Optimization of a Mammalian Cell Bioprocess using a Semi-Mechanistic Model

中国仓鼠卵巢（CHO）细胞被广泛应用于制药行业，用于生产治疗一系列疾病的重组蛋白。这些蛋白质的生产效率和治疗效果受到用于培养细胞的化学培养基的化学成分和生物反应器操作条件的高度影响。优化培养基组成是非常具有挑战性的，因为其复杂的性质涉及数十种不同的成分，如氨基酸和非氨基酸物种（盐，维生素，金属等）。我们的工业合作伙伴（Sigma-Aldrich）是细胞培养基的主要制造商，目前正在试验一个先进的48个微反应器系统（Sartorius-AMBR 15,15毫升体积），这是我们实验室无法提供的，用于开发新的细胞系和生长培养基。虽然工业抗体生产涉及补料批或灌注操作，但AMBR微生物反应器系统相对而言更容易、更快、更便宜地用于培养基开发。我的团队在哺乳动物细胞建模方面的工作（Nikdel和Budman， 2016）可以直接应用于Sigma对AMBR系统的媒体开发感兴趣的两个方面：i-扩大在滑铁卢开发的模型的有效性，该模型是在第一个Engage项目下开发的，用于描述使用不同媒体配方的AMBR和灌注操作；i-开发一种新的混合建模方法。其中描述氨基酸动力学的代谢模型与描述非氨基酸组分（盐、维生素、金属等）的经验模型相结合。该模型稍后将用于媒体和流程优化。

Chinese hamster ovary (CHO) cells are used extensively in the pharmaceutical industry to produce recombinantproteins to treat an array of diseases. The productivity and the therapeutic effectiveness of these proteins arehighly influenced by the chemical composition of the chemical medium and the bioreactor operating conditionsthat are used to grow the cells.Optimizing the media composition is very challenging due to its complex nature involving tens of differentcomponents such as aminoacids and non-aminoacid species (salts, vitamins, metals etc.).Our industrial collaborator (Sigma-Aldrich), a main manufacturer of cell culture media, is currentlyexperimenting with an advanced system of 48 micro-reactors (Sartorius-AMBR 15, 15 ml volume each), that itis not available in our lab, to develop new cell lines and growth media. Although industrial antibodymanufacturing involves either fed-batch or perfusion operations, the AMBR micro-bioreactor system iscomparatively much easier, quicker and cheaper to run for the purpose of media development.The work of my group in modeling of mammalian cells (Nikdel and Budman, 2016) can be directly applied totwo aspects that are of interest for Sigma for the purpose of media development with the AMBR system:i- To expand the validity of the model developed at Waterloo under the first Engage project for describing bothAMBR and perfusion operations using different media formulations andii- To develop a novel hybrid modeling approach, where the metabolic model describing the dynamics ofaminoacids is combined with an empirical model to describe non-aminoacid components (salts, vitamins,metals, etc.). This model will be later used for media and process optimization.