Preclinical characterization of CNS-active, mutant-selective fourth-generation EGFR inhibitors with potent activity against single, double, and triple mutant EGFR variants including T790M and C797S

Background: EGFR activating mutations are observed in 10–50% of NSCLC patients and the common mutations (L858R [L] and exon 19 deletions [D]) are initially sensitive to first- second- and third-generation EGFR inhibitors (e.g., erlotinib [1G], afatinib [2G], and osimertinib [3G], respectively). However, on-target resistance is observed in a substantial percentage of patients, with T790M (T) and C797S (C) mutations observed most frequently (post-1G/2G and post-3G, respectively). Furthermore, EGFR mutational heterogeneity can increase during treatment with existing inhibitors, activity against wild type (WT) EGFR leads to dose-limiting toxicities, and tumors commonly metastasize to the brain. We therefore set out to identify a fourth-generation (4G) EGFR inhibitor with 1) potent activity against all 8 major single (L and D), double (LT, DT, LC, DC), and triple (LTC, DTC) mutant variants, 2) selectivity over WT EGFR, and 3) central nervous system (CNS) activity. Materials and Methods: In vitro drug activity was determined by assessing effects (IC50 s) in kinase assays and on tumor and BaF3 cell lines expressing WT or mutant forms of EGFR. Efficacy studies were conducted using engineered BaF3 and human lung cancer cells, or patient-derived models; mice were dosed orally once daily (QD) at doses that did not exceed the maximum tolerated dose. Results: We have identified multiple advanced lead compounds with the profiles of a 4G EGFR inhibitor. A representative compound, Cmpd A, potently inhibited the kinase activity of EGFR L, LT, and LTC variants with IC50 s £1 nM, compared to IC50 s for osimertinib of 0.4, 0.2, and 360 nM respectively. In engineered BaF3 cells, Cmpd A inhibited viability of all 8 EGFR single, double, and triple mutant variants with IC50 s <12 nM, while erlotinib and osimertinib were inactive against all 4 T790M- and all 4 C797S-containing variants, respectively. Cmpd A inhibited the activity of all 8 variants with 11- to 69-fold selectivity over WT EGFR, which compared favorably to the 6- to 7-fold selectivity erlotinib exhibited for L and D. In vivo, Cmpd A (30–40 mg/kg QD) induced tumor regression in all 8 tumor models tested, including those containing EGFR single (L or D), double (LC or DT), and triple (LTC or DTC) mutant variants. Finally, Cmpd A had substantial antitumor activity and significantly prolonged survival in an intracranial tumor model. Conclusions: We have identified potent mutant-selective and CNS-active 4G EGFR inhibitors. These data strongly support that inhibition of both major activating mutants, and key mutants associated with clinical resistance to all existing inhibitors - including C797S-containing variants associated with both 1st and 2nd line osimertinib failure - is achievable with a single molecule. Conflict of interest: Ownership: Employee and stock holder of Theseus Pharmaceuticals, Inc.