Computational Modelling of Natural killer Cell Function in Vaccine Efficacy

使用先进的动物模型，国家的最先进的成像和遗传分析提供了重要的见解如何免疫反应被激活，以响应模式识别受体的刺激。这导致了用于人类和动物健康的“第二代”佐剂的发展，其中包括Toll样受体（TLR）激动剂。AS01是葛兰素史克公司开发的一种新型佐剂，用于带状疱疹疫苗和疟原虫（疟疾）疫苗。它含有两种关键的免疫增强剂，QS-21（皂苷）和TLR4激动剂MPL，它们协同激活免疫反应。确定AS01免疫应答早期特征的关键之一是细胞因子IFN的早期产生，这依赖于QS-21和MPL之间的协同作用。通过动物模型研究表明，早期干扰素的产生依赖于自然杀伤（NK）细胞，这在as01佐剂疫苗免疫后多功能T细胞的产生中起关键作用。此外，利用转录组学分析，在一项疟疾疫苗接种/挑战研究中证实了IFN-在as01驱动的免疫应答中的作用。虽然NK细胞是naïve小鼠淋巴结中的一小部分细胞，但它们在空间上组织在包膜下区域，在那里它们通过与包膜下巨噬细胞相互作用来应对感染。它们可以通过含有佐剂的MPL在CCR7依赖过程中被募集到淋巴结。尽管传统上认为NK细胞参与了对肿瘤细胞和病毒感染细胞的保护，但最近的研究表明，细胞内细菌和寄生虫感染的控制依赖于NK细胞衍生的干扰素的产生，除了诱导强烈的TH1免疫反应外，还部分通过启动树突状细胞的发育和成熟以及刺激巨噬细胞的功能来起作用。因此，了解NK细胞在免疫应答过程中的动力学和作用机制对于使用合理的设计过程产生更好的疫苗递送策略非常重要。在约克大学的Coles和Timms小组中，作为约克计算免疫学实验室（www.ycil.org）的一部分，我们一直在开发基于agent的免疫系统功能计算模型，通过透明的建模过程为控制免疫反应的机制提供新的见解。计算机模拟可以捕捉生物系统的复杂性（例如对佐剂的局部反应），并可以详细描述单个成分的作用，包括细胞、可溶性因子（例如细胞因子、趋化因子）和局部组织微环境（例如淋巴结或肌肉组织）对高亲和力抗体反应和CTL反应的启动。利用三级淋巴组织（TLT）的计算模拟，我们已经表明，依赖于TNF炎症信号的强度可以导致产生非常不同的病理结果，从弥漫性淋巴浸润到具有滤泡树突状细胞网络的高度组织化的三级结构。利用统计分析技术，我们已经能够解决模型中发生这种情况的原因，并预测tlt将如何应对各种潜在的干预策略。我们最近一直在应用这项技术来了解不同佐剂是如何起作用的，2关注淋巴结重塑和生发中心形成的过程，关注不同免疫细胞之间的相互作用如何推动这一过程。通过将关键系统工程原理应用于模拟，可以利用称为多目标优化的过程来确定模拟中导致最佳免疫反应的关键参数，这些参数可以用作进一步实验的基础。

The use of advanced animal models, state-of-the art imaging and genetic analysis has provided significant insights into how the immune response is activated in response to pattern recognition receptor stimulation. This has led to the development of "second generation" adjuvants for applications in human and animal health that include Toll Like Receptor (TLR) agonists. AS01 is one such new adjuvant developed by GSK that is used in both VZV-Herpes Zoster and plasmodium (malaria) vaccines. It contains two key immune enhancers, QS-21 (saponin) and the TLR4 agonist MPL that synergistically activate the immune response. One of the key defining early features of immune responses to AS01 is the early production of the cytokine IFN which is dependent on the synergy between QS-21 and MPL. Through use of animal models it has been shown that early interferon production is natural killer (NK) cell dependent and this has a key role in the generation of polyfunctional T cells upon immunization with AS01-adjuvant vaccines. Moreover, a role for IFN- in AS01-driven immune responses was confirmed in a malaria vaccination/challenge study using transcriptomics analysis.Although NK cells are a small population of cells in naïve murine lymph nodes they are spatially-organised in the subcapsular region where they are primed to respond to infection through interacting with subcapsular macrophages. They can be recruited to lymph nodes in a CCR7 dependent process by MPL containing adjuvants. Although classically NK cells are thought to be involved in protection against tumour cells and virally infected cells, more recently control of intracellular bacterial and parasitic infection has been shown to depend on NK cell derived interferon production working in part through priming dendritic cell development and maturation and stimulating macrophage function in addition to inducing strong TH1 immune responses. Thus understanding the dynamics and mechanisms of action of NK cells during immune responses is important to generating better vaccine delivery strategies using a rational design process.In the Coles and Timms groups at the University of York, as part of the York Computational Immunology Laboratory (www.ycil.org), we have been developing agent based computational models of immune system function to provide new insights into mechanisms controlling immune responses using a transparent modelling process. In silico simulations can capture the complexity of the biological system (e.g. localised responses to adjuvants) and can detail the role of individual components including cells, soluble factors (e.g. cytokines, chemokines) and the localised tissue microenvironment (e.g. lymph node or muscle tissue) has on the priming of high affinity antibody responses and CTL responses. Using a computational simulation of tertiary lymphoid tissue (TLT) we have shown that depending on the strength of the inflammatory signal by TNF can lead to the generation of very different pathological outcomes from diffuse lymphoid infiltrates to highly organised tertiary structures with follicular dendritic cell networks. Using statistical analysis techniques we have been able to resolve why this occurs in the model and predicted how TLTs will respond to various potential intervention strategies. We have more recently been applying this technology to understand how different adjuvants work,2focusing on the process of lymph node remodelling and germinal centre formation focusing on how the interplay between different immune cells drives this process. Through application of critical systems engineering principles to the simulation it is possible to utilise a process called multi-objective optimisation to determine key parameters that lead to optimal immune responses within the simulation that can be used as a basis for further experimentation.