Applying a systems pharmacology approach to understanding glucagon-like peptide 1 receptor signalling bias

开发新的有效药物是制药业面临的一个重大挑战。尽管生产不断改进，但成本继续增加，而批准的药物数量却在下降。这在针对G蛋白偶联受体（gpcr）的药物开发中尤为明显，这是一种主要的药物靶点。因此，需要新的方法。系统药理学（SP）是一门结合数学和计算技术的新兴学科，旨在提供更全面的药理学观点。在这里，我们首次提出开发和应用SP方法来定量模拟来自GPCR信号转导的剂量依赖性时间过程数据，特别是胰高血糖素样肽1 （GLP1）受体。只有通过建立最精确的定量GPCR信号模型，我们才能最终使用计算机来预测药物对普通人群的反应。生成GPCR信号的定量模型需要高质量、可重复的时间过程数据，以及以绝对置信度估计无法直接测量的参数的能力。SP方法将允许我们执行这些类型的分析。生物学数据将使用从哺乳动物细胞中获得的GLP1受体的一系列可靠的第二信使测定来获得。在计算上，我们将利用结构可识别性分析来确定未知模型参数的唯一性，确保我们的参数估计尽可能稳健。我们方法的真正优势在于“湿”实验和“干”模型之间的协同作用，确保进行最合适的实验。

The development of new efficacious drugs is a major challenge to the pharmaceutical industry. Despite continued improvements in production, costs continue to increase, while the number of approved drugs declines. This has been particularly evident in the development of drugs aimed at G protein-coupled receptors (GPCRs), a leading pharmaceutical target. Consequently, new approaches are required. Systems pharmacology (SP) is an emerging discipline combining mathematical and computational techniques to provide a more holistic view of pharmacology. Here we propose to, for the first time, develop and apply SP approaches to quantitatively model dose-dependent time-course data derived from GPCR signal transduction, specifically the Glucagon-like peptide 1 (GLP1) receptor. Only through producing the most quantitatively accurate models of GPCR signalling may we eventually be able to use computers to predict how drugs will react when administered to the general population. Generating quantitative models of GPCR signalling requires high quality, reproducible time-course data coupled to the ability to estimate, with absolute confidence, parameters that cannot be measured directly. SP approaches will allow us to perform these types of analyses. Biological data will be obtained using a range of robust second messenger assays for the GLP1 receptor obtained from mammalian cells. Computationally, we will utilise structural identifiability analysis to ascertain the uniqueness of the unknown model parameters, ensuring that our parameter estimation is as robust as possible. The true strength of our approach is the synergy between 'wet' experiments and 'dry' modelling, ensuring that the most appropriate experiments are performed.