Fatima Zahra Marok (1), Jan Georg Wojtyniak (1,2), Matthias Schwab (2,3,4), Thorsten Lehr (1)
(1) Clinical Pharmacy, Saarland University, Saarbrücken, Germany (2) Dr. Margarete Fischer-Bosch-Institut für Klinische Pharmakologie, Stuttgart, Germany, (3) Department of Clinical Pharmacology, University Hospital Tübingen, Tübingen, Germany, (4) Department of Pharmacy and Biochemistry, University Tübingen, Tübingen, Germany
Introduction: The intravenously administered cytotoxic drug 5-fluorouracil (5-FU) and its oral prodrug capecitabine (CCB) are used as first line agents for the treatment of colorectal cancer, one of the most common tumor types today [1]. Dihydopyrimidine dehydrogenase (DPD) is the rate-limiting enzyme in the transformation of 5-FU, CCB and further fluoropyrimidine based drugs [3]. Polymorphisms in DPYD, the gene encoding for DPD, lead to a genotype dependent enzyme activity which can increase the occurrence of 5-FU and CCB related adverse drug effects leading to life-threatening toxicities [4]. The general use of fluoropyrimidine analogues in the anticancer therapy is broad and thus, it is important to understand the impact of drug-gene and drug-food interactions, to optimize dosing recommendations and consequently limit the occurrence of adverse drug effects.
Objectives:
- To build a physiologically-based pharmacokinetic (PBPK) model for CCB, 5-FU and their respective metabolites 5’-desoxy-5-fluorocytidine (DFCR), 5’-desoxy-5-fluorouridine (DFUR), 5,6-dihydrofluorouracil (DHFU), α-fluoro-β-ureidopropionicacid (FUPA) and α-fluoro-β-alanine (FBAL)
- To predict drug-food-interactions (DFIs) for CCB
- To predict drug-gene-interactions (DGIs) of all compounds for the clinically relevant DPYD gene variants c.1905+ 1G>A (DPYD*2A), c.1679T>G, c.2846A>T and c.1129-5923C>G
- To develop dose recommendations for various DGI combinations
Methods: PBPK model development was performed with PK-Sim® and MoBi® (version 7.4.0) as part of the Open Systems Pharmacology Suite [6]. Data for model development were extracted from literature, including physicochemical parameters and plasma concentration-time profiles for all compounds and for various DPYD genotypes. Data were separated in an internal and an external dataset for model development and evaluation, respectively. Additionally, an uracil (U) model was developed and used for the evaluation of the genotype implementation, where the enzyme activity of genetic variants of DPYD are described through the ratio of U and its first metabolite dihydrouracil (DHU). The final models were used for dose optimization. For this purpose, exposure of 300 mg/m2 5-FU as the area under the plasma concentration-time curve (AUC) at steady-state was simulated as reference value. Exposure were simulated for different DPYD genotypes at steady-state adapting the dose stepwise until matching exposure compared to wildtype was reached.
Results: Nine whole-body PBPK models for CCB, 5-FU and their five metabolites, as well as U and DHU were developed. The compiled data consists of 12 CCB studies (dose range 1260-2372 mg as Xeloda® tablet in 183 patients), 21 5-FU studies (250-1134 mg as bolus injection, continuous infusion or peroral solution in 226 patients) and two peroral U studies (as peroral solutions in 42 patients). The models precisely predict the PK of CCB and 5-FU in wildtype patients and heterozygous patients for the gene variant *2A for fed and fasted patients. Predicted and observed AUC ratio of fed versus fasted was 1.23 for CCB. Additionally, mean predicted to observed AUC values for fed and wildtype individuals show ratios of 1.0 and 1.1 (range 0.73 – 1.87) for CCB and 5-FU, those of the remaining metabolites as well as the DHU/U ratios lie within the two-fold acceptance limits. Dose recommendations were derived for combinations of hetero- and homozygous variants of the four relevant polymorphisms as well as under fasted or fed conditions. For example, model predicts a dose reduction of 35% for DPYD*2A heterozygous variants compared to 50% which is recommended by the current CPIC guideline [5].
Conclusion: A comprehensive set of PBPK models for CCB, 5-FU and their respective metabolites were successfully developed. The models capture the important impact of drug-food and drug-gene interactions and can play an important role in decreasing the occurrence of potential life threatening adverse drug effects by deriving alternative dosing regimens for DPYD deficient patients.
Funding: Supported by the Robert Bosch Stiftung (Stuttgart, Germany) and the European Commission Horizon 2020 UPGx grant 668353
References:
[1] F. Casale et al. Plasma concentrations of 5-fluorouracil and its metabolites in colon cancer patients. Pharmacol Res. (2004) 50(2): 173-9
[2] B. Reigner et al. Effect of food on the pharmacokinetics of capecitabine and its metabolites following oral administration in cancer patients. Clin Cancer Res. (1998) 4(4): 941-8
[3] J. Sistonen et al. Predicting 5-fluorouracil toxicity: DPD genotype and 5,6-dihydrouracil:uracil ratio. Pharmacogenomics (2014) 15(13): 1653–66
[4] E. Hishinuma et al. Functional characterization of 21 allelic variants of dihydropyrimidinase Biochem. Pharmacol. (2017) 143: 118–128
[5] J. A. Johnson et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) Guideline for Dihydropyrimidine Dehydrogenase Genotype and Fluoropyrimidine Dosing: 2017 Update Clin. Pharmacol. Ther (2017) 102(3): 397–404
[6] Eissing T et al. A computational systems biology software platform for multiscale modeling and simulation: integrating whole-body physiology, disease biology, and molecular reaction networks. Front Physiol (2011) 2: 4.
Reference: PAGE 28 (2019) Abstr 8927 [www.page-meeting.org/?abstract=8927]
Poster: Drug/Disease Modelling - Absorption & PBPK