Combined PET and microdialysis for in vivo estimation of drug blood-brain barrier transport and brain unbound concentrations

Highlights

• Oxycodone was studied in rat brain by simultaneous microdialysis and PET imaging.

• Total concentrations measured by PET were converted to unbound drug concentrations.

• Active brain uptake of oxycodone was verified with both PET and microdialysis.

• PET has the potential to accurately determine blood-brain barrier drug transport.

• Neuro PET enables animal to human translation of effective drug concentrations.

Abstract

Methods to investigate blood-brain barrier transport and pharmacologically active drug concentrations in the human brain are limited and data translation between species is challenging. Hence, there is a need to further develop the read-out of techniques like positron emission tomography (PET) for studying neuropharmacokinetics. PET has a high translational applicability from rodents to man and measures total drug concentrations in vivo. The aim of the present study was to investigate the possibility of translating total drug concentrations, acquired through PET, to unbound concentrations, resembling those measured in the interstitial fluid by microdialysis sampling.

Simultaneous PET scanning and brain microdialysis sampling were performed in rats throughout a 60 min infusion of [N-methyl-¹¹C]oxycodone in combination with a therapeutic dose of oxycodone and during a 60 min follow up period after the end of infusion. The oxycodone concentrations acquired with PET were converted into unbound concentrations by compensating for brain tissue binding and brain intracellular distribution, using the unbound volume of distribution in brain (V_u,brain), and were compared to microdialysis measurements of unbound concentrations.

A good congruence between the methods was observed throughout the infusion. However, an accumulating divergence in the acquired PET and microdialysis data was apparent and became more pronounced during the elimination phase, most likely due to the passage of radioactive metabolites into the brain. In conclusion, the study showed that PET can be used to translate non-invasively measured total drug concentrations into unbound concentrations as long as the contribution of radiolabelled metabolites is minor or can be compensated for.

Introduction

The pharmaceutical industry is facing difficulties within their central nervous system (CNS) related drug development programs, with high attrition rates late in clinical development (Arrowsmith and Miller, 2013). A major hindrance in CNS drug development is the exceptional ability of the blood-brain barrier (BBB) to restrict drug molecules from entering the CNS. The BBB is a highly dynamic barrier responsible for maintaining the brain microenvironment through its tight junctions, low pinocytotic activity, efficient transporters, and its metabolic properties. As a consequence of the complexity of brain drug delivery as well as the sensitivity of the brain as an organ, neuropharmacokinetics (neuroPK) of drugs is both difficult to investigate and to accurately translate between animals and humans. While some of this is due to anatomical and physiological species differences, it primarily owes to the limitations of the techniques currently used to investigate CNS drug concentrations.

Total drug concentration in brain and its relation to total plasma drug concentration has long been investigated and used to generate a ratio usually designated K_p,brain or logBB, to estimate CNS drug exposure and brain drug delivery across the BBB for further predictions of drug effect. However, given the free drug hypothesis, it is generally accepted that it is the unbound drug that interacts with therapeutic targets and hence it is the unbound concentrations that drive the effect and should be used for drug-effect relationship estimates (Kalvass et al., 2007b, Liu et al., 2009, Watson et al., 2009). Accordingly, the unbound concentrations of drugs in brain and plasma should be used for exposure assessment. The ratio resulting from these measures, K_p,uu,brain, gives a more accurate description than K_p,brain of net transport at the BBB, as the latter is dependent on drug binding to constituents in brain and plasma in addition to BBB transport mechanisms (Gupta et al., 2006, Summerfield et al., 2006, Kalvass et al., 2007a, Hammarlund-Udenaes et al., 2008).

While being applicable in both animals and humans, positron emission tomography (PET) represents a technique with high translatable potential for the use in pharmacokinetic (PK) and pharmacodynamic (PD) studies. Nevertheless, the technique suffers from the limitation that only total drug concentrations can be measured through biodistribution studies. Thus, the total brain to plasma partition coefficient, denoted volume of distribution (V_T) within PET terminology (equal to the K_p,brain terminology in PK), is used for the assessment of tissue distribution and brain exposure. Microdialysis is currently the only method for the direct measurement of unbound drug concentrations in brain. However, it has a highly restricted applicability in humans and is also limited by technical demands and surgical expertise. As a consequence, in vitro and preclinical techniques have been developed and optimized for the combined use in the screening of CNS compounds to assess unbound drug concentrations in brain and to estimate BBB transport in animals (Kalvass and Maurer, 2002, Friden et al., 2007, Friden et al., 2009, Loryan et al., 2014). Modeling and simulations have also become important tools for the interpretation and prediction of PK and PD data. Even so, development of new experimental designs and optimization of current methods are required for the translation of neuroPK studies from rodents to humans.

The concept of the unbound volume of distribution in brain (V_u,brain) was introduced by Wang and Welty in 1996 (Wang and Welty, 1996). It relates unbound drug concentrations in the interstitial fluid to the total amount of drug in brain by describing the distribution of drug inside the brain. Hence, this parameter has been used as a complement to PK experiments examining total brain concentrations in animals (Friden et al., 2007). In this sense, V_u,brain can be applied to compensate total drug concentrations for tissue binding as well as for intracellular distribution, resulting in an estimation of unbound brain concentrations. This supports the notion that V_u,brain can be used for conversion of total drug concentrations acquired through PET to unbound drug concentrations. This in turn, would enable preclinical to clinical translation of neuroPK properties and possible estimations of unbound brain drug concentrations and BBB transport in humans.

The present study aimed to elucidate the ability of using PET imaging, in combination with PK theory, to estimate brain unbound drug concentrations and net transport across the BBB, with verification of data by microdialysis. With the potential of being extrapolated from animals to humans, the study also aims to aid in the development of translational experimental designs for enhanced interpretation of brain PK in humans. The opioid agonist oxycodone was chosen as a model compound for the use in this study as its neuroPK properties have been investigated both in vivo and in silico in rats (Bostrom et al., 2005, Bostrom et al., 2006, Bostrom et al., 2008, Okura et al., 2008, Ball et al., 2012). In addition, oxycodone represents an interesting compound given its active uptake at the BBB (Bostrom et al., 2006).