The lytic nature of virus-based expression systems, used for production of many vaccines and viral vectors, limits their transition from traditional batch-based to intensified continuous bioprocessing. In this work, we implemented a multi-stage bioreactor setting to leverage continuous operation using virus-based expression systems. The multi-stage bioreactor set-up was used for continuous production of influenza virus-like particles (iVLPs) as vaccine candidate, using the insect cell-baculovirus expression vector system (IC-BEVS). Additionally, this system is being validated to produce another two biopharmaceutical products: (i) adenovirus-like particles (ADDomer) as a nanoparticle-based snake-bite therapy alternative, using IC-BEVS, and (ii) recombinant adeno-associated viruses (rAAV) as gene therapy vector, using the HeLaS3 packaging cell line (PCL) infected with wild-type adenovirus type 5 (wtAd5). Critical process parameters (CPP) being fine-tuned to maximize productivity include cell line, virus construct, cell concentration at infection (CCI), residence time (RT) in the production bioreactor, and perfusion rates (PR). For iVLPs, optimization of cell line, baculovirus construct, and RT, allowed for consistent iVLPs titer (34 ±14 HA titer/mL) throughout the period of continuous operation (20 days). For ADDomer, cell line and PR have been optimized to achieve 8-fold higher ADDomer expression via high cell density (HCD) culture (10-20 x106 cell/mL) using perfusion. This HCD approach will be employed into the continuous production using the multi-stage bioreactor set-up. A similar approach is being undertaken for continuous rAAV production, but using lower CCI (0.5-1 x106 cell/mL); RT will be further fine-tuned for optimal rAAV expression using the HeLaS3-PCL expression system. This work showcases the potential of multi-stage bioreactors for continuous production of different types of biopharmaceutical using virus-based expression systems across multiple cell hosts. This approach allows seamless integration with continuous downstream processing, paving the way for fully integrated, end-to-end biomanufacturing.

Chinese hamster ovary (CHO) cells remain the most widely used host cell line for biotherapeutics production. Despite their widespread use, understanding endoplasmic reticulum (ER) stress conditions in recombinant protein production remains limited, often creating bottlenecks preventing improved production titers and product quality. Excessive protein production, nutrient deprivation, and other cell culture media conditions (e.g. waste build-up) in CHO cell cultures can lead to protein unfolding and misfolding, which in turn triggers ER stress. When ER stress occurs, cells activate the unfolded protein response (UPR) to restore protein homeostasis and maintain ER folding capacity. However, if ER stress persists and remains unresolved, the UPR can trigger apoptosis, leading to cell death. Monitoring UPR levels is therefore essential for maintaining high productivity and ensuring product quality. In our study we have carried out a comprehensive whole cell proteomic and ubiquitinated proteomic investigation of CHO cells in a range of bioprocess conditions such as temperature shift, culture longevity, waste build-up, and high and low productivity, to identify potential protein biomarkers of ER-stress. Using open-source R packages, the proteomic data compiled from the various stress conditions was analysed using PCA and regression analysis; and randomForest and Support Vector Machine machine learning models. We identified a panel of ER-stress related biomarkers, including well known markers of ER-stress such as BiP and ATF6, and developed a targeted mass spectrometry based assay, i.e. parallel reaction monitoring (PRM), to monitor stress related biomarkers in CHO cell cultures. Using PRM, we were able to accurately quantify a well-defined set of peptides associated with a panel of protein biomarkers linked with ER stress. Understanding UPR regulation under varying culture conditions could establish this assay as a screening tool for assessing media composition or process changes, and also during cell line development for the selection of high-performing clones.

Suspension MDCK cells are a highly relevant cell substrate for scalable and efficient production of influenza A virus (IAV). Considering the high heterogeneity within conventional cell populations, the development of clonal cell lines has resulted in candidates with superior growth characteristics and high IAV yields (Zinnecker et al., 2024). However, proteomic analysis could help to further understand specific properties of such clonal cell lines and help to identify best producers for vaccine manufacturing. In the present study, we compare proteome alterations between two IAV-infected suspension MDCK cell clones (C59 & C113, Sartorius, Germany) to elucidate differences in cell growth, size, metabolism and IAV productivity. Using advanced mass spectrometry, a total of 5177 host cell proteins were detected in both cell clones. Protein network analysis of the differentially expressed proteins with respect to cell growth revealed that fatty acid oxidation and branched-chain amino acid degradation were upregulated in the highly productive cells, whereas steroid biosynthesis and DNA replication were more active in the faster growing cells. After infection, 122 proteins were significantly upregulated (log2 fold change ≥1) in the high-producing cell line, including proteins associated with membrane trafficking. In addition, proteins that have cross-links to the IAV-NS1 protein and proteins that support virus production were identified. In addition, 98 proteins associated with antiviral signaling pathways such as Met and TNF signaling were downregulated (log2 fold change ≤1). In the less producing cell line, 77 proteins were downregulated and 57 upregulated after infection. Here, RNA metabolism seemed to be downregulated, whereas the TCA cycle and stress response were upregulated. Overall, we were able to identify important differences between a fast-growing and a high-producing clonal MDCK cell line, revealing potential bottlenecks and providing further insights into the efficient production of IAV in cell cultures.

Perfusion bioreactors are emerging as a powerful platform for continuous monoclonal antibody (mAb) production in Chinese Hamster Ovary (CHO) cells with advantages in productivity, product quality, and operational efficiency. However, conventional perfusion processes typically rely on offline sampling and delayed analytics to guide process decisions. This piecewise approach often leads to reactive rather than proactive control, where flow rates, feed concentration and downstream adjustments are only made after metabolite imbalances or performance issues arise. In this study, we present the design of an advanced perfusion bioreactor platform built around the Sartorius Biostat® B-DCU 2L bioreactor system coupled with a Repligen ATF (alternating tangential flow) controller. Enhanced with integrated online sensors, the system enables real-time process awareness and more responsive control. Alongside standard sensors (pH, dissolved oxygen, temperature) the system incorporates a multi-spectrum UV sensor and a capacitance probe to provide continuous data on metabolic state, nutrient consumption, cell density, and product yield. The central aim of this work is to demonstrate how a real-time sensing strategy can overcome the limitations of traditional offline workflows. Leveraging the Sartorius Cellca 2 CHO strain for maximum protein production, the system uses continuous monitoring to enable early detection of metabolic shifts to support dynamic control of perfusion rates, nutrient supplementation, and process parameter shifts including temperature and dissolved oxygen. This level of process visibility has the potential to boost productivity over extended culture periods, improve overall titre, and maximize space-time yield while simultaneously informing downstream operations. By transitioning from manual adjustments to adaptive control, this platform reflects a broader push toward integrated and data-driven biomanufacturing. Future work will focus on leveraging these sensor outputs to develop predictive control algorithms and fully closed-loop systems for robust, scalable mAb production.