Joachim Almquist, Jacob Leander, Waqas Sadiq, Tove Hegelund Myrbäck, Susanne Prothon, Ulf Eriksson
Quantitative Clinical Pharmacology, Early Clinical Development, IMED Biotech Unit, AstraZeneca, Gothenburg, Sweden
Introduction:
AZD9567 is an oral selective glucocorticoid receptor modulator (oSGRM) aimed for the treatment of inflammatory disease [1]. It has been designed to deliver similar efficacy to oral steroids such as prednisolone, but to differentiate on safety.
Here, we present a model-based approach to estimate equipotent doses of AZD9567 and prednisolone in terms of equally large ex vivo inhibition of the anti-inflammatory biomarker TNFα. The resulting equipotency relationship makes it possible to translate doses of one compound into the other based on their anti-inflammatory effect.
Objectives:
The objective was to estimate an equipotent dose relationship between AZD9567 and prednisolone with respect to ex vivo TNFα inhibition. It is particularly interesting to estimate the dose of AZD9567 equipotent to the commonly used 20 mg prednisolone. Key intermediate steps towards this goal are to develop two popPK models, one for each compound, and to develop a joint TNFα PD model for both compounds.
Methods:
Models were based on clinical data from a SAD and a MAD study, which include measurements of plasma concentration of the respective compound and concentrations of the cytokine TNFα as released in response to an ex vivo whole blood lipopolysaccharide (LPS) stimulation [2].
Nonlinear mixed effect PKPD modeling was done sequentially, fixing both fixed and random effects of the PK models before fitting the PD model. The PD model contained both common and compound-specific parameters and was fitted for both compounds simultaneously. All models were estimated with NONMEM 7.3.0 using the FOCE method [3].
Dose-response curves for each compound were determined by simulation. The response was defined as the average inhibition of TNFα over 24 h following 5 consecutive daily doses. Confidence intervals for the equipotent dose relationship were determined by simultaneously accounting for the uncertainty in the dose-response of each compound.
Results:
A two-compartment model with first order absorption and a lag time was used to describe the PK of AZD9567. The dominant phase was associated with a half-life of 4.4 h. Since the PK was not dose proportional, the oral dose was used in a linear covariate model for the relative bioavailability.
Due to nonlinear plasma protein binding of prednisolone, a static model was first set up to convert total concentrations into unbound concentrations. This model showed that the unbound fraction goes from 6% in the limit of low concentrations to 38% in the high limit. Subsequently, a two-compartment model with a transit compartment approximation [4] was used to describe the PK of unbound prednisolone. The dominant phase was associated with a half-life of 1.9 h.
TNFα inhibition was modeled by an inhibitory Imax-model with sigmoidicity parameter, but in which the driving concentration was a weighted sum of the plasma concentration and the concentration in an effect compartment. Since the weighting parameter was estimated from data, this model had the ability to revert into a direct response model or into an effect compartment model, or to adopt an intermediate setting. The estimate of the weighting parameter (the plasma proportion) was 22%, thus favoring the effect compartment as the main PD driver. The estimate of IC50 for total AZD9567 was 765 nM (95% CI of 610 – 920 nM). Given an unbound fraction 0.637%, this corresponds to a IC50 of 4.87 nM for unbound AZD9567. The estimate of IC50 for unbound prednisolone was 17.0 nM (95% CI of 13.5 – 20.4 nM). Considering unbound concentrations, AZD9567 is a more potent inhibitor of TNFα release than prednisolone.
Simulated doses-response curves for AZD9567 and prednisolone were compared to identify equal levels of TNFα inhibition. For example, it was estimated that a dose of 40 mg AZD9567 (95% CI 29 – 54 mg) was equipotent to 20 mg prednisolone, with both doses resulting in an average TNFα inhibition of 43%. When performed over the whole range of studied doses, such comparisons defined the full equipotency relationship. The full relationship is nonlinear and there is no simple formula such as a factor that relates the equipotent doses.
Conclusion:
An equipotency relationship between AZD9567 and prednisolone was estimated using PKPD modeling of the anti-inflammatory biomarker TNFα. Not only is this relationship a critical component of a dose selection strategy with respect to efficacy, but it also provides a common reference system that will facilitate future assessment of safety biomarkers.
References:
[1] Ripa et al. Discovery of a novel oral glucocorticoid receptor modulator (AZD9567) with improved side effect profile. J Med Chem 2018; 61(5):1785-1799
[2] Schmolz M and Eisinger D. TruCulture®: A simple whole blood collection and culture system for quantifying physiological interactions of the human immune system in the clinic. https://myriadrbm.com/scientific-media/truculture-white-paper/[3] Beal S and Sheiner L, “NONMEM User’s Guides. (1989-2009).” Icon Development Solutions, Elliot City, MD, USA, 2009.
[4] Savic R, Jonker D, Kerbusch T, Karlsson M. Implementation of a transit compartment model for describing drug absorption in pharmacokinetic studies. J Pharmacokinet Pharmacodyn 2007; 34:711-726
Reference: PAGE 28 (2019) Abstr 9043 [www.page-meeting.org/?abstract=9043]
Poster: Drug/Disease Modelling - Other Topics