The immune system as a therapeutic tool
The immune system has evolved to recognise and eliminate cells that are infected with viral or bacterial pathogens or abnormal such as cancer cells. Accumulating evidence suggests that the immune response, in particular T cells, plays an important role in the clearance of cancers. However, these responses do not always result in resolution of disease, as cancerous cells are often able to ‘hide’ from the immune response, resulting in tumour growth and progression.
The induction of an effective anti-tumour T cell response requires engagement between peptide-MHC (pMHC) presented on the surface of tumour cells and the T cell receptor (TCR) on circulating T cells. The ability of peripheral T cells to respond to the presented antigen is determined during their development by the strength of interaction of their TCR with self-MHC. If the interaction is too strong, the T cell is deleted from the repertoire (negative selection), preventing recognition of self in the periphery leading to autoimmunity. As tumours are derived from normal cells, the vast majority of pMHC presented on tumour cells consists of self-peptides preventing their detection. Thus, a much greater amount of a self-peptide or a new peptide distinct to the tumour must be presented in order for the T cells to detect them. Ultimately, the success of harnessing T cells to target tumour cells is dependent upon their ability to distinguish tumour from normal cells.
As cancers develop, they acquire a number of somatic mutations in DNA which, if they occur in exons of expressed genes, may alter the sequence and result in formation of neo (new) antigens. The recognition of these neo-antigens by T cells indicates a possible ‘Achilles heel’ of tumour cells, ideal for therapeutic exploitation. Our understanding of antitumour T cell responses has paved the way for immunotherapies which utilise patient T cells, such as adoptive transfer of autologous tumour infiltrating lymphocytes (TIL) and re-activation of TIL by blocking inhibiting molecules, termed checkpoint inhibition. In both these therapies, T cells specific for somatic mutations in tumours have been shown to play a key role in their efficacy.
Mutanome based personalised immunotherapy
The ability to identify somatic mutations within an individual’s tumour represents a highly attractive target for potential immunotherapy. However, until recently, the ability to scan a tumour genome to identify these potential targets was limited. The emergence of next generation sequencing (NGS), a collective term for a number of different high throughput sequencing technologies, has revolutionised the study of cancer genomics. This technology allows sequencing of each patient’s tumour genome, identifying tumour-specific mutations, termed the mutanome. The reduction in both cost and time of NGS, alongside ever increasing bioinformatics and protein prediction tools, has led to advances in tumour genome analysis and neo-antigen prediction. Currently, many collaborative efforts are underway, such as the 100,000 genomes project and The Cancer Genome Atlas (TCGA), to catalogue cancer mutanomes across an array of tumour types. Employing this technology has led to the discovery of a greater number of tumour-specific mutations than initially thought, with up to 100s of mutations identified in different tumours. These analyses also revealed that the mutational load of a tumour correlates with the efficacy of immunotherapy, with a greater number of mutations giving better responses. It is currently thought that >95% mutations identified are unique to the individual patient’s tumour and are not shared between tumour type or healthy cells, a feature that has hampered the ability to exploit differences therapeutically. As a result, personalised vaccines may be required to target neo-epitopes regardless of the mutational load of the tumour to maximise therapeutic efficacy. Therefore a key challenge is how immunogenic cancer mutations of relevance can be identified and therapeutically exploited for an individual.
Identifying tumour antigens and future directions
While a vast number of mutations have been identified from mutanome analysis of tumours, only a small proportion of these result in the formation of a neo-antigen presented on MHC class I (MHC I) or II molecules. Therefore, to enable prediction of neo-antigens many aspects have to be considered such as i) whether the mutated gene is transcribed and translated into protein that is expressed within the cell; ii) the many processing steps that generate the peptide epitopes; and iii) how well these will bind to MHC I to allow stable presentation at the cell surface. The use of biochemical assays, prediction algorithms and molecular modelling techniques have allowed the accurate prediction and identification of a number of cancer neo-antigens. However, the final and possibly most complex aspect that is yet to be well defined is whether these presented neo-antigens are immunogenic. Proof of concept studies, using current methodologies, have identified neo-antigens from both mouse and human tumours, although, the only way to determine immunogenicity was to examine responses in vivo, with varying success of predicting responses providing little correlation. This suggests that the propensity of neo-antigens to elicit an immune response may only occur in a small fraction and is difficult to predict.
In order to better predict immunogenic neo-antigens, each aspect needs to be carefully considered. One parameter that excludes many potential neo-antigens is selection of high affinity MHC binding peptides, since they would be expected to have a longer half-life at the cell surface and maximise T cell recognition. Emerging evidence, however, shows that low affinity antigens are as efficacious in activating tumour-specific T cell responses as their high affinity cousins. It is therefore important to identify ways to predict immunogenicity that do not rely on binding affinity alone. Developments in our understanding of pMHC and TCR interactions through crystallography and molecular modelling provide us with valuable insights for predicting immunogenicity of neo-antigens. Nevertheless, at present the time taken for these analyses does not make it amenable to identify good neo-antigen candidates for clinical adoption. A greater understanding of how antigens are generated and the biochemical properties that underpin immunogenicity is needed to better predict neo-antigens. These advances will allow development of personalised cancer vaccines and advance the success of immunotherapy in combatting cancer.
Associate Professor in Cancer Immunology
Postdoctoral Research Fellow University of Southampton
Cancer Research UK Immunology Award
Cancer Research UK are inviting applications from non-cancer immunologists – to encourage these scientists to explore how their area of research could advance our understanding of the role the immune system can play in fighting cancer. They are providing three years of funding, up to £300,000, that can be used to fund scientific posts and associated running costs. You can find more information on the BSI blog or at www.cruk.org.uk/immunology-award.
Image credit: Shutterstock