Simeon Rüdesheim (1,2), Jan-Georg Wojtyniak (1,2,3), Roman Tremmel (4), Matthias Schwab (2,4,5) and Thorsten Lehr (1)
(1) Clinical Pharmacy, Saarland University, Saarbrücken, Germany, (2) Dr. Margarete Fischer-Bosch-Institut fuer Klinische Pharmakologie, Stuttgart, Germany, (3) Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim, Germany, (4) Department of Clinical Pharmacology, University Hospital Tübingen, Tübingen, Germany, (5) Department of Pharmacy and Biochemistry, University Tübingen, Tübingen, Germany
Introduction: Although still relevant, tricyclic antidepressants (TCAs) are not considered a first-line therapy option in the treatment of depressions due to severe adverse drug effects (ADEs) and resulting safety concerns [1]. Among TCAs and other clinically used antidepressants, desipramine stands out with a relatively high relative risk for seizures, arrhythmias, hypotension and – most notably – an increased risk of death due to overdose [2]. Moreover, as desipramine is mainly metabolized through CYP2D6, its pharmacokinetics are highly susceptible to drug-gene interactions (DGIs) and drug-drug interactions (DDIs). Consequently, physiologically based pharmacokinetic (PBPK) modeling can deliver valuable insights into the impact and nature of CYP2D6 DGIs and DDIs and contribute towards model-based precision dosing.
Objectives:
- Development of PBPK models for CYP2D6 probe substrate desipramine, its metabolite 2-hydroxydesipramine as well as CYP2D6 inhibitor quinidine and its metabolite 3-hydroxyquinidine
- Prediction of desipramine CYP2D6 DGI effects and DDI with quinidine
- Development of desipramine dose recommendations for various DDI/DGI scenarios
Methods: PBPK model development was performed with PK-Sim® (version 8.0) as part of the Open Systems Pharmacology Suite [3]. Physicochemical parameters were extracted from literature as well as plasma concentration-time profiles, which were subsequently separated in internal and external datasets for model development and evaluation, respectively. Model performance was assessed via calculation of geometric mean fold errors (GMFE) of predicted versus observed area under the plasma concentration-time curve (AUC) ratios as well as mean relative deviation (MRD) of predicted versus observed plasma concentration-time values. Individually developed desipramine and quinidine models were linked to predict the DDI effects of concomitant use. Finally, the models were used for dose optimizations. For this purpose, AUC of 50 mg desipramine single dose (s.d.) was used as reference exposure. Subsequently, simulated exposures for different CYP2D6 genotypes with and without concomitant administration of 100 mg quinidine once daily (q.d.) over a wide dose-range (25-250 mg desipramine) were calculated and compared to the reference exposure.
Results:
The final model was able to describe all plasma concentration-time profiles satisfactorily with an overall GMFE of AUC of 1.39 for desipramine (35/36 values ≤ 2) and 1.29 for quinidine (16/16 ≤ 2) as well as MRD of plasma concentration-time profiles of 1.89 for desipramine (16/30 ≤ 2) and 1.63 for quinidine (10/12 ≤ 2). Based on CYP2D6 activity scores, CYP2D6 poor metabolizer (PM), extensive metabolizer (EM) and ultrarapid metabolizer (UM) plasma concentration-time profiles could be predicted successfully with mean DGI ratios (predicted AUC ratio PM or UM to EM versus observed AUC ratio PM or UM to EM) of 0.82 for PM and 0.99 for UM. Moreover, the coupled models were able to accurately predict the quinidine and desipramine DDI. The DDI ratio (predicted AUC ratio desipramine and quinidine to desipramine and placebo versus observed AUC ratio desipramine and quinidine to desipramine and placebo) was 1.42, well within the acceptance criterion for DDI predictions [4]. The model predicted that a 75% dose reduction in PMs and a 300% dose increase in UMs were necessary to obtain an exposure comparable to the reference exposure. In contrast, the Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline for TCAs recommends a dose reduction of 50% for PMs and a usage of alternative antidepressants not metabolized through CYP2D6 for UMs [5]. Finally, simulations have shown that the desipramine dose should be reduced for all genotypes by 75% if desipramine is given together with quinidine.
Conclusions: Comprehensive PBPK models of desipramine and quinidine were developed and coupled to predict CYP2D6 DDIs and DGIs. Overall, the model holds the potential for a more personalized medicine by reducing the risk of ADEs or therapy failure.
Funding: Supported by the Robert Bosch Stiftung (Stuttgart, Germany) and the European Commission Horizon 2020 UPGx grant 668353
References:
[1] Koenig A M et al. First-line pharmacotherapies for depression – what is the best choice? Pol Arch Med Wewn. 2009 Jul-Aug;119(7-8):478-86.
[2] White N et al. Suicidal antidepressant overdoses: a comparative analysis by antidepressant type. J Med Toxicol. 2008 Dec;4(4):238-50.
[3] 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.
[4] Guest E J et al. Critique of the two-fold measure of prediction success for ratios: application for the assessment of drug-drug interactions. Drug Metab Dispos. 2011 Feb;39(2):170-3.
[5] Hicks J K et al. Clinical pharmacogenetics implementation consortium guideline (CPIC) for CYP2D6 and CYP2C19 genotypes and dosing of tricyclic antidepressants: 2016 update. Clin Pharmacol Ther. 2017 Jul;102(1):37-44.
Reference: PAGE () Abstr 9508 [www.page-meeting.org/?abstract=9508]
Poster: Drug/Disease Modelling - Absorption & PBPK