The technologies that are reshaping biotherapeutics manufacturing, an EU perspective

By David Murray, Principal Consultant, Regulatory Strategy at Parexel International

The development of pharmaceuticals is an area where neither science, nor technology stand still’[i], as reflected in the recent proposal to repeal the current EU medicines directive. This includes current and emerging innovative technologies in biological medicinal product development, manufacture and testing. While many of these innovations show great promise, they do, as outlined herein, face some regulatory challenges in their translation to patient benefit. 

Rapid sterility testing

Rapid sterility testing is an attractive approach, in particular for cell and gene therapy products, due to their often-short shelf lives. Nonetheless, parenteral preparations in the EU must currently comply with the Ph. Eur. 2.6.1 test for sterility which requires microbiological culture and is, therefore, time-consuming.

However, it can be acceptable to use alternative methods, with the agreement of the competent authority – provided the alternative method enables an unequivocal decision to be made that compliance with the Ph. Eur. standards would be achieved if official methods were used (as per Ph. Eur. General Notices). While Ph. Eur. 5.1.6 (Alternative Methods for Control of Microbiological Quality) provides some useful details on the validation of such methods, clearer regulatory guidance is needed on what is required to demonstrate that rapid sterility methods are suitable for their intended purposes, as well as what the requirements are in relation to the identification of any potential contaminating microorganism. These developments would greatly enable the use of rapid sterility testing and help patients gain timely access to life-saving treatments. 

Multi-attribute methodologies

The promise of Multi-Attribute Methods (MAMs) by mass spectrometry (MS) is the ability to test multiple quality attributes (e.g., charge variants, glycans, oxidation and identity) within a single method run[ii]. Furthermore, these methods generate rich data sets that can be retrospectively reassessed for new peaks / new quality attributes that may be identified in the future which could be very useful for demonstrating comparability following process changes. In addition, these methods fit the enhanced method development approach outlined in the draft ICHQ14 guidance[iii], thereby easing their lifecycle management. As the performance characteristics of these methods may not be specific to the analyte, an established platform MAM when used for a new purpose will likely benefit from an abbreviated validation approach, if scientifically justified, in line with the draft revision to ICHQ2[iv].

Some regulatory challenges exist for the introduction and acceptance of these method types. One challenge is the frequent lack of correlation between MAMs and conventional methods due to the fundamental differences in the underlying measurement techniques. Importantly, it will be important that companies satisfy Agency reviewers that all key attributes of either the product or of impurities are detected with the updated MAM method. Guidance on the data requirements for the introduction of MAMs would prove useful in this regard. 

Digitization: modelling for process and product understanding and control, artificial intelligence / digital twins

Traditionally, process characterisation studies and control strategy development for biologics have been performed experimentally in accordance with 2016 EMA guidance[v]. For many medicines being developed for orphan or unmet needs, the use of modelling represents an attractive approach for shortening the development times associated with standard CMC development. Modelling can be applied to process characterisation and understanding, process control and product stability. Recent examples of modelling in authorised products include the use of modelling to justify specifications[vi] and shelf life[vii].  Nonetheless, a recent EMA survey on innovation highlighted how the use of modelling is an area where barriers are perceived in the current legislation/guidance[viii]. While some direction exists, more thorough guidance is needed, for example on real-time predictive modelling. Another area where guidance is needed is on how the models themselves, and subsequent updates to them, will fit the current EU variations legislation. What’s likely expected is that updates to the design and maintenance of these models will be considered a lower reporting category than updates that affect the actual impact of the model on control of the commercial process. The recently established Quality Innovation Group (QIG) at the EMA has included process models as a topic of focus, and a Q&A on models is expected in 2023[ix][x]. In the interim, biomanufacturers should continue to use the guidance that currently exists on models including the ‘Role of Models’ section in the EMA’s Points to consider for ICH Q8/Q9/Q10 guidelines[xi] and the ‘process models’ sections in ICHQ13[xii]. Regarding stability models, useful guidance exists in the EMA toolbox for Priority Medicines[xiii] and stability modelling approaches will be included in the forthcoming combined ICH Q1A-F, ICH Q5C stability guideline[xiv].

Digital twins replicate complex systems such as biopharmaceutical manufacturing processes. They use data and models to mirror the behaviour of their corresponding twin. The use of digital twins for process design and optimisation coupled with prior platform knowledge represents a less wasteful and less time-consuming approach to product and process development compared to the traditional execution of many scale-down experimental runs. While a recent discussion paper from the US-FDA[xv] explored some regulatory considerations for the use of digital twins / artificial intelligence (AI) in the development of biological products, there is still no clarity on regulatory expectations about where CMC AI can be used and how any supporting data should be presented in regulatory submissions. Promisingly, the forthcoming update to ICH M4Q (the common technical document on quality) is expected to provide guidance on the inclusion of information supporting artificial intelligence[xvi]. 

Continuous manufacture

To date, for biotechnological medicinal products, manufacturing typically occurs using a batch process. Despite the potential for increased process efficiency and reduced cost and time, end-to-end continuous manufacture (CM) has yet to be approved in an EU-licenced biologic product. The recent publication of ICHQ13xii and the recent revision to ICHQ5A(R2)[xvii] provide useful guidance for biomanufacturers developing CM processes, including guidance on viral safety. One barrier to the implementation of CM for biomanufacturing is the time-to-result for in-process tests applied to critical process steps. The use of surrogate markers, particularly in upstream processing, could allow the process to run continuously without significant holds as well as allowing rapid re-routing of non-compliant material (i.e., away from the product for release). Adequate guidance is needed on this alternative type of in-process control. Furthermore, a barrier recently identified is the lack of a global and harmonised framework allowing companies to use systems models to link CM and real-time releasex. It’s likely the aforementioned Q&A on models due in 2023 will address this. 

These are just some of the technologies reshaping biomanufacturing. A general challenge is in ensuring EU assessors and inspectors are equipped with the skills, training and relevant tools to regulate these new technologies. An even greater challenge is in obtaining global agreement amongst regulators on what is required. The onus is on developers to keep the agencies up to date by taking them on the journey with them during development. The recently established QIG at the EMA is an entry point into the regulatory system for CMC innovators and they are open to 1-to-1 meetings to discuss challenges and pathways.

Finally, an important challenge, that needs to be addressed by both innovators and regulators, is in communicating to the public that innovations in biomanufacturing, including artificial intelligence, does not result in lower quality medicines or negatively impact the safety and efficacy of medicines on the EU market and that the same legislative requirements apply. 

Return to Insights Center