Juggling CMC trade-offs with personalised cancer vaccines

Personalised cancer vaccines offer significant potential to address an unmet medical need in treatment-resistant disease with fewer side effects and less toxicity compared to standard treatment, such as chemotherapy. They achieve this by specifically targeting tumour cells and not the healthy cells.¹

Many clinical trials for personalised cancer vaccines are underway globally, with several in late-stage development and showing promising results.² There are, however, many challenges that innovators need to overcome from a development, regulatory, and CMC (chemistry, manufacturing, and controls) perspective.

As a still-novel class of products, personalised cancer vaccines currently lack clear regulatory guidance, which may lead to some discrepancy in the oversight and opinion of the different regulatory authorities regarding development.³

Additionally, while the US Food and Drug Administration (FDA) classifies all these products as therapeutic cancer vaccines, a comparable classification does not exist in the EU. Instead, products are classified according to composition. Most products will be classified as vaccines or chemical medicinal products, but some products, such as those composed of non-synthetic nucleic acids, are considered an advanced therapy medicinal product (ATMP) and therefore ATMP regulations apply. Regulatory changes currently being discussed may, in the future, result in the classification of all products containing synthetic nucleic acids as ATMPs. This would place the majority of cancer vaccines in the EU under the ATMP regulations.⁴

There are also challenges with identifying the right target(s) for the vaccine to ensure a good immunologic response. Here, artificial intelligence, machine learning, and other bioinformatic tools are highly beneficial.⁵ Given this complex and fast-changing environment, a multidisciplinary approach that brings together experts in vaccines, ATMPs, AI and machine learning, regulatory requirements, and both US and EU knowledge will be key.

Tackling CMC challenges

Perhaps the most complex part of personalised cancer development is CMC. This is partly due to the nature of personalised medicines and partly due to the very fast turnaround needed from identification of the patient to drug administration (usually two to three months). The time is of critical importance since these products target late-stage cancer patients. It’s important to note, however, that this short timeline must encompass patient biopsy, drug design, manufacture, control, and release of the product – all of which involve a complex logistical chain and careful timing and planning. Several areas of CMC development that have proven to be particularly challenging are presented below.

Complex bioinformatics

Personalised medicine follows a non-conventional production process, starting with the identification, selection and preparation of patient-specific input material using Next Generation Sequencing (NGS) techniques and vaccine design using artificial intelligence and machine learning.⁶ These are the areas that fall under regulatory scrutiny.

For example, NGS analysis must be conducted on validated protocols, qualified equipment, and by trained staff. Processes should be accredited according to globally recognised standards, including, for example, ISO 15189⁷ for quality management specific to medical laboratories, ISO 17025⁸ testing and calibration laboratory standards, College of American Pathologists (CAP) laboratory standards,⁹ and on-site audits. However, since NGS is well-established in clinical trials, these requirements are unlikely to be a major obstacle.

The bigger challenge lies with AI and machine learning, since there is currently no internationally approved regulatory framework for assessing the use of these innovative algorithms in the design of these types of products.¹⁰

Additionally, our experience has found that the in silico pipeline should remain unchanged during the clinical trial run. The challenge here is that, because the system is self-learning, using data to train and optimise, there needs to be a careful balance between the modifications and training of the system and keeping it in a steady state so that it is possible to make a comparison between patients participating in a clinical trial, as well as between different stages of clinical trials. Developers are advised to discuss with the regulators to what extent and when certain modifications can be introduced.

Potency of the investigational drugs

The next big challenge for CMC is how to establish the potency assay. These are normally used to quantitatively measure the biological activity of the drug in the disease-relevant system.¹¹ However, since these vaccines are designed for each patient-specific tumour, neither the vaccine nor the disease is comparable between the trial subjects, and no surrogate models exist to investigate biological activity of these products. As such, standard potency assays are not feasible and alternative solutions must be discussed with regulators beforehand.

Sterility testing

As a time-consuming process, sterility testing can be a bottleneck; however, there are several alternative options. One, where applicable, is to follow the principles of real-time or parametric release, where testing is not performed on each batch, but rather is dependent on demonstrating that pre-determined, validated sterilising conditions have been achieved throughout the manufacturing process.¹² Another would be to use alternative rapid sterility testing,¹³ though this typically requires bringing in another service provider, which may not help to shorten the timeline.

Another option might be to consider “sterility by design”, leveraging a risk-based approach, where the manufacturing process is designed to minimise any risk of microbial contamination.¹⁴

In some situations, regulators may be open to potentially allowing distribution to clinical sites without the final sterility test, on condition that the results are provided as soon as possible and ensuring the product is not used before final release.

Stability barriers

The next big CMC challenge is stability testing. Since vaccines are personalised and can only be produced after enrolment of a patient in the trial, “traditional” stability testing is not feasible. Developers must present well thought-out solutions to the regulators. Some approaches that we have seen regulators being open to include a range of studies where the manufacturer tests a whole spectrum of the possible product make-up. For example, peptides where the developer may test the most hydrophobic and the most hydrophilic ones.

Another possibility might be providing a mix of stability data of engineering and clinical batches manufactured using the same process. Also, leveraging prior knowledge on the platform technology might prove beneficial, specifically when the manufacturer has stability data on products manufactured using similar processes.

Changing regulatory environment

Due to the novelty of this product class, few regulatory guidances have been released so far, and the highly dynamic nature of the field means they are likely to be constantly updated.

Additionally, the lack of global harmonisation means approaches accepted by some agencies may not be accepted by others. This can make it difficult to navigate the personalised cancer vaccine space.

It is therefore important that developers regularly engage with the authorities and seek their endorsement for any proposed solutions. Developers should also closely monitor guidances, position papers, and news in the field. Additionally, to address discrepancies between authorities, it’s important to seek parallel or joint scientific advice, especially within the EU, where it is possible to leverage and use the opinion of one regulatory agency to inform the other agency about the current plans.

Disclaimer: The information provided in this article does not constitute legal advice. PharmaLex and Cencora strongly encourage readers to review available information related to the topics discussed herein and to rely on their own experience and expertise in making decisions related thereto.

About the authors

Dr Ilona Baraniak-Lang is associate principal consultant at PharmaLex. She is a vaccinologist and virologist, and specialises in global regulatory affairs strategies. She has successfully supported 200+ vaccine, ATMP, and other biopharmaceutical development projects, with clients including big pharma, SMEs, start-ups, governmental institutions, non-profit organisations, and academia.

Dr Anna-Lena Amend is a consultant at PharmaLex. She is a molecular biologist and specialises in global regulatory affairs strategies for ATMPs/cell and gene therapies, other biopharmaceutical development projects, and also nucleic acid-based vaccines. Typical clients encompass start-ups, SMEs, big pharma, and academia.