Integrated Population Pharmacokinetic, Pharmacodynamic, and Safety Analyses to Inform Dosage Selection in the Clinical Development of the ATR Inhibitor Tuvusertib.

Abstract

We present model-informed selection of the recommended dose for expansion (RDE) of investigational oral ATR inhibitor tuvusertib, by integrating clinical pharmacokinetics (PK), pharmacodynamics (PD), and safety data from DDRiver Solid Tumors 301 trial Part A1 (NCT04170153). A population PK (POPPK) model was developed to characterize PK and hemoglobin (HGB) reduction after multicycle treatment was simulated using a semi-mechanistic, multivariate POPPK/PD model of reticulocyte (RET), red blood cell (RBC), and HGB dynamics. A semi-mechanistic PK-efficacy model characterized concentration-dependent tumor growth inhibition (TGI) in ARID1A mutant xenograft models. The clinical exposure-PD relationship was described for phosphorylated Ser-139 residue of the histone variant H2AX (γH2AX) as a biomarker of ATR inhibition. POPPK simulations predict the average steady-state concentrations to exceed phosphorylated checkpoint kinase 1 (pCHK1) IC₉₀ at 100-180 mg once daily (QD) and 180 mg QD 2 weeks (w) on/1w off. Exposure-related PD suggested target engagement of ≥80% at ≥130 mg QD. POPPK/PD modeling showed partial HGB recovery and lower rates of Grade ≥3 anemia after multicycle treatment with 180 mg QD 2w on/1w off vs. 130 mg and 180 mg QD. Lesser HGB reduction was predicted for 100 mg QD vs. higher QD doses. Translational modeling indicated no effect of the one-week dosing break on TGI at 180 mg QD. The analysis supports tuvusertib 180 mg QD 2w on/1w off as the RDE and 100 mg QD as the no-regret dose for clinical evaluation. This example underscores the value of integrated quantitative pharmacology analyses to inform dose selection using a totality of evidence approach.