218P - A novel tumor adenosine signature to guide indication selection for adenosine pathway inhibitors

Background

Adenosine (ADO) is a major immunosuppressive player within the tumor micro-environment (TME), acting via A2A and A2B receptors or through the equilibrative nucleoside transporter 1 (ENT1). Quantifying ADO in the TME could help identify tumors more likely to respond to the several ADO pathway-targeting agents that have entered clinical trials. However, several limitations make ADO measurement impractical. Gene signatures have been proposed to overcome this limitation, but none truly reflects ADO content in the TME. Here, we present the first ADO gene signature based on spatial ADO quantification in human tumors and demonstrate its potential for indication selection.

Methods

Thirteen tumor biopsies from six cancer types were analyzed for ADO using quantitative mass spectrometry imaging (qMSI). A total of 183 regions of interest (ROIs) were selected based on ADO content (high vs low) and profiled by spatial transcriptomics (GeoMx DSP). ROIs were split into training (70%) and test (30%) sets. A linear mixed model adjusted for tumor type and patient was applied to the training set to identify differentially expressed genes (DEGs) in high vs low ADO regions (FDR <0.05). g:profiler was used to perform pathway enrichment. Lasso regression identified the most predictive DEGs, and the resulting gene signature predictive accuracy was evaluated using the Area Under the Receiver Operating Characteristics (AUROC) on the test set.

Results

We identified 249 DEGs when comparing high vs low ADO ROIs across tumor types. Pathway enrichment analysis revealed associations with metabolism and immune activation. Twenty-five DEGs had the highest cross-tumor predictive power, as validated in the test set (AUC = 0.905). In TCGA, the new ADO signature was higher in tumors compared to healthy tissue and showed variable expression and prognostic value across tumor subtypes.

Conclusions

We developed the first ADO signature based on ADO content in human tumors. The novel signature is a powerful tool to prioritize indications that would benefit the most from ADO-targeting drugs, and holds potential to unveil the mechanisms of ADO immunosuppression and to identify individual patients responsive to treatment.

Legal entity responsible for the study

iTeos Therapeutics.

Funding

iTeos Therapeutics.

Disclosure

S. Dekoninck, N. Rosewick, T. Sanders, L. Chaible, J. Marchante, H. Shehade, F. Strozzi, E. Houthuys, Y. McGrath, R. Marillier, M. Rossetti: Financial Interests, Personal, Full or part-time Employment: iTeos Therapeutics; Financial Interests, Personal, Stocks/Shares: iTeos Therapeutics. C. Martinoli: Financial Interests, Personal, Advisory Role: iTeos Therapeutics.