Quality By Design (QbD) Based Antibody Drug Cell Culture Process Development

Objective: There are numerous regulatory parameters of cell culture process in the development and production of antibody drugs,which may have a significant impact on the quality and yield of antibody drugs.In the traditional development and production processes,there is a lack of preset control over the quality of antibody drugs,relying only on For the ultimate batch-based quality control.In this project,the Qb D concept was used to propose a preset standard for antibody quality,and the effects of different cell culture regulatory parameters on the antibody quality were studied in order to establish a controllable,robust,and amplifiable quality that meets the current requirements for registration of biological drugs.Cell culture process.Method: 1.Establish target quality attributes and evaluate critical process parameters Based on the concept of Qb D,on the one hand,by referring to PD-1 antibody-related clinical data,pharmacopoeia and other antibody drugs and predecessor's process experience,on the other hand,through the risk assessment tool to establish target quality attributes and evaluate key process parameters.2.Identify the best process plan for cell culture and create a design space The partial factor experiment was used to screen the process parameters.Based on this,the central composite experimental design was used.A regression model was established with multiple quality indicators as the dependent variables,and the design space and operating space for cell culture were determined.The specific implementation process is as follows: The DOE software was used to design the experimental protocol and high-throughput parallel reactors were used for the cultivation of different process parameters.Cell culture After harvest,the cell supernatant was subjected to Protein A affinity chromatography and Superdex 200 gelfiltration chromatography to obtain a final purified sample,and the purified protein was subjected to quality analysis.The specific analysis methods are as follows: 2.1 Detection of protein concentration: The protein was quantified using ultra-high performance liquid protein A column analysis method.2.2.Detection of antibody aggregates: The ultra-high performance liquid phase SEC column analysis method was used to determine the antibody aggregate content.2.3 Antibody Heterogeneity Detection: The capillary and isoelectric focusing electrophoresis were used to determine the acidic and alkaline peaks of the antibody.2.4 Detection of glycosylation of the antibody: The ultra-high performance liquid phase Glycan column analysis method was used to qualitative and determine the content of the N-glycan structure.The data was analyzed using DOE software and a model was established.3.Model Validation: The reliability and predictability of the regression model were verified by establishing parallel and scale-up experiments on the microreactor and the 5L bioreactor,respectively.4.Detection of in vitro affinity: Two-step purification of antibodies in different cell culture conditions was performed using a protein activator for in vitro affinity detection.Result: 1 Critical Quality Attributes: Through partial factor experiments,it was found that the five potential CPPs had significant effects on different observations.Comprehensive consideration was given to the determination of dissolved oxygen(DO),feed(Feed),and p H as the final CPPs.2.Determination of optimum process conditions: The optimal experimental conditions found in the combinatorial experiment at the center are: p H 6.92,dissolved oxygen 51.07%,feed 24.57%,model predictions are as follows: galactosylation 39.13%,aggregates 15%,alkali The peak was 4.43%,and the protein yield was 1.83 g/L.The design space is: p H 6.92 to 6.95,DO 43% to 53%,and Feed 20.00% to 25.00%.3.Microreactor regression model reliability and predictability: Ambr15 validation results showed galactosylation(36.83±0.73)%,aggregates(14.71±1.10)%,basic peak(5.433±0.29)%,protein concentration(1.64± 0.12)g/L.The reliability and predictability of the 4.5L reactor regression model: 5L reactor magnified 51.60% of galactosylation,14.70% of aggregates,4.973% of basic peak,and protein concentration of 1.65 g/L.5.Affinity in vitro: The affinity of the antibody in vitro was better under the central combinatorial experimental design,and the difference in affinity could not be observed within the instrument detection range.Conclusion: The antibody drug cell culture process based on the concept of Qb D was established.The design space is: p H 6.92 to 6.95,DO 43% to 53%,and Feed 20.00% to 25.00%.After amplification and verification,the amplification of the process can be achieved,and the quality can be improved.control.