Molecular Psychiatry (2002) 7, 317–321  2002 Nature Publishing Group All rights reserved 1359-4184/02 $25.00 www.nature.com/mp ORIGINAL RESEARCH ARTICLE Significant of brain and plasma kinetics with J Tauscher1,2,4, C Jones1, G Remington2,3, RB Zipursky2,3 and S Kapur1,2,3

1Schizophrenia-PET Program, Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; 2Schizophrenia and Continuing Care Program, Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; 3Department of Psychiatry, University of Toronto, Toronto, Canada; 4Department of General Psychiatry, University of Vienna, Vienna, Austria

Keywords: antipsychotics; ; positron withdrawal from oral doses of conventional neurolep- emission tomography; occupancy; dopamine tics,8 and up to 6 months after depot injection of halo- receptors; receptors; peridol decanoate.9 Current dosing regimens of psychotropic are We measured the time course of plasma levels based on plasma kinetic considerations, although it is and brain D2 receptor occupancy in healthy volunteers unclear whether plasma levels faithfully reflect brain after a single dose of 3–4 mg (n = 4) or kinetics of drugs.1,2 To examine this, we compared the 15 mg (n = 4). We used [11C]raclopride as a kinetics of plasma levels of two widely used antipsy- PET- to estimate central D2 receptor occupancy chotics, olanzapine and risperidone, vs the time course in the striatum,1 and [11C]FLB457 for quantification of of their effects in the brain. We used positron emission extra-striatal D blockade in the thalamus.10 Addition- tomography (PET) and [11C]-labeled ligands to quantify 2 ally, we investigated the time course of plasma levels striatal and extra-striatal dopamine-2 (D2), and cortical and striatal D2 occupancy in five stable patients suffer- serotonin-2A (5-HT2A) receptor occupancy in healthy subjects after a single dose, and in patients chronically ing from schizophrenia. These patients had received a − treated for psychosis. We found a significant dis- monotherapy with olanzapine 15–20 mg day 1 (n = 3) sociation of brain and plasma kinetics. Mean plasma or risperidone 3 mg day−1 (n = 2) for at least one month. elimination half-lives of single doses of olanzapine and We scanned them immediately after withdrawal from risperidone were 24.2 and 10.3 h, respectively, whereas their , and again after 24 it took on average 75.2 h with olanzapine, and 66.6 h and 48 h. with risperidone to decline to 50% of their peak striatal In the single-dose experiments in healthy volunteers, D receptor occupancy. We found similar discrepancies 2 mean peak olanzapine plasma levels occurred 6 h after between the time course of plasma levels and extra-stri- intake (23 ng ml−1 ± 4 SD). The plasma elimination atal D2 as well as 5-HT2A receptor occupancy. Our results question the current reliance on plasma kinetics half-life time (t1/2) was 24.2 h (95% confidence interval as the main basis for dosing regimens of antipsy- (CI): 21.2–28.1). After 72 h, plasma levels declined on chotics. Studies of brain kinetics may provide a sounder basis for determining dosing schedules of psycho- tropic . Molecular Psychiatry (2002) 7, 317–321. DOI: 10.1038/ sj/mp/4001009

Novel antipsychotics are currently used to treat diverse medical conditions such as schizophrenia, mania, depression, and dementia. It is well established that antagonism of D2 receptors is critical for antipsychotic efficacy,3,4 and that at clinical doses antipsychotics 5 block a substantial proportion of D2 receptors in vivo. Positron emission tomography and selective ligands such as [11C]raclopride have provided valuable insights regarding receptor occupancy during treatment with olanzapine and risperidone.6,7 However, in most stud- ies PET scans were obtained only at a single time point Figure 1 Single dose experiments in controls. Time course after medication intake, typically 12 h post-dose. of plasma levels (olanzapine: , grey dashed line; risperidone Hence, little is known about the time course of brain plus 9-OH-risperidone: , grey solid line), and striatal D2 receptor blockade with these atypical antipsychotics. receptor occupancy (olanzapine: , black dashed line; risperi- This is of particular interest as it has been shown that done: , black solid line). All results are normalized to 100% D2 receptor occupancy can persist up to 15 days after of their peak value. Error bars denote one standard deviation. Significant dissociation of brain and plasma kinetics J Tauscher et al 318 average to 4 ng ml−1, corresponding to 19% of the peak with risperidone. In patients under steady state con-

plasma level (Figure 1). The active moiety of risperi- ditions, the brain t1/2 at striatal D2 receptors was 88.4 h done, risperidone plus 9-hydroxyrisperidone, peaked (95% CI: 61.1–159.3) for olanzapine and 45.5 h (95% −1 ± on average after 2 h (37 ng ml 14 SD), reached t1/2 CI: 44.8–46.2) for risperidone. after 10.3 h (95% CI: 8.1–14.3) and declined to 5% of The extra-striatal D2 receptor occupancy, as mea- peak levels within 72 h (Figure 1). The observed sured with [11C]FLB457 and PET in the thalamus, plasma kinetics were in close correspondence to pre- closely resembled the time course seen in the striatum. viously reported results obtained in healthy volun- Single doses of olanzapine or risperidone led to com- 11–13 14 teers and patients with schizophrenia. parably high peak D2 occupancy of extra-striatal D2 Both single doses of risperidone and olanzapine led receptors in the thalamus ranging from 72% to 82%

to high striatal D2 receptor blockade, with peak occu- and declining to 51–69% after 24 h, and 27–37% after pancy rates at the time of their respective peak plasma 72 h. Table 1 shows a synopsis of plasma levels and D2 levels ranging from 80–88%. Subsequent serial PET receptor occupancy in healthy volunteers and patients.

scans revealed sustained D2 receptor occupancy with Each of the healthy volunteers experienced sedation, both drugs in the range of 71–77% after 24 h, 56–60% which was reversible within hours. Additionally, three

after 48 h, and 36–42% after 72 h. Brain t1/2 at the stria- of the four subjects receiving risperidone showed mild tal D2 receptor was 75.2 h (95% CI: 67.9–84.4) for a sin- to moderate akathisia, which spontaneously resolved gle dose of olanzapine and 66.6 h (95% CI: 54.0–86.8) in two cases, and had to be treated with 1 mg of lora-

for risperidone. These brain receptor t1/2 values of both zepam in one case. None of the patients reported any drugs, observed after only one single dose, are substan- clinically relevant side effects at the time of the PET tially longer than the currently recommended dosing scans. interval of 24 h with both drugs. To examine whether the brain effects of antipsy- In olanzapine-treated patients, plasma-levels peaked chotics follow a similar time course on other neuro- −1 ± at 66 ng ml ( 4 SD), showed a t1/2 of 19.5 h (95% CI: transmitter receptors, and since 5-HT2A receptors have 12.4–45.9), and declined to 16 ng ml−1 (±3 SD) after been implicated in the genesis and treatment of psy- 48 h, corresponding to 25% of the peak plasma level. chosis,15 we additionally examined the time course of

After discontinuation of risperidone in patients, 5-HT2A receptor occupancy in two healthy volunteers plasma levels declined with a t1/2 of 13.4 h (95% CI: receiving a single dose of 4 mg risperidone or 15 mg 9.8–20.8) to 16% of their peak values within 48 h olanzapine. Both drugs led to a complete saturation of

(Figure 2). 5-HT2A receptors within 6 h. After 48 h risperidone −1 Continuous treatment with 15–20 mg day olanza- still led to a complete blockade of 5-HT2A receptors and pine or 3 mg day−1 risperidone resulted in high striatal olanzapine to 92% occupancy, indicating a persisting

D2 receptor blockade ranging from 81% to 88% at the cortical 5-HT2A receptor blockade, despite the fact that time of peak plasma levels. Subsequent serial PET drug plasma levels had declined to less than 10% of scans after discontinuation of the antipsychotic medi- their peak plasma level with both drugs.

cation revealed mean striatal D2 occupancy values of In summary, both single doses of olanzapine and ris- 78% with olanzapine and 66% with risperidone after peridone in healthy volunteers and continuous treat-

24 h, and after 48 h 61% with olanzapine and 42% ment in patients led to substantial and prolonged D2 and 5-HT2A receptor occupancy. Plasma drug levels declined significantly faster than brain receptor occu-

pancy. The 95% CI of brain and plasma t1/2 were dis- tinctly different and non-overlapping for both atypical antipsychotics in single-dose experiments in controls and under steady-state conditions in patients. Thus, our study convincingly demonstrates a significant dis- sociation of declining drug plasma levels and the time course of brain effects in healthy volunteers and patients, after single doses or continuous treatment with risperidone or olanzapine, and both, in striatal and extra-striatal brain regions. Several factors may account for the significant differ- ence in the time courses of drug plasma levels and brain receptor occupancy. The relation between plasma drug levels and occupancy is non-linear and can be best described by a rectangular hyperbola.1 In Figure 2 Discontinuation experiments in patients. Time the case of almost saturated receptors (Ͼ90% course of plasma levels (olanzapine: , grey dashed line; ris- peridone plus 9-OH-risperidone: , grey solid line) and stria- occupancy), a 50% reduction of drug plasma concen- tration translates into a very minor receptor occupancy tal D2 receptor occupancy (olanzapine: , black dashed line; risperidone: , black solid line). All results are normalized to decline. With lower values for both plasma concen- 100% of their peak value. Error bars denote one standard tration and occupancy—in the steeper part of the deviation. hyperbola—a similar reduction in plasma concen-

Molecular Psychiatry Significant dissociation of brain and plasma kinetics J Tauscher et al 319 Table 1 Time course of plasma levels and dopamine-2 receptor occupancy

Single Dose Experiments in Controls −1 ± Mean plasma levels (ng ml ) SD n Peak 24 h 48 h 72 h Mean t1/2 in h (95% CI) Olanzapine 4 23 ± 414± 47± 14± 0 24.2 (21.2–28.1) Risperidone 4 37 ± 14 10 ± 732± 1 10.3 (8.1–14.3) ± Mean striatal D2 occupancy SD Olanzapine 2 81% ± 1 72% ± 4 58% ± 3 42% 75.2 (67.9–84.4) Risperidone 2 87% ± 1 77% ± 0 59% 36% 66.6 (54.0–86.8)

Mean extra-striatal D2 occupancy ± SD Olanzapine 2 79% ± 4 63% ± 9 32% ± 7 60.3 (52.9–70.0) Risperidone 2 74% ± 3 51% 29% 55.7 (41.2–85.9)

Discontinuation Experiments in Patients −1 ± Mean plasma levels (ng ml ) SD n Peak 24 h 48 h Mean t1/2 in h (95% CI) Olanzapine 2 66 ± 428± 416± 3 19.5 (12.4–45.9) Risperidone 3 31 ± 12 11 ± 86± 5 13.4 (9.8–20.8) ± Mean striatal D2 occupancy SD Olanzapine 2 83% ± 2 78% ± 4 61% ± 6 88.4 (61.1–159.3) Risperidone 3 88% ± 1 72% ± 9 47% ± 16 45.5 (44.8–46.2)

tration results in a much higher decline of receptor patients with schizophrenia despite only transient D2 occupancy. This non-linear relationship might explain receptor occupancy.17 Furthermore, low-dose treat- the prolonged occupancy seen with olanzapine. How- ment with decanoate has proven clinical ever, risperidone led to substantial receptor blockade efficacy for relapse prevention in schizophrenic with almost no measurable plasma levels. patients, even when the mean D2 receptor occupancy The presence of longer acting, active metabolites, had decreased to 52% 4 weeks after injection.18 The such as 9-OH-risperidone, or altered brain metabolism clinical efficiency and other benefits of longer dosing of the parent compound may contribute to the pro- intervals, such as a substantial medication cost longed receptor occupancy.5 In addition, a relatively reduction, need to be examined. While the final answer low dissociation rate at brain receptors may lead to can only be empirically determined, the current data both drugs being attached for a relatively long time to should lead one to question the conventional approach their binding sites, even after plasma levels have of relying on plasma elimination half-lives for dosing declined.2 Furthermore, the tissue concentration of schedules of antipsychotics. We think that controlled lipophilic compounds such as olanzapine or risperi- clinical trials based on central kin- done is known to be 4- to 46-fold higher than the etics are warranted to develop innovative and rational plasma concentration in the rat.16 Therefore, the effec- dosing schedules for psychotropic medications. tive concentration at their molecular target in the human brain is likely much higher than their plasma Methods concentration. Using PET and [11C]-labeled ligands in humans in Ten healthy volunteers (four females, six males; mean vivo, we observed sustained brain receptor effects with age 31.4 ± 7.6 years SD) and five patients with a DSM- olanzapine and risperidone: (1) on D2 and 5-HT2A IV diagnosis of schizophrenia (five males; mean age receptors; (2) in different brain regions: striatum, thala- 30.5 ± 8.7 years SD) were included after they provided mus and frontal cortex; (3) after single doses in healthy written consent using forms and procedures approved volunteers; and (4) after discontinuation of these medi- by the University of Toronto Human Subjects Review cations in patients suffering from schizophrenia. All Committee. four lines of evidence call for a revision of current rec- The healthy subjects had no history of psychiatric or ommendations and practice to dose risperidone and neurologic disorder, did not suffer from a significant olanzapine once- or twice-a-day. The finding that brain medical condition or psychiatric axis I disorder, as kinetics are so prolonged is even more important given established by a structured clinical interview (SCID- recent findings that only ‘transiently’ high D2 occu- I/NP), and had not taken any psychotropic medication pancy is sufficient for obtaining17 and maintaining18 3 months prior to this study. All patients were under antipsychotic efficacy. In contrast to constant high anti- steady state conditions as they had been treated with biotic plasma levels being required for effective treat- an antipsychotic monotherapy in the same dose for at ment of bacterial infections, continuous high D2 recep- least one month prior to the PET study. Two of the tor occupancy may not be necessary for antipsychotic patients received 3 mg day−1 of risperidone and the efficacy. showed antipsychotic efficacy in three other patients were on 15–20 mg day−1 of olanza-

Molecular Psychiatry Significant dissociation of brain and plasma kinetics J Tauscher et al 320 pine. All had responded to their medication and were plasma levels in the single dose experiments with heal- regarded as clinically stable enough to tolerate a 48- thy subjects (P = 0.14). h withdrawal from their medication, during which we We determined olanzapine and risperidone plasma

assessed drug plasma levels and striatal D2 receptor levels using liquid chromatography/mass spec- occupancy. No subject had received any other antipsy- trometry.21,22 Risperidone plasma levels were meas- chotic in the last 3 months prior to this study, and none ured by means of a radioimmunoassay for the active had received any depot antipsychotics in the year moiety, which reflects both risperidone and 9-OH-ris- 22 before. peridone. To compute plasma t1/2 values, we fitted All healthy volunteers underwent a baseline PET the plasma levels to a nonlinear regression using a sin- scan, either with [11C]raclopride to obtain their gle 2-parameter exponential decay function, which can = × −b × t = unmedicated striatal D2 receptor binding potential (BP) be described as f a e (a slope of plasma level as described in detail elsewhere,7 or with [11C]FLB decline; b = rate constant for elimination from the 45710 to calculate their baseline extra-striatal BP in the plasma; t = time). For these exponential decay fits, the thalamus, or with [11C]setoperone to assess their base- r2 values ranged from 0.90 to 0.99, and all were statisti- Ͻ line 5-HT2A BP in the prefrontal cortex. Subsequently, cally significant with P-values ranging from 0.001 they were randomly assigned to a single dose of 3–4mg to 0.01. risperidone (n = 5), or 15 mg olanzapine (n = 5). Medi- cation induced side effects were rated according to the UKU Side Effect Rating Scale.19 The BP after the sub- Acknowledgements jects had taken their medication was compared to the The authors thank the patients and the volunteers for baseline value obtained before treatment was started, their participation: Doug Hussey, Kevin Cheung, Jenny and the proportion of receptors blocked by a drug Li and Armando Garcia for technical assistance; Barb (‘receptor occupancy’) was calculated using the follow- Brownlee for critical review of the manuscript; and Dr ing equation: Alan Wilson for supervising the radiochemical synth- BP during treatment eses. The research was supported in part by a grant % Receptor occupancy=ͫ1−ͩ ͪͬ×100 baseline BP from Eli Lilly Canada. Astra Zeneca provided the pre- cursor used in the synthesis of [11C]raclopride. As we had no baseline value in patients treated with antipsychotics, we used an age-corrected estimate obtained from a separate group of 12 antipsychotic References naive patients with schizophrenia and 15 age-matched 1 Farde L, Wiesel FA, Halldin C, Sedvall G. Central D2-dopamine healthy subjects. This procedure has been described in receptor occupancy in schizophrenic patients treated with antipsy- 7 detail elsewhere. Time courses of the levels of D2 chotic drugs. Arch Gen Psychiatry 1988; 45:71–76. occupancy were assessed by a series of PET scans over 2 Gifford AN, Gatley SJ, Volkow ND. Evaluation of the importance the next 4 days, using images obtained with a GE- of rebinding to receptors in slowing the approach to equilibrium 2048–15B head scanner using the striatal/cerebellar of high-affinity PET and SPECT radiotracers. Synapse 1998; 28: 11 167–175. ratio minus one, obtained with a [ C]raclopride bolus- 3 Seeman P, Lee T, Chau-Wong M, Wong K. Antipsychotic drug plus-infusion protocol, as an estimate for striatal D2 doses and neuroleptic/dopamine receptors. Nature 1976; 261: 7 BP. For the quantification of extra-striatal D2 BP in the 717–719. thalamus, we applied a simplified reference tissue 4 Seeman P, Tallerico T, Corbett R, Van Tol HH, Kamboj RK. Role 20 of dopamine D2, D4 and serotonin(2A) receptors in antipsychotic model to 60-min time-activity curves obtained with and anticataleptic action. J Psychopharmacol 1997; 11:15–17. 11 [ C]FLB 457 in the thalamus and the cerebellum. We 5 Schotte A, Janssen PF, Gommeren W, Luyten WH, Van Gompel P, merged values from both hemispheres for the striatal Lesage AS et al. Risperidone compared with new and reference antipsychotic drugs: in vitro and in vivo receptor binding. Psycho- and the extra-striatal BP quantification. The 5-HT2A scans were obtained with the use of a bolus injection pharmacol (Berl) 1996; 124:57–73. 11 6 Nyberg S, Farde L, Halldin C. A PET study of 5-HT2 and D2 dopa- of high specific activity [ C]setoperone, and 5-HT2A mine receptor occupancy induced by olanzapine in healthy sub- occupancy was determined in the prefrontal cortex as jects. 1997; 16:1–7. described in detail elsewhere.7 In analogy to the term 7 Kapur S, Zipursky RB, Remington G. Clinical and theoretical impli- ‘plasma elimination half-live’ we used the term ‘brain cations of 5-HT2 and D2 receptor occupancy of , risperi- done, and olanzapine in schizophrenia. Am J Psychiatry 1999; 156: elimination half-live’ (brain t1/2) to describe the time 286–293. needed until brain receptor occupancy declined to 8 Baron JC, Martinot JL, Cambon H, Boulenger JP, Poirier MF, Cail- 50%. In order to compute brain t1/2 values for striatal, lard V et al. Striatal dopamine receptor occupancy during and fol- extra-striatal and prefrontal regions, after normalizing lowing withdrawal from neuroleptic treatment: correlative evalu- time course occupancies to 100% of their respective ation by positron emission tomography and plasma prolactin levels. 1989; 99:463–472. peak value, we fitted the time courses of receptor occu- 9 Nyberg S, Farde L, Halldin C. Delayed normalization of central D2 pancy with olanzapine and risperidone to a linear dopamine receptor availability after discontinuation of haloperidol polynomial regression, which can be described as decanoate. Preliminary findings. Arch Gen Psychiatry 1997; 54: = + × = = 953–958. f (y0 a t (y0 intercept; a slope of occupancy = 2 10 Farde L, Suhara T, Nyberg S, Karlsson P, Nakashima Y, Hietala J decline; t time). The r values of the linear regression 11 et al. A PET-study of [ C]FLB 457 binding to extrastriatal D2-dopa- analyses ranged from 0.95 to 0.99, and their P-values mine receptors in healthy subjects and antipsychotic drug-treated were Ͻ0.05, except for the one describing olanzapine patients. Psychopharmacol (Berl) 1997; 133:396–404.

Molecular Psychiatry Significant dissociation of brain and plasma kinetics J Tauscher et al 321 11 Callaghan JT, Bergstrom RF, Ptak LR, Beasley CM. Olanzapine. 18 Nyberg S, Farde L, Halldin C, Dahl ML, Bertilsson L. D2 dopamine Pharmacokinetic and pharmacodynamic profile. Clin Pharmacoki- receptor occupancy during low-dose treatment with haloperidol net 1999; 37: 177–193. decanoate. Am J Psychiatry 1995; 152:173–178. 12 Heykants J, Huang ML, Mannens G, Meuldermans W, Snoeck E, 19 Lingjaerde O, Ahlfors UG, Bech P, Dencker SJ, Elgen K. The UKU Van Beijsterveldt L et al. The pharmacokinetics of risperidone in side effect rating scale. A new comprehensive rating scale for psy- humans: a summary. J Clin Psychiatry 1994; 55 Suppl: 13–17. chotropic drugs and a cross-sectional study of side effects in neuro- 13 Huang ML, Van Peer A, Woestenborghs R, De Coster R, Heykants J, leptic-treated patients. Acta Psychiatr Scand Suppl 1987; 334:1– Jansen AA et al. Pharmacokinetics of the novel antipsychotic agent 100. risperidone and the prolactin response in healthy subjects. Clin 20 Lammertsma AA, Hume SP. Simplified reference tissue model for Pharmacol Ther 1993; 54: 257–268. PET receptor studies. Neuroimage 1996; 4:153–158. 14 Borison RL, Diamond B, Pathiraja A, Meibach RC. Pharmacokinet- 21 Berna M, Shugert R, Mullen J. Determination of olanzapine in ics of risperidone in chronic schizophrenic patients. Psychopharm- human plasma and serum by liquid chromatography/tandem mass acol Bull 1994; 30:193–197. spectrometry. J Mass Spectrom 1998; 33:1003–1008. 15 Meltzer HY, Matsubara S, Lee JC. Classification of typical and 22 Woestenborghs R, Lorreyne W, Van Rompaey F, Heykants J. Deter- drugs on the basis of dopamine D-1, D-2 and mination of risperidone and 9-hydroxyrisperidone in plasma, urine serotonin2 pKi values. J Pharmacol Exp Ther 1989; 251:238–246. and animal tissues by high-performance liquid chromatography. J 16 Aravagiri M, Teper Y, Marder SR. Pharmacokinetics and tissue dis- Chromatogr 1992; 583:223–230. tribution of olanzapine in rats. Biopharm Drug Dispos 1999; 20: 369–377. Correspondence: J Tauscher, MD, Dept of General Psychiatry, Univer- 17 Kapur S, Zipursky R, Jones C, Shammi CS, Remington G, Seeman sity of Vienna, Wa¨hringer Gu¨ rtel 18–20, A-1050 Vienna, Austria. E- P. A positron emission tomography study of quetiapine in schizo- mail: johannes.tauscherȰakh-wien.ac.at phrenia: a preliminary finding of an antipsychotic effect with only Received 2 August 2001; accepted 9 October 2001 transiently high dopamine D2 receptor occupancy. Arch Gen Psy- chiatry 2000; 57:553–559.

Molecular Psychiatry