Digital Twin for the Production of hMSC-Derived Extracellular Vesicles for Applications in Cell and Gene Therapy toward Autonomous Operation.

Cell and gene therapy are innovative advanced therapy medicinal products (ATMPs) whose importance for the treatment of various rare and challenging diseases has increased significantly in recent years. Due to their versatility and more flexible application, extracellular vesicles (EVs) show great potential compared to classical cell therapy. Autonomous processes help to reduce the facility footprint and the cost of goods (COG). This requires calibrated predictive process models which, in combination with state-of-the-art process analytical technologies (PAT) and an advanced process control (APC) strategy, guarantee consistent product quality in line with the quality-by-design (QbD) approach required by the regulatory authorities. In this study, a process was therefore developed that consists of two chromatographic purification steps after 3D cultivation of human mesenchymal stem cells (hMSC) on microcarriers and harvesting by depth filtration. The optimized multimodal size exclusion chromatography (SEC) is used for the efficient removal of proteins, and in the subsequent anion exchange chromatography (AEX), the strengths in DNA removal as well as the separation of intact and nonintact EVs are exploited. With a clearance of >98% of DNA and proteins, the regulatory thresholds with regard to DNA per dose and protein concentration are met. The required physicochemical-based mechanistic process models were calibrated for all unit operations. Furthermore, the potential of in situ microscope imaging to determine the viable cell density for adherent growing hMSC during the upstream processing (USP) stage is demonstrated. Fourier transform infrared (FTIR) spectroscopy is used to predict the concentration of the critical metabolite glucose. Inline multi-angle light scattering (MALS) is established in the downstream stage to determine the total particle concentration. The developed predictive process models and PAT tools resulted in the advanced process control (APC) strategy and pave the way for a fully automated process that enables additional productivity increases of 20% at 99.9% reliability with a time saving factor of 2. In summary, this study helps to further improve process economics by reducing batch failures, time-to-market, and COG, which makes these innovations accessible to more patients. Robust, scalable processes are required for commercial production.