Downstream design space development for the production of recombinant BMP-2 across unit operations

Abstract

This study shows the application of a novel and innovative risk based approach for downstream pro-cess development across unit operations along Quality by Design (QbD) principles. The use of a novel risk based roadmap for knowledge space development was exemplified for the model protein re-combinant human bone morphogenetic protein-2 (rhBMP-2) produced in Escherichia coli. The ap-plied methodology involves risk assessment tools, interaction matrix across unit operations design, Design of Experiments (DoE) execution and multivariate data analysis (MVDA). A unit operation, as a part of studies multiple-step downstream process, can be considered to have a single function. But the performance of a unit operation is frequently influenced by varying process parameters in other unit operations. A block-wise investigation of these unit operations as a collective arrangement will deliver more process understanding in contrast to a consecutive investigation. The process was based on prior knowledge of an industrial partner. A risk assessment was performed to define poten-tial Critical Process Parameters (CPPs), Critical Quality Attributes (CQAs) and Key Performance Indica-tors (KPIs). Purity was defined as a critical quality attribute (CQA) in respect to further downstream unit operations and final product demands. The specific rhBMP-2 titer was identified as a key per-formance indicator (KPI). In a second step of the risk assessment possible interactions across unit operations were identified.<br />The conventional consecutive approach was performed for unit operations where no interactions with other unit operations were anticipated. Unit operations solubilization and tangential flow filtra-tion were decided to be evaluated consecutively. Potential critical process parameters across unit operations were investigated block-wise on their impact on CQAs and KPIs.<br />The unit operations homogenization, centrifugation and washing - unit operations mainly focusing on isolation in downstream processing - were investigated block-wise.<br />Based on the outcome of the risk assessment an interaction matrix design was developed to collect and rate potential interactions of CPPs across unit operations.<br />In the next step the results of risk assessment (potential CPPs, KPIs, CQAs) and interaction matrix were grouped in Design of experiments. The regression was carried out in the form of process analy-sis DoEs across unit operations or for single unit operations. In data evaluation a multivariate data analysis was done. A core finding was the identification of optimal operating conditions for the inves-tigated downstream processing unit operations of rhBMP-2. Evaluation of the experiments showed following results regarding to purity and titer. An increase of purity in the investigated DSP steps from 15,7% before homogenization to 60% after the precipitation step by dH2O was reached.<br />More-over, till the solubilization step a recovery of 98,9% was achieved.<br />Furthermore, the fundament of a Process Analytical Technology (PAT) strategy for precipitation of the target protein via Tangential Flow Filtration (TFF) was implemented based on in-line conductivity measurement. Conductivity in the retentate correlated with remaining concentration of detergent Guanidine Hydrochloride (GnHCl). GnHCl concentration in retentate had an impact on rhBMP-2 pre-cipitation and following on product titer.<br />The presented work demonstrates that a high degree of mechanical and product specific insight of the process can be extracted. The understanding is gained by means of statistical regression of criti-cal process parameters and of critical quality attributes. This additional information obtained during process development can be used for the interpretation of the knowledge space across unit opera-tions. The creation of a higher level of understanding can be included into the regulatory filing. Proof of process understanding is considered a key element in QbD context for the regulatory submission of pharmaceutical processes by U.S. Food and Drug Administration (FDA) [1]. Process development across unit operations using a DoE approach delivered additional process knowledge, which cannot be gained in a consecutive investigation.<br />To the author's knowledge, with this method it was possible for the first time to take into account interdependencies of process parameters across unit operations: the interaction of Triton X-100 washing with g-number in centrifugation was detected. The variation of interacting CPPs had an in-fluence on the results in another unit operation.<br />The obtained results led to the development of a major knowledge space across unit operations, which allowed manufacturing in a more flexible design/operating space.