Dominik Lott (1), Pierre-Eric Juif (1), Jasper Dingemanse (1), Andreas Krause (1)
(1) Idorsia Pharmaceuticals Ltd, Department of Clinical Pharmacology, Allschwil, Switzerland
Objectives: Development of a population model describing the pharmacokinetics (PK) and pharmacodynamics (PD), i.e., effect on total lymphocyte count, of the selective sphingosine-1-phosphate 1 (S1P1) receptor modulator cenerimod [1], including effects of demographics and differences between healthy subjects and systemic lupus erythematosus (SLE) patients. Assessment of attainment of steady-state conditions with a terminal half-life of up to 22 days [2]. Prediction of incidence of low total lymphocyte counts, a safety parameter.
Methods: Data from four Phase 1 studies in healthy subjects [2] and one Phase 2 study in SLE patients [3] were pooled for analysis. These included single-dose administration of 1 to 25 mg and multiple-dose administration of 0.5 to 4 mg cenerimod once daily (o.d.). The modelling data set comprised 2393 cenerimod concentration and 1906 total lymphocyte count measurements from 110 (64 healthy, 46 with SLE) and 130 (67 healthy, 63 with SLE) subjects, respectively. PK/PD model development was conducted sequentially using nonlinear-mixed effects modelling techniques. Parameters were estimated using SAEM in Monolix 2018R2 [4]. Age, body weight, body mass index, fat mass, food intake, and effect of disease were investigated in covariate analyses for their impact on model parameters.
Results: The PK of cenerimod were well described by a 3-compartment model with estimates of 121, 196, and 306 L for apparent volumes of distribution of the central compartment and peripheral compartments, respectively, and a low apparent clearance of 1.42 L/h. Body weight was added as a covariate on all volumes of distribution, inter-compartmental flows, and drug clearance with covariate coefficients fixed to allometric coefficients [5]. Presence of food was identified to delay cenerimod absorption by 0.9 h and to slightly decrease the rate of absorption by 15%. There was no evidence for differences in PK between healthy subjects and SLE patients.
Model-based simulations of cenerimod administration o.d. for 365 days showed that 90, 95, and 99% of the exposure after 365 days is reached after 49, 75, and 134 days, respectively, compared to 85% on Day 35. Exposure to cenerimod is predicted to be 20% lower and 37% higher in subjects with a body weight of 100 and 50 kg, respectively, compared to the population-typical subject (75 kg). Food intake had no relevant effect on the exposure to cenerimod.
The model that characterized the relationship between cenerimod concentration and total lymphocyte count best was an indirect-effect Imax model with Imax proportional to the estimated baseline total lymphocyte count. A periodic function was used to describe the variation in total lymphocyte count (circadian rhythm) over the course of a day. Maximum drug effect (Imax) and concentration at half-maximum effect (IC50) were estimated as 92.3% and 25.6 ng/mL, respectively. In covariate analyses, as expected due to disease, SLE patients were identified to have a lower total lymphocyte count at baseline (1.63×109/L) than healthy subjects (1.96×109/L).
Simulations predicted average total lymphocyte count decreases from baseline to Day 35 of 66.9 and 77.5% for o.d. doses of 2 and 4 mg, respectively. The respective decreases on Day 365 were 69.7 and 79.5%, suggesting that the maximum PD effect is essentially achieved after 35 days of treatment.
Total lymphocyte counts below 0.5 and 0.2×109/L were predicted to increase with dose and to be higher with a lower total lymphocyte count at baseline. Following o.d. doses of 2 mg for 365 days, the model predicted that 0.4 and 0.6% of healthy subjects and SLE patients, respectively, would experience total lymphocyte counts below 0.2×109/L at least once with 4 predose measurements at 3, 6, 9, and 12 months.
Conclusions: Relevant demographic differences were limited to body weight. Predicted changes in exposure and effect on total lymphocyte count beyond 35 days of o.d. cenerimod administration are negligible. This indicates that long-term administration for, e.g., 365 days, should not result in drug exposure substantially different from that observed in completed Phase 1 and Phase 2 studies. Due to the drug’s long half-life the in silico model of long-term treatment scenarios that have yet to be clinically tested proved to be useful. The model provides a robust basis to support planning of late-phase studies and answering of questions from health authorities via model-based simulations.
References:
[1] Piali L, Birker-Robaczewska M, Lescop C, et al. Cenerimod, a novel selective S1P1 receptor modulator with unique signaling properties. Pharmacol Res Perspect. 2017;5(6):1-12.
[2] Juif PE, Baldoni D, Reyes M, et al. Pharmacokinetics, pharmacodynamics, tolerability, and food effect of cenerimod, a selective S1P1receptor modulator in healthy subjects. Int J Mol Sci. 2017;18(12).
[3] Clinicaltrials.gov. Clinical Study to Investigate the Biological Activity, Safety, Tolerability, and Pharmacokinetics of ACT-334441 in Subjects With Systemic Lupus Erythematosus. https://clinicaltrials.gov/ct2/show/NCT02472795?cond=ACT-334441&rank=1. Published 2015. Accessed February 4, 2019.
[4] Lixoft-Incuballiance. Monolix User Guide 2018. http://monolix.lixoft.com/single-page/.
[5] Anderson BJ, Holford NHG. Mechanism-Based Concepts of Size and Maturity in Pharmacokinetics. Annu Rev Pharmacol Toxicol. 2008;48(1):303-332.
Reference: PAGE 28 (2019) Abstr 8977 [www.page-meeting.org/?abstract=8977]
Poster: Drug/Disease Modelling - Other Topics