Abstract 23P
Background
Novel immunotherapies targeting the adenosine pathway are in clinical trials, however, only modest monotherapy activity has been observed in non-prioritised patient groups. To optimise the chance of success with our A2AR-selective antagonist, EXS-21546 (‘546; NCT04727138, discovered in collab. with Evotec), we work to identify an adenosine-induced immunosuppression biomarker signature for clinical trial patient selection. Here we present initial transcriptional and functional data mapping the adenosine suppressed immune potential at the single cell level, and subsequent modulation through antagonism of A2AR with ‘546, along with first patient-selection modelling to prioritise patients for ‘546 therapy.
Methods
By leveraging disease-relevant primary human tissues, we model the patient specific anticancer immune potential, and begin to validate patient selection methods functionally with a translatable high content imaging platform amenable to primary human material. This platform is supported by end-to-end deep learning-driven image analysis; work is combined with orthogonal multi-omic characterisation of single cell effects induced by ‘546.
Results
We present preclinical mechanistic studies of A2AR antagonism on tumour infiltrating immune cells, leading to a foundational predictive adenosine suppression signature. The patient selection gene signature is correlated to functional profiling using a high content imaging platform with proven translational capabilities (Kornauth et al, Cancer Disc, 2022), demonstrating association of immune activity with antagonism of adenosine signalling by ‘546. Signatures and patient selection algorithms are cross-validated with publicly available data.
Conclusions
Combining deep learning of single cell functional and multi-omics profiling data of disease relevant primary model systems, we model the association of the immune response potential to A2AR antagonism in cancer to define a biomarker signature with the potential to identify patients likely to benefit from A2AR antagonism. This could be implemented in future studies of our clinical candidate ‘546 to deliver the right drug at the right time to the right patients.
Legal entity responsible for the study
Exscientia.
Funding
Exscientia.
Disclosure
G. Vladimer, I. Alt, R. Sehlke, A. Lobley, C. Baumgärtler, M. Stulic: Financial Interests, Personal, Full or part-time Employment: Exscientia; Financial Interests, Personal, Stocks/Shares: Exsciencia. K. Hackner, L. Dzurillova, E. Petru, L. Hadjari, J. Lafleur, J. Singer: Non-Financial Interests, Institutional, Other: Exscientia. N. Krall: Financial Interests, Personal, Full or part-time Employment: Exscientia; Financial Interests, Personal, Stocks/Shares: Exsciencia. J. Šufliarsky, L. Hefler, T. Fuereder: Non-Financial Interests, Institutional, Other: Exscientia. C. Taubert: Financial Interests, Personal, Stocks/Shares: Exscientia; Financial Interests, Personal, Full or part-time Employment: Exscientia. C. Boudesco, A. Payne: Financial Interests, Personal, Full or part-time Employment: Exscientia; Financial Interests, Personal, Stocks/Shares: Exscientia.