构域相比，变异位点位于CTD者多＞1岁起病；与LassoB非结构域相比，变异位点位于LassoA者DRE更多；变异位点相同，表型亦可轻重不一。≤1岁起病及头颅MRI异常是药物难治的独立危险因素。

第二部分 DEPDC5基因相关癫痫的iPSC模型建立及发病机制研究

目的：DEPDC5基因相关癫痫药物难治者高达65%，其发病机制不明，本研究旨在建立DEPDC5变异患者来源的诱导性多能干细胞（Induced Pluripotent Stem Cell, iPSC）并探究其致痫机制。

方法：利用1例于2月龄起病且存在FCD的DEPDC5基因相关癫痫患者（c.247C>T/p.Gln83*）及健康父亲的外周血单核细胞（Peripheral Blood Mononuclear Cell, PBMC）构建iPSC，通过免疫荧光染色检测细胞表面标志物、核型分析、体外分化潜能鉴定、变异分析及支原体检测进行iPSC鉴定；将iPSC诱导分化为谷氨酸能神经元，免疫荧光染色检测神经元标志物，PCR及Western blot检测DEPDC5 mRNA及蛋白表达量；利用高内涵成像观测DEPDC5变异对神经元形态发育的影响及其对细胞抗氧化应激损伤能力的影响；利用多微电极阵列（Multi-electrode array, MEA）监测DEPDC5变异对神经元电生理功能的影响。

结果：利用DEPDC5基因相关癫痫患者及健康父亲的PBMC构建iPSC并成功诱导分化为谷氨酸能神经元。与野生型相比，变异型神经元DEPDC5 mRNA和蛋白表达水平明显下降。与野生型相比，变异型神经元在神经元诱导分化24h、72h和96h胞体面积显著增大，在0h、24h、48h和72h神经突总长度及平均长度均显著增长，在0h、24h和72h神经突宽度显著增宽，在48h和72h神经元突触分支节点总数、分支节点平均数及分支节点至胞体距离均显著增加（均P＜0.05）；神经元抗氧化应激损伤能力两组间比较差异无统计学意义。在神经元诱导分化第27天，变异型神经元平均放电频率、脉冲数量、脉冲频率及脉冲持续时间均显著高于野生型（均P＜0.05）；神经网络爆发变异系数及神经元放电的同步化指数组间比较差异无统计学意义。

结论：构建人源DEPDC5变异iPSC细胞模型，为探究DEPDC5基因相关癫痫的发病机制提供可靠的细胞模型；在人源模型证实DEPDC5变异导致谷氨酸能神经元形态发育异常，即胞体面积增大，神经突长度、宽度及突触分支复杂性增加，并导致神经元电生理功能异常，即兴奋性增加。  

PARTI  The Genotypic and phenotypic spectrum of DEPDC5-related epilepsy

Objective: More than 50% of probands remain refractory despite combination therapy. The study aimed to analyse the clinical and genetic characteristics and prognosis of DEPDC5-related epilepsy in multidisciplinary team. In addition, we integrated previously published cases to summarise the genotype-phenotype spectrum and elucidated the genotype-phenotype correlation and risk factors for drug resistance.

Methods: The clinical and genetic characteristics of probands with DEPDC5-related epilepsy in our cohort were extracted and analyzed. A literature search was performed using PubMed, Web of Science, CNKI, and Wanfang databases with the key words "DEPDC5" or "GATOR1" (up to February 2024). Patient demographics, age of onset, epilepsy phenotype, developmental milestones, EEG, MRI findings, outcomes, and genotype were summarized. Correlations between different inheritance modes, mutations types, and mutation sites with age of epilepsy onset, psychomotor development, MRI findings, and drug effect were explored using the χ² test or Fisher's exact test. Risk factors associated with drug resistance were analyzed using a multivariable logistic regression model.

Results: Of the 19 cases in our cohort, 37% (7/19) had onset at ≤ 1 year of age and the seizure types varied, with 57.9% (11/19) presenting focal motor seizures. 5 probands presented with FFEVF, 2 with IS and 1with SHE. Developmental delay was reported in 37% (7/19) of the probands and comorbidities in 15.8% (3/19). Interictal EEG suggested epileptiform discharges in 94% (16/17) of probands, of which 41.2% (7/17) were focal, 41.2% (7/17) multiple forms, 1(6%) multifocal, 1 (6%) unilateral and 1 (6%) normal. MRI abnormalities in 64.7% (11/17). 13 mutations were previously unreported; 84.2% (16/19) were truncating mutations, 10% (2/19) were missense mutations; 5.3% (1/19) were in-frame deletions, 89.5% (17/19) were inherited mutations. Follow-up ranged from 6 months to 7 years and 5 months, 68% (13/19) of probands were seizure-free in the multidisciplinary treatment, of which 1 case (5%) remained seizure-free for 5 years without treatment for sporadic seizures; 21.1% (4/19) were seizure-free with monotherapy, and 15.8% (3/19) with≥2 ASMs. The remaining 57.9% (11/19) were DRE. 5 of 11 underwent epilepsy surgery after comprehensive preoperative evaluation, and they were all seizure-free at postoperative follow-up from 2 years and 3 months to 5 years and 8 months, with postoperative pathology of FCDIIa in 2 cases, FCD IIb in 1 

case, FCD III in 1 case and gliocytosis in 1 case.

A total of 255 probands from our corhort and literature were reviewed. About one-third had onset at ≤1 year of age.The epilepsy type was mainly focal epilepsy with variable onset foci, with SHE being the most common epilepsy syndrome. MRI revealed abnormalities in 37.7% of the probands. About 67% had focal or regional discharges, and about 65% had DRE. Truncating mutations predominated (82%), followed by missense mutations (13.7%). Mutations were predominantly inherited from parents (87.6%). p.Arg243*, p.Arg843*, and p.Arg1087* were the most common mutations. Genotype- phenotype associations showed that the proportion of DRE was higher in probands with truncating mutations (67.6%) than in those with missense (46.7%). Mutations in the CTD domain have an age of onset >1 year compared to other domains (P<0.05). A higher proportion of DRE (81.3%) was found in probands with mutations in LassoA non-domains compared to LassoB. Recurrent mutations were associated with both severe and less severe phenotypes.Onset at age ≤ 1 year of age (OR=0.370, 95%CI: 0.171~ 0.798, P=0.011) and abnormal MRI (OR=0.459, 95%CI: 0.212~0.990, P=0.047) were independent risk factors for DRE.

Conclusions: About one-third of patients with DEPDC5-related epilepsy have onset at ≤1 year of age, with predominantly focal epilepsy with variable onset foci, and interictal EEG often shows multifocal or diffuse discharges and even multiple forms. 37.7% have brain abnormality. About 65% of probands are DRE and may benefit from epilepsy surgery after after comprehensive multidisciplinary evaluation. DEPDC5 mutations are predominantly truncating, followed by missense mutations. DRE rates are higher in probands with truncating mutations than missense mutations. Mutations in the CTD domain have an age of onset >1 year compared to other domains. A higher proportion of DRE was found in probands with mutations in LassoA non-domains compared to LassoB. Recurrent mutations are associated with both severe and less severe phenotypes. Onset at age ≤1 year of age and abnormal MRI are independent risk factors for DRE.

PARTII  Generation of iPSC and pathogenesis of

DEPDC5-related epilepsy

Objective: DEPDC5-related epilepsy may exhibit drug-resistance rates as high as 65%. The study aimed to establish iPSC derived from probands with DEPDC5 mutation and to investigate the epileptogenic mechanism.

Methods: The second section used PBMC from a patient with DEPDC5-related epilepsy, manifesting onset at 2 months of age and FCD to generate iPSC. The iPSC was then differentiated into glutamatergic neurons and immunofluorescence staining was performed to detect neuronal markers. High-content analysis was employed to observe morphology and to assess the impact of DEPDC5 mutation for cell on anti-oxidative stress. MEA was utilized to monitor neuronal electrophysiology, whereas PCR and Western blot analyses were performed to assess DEPDC5 expression.

Results: iPSC from PBMC obtained from a patient with DEPDC5-related epilepsy, followed by differentiation into glutamatergic neurons. The levels of DEPDC5 mRNA and protein expression were significantly diminished in the patient's neurons. Atypical neuronal morphology was noted. In comparison to the control group, neurons derived from DEPDC5 exhibited larger soma size, increased total and mean neurite length, wider neurites, and more intricate neurite branching at various time points, respectively. There were no significant differences observed in anti-oxidative stress between the two groups.  On day 27 of differentiation, DEPDC5 neurons demonstrated significantly higher mean firing rate, number of bursts, burst frequency, and burst duration compared to the control group (all P < 0.05). No statistically significant differences in synchrony between the two groups.

Conclusions: The hiPSC platform offers a reliable model for investigating the pathogenesis of DEPDC5-related epilepsy. DEPDC5 mutation leads to abnormal morphological development and increases excitability of glutamatergic neurons.