Cytokine network ecology: towards a dynamic understanding of immune responses to co-infection

细胞因子是组织免疫反应的重要分子：它们激活细胞分裂和攻击感染因子，有时它们直接作用于病原体。因此，细胞因子强烈影响寄生虫被杀死的方式（和速度）。每一种免疫反应都涉及多种细胞因子信号，这是一个复杂的正反馈和负反馈循环网络，决定了宿主抵抗感染的能力。这些反馈回路，以及这种信号系统固有的时间滞后，使得研究免疫系统变得困难。在任何一个时间点收集的数据或一次只分析一个变量根本无法揭示不同细胞因子必须相互作用才能产生观察到的全系统免疫反应的方式。在这个项目中，我建议将宿主视为一个封闭的生态系统，其中生态和进化分析可以确定细胞因子如何共同作用以产生有效的与病理的免疫反应。为什么要通过研究细胞因子来了解整个免疫系统？细胞因子的吸引力有三个方面：细胞因子与免疫相关——例如，几乎所有免疫学家都测量细胞因子来帮助他们推断细胞的功能，无论研究的是哪种宿主细胞类型，也无论研究的是哪种感染或自身免疫性疾病。-细胞因子在分析上易于处理-相同的<10细胞因子与所有描述的感染性和自身免疫性疾病有关。涉及哪些细胞和膜结合分子（以及较少丰富或较少描述的细胞因子）的细节确实在这些系统中发生变化，但相同的细胞因子总是重要的这一事实令人震惊。例如，对这些<10的细胞群进行建模比对许多不同环境的细胞群进行建模要好。——而且（对于进化研究而言），由于细胞因子选择了杀死寄生虫的机制，它们强烈影响宿主的健康和生存。该项目旨在利用基于最优性和概率的数学方法，将共感染小鼠多种细胞因子的真实数据整合到网络框架中。最优性方法是合适的，因为我们预计免疫系统存在自然选择，特别是优化其多任务能力/例如，优化管理原生动物-蠕虫共感染。同时，概率方法也是合适的，因为细胞因子形成了一个概率网络，有很多变异性和偶然事件。我建议使用的概率统计方法已经成功地应用于环境科学（例如，在多种相互作用的环境因素下预测对珊瑚礁的破坏）。测试最优预测极大地加深了我们对进化生物学的理解，从鸟类鸣叫到抗生素耐药性的发展。将这些预测分析方法结合应用于免疫分子也将取得成果：对与宿主健康相关的免疫系统功能的综合理解。

Cytokines are important molecules for the organisation of immune responses: they activate cells to divide and to attack infectious agents, and they sometimes act directly against pathogens. Cytokines thus strongly influence how (& how quickly) parasites are killed. Multiple cytokine signals are involved in every immune response -- an intricate network of positive and negative feedback loops that determine how well the host manages to fight infection. These feedback loops, and the time lags inherent in such a signalling system, can make the immune system difficult to study. Data collected at any one time point or analysed only one variable at a time simply cannot reveal the way that different cytokines must interact to generate an observed system-wide immune response. With this project, I propose to treat the host as a closed ecological system in which ecological and evolutionary analyses can identify how cytokines work together to generate effective versus pathological immune responses. Why study cytokines to learn about the whole immune system? The appeal of cytokines is three-fold: - cytokines are immunologically relevant -- e.g., nearly all immunologists measure cytokines to help them infer the function of cells, no matter which host cell type nor which infection or autoimmune condition is under study. - cytokines are analytically tractable -- the same <10 cytokines are implicated in all described infectious and autoimmune diseases. Details of which cells and membrane bound molecules (& less abundant or less well described cytokines) are involved does change across these systems, but the fact that the same cytokines are always important is striking. Better to model these <10 than to model the many cell populations, for example, that vary across context. - and (for evolutionary studies), because cytokines choose parasite-killing mechanisms, they strongly influence host health and survival. The proposed project aims to integrate real data on multiple cytokines of co-infected mice into a network framework, making use of optimality-based and probabilistic mathematical methods. Optimality methods are appropriate because we expect that there has been natural selection on immune systems, particularly to optimise their ability to multi-task / for example, to optimally manage protozoan-helminth co-infection. At the same time, probabilistic methods are also appropriate because cytokines form a probabilistic network, with a lot of variability and chance events. The probabilistic statistical methods that I propose to use have been successfully applied to environmental science (for example, to predict damage to coral reefs given multiple interacting environmental factors). Testing optimality predictions has greatly deepened our understanding of the evolutionary biology of everything from bird song to the development of antibiotic resistance. The combined application of these predictive analytical methods to immunological molecules will also bear fruit: an integrated understanding of immune system functioning that links to the health of hosts.