8
Pharmacokinetic-Pharmacodynamic Modeling of Buspirone and Its Metabolite 1-(2-Pyrimidinyl)-piperazine in Rats KLAAS P. ZUIDEVELD, 1 JASNA RUSIC ¸ -PAVLETIC ¸, 2 HUGO J. MAAS, LAMBERTUS A. PELETIER, PIET H. VAN DER GRAAF, 3 AND MEINDERT DANHOF Leiden/Amsterdam Center for Drug Research, Division of Pharmacology, Gorlaeus Laboratories, Leiden, The Netherlands (K.P.Z., J.R.-P., H.J.M., P.H.V., M.D.); and Mathematical Institute, Leiden University, Niels Bohrweg, Leiden, The Netherlands (L.A.P.) Received April 4, 2002; accepted July 3, 2002 ABSTRACT The objective of this investigation was to compare the in vivo potency and intrinsic activity of buspirone and its metabolite 1-(2-pyrimidinyl)-piperazine (1-PP) in rats by pharmacokinetic- pharmacodynamic modeling. Following intravenous adminis- tration of buspirone (5 or 15 mg/kg in 15 min) or 1-PP (10 mg/kg in 15 min), the time course of the concentrations in blood were determined in conjunction with the effect on body temperature. The pharmacokinetics of buspirone and 1-PP were analyzed based on a two-compartment model with metabolite formation. Differences in the pharmacokinetics of buspirone and 1-PP were observed with values for clearance of 13.1 and 8.2 ml/min and for terminal elimination half-life of 25 and 79 min, respec- tively. At least 26% of the administered dose of buspirone was converted into 1-PP. Complex hypothermic effects versus time profiles were observed, which were successfully analyzed on the basis of a physiological indirect response model with set- point control. Both buspirone and 1-PP behaved as partial agonists relative to R-()-8-hydroxy-2-(di-n-propylamino)tetra- lin (R-8-OH-DPAT) with values of the intrinsic activity of 0.465 and 0.312, respectively. Differences in the potency were ob- served with values of 17.6 and 304 ng/ml for buspirone and 1-PP, respectively. The results of this analysis show that bu- spirone and 1-PP behave as partial 5-hydroxytryptamine 1A agonists in vivo and that following intravenous administration the amount of 1-PP formed is too small to contribute to the hypothermic effect. Buspirone is a selective 5-HT 1A agonist that is used clini- cally as an anxiolytic drug (for reviews on its pharmacological properties, see New, 1990 and Fulton and Brogdon, 1997). In vivo buspirone is metabolized into 1-(2-pyrimidinyl)-pipera- zine (1-PP), which also possesses affinity to the 5-HT 1A re- ceptor. This metabolite could therefore contribute to the ef- fect of buspirone (Bianchi et al., 1988; Manahan-Vaughan et al., 1995; Cao and Rodgers, 1997). At present there is limited quantitative information on the magnitude of this effect. In theory the contribution of the metabolite to the effect is determined by the relative concentrations of buspirone and 1-PP, their relative in vivo potency and intrinsic activity, and the nature of the pharmacodynamic interaction between the two. Recently we have developed a physiological pharmacoki- netic-pharmacodynamic (PK-PD) model to quantitatively characterize the pharmacodynamics of 5-HT 1A receptor ago- nists in vivo using the hypothermic effect as a pharmacody- namic endpoint (Zuideveld et al., 2001, 2002a). The hypother- mic response is considered a robust marker of 5-HT 1A activity (Millan et al., 1993; Cryan et al., 1999). It has been shown that the PK-PD model allows for the estimation of the in vivo potency and intrinsic activity of a wide array of different 5-HT 1A agonists such as R-8-OH-DPAT, S-()-8- hydroxy-2-(di-n-propylamino)tetralin (S-8-OH-DPAT), N-[2- [4-(2-methoxyphenyl)-1-piperzinyl]ethyl]-n-2-pyridinyl-cy- clohexanecarboxamide (WAY-100,635), and Flesinoxan (Zuideveld et al., 2001, 2002a, 2002b). In the present inves- tigation we apply this model to determine the relative in vivo potency and intrinsic activity of buspirone and its metabolite 1-PP, taking into consideration the potential pharmacody- namic interaction between the two. In this respect it is im- portant to consider the mechanism of the hypothermic re- sponse of buspirone and 1-PP. The hypothermic effect of buspirone has been widely studied (Higgins et al., 1988; 1 Present address: Pharsight Corporation, Argentum, 2 Queen Caroline St., Hammersmith, W6 9DT London, UK. 2 Present address: Pliva I ¨ stra ivae `ki Institut, Division of Pharmacology, Prilaz Baruna Filipoviaæ 25, 10000 Zagreb, Croatia. 3 Pfizer Global Research & Development, Discovery Biology, Ramsgate Road, Sandwich, Kent CT13 9NJ, UK. Article, publication date, and citation information can be found at http://jpet.aspetjournals.org. DOI: 10.1124/jpet.102.036798. ABBREVIATIONS: 5-HT 1A , 5-hydroxytryptamine 1A ; 1-PP, 1-(2-pyrimidinyl)-piperazine; PK-PD, pharmacokinetic-pharmacodynamic; R-8-OH- DPAT, R-()-8-hydroxy-2-(di-n-propylamino)tetralin; S-()-8-hydroxy-2-(di-n-propylamino)tetralin; WAY-100,635, N-[2-[4-(2-methoxyphenyl)-1- piperazinyl]ethyl]-n-2-pyridinyl-cyclohexanecarboxamide; PVP, polyvinylpyrrolidone; HPLC, high performance liquid chromatography; 2 -AR, 2 -adrenergic receptor. 0022-3565/02/3033-1130 –1137$7.00 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 303, No. 3 Copyright © 2002 by The American Society for Pharmacology and Experimental Therapeutics 36798/1021263 JPET 303:1130–1137, 2002 Printed in U.S.A. 1130 at ASPET Journals on July 29, 2018 jpet.aspetjournals.org Downloaded from

Pharmacokinetic-Pharmacodynamic Modeling of Buspirone …jpet.aspetjournals.org/content/jpet/303/3/1130.full.pdf · Pharmacokinetic-Pharmacodynamic Modeling of Buspirone and Its Metabolite

  • Upload
    ngohanh

  • View
    227

  • Download
    2

Embed Size (px)

Citation preview

Pharmacokinetic-Pharmacodynamic Modeling of Buspironeand Its Metabolite 1-(2-Pyrimidinyl)-piperazine in Rats

KLAAS P. ZUIDEVELD,1 JASNA RUSIC-PAVLETIC,2 HUGO J. MAAS, LAMBERTUS A. PELETIER,PIET H. VAN DER GRAAF,3 AND MEINDERT DANHOF

Leiden/Amsterdam Center for Drug Research, Division of Pharmacology, Gorlaeus Laboratories, Leiden, The Netherlands (K.P.Z., J.R.-P.,H.J.M., P.H.V., M.D.); and Mathematical Institute, Leiden University, Niels Bohrweg, Leiden, The Netherlands (L.A.P.)

Received April 4, 2002; accepted July 3, 2002

ABSTRACTThe objective of this investigation was to compare the in vivopotency and intrinsic activity of buspirone and its metabolite1-(2-pyrimidinyl)-piperazine (1-PP) in rats by pharmacokinetic-pharmacodynamic modeling. Following intravenous adminis-tration of buspirone (5 or 15 mg/kg in 15 min) or 1-PP (10 mg/kgin 15 min), the time course of the concentrations in blood weredetermined in conjunction with the effect on body temperature.The pharmacokinetics of buspirone and 1-PP were analyzedbased on a two-compartment model with metabolite formation.Differences in the pharmacokinetics of buspirone and 1-PPwere observed with values for clearance of 13.1 and 8.2 ml/minand for terminal elimination half-life of 25 and 79 min, respec-tively. At least 26% of the administered dose of buspirone was

converted into 1-PP. Complex hypothermic effects versus timeprofiles were observed, which were successfully analyzed onthe basis of a physiological indirect response model with set-point control. Both buspirone and 1-PP behaved as partialagonists relative to R-(�)-8-hydroxy-2-(di-n-propylamino)tetra-lin (R-8-OH-DPAT) with values of the intrinsic activity of 0.465and 0.312, respectively. Differences in the potency were ob-served with values of 17.6 and 304 ng/ml for buspirone and1-PP, respectively. The results of this analysis show that bu-spirone and 1-PP behave as partial 5-hydroxytryptamine1Aagonists in vivo and that following intravenous administrationthe amount of 1-PP formed is too small to contribute to thehypothermic effect.

Buspirone is a selective 5-HT1A agonist that is used clini-cally as an anxiolytic drug (for reviews on its pharmacologicalproperties, see New, 1990 and Fulton and Brogdon, 1997). Invivo buspirone is metabolized into 1-(2-pyrimidinyl)-pipera-zine (1-PP), which also possesses affinity to the 5-HT1A re-ceptor. This metabolite could therefore contribute to the ef-fect of buspirone (Bianchi et al., 1988; Manahan-Vaughan etal., 1995; Cao and Rodgers, 1997). At present there is limitedquantitative information on the magnitude of this effect. Intheory the contribution of the metabolite to the effect isdetermined by the relative concentrations of buspirone and1-PP, their relative in vivo potency and intrinsic activity, andthe nature of the pharmacodynamic interaction between thetwo.

Recently we have developed a physiological pharmacoki-netic-pharmacodynamic (PK-PD) model to quantitativelycharacterize the pharmacodynamics of 5-HT1A receptor ago-nists in vivo using the hypothermic effect as a pharmacody-namic endpoint (Zuideveld et al., 2001, 2002a). The hypother-mic response is considered a robust marker of 5-HT1A

activity (Millan et al., 1993; Cryan et al., 1999). It has beenshown that the PK-PD model allows for the estimation of thein vivo potency and intrinsic activity of a wide array ofdifferent 5-HT1A agonists such as R-8-OH-DPAT, S-(�)-8-hydroxy-2-(di-n-propylamino)tetralin (S-8-OH-DPAT), N-[2-[4-(2-methoxyphenyl)-1-piperzinyl]ethyl]-n-2-pyridinyl-cy-clohexanecarboxamide (WAY-100,635), and Flesinoxan(Zuideveld et al., 2001, 2002a, 2002b). In the present inves-tigation we apply this model to determine the relative in vivopotency and intrinsic activity of buspirone and its metabolite1-PP, taking into consideration the potential pharmacody-namic interaction between the two. In this respect it is im-portant to consider the mechanism of the hypothermic re-sponse of buspirone and 1-PP. The hypothermic effect ofbuspirone has been widely studied (Higgins et al., 1988;

1 Present address: Pharsight Corporation, Argentum, 2 Queen Caroline St.,Hammersmith, W6 9DT London, UK.

2 Present address: Pliva Istra ivaeki Institut, Division of Pharmacology,Prilaz Baruna Filipoviaæ 25, 10000 Zagreb, Croatia.

3 Pfizer Global Research & Development, Discovery Biology, RamsgateRoad, Sandwich, Kent CT13 9NJ, UK.

Article, publication date, and citation information can be found athttp://jpet.aspetjournals.org.

DOI: 10.1124/jpet.102.036798.

ABBREVIATIONS: 5-HT1A, 5-hydroxytryptamine1A; 1-PP, 1-(2-pyrimidinyl)-piperazine; PK-PD, pharmacokinetic-pharmacodynamic; R-8-OH-DPAT, R-(�)-8-hydroxy-2-(di-n-propylamino)tetralin; S-(�)-8-hydroxy-2-(di-n-propylamino)tetralin; WAY-100,635, N-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-n-2-pyridinyl-cyclohexanecarboxamide; PVP, polyvinylpyrrolidone; HPLC, high performance liquid chromatography; �2-AR,�2-adrenergic receptor.

0022-3565/02/3033-1130–1137$7.00THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 303, No. 3Copyright © 2002 by The American Society for Pharmacology and Experimental Therapeutics 36798/1021263JPET 303:1130–1137, 2002 Printed in U.S.A.

1130

at ASPE

T Journals on July 29, 2018

jpet.aspetjournals.orgD

ownloaded from

Koenig et al., 1988; Young et al., 1993) and is generallyconsidered a specific 5-HT1A receptor-mediated response(Koenig et al., 1988; Kommisarev et al., 1990; Cryan et al.,1999). 1-PP however has affinity for both the 5-HT1A receptor(KD � 1035 nM � 94 ng/ml�1) and the �2-adrenoceptor(�2-AR) (KD � 40 nM � 3.6 ng/ml�1) (Gobert et al., 1995) andit is therefore possible that this effect is not mediated by the5-HT1A receptor (Kommisarev et al., 1990).

Materials and MethodsExperiments that were approved by the Ethics Committee of the

University of Leiden were performed on male Wistar rats (BroekmanBV, Someren, The Netherlands) weighing 315 � 3 g (mean � S.E.M.,n � 27). They were housed in standard plastic cages (six per cagebefore surgery and individually after surgery). They were housedunder normal 12-h light/dark cycle (lights on at 7 AM and lights offat 7 PM) and a temperature of 21°C. During the light period, a radiowas turned on for background noise. Acidified water and food (labo-ratory chow; Hope Farms, Woerden, The Netherlands) were providedad libitum before the experiment.

Surgical Procedure

Eight days before the experiment, the rats were operated upon.The animals were anesthetized with an intramuscular injection of0.1 ml/kg Domitor (1 mg/ml medetomidine hydrochloride; Pfizer,Capelle a/d IJsel, The Netherlands) and 1 ml/kg Ketalar (50 mg/mlketamine base; Pfizer, Hoofddorp, The Netherlands). Indwelling py-rogen-free cannulae for drug administration were implanted into theright jugular vein (Polythene, 14 cm, 0.52-mm i.d., 0.96-mm o.d.) forinfusions of buspirone and 1-PP. For blood sampling, the left femoralartery (Polythene, 4 cm, 0.28-mm i.d., 0.61-mm o.d. � 20 cm,0.58-mm i.d., 0.96-mm o.d.) was cannulated. Cannulae were tun-neled subcutaneously to the back of the neck and exteriorized. Toprevent coagulation of blood, the cannulae were filled with a 25%(w/v) solution of polyvinylpyrrolidone (PVP) (Brocacef, Maarssen,The Netherlands) in a 0.9% (w/v) pyrogen-free sodium chloride solu-tion (NPBI, Emmer-Compascuum, The Netherlands) that contained50 IU/ml heparin (Leiden University Medical Center, Leiden, TheNetherlands). Just before the experiment, the PVP solution wasremoved and the cannulae were flushed with saline containing 20IU/ml heparin. The skin in the neck was stitched with normal su-tures, and the skin in the groin was closed with wound clips. Atelemetric transmitter (Physiotel implant TA10TA-F40 system; DataSciences International, St. Paul, MN) with a weight of approximately7 g, which had been made pyrogen free with CIDEX (22 g/l glutar-aldehyde; Johnson and Johnson Medical Ltd., Gargrave, Skipton,U.K.) for at least 2 h, was implanted into the abdominal cavity for themeasurement of core body temperature. After surgery, an injection ofthe antibiotic ampicillin (0.6 ml/kg of a 200 mg/ml, A.U.V., Cuijk, TheNetherlands) was administered.

Experimental Protocol

Dosage Regimen. The experiments were performed 8 days aftersurgery. Rats received infusions of buspirone and 1-PP. For buspi-rone, infusions of 5 and 15 mg/kg in 15 min (12.96 and 38.91 �mol/kgfor n � 6 and n � 7, respectively) were given. For 1-PP, an infusionof 10 mg/kg (60.90 �mol/kg for n � 6) in 15 min was administered.Six rats received vehicle treatments in which an equivalent amountof saline was infused. For the infusions, an external cannula with aspecific volume was filled with a solution of the drug in an amount ofsaline, which was calculated according to the weight of the rat andconnected to the infusion pump (BAS BeeHive, BAS BioanalyticalSystems Inc., West Lafayette, IN). All the experiments started be-tween 9:00 and 9:30 AM.

Blood Sampling. Approximately 15 to 18 serial blood samples of50 �l were taken according to a fixed time schedule to determine the

pharmacokinetics of the drug. The exact amount was measured witha capillary (Servoprax, Wesel, Germany) and transferred into a glasscentrifuge tube containing 400 �l of Millipore water for hemolysis.During the experiment the samples were kept on ice. After theexperiment, samples were stored at �20°C pending analysis.

Data Acquisition

Temperature Measurements. To measure the body tempera-ture of the rat a telemetric system (Physiotel Telemetry system;Data Sciences International) was used. A telemetric transmitter(Physiotel implant TA10TA-F40; Data Sciences International) wasimplanted in the abdominal cavity of the rat. The transmitter mea-sured the body temperature every 30 s for a 2-s period and signaledthe data to a receiver (Physiotel receiver, model RPC-1; Data Sci-ences International). The receiver was connected to the computerthrough a BCM 100 consolidation matrix (Data Sciences Internation-al). The temperature profiles and the room temperature data (C10Ttemperature adapter; Data Sciences International) were processedand visualized using the Dataquest LabPro software (Data SciencesInternational, running under OS/2 Warp, IBM).

HPLC Analysis of Buspirone and 1-PP. The blood concentra-tions of buspirone and 1-PP were determined simultaneously byHPLC using UV detection. The HPLC system consisted of a Shi-madzu LC-10 AD pump (flow rate: 1 ml/min; Shimadzu Corporation,Kyoto, Japan), a Waters 712WISP autosampler (Waters, Milford,MA), a Waters Sperisorb S5 OD/CN column (mixed phase 4.6 mm �250 mm) together with a Sperisorb S5CN guard column (all Waters),a Spectroflow 757 UV absorbance detector set at 240 nm (KratosAnalytical, Ramsey, NJ), and a C-R3A Chromatopac integrator (Shi-madzu Corporation). The mobile phase consisted of a mixture of a0.05 M potassium phosphate buffer (pH 5) and acetonitrile (60:40,v/v), to which 0.005 M KCl was added. The analytes were extractedfrom blood using an extensive liquid-liquid extraction to allow for thesimultaneous analysis of both buspirone and 1-PP. To 50 �l of bloodhemolyzed in 400 �l of water, 100 �l of the internal standard solution(5000 ng/ml 4-[3-(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)-pi-perazine) was added. The mixture was deproteinized by adding 2 mlof acetonitrile and centrifuged. The supernatant was transferred to aclean tube and volumes of 50 �l of hydrochloric acid (1 M), 1 ml ofwater and 3 ml of distilled ethyl acetate were added. After separa-tion, 2 ml of borate buffer (0.2 M, pH 10) and 3 ml of dichloromethanewas added to the aqueous solution. The organic phase was trans-ferred to a clean tube and evaporated to dryness under vacuum at37°C. The residue was re-dissolved in 100 �l of 30% acetonitrile, ofwhich 50 �l was injected into the HPLC system. On the day ofanalysis, a 9-point calibration curve was prepared for both buspironeand 1-PP by spiking 50 �l of blood hemolyzed in 400 �l of water with50 �l of buspirone and 1-PP. This resulted in a blood concentrationrange of 10 to 10,000 ng/ml for buspirone and 1 to 10,000 ng/ml for1-PP. Samples were processed as described, and the peak-area ratioof either buspirone or 1-PP over the internal standard were calcu-lated. Calibration curves were constructed by weighted linear regres-sion ([weight factor � 1/(peak-area ratio)]2). Recovery, intra-, andinter-assay variation were determined using spiked blood of 500 and1500 ng/ml for buspirone and 50, 500, and 1500 ng/ml for 1-PP. Forbuspirone, recovery was (n � 3, mean � S.E.M., corrected for volumeloss) 41 � 3.6% and 45 � 8.6% for 500 and 1500 ng/ml, respectively.For 1-PP, recovery was (n � 3, mean � S.E.M., corrected for volumeloss) 88 � 9.8%, 82 � 3.8%, and 88 � 16.5% for 50, 500, and 1500ng/ml, respectively. Inter-assay variation (n � 10) was 9.9% and19.2% (accuracy: �8.9% and 1.2%) for 500 and 1500 ng/ml buspironeand 9.3, 14.4, and 16.3% (accuracy: �12.6, �0.4, and �0.8%) for 50,500, and 1500 ng/ml 1-PP. Intra-day assay variation (n � 4) was 4.4and 3.4% (accuracy: �2.5 and 1.4%) for 500 and 1500 ng/ml buspi-rone and 7.4, 9.8, and 7.9% (accuracy: 12.6, �1.7, and �17.3%) for 50,500, and 1500 ng/ml 1-PP. Detection limit (signal-to-noise ratio 3)using 50 �l of blood was 22.2 and 5.2 ng/ml for buspirone and 1-PP,respectively.

PK-PD Modeling of Buspirone- and 1-PP-Mediated Hypothermia 1131

at ASPE

T Journals on July 29, 2018

jpet.aspetjournals.orgD

ownloaded from

Chemicals. Buspirone and 1-PP were generously donated byBristol-Myers Squibb (Princeton, NJ). 4-[3-(Benzotriazol-1-yl)pro-pyl]-1-(2-methoxyphenyl)-piperazine maleate was purchased fromTocris (Lanford, Bristol, UK). Dichloromethane was purchased fromRiedel-de Haen (Seelze, Germany). All other chemicals used were ofanalytical grade (Baker, Deventer, The Netherlands).

Data Analysis

A population approach was used to quantify both the pharmaco-kinetics and pharmacodynamics of buspirone and 1-PP. Using thisapproach, the population is taken as the unit of analysis while takinginto account both intra-individual variability in the model parame-ters as well as inter-individual residual error. Modeling was per-formed using the nonlinear mixed effects modeling software NON-MEM developed by Sheiner & Beal (version V 1.1, NONMEM projectGroup, University of California, San Francisco, CA). Individual pre-dictions were obtained in a Bayesian post hoc step.

Pharmacokinetic Analysis. The concentration-time profiles ofthe buspirone administration, the 1-PP formed during the buspironeadministration, and the 1-PP following direct administration werebest described using a two-compartment model with metabolite for-mation. In this model, the input function for the metabolite followingadministration of the parent drug is the fraction of buspirone that isconverted into 1-PP. This model is mathematically described asfollows (Houston, 1985)

�V1,B �

dC1,B

dt� in � CLB � C1,B � CLd,B � C1,B�CLd,B � C2,B�CLB31-PP � C1,B ,

V2,B �dC2,B

dt� CLd,B � C1,B�CLd,B � C2,B ,

V1,1-PP �dC1,1-PP

dt� CLB31-PP � C1,B�CL1-PP � C1,1-PP�CLd,1-PP� C1,1-PP�CLd,1-PP � C1,1-PP ,

V2,1-PP �dC2,1-PP

dt� CLd,1-PP � C1,1-PP�CLd,1-PP � C2,1-PP ,

�(1)

where C1 represents the concentration in the central compartmentand C2 the concentration in the distribution compartment. V1 and V2

are the compartments’ volumes of distribution. CL represents theclearance and CLd the inter-compartmental clearance (see Fig. 1 fora schematic overview of the model). The subscripts B and 1-PPdenote buspirone and the metabolite 1-PP respectively. For the in-fusion, in is defined as

in �Doseinf

Tinfwhen t � Tinf ,

(2)

in � 0 when t � Tinf ,

where Tinf is the duration of the infusion and Doseinf is the totalamount of drug infused. CLB31-PP in eq. 1 represents the clearancefor the conversion of buspirone into 1-PP. The value of this clearancecan be used to determine the fraction of buspirone that is convertedinto 1-PP, according to

fB31-PP �CLB31-PP

CLB31-PP�CLB(3)

The fraction calculated based on this equation does not take intoaccount the amount of 1-PP metabolized directly after being formedand is therefore smaller than the actual amount formed (Houston,1985). The model (eq. 1) was implemented in NONMEM using AD-VAN6. The concentration-time profiles of the separate buspirone and1-PP infusions were also modeled using a standard two-compart-ment pharmacokinetic model. Concentrations were all entered inmolar units. The model as implemented in NONMEM’s ADVAN3TRANS4 was used for this purpose. For all models, inter-individualvariability on the parameters was modeled by an exponential equa-tion,

Pi � � � exp��i�, (4)

where � is the population value for parameter P, Pi is the individualvalue and �i is the random deviation of Pi from P. The values of �i areassumed to be independently normally distributed with mean zeroand variance 2. Residual error was characterized by a proportionalerror model,

Cmij � Cpij � �1 �ij�, (5)

where Cpij is the jth plasma concentration for the ith individualpredicted by the model, Cmij is the measured concentration, and �

accounts for the residual deviance of the model predicted value fromthe observed concentration. Different � values were assigned for theobservations of buspirone, 1-PP during buspirone infusions, and1-PP when modeled simultaneously. The values for � are assumed tobe independently normally distributed with mean zero and variance�2. The values for the population �, 2, and �2 are estimated usingthe first order method in NONMEM.

Pharmacodynamics Analysis. Recently we have proposed aphysiological model, to describe the complex time course of the hy-pothermic response in vivo following the administration of 5-HT1A

receptor agonists (Zuideveld et al., 2001). The model contains aset-point control that can be attenuated by the drug receptor inter-action and utilizes the concept of an indirect physiological responsemodel (Dayneka et al., 1993). This model considers a 0th-order rateconstant associated with the warming of the body (kin) and a first-order rate constant associated with the cooling of the body (kout). Themodel is further characterized by the phenomenological system pa-rameter A and a parameter , which determines the amplification inthe system (for details, see Zuideveld et al., 2001). In this model, thestimulus generated by the drug-receptor interaction, which drivesthe physiological processes that lower temperature and characterizethe drug-receptor interaction in terms of the drugs’ potency andintrinsic activity, is described by the function f(C),

f�C� � S �Smax � Cn

SC50n Cn , (6)

where S is the stimulus, Smax is the maximum stimulus the drug canproduce, C is the drug concentration, SC50 is the concentrationrequired to produce 50% of the maximum stimulus and n is the slopefactor, which determines the steepness of the curve. The parameterswere estimated for both 1-PP and buspirone.

However because 1-PP is formed during the buspirone adminis-tration, both compounds will be present simultaneously. Hence apharmacodynamic interaction model may be needed to quantify thepharmacodynamics in this situation. The contribution of 1-PP to theeffect of buspirone was evaluated considering both a non-competitiveand a competitive interaction model. To characterize the non-com-

Fig. 1. Schematic representation of the two-compartment model for in-travenous administration of buspirone and a two-compartment model for1-PP, where the input is the fraction of buspirone converted into 1-PP (eq.1). C1 represents the concentration in the central compartment and C2the concentration in the distribution compartment. V1 and V2 are thecompartments’ volumes of distribution. CL represents the clearance andCLd the inter-compartmental clearance. The subscripts B and 1-PP de-note buspirone and the metabolite 1-PP, respectively. For an infusion, inis defined as described in eq. 2. CLB31-PP represents the clearance for theconversion of buspirone into 1-PP.

1132 Zuideveld et al.

at ASPE

T Journals on July 29, 2018

jpet.aspetjournals.orgD

ownloaded from

petitive pharmacodynamic interaction between the two drugs, thefollowing equation has been proposed (Ariens and Simonis, 1964),

SB�1-PP �SB

Sm

S1-PP

Sm�

SB � S1-PP

Sm2 , (7)

where SB�1-PP is the combined stimulus of both drugs, SB and S1-PP

are the single drug stimuli and Sm is the maximum achievablestimulus of the system. As the maximum achievable stimulus in thesystem studied equals 1, substituting eq. 6 for both drugs in eq. 7,yields

SB�1-PP �SB � CB

nB

SC50B

nB CBnB

S1-PP � C1-PP

n1-PP

SC501-PP

n1-PP C1-PPn1-PP

�SB�CB

nB�S1-PP�C1-PPn1-PP

(SC50B

nB �CBnB)�(SC501-PP

n1-PP �C1-PPn1-PP)

. (8)

The competitive pharmacodynamic interaction between the twodrugs acting at the same receptor was characterized by the equationoriginally proposed by Holford and Sheiner (1981),

SB�1-PP �

SmaxB �� C1,B

SC50B�nB�Smax1-PP ��C1,1-PP

SC501-PP�n1-PP

1 � C1,B

SC50B�nB

�C1,1-PP

SC501-PP�n1-PP

, (9)

where the stimulus of both drugs is combined. Recently, a reducedversion of this model has been applied successfully to characterizethe interaction between R-8-OH-DPAT and the competitive antago-nist WAY-100,635 (Zuideveld et al., 2002a). The pharmacodynamicparameters found for buspirone with eqs. 8 and 9 were comparedwith parameters obtained when considering each drug separately onthe basis of eq. 6. This allows estimation of the relative contributionof 1-PP to the observed effect during and after the buspirone infu-sion.

The pharmacodynamic model was implemented in NONMEM us-ing ADVAN6. The maximum stimulus Smax was assumed to be 1 andtherefore fixed to this value (Zuideveld et al., 2001). Inter-individualvariability on the parameters was modeled to an exponential equa-tion, as described in eq. 4. Residual error was characterized by aproportional error model:

ymij � ypij � �1 �ij�, (10)

where ypij is the jth prediction for the ith individual predicted by themodel, ymij is the measurement, and � accounts for the residualdeviance of the model predicted value from the observed value. Thevalues for the population �, 2, and �2 are estimated using thecentering first-order conditional estimation method with the first-order model in NONMEM. A conditional estimation method wasused due to the high degree of nonlinearity of the model and the highdensity of the data. The centering option gives the average estimateof each element of � together with a P value that can be used toassess whether this value is sufficiently close to zero. The occurrenceof an average � that is significantly different from zero indicates anuncentered or a biased fit. This method was not chosen because theaverage estimates of each element of � were expected to be differentfrom zero, but rather to greatly decrease computing time as requiredwith just the conditional estimation method. To further decreasecomputing time only 1/16th of the temperature data set was used formodeling, reducing the temperature measurements from over 900measurements per individual to approximately 60. The implicationof this reduction is that there is a data point every 8 min, as opposedto every 0.5 min. This reduction did not void the integrity of the dataprofiles.

Statistical Analysis. Goodness-of-fit was analyzed using the ob-jective function and various diagnostic methods. Model selection wasbased on the Akaike Information Criterion (Akaike, 1974) and as-sessment of the accuracy and precision of parameter estimates and

correlations between the parameter estimates. Comparisons of max-imal decreases in body temperature were performed using an un-paired t test, where a P value �0.05 was considered significant.

ResultsPharmacokinetics. During the experiments blood sam-

ples were taken to construct individual concentration-timeprofiles for both buspirone and 1-PP. Individual concentra-tion profiles were predicted by fitting a population pharma-cokinetic model to the data (eq. 1). On the basis of the con-centration curves, the goodness-of-fit plots and the minimumvalue of the objective function, two-compartment modelswere selected for buspirone, 1-PP, and 1-PP following admin-istration of buspirone. The concentration-time profiles of the1-PP administration were fitted both separately and simul-taneously with the “1-PP during the buspirone administra-tion” profiles. Ultimately the values of the pharmacokineticparameters of 1-PP were fixed to those obtained upon theseparate administration. These values were subsequentlyentered as constants in the analysis of the concentrationprofiles of buspirone and 1-PP on the basis of the two-com-partment metabolite model (eq. 1), which resulted in a sig-nificant improvement of the objective function. The obtainedpharmacokinetic parameter estimates were used as input inthe pharmacodynamic analysis because it is believed thatthese concentrations are closest to the “true” concentrationsat each observation of the body temperature. Both the indi-vidually measured concentrations and the population pre-dicted profiles following administration of buspirone and1-PP are depicted in Fig. 2. All parameters were estimated intheir mixed effect form, the random effect being incorporatedin exponential error models. The results of the analysis arerepresented in Table 1. The clearance found for buspironewas 13 ml/min resulting in a half-life of 25 min. For 1-PP, aclearance of 8.2 ml/min was found, which corresponds to ahalf-life of 79 min. Based on the pharmacokinetic analysis(eq. 3), 26% of the administered dose of buspirone is metab-olized to 1-PP. Pharmacokinetic parameter estimates were

Fig. 2. Blood concentration versus time profiles of buspirone (open cir-cles) and 1-PP (closed circles) following administration of buspirone (5mg/kg in 15 min) (panel A), buspirone (15 mg/kg in 15 min) (panel B), and1-PP (10 mg/kg in 10 minutes) (panel C). Depicted are the individuallymeasured concentrations (circles) and the population-predicted concen-tration profiles (solid line, buspirone; dashed line, 1-PP).

PK-PD Modeling of Buspirone- and 1-PP-Mediated Hypothermia 1133

at ASPE

T Journals on July 29, 2018

jpet.aspetjournals.orgD

ownloaded from

independent of administered dose and no statistically signif-icant correlations between the parameter estimates and be-tween the parameters from the different dosing groups weredetected. Table 1 further displays the inter-individual vari-ation, and the intra-individual variation, which are bothfound to be reasonable for both buspirone and 1-PP. For thedifferent drugs and administrations, different intra-individ-ual variation was predicted, which significantly improved thefit. The precision of the parameters prediction, as repre-sented by their 95% confidence intervals, is good.

Pharmacodynamics. The average time-effect profiles forthe effect on body temperature following administration ofvehicle, buspirone, and 1-PP are depicted in Fig. 3. Theaverage baseline temperature (S.E.M., n � 33) was 38.0 �0.005°C. Following administration of 5 and 15 mg/kg buspi-rone in a 15-min infusion, a maximal decrease of 2.8 � 0.3°Cwas observed after approximately 55 min. Administration of10 mg/kg 1-PP in a 15-min infusion resulted in a maximumtemperature decrease of 1.6 � 0.2°C after approximately 36

min. The temperature decreases were all statistically signif-icantly (P � 0.05) different from the vehicle treatment, con-sisting of saline. After reaching a maximal decrease, a rapidrecovery was observed followed by a plateau phase before thebody temperature returned to baseline.

The time-effect profiles for buspirone and 1-PP were ana-lyzed on the basis of the set-point model combined with 1) thesigmoidal pharmacodynamic model, 2) the noncompetitive,and 3) the competitive interaction model. This analysis re-sults in two types of pharmacodynamic parameters: drug-specific parameters (SC50, Smax, and n) and system-specificparameters (kin, A, and ). These parameters were estimatedusing the centering first-order conditional estimation methodwith a first-order model. The average values for all � valueswere not significantly different from 0. Population parameterestimates are depicted in Table 2. Individual post hoc pre-dictions of the parameters were not biased with respect to thedifferent treatment groups. Interestingly, no difference wasfound in the pharmacodynamic parameter estimates for bu-spirone regardless of whether the sigmoid pharmacodynamicmodel (eq. 6) or the noncompetitive interaction model (eq. 8)were used. The parameters obtained with the competitiveinteraction model (eq. 9) differ substantially, in particularwith respect to the values of SC50 and Smax. No differencewas found in the system-specific parameters between thevarious models or various compounds. The inter-individualvariations of the pharmacodynamic parameter estimates for1-PP were larger than for buspirone. Figure 4 depicts theindividual observations, predictions, and the population pre-dictions for each model and two typical individual rats.

DiscussionThe present study provides novel information on the phar-

macokinetic and pharmacodynamic relationship of buspironeand its metabolite 1-PP in vivo. Specifically, the presentinvestigation has also focused on the relative potency andintrinsic activity of the metabolite 1-PP and the extent towhich 1-PP contributes to the effect of the buspirone. Thisstudy shows that with values of the relative intrinsic activityof 0.465 and 0.312 both buspirone and 1-PP behave as partialagonists relative to R-8-OH-DPAT. Furthermore it is shownthat the potency of 1-PP is 20-fold lower compared with

TABLE 1Population pharmacokinetic parameters and inter- and intra-individual variabilities of buspirone and 1-PP

Drug Parameter ValueCV Inter-Individual

95% CIa Unit1-PP Fixed 1-PP

%

Buspirone CLB 13.1 33 7.96–18.2 ml/minCLd,B 41.7 35 23.9–59.4 ml/minV1,B 8.94 96 4.92–12.9 mlV2,B 626 38 467–785 ml

1-PP CLB31-PP 4.53 51 1.63–7.43 ml/minCL1-PP 8.22 31 35 6.51–9.89 ml/minCLd,1-PP 123 35 24 84.2–162 ml/minV1,1-PP 8.32 131 213 2–14.6 mlV2,1-PP 924 66 34 722–1130 mlfB31-PP 25.7 N.A. N.A. N.A. N.A.

Buspirone CV intra-individual: 28%1-PP (metabolite) CV intra-individual: 51%1-PP (administered) CV intra-individual: 26%

CV, coefficient of variation; N.A., not applicable.a 95% CI is the 95% confidence interval over the precision of the estimated parameter.

Fig. 3. Average temperature-time profiles (�S.E.M.) for 5 and 15 mg/kgbuspirone in 15-min infusion, closed squares and triangles, respectively(n � 6 and n � 7), 10 mg/kg 1-PP in a 15-min infusion; closed circles (n �6) and a control experiment consisting of a regular infusion containing anequivalent amount of saline (thick line, n � 12). All infusions started att � 0.

1134 Zuideveld et al.

at ASPE

T Journals on July 29, 2018

jpet.aspetjournals.orgD

ownloaded from

buspirone and that following intravenous administration ofthe parent drug, formation of 1-PP does not contribute to theobserved effect.

In comparative pharmacodynamic investigations in vivo, itis important to take differences in pharmacokinetics betweenthe compounds into account. Therefore in the present inves-tigation, the effects of buspirone and 1-PP were determinedby an integrated PK-PD approach. In the present study, thevalue of the clearance was 13 ml/min resulting in a half-lifeof 25 min for buspirone and a clearance of 8.2 ml/min with acorresponding half-life of 79 min for the metabolite 1-PP.Both values are in close agreement with the values reportedby Caccia et al. (1985). In the present study, it was estimatedon the basis of the two-compartment model with metaboliteformation that at least 26% of the administered dose ofbuspirone is metabolized into 1-PP. Again, this appears to beconsistent with findings of Caccia et al. (1986) and Jajoo et al.(1989a, 1989b) who showed that 25% of the administereddose is excreted into urine as 1-PP. Overall the pharmacoki-netics of both buspirone and 1-PP could be characterizedsuccessfully using the metabolite model.

To quantify the pharmacodynamic response of both buspi-rone and 1-PP the hypothermic response was chosen as apharmacodynamic endpoint. Recently, we have developed anintegrated pharmacokinetic-pharmacodynamic model to

characterize 5-HT1A receptor-mediated hypothermia, interms of potency and intrinsic activity of the agonists used(Zuideveld et al., 2001). The 5-HT1A agonist-induced hypo-thermic response was chosen, because it is considered to be arobust marker for 5-HT1A activity (Millan et al., 1993; Cryanet al., 1999) and behaves as an ideal biomarker, which iscontinuous, reproducible, and sensitive enough to discrimi-nate between a full and a partial agonist (Hadrava et al.,1996) and is selective (Millan et al., 1993; Cryan et al., 1999).The PK-PD model itself combines an indirect physiologicalresponse model (Dayneka et al., 1993) with a set-point con-trol. It is believed that the 5-HT1A receptors play a role inmaintaining the body’s set-point temperature (Lin et al.,1983; Schwartz et al., 1998; Zeisberger, 1998). In additionnumerous reports suggest that this set-point is regulatedthrough an interplay between the 5-HT1A (hypothermia) andthe 5-HT2A/C (hyperthermia) receptor system (Gudelsky etal., 1986; Salmi and Ahlenius, 1998). Therefore the proposedmodel is considered to reflect physiology. This is confirmed bythe observation that a competitive interaction model accu-rately characterizes the interaction between R-8-OH-DPATand WAY-100,635 (Zuideveld et al., 2002a). In addition, thismodel has been applied successfully to characterize the invivo pharmacodynamics of a range of 5-HT1A ligands includ-ing R- and S-8-OH-DPAT, WAY-100,635 and flesinoxan(Zuideveld et al., 2001, , 2002b) showing differences in bothin vivo potency as well as intrinsic activity.

Regarding the hypothermic response induced by buspi-rone, much evidence has been presented indicating that it ismediated through the 5-HT1A receptor (Higgins et al., 1988;Young et al., 1993). Furthermore its effect can be antago-nized by selective 5-HT1A receptor antagonists (Koenig et al.,1988; Cryan et al., 1999) and its selectivity over other recep-tors that may also mediate a hypothermic response, e.g., the�2-AR (Dilsaver et al., 1989; MacDonald et al., 1989), �-opi-oid (OP3) (Spencer et al., 1988), dopamine D3 (Barik and deBeaurepaire, 1998) and adenosine A1 receptors (Malhotraand Gupta, 1997) is high (New, 1990; Gobert et al., 1995).1-PP on the other hand has affinity for both the 5-HT1A

(KD � 1035 nM � 94 ng/ml) and the �2-AR (KD � 40 nM � 3.6ng/ml) (Gobert et al., 1995). The mechanism of the hypother-mic response of 1-PP is therefore subject to debate. Despitethe fact that 1-PP behaves as an antagonist (rather than anagonist) at the �2-AR, evidence has been presented thatsuggests that its hypothermic response may not be mediatedby 5-HT1A receptors (Komissarev et al., 1990). In this respectit is of interest that in a separate set of experiments it has

TABLE 2Population pharmacodynamic parameters and inter- and intra-individual variabilities of 1-PP and buspironeBuspirone was fitted assuming no interaction with 1-PP, a non-competitive interaction model and a competitive interaction model. The inter-individual coefficient of variationis displayed between brackets (%).

System-Specific Drug-SpecificIntra-Individual Error

kin A SC50 Smax n

°C/min min�1 ng/ml %

1-PP 1.34 0.0212 6.8 304 0.312 1.18 5(66) (45) (39) (88) (78) (77)

Buspirone 1.24 0.0201 5.87 17.6 0.465 0.89 9(No interaction) (44) (26) (10) (79) (19) (61)Buspirone 1.23 0.0199 5.87 15.7 0.498 0.89 10(Noncompetitive interaction) (38) (22) (112) (57) (23) (61)Buspirone 1.23 0.0207 5.88 43.3 0.613 0.89 8(Competitive interaction) (43) (31) (9) (143) (27) (63)

Fig. 4. Selected representative fits of different treatments of buspironeand 1-PP. Closed circles represent measured body temperature, the solidline represents individual prediction, and the dashed line the populationprediction for 10 mg/kg infusion of 1-PP in 15 min (C), 5 (A) and 15 (B)mg/kg buspirone in 15-min infusions. Infusions all started at t � 60.

PK-PD Modeling of Buspirone- and 1-PP-Mediated Hypothermia 1135

at ASPE

T Journals on July 29, 2018

jpet.aspetjournals.orgD

ownloaded from

been shown that the selective 5-HT1A receptor agonist WAY-100,635, when administered in doses that inhibit the effect of8-OH-DPAT in a competitive manner, does not influence thehypothermic response of 1-PP (Zuideveld et al. unpublishedobservations). This shows that 1-PP apparently induces hy-pothermia via a mechanism that is different from that ofbuspirone and other selective 5-HT1A agonists.

In the present study, the PK-PD relationships of the hypo-thermic effect for both buspirone and 1-PP are characterizedin a quantitative manner using the set-point model devel-oped for characterization of 5-HT1A receptor agonists. Al-though the mechanism of the 1-PP-induced hypothermic re-sponse is unclear, the set-point model could be appliedsuccessfully to describe the 1-PP-induced hypothermic re-sponse. This seems justified since the observed complex time-effect profile bears a great similarity to that following admin-istration of a selective 5-HT1A agonist. Furthermore, it is wellestablished that the mechanism of �2-AR-mediated hypo-thermia is very similar to that of 5-HT1A-mediated hypother-mia. Both receptors are connected to the regulation of thebody’s set-point and utilize the same effector mechanisms(Frank et al., 1997; Sallinen et al., 1997). Pharmacodynamicanalysis revealed that the hypothermic response of 1-PPcould be well characterized using this model, showing that1-PP behaves as a partial agonist with a value of the param-eter Smax of 0.312 relative to R-8-OH-DPAT. Furthermorethe potency is roughly 20 times lower than that of buspironeitself.

In the pharmacodynamic analysis, we have utilized both anon- and a competitive interaction model to describe thepossible interaction between buspirone and 1-PP. The non-competitive interaction model as proposed by Ariens andSimonis (1964) was chosen because of its simplicity, andsince the maximal stimulus of the system is known, as de-fined by the maximal response of R-8-OH-DPAT, the param-eters are identifiable. A specific feature of this model is thatit assumes the interaction to be additive, with the conse-quence that neither synergism nor antagonism can be esti-mated. Therefore, the pharmacodynamics of buspirone hasalso been characterized with a regular sigmoidal model, dis-regarding 1-PP. As can be concluded from the pharmacody-namic analysis, the parameter estimates for potency, intrin-sic activity, and the slope factor are very similar between theinteraction model and the regular sigmoidal model, suggest-ing that 1-PP hardly plays a significant role in buspirone’shypothermic response, essentially adding zero to the effect.The data was also analyzed using a competitive interactionmodel as proposed by Holford and Sheiner (1981). The majordisadvantage of this model is that essentially the slope factorn determines whether either synergism or antagonism ispredicted. Despite the fact that the parameters obtained withthe competitive interaction model deviate the most from theones found with the noncompetitive and sigmoidal model, theslope factors are relatively close to 1, and it is concluded thatthe competitive interaction model is more sensitive to a rel-atively small amount of 1-PP in the body. The peak concen-tration of 1-PP during the 15 mg/kg dose of buspirone is210 � 34 ng/ml after 60 min, which only represents a mere69% of its SC50, and therefore too low to contribute signifi-cantly. However, as buspirone is subjected to large scalepresystemic metabolism (Caccia et al., 1985), which is ex-pected given its high clearance (present investigation), it

should be anticipated that in humans higher concentrationsof 1-PP will be found after an oral administration, whichcould ultimately interfere with buspirone effect in a signifi-cant manner. In this respect it is interesting to note thatbased on the potency and its ratio with its affinity for the5-HT1A receptor [potency, 304 ng/ml]/[affinity, 94 ng/ml] �3.2 for 1-PP versus [potency, 17.6 ng/ml]/[affinity, 2.7 ng/ml]� 6.5 for buspirone, it is conceivable that 1-PP mediates itshypothermic effect through the 5-HT1A receptor.

In the present study, the PK-PD relationship of the hypo-thermic effect for both buspirone and 1-PP is characterized ina quantitative manner using the set-point model. The resultsshow that buspirone and 1-PP behave as partial 5-HT1A

agonists in vivo relative to 5-HT1A agonists such as R-8-OH-DPAT and flesinoxan. Pharmacodynamic analysis using botha sigmoidal Emax, a non- and a competitive interaction modelrevealed that in the rat, the peak 1-PP concentrationsreached are too low to contribute significantly to the buspi-rone effect after intravenous administration.

Acknowledgments

We thank Erica Tukker for technical assistance in the animalexperiments. The generous donation of buspirone and 1-PP by Bris-tol-Myers Squibb is highly appreciated.

ReferencesAkaike H (1974) A new look at the statistical model identification. IEEE (Inst Electr

Electron Eng) Trans Automatic Control 19:716–723.Ariens EJ and Simonis AM (1964) A molecular basis for drug action. The interaction

of one or more drugs with different receptors. J Pharm Pharmacol 16:289–312.Barik S and de Beaurepaire R (1998) Hypothermic effects of dopamine D3 receptor

agonists in the island of Calleja Magna. Potentiation by D1 activation. PharmacolBiochem Behav 60:313–319.

Bianchi G, Caccia S, Della VF, and Garattini S (1988) The �2-adrenoceptor antago-nist activity of ipsapirone and gepirone is mediated by their common metabolite1-(2-pyrimidinyl)-piperazine (PmP). Eur J Pharmacol 151:365–371.

Caccia S, Conti I, Vigano G, and Garattini S (1986) 1-(2-pyrimidinyl)-piperazine asactive metabolite of buspirone in man and rat. Pharmacology 32:46–51.

Caccia S, Fong MH, and Guiso G (1985) Disposition of the psychotropic drugsbuspirone, MJ-13805 and piribedil, and their common active metabolite 1-(2-pyrimidinyl)-piperazine in the rat. Xenobiotica 15:835–844.

Cao BJ and Rodgers RJ (1997) Comparative behavioural profiles of buspirone and itsmetabolite 1-(2-pyrimidinyl)-piperazine (1-PP) in the murine elevated plus-maze.Neuropharmacology 36:1089–1097.

Cryan JF, Kelliher P, Kelly JP, and Leonard BE (1999) Comparative effects ofserotonergic agonists with varying efficacy at the 5-HT1A receptor on core bodytemperature: modification by the selective 5-HT1A receptor antagonist WAY-100,635. J Psychopharmacol 13:278–283.

Dayneka NL, Garg V, and Jusko WJ (1993) Comparison of four basic models ofindirect pharmacodynamic responses. J Pharmacokinet Biopharm 21:457–478.

De Vry J (1995) 5-HT1A receptor agonists: recent developments and controversialissues. Psychopharmacology Ser (Berl) 121:1–26.

Dilsaver SC, Davidson R, and Majchrzak M (1989) Chronic treatment with amitrip-tyline produces subsensitivity to the hypothermic effects of clonidine. Prog Neu-ropsychopharmacol Biol Psychiatry 13:297–302.

Frank SM, Raja SN, Wu PK, and El-Gamal N (1997) Alpha-adrenergic mechanismsof thermoregulation in humans. Ann N Y Acad Sci 813:101–110.

Fulton B and Brogden RN (1997) Buspirone; An updated review of its clinicalpharmacology and therapeutic applications. CNS Drugs 7:68–88.

Gobert A, Newman Tancredi A, Rivet JM, Audinot V, and Millan MJ (1995) Yohim-bine is a potent, partial agonists at rat and cloned human serotonin1A receptors; acomparison to buspirone and its metabolite 1-PP. Affective Disorders Antidepres-sants P.1.047.

Gudelsky GA, Koenig JI, and Meltzer HY (1986) Thermoregulatory responses toserotonin (5-HT) receptor stimulation in the rat. Evidence for opposing roles of5-HT2 and 5-HT1A receptors. Neuropharmacology 25:1307–1313.

Hadrava V, Blier P, and Demontigny C (1996) Partial agonistic activity of R and Senantiomers of 8-OH-DPAT at 5-HT1A receptors. J Psychiatry Neurosci 21:101–108.

Higgins GA, Bradbury AJ, Jones BJ, and Oakley NR (1988) Behavioural and bio-chemical consequences following activation of 5HT1-like and GABA receptors inthe dorsal raphe nucleus of the rat. Neuropharmacology 27:993–1001.

Holford NH and Sheiner LB (1981) Pharmacokinetic and pharmacodynamic model-ing in vivo. CRC Crit Rev Bioeng 5:273–322.

Houston JB (1985) Drug metabolite kinetics, in Pharmacokinetics: Theory and Meth-odology (Rowland M and Tucker GT eds) pp 131–162, Pergamon Press, Oxford.

Jajoo HK, Mayol RF, LaBudde JA, and Blair IA (1989a) Metabolism of the anti-anxiety drug buspirone in the rat. Drug Metab Dispos 17:625–633.

1136 Zuideveld et al.

at ASPE

T Journals on July 29, 2018

jpet.aspetjournals.orgD

ownloaded from

Jajoo HK, Mayol RF, LaBudde JA, and Blair IA (1989b) Metabolism of the anti-anxiety drug buspirone in human subjects. Drug Metab Dispos 17:634–640.

Koenig JI, Meltzer HY, and Gudelsky GA (1988) 5-HT1A receptor-mediated effects ofbuspirone, gepirone and ipsapirone. Pharmacol Biochem Behav 29:711–715.

Komissarev IV, Dolzhenko AT, Obraztsova OG, and Zin’kovskya LYa (1990) Analysisof the hypothermic action of 8-hydroxy-2-(di-N-propylamino)-tetralin, and of 1-(2-pyrimidynyl)-piperazine and its derivatives. Bull Exp Biol Med 109:275–277.

Lin MT, Wu JJ, and Tsay BL (1983) Serotonergic mechanisms in the hypothalamusmediate thermoregulatory responses in rats. Naunyn Schmiedebergs Arch Phar-macol 322:271–278.

MacDonald E, Haapalinna A, Virtanen R, and Lammintausta R (1989) Effects ofacute administration of medetomidine on the behaviour, temperature and turn-over rates of brain biogenic amines in rodents and reversal of these effects byatipamezole. Acta Vet Scand Suppl 85:77–81.

Malhotra J and Gupta YK (1997) Effect of adenosine receptor modulation on penty-lenetetrazole-induced seizures in rats. Br J Pharmacol 120:282–288.

Manahan-Vaughan D, Anwyl R, and Rowan MJ (1995) The azapirone metabolite1-(2-pyrimidinyl)piperazine depresses excitatory synaptic transmission in the hip-pocampus of the alert rat via 5-HT1A receptors. Eur J Pharmacol 294:617–624.

Millan MJ, Rivet JM, Canton H, Le Marouille Girardon S, and Gobert A (1993)Induction of hypothermia as a model of 5-HT1A receptor-mediated activity in therat: a pharmacological characterization of the actions of novel agonists and antag-onists. J Pharmacol Exp Ther 264:1364–1376.

New JS (1990) The discovery and development of buspirone: a new approach to thetreatment of anxiety. Med Res Rev 10:283–326.

Sallinen J, Link RE, Haapalinna A, Viitamaa T, Kulatunga M, Sjoholm B, Mac-Donald E, Pelto HM, Leino T, Barsh GS, Kobilka BK, and Scheinin M (1997)Genetic alteration of �2C-adrenoceptor expression in mice: influence on locomotor,hypothermic, and neurochemical effects of dexmedetomidine, a subtype-nonselective �2-adrenoceptor agonist. Mol Pharmacol 51:36–46.

Salmi P and Ahlenius S (1998) Evidence for functional interactions between 5-HT1Aand 5-HT2A receptors in rat thermoregulatory mechanisms. Pharmacol Toxicol82:122–127.

Schwartz PJ, Rosenthal NE, and Wehr TA (1998) Serotonin1A receptors, melatonin,and the proportional control thermostat in patients with winter depression. ArchGen Psychiatry 55:897–903.

Spencer RL, Hruby VJ, and Burks TF (1988) Body temperature response profiles forselective mu, delta and kappa opioid agonists in restrained and unrestrained rats.J Pharmacol Exp Ther 246:92–101.

Young AH, McShane R, Park SB, and Cowen PJ (1993) Buspirone-induced hypo-thermia in normal male volunteers. Biol Psychiatry 34:665–666.

Zeisberger E (1998) Biogenic amines and thermoregulatory changes. Prog Brain Res115:159–176.

Zuideveld KP, Maas HJ, Treijtel N, Hulshof J, Der Graaf PH, Peletier LA, andDanhof M (2001) A set-point model with oscillatory behavior predicts the timecourse of 8-OH-DPAT-induced hypothermia. Am J Physiol Regul Integr CompPhysiol 281:R2059–R2071.

Zuideveld KP, Treijtel N, Maas HJ, Gubbens-Stibbe JM, Peletier LA, Der Graaf PH,and Danhof M (2002a) A competitive interaction model predicts the effect ofWAY-100,635 on the time course of R-(�)-8-Hydroxy-2-(di-n-propylamino)tetralin-induced hypothermia. J Pharmacol Exp Ther 300:330–338.

Zuideveld KP, Van Gestel A, Peletier LA, Van der Graaf PH, and Danhof M (2002b)Pharmacokinetic-pharmacodynamic modeling of the hypothermic and corticoste-rone effects of the 5-HT1A-receptor agonists flesinoxan. Eur J Pharmacol 445:43–54.

Address correspondence to: Meindert Danhof, Leiden/Amsterdam Center forDrug Research, Division of Pharmacology, Gorlaeus Laboratory, P.O. Box 9502,2300 RA Leiden, The Netherlands. E-mail: [email protected]

PK-PD Modeling of Buspirone- and 1-PP-Mediated Hypothermia 1137

at ASPE

T Journals on July 29, 2018

jpet.aspetjournals.orgD

ownloaded from