Abstract 7143: Development of a peripheral dazostinag (TAK-676) activation RNASeq signature to evaluate STING pharmacodynamic modulation in human solid tumor clinical trials

Abstract Dazostinag, a novel STimulator of INterferon Genes (STING) agonist, is currently being evaluated with pembrolizumab for colorectal and head & neck cancers (iintune-1, NCT04420884). RNA sequencing (RNASeq) is a valuable tool for pharmacodynamic (PD) biomarker development, offering comprehensive transcriptome profiling and a wide dynamic detection range. For the iintune-1 Phase 1/2 study, we developed an RNASeq peripheral PD signature for STING activation and demonstrated its utility as a secondary endpoint to measure biological responses to dazostinag. To develop the PD signature, peripheral blood mononuclear cells (PBMCs) from 7 healthy donors were treated ex vivo with vehicle or dazostinag (0.33-10 μM, 6 hours, 3 replicates) and samples were processed for RNASeq. Protein-coding genes with a maximum FPKM ≥ 1 and low variance between replicates were filtered and analyzed utilizing dose-response mixed model for repeated measures. Genes with significant dose responses (FDR ≤ 0.05 and fold change ≥ 1.5 at 0.33 μM) underwent network analysis to identify biologically connected genes. A 19-gene network with significant response to dazostinag was identified and termed STING_19. Among these genes, 17 are in the interferon (IFN) pathway, consistent with the known IFN activation effect of STING agonists. An independent RNASeq dataset from human PBMCs treated ex vivo with dazostinag (0.12-10 μM) was used to evaluate STING_19 against hallmark IFN-α and IFN-γ signatures. STING_19 demonstrated a robust dose-dependent induction by dazostinag, with a median 70-fold increase at 10 μM, greater than the 4 to 6-fold induction of IFN-α and IFN-γ signatures. Additionally, STING_19 induction was confirmed in a syngeneic mouse model, showing a 55-fold increase at 6 hours post-treatment with 1 mg/kg dazostinag. In the iintune-1 study, for dazostinag single agent arm, STING_19 exhibited a dose-dependent induction in blood, starting from 0.8 mg/kg with a median 18-fold increase at the highest dose of 14mg/kg, indicating a robust PD response. Pembrolizumab combination arm showed a similar induction of STING_19 in response to dazostinag, with pembrolizumab alone at low dazostinag dose (as a surrogate for monotherapy) showing minimum STING_19 induction. Within 0 to 24-hours post-treatment, maximum induction of STING_19 was typically at 24 hours for doses up to 5 mg, and at 10 hours for doses between 7-14 mg, suggesting dose-dependent STING activation dynamics by dazostinag. This induction returned to baseline prior to subsequent weekly doses. In summary, we developed a robust RNASeq-based peripheral STING activation PD biomarker using ex vivo treated PBMCs to evaluate the biological response to STING agonist treatment. This framework could be applied for the development of RNASeq based peripheral PD biomarkers for other investigational assets. Citation Format: Kai Ding, Xin Tong, Radha Ramesh, Samanthi Perera, Yu-Chung Huang, Alex Parent, Jeffrey Raizer, John Gibbs, Neil Lineberry, Vicky Appleman, Somnath Bandyopadhyay, Richard C. Gregory. Development of a peripheral dazostinag (TAK-676) activation RNASeq signature to evaluate STING pharmacodynamic modulation in human solid tumor clinical trials [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 7143.