The international neoantigen initiative Tumour Neoantigen Selection Alliance (TESLA) has identified key parameters for immunotherapy advancement.
Neoantigens
Somatic alterations are a key hallmark of cancer. Alterations result in the generation of mutated peptide fragments, also known as neoantigens. These peptides are presented on MHC class I molecules in order to elicit a protective anti-tumour immune response. Experts have long viewed neoantigens as promising therapeutic targets as they are tumour specific. However, hundreds of mutations exist in a tumour and identifying the ones that trigger an immune response is critical. Many approaches have attempted to identify therapeutically relevant neoantigens, including combining tumour sequencing data with bioinformatic algorithms. Nevertheless, there has been no reference data to compare these approaches and parameters that govern tumour epitope immunogenicity remain unclear.
TESLA
The Parker Institute for Cancer Immunotherapy (PICI) and the Cancer Researcher Institute launched an initiative known as TESLA in 2016. This consortium brings together over 36 leading neoantigen research groups in academia, non-profit and industry. Using predictive algorithms and machine learning, the group are attempting to search for which cancer neoantigens can be recognised and stimulate an immune response.
In a study, published in Cell, researchers reported results obtained by TESLA. The team identified key parameters governing tumour epitope immunogenicity. Additionally, they analysed predictions made by multiple independent pipelines on a common set of tumours samples and used a centralised set of validation experiments.
Specifically, the team assessed 608 epitopes for T cell binding in patient-matched samples. They developed a model based on five peptide features associated with presentation and recognition. The features included: binding affinity, tumour abundance, binding stability, fraction hydrophobic and mutational position. The model filtered out 98% of non-immunogenic peptides with a precision above 0.70. The team also validated these findings in an independent cohort of 310 epitopes prioritised from tumour sequencing data and assessed for T cell binding.
Implications
This research provides insight into what features may be important in neoantigen prediction. Moreover, it also has the potential to improve drug developers’ and researchers’ mathematical algorithms.
Daniel Wells, corresponding author, stated:
“Our aim is that data produced from TESLA becomes the reference standard when developing a new neoantigen-based treatment.
If every method, old and new, used the data to benchmark their predictions, the whole field would be able to collaborate and iterate on new methods much more quickly.”
The team noted that the data from TESLA is available for download and serves as an open benchmark to accelerate the development of neoantigen-based therapies.
Image credit: Meletrios Verras – canva.com
