A comparison of two semi-mechanistic models for prolactin release and prediction of receptor occupancy following administration of dopamine D₂ receptor antagonists in rats

Abstract

We compared the model performance of two semi-mechanistic pharmacokinetic-pharmacodynamic models, the precursor pool model and the agonist-antagonist interaction model, to describe prolactin response following the administration of the dopamine D₂ receptor antagonists risperidone, paliperidone or remoxipride in rats. The time course of pituitary dopamine D₂ receptor occupancy was also predicted.

Male Wistar rats received a single dose (risperidone, paliperidone, remoxipride) or two consecutive doses (remoxipride). Population modeling was applied to fit the pool and interaction models to the prolactin data. The pool model was modified to predict the time course of pituitary D₂ receptor occupancy. Unbound plasma concentrations of the D₂ receptor antagonists were considered the drivers of the prolactin response. Both models were used to predict prolactin release following multiple doses of paliperidone.

Both models described the data well and model performance was comparable. Estimated unbound EC₅₀ for risperidone and paliperidone was 35.1 nM (relative standard error 51%) and for remoxipride it was 94.8 nM (31%). KI values for these compounds were 11.1 nM (21%) and 113 nM (27%), respectively. Estimated pituitary D₂ receptor occupancies for risperidone and remoxipride were comparable to literature findings. The interaction model better predicted prolactin profiles following multiple paliperidone doses, while the pool model predicted tolerance better.

The performance of both models in describing the prolactin profiles was comparable. The pool model could additionally describe the time course of pituitary D₂ receptor occupancy. Prolactin response following multiple paliperidone doses was better predicted by the interaction model.

Introduction

Schizophrenia is a psychiatric disorder, characterized by dopamine dysregulation, resulting in a spectrum of positive and negative cognitive symptoms such as delusions, hallucinations, a motivation and social withdrawal among others (Di Forti et al., 2007). It is treated with dopamine D₂ receptor antagonists, also known as antipsychotics. Dopamine D₂ receptor antagonists act by binding to central as well as peripheral D₂ receptors located in the cortical and tuberoinfundibular regions of the brain respectively (Beaulieu and Gainetdinov, 2011, Missale et al., 1998). Treatment by D₂ receptor antagonists leads to hyperprolactinemia, due to loss of tonic inhibitory control of dopamine on the release of prolactin (Di Forti et al., 2007). Hyperprolactinemia leads to commonly reported adverse events such as galactorrhea, gynecomastia, and sexual dysfunction. Kapur et al. (2002) have shown that the potency of antipsychotic drugs correlates strongly with dopamine D₂ antagonism. The hyperprolactinemic effects of antipsychotics are mostly correlated with their affinity for dopamine D₂ receptors at the level of the anterior pituitary lactotrophs (Ben-Jonathan et al., 2008, Peuskens et al., 2014). Taken together, and the fact that prolactin is easily assayed in plasma, prolactin becomes a suitable and attractive biomarker to study the effects of dopamine D₂ antagonism. However, it should be considered that several atypical antipsychotics exhibit a higher peripheral-to-central dopamine D₂ receptor occupancy as a result of active efflux at the blood-brain barrier. Therefore these compounds cause sustained prolactin elevations, whereas antipsychotics that easily cross the blood-brain barrier and exhibit rapid disassociation from the dopamine receptor only cause transient hyperprolactinemia (Fitzgerald and Dinan, 2008).

Mechanism-based modeling enables the translational step from rat to humans and provides insights into therapeutic dose ranges for newer compounds during drug development. Two semi-mechanistic models have been proposed in the literature to model the time course of prolactin following administration of D₂ receptor antagonists. The first is the precursor pool model, which has been used to model the prolactin release in healthy volunteers and rats (Ma et al., 2010, Movin-Osswald and Hammarlund-Udenaes, 1995, Stevens et al., 2012). The second model is the agonist-antagonist interaction model, which has been used to clinical data (Bagli et al., 1999, Friberg et al., 2009, Ma et al., 2010). To our knowledge, so far, no published preclinical study has examined how these two models compare with each other with regard to their ability to describe prolactin release following D₂ receptor antagonist administration in rats.

The first aim of this study was to compare the aforementioned models in describing prolactin release in rats. The second aim was to extend the pool model to estimate pituitary dopamine D₂ receptor occupancy, which is the driver of prolactin release. Thirdly, we aimed to compare the ability of the models to predict prolactin release following multiple doses.