Xiao Zhu (1), David B Finlay (2), Jamie J Manning (2), Michelle Glass (2), Stephen B Duffull (1)
(1) Otago Pharmacometrics Group, School of Pharmacy, University of Otago, Dunedin, New Zealand; (2) Department of Pharmacology and Toxicology, University of Otago, Dunedin, New Zealand
Objectives: Positive allosteric modulator (PAM)-antagonists exhibit paradoxical pharmacological profiles by simultaneously increasing agonist affinity whilst decreasing the efficacy of the agonist they modulate [1]. Due to their unique characteristics, PAM-antagonists are of great therapeutic interest for G-protein coupled receptors (GPCR) drug discovery.
The current trend in quantitative systems pharmacology (QSP) sees the development of more mechanism-based models, which are inevitably more sophisticated but are still assembled from the building blocks provided by receptor theory [2-3]. PAM-antagonists represent a challenge to standard receptor theory and QSP model development. PAM-antagonist can produce complicated profiles that include time dependent non-monotonicity in receptor effect which varies at different PAM-antagonist and ligand concentrations. Current approaches typically employ the operational model of allosteric modulation (i.e. an equilibrium model) to describe the responses of PAM-antagonists from snapshots of bioassays [4]. This framework overlooks the time dependent modulation by PAM-antagonists and hence does not work for non-equilibrium conditions [5]. In some PK-PD analyses [6-7], empirical kinetic solutions have been shown to adequately describe the data but do not allow mechanistic insights as every response curve arising from a combination of PAM and ligand must be fitted separately. There is a need for an overarching mechanistic model that can be used to provide insights into the underlying biological mechanisms that arise from a diverse set of observed phenomena.
The overarching aim of this work was to develop a mathematical model for a PAM-antagonist as a building block for the future development of QSP models. ORG27569, a well-characterised PAM-antagonist of the cannabinoid-1 (CB1) receptor [8-9], was used as an illustrative example. This work encompasses three specific objectives:
(1) to generalise the equilibrium model for PAM-antagonist into a kinetic model and evaluate its performance using the observed signature profiles,
(2) to extend the kinetic model by incorporating a hypothesis that ORG27569 could stabilise an intermediate status before full receptor inactivation and evaluate its performance,
(3) to test the extended PAM-antagonist model with a newly designed experiment.
Methods: A kinetic operational model of allosteric modulation was constructed for ORG27569 (i.e. PAM-antagonist). Subsequently, the performance of this kinetic model was assessed using the signature profiles reported in the literature [8-9]: i) enhanced receptor binding, ii) reduced internalisation and iii) time-dependent modulation of cAMP signalling. Later, the kinetic model was extended to incorporate the “intermediate status” hypothesis (where ORG27569 could stabilise an intermediate status on the pathway to full receptor inactivation). The performance of the extended PAM-antagonist model was evaluated. In addition, a newly designed cAMP signalling assay was conducted to verify the non-intuitive signature profiles simulated from the extended model.
Results: When ignoring the dependence among different signature profiles, the kinetic PAM-antagonist model could describe the observations from each signalling pathway on their own. However, it was noted that the onset rate of ORG27569 to ligand-receptor complex tended to be 100-fold different among different assays (0.007 µM-1·min-1 for receptor binding and internalisation, 0.7 µM-1·min-1 for cAMP). Hence, the kinetic PAM-antagonist model failed to simultaneously describe all the signature profiles from ORG27569 and was incapable of predicting the outcomes of new experiments. By incorporating the “intermediate status” hypothesis, the extended PAM-antagonist model could reproduce all of the signature profiles from the allosteric modulation of CB1 receptor. Furthermore, a newly designed cAMP signalling assay managed to reproduce the corresponding simulation results from the extended model. All of these demonstrated that the extended PAM-antagonist model could dynamically represent the unique modulation behaviours of ORG27569.
Conclusions: We have developed and extensively evaluated an extended kinetic model for PAM-antagonist with ORG27569 as a motivating example. The developed PAM-antagonist model has generality beyond the CB1 system and could be used as a building block for PAM-antagonist within QSP models in the future.
References:
[1] Kenakin T, Strachan RT. Trends Pharmacol Sci. 2018; 39(8): 748-765.
[2] Wajima T, et al. Clin Pharmacol Ther. 2009; 86(3): 290-298.
[3] Peterson MC et al. CPT Pharmacometrics Syst Pharmacol. 2012; 1(11): 1-8.
[4] Price MR, et al. Mol Pharmacol. 2005; 68(5): 1484-1495.
[5] Zhu X et al. Br J Pharmacol. 2019; 176(14): 2593-2607.
[6] Bursi R et al. Psychopharmacology. 2011; 218(4): 713-724.
[7] Eleveld D et al. Anesthesiology. 2016; [A5009]
[8] Baillie GL, et al. Mol Pharmacol. 2013; 83(2): 322-338.
[9] Cawston EE, et al. Br J Pharmacol. 2013; 170(4): 893-907.
Reference: PAGE () Abstr 9344 [www.page-meeting.org/?abstract=9344]
Poster: Oral: Methodology - New Modelling Approaches