BRIC DOCTORATE PROGRAMME on Linking High Throughput Cell Culture Multivariate Analysis and Economics for More Effective Process Integration

随着抗体部门的成熟，上游(USP)产品显著增加，这为细胞培养套件的设计和操作带来了根本性变化的可能性。然而，由于一组可能影响细胞培养性能的内在复杂的相互作用，IT很难预测随着滴度的增加对杂质分布的影响，因此很难预测下游(DSP)操作的稳健性。设计质量倡议推动了更好地了解细胞培养战略对下游加工设备职责的影响的需要，以实现有效的过程集成，从而实现持续改进。该项目将探索最先进的高通量细胞培养和多元数据分析技术，以表征细胞培养操作，不仅在生长和生产力方面，而且在杂质分布和质量方面。由此产生的细胞培养统计因果关系将被集成到UCL决策工具团队开发的过程经济学模型中，以确定未来最具成本效益的USP和DSP集成制造战略。拟议的计划将把MedImmune在抗体生产方面的领先地位与UCL在生物过程决策工具和缩小规模技术方面的领先地位联系起来，以解决这些复杂的过程商业决策。项目阶段：1.微尺度数据生成(年1、年2)。滴度的增加可能是细胞密度或比细胞生产率增加的不同组合的结果，每一种组合对DSP的杂质负荷都有不同的影响。这取决于细胞系特性、生物反应器操作参数、培养基或进料类型以及进料策略等因素。这些对细胞生长、生产力、活力、杂质分布有重大影响，也可能对产品质量产生影响。最初，表征细胞培养策略的大型数据集将通过a)利用MedImmune的历史数据和b)使用‘AMBR微型生物反应器系统’(由Automation Partnership和MedImmune合作开发)和来自UCL EPSRC IMRC当前活动的微孔系统进行高通量实验来生成。2.多因素分析(年2、年3)。下一个挑战将是探索有效的技术来分析如此庞大的数据集，并降低到预测性相关性。将研究先进的多变量分析技术，以帮助预测不同细胞培养策略产生的产品和杂质分布。然后将得出统计的(而不是机械的)因果关系，以将杂质分布与细胞密度、比细胞生产率、培养时间和效价等关键因素联系起来。将重点关注的关键质量属性是HCP(宿主细胞蛋白)、HMW(聚集体)、细胞活力和产品潜力的水平。3.与过程经济学模型的联系(3年)。由此产生的细胞培养预测相关性将与伦敦大学学院与医学免疫公司合作开发的TSB/EPSRC开发的整个生物过程成本模型联系起来，以预测与不同细胞培养策略相关的设备尺寸、COG和风险。5.情景分析(第3年、第4年)。几种细胞培养策略将被插入到综合细胞培养和过程经济模型中，以快速确定最有希望和最强大的USP和DSP活动组合，以便在现有和新设施中进行更精简的开发。这项研究的总体成果将是一个系统的框架，将微型实验与统计相关性和成本建模相结合，以便能够在开发周期的早期选择创新的细胞培养策略，平衡上下游可制造性、对工艺可变性的稳健性和成本效益的需求。

As the antibody sector has matured, it has seen SIGNIFICANT INCREASES IN UPSTREAM (USP) PRODUCTIVITES that have opened up the possibility for radical changes to the design and operation of cell culture suites. However, due to the inherently complex set of interactions that can affect cell culture performance, IT IS HARD TO PREDICT THE CONSEQUENCES ON THE IMPURITY PROFILES AND HENCE ROBUSTNESS OF DOWNSTREAM (DSP) OPERATIONS as titres increase. Quality by Design initiatives are driving the need for greater understanding of the impact of cell culture strategies on the downstream processing equipment duties so as to enable EFFECTIVE PROCESS INTEGRATION and hence CONTINUOUS IMPROVEMENTS. This project will explore STATE-OF-THE-ART HIGH THROUGHPUT CELL CULTURE and MULTIVARIATE DATA ANALYSIS techniques to characterise cell culture operations, not only in terms of growth and productivity BUT ALSO IMPURITY PROFILES AND QUALITY. The resulting cell culture statistical cause-and-effect correlations will be integrated into PROCESS ECONOMICS MODELS developed in the UCL Decisional Tools team so as to TO IDENTIFY THE MOST COST-EFFECTIVE INTEGRATED USP AND DSP MANUFACTURING STRATEGIES FOR THE FUTURE. The proposed programme will link MedImmune's leadership in antibody production and UCL's leadership in bioprocess decisional tools and scale-down techniques to tackle these intricate process-business decisions. Project stages: 1. MICROSCALE DATA GENERATION (Yr 1, 2). Increases in titres can be a result of different combinations of increases in cell densities or specific cell productivities, each with a different impact on impurity burdens on DSP. This depends on factors such as the cell line characteristics, bioreactor operating parameters, medium or feed type and the feeding strategy. These have a significant impact on cell growth, productivity, viability, impurity profiles and may also have an impact on product quality. Initially large datasets characterising cell culture strategies will be generated by a) leveraging historical data from MedImmune and b) high throughput experimentation using both the 'ambr microscale bioreactor system' (developed in a collaboration between The Automation Partnership and MedImmune) and microwell systems from current UCL EPSRC IMRC activities. 2. MULTIVARIATE ANALYSIS (Yr 2, 3). The next challenge will be exploring effective techniques to analyse such large datasets and reduce to predictive correlations. Advanced multivariate analysis techniques will be investigated to help predict the product and impurity profiles resulting from different cell culture strategies. Statistical (rather than mechanistic) cause-and-effect correlations will then be derived to link the impurity profile to key factors such as the cell density, specific cell productivity, culture duration, and titre. Key quality attributes that will be focused on are the levels of HCP (host cell proteins), HMW (aggregates), cell viability and product potency. 3. LINKAGE TO PROCESS ECONOMICS MODELS (Yr 3). The resulting cell culture predictive correlations will be linked to a whole bioprocess cost model developed at UCL in a TSB/EPSRC collaboration with MedImmune so as to predict the equipment sizes, COG and risks associated with different cell culture strategies. 5. SCENARIO ANALYSIS (Yr 3, 4). Several cell culture strategies will be plugged into the integrated cell culture and process economics model to enable rapid identification of the most promising and robust combinations of USP and DSP activities for more streamlined development in both existing and new facilities. The overall outputs of this research will be a systematic framework combining microscale experimentation with statistical correlations and cost modelling so as to enable selection of innovative cell culture strategies early in the development cycle that BALANCE THE NEEDS OF UPSTREAM AND DOWNSTREAM MANUFACTURABILITY, ROBUSTNESS TO PROCESS VARIABILITIES AND COST-EFFECTIVENESS.