A Personalized Shot at Cancer

Personalized therapeutic cancer vaccines are on the rise

Genetic instability is a critical hallmark of cancer. Mutations in cancer cells can lead to the expression of tumour-specific neoantigens - often exquisitely unique to an individual - where mutated peptide products are broken down, presented by MHC molecules, and subsequently recognized by T cells. Such neoantigens can be used to develop personalized therapeutic cancer vaccines that trigger robust anti-tumour immune responses yet spare healthy tissues in a patient.

Importantly, a body of evidence strongly suggests the number of somatic mutations within a tumour directly correlates with T cell infiltration and often predicts overall effectiveness of non-vaccine immunotherapies due to the increased number of potential neoantigens.

Over several decades, there have been numerous attempts to bring effective therapeutic cancer vaccines to the market, but the field has been plagued with significant failure. So far, the only approval - a personalized dendritic cell vaccine called sipuleucel-T, used to treat prostate cancer - has not been as efficacious as trials promised. However, there has been a surge of activity in recent years based on technological advances in different therapeutic modalities and improved understanding of tumour biology. Better sequencing technology, development of sophisticated antigen selection platforms and combination immunotherapies have all contributed to a renewed optimism around neoantigen vaccines.

Recent success from Phase I and II trials reported by Moderna (Weber et al, 2024. Lancet) and BioNTech (Rojas et al, 2023 Nature) that demonstrate vaccine-induced immunogenicity and evidence of prolonged remission of tumours in patients has particularly reinvigorated the hope that effective personalized cancer vaccines could be a reality, especially given the vast array of modalities being pursued. However, questions persist as to how immunogenic each of these different vaccine platforms that include mRNA, DNA, viral vectors, and peptides can be, and to how to best position each therapy. Moreover, investigators are still refining which is the ideal disease setting for the best chances of success (early vs. late stage) and which neoantigens should be prioritized for each individual patient.

The rivalry that is shaping cancer vaccines

The vaccine strategy that triumphs remains to be seen but two of the key players are firmly engaged on the battle ground of oncology vaccine clinical trials. They also happen to be the same companies who went head-to-head in the development and scale-up of the world's first COVID-19 vaccines. Moderna's mRNA-4157 encodes 34 carefully selected neoantigens in a single base-modified mRNA. Delivered using lipid nanoparticles (LNPs) via intramuscular injection, the vaccine is clearly making headway in the clinic with prolonged recurrence-free survival in addition to favourable safety profiles (Weber et al, 2024 Lancet). In a starkly different approach, BioNTech's candidate autogene cevumeran - in collaboration with Genentech - is a mix of unmodified mRNAs formulated into lipoplexes that encode 20 neoantigens, which are dosed intravenously. Although too early to compare, the inherent property of unmodified mRNAs to be more immunogenic may eventually be a risky but decisive strategy that improves the potency of mRNA-based cancer vaccines.

Formulation, dosing, and mRNA format are not the only differences. Moderna tested its vaccine in what is termed the adjuvant setting, treating patients after resectable tumours had been surgically removed. In contrast, BioNTech chose patients with more advanced, later-stage melanoma that had metastasized or that was otherwise inoperable. Cancer drug developers with new modalities often chase later-stage disease because the unmet need is far greater, trials are shorter and safety liabilities are better tolerated. However, in earlier stage disease, tumours are often smaller and slow-growing so investigators have time to generate custom, patient-centric vaccines and crucially, once administered, there is sufficient time to mount an immune response before the tides turn in favour of the tumour. Although, BioNTech's rationale to target stage IV solid tumours is laudable, disclosure of preliminary data suggests clinical impact is unclear beyond immunogenicity.

However, a parallel trial where autogene cevumeran was tested in an adjuvant setting in patients with resectable pancreatic cancer showed promising results (Rojas et al, 2023, Nature). Among those receiving the vaccine, over half developed neoantigen specific T cells and of those, the majority experienced no tumour recurrence compared to patients who did not mount an immune response to the vaccine. Not only does this suggest that preventing cancer from 'coming back' may be the best setting for neoantigen vaccines, but it may also be possible to initiate immune surveillance to limit tumour recurrence. The disappearance of micro-metastases seen pre-vaccination in some of the patients dosed in this trial implies this may well be the case and hints at the systemic function of the vaccine.

A cut above the rest

Generating effective personalized cancer vaccines requires drug developers to pick the most patient-relevant and highest quality of neoantigens. They are routinely chosen based on their expression levels in tumour biopsies and their predicted binding affinity to HLA molecules as well as other parameters such as level of dissimilarity to non-mutated self-antigens (the greater the better), clonality and surface density and stability of peptide-HLA complexes, all of which impact recognition by T cells. Although much has been gleaned following extensive study in the last decade, what makes a good neoantigen in terms of immunogenicity and clinical efficacy is being debated (Lang et al, 2022, Nature Reviews Drug Discovery).

Identifying neoantigens can be laborious, particularly where there is a need to isolate and analyze neoantigen-specific T cells from cancer patient blood samples- the rationale being that these cells have been primed to recognize existing tumours-and to then deconvolute the identity of the peptides these T cell recognise. Harnessing this approach, neoantigens are incorporated into vaccines to boost a patient's pre-existing immunity rather than to generate de novo anti-tumour responses.

Alternate strategies under intense focus include AI-based algorithms that help predict which tumour-specific peptides are expressed and presented by the HLA molecules and crucially, whether cytotoxic T cells will react to kill the tumour. These sophisticated computational tools can be supplemented with empirical data from tumour samples or trials to consolidate predictions; current thinking suggests this is essential to pick the best neoantigens. Indeed, relying solely on machine learning is likely to give sub-optimal predictions given the risky of false positives. This is partly due to the lack of understanding of how to weigh these algorithms in terms of favourable peptide characteristics and other parameters that would prompt a sufficient anti-tumour immune response.

Conclusion

So where are we with this new treatment strategy and will they ever be widely available for patients? Undoubtedly, as many more vaccine candidates enter the clinic, the wealth of ensuing efficacy and safety data-such as what constitutes a good set of neoantigens to encode, success rate compared to disease setting and choice of modality will no doubt inform and improve design of the next-generation of effective personalized therapies in oncology. That shot at cancer could finally reach its mark.

Dan Rocca, PhD, is Senior Research Leader with the In Vitro Biology section of Charles River Laboratories' Discovery business.