肝脏是执行代谢、消化、免疫、合成等生理功能的主要场所，也是罕见的在机体成熟后仍维持强大再生能力的器官。当肝脏丢失部分质量后，如何启动自我增殖、恢复原有水平，是一个经久不衰的重要科学问题。已有研究对肝再生过程中的生化信号通路及相关时间节点进行了深入的讨论；近年来，力学调控在肝再生中的角色也逐渐被发掘和探索。在经典的70% 肝切除引起的肝再生模型中，血流动力学变化被认为是触发肝再生的主要因素之一。肝部分切除后单位体积的肝脏内急剧增大的血流会对肝血窦内皮细胞产生机械拉伸和流体剪切两种力学刺激。在体模型中，这两种力学刺激是同时存在的，因此无法阐明其中任意一种对肝再生进程的贡献程度和具体机制。体外实验中，大多采用二维的机械拉伸或流体剪切模式对肝血窦内皮细胞施加单一的力学刺激，这种二维的力学加载模式或与生理三维的情况有所差异。已有文献阐述了二维拉伸对肝血窦内皮细胞旁分泌的激活机制，但流体剪切是否影响肝血窦内皮细胞调控肝再生，以及与机械拉伸间的耦合关系，目前仍不清楚。

为了研究流体剪切与机械拉伸在肝脏再生中的作用机制，本文提出了一种创新性的体外仿生肝再生芯片（liver regeneration chip, LRC）系统，可在模拟肝脏生理微结构的同时，实现机械拉伸和流体剪切的单独加载和耦合加载。首先，该肝再生芯片由聚二甲基硅氧烷和盖玻片键合形成微流控装置，并利用针灸针和胶原胶形成胶原胶中空管道。在肝再生芯片中，小鼠原代肝血窦内皮细胞和肝细胞生长状态良好。其中，肝血窦内皮细胞形成三维的血管结构，并通过右旋糖酐渗透实验证明了其屏障功能；肝细胞在肝血窦内皮细胞形成的血管两侧生长为连续单层，免疫荧光染色和过碘酸希夫反应分别证明了肝细胞胞间形成了胆小管结构，同时具备合成糖原的基本功能。进一步，建立了高精度压力泵-肝再生芯片-荧光显微镜联合系统，对芯片中血管进行力学加载的同时可以对血管实时拍摄。该系统可实现拉伸-流动耦合、单独拉伸、单独剪切等三种力学加载模式。第一，通过压力泵对低浓度胶原胶填充的芯片进行流量加载，可以在保证血管内缓冲液流动的同时，引起血管直径扩张，从而实现机械拉伸和流体剪切的耦合加载。第二，改变流量加载模式为压力加载模式，并密闭肝再生芯片的出口，血管仍然随着压力的增大而扩张，但经粒子示踪实验证明，此时血管内粒子运动速度极低，证明剪切应力的作用相对可以忽略不记，此时解耦得到三维加载的单独拉伸模式。第三，提高胶原胶浓度以增大其刚度，仍然使用流量加载模式，发现较低流量加载范围内，血管应变极小可忽略，证明此时解耦得到单独剪切模式。

基于上述所构建的力学模态可控的肝再生芯片系统，本文解耦了机械拉伸和流体剪切对肝系细胞典型功能的影响机制。分析三种力学加载模态下肝血窦内皮细胞的转录组学数据，发现在促进肝再生的旁分泌因子上，三种模态表现出多样性。其中，单独拉伸显著提高肝细胞生长因子（Hgf）的基因水平，单独剪切显著提高肝素结合性表皮生长因子（Hbegf）的基因水平，耦合加载提高了趋化因子12（Cxcl12）的基因水平。同时，三种力学加载引起了不同的力学敏感通路富集，为进一步实验验证提供了线索。

上述转录组数据还显示，单独剪切模式引起的差异基因数目是三种模式中最高的。为进一步开展验证，本文对体外二维培养的肝血窦内皮细胞施加不同梯度的流体剪切，发现多种力学敏感通路的激活。以典型的10 dyne/cm²条件为例，剪切应力激活了肝血窦内皮细胞的Nrf2通路，促进了Bmp6的表达，因而肝血窦内皮细胞上清培养的肝细胞展现出铁调素（Hepcidin）的高表达，也体现出部分药物代谢、蛋白转运功能的改善。

肝切除后肝内也会发生强烈的免疫反应，一个典型的事件是中性粒细胞会被募集在肝脏中。流体剪切下中性粒细胞的招募，不仅受到胞内钙离子水平的调控，而且还通过表面的β2整合素与内皮细胞表达的多种配体分子相结合，介导两者的相互作用。本文通过简化的平行平板流动腔系统和经典的抗体阻断实验，证明了β2整合素的两种亚型，LFA-1和Mac-1，与ICAM-1、RAGE和JAM-A为代表的配体分子结合时，与各自的分子键强度决定了PMN胞内钙响应强度，为在肝再生芯片系统中验证肝切除后肝内的免疫反应奠定了基础。

综上所述，本文以肝再生为生理背景，创建了一种模拟肝脏结构、解耦力学模态的肝再生芯片，并以此诠释了不同力学模态对肝血窦内皮细胞的调控作用。在此基础上，本文也探究了流体剪切调控肝血窦内皮细胞旁分泌功能，以及调控中性粒细胞钙响应的分子机制。本文工作将为从力学可控的微器官/类器官视角认识不同生理力学刺激下肝切除后再生的机制，提供新颖的技术平台和重要的基础数据。

The liver is the main site for implementing those physiological functions such as metabolism, digestion, immunity, and synthesis, and also serves as the sole organ that maintains strong regenerative capacity after organ maturation. How to initiate self proliferation and restore the original level of liver mass after partial hepatectomy is a long-standing and key issue. Previous studies have conducted in-depth investigations on the biochemical signaling pathways and the related time nodes in the process of liver regeneration. Recently, the roles of mechanical regulation in liver regeneration have gradually been unraveled and explored. In the typical liver regeneration model induced by classic 70% hepatectomy, hemodynamic change is considered to be one of the main factors that trigger liver regeneration. Rapid increase in blood flow per unit volume of the liver generates two modes of mechanical stimuli, namely mechanical stretch and shear stress, on the endothelial cells of the hepatic sinusoids. In an in vivo models where two mechanical stimuli co-exist, it is impossible to de-couple the contributions and specific mechanisms for either one of them in liver regeneration. While a single mechanical stimulus of mechanical stretch or shear stress is able to be applied to liver sinusoidal endothelial cells (LSECs) in an in vitro two-dimensional loading model, this modeling may differ from the physiological three-dimensional loading mode. Meanwhile, although it is known in literatures that the two-dimensional stretch can activate the paracrine secretion of LSECs, whether shear stress could also directly contribute to the liver regeneration through LSECs remains unclear.

Based on the above reasoning, this dissertation aims to develop an innovative in vitro biomimetic liver regeneration chip or LRC. It was designed to mimic the in vivo physiological microstructure of the liver and also de-couple the impacts of mechanical stretch and shear stress. First, this LRC was composed of polydimethylsiloxane base material and a cover glass, which were bonded together, with collagen gel filled in between. Inside the LRC, both mouse primary LSECs and hepatocytes were found to proliferate perfectly. Among them, the LSECs formed a three-dimensional vascular structure, and the barrier function was well illustrated through dextran permeation tests. Hepatocytes so grown exhibited the morphology of continuous monolayers on both sides of the blood vessel formed by LSECs. Immunofluorescence staining and periodic Schiff reaction demonstrated, respectively, the intercellular bile canaliculus structure of hepatocytes and the basic function of glycogen synthesis. Next, the combined system of pressure pump-LRC-fluorescence microscope was established, enabling to perform mechanical loading on the blood vessel in LRC with real-time image acquisition. This system was able to produce three mechanical loading modes, that is, stretch-shear coupling, stretch alone, and shear alone. First, the pressure pump was used to perfuse the medium flow into LCR that was filled with low concentration collagen gel, in order to ensure the accessibility of buffer solution flow in the blood vessel when varying the vessel diameter is required for achieving the coupled loading of mechanical stretch and shear stress. Second, by altering the flow loading mode to a pressure loading mode and sealing the outlet of LRC, the blood vessels still expanded with the increase of pressure. Particle track velocimetry tests inside the blood vessels indicated that the velocity of particles was extremely low at this time, supporting that the rols of shear stress can be ignored and an isolated three-dimensional stretch mode was decoupled. Third, by increasing the concentration of collagen gel to enhance its stiffness in the flow loading mode, the vascular strain was quite small and could be neglected within the low flow rate range, implying that an isolated shear stress mode was decoupled at this time.

Based on the above developed LRC system, this dissertation furher decoupled the impacts of mechanical stretch and shear stress on typical functions of liver cells. Analyzing the transcriptomic data of LSECs under three different mechanical loading modes indicated that these three modes exhibited the diversity in the paracrine factors that promote liver regeneration. Typically, stretch alone significantly increased the gene expression of hepatocyte growth factor (Hgf), shear alone remarkably enhanced the gene expression of heparin binding epidermal growth factor (Hbegf), and stretch-shear coupling increased the gene expression of C-X-C motif chemokine ligand 12 (CXCL12). Meanwhile, different mechanosensitive pathways were enriched under the three loading modes, respectively, providing clues for further experimental validation.

The above transcriptome data also showed that the number of differentially expressed genes caused by shear alone was the highest among the three modes. To further validate their potential roles, different gradients of shear stress were applied experimentally to mouse LSECs cultured in vitro, revealing the activation of various mechanosensitive pathways. Taking a typical stress of 10 dyne/cm² as an example, shear stress activated the Nrf2 pathway of LSECs and thus promoted the expression of Bmp6. Mouse hepatocytes cultured with supernatants of these shear-exposed LSECs showed the high expression of hepcidin, also implying the potential improvements in the related drug metabolism and protein transportation.

Liver resection is also known to be accompanied with strong immune responses, as typically represented by neutrophils recruitment into the liver. This recruitment is regulated by intracellular calcium ion levels and β2 integrin binding on neutrophils to various ligands expressed by endothelial cells for mediating their interactions. This dissertation further demonstrated, using a simplified parallel plate flow chamber assay and conventional antibody blocking experiments, the bindings the two subtypes of β2 integrin, LFA-1 and Mac-1, to their respective ligands of ICAM-1, RAGE, and JAM-A. Results indicated that the molecular bond strength formed between these counterparts determined the intracellular calcium response intensity of neutrophils, providing the bases for further eludicating the immune response inside the LRC system after liver resection.

In summary, this dissertation, taking the liver regeneration as the physiological background, developed a novel liver regeneration chip that simulated liver structure and decoupled mechanical loading modes and deciphered the regulatory effects of different modes on tyical functions of liver sinusoidal endothelial cells. On this basis, this dissertation also explored the specific mechanisms by which shear stress regulates the paracrine function of liver sinusoidal endothelial cells and regulates the calcium response of neutrophils. This work provided a new technical platform and key basic data in understanding the roles of in vivo mechanical stimuli in liver regeneration mechanisms after liver resection, from the viewpoint of mechanically-modulatory micro-organs or organoids.