唾液外泌体是发掘新型癌症早期诊断生物标志物的潜在来源。目前分离唾液外泌体最常用的超速离心法（Ultracentrifugation, UC）和聚乙烯二醇沉淀法（Polyethylene glycol precipitation, PEG）需要大体积样本、耗时并且无法去除唾液淀粉酶和白蛋白等高丰度蛋白，导致无法从唾液样品中分离出高纯度、高产量的外泌体，不能满足蛋白质组学的研究。为了解决这一问题，本文采用了基于负压振荡技术（Negative pressure oscillation technology, NPOT）的无标签外泌体自动纯化系统（Label-free exosome automatic purification system, LEAPS），建立了一种高效高纯地从唾液中分离外泌体的新工艺和新技术。该系统由一个自动控制平台和一个带有两个平行纳米多孔膜的微流控芯片组成，可实现自动化非标记地分离纯化唾液外泌体。基于该系统，本论文主要做了以下几方面的研究：

（1）通过LEAPS技术，对唾液外泌体分离条件（分离负压、转换时间和洗涤次数）进行优化。研究表明，在负压为30 kPa、转换时间为10 s和洗涤次数为2次的条件下，LEAPS能够在40 min内从1 mL唾液中迅速分离纯化外泌体。

（2）与UC和PEG方法相比，LEAPS技术将分离时间分别缩短了6倍和24倍，外泌体的产量分别提高了2倍和1.5倍，外泌体的纯度分别提高了12倍和60倍，RNA产量分别提高了10倍和5倍。此外，通过LEAPS可以有效去除唾液中高丰度蛋白（例如唾液淀粉酶和白蛋白）的影响。蛋白质组学分析结果显示，LEAPS分离的外泌体携带更多的外泌体标志物，这表明LEAPS在基于唾液外泌体的生物分子研究上具有更大的价值。

（3）采用LEAPS技术对健康人和食管癌患者的唾液样本进行分离研究。结果发现，在食管癌患者唾液蛋白浓度远高于健康人的情况下，通过LEAPS分离的食管癌患者和健康人的唾液外泌体在分离时间，粒子浓度或粒径大小等方面无显著差异。蛋白免疫印迹结果显示，食管癌患者和健康人的唾液外泌体均表达CD81、CD9、CD63等外泌体特异性蛋白。TMT（Tandem Mass Tags）蛋白质组学结果显示，在健康对照组和食管癌组的外泌体中共检测出658种蛋白，其中差异表达蛋白共105种。因此，基于LEAPS的唾液外泌体分离对于发现唾液中的外泌体生物标志物并用于癌症诊断具有重要的潜力。

Salivary exosomes are potential sources for discovering new biomarkers for early diagnosis of cancer. However, ultracentrifugation (UC) and polyethylene glycol precipitation (PEG), the most commonly used methods for separating salivary exosomes, require large volume samples, are time-consuming, and are unable to remove high abundance proteins, such as salivary amylase and albumin, resulting in the inability to isolate high-purity, high-yield exosomes from saliva samples, which is not sufficient for proteomics research. To solve this problem, this paper reports a label-free exosome automatic purification system (LEAPS) based on negative pressure oscillation technology (NPOT) for separating exosomes from saliva with high efficiency and high purity. The system is made up of an automatic control platform and a microfluidic chip with two parallel nano-porous membranes, which can realize automatic label-free separation and purification of saliva exosomes. Based on this system, this thesis mainly did the following researches:

（1）Using the LEAPS technology, the conditions of saliva exosome separation (separation negative pressure, conversion time and washing times) were optimized. The results show that under the negative pressure of 30 kPa, conversion time of 10 s and washing times of 2 times, LEAPS can rapidly separate and purify exosomes from 1 mL of saliva within 40 min.

（2）Compared with the UC and PEG methods, the LEAPS reduces the separation time by 6 times and 24 times, the yield of exosomes increases by 2 times and 1.5 times, the purity of exosomes increases by 12 times and 60 times, and the RNA yield was increased by 10 times and 5 times, respectively. In addition, LEAPS can effectively remove the high-abundance proteins (such as salivary amylase and albumin) in saliva. Proteomic results suggest that LEAPS-isolated exosomes are capable of carrying more exosome-associated biomarkers, which indicates that LEAPS has greater value in the study of biomolecules based on salivary exosomes.

（3）The LEAPS technology was used to separate saliva samples from healthy people and esophageal cancer patients. It was found that although the salivary protein concentration of esophageal cancer patients was much higher than that of healthy people, there was no significant difference in the separation time, particle concentration or particle size of salivary exosomes between esophageal cancer patients and healthy people. Western blot results showed that salivary exosomes of patients with esophageal cancer and healthy people both expressed exosome-specific proteins such as CD81, CD9 and CD63. TMT (Tandem Mass Tags) proteomics results showed that a total of 658 proteins were detected in the exosomes of the healthy control group and esophageal cancer group, of which 105 differentially expressed proteins. Therefore, LEAPS-based salivary exosome isolation proves to be valuable for the discovery of exosomal biomarkers in saliva and for the development of cancer diagnosis.