Adriana Savoca1, Diana Zindel2, Radek Polanski7, Demetrios Kostomiris3, Mohammad Pirmoradian4, Sara Talbot5, Nicolas Floc’h5, Aisha Swaih5, Poppy Winlow2, Paul Davey6, HIlary Lewis5, Clare Thomson6, William McCoull6
1Clinical Pharmacology and Quantitative Pharmacology, Clinical Pharmacology and Safety Sciences, R&D, AstraZeneca, Cambridge Biomedical Campus, 2Mechanistic Biology and Profiling, Discovery Sciences, R&D, AstraZeneca, Cambridge Biomedical Campus, 3Mechanistic and Structural Biology, Discovery Sciences, R&D, AstraZeneca, 4Biomarker, Discovery and Development, Translational Science and Experimental Medicine, R&D, AstraZeneca, 5Bioscience, Oncology OTD, AstraZeneca, Cambridge Biomedical Campus, 6Chemistry, Oncology OTD AstraZeneca, 7Clinical Trial Transparency, R&D Clinical Science, AstraZeneca
Introduction/Objectives: EGFR Exon20 insertions (Exon20Ins) are the third most common EGFR activating mutation in non-small cell lung cancer after the single point L858R mutation in Exon21 and Exon19 deletions [1]. Because patients with EGFR Exon20Ins are generally less sensitive to traditional EGFR Tyrosine Kinase Inhibitors (TKIs) and have poor prognosis [2], this target is of clinical interest as shown by recent approval of the bispecific antibody amivantamab [3]. We wanted to build a modelling framework linking pharmacokinetics (PK)/ pharmacodynamics (PD) and efficacy of irreversible inhibitors targeting EGFR Exon20Ins mutation to support discovery of new drugs and gain insights into clinical efficacy observed for inhibitors that are currently in development. Methods: We combined in vitro data on proprietary irreversible inhibitors targeting EGFR Exon20Ins from various experiments to inform a PD model featuring a mechanistic description of protein turnover and phosphorylation inhibition, aimed at describing phosphorylated EGFR (phosEGFR) reduction following inhibitor treatment. Specifically, in NCI-H2073 cells engineered to host Exon20 SVDIns mutation, EGFR turnover was studied via SILAC/MS experiments and time-course of phosEGFR inhibition after dosing a few inhibitors was monitored using ELISA. Kinetic parameters represent the only compound-dependent parameters in the model and were determined from a biochemical binding assay for the tested inhibitors. Next, we validated our model predictions in vivo and examined in vitro/in vivo correlation using PD data from PK/PD studies in mouse NCI-H2073 SVDIns and LXF2478 (ASVIns) cell line-derived xenografts (CDXs) and suitable compartmental pharmacokinetic models. We also incorporated tumour growth inhibition data from mouse xenograft studies into our analysis to determine the target engagement level required for significant tumour regression. Finally, we combined published clinical PK data and models [4, 5] with the proposed PD model to generate clinical simulations of phosEGFR reduction following doses of i) osimertinib (third generation EGFR TKI) and ii) zipalertinib (TKI targeting EGFR Exon20Ins mutation) at clinically investigated levels in patients with EGFR Exon20Ins. Results: The PD model successfully integrated in vitro data to describe phosEGFR reduction in EGFR Exon20 SVDIns cells for a few irreversible inhibitors by adapting their respective kinetic parameters. The model was able to predict in vivo datasets across different compounds and tumour models, demonstrating a good in vitro/in vivo correlation. We discovered that achieving high (>~85-90%) and sustained levels of phosEGFR reduction is necessary for significant tumour regression, in line with target engagement requirements for other TKIs (e.g., targeting other EGFR mutations and MET [6, 7]). Clinical PD simulations for the irreversible inhibitors osimertinib and zipalertinib aligned well with their respective clinical efficacy observations in patients with EGFR Exon20Ins, suggesting that i) limited target engagement may have contributed to the low-modest efficacy of osimertinib in this population [8], ii) high and sustained levels of target engagement were achieved across efficacious clinically tested doses for zipalertinib. Conclusions: The developed model is a valuable tool to interpret in vitro data of EGFR Exon20Ins irreversible inhibitors and support screening and selection of new molecules. Our findings on target engagement requirements can inform efficacious dose predictions in the clinic. Clinical PD simulations suggest the model can provide insights into clinical efficacy of in-development inhibitors and help understanding competitive differentiation. Overall, our results enhance confidence in the model as a tool to bridge the preclinical to clinical PKPD/anti-tumour activity relationship for irreversible EGFR Exon20Ins inhibitors and guide discovery of new molecules.
1. Yasuda, H., et al. (2012). Lancet Oncol 13(1): e23-31. 2. Yang, G., et al. (2022). Front Pharmacol 13: 984503. 3. Chon, K., et al. (2023). Clinical Cancer Research, 29(17), 3262-3266. 4. Brown, K., et al. (2017). Br J Clin Pharmacol 83(6): 1216-1226. 5. Piotrowska, Z., et al. (2023). Journal of Clinical Oncology, 41(26), 4218-4225. 6. Yates, J. W. T., et al. (2016). Molecular Cancer Therapeutics 15(10): 2378-2387. 7. Jones, R. D. O., et al. (2021). British Journal of Pharmacology 178(3): 600-613. 8. Yasuda, H., et al. (2021). Lung Cancer, 162, 140-146.
Reference: PAGE 33 (2025) Abstr 11668 [www.page-meeting.org/?abstract=11668]
Poster: Drug/Disease Modelling - Oncology