The Rise of Predictive Mice: How Pharmacological Humanization Advances Preclinical Drug Development

Abstract

Despite major advances in anticancer drug development, the successful translation of promising preclinical findings into effective clinical therapies remains a major challenge in oncology. Many drug candidates demonstrate strong efficacy in experimental models but ultimately fail during clinical development due to limited therapeutic benefit, unexpected toxicity, or poor reproducibility of preclinical outcomes in patients. Rather than attributing these failures solely to limitations of preclinical models, this thesis demonstrates that translational failures largely arise from the way models are designed, interpreted, and applied. Improving predictive performance therefore requires experimental strategies that more accurately reproduce the pharmacological and biological conditions encountered in patients. This research was based on the hypothesis that integrating clinically relevant pharmacokinetic-pharmacodynamic (PKPD) principles with biologically patient representative tumor models can improve translational predictability. The work first identifies key factors contributing to the disconnect between preclinical and clinical outcomes, including species-specific differences in drug disposition, pharmacological response, and conventional dosing strategies. Current approaches frequently rely on maximum tolerated dose (MTD) regimens in mice, resulting in drug exposures that exceed clinically achievable levels and may overestimate therapeutic efficacy. These findings support a pharmacological humanization framework in which patient-relevant drug exposure becomes the primary determinant of preclinical study design. Implementation of this strategy required the development of robust analytical methodologies. Two highly sensitive liquid chromatography–tandem mass spectrometry (LC–MS/MS) methods were developed and validated, enabling accurate quantification of ABT-751 and multiple pharmacologically relevant compounds across biological matrices. These analytical platforms supported comprehensive pharmacokinetic studies, exposure-response analyses, and more efficient experimental designs. Application of these methods demonstrated that the limited clinical efficacy of ABT-751 was not caused by insufficient systemic or intratumoral exposure. Instead, resistance was primarily associated with biological characteristics of the tumor microenvironment, particularly hypoxia. These findings highlight the importance of clinically relevant tumor models capable of identifying resistance mechanisms that remain undetected in conventional xenograft systems. Pharmacologically humanized mouse models were subsequently applied to investigate anticancer therapies under clinically relevant exposure conditions. Studies with trametinib and abemaciclib showed that patient-equivalent exposures produced less pronounced tumor responses than conventional MTD-based dosing but more accurately reflected clinical outcomes. These findings demonstrate that maximizing antitumor effects in mice does not necessarily improve clinical translation and that accurate exposure matching between species is essential for predictive preclinical research. The same principles were extended to central nervous system drug delivery. Investigation of the ATP-binding cassette transporters ABCB1 and ABCG2 demonstrated that effective inhibition of blood–brain barrier transport requires higher dual inhibitor exposures than previously achieved clinically. Furthermore, evaluation of approved JAK-STAT3 inhibitors for brain metastases revealed insufficient target inhibition at clinically relevant exposures, limiting their therapeutic potential. Overall, this thesis demonstrates that improving oncology drug development requires refinement of preclinical models through pharmacological humanization and biologically relevant experimental design. Integrating clinically representative drug exposure with translational tumor models provides a more reliable framework for predicting therapeutic efficacy, improving decision-making before clinical trials, and increasing the likelihood of successful translation of novel anticancer therapies.