第一部分 蛋白质组学揭示磷脂代谢在小细胞肺癌抗肿瘤免疫和转移复发中的重要性

【背景】 小细胞肺癌（small-cell lung cancer，SCLC）是一种高度恶性的神经内分泌肿瘤，近70%的患者在确诊时已经发生转移，整体预后较差。脑转移在SCLC很常见，整个疾病过程中，脑转移患者的占比能达到近80%，因此制定了预防性脑放疗方案，虽然可以降低脑转移的发生率，但并不能改善患者的无病生存期和总生存期。关于允许SCLC在特定器官环境中归巢、种植和形成转移灶的机制仍知之甚少。因此，探索转移相关的分子生物学机制并确定抑制肿瘤细胞转移扩散或能够杀伤转移灶肿瘤的新靶点对于开发SCLC新型治疗策略至关重要。

【方法】 我们对90例既往未接受过治疗的SCLC患者原发灶肿瘤样本、15例与原发灶肿瘤样本配对的淋巴结转移灶肿瘤样本以及15例广泛期SCLC患者的脑转移灶肿瘤样本进行了基于液相色谱-串联质谱技术的蛋白质组学分析。通过通路富集分析、二维富集分析和加权基因共表达网络分析（weighted gene co-expression network analysis，WGCNA）探索与SCLC转移相关的生物学功能和分子信号通路，再使用机器学习算法确定与SCLC转移相关的关键蛋白质，最后通过免疫组化和免疫荧光进一步验证关键蛋白质的生物学意义和临床价值。

【结果】 无监督聚类分析结果提示部分原发灶肿瘤样本具有与转移灶肿瘤样本相似的蛋白质表达特征，提示这些原发灶肿瘤样本的潜在转移能力。通路富集分析和二维富集分析结果表明潜在转移的原发灶肿瘤样本与转移灶肿瘤样本在生物学功能和分子信号通路改变方面高度一致，表现为细胞周期、DNA复制和细胞分裂等信号通路的上调，细胞粘附、脂质分解代谢、磷脂酶A2活性和免疫刺激等信号通路的下调。WGCNA结果同样提示免疫反应、脂质代谢和成纤维细胞迁移等是不利于SCLC转移的生物学过程。临床层面进一步证实，血脂水平是SCLC患者的独立预后预测因素，高密度脂蛋白胆固醇水平升高与SCLC患者更好的预后相关，低密度脂蛋白胆固醇水平升高与SCLC患者更差的预后相关。最后通过支持向量机算法确定的SCLC转移相关十种关键蛋白质，显示出与肿瘤免疫细胞浸润和患者预后之间的相关性。最后，免疫组化结果证实，肿瘤间质IIA型磷脂酶A2（phospholipase A2 group IIA，PLA2G2A）高表达与局限期SCLC患者的延迟复发相关。

【结论】 本研究通过蛋白质组学分析发现了磷脂代谢及免疫反应对于SCLC转移和患者疾病复发的关键作用，并确定了参与磷脂代谢通路和先天免疫反应的PLA2G2A对于SCLC患者的预后预测意义。

第二部分 基于脂质代谢相关蛋白质的小细胞肺癌预后模型的构建

【背景】 小细胞肺癌（small-cell lung cancer，SCLC）是一种高度侵袭性的恶性肿瘤，患者远期预后通常较差。近年来基于分子分型和肿瘤异质性的研究并未显示出对于SCLC患者预后生存的预测价值，因此迫切需要开发新型的生物标志物来识别高危SCLC患者，并优化个体治疗策略，从而改善预后。细胞代谢转变已经逐渐成为恶性肿瘤的标志性特征，脂质代谢是能量产生、生物膜形成和生物信号转导所必需的，高脂饮食与肿瘤发病风险的升高存在相关性，脂质的摄取、储存和代谢改变深度参与了肿瘤细胞的恶性转变。我们在前期的研究中发现了脂质代谢与SCLC疾病复发和转移的相关性，本研究将尝试通过脂质代谢相关蛋白质构建SCLC的预后预测风险模型，并对模型背后的分子生物学机制进行探索。

【方法】 首先通过公共数据库确定脂质代谢相关蛋白质，使用无监督一致性聚类算法探索SCLC不同的脂质代谢模式及相关分子分型，然后使用LASSO回归模型确定用于构建风险评分模型的最稳健蛋白质数量，并通过多因素Cox回归模型构建预后风险评分模型，对风险评分模型的预后价值进行验证后，进一步探索预后模型的潜在分子生物学机制，及其与药物敏感性之间的潜在相关性。

【结果】 我们首先确定了940个与脂质代谢相关的蛋白质，其中51个与SCLC患者的预后相关，对于51个预后相关脂质代谢蛋白质的一致性聚类结果显示，SCLC存在两种不同的脂质代谢模式，并与患者预后和免疫细胞浸润相关。通过LASSO回归确定了8个与生存相关的脂质代谢蛋白质用来构建SCLC的预后风险评分模型，该风险评分是SCLC患者的独立预后预测因素，并在独立的SCLC患者队列中得到了验证。通过对低风险组和高风险组的差异表达蛋白质进行通路富集分析揭示了该风险评分与SCLC神经元特征和抗肿瘤免疫反应之间的相关性。最后根据蛋白质表达谱预测的药物敏感性也显示出与风险评分之间的关联，低风险组患者对于几乎全部化疗药物和靶向治疗药物的敏感性均显著高于高风险组。

【结论】 SCLC存在两种不同的脂质代谢模式，对于患者具有预后预测价值。根据脂质代谢蛋白质构建的预后风险评分模型是SCLC患者的独立预后预测因素，该风险评分显示出与SCLC抗肿瘤免疫反应之间的相关性，并且可能对药物治疗的敏感性具有预测价值。

第三部分 循环肿瘤细胞/循环肿瘤内皮细胞在小细胞肺癌一线化疗疗效中的预测价值

【背景】 小细胞肺癌（small-cell lung cancer，SCLC）具有快速生长和早期转移的特点，患者在确诊时多已处于疾病晚期，化疗是SCLC患者的主要治疗手段。尽管SCLC对于化疗较为敏感，但大部分患者会在短时间内不可避免地出现耐药和疾病进展，整体预后较差。以循环肿瘤细胞（circulating tumor cell，CTC）为代表的液体活检技术能够通过外周血获得患者体内的肿瘤信息，目前已经在多种类型、不同疾病阶段的肿瘤患者中显示出对于疗效和预后的预测价值。本研究旨在确定CTC、循环肿瘤内皮细胞（circulating tumor endothelial cell，CTEC）及其亚型对于SCLC患者一线治疗的临床意义。

【方法】 本研究入组既往未接受过治疗的初诊SCLC患者，入组患者均接受一线标准化疗（联合或不联合放疗）治疗，采集患者在第一周期化疗前、第二周期化疗结束时、一线化疗结束时和疾病进展时的外周血。通过差相富集方法对外周血中的CTC和CTEC进行富集，然后使用免疫染色荧光原位杂交方法对CTC和CTEC进行分子检测，以确定不同的细胞类型。在获得CTC和CTEC及其各自亚型的数据后，分析不同类型细胞或细胞团与患者临床病理特征、疗效、无进展生存期（progression-free survival，PFS）和总生存期（overall survival，OS）之间的相关性。

【结果】 2018年11月至2020年1月，共入组33例患者，中位随访时间为20.0个月（范围，2.8-30.2个月），中位PFS和中位OS分别为11.5个月（范围，1.7-25.9个月）和19.7个月（范围，2.8-30.2个月）。CTC和CTEC的检出率分别为96.6%和65.5%，CTC检出水平显著高于CTEC检出水平（p<0.05），广泛期患者具有显著更高的CTEC检出水平（p=0.035），循环肿瘤细胞-白细胞（circulating tumor cell white blood cell，CTC-WBC）细胞团的检出与更差的一线治疗疗效有关（p=0.030）。第一周期化疗前（17.3个月 vs. 22.6个月，p=0.041）和第二周期化疗结束时（19.9个月 vs. 25.2个月，p=0.018）检出CTC-WBC细胞团的患者OS较未检出的患者明显更差。CTC-WBC细胞团的检出是患者较差OS的独立预测因素（第一周期化疗前：HR 9.3，95% CI 1.4-48，p=0.0079；第二周期化疗结束时：HR 4.4，95% CI 1.1-18，p=0.041）。此外，一线化疗结束时CTC水平较高、CTC细胞团的检出以及CTEC的检出均是SCLC患者的不良预后因素。

【结论】 第一周期化疗前和第二周期化疗结束时CTC-WBC细胞团的检出是SCLC患者一线治疗的独立预后预测因素，此外，化疗结束时CTC水平较高、CTC细胞团的检出以及CTEC的检出也都与患者的不良预后相关。

Section I Proteomic profiling reveals the significance of lipid metabolism in anti-tumor immunity, recurrence and metastasis of small-cell lung cancer

Background: Small-cell lung cancer (SCLC) is an aggressive neuroendocrine tumor. Nearly 70% of SCLC are diagnosed with metastasis, exhibiting poor prognosis. Brain metastasis is common in SCLC, and the proportion of patients with brain metastasis can reach to 80% throughout the entire disease process. Prophylactic cranial irradiation could reduce the incidence of brain metastasis, while the disease-free survival and overall survival were not improved. Little is known about the molecular and cellular mechanisms that allows SCLC to home to, seed, and grow into metastases in specific organ environments. Therefore, exploring the molecular mechanisms related to SCLC metastasis and identifying new targets that inhibit tumor cell metastasis or eliminate metastatic tumors are crucial for developing novel therapeutic strategies for SCLC.

Methods: We conducted comprehensive proteomic analysis based on liquid chromatography-mass spectrometry on primary tumor samples from 90 previously untreated patients with limited-stage or extensive-stage SCLC, 15 lymph node metastatic tumor samples paired with primary tumor samples, and 15 brain metastatic tumor samples from patients with extensive-stage SCLC. Pathway enrichment analysis, two-dimension enrichment analysis, and weighted gene co-expression network analysis (WGCNA) were conducted to explore the biological functions and molecular signaling pathways associated with SCLC metastasis. Then, hub proteins associated with SCLC metastasis were identified using machine learning algorithms. Finally, the biological significance and clinical value of hub proteins were further validated through immunohistochemistry and immunofluorescence.

Results: The results of unsupervised clustering suggested that partial primary tumor samples had protein expression characteristics similar to those of metastatic tumor samples, indicating the metastatic potential of these primary tumor samples. The results of pathway enrichment analysis and two-dimension enrichment analysis also indicated a high degree of consistency in the biological function and signaling pathway changes between potential metastatic primary tumor samples and metastatic tumor samples, manifested by up-regulation of signaling pathways related to cell cycle, DNA replication, and cell division, as well as down-regulation of signaling pathways related to cell adhesion, lipid metabolism, phospholipase A2 activity, and immune stimulation. The WGCNA results also indicated that immune response, lipid metabolism, and fibroblast migration were biological processes unfavorable for SCLC metastasis. It has been further confirmed that blood lipid levels were independent prognostic factors for SCLC patients. Elevated levels of high-density lipoprotein cholesterol were associated with better prognosis, while low-density lipoprotein cholesterol was associated with worse prognosis. Finally, 10 hub proteins related to SCLC metastasis were identified using support vector machine algorithm, demonstrating correlations with immune cell infiltration and patient prognosis. Finally, immunohistochemical results further confirmed the association between high expression of phospholipase A2 group IIA (PLA2G2A) in tumor stroma and delayed recurrence in patients with limited-stage SCLC.

Conclusions: This study demonstrated the significance of phospholipid metabolism and immune response in SCLC metastasis and disease recurrence through proteomic analysis, and determined the prognostic significance of PLA2G2A for SCLC patients, which is involved in both phospholipid metabolism and innate immune response.

Section II Construction of prognostic model for small-cell lung cancer based on lipid metabolism related proteins

Background: Small-cell lung cancer (SCLC) is a highly invasive malignant tumor, and patients always have poor prognosis. Recent studies of molecular subtypes and tumor heterogeneity have not shown prognostic value for SCLC patients. There is an urgent need to develop novel biomarkers to identify high-risk SCLC patients and optimize individual treatment strategies, ultimately improving prognosis. Metabolism transformation has become a hallmark of cancer. Lipid metabolism is necessary for energy production, biofilm formation, and biological signal transduction. High-fat diet is associated with an increased risk of tumor development. Changes in lipid uptake, storage, and metabolism are deeply involved in the malignant transformation of tumor cells. In previous research, we also found a correlation between lipid metabolism and the recurrence and metastasis of SCLC. This study attempted to construct a prognostic risk model for SCLC based on lipid metabolism related proteins and explored the biological mechanisms behind the model.

Methods: Lipid metabolism related proteins were identified from public databases. Unsupervised consensus clustering was used to explore different lipid metabolism patterns and related molecular subtypes in SCLC. Then, LASSO regression model was used to determine the most robust proteins for constructing risk score, which was constructed using multi-variable Cox regression model. After verifying the prognostic value of the risk scoring model, the potential molecular mechanisms of the prognostic model and its potential correlation with drug sensitivity were analyzed.

Results: We identified 940 proteins related to lipid metabolism, among which 51 proteins were correlated with prognosis of SCLC patients. The consensus clustering results of 51 prognostic related lipid metabolism proteins showed that SCLC had two different lipid metabolism patterns, which was associated with patient prognosis and immune cell infiltration. Eight survival related lipid metabolism proteins were identified through LASSO regression to construct a prognostic risk scoring model for SCLC. This risk score was an independent prognostic factor for SCLC patients and was validated in an independent SCLC patient cohort. It was demonstrated that this risk score was associated with neuronal characteristics and anti-tumor immune response through pathway enrichment analysis of differentially expressed proteins between low-risk and high-risk groups. Finally, drug sensitivity prediction based on protein expression profiles also showed a correlation with this risk score, with low-risk patients showing significantly higher sensitivity to almost all chemotherapy and targeted therapy drugs compared to high-risk groups.

Conclusions: SCLC had two different lipid metabolism patterns, which showed prognostic value for patients. The prognostic risk scoring model constructed based on lipid metabolism proteins was an independent prognostic factor for patients with SCLC. This risk score also showed correlation with anti-tumor immune response of SCLC and might be a predictor for drug sensitivity.

Section III Prognostic value of circulating tumor cell, circulating tumor endothelial cell and their subtypes in first-line treatment of small-cell lung cancer

Background: Small-cell lung cancer (SCLC) has the characteristics of rapid growth and early metastasis. Patients are often diagnosed with metastasis, and chemotherapy is the main treatment strategy. Although SCLC is sensitive to chemotherapy, most patients will inevitably develop resistance and disease progression, resulting in poor prognosis. Liquid biopsy, represented by circulating tumor cell (CTC), can obtain tumor information through peripheral blood. CTC had shown high value for efficacy and prognosis evaluation in patients with various types of tumors in different disease stages. The aim of this study was to determine the clinical significance of CTC, circulating tumor endothelial cell (CTEC), and their subtypes in first-line treatment of SCLC patients.

Methods: This study enrolled patients with newly diagnosed SCLC who had not received previous treatment. All enrolled patients received first-line standard chemotherapy (with or without radiotherapy). Peripheral blood samples were collected before the first cycle of chemotherapy, at the end of the second cycle of chemotherapy, at the end of first-line chemotherapy, and during disease progression. CTC and CTEC in peripheral blood were enriched using subtraction enrichment method. Then, molecular detection of CTC and CTEC was performed using immunostaining fluorescence in situ hybridization to determine different cell types. After obtaining data of CTC, CTEC and their subtypes, the correlation of different types of cells or cell clusters with the clinical characteristics, efficacy, progression-free survival (PFS), and overall survival (OS) of SCLC patients were analyzed.

Results: From November 2018 to January 2020, a total of 33 patients were enrolled in this study. The median follow-up time for all patients was 20.0 months (range, 2.8-30.2 months), with median PFS and median OS of 11.5 months (range, 1.7-25.9 months) and 19.7 months (range, 2.8-30.2 months), respectively. The detection rates of CTC and CTEC were 96.6% and 65.5%, respectively. The level of CTC was significantly higher than that of CTEC (p<0.05), and patients with extensive-stage SCLC had significantly higher levels of CTEC detection (p=0.035). The detection of circulating tumor cell white blood cell (CTC-WBC) clusters was associated with poorer efficacy (p=0.030). The OS of patients with CTC-WBC clusters detected before the first cycle of chemotherapy (17.3 vs. 22.6 months, p=0.041) and at the end of the second cycle of chemotherapy (19.9 vs. 25.2 months, p=0.018) were significantly shorter than that of patients without detection. The detection of CTC-WBC clusters was an independent factor for poor OS (before the first cycle of chemotherapy: HR 9.3, 95% CI 1.4-48, p=0.0079; at the end of the second cycle of chemotherapy: HR 4.4, 95% CI 1.1-18, p=0.041). In addition, higher CTC levels, the detection of CTC clusters and the detection of CTEC at the end of first-line chemotherapy was also predictors for poor prognosis of patients with SCLC.

Conclusions: The detection of CTC-WBC clusters before the first cycle of chemotherapy and at the end of the second cycle of chemotherapy were independent prognostic factor for SCLC patients with first-line chemotherapy. In addition, higher CTC levels, the detection of CTC clusters and the detection of CTEC at the end of the first-line chemotherapy were also associated with poor prognosis.