CHO cell engineering directed by genomescale modelling

生物药物（由（糖）蛋白和/或核酸组成的药物产品）已经彻底改变了许多使人衰弱和危及生命的疾病的治疗，部分原因是与传统的小分子疗法相比，它们具有高特异性和降低的毒性。因此，生物制药的增长超过了市场增长，刺激了制药行业全球领导者的关注。这就扩大了对强化和改进生产工艺的策略的研究，旨在降低成本，同时提高效率和产品质量。生物制药产品主要通过将重组DNA转染到宿主细胞系中，从而表达、收获和纯化产品来生产。中国人卵巢（CHO）细胞是用于这种表达的最广泛使用的哺乳动物宿主，占市场上所有单克隆抗体的80%。这在很大程度上是由于它们的长期研究、人类相容的糖型、对生物反应器的适应性和易于遗传操作。尽管它们的长期和广泛的使用，然而，仍然有显着的余地，以改善CHO表达系统，通过定向细胞系工程，允许强化和改进的制造过程。基因组规模建模（GeM）是一种计算方法，代表了一个结构化的数据库，所有已知的代谢过程中发生的细胞内。通过整合每个代谢过程中涉及的代谢物、酶和基因，这些模型可以应用于计算细胞内代谢通量、基因表达调控和蛋白质分泌。当与优化算法相结合时，GeMs允许识别潜在的遗传工程策略。例如，通过鉴定非必需代谢途径，可以鉴定候选基因用于敲除/敲低，从而潜在地释放细胞资源并减少宿主细胞蛋白质，提高产品质量和生产率。已经为建立CHO GeM投入了大量努力。值得注意的是，2016年发表了一种通用CHO GeM模型iCHO 1766，其中包含大多数已知的CHO基因、酶和代谢物。Gutierrez等人最近扩展了该模型，将分泌途径（iCHO 2048）包括在内，从而能够计算每种分泌蛋白的能量成本和机械需求。最近，Kol等人利用这种iCHO 2048 GeM产生宿主细胞蛋白质的敲除克隆，报告了特定克隆的更高生产率和改善的生长特性，同时降低了下游加工的压力。然而，在这项工作之外，仍然很少有例子利用这些模型来指导CHO细胞工程策略，使其成为一个有吸引力的研究领域。因此，本博士项目旨在探索这一研究空白。首先，通过扩展iCHO 2048 GeM，例如通过包括翻译后修饰和蛋白质降解途径，以提高生物相关性并允许上级建模能力。至关重要的是，对于有效的建模，因此有效的定向细胞工程，用实验数据约束模型至关重要。因此，在湿实验室中验证选定的靶标之前，应使用代谢进展对该模型进行约束和简化，以将优化算法应用于模型，从而鉴定用于基因工程的靶标，主要旨在优化细胞比生产率。

Biopharmaceuticals (pharmaceutical products consisting of (glyco)proteins and/or nucleic acids) have revolutionised the treatment of many debilitating and life threating diseases, thanks in part to their high specificity and reduced toxicity compared to traditional small-molecule therapeutics. As a result, biopharmaceuticals boast above market growth, stimulating increased focus from global leaders within the pharmaceutical industry. This has amplified research into strategies for the intensification and improvement of manufacturing processes, aiming to reduce cost while increasing efficiency and product quality.Biopharmaceutical products are predominantly manufactured by transfecting recombinant DNA into a host cell line, whereby the product is expressed, harvested, and purified. Chinese Hamster Ovary (CHO) cells are the most widely used mammalian host for such expression, accounting for the production of 80% of all monoclonal antibodies in the market. This is largely due to their long-term study, human compatible glycoforms, adaptability to bioreactors and ease of genetic manipulation. Despite their long-term and widespread use however, there remains significant scope to improve CHO expression systems through directed cell line engineering, allowing for the intensification and improvement of manufacturing processes.Genome-scale modelling (GeM) is a computational methodology representing a structured database of all known metabolic processes that take place within a cell. By integrating the metabolites, enzymes and genes involved within each metabolic process, these models can be applied to compute intracellular metabolic fluxes, gene expression regulation and protein secretion. When coupled with optimization algorithms, GeMs allows for the identification of potential genetic engineering strategies. For instance, by identifying non-essential metabolic pathways, candidate genes may be identified for knockout/knockdown, potentially freeing up cellular resources and reducing host cell proteins, improving product quality and productivity. Substantial effort has already been invested into building a CHO GeM. Significantly, a generic CHO GeM model, iCHO1766, containing most known CHO genes, enzymes and metabolites was published in 2016. This model was recently expanded by Gutierrez et al. to include the secretory pathway(iCHO2048), enabling the computation of energetic costs and machinery demands of each secreted protein.These modelling efforts clearly pave the way for research into directed CHO cell engineering. Recently, Kol et al. utilised this iCHO2048 GeM to generate knockout clones for host cell proteins, reporting higher productivity and improved growth characteristics in specific clones, while reducing pressure in downstream processing. Outside of this work however, there remains few examples utilising these models to directed CHO cell engineering strategies, making it an attractive area for investigation.This PhD project therefore aims to explore this research gap. Firstly, by expanding the iCHO2048 GeM, for instance by including post translational modification and protein degradation pathways, to improve biological relevance and allow superior modelling capabilities. Critically, for effective modelling, and therefore effective directed cell engineering, it is vital to constrain models with experimental data. This model shall therefore be constrained and reduced using metabolic, progress to apply optimisation algorithms to models to identify targets for genetic engineering, primarily aiming to optimise cell specific productivity, before selected targets are validated in the wet lab.