研究背景

睑裂狭小-上睑下垂-倒向型内眦赘皮综合征（blepharophimosis-ptosis-epicanthus inversus syndrome, BPES）共分为两型：I型和II型，两种类型均包括眼部发育异常，其中I型伴有女性不孕。目前已经很多报道了许多手术矫正的方法，但均存在局限性，分一期还是二期进行手术也留有很多争议。作为一种罕见的常染色体显性遗传病，在致病原因方面，主流观点认为BPES是由于Forkhead box L2（FOXL2）基因突变所导致的。人类的FOXL2基因是位于染色体3q22.3区域的单外显子基因，通过调控下游基因StAR在维持卵泡发育、卵巢正常功能方面起着重要作用。突变基因如何对下游基因启动子的转录功能产生影响，导致BPES表型的具体分子机制尚未完全明确。有学者在动物模型中敲除FOXL2基因后观察到了与人类BPES相似的眼睑表型，并发现在小鼠神经嵴细胞（neural crest cells, NCCs）和中胚层细胞中，FOXL2基因对眼睑发育和提上睑肌的发育至关重要，但是尚未发现有关该基因在人神经嵴细胞和中胚层中对眼睑发育影响的相关报道。

研究目的

1.通过对我们改良的一期小睑裂畸形手术方案进行回顾性分析，评估其有效性和安全性。

2.收集临床患者家系的外周血样本，利用全外显子测序技术（whole exome sequencing, WES）筛查FOXL2基因的新突变位点，并检测这些突变对蛋白质结构、下游基因表达和细胞内定位的影响。

3.为了模拟体外发育过程，我们将人胚胎干细胞诱导神经嵴细胞，并将神经嵴细胞诱导为间充质干细胞（mesenchymal stem cells, MSC）。使用慢病毒载体在神经嵴细胞和间充质干细胞中敲除FOXL2基因，以探究该基因对细胞凋亡的影响。

研究方法

1.回顾性分析了2017年1月至2022年6月期间就诊的25例BPES患者（50只眼）。我们使用改良反向Z字成形联合内眦韧带切开法矫正内眦畸形，使用 “V-Y”推进皮瓣联合睑缘结膜瓣开大外眼角，采用眼轮匝肌-额肌瓣复合组织瓣矫正上睑下垂。术前和术后进行眼睑周围数据测量，包括内眦间距（inner intercanthal distance, IICD）、瞳距（interpupillary distance, IPD）、水平睑裂长度（horizontal palpebral fissure length , HPFL）、上睑缘角膜映光距离（marginal reflex distance-1, MRD-1）、IICD/IPD比值、IICD/HPLF比值、温哥华瘢痕量表（vancouver scar scale, VSS）和总体术后满意度。术前和术后采集患者面部照片。术后6-15个月进行随访，使用温哥华瘢痕量表对患者的瘢痕进行评分，评分范围为0-18分，分数越高表明瘢痕越严重。

2.采集患者家系成员的外周血进行全外显子测序，使用PolyPhen-2、SIFT和MutationTaster软件对突变位点进行预测，并使用AlphaFold2进行蛋白质构象预测。构建转染质粒：野生型为pEGFP-WT和pcDNA-WT，携带c.292T>A、c.383G>T的突变质粒分别命名为pEGFP-MT1（c.292T>A）和pEGFP-MT2（c.383G>T）、pcDNA-MT1（c.292T>A）和pcDNA-MT2（c.383G>T）。将人的StAR的启动子构建到pGL3-basic中作为荧光素酶报告载体。转染到HEK-293T细胞中进行双荧光素酶报告实验、细胞免疫荧光实验、实时（RT）荧光定量PCR实验。

3.我们将胚胎干细胞（H9 hESCs）注射到NOD-SCID鼠大腿肌肉间隙进行畸胎瘤成形实验。随后将H9 hESCs诱导分化形成NCCs（H9-NCCs），再将H9-NCCs分化为间充质干细胞（H9-NCCs-MSC），使用流式细胞术、免疫荧光和实时（RT）荧光定量 PCR对这两种细胞进行鉴定。并将构建好的慢病毒载体对这两种细胞进行转染，检测两种细胞的凋亡情况。

研究结果

1.共纳入25名患者（12名女性，13名男性）。术后随访时间为12.7±3.5个月。IPD的平均值为50.15±7.43mm。IICD由术前的39.38±2.39mm下降至术后的31.64±2.37mm；HPLF由术前20.08±3.53mm增加至术后26.04±1.36mm（p<0.01）；MRD-1从术前-1.12±0.71mm增加到术后3.24±0.47mm（p<0.01）。IICD/IPD由术前0.80±0.12下降至术后0.64±0.07（p<0.05），IICD/HPLF由术前1.98±0.33下降至术后1.22±0.10（p<0.01）。术后VSS评分为3.68±1.07，总体评分较低。患者满意度为8.8±1.08，总体满意度较高。

2.WES结果表明，F-1中家族III(1)和II(1)中存在c.292T>A突变，蛋白突变为p.W98R，F-2中家族II(1)中存在c.383G>T突变，蛋白突变为p.W128L，PolyPhen-2、SIFT和MutationTaster预测这两种突变氨基酸都会损害叉头蛋白。蛋白质结构分析表明，对于MT1（c.292T>A，p.W98R），色氨酸等电位为5.89，色氨酸突变为精氨酸导致等电位变为10.76，氨基酸之间的电位差和氢键增加，导致蛋白质结构的变化和蛋白质功能的改变。MT2（c.383G>T，p.W128L）与WT相比，亮氨酸等电位为6.01。尽管电位差基本保持不变，但观察到叉头结构域局部空间结构发生变化。细胞免疫荧光观察到突变质粒（p.W98R和p.W128L）显示出非WT分布。双荧光素酶实验表明WT-FOXL2抑制StAR启动子活性，而在用等量突变FOXL2质粒（c.292T>A和c.383G>T）转染的细胞中，StAR启动子活性并未受到显著抑制（p<0.05）。RT-PCR实验也表现出同样的结果，c.292T>A和c.383G>T的细胞中StAR mRNA表达显著高于WT-FOXL2细胞。

3.hESC畸胎瘤形成实验结果显示瘤体内包含内胚层，外胚层以及中胚层组织，hESC具有良好的多能性。hESC成功向H9-NCCs分化，Real-time PCR结果显示，与hESCs相比，H9-NCCs高表达NCCs特定标志物基因P75，AP2，HNK-1和 ZIC1，低表达hESCs多能性基因OCT4和NANOG。免疫荧光染色结果显示 H9-NCCs中P75和AP2蛋白均为阳性。成功分化的H9-NCCs-MSCs可以进行成骨和成软骨诱导分化，但尚未能成脂分化。成功构建FOXL2基因敲除的H9-NCCs-MSC后，对照（NC）组中早期凋亡率为30.2%，晚期为37.9%，shRNA组早期凋亡率为28.3%，晚期凋亡为25.6%，NC组早期凋亡率与晚期凋亡率均小于shRNA组，其中晚期凋亡率NC组显著高于shRNA组。

研究结论

1.在本研究中，我们采用创新的手术方法对BPES患者进行了一期手术矫正，并取得了显著的效果。该技术设计合理，明显改善了患者眼部外观，术后瘢痕不明显，患者满意度高。

2.我们报告了两种致病性FOXL2变异（c.292T>A和c.383G>T），并首次证实这两种错义突变都会导致FOXL2蛋白表达和活性的降低。这项研究扩展了已知 FOXL2 变异的范围，有助于理解BPES的病因。

3.我们成功分化了H9-NCCs和H9-NCCs-MSC，并且hESC的细胞多能性在畸胎瘤实验中得到了验证。我们进一步通过慢病毒载体介导的方法成功构建FOXL2基因敲除的H9-NCCs-MSC，并且发现FOXL2基因敲减可以降低H9-NCCs-MSC的早期和晚期凋亡率。我们推测，FOXL2基因在MSC细胞中起着促进凋亡的作用以维持眼睑的正常发育。

Background

The blepharophimosis-ptosis-epicanthus inversus syndrome (BPES) is divided into two types: type I and type II, both of which include ocular developmental anomalies, with type I being associated with female infertility. Currently, many surgical correction methods have been reported, but all have their limitations. However, these procedures have some limitations, and there is considerable controversy regarding whether to perform the surgery in one or two stages. As a rare autosomal dominant genetic disorder, in terms of its pathogenesis, the prevailing view suggests that BPES is caused by mutations in the Forkhead box L2 (FOXL2) gene. The human FOXL2 gene is a single-exon gene located in the chromosome 3q22.3 region and plays an important role in maintaining follicular development and normal ovarian function by regulating downstream genes such as StAR. The specific molecular mechanism of BPES remains unclear, particularly how mutations in the gene affect the transcriptional function of downstream gene promoters. Some scholars have observed a BPES-like eyelid phenotype in animal models after knocking out the FOXL2 gene. They discovered that FOXL2 gene is crucial for eyelid development and levator palpebrae superioris muscle development in mouse neural crest cells (NCCs) and mesodermal cells. However, there have been no reports on the impact of this gene on eyelid development in human neural crest cells and mesodermal cells.

Objectives  

1. Conduct a retrospective analysis of our refined one-stage surgical approach for BPES, evaluating its efficacy and safety.

2. Collect peripheral blood samples from clinical patient families and employ whole exome sequencing(WES) technology to screen for novel mutation sites of the FOXL2 gene. Investigate the effects of these mutations on protein structure, downstream gene expression, and intracellular localization.

3. To simulate the in vitro developmental process, we induced human embryonic stem cells into neural crest cells and further induced these neural crest cells into mesenchymal stem cells (MSCs). Utilize a lentiviral vector to knock out the FOXL2 gene in both neural crest cells and mesenchymal stem cells, examining the gene's influence on cell apoptosis.

Methods 

1. A retrospective analysis was conducted on 25 patients (50 eyes) with BPES who were treated between January 2017 and June 2022. We corrected the medial canthal deformity using an improved reverse Z-plasty combined with incision of the medial canthal ligament. For the lateral canthal area, we enlarged the outer canthus using a "V-Y" advancement flap combined with incision of the tarsal conjunctiva. To address upper eyelid ptosis, we employed a composite tissue flap consisting of the orbicularis oculi muscle and frontalis muscle. Pre- and post-operative periocular measurements were performed, including interpupillary distance (IPD), inner intercanthal distance (IICD), horizontal palpebral fissure length (HPFL), margin reflex distance-1 (MRD-1), IICD/IPD ratio, IICD/HPFL ratio, Vancouver Scar Scale (VSS), and overall postoperative satisfaction. Facial photographs of patients were taken before and after surgery. Follow-up was conducted 6-15 months postoperatively, and scar quantification was performed using the Vancouver Scar Scale, with scores ranging from 0-18, where higher scores indicate more severe scarring.

2. Peripheral blood samples were collected from family members of the patients for whole-exome sequencing. Predictions for mutation sites were conducted using PolyPhen-2, SIFT, and MutationTaster software, and protein structure predictions were performed using AlphaFold2. Transfection plasmids were constructed, with wild-type versions labeled as pEGFP-WT and pcDNA-WT, and mutant plasmids carrying c.292T>A and c.383G>T mutations named pEGFP-MT1 (c.292T>A) and pEGFP-MT2 (c.383G>T), pcDNA-MT1 (c.292T>A), and pcDNA-MT2 (c.383G>T). The human StAR promoter was constructed into pGL3-basic as a luciferase reporter vector. Transfections were performed in HEK-293T cells for dual-luciferase reporter assays, cell immunofluorescence experiments, and real-time quantitative PCR assays.

3. Human embryonic stem cells (H9 hESCs) were injected into the thigh muscle gap of NOD-SCID mice for teratoma formation experiments. Subsequently, H9 hESCs were induced to differentiate into neural crest cells (H9-NCCs), which were further differentiated into mesenchymal stem cells (H9-NCCs-MSCs). Flow cytometry, immunofluorescence, and real-time (RT) quantitative PCR were used to identify these two types of cells. Lentiviral vectors were then transfected into these two types of cells to detect apoptosis.

Results 

1. A total of 25 patients, consisting of 12 females and 13 males, were included in the study. The mean follow-up period was 12.7±3.5 months. The mean IPD was 50.15±7.43 mm. The IICD decreased from 39.38±2.39 mm preoperatively to 31.64±2.37 mm postoperatively; HPLF increased from 20.08±3.53 mm preoperatively to 26.04±1.36 mm postoperatively (p<0.01); MRD-1 increased from -1.12±0.71 mm preoperatively to 3.24±0.47 mm postoperatively (p<0.01). The IICD/IPD decreased from 0.80±0.12 preoperatively to 0.64±0.07 postoperatively (p<0.05), and the IICD/HPLF decreased from 1.98±0.33 preoperatively to 1.22±0.10 postoperatively (p<0.01). The postoperative VSS score was 3.68±1.07, indicating relatively low scarring. The patient satisfaction score was 8.8±1.08, indicating overall high satisfaction.

2. Whole-exome sequencing (WES) results revealed a c.292T>A mutation in family III(1) and II(1) of F-1, resulting in a protein mutation p.W98R. In family II(1) of F-2, a c.383G>T mutation was identified, leading to a protein mutation p.W128L. Predictions by PolyPhen-2, SIFT, and MutationTaster suggested that both mutations would impair the forkhead protein. Protein structure analysis showed that for MT1 (c.292T>A), the potential was changed from 5.89 due to the substitution of tryptophan with arginine, resulting in a potential of 10.76 and changes in protein structure and function. For MT2 (c.383G>T), the potential was 6.01, with observed local spatial structural changes in the forkhead domain compared to WT. Immunofluorescence staining of mutant plasmids (p.W98R and p.W128L) showed non-WT distribution. Dual-luciferase reporter assays demonstrated that WT-FOXL2 inhibited StAR promoter activity, while in cells transfected with equivalent mutant FOXL2 plasmids (c.292T>A and c.383G>T), StAR promoter activity was not significantly inhibited (p<0.05). RT-PCR experiments also showed similar results, with significantly higher StAR mRNA expression in cells with c.292T>A and c.383G>T mutations compared to WT-FOXL2 cells.

3. Results of the hESC teratoma formation experiment showed the presence of endoderm, ectoderm, and mesoderm tissues within the tumor mass. This demonstrates the pluripotency of hESCs. Successful differentiation of hESCs into H9-NCCs was achieved, with real-time PCR showing high expression of NCC-specific marker genes P75, AP2, HNK-1, and ZIC1, and low expression of pluripotent marker genes OCT4 and NANOG compared to hESCs. Immunofluorescence staining showed positive P75 and AP2 protein expression in H9-NCCs. Successful differentiation of H9-NCC-MSCs was achieved, capable of osteogenic and chondrogenic induction but not adipogenic induction. After the successful construction of FOXL2 gene knockout H9-NCCs-MSC, the early apoptosis rate in the control (NC) group was 30.2%, and the late apoptosis rate was 37.9%. The early apoptosis rate in the shRNA group was 28.3%, and the late apoptosis rate was 25.6%. The early and late apoptosis rates in the NC group were lower than those in the shRNA group, and the late apoptosis rate in the NC group was significantly higher than that in the shRNA group.

Conclusion

1. In this study, we performed a one-stage surgical correction in patients with BPES using an innovative surgical approach and achieved remarkable results. The technique was well-designed and significantly improved the patients' ocular appearance with insignificant postoperative scarring and high patient satisfaction.

2. We reported two pathogenic FOXL2 variants (c.292T>A and c.383G>T), and for the first time confirmed that both missense mutations lead to decreased expression and activity of the FOXL2 protein. This study expands the spectrum of known FOXL2 mutations, aiding in the understanding of BPES pathogenesis.

3. We successfully differentiated H9-NCCs and H9-NCCs-MSCs, and the pluripotency of hESCs was validated in the teratoma experiment. We further successfully constructed FOXL2 knockdown H9-NCCs-MSC by a lentiviral vector-mediated approach and found that FOXL2 knockdown reduced the early and late apoptosis rates of H9-NCCs-MSC. We speculate that the FOXL2 gene maintains normal eyelid development by promoting apoptosis in MSCs.