|In-depth neuropharmacokinetic analysis of antipsychotics based on a novel approach to estimate unbound target-site concentration in CNS regions: Link to spatial receptor occupancy
Irena Loryan1*, MD, PhD, Erik Melander1, MSc, Marielle Svensson1, MSc, Maryam Payan1,2, PhD, Frida König1, MSc, Britt Jansson1, BSc, and Margareta Hammarlund-Udenaes1, PhD
1 – Translational PKPD Group, Department of Pharmaceutical Biosciences, Associate member of SciLifeLab, Uppsala University, Sweden
2 – Current affiliation: Food and Drug Control Laboratories Research Center (FDLRC); Ministry of Health and Medical Education (MOH); Tehran, Iran
*Dr. Irena Loryan
Translational PKPD group, Department of Pharmaceutical Biosciences
Box 591, 751 24 Uppsala, Sweden
+46 18 471 4995; +46 73 8914982 (mobile)
Neuropharmacokinetic analysis of antipsychotics
Materials and Methods
Haloperidol, clozapine, paliperidone, risperidone, olanzapine, quetiapine, 2-hydroxypropyl-β-cyclodextrin (HPβCD), 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), human serum albumin, sodium deoxycholate and sodium dodecyl sulfate (SDS) were obtained from Sigma-Aldrich (Stockholm, Sweden). Acetonitrile, ammonium formate, trichloroacetic acid and formic acid were purchased from Merck (Darmstadt, Germany). PierceTM BCA protein assay kit and rat serum albumin ELISA kit were purchased from Thermo Scientific (MA, USA) and GenWay Biotech Inc. (San Diego, CA, USA), respectively. Water used in all experiments was purified using a Milli-Q Academic system (Millipore, Bedford, MA, USA).
All experiments were performed on drug-naïve male Sprague-Dawley rats (Taconic, Lille Skensved, Denmark) in accordance with guidelines from Swedish National Board for Laboratory Animals, and were approved by the Animal Ethics Committee of Uppsala, Sweden. In total, 40 drug-naïve rats weighing 282±34 g were included in in vivo neuroPK studies (ethical approval C16/12). In vitro brain tissue binding studies and brain tissue uptake studies were conducted on 16 and 18 rats, respectively, weighing 250-300 g (ethical approval C351/11). All rats were housed in groups at 18 to 22°C under a 12-h light/dark cycle with ad libitum access to food and water.
The CMA-ROI toolbox
The methodological platform of CMA-ROI includes i) in vivo neuroPK studies performed in rodents followed by measurement of brain discrete region/spinal cord partitioning coefficient (Kp,ROI); ii) in vitro drug tissue binding studies for determination of fraction of unbound drug in CNS regions (fu,ROI) and blood (fu,plasma); iii) in vitro brain slice studies for evaluation of unbound volume of distribution in discrete brain regions (Vu,ROI). The extent of unbound drug transport across the BBB/BSCB in the various CNS regions (Kp,uu,ROI) is thereafter assessed through a combination of the experimentally determined Kp,ROI, fu,ROI and fu,plasma using the following equation:
where Kp,ROI is the ratio of total CNS regional to total plasma drug concentrations measured at steady state conditions.
In addition, the extent of unbound drug transport into the neural cells (Kp,uu,cell,ROI) in various brain regions, below exemplified by cortex and striatum, was estimated as:
Below CMA-ROI key methodological platforms are discussed in details.
In vivo neuropharmacokinetic study
The CNS exposure of antipsychotic drugs was assessed in vivo at steady state conditions. Animals were not randomized and study was not blinded. The femoral vein and artery were surgically catheterized 24 h before the experiment. Rats were thereafter individually placed into the CMA120 system (CMA, Solna, Sweden) for freely moving animals with ad libitum access to food and water. The dosing solutions of antipsychotic drugs were prepared individually the day before experiment by diluting the respective stock solution in a vehicle containing 10% HPβCD in 0.9% saline for parenteral use (Uppsala University Hospital Pharmacy, Uppsala, Sweden). Rats were initially given an intravenous loading dose of haloperidol, clozapine, paliperidone, risperidone, olanzapine or quetiapine during five minutes using a fast-rate infusion followed by a four hour constant-rate infusion with a flow rate of 1 mL/kg · h-1 (Supplementary Table 1). The dosing regimen was chosen based on a simulation of the plasma profile of the drug, targeting a total plasma concentration of 100 ng/mL, using the Berkeley Madonna software package (version 8.3.18 for Windows, Berkeley, CA).
/Supplementary Table 1/
Blood (300 µL) was sampled from arteria femoralis and collected into heparin containing tubes at 0, 150, 210 and 240 min after the start of the infusion, to confirm attainment of plasma steady state conditions. At the end of the experiment, the rats were anaesthetized by inhalation of 5% isoflurane (Abbot Scandinavia, Solna, Sweden). CSF was collected from the cisterna magna using BD insulin syringe (BD Biosciences, Plymouth, UK). Exclusion criterion for CSF samples contaminated with the blood was pre-established. To reduce the residual cerebral blood volume, up to 10 ml of blood was taken intracardially using a 10 ml EDTA containing vacutainer (BD Biosciences, Plymouth, UK) (1). After decapitation, the brain was removed and the structures of interest, hypothalamus, frontal cortex, cerebellum, striatum, hippocampus, and brainstem, were dissected according to Glowinski and Iversen (2) and Chiu et al. (3). The spinal cord was removed from the column using a ventral approach. Blood samples were centrifuged for 5 minutes at 1000 rpm and plasma was transferred to Eppendorf tubes. Plasma, brain regions/spinal cord and CSF samples were immediately frozen on dry ice and thereafter stored at -80°C pending bioanalysis. Prior to analysis the brain structures/spinal cords were individually homogenized on ice in 1:4 (w:v) phosphate buffered saline (PBS) pH 7.4 using ultrasonic processor VCX-130 (Sonics, Chemical Instruments AB, Sweden).
Validation using cerebral microdialysis technique would be very valuable for the current study and was planned for clozapine and risperidone. For both drugs cerebral microdialysis data were previously reported. However, an extensive nonspecific sticking of compounds to the plastic tubing (fluorinated ethylene propylene) was observed for both drugs. The use of 0.5 % albumin in perfusate significantly prevented the sticking to the tubing for risperidone; however albumin was affecting the “true” in vivo recovery estimates. Therefore, the validation was not completed in our lab. In fact, often in published literature experimental data on cerebral microdialysis presented without thorough assessment of sticking of substance to the plastic tubing and probe, in vitro and in vivo recovery. Consequently, conclusions made based on this kind of “quick and dirty” experiments may be misleading.
In vitro drug tissue binding study
Equilibrium dialysis was used to assess the fraction of unbound antipsychotic drugs in plasma (fu,plasma), in whole brain homogenate (fu,brain) and in discrete brain regions and spinal cord (fu,ROI). A Teflon 96-well plate (model HTD96b, HTDialysis LLC, Gales Ferry, CT, USA) fitted with a regenerated cellulose membrane (molecular weight cutoff 12-14 kDa) was used in all tissue binding experiments. The cellulose membrane was treated according to the recommendations from manufacturer before starting the experiment (HTDialysis LLC, Gales Ferry, CT, USA).
For the plasma protein binding measurements, frozen blank plasma was thawed and pH was adjusted to 7.4 using 1 M phosphoric acid or 1 M sodium hydroxide. Plasma (n=4), at volumes of 125 µl, were individually spiked to a final drug concentration of 200 nM. The plasma samples were dialyzed against equal volumes of PBS, pH 7.4, for 6 hours at 37°C at 200 rpm in an incubator with orbital shaking (MaxQ4450 Thermo Fisher Scientific, NinoLab, Sweden). At the end of the experiment, 80 µl was sampled from the spiked plasma (donor side) and from the buffer (receiver side), and subsequently mixed with 80 µl of either PBS pH 7.4 or with undiluted plasma, respectively. Samples were stored at -20oC pending bioanalysis. The fraction of unbound drug in plasma (fu,plasma) was assessed using Equation 3:
For the brain tissue binding assay, dilution factors 5 and 10 were initially tested using whole brain tissue homogenate. The frozen blank whole brain tissue was homogenized with an ultrasonic processor VCX-130 (Sonics, Chemical Instruments AB, Sweden) in PBS using 1:4 (w:v) or 1:9 (w:v) dilution. Whole brain tissue homogenate was treated similarly as plasma. No significant differences were found in fraction of unbound drug in whole brain homogenate diluted 5-fold or 10-fold (data not shown); moreover, the results were comparable to literature fu,brain values. Therefore, the decision was made to use a 10-fold diluted homogenate in the regional tissue binding study. The brain regions and spinal cord were isolated as described above. Tissues from rats were pooled (2 different pools, n=8 biological replicates in each pool) to study the fraction of unbound drug in brain regions and spinal cord with 4 technical replicates in each experiment, except the experiments with homogenates of hypothalamus and frontal cortex that were studied in duplicates and triplicates, respectively). Whole brain tissue homogenate was used in each experiment as a reference. Eighty µl was taken from both sides of the membrane at 6 h, assuming equilibrium. To minimize the occurrence of potential matrix effect interference during the bioanalysis the samples from receiver side were mixed with 80 µl of 1:9 (w:v) homogenate of the respective brain region or spinal cord. Due to the limited volume of homogenates of hypothalamus and frontal cortex those samples were mixed with 1:9 (w:v) whole brain homogenate. The brain/spinal cord tissue samples were mixed with 80 µl of PBS. Drug recovery and stability in the plasma and brain/spinal cord homogenates were tested in each experiment. Samples were stored at -20oC pending bioanalysis. Olanzapine showed low stability during the 6 hour incubation of the brain and spinal cord homogenates. Equation 4 was used to assess the fraction of unbound drug in brain tissue, according to
Where fu,ROI,D is the fraction of unbound drug in diluted homogenate of brain regions or spinal cord. D is the dilution factor (in these experiments, D=10). Equation 5 was used to correct fu,ROI,D for the dilution (4):
In vitro brain slice uptake study
The volume of distribution of unbound antipsychotic drug in brain (Vu,brain, mL · g brain-1) was estimated using the brain slice method according to previously published protocols (5, 6). In order to measure Vu,brain in cortex (Vu,cortex) and striatum (Vu,striatum), microdissection of cortex and striatum was performed from the brain slices at the end of the incubation period (Figure 2).
/Manuscript Figure 2/
Briefly, six 300 µm brain slices were cut using a vibrating blade microtome Leica VT1200 (Leica Microsystems AB, Sweden), and incubated in a HEPES-buffered artificial extracellular fluid (aECF) containing the antipsychotic drug with an initial concentration of 200 nM. Two approaches were used. First, the cassette (mixture of all drugs in aECF) incubation approach was used to study the Vu,brain in the entire brain slice (n=3 brains per cassette, i.e., 18 slices). Second, the single drug incubation approach was used to study the brain tissue uptake in cortex and striatum (n=2 rat brains per drug, i.e., 12 slices). Incubation was performed during 5 h at 37°C in a shaker (MaxQ4450 Thermo Fisher Scientific, NinoLab, Sweden) with a rotation speed of 45 rpm and constant oxygen flow of about 75-80 mL per minute through a glass frit. Buffer and brain slices were sampled by the end of the incubation. The slices were weighed and homogenized in 9 volumes (w:v) of aECF using ultrasonic processor VCX-130 (Sonics, Chemical Instruments AB, Sweden). The viability of the brain slices was assessed by using one of the slices per incubation to measure the activity of released lactate dehydrogenase using a cytotoxicity detection kit (Roche Diagnostics GmbH, Germany). The other slices were analyzed for their drug content. The brain slices were microdissected, cortex and striatum were isolated, dried on the filter paper, individually weighted and homogenized. In each experiment one brain slice was not dissected and was used for estimation of Vu,brain in the entire brain slice as a reference control.
Assuming that the concentration of the drug in virtually protein free aECF is equal to the interstitial fluid concentration in the brain slice at equilibrium, Vu,ROI was estimated as the ratio of the amount of drug in the entire brain slice or in cortex/striatum (AROI, nmole · g brain-1) to the measured final aECF concentration (Cbuffer, µmole · mL-1) using Equation 4. The brain tissue density was assumed to be 1 g · mL-1.
where Vi (mL · g brain-1) is the volume of the surrounding brain slice layer of aECF. A volume of 0.094 mL ∙g brain-1 was obtained using [14C]inulin (5).