Heinrich Huber1, Hitesh Mistry
1Boehringer-Ingelheim RCV, 2SEDA Pharmaceutical Development Services
Objectives In oncology, the use of in-vitro to in-vivo correlations (IVIVC) is prevalent, linking in-vitro parameters like IC50 to in-vivo drug exposure in plasma and tumor growth. These correlations are crucial for experimental design and for dosage decisions of tumour xenograft studies based on in vitro efficacy from cell line models. We provide a mathematical analysis that primarily focuses on exploring the connection between these empirical IVIVCs of exposure-response relationships and a PD model of semi-mechanistic tumor-growth. The employed model that is used routinely at preclinical research in Boehringer Ingelheim to predict tumor growth inhibition by small-molecule kinase inhibitors. Method To accomplish this, rather than focusing on numerical evaluations, we provide a rigorous mathematical analysis of linking drug PK to tumour PD. This analysis integrates parameters including the drug’s peak-to-trough ratio (PTR), the Hill coefficient from in vitro dose-response curves, and pharmacodynamic characteristics unique to the xenograft model employing a specific tumor cell type. Specifically, we use a PD model of tumour radius growth (with a xenograft specific unperturbed growth rate g), whereby the growth is attenuated by an EMAX function. Thereby, the model describes the change of the tumour radius over time (dR/dt) in the following way, dR/dt = g – d_MAX* (conc_t(t)^hill/ (conc_t^hill(t) + IC50)), whereby IC50 refers to the free in vitro IC50, dMAX (dMAX > g) describes the maximum inhibition effect, specific to the xenograft and hill the Hill coefficient of concentration. Our study, therefore, quantifies the dependency of anti-tumor response on the interaction between drug properties and the type of xenograft used. For further details and justifications of the PD model, please refer to [1]. Result We find that the determinant of anti-tumor response—whether it’s the area under the curve (AUC) or the peak-trough ratio—may depend on the kind of xenograft employed. Our study quantifies this dependency by focusing on the interaction between drug properties and xenograft type. Specifically, we show that Hill coefficients >1 and very resistant models (characterized by g/dMAX > 0.8) foster peak-trough drivenness rather than AUC-drivenness in the anti-tumor effect of the compound. We illustrated our findings using a legacy dataset from 86 mouse xenograft experiments, featuring unbound IC50 values derived from anti-cell proliferation assays conducted over 5 days of drug incubation, ranging from 0.2-700 nM (median: 7.7). The total AUC varied from 1.84-4141 nM*h (median: 55), with the unbound fraction ranging from 0.04-2.3% (median: 0.53) over a study duration of 14 days with either once daily (qd) or twice daily (bid) dosing. Furthermore, we propose a method to predict population dose ranges in mouse clinical trials. Conclusion Our study suggests that drivers of efficacy in anti-tumor growth depend on the PK properties of the compound in interplay with the PD properties of the xenograft model used.
(1) Huber and Mistry, Journal of Pharmacokinetics and pharmacodynamics 2024
Reference: PAGE 33 (2025) Abstr 11475 [www.page-meeting.org/?abstract=11475]
Poster: Drug/Disease Modelling - Oncology