Everything looks promising in the laboratory. The formulation meets the assumptions, the parameters are within acceptable limits, and the test reports are consistent. However, the final verdict will not be made in the laboratory. The real test of technology is production: working at full scale under pressure from time, costs, and regulatory requirements.
The differences between a 'working solution' and a 'repetitive process' become apparent when the product is transferred from R&D to the production plant. Scaling up changes the physics of the system: the mixing energies, temperature gradients, flow, and humidity dynamics are all different. Parameters that were stable at laboratory scale begin to behave unpredictably at industrial scale. Projects intended to reduce time-to-market can become stuck in the validation or technological correction phase for months.
The paradox is that technology transfer presents both risks and opportunities. Projects are sometimes lost at this stage, but new process solutions are also created that were not envisaged during the research phase. Industrialisation is not the mere implementation of a concept. Rather, it is the confrontation of the concept with reality, and often the second, equally creative stage of technology development.
In this article, we will examine why technology projects lose momentum between R&D and production, the most common errors in the transfer process, and how to leverage this stage to build a lasting technological advantage for the plant.
Simplification of the process architecture is a prerequisite for global transfer
The invention, which is described in the international patent application published under the number WO2022/123033 A1 in 2022 and is entitled 'Method for producing a virus', was created by the University of Oxford in response to issues that arose during the scaling up of the production of adenovirus vectors, particularly those used in vaccines. While the methods used in laboratory testing, including perfusion or medium exchange, were controllable, significant limitations were revealed when moving to a large industrial scale and transferring technology to many production plants. The process was expensive, operationally complex, and difficult to standardise between different locations. One particular challenge was achieving reproducible virus performance in large bioreactors while maintaining a high viral particle titre without implementing complex perfusion systems. These problems only became apparent at the time of real industrialisation and global technology transfer.
The solution is to use a modified fed-batch model, which involves the controlled addition of a fresh medium without necessitating a full medium exchange or perfusion. Eliminating these stages significantly reduces the operational complexity of production, the number of critical points requiring monitoring, and the time taken to prepare the installation for commercial production. An additional advantage is the increased ability to transfer technology between different production plants. This helps to simplify the process architecture and increase its scalability and replicability across multiple plants.
Scaling changes kinetics – stabilisation of UF/DF processes
The invention, belonging to the Bristol-Myers Squibb Company and described in US patent application number US2023/0374064 A1 of 2023, entitled 'Methods for Concentrating Proteins', was created in response to problems observed during the scaling of ultrafiltration/diafiltration (UF/DF) processes involving biological proteins. While the concentration process was stable and predictable in laboratory conditions, significant differences in process time, flow dynamics, and the level of protein aggregates and particles in the micrometric range were observed when moving to larger volumes.
When transferring from a small-scale model to commercial-scale production, it was noted that traditional batch or fed-batch modes led to changes in concentration kinetics and increased high-molecular-weight aggregation. This meant that the process could not be fully predicted in the event of a change in volume or between different plants.
In response, the ‘pseudo-batch loading’ configuration was developed, which functionally separated the concentration stage from the initially loaded retentate. This solution enabled a more stable and repeatable process profile to be obtained between scales, reducing particle variability and improving data compatibility between the laboratory model and commercial production. This invention is the direct result of research carried out to solve problems that arose during the process transfer.
Shortening the process in response to GMP restrictions
The 2024 invention entitled 'One-step method for producing circular RNA', published under No. WO2024/055941 A1 and belonging to Suzhou Abogen Biosciences Co., addressed the limitations of traditional circular RNA (circRNA) production processes. These processes involved multi-stage synthesis and purification of precursors under research and development conditions. In industrial conditions, however, this process was found to be too lengthy, complicated, and susceptible to changes in parameters such as temperature, reaction time, and magnesium ion concentration.
When attempting to scale up the process for commercial RNA production, including for vaccine and therapeutic applications, it was found that the need to purify the precursors separately before circularisation increased the risk of error, extended the production cycle and complicated GMP validation.
The main advantage of the invention, which involves the one-step synthesis of circular RNA, is that it significantly simplifies the production process. The elimination of the need to purify the precursors separately before circularisation reduces the number of unit operations, shortening the production time and reducing the risk of operational error.
Shortening the production cycle is particularly advantageous in an industrial environment, especially for the production of therapeutic RNA and vaccines. The single-vessel reaction also minimises exposure to degradation and makes the process less susceptible to temperature fluctuations and reaction parameters at large volumes.
Resistance to hardware differences in the lyophilisation process
The invention described in the international application published under the number WO2022/101461 A1 in 2022 and entitled 'Enhanced formulation stabilisation and improved lyophilisation processes’ (Biontech SE) was created against the background of problems related to the transfer of lyophilisation cycles from research to commercial production. While the freeze-drying process was stable in laboratory conditions, significant deviations in shelf temperature, chamber pressure and primary drying time were observed when moving to full-scale production and changing equipment between production plants.
These differences led to the collapse temperature being exceeded, the lyophilisate structure becoming unstable, and the risk of product quality being lost. The problem became apparent during the transfer of the cycle to other GMP installations, where even minor structural differences in the devices resulted in significant process variability.
The solution consists in developing a mathematical model based on heat and mass balances, which can be used to predict product temperature and optimise temperature ramps. This helped to create a lyophilisation cycle that is resistant to deviations in hardware parameters. The solution reduces the risk of the lyophilisate structure collapsing, of quality being lost, and of the batch having to be repeated, making the transfer of technology between plants much easier.
Digital technology transfer
The invention of 2024, entitled 'Biosecure Digital Twin for Cyber-Physical Anomaly Detection and Biological Process Modelling' and belonging to National Resilience, Inc., was created in response to the growing complexity of biopharmaceutical production environments, particularly in the context of process transfer between multiple sites and the integration of OT/IT systems. During the technology transfer process, issues relating to inconsistent data monitoring, sensor discrepancies, different SCADA/DCS systems, and infrastructure vulnerability to cybersecurity issues were identified.
Using a digital twin in the bioproduction environment offers benefits relating to the monitoring, safety, and control of industrial processes. One of its main advantages is its ability to detect real-time anomalies by aggregating sensor data and applying machine learning algorithms.
The system enables data integration from multiple production sites and federated learning, facilitating process comparison and synchronisation without centralising sensitive data. This solution significantly improves the security of technology transfer between parties and enables the swift identification of operational discrepancies.
Another benefit is strengthening the cybersecurity of the production environment by monitoring cyber-physical incidents and isolating IT and OT layers. In practice, this improves the quality and stability of the process and increases the resilience of the industrial infrastructure to external threats.
Technology transfer is neither an administrative stage of implementation nor a technical addition to completed research. Rather, it is the moment at which the technology's maturity is verified, and its real limitations are revealed. Only in a full-scale production environment, under the constraints of GMP and the associated time and cost pressures, can it be determined whether the process is truly repeatable, resilient, and scalable. The line between a failed transfer and a breakthrough improvement can be thin. Whether it is a failure or a breakthrough is determined by whether the organisation treats industrialisation as the second, equally important inventive step.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
