第一部分：淋巴管肌瘤病疾病严重度分级研究

背景：

LAM（lymphangioleiomyomatosis），即淋巴管肌瘤病，是一种发病率较低且恶性程度不高的肿瘤性疾病。罕见病注册登记系统的建立有助于了解LAM自然病程和疾病特征。目前被诊断的LAM病例数越来越多，其发病率可能远高于既往研究。但是目前没有LAM 的疾病严重度分级，难以实现疾病的精准、高效管理。

方法：

本研究基于北京协和医院牵头开展的中国淋巴管肌瘤病注册登记研究，构建LAM严重度评分（LAM severity score，LSS）。筛选基线信息较完整、不合并急性气胸和乳糜胸且随访期间有符合本研究要求的肺功能结果的LAM确诊患者。疾病严重度评分变量筛选主要基于本队列对疾病进展和死亡危险因素研究的前期结果。研究主要分析了不同严重度评分的患者组间和组内的肺功能情况，以及患者的死亡情况。构建患者第一秒用力呼气量（forced expiratory volume in one second，FEV1）变化曲线和计算年变化率，比较不同严重度患者组间差异；通过广义估计方程（generalized estimating equation，GEE）探索每个LSS组中的患者随访过程中FEV1进展情况；使用Kaplan-Meier分析不同严重度患者死亡情况。

结果：

LAM严重度评分通过Borg评分、血清VEGF-D（vascular endothelial growth factor D）水平、FEV1占预计值百分比和CT分级四个维度将患者分为I、II、III级，以评估疾病严重度。研究发现，LSS III级患者腹膜后LAM比例更高，而LSS I级患者合并肾脏血管平滑肌脂肪瘤（angiomyolipoma，AML）比例更高。LSS 具有良好疾病区分度，随访过程中三级患者的FEV1相比有显著差异，2年的FEV1年变化率也有显著差异；在未进行西罗莫司治疗的患者中，1年及2年的FEV1的年变化率组间均有显著差异。LSS I组患者的肺功能FEV1稳定，西罗莫司治疗对疾病发展无明显影响；LSS II和III级患者肺功能呈下降趋势，西罗莫司治疗对LSS II和III级患者肺功能下降有显著延缓作用，尤其在LSS III级患者中，使用西罗莫司改善FEV1获益更大并降低死亡风险（P=0.008）。

结论：

LSS有效区分LAM患者疾病严重度和进展。LSS I级患者肺功能稳定，西罗莫司无显著益处；LSS II级患者肺功能呈下降趋势，西罗莫司治疗可以改善病情进展；LSS III级患者肺功能下降快、死亡率高，西罗莫司治疗显著延缓其肺功能下降并降低死亡风险。 LAM严重度评分为LAM患者提供个体化管理策略。

第二部分：LAM细胞和淋巴内皮细胞间的相互作用研究

背景：

罕见病淋巴管肌瘤病（lymphangioleiomyomatosis，LAM），具有低度恶性的特征，属于肿瘤性疾病范畴。这种疾病的病理学特征是TSC2突变非典型平滑肌样肿瘤细胞即LAM细胞，过度增殖。血管内皮生长因子D（vascular endothelial growth factor D，VEGF-D）是重要生物标志物，可以刺激血管内皮生长因子受体3 （vascular endothelial growth factor receptor 3，VEGFR3）。淋巴循环系统是疾病全身转移的重要途经，淋巴管内肿瘤细胞的增殖引起淋巴引流通道的阻塞，这在约20%的LAM患者中出现了淋巴循环异常的表现。在淋巴管和乳糜性积液中可以发现由单层淋巴内皮细胞（lymphatic endothelial cell，LEC）包裹LAM细胞周围的类球形聚集体即LAM细胞簇，基于此科学家也提出了淋巴管肌瘤病的病理假说，LAM细胞簇是疾病全身进展的重要途径。本部分研究拟以病理假说为基础，系统地研究LAM细胞和淋巴内皮细胞之间的相互作用。

方法：

使用GEO数据库GSE135851中的2名LAM患者肺组织的单细胞数据和1名器官捐献者的单细胞数据进行分析。经过降维注释后，分别计算两组数据中淋巴内皮细胞的比例，并提取差异表达基因，进行基因本体（gene oncology，GO）分析和京都基因与基因组百科全书（Kyoto Encyclopedia of Genes and Genomes，KEGG）通路富集分析。细胞实验中，用LAM患者的TSC2缺失的621-101细胞和Tsc2^-/-/p53^-/- 小鼠胚胎成纤维细胞作为LAM细胞系，分别研究了两种LAM细胞系和淋巴内皮细胞之间的相互作用。

结果：

对单细胞数据进行质控、标准化、降维注释后，将单细胞数据分成了28个细胞簇，根据LAM细胞标志物包括PMEL和ACTA2，识别出LAM细胞，其主要位于平滑肌细胞簇中，LAM细胞高表达VEGF-D。根据淋巴内皮细胞的标志物PROX1和FLT4，提取出LAM组和对照组中的LEC，发现 LAM患者组的淋巴内皮细胞浸润约1.609%，而对照组的淋巴内皮细胞浸润约0.196%。经过差异基因表达分析提示，LAM组中的淋巴内皮细胞淋巴管生成相关通路被高度激活。细胞实验中发现，LAM细胞系表现出雷帕霉素靶蛋白（mechanistic target of rapamycin，mTOR）通路激活，VEGF-D 分泌增加。LAM细胞通过分泌VEGF-D，募集淋巴内皮细胞形成LAM球体结构，促进淋巴内皮细胞管生长。LAM细胞分泌的VEGF-D，作用于LEC表面的VEGFR3，使AKT、ERK磷酸化。

结论：

在本部分的研究中，通过单细胞数据发现，LAM患者肺中淋巴内皮细胞增多，淋巴管生成相关通路激活。细胞实验提示，LAM细胞和淋巴内皮细胞之间存在趋化、促进成管的相互作用。LAM细胞可以通过分泌VEGF-D，作用于LEC表面的VEGFR3，引起下游的AKT、ERK磷酸化，这是两者相互作用的关键机制之一。

第三部分：雷帕霉素和VEGFR3抑制剂在淋巴管肌瘤病的协同作用研究

背景：

淋巴管肌瘤病（lymphangioleiomyomatosis，LAM）是一种罕见肿瘤性疾病，是由结节性硬化症基因（TSC1/TSC2）突变驱动引起雷帕霉素靶蛋白通路（mechanistic target of rapamycin，mTOR）激活引起。目前治疗手段有限，唯一的是mTOR抑制剂：西罗莫司（又称雷帕霉素）。西罗莫司可以抑制LAM细胞生长，但不能起杀细胞作用。停止治疗后，肺功能下降恢复，因此，稳定肺功能需要持续药物治疗。西罗莫司使用过程中会发生一些不良事件。药物毒性和耐药性的发展是mTOR抑制剂治疗的潜在问题，而目前淋巴管肌瘤病的治疗除了mTOR通路抑制剂，尚无可供患者选择的其他药物。本部分旨在LAM细胞与淋巴内皮细胞（lymphatic endothelial cell，LEC）的相互作用研究基础上探索新的治疗策略。

方法：

在细胞实验上，使用LAM患者的TSC2缺失的621-101细胞和TSC2过表达的621-103细胞和人淋巴内皮细胞进行共培养，Tsc2^+/+/p53^-/- 小鼠胚胎成纤维细胞 （murine embryo fibroblast，MEF）和 Tsc2^-/-/p53^-/- MEF与小鼠淋巴内皮细胞进行共培养，使用雷帕霉素和血管内皮生长因子受体3（vascular endothelial growth factor receptor 3，VEGFR3）抑制剂治疗，探索对细胞间相互作用的影响。在动物实验中，大鼠子宫平滑肌瘤的TSC2缺失的ELT3细胞构建裸鼠皮下荷瘤模型，并用雷帕霉素和VEGFR3抑制剂进行治疗，观察皮下荷瘤的变化和淋巴结转移情况。

结果：

在LAM细胞和LEC的相互作用中，雷帕霉素对趋化作用有抑制效果，但对管形成无抑制作用，VEGFR3抑制剂可以显著减少间接共培养中向LAM细胞迁移的LEC数量和显著抑制LEC的成管，联合使用的时候可以起到协同作用。在动物实验中，雷帕霉素可以显著大幅度减小皮下荷瘤体积，但不能抑制向前哨淋巴结的转移，VEGFR3抑制剂对肿瘤大小抑制效果不佳，但可以抑制淋巴结转移，两者联合使用时，可以弥补雷帕霉素单独使用的不足。

结论：

本研究首次揭示了雷帕霉素和VEGFR3抑制剂协同治疗的应用前景。两者联合使用在阻断LAM细胞和LEC之间的相互作用上具有协同效果，能更有效延缓疾病系统进展。

Part I: Research on The Lymphangioleiomyomatosis Severity Score

Background

Lymphangioleiomyomatosis (LAM) is a rare and low-grade malignant tumor disease. The establishment of a rare disease registration system is conducive to understanding the natural course and disease characteristics of LAM. Currently, the number of diagnosed LAM cases is increasing, and its incidence may be much higher than that in previous studies. However, there is currently no disease severity classification for LAM, hindering precise and efficient disease management.

Methods

This study is based on the LAM registry study initiated by Peking Union Medical College Hospital, aimed at developing LAM severity score (LSS). We selected LAM patients with complete baseline information, without pneumothorax or chylothorax at baseline, and lung function results that met the requirements of this study during follow-up. The selection of variables for LSS was primarily based on the previous findings of our cohort regarding risk factors for disease progression and mortality. The study assessed pulmonary function (via intra-group and inter-group comparisons) and mortality outcomes, with all analyses stratified by disease severity scores. Specifically, we constructed change curves for forced expiratory volume in one second (FEV1) and calculated annual change rates to compare differences among patients with varying severity. Furthermore, we explored the progression of FEV1 during follow-up in each LSS group employing generalized estimating equation (GEE). Additionally, we utilized Kaplan-Meier analysis to assess the mortality among patients.

Results

The LSS categorizes patients into stages I, II, and III based on four dimensions: Borg scores, serum vascular endothelial growth factor D (VEGF-D) levels, the percentage of FEV1 predicted values, and CT grades, to evaluate disease severity. The study found that patients in LSS III exhibited a higher proportion of retroperitoneal LAM, while those in LSS I had a higher incidence of renal angiomyolipomas (AML). The LSS demonstrated good discriminatory ability for disease severity, with significant differences in FEV1 among the three LSS categories during follow-up. After two years, the annual change rates of FEV1 among groups were significantly different. In patients who did not receive sirolimus treatment, there were significant differences in FEV1 annual change rates among groups at both one and two years. Patients in the LSS I group maintained stable lung function (FEV1), and sirolimus treatment showed no significant impact on disease progression. Conversely, patients in the LSS II and III categories exhibited a declining trend in lung function. Sirolimus treatment significantly delayed the decline in lung function for LSS II and III patients, particularly in LSS III patients, where sirolimus use resulted in greater improvements in FEV1 and a reduction in mortality risk (P=0.008).

Conclusions

The LSS effectively distinguishes the severity and progression of LAM. Patients classified as LSS I maintain stable lung function, and sirolimus shows no significant benefit. LSS II patients exhibit a declining trend in lung function, and sirolimus treatment can help slow disease progression. Patients in LSS III experience a rapid decline in lung function and have a higher mortality rate; sirolimus significantly delays the decline in lung function and reduces the mortality risk. The LAM severity score provides a personalized management strategy for LAM patients.

Part II: Exploring the Interaction Between LAM Cells and Lymphatic Endothelial Cells

Background

Lymphangioleiomyomatosis (LAM), a rare disease, has the characteristics of low-grade malignancy. The pathological hallmark of this disease is the excessive proliferation of atypical smooth muscle-like tumor cells characterized by TSC2 mutations, referred to as LAM cells. Vascular endothelial growth factor D (VEGF-D) is an important serum biomarker for LAM, capable of stimulating vascular endothelial growth factor receptor 3 (VEGFR3). The lymphatic system serves as a critical pathway for systemic metastasis of the disease. The proliferation of tumor cells within the lymphatic vessels leads to obstruction of lymphatic drainage pathways, manifesting in approximately 20% of LAM patients as abnormalities in lymphatic circulation. In lymphatic vessels and chylous effusions, spherical aggregates known as LAM cell clusters can be observed, with LAM cells enveloped by a monolayer of lymphatic endothelial cells (LECs). Based on this observation, scientists have proposed a pathological hypothesis for LAM, suggesting that these LAM cell clusters represent a significant pathway for systemic progression of the disease. This part of the study aims to investigate the interactions between LAM cells and LECs, based on the pathological hypothesis.

Methods

Single-cell data from the lung tissues of two LAM patients and one organ donor, as provided in the GEO database (GSE135851) were used for analysis. Following dimensionality reduction and annotation, we calculated the proportions of LECs in both cohorts and identified differentially expressed genes. Subsequently, we did gene ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. In cell experiments, TSC2-deficient 621-101 cells derived from LAM patients and Tsc2^-/-/p53^-/- mouse embryonic fibroblasts were used as LAM cell models to investigate the interactions between LAM cell lines and LECs.

Results

After performing quality control, normalization, dimensionality reduction, and annotation on the single-cell data, the cells were divided into 28 clusters. Utilizing the LAM cell markers PMEL and ACTA2, we identified LAM cells primarily located within the smooth muscle cell cluster, which exhibited high expression of VEGF-D. By employing the LEC markers PROX1 and FLT4, we extracted LECs from both the LAM and control groups. Our findings revealed that the infiltration of lymphatic endothelial cells in the LAM patient group was approximately 1.609%, compared to about 0.196% in the control group. Differential gene expression analysis indicated that lymphangiogenesis-related pathways were highly activated in the LECs of the LAM group. In our cellular experiments, analysis of LAM cell lines indicated a marked increase in both mechanistic target of rapamycin (mTOR) signaling pathway activation and VEGF-D expression levels. The LAM cells were found to secrete VEGF-D, which recruited LECs to form LAM spheroids and facilitated the growth of lymphatic endothelial cell tubes. VEGF-D secreted by LAM cells acted on VEGFR3 of LECs, triggering downstream phosphorylation of AKT and ERK.

Conclusions

In this part of the study, single-cell data revealed an increase of lymphatic endothelial cells in the lungs of LAM patients, along with activation of lymphangiogenesis-related pathways. Cellular experiments suggested a chemotactic and tube-promoting interaction between LAM cells and LECs. LAM cells secrete VEGF-D, which acts on VEGFR3 present on the surface of LECs, leading to the downstream phosphorylation of AKT and ERK. This represents one of the key mechanisms underlying the interaction between the two cell types.

Part III: Synergistic effects of rapamycin and VEGFR3 inhibitors on lymphangioleiomyomatosis

Background

Lymphangioleiomyomatosis (LAM) is a rare, low-grade malignant tumor driven by mutations in TSC1/TSC2, leading to the activation of the mTOR (mechanistic target of rapamycin) pathway. Currently, treatment options are limited, with the only available therapy being mTOR inhibitors, such as sirolimus (also known as rapamycin). While sirolimus can inhibit the growth of LAM cells, it does not exhibit cytotoxic effects. Upon discontinuation of treatment, lung function declines, indicating that continuous drug treatment is necessary to maintain stability in lung function. During sirolimus therapy, several adverse events can occur. Drug toxicity and the development of resistance are potential challenges associated with mTOR inhibitor therapy, and currently, there are no alternative pharmacologic options available. This part aims to explore new therapeutic strategies based on the interaction research between LAM cells and LECs (lymphatic endothelial cells).

Methods

In cellular experiments, co-cultures were established using TSC2-deficient 621-101 cells and TSC2-overexpressing 621-103 cells from LAM patients, alongside human LECs. Additionally, Tsc2^+/+/p53^-/- murine embryo fibroblasts (MEFs) and Tsc2^-/-/p53^-/- MEFs were co-cultured separately with mouse LECs. The efficacy of treatment with rapamycin and vascular endothelial growth factor receptor 3 (VEGFR3) inhibitors was examined to investigate their effects on intercellular interactions. In animal experiments, a subcutaneous tumor model was established in nude mice using TSC2-deficient ELT3 cells derived from rat uterine leiomyomas. Treatment with rapamycin and VEGFR3 inhibitors was administered, and the influences on subcutaneous tumors as well as lymph node metastasis were observed.

Results

In the interaction between LAM cells and LECs, rapamycin exhibited an inhibitory effect on chemotaxis but did not inhibit tube formation. VEGFR3 inhibitors could significantly reduce the number of LECs migrating toward LAM cells in indirect co-cultures and markedly suppress LEC tube formation. The combination of both treatments demonstrated a synergistic effect. In animal experiments, rapamycin significantly reduced the volume of subcutaneous tumor growth; however, it did not inhibit the metastasis to sentinel lymph nodes. While the VEGFR3 inhibitor showed limited efficacy in suppressing tumor size, it effectively inhibited lymph node metastasis. The animal study suggested that the combined use of rapamycin and VEGFR3 inhibitors could overcome the limitations of rapamycin used alone.

Conclusions

This study has for the first time revealed the application prospects of the combined treatment with rapamycin and VEGFR3 inhibitors. The combined use of the two drugs has a synergistic effect in blocking the interaction between LAM cells and LECs, and can more effectively delay the progression of the disease.