Ye Yao (1), Junsheng Xue (1), Daming Kong (1), Mengyi Han (1), Yuchen Guo (1), Xunqiang Wang (2), Fan Feng (2), Tianyan Zhou (1)
(1) Department of Pharmaceutics, School of Pharmaceutical Sciences, Peking University, China, (2) Chia Tai Tianqing Pharmaceutical Group Co., Ltd, China
Objectives: TQB-3395 is an orally bioavailable, novel small-molecule candidate in the same class of afatinib investigated by Chia Tai Tianqing Pharma, which potentially blocks cellular key signaling resulting in tumor growth inhibition or tumor regression. Its preclinical anti-cancer activity has been demonstrated in vitro and in vivo. This study aimed to preliminarily anticipate its human efficacious dosage through pharmacokinetic/pharmacodynamic (PK/PD) modeling.
Methods: PK in mouse, rat, and dog following intravenous and oral administration of TQB-3395 were characterized using population modeling methods, which were then extrapolated to human through allometric scaling (AS) [1]. Preclinical PK/PD models were established to quantify drug efficacy in inhibiting tumor growth using mouse PD data collected from several types of xenograft, i.e. lung cancer HCC827, NCI-H292, H1975, esophageal cancer Eca-109, pharyngeal squamous carcinoma FaDu. Classical exponential model, two-phase Simeoni and Koch model, as well as Logistic and Gompertz model where a theoretical upper limit of tumor size exists were attempted to fit the unperturbed tumor growth [2-5], and a linear or nonlinear Emax function was introduced to describe the drug effect. Model-based simulations were performed to identify a threshold steady state concentration for tumor stasis (when the drug exhibited a linear inhibitory effect on tumor growth) or threshold daily dose (in the case of nonlinear inhibition) in mice for each cancer type. Assuming the same unbound plasma concentration at steady state was required to achieve tumor stasis between species, clinical efficacious doses were predicted [6].
Results: Clearance and volume of distribution in human were predicted by single species AS, simple AS, and Rule of Exponents (ROE) method to be 7.33 L/h and 28.1 L on average, and a mean oral bioavailability of 26.8% was adopted. Since similar plasma protein binding was observed among a variety of species, the unbound fraction was assumed to be the same between mouse and human. The PK/PD models were established and validated for each cancer type to simulate the threshold concentration or dose in mice, and the human efficacious dose was calculated accordingly after incorporating the predicted PK parameters. A murine threshold concentration (or dose) of 0.00826 mg/L, 0.0277 mg/L, 0.65 mg/kg, and 1.15 mg/kg was simulated for HCC827, Eca-109, NCI-H292, and FaDu, respectively, and the corresponding human oral dose was 5-8 mg, 18-26 mg, 33-48 mg, and 59-85 mg per day. TQB-3395 was less effective in mice bearing H1975 tumors with a threshold concentration of 1.39 mg/L, so up to 911-1324 mg per day was expected to be clinically efficacious.
Conclusions: PK of TQB-3395 was extrapolated to human (and was comparable to observed apparent clearance in cancer patients). Based on preclinical PK/PD relationships, efficacious doses to treat patients of certain cancer types were anticipated, setting safety issue aside, which may benefit future clinical trial design.
References:
[1] Mahmood I. Adv Drug Deliv Rev (2007) 59, 1177-92
[2] Simeoni M et al. Cancer Res (2004) 64, 1094-101
[3] Koch G et al. J Pharmacokinet Pharmacodyn (2009) 36, 179-97
[4] Vaidya VG et al. Int J Biomed Comput (1982) 13, 19-36
[5] Laird AK. Br J Cancer (1964) 13, 490-502
[6] Eigenmann MJ et al. Mol Cancer Ther (2016) 15, 3110-9
Reference: PAGE () Abstr 9384 [www.page-meeting.org/?abstract=9384]
Poster: Drug/Disease Modelling - Oncology