Human positron emission tomography studies of brain neurokinin 1 receptor occupancy by aprepitant

Human Positron Emission Tomography Studies of Brain Neurokinin 1 Receptor Occupancy by Aprepitant Mats Bergstro¨m, Richard J. Hargreaves, H. Donald Burns, Michael R. Goldberg, David Sciberras, Scott A. Reines, Kevin J. Petty, Mattias O¨gren, Gunnar Antoni, Bengt La˚ngstro¨m, Olli Eskola, Mika Scheinin, Olof Solin, Anup K. Majumdar, Marvin L. Constanzer, Wendy P. Battisti, Thomas E. Bradstreet, Cynthia Gargano, and Jarmo Hietala Background: Aprepitant is a highly selective substance P (neurokinin 1 [NK1] receptor) antagonist that significantly improves the pharmacotherapy of acute and delayed highly emetogenic chemotherapy–induced nausea and vomiting, probably through an action in the brain stem region of the central nervous system. Here, we report the use of positron emission tomography imaging with the NK1 receptor binding–selective tracer [18F]SPA-RQC to determine the levels of central NK1 receptor occupancy achieved by therapeutically relevant doses of aprepitant in healthy humans. Methods: Two single-blind, randomized, placebo-controlled studies in healthy subjects were performed. The first study evaluated the plasma concentration–occupancy relationships for aprepitant dosed orally at 10, 30, 100, or 300 mg, or placebo (n ⫽ 12). The second study similarly evaluated oral aprepitant 30 mg and placebo (n ⫽ 4). In each study, dosing was once daily for 14 consecutive days. Data from both studies were combined for analyses. The ratio of striatal/cerebellar [18F]SPA-RQ (high receptor density region/reference region lacking receptors) was used to calculate trough receptor occupancy 24 hours after the last dose of aprepitant. Results: Brain NK1 receptor occupancy increased after oral aprepitant dosing in both a plasma concentration–related (r ⫽ .97; 95% confidence interval [CI] ⫽ .94 –1.00, p ⬍ .001) and a dose-related (r ⫽ .94; 95% CI ⫽ .86 –1.00, p ⬍ .001) fashion. High (ⱖ90%) receptor occupancy was achieved at doses of 100 mg/day or greater. The plasma concentrations of aprepitant that achieved 50% and 90% occupancy were estimated as approximately 10 ng/mL and approximately 100 ng/mL, respectively. Conclusions: Positron emission tomography imaging with [18F]SPA-RQ allows brain NK1 receptor occupancy by aprepitant to be predicted from plasma drug concentrations and can be used to guide dose selection for clinical trials of NK1 receptor antagonists in central therapeutic indications. Key Words: Substance P, depression, neurokinin 1 receptor (NK1), aprepitant

N

eurokinin 1 (NK1) receptors and their endogenous ligand, substance P, are found in brain regions that are critical for the regulation of the vomiting reflex (e.g., brain stem nuclei of the dorsal vagal complex) and are also highly expressed in brain regions that are important for the regulation of affective behavior and neurochemical responses to stress (e.g., nucleus accumbens, striatum, amygdala, septum, hippocampus, hypothalamus, locus coeruleus, raphe nucleus, and periaqueductal gray matter) (Arai and Emson 1986; Brodin et al 1987; Hokfelt et al 1987; Mantyh et al 1984; Pioro et al 1990). Aprepitant (EMEND, also known as MK-869 and L-754030; Merck & Co., West Point, Pennsylvania,) is a highly selective substance P (NK1) receptor antagonist (IC50 ⫽ 100 pM) that crosses the blood– brain barrier. Aprepitant is chemically described as 5-[[(2R,3S]-2-[(1R)-1-[3,5-bis(trifluoromethyl) phenyl]ethoxy]-3(4-fluorphenyl)-4-morpholinyl]methyl]-1,2-dihydro-3H-1,2,4-triazol-3-one, with a molecular weight of 534.43 d. The current market version of aprepitant contains the drug as a nanoparticle formulation. During early development of aprepitant, tablet formulations were used. Due to these changes in formulation From the Uppsala Positron Emission Tomography Centre (MB, MO¨, GA, BL), Uppsala, Sweden; Merck Research Laboratories (RJH, HDB, MRG, DS, SAR, KJP, AKM, MLC, WPB, TEB, CG), West Point, Pennsylvania; Turku Positron Emission Tomography Centre (OE, OS, JH); and Department of Psychiatry (JH) and Pharmacology and Clinical Pharmacology (MS), Turku University, Turku, Finland. Address reprint requests to Richard Hargreaves, Ph.D., Merck & Co. Inc., P.O. Box 4, 770 Sumneytown Pike, WP42-213, West Point, PA 19486. Received August 27, 2003; revised February 3, 2004; accepted February 4, 2004.

0006-3223/04/$30.00 doi:10.1016/j.biopsych.2004.02.007

during development, all measurements were primarily linked to plasma concentration rather than dose. Recent studies have demonstrated that treatment with aprepitant significantly improves preventive pharmacotherapy of acute and delayed highly emetogenic chemotherapy–induced nausea and vomiting (Campos et al 2001; Chawla et al 2003; Hesketh et al 2003; Navari et al 1999; Patel and Lindley 2003; Poli-Bigelli et al 2003; Van Belle et al 2002) through an action at central sites (Huskey et al 2003; Rupniak et al 1997). Furthermore, early clinical studies, the results of which were not subsequently reproduced in phase III clinical trials, suggested that substance P (NK1 receptor) antagonists might have the potential to relieve the symptoms of depression (Kramer et al 1998). Preclinical pharmacologic, pharmacodynamic, and radioligand binding studies suggest that high central nervous system (CNS) NK1 receptor occupancy is required to block central substance P–mediated behaviors (Duffy et al 2002; Merck Research Laboratories, unpublished data). Until now, however, it has not been possible to measure NK1 receptor occupancy directly in the brain of living humans. Such information would be extremely valuable to many decision-making processes during the development of centrally acting drugs targeted at substance P, to ensure adequate proof-of-concept testing and to optimize dosing regimens. In humans, positron emission tomography (PET) with specific radiolabeled ligands has been used to quantitatively assess in vivo occupancy of a number of neurotransmitter receptors, including those for dopamine, serotonin, and benzodiazepines (Talbot et al 2002). The recent development and validation of [18F]SPA-RQ ([[18F] 2-fluoromethoxy-5-(5-trifluoromethyl-tetrazol-1-yl)-benzyl]([2S,3S]2-phenyl-piperidin-3-yl)-amine) (Solin et al, in press) in preclinical and clinical models has established this tracer as a highly selective, high-affinity (67 pM), NK1 receptor ligand that binds specifically and reversibly and can be used BIOL PSYCHIATRY 2004;55:1007–1012 © 2004 Society of Biological Psychiatry

1008 BIOL PSYCHIATRY 2004;55:1007–1012

M. Bergstrom et al

Table 1. Percent Occupancy of Tachykinin NK1 Receptors in the Striatum, as Estimated by Positron Emission Tomography Subject No. Study 1 02 09 01 03 08 07 10 04 12 06 05 11 Study 2 03 01 02 04

Aprepitant Dose (mg/day)a

Plasma Aprepitant Concentration (ng/mL)b

Striatal NK1-Receptor Occupancy (%)

300 300 100 100 100 30 30 30 10 10 0 0

3565 436 1053 502 75 58 47 37 19 14.6 Placebo Placebo

102 96 94 95 91 82 81 78 63 65 9 ⫺1

30 30 30 0

40 27 18 Placebo

73 67 59 11

NK1, neurokinin 1; PET, positron emission tomography. a The 10, 30, and 100-mg tablet formulations used in this study provide trough plasma concentrations very similar to those obtained with the marketed capsule formulation. b Average of pre-PET and post-PET scan values on day 15.

successfully in PET imaging (Hargreaves 2002). In monkey PET studies, Bergstro¨m et al (2000) have also imaged substance P receptor expression with 11C-labeled radiotracer. The objective of the present studies was to use [18F]SPA-RQ tracer and PET imaging in healthy humans to estimate the levels of CNS NK1 receptor occupancy produced by therapeutically relevant plasma levels of aprepitant, to guide dose selection for trials in which the potential therapeutic action of aprepitant was centrally mediated.

Methods and Materials Subjects Healthy male subjects, aged between 18 and 45 years, were enrolled at two study sites (Uppsala University, Uppsala, Sweden and University of Turku, Turku, Finland). History, physical examination, and laboratory assessments were conducted within 3 weeks before study initiation to assess health status. All volunteers were medication free for at least 2 weeks before the

initiation of the study and were nonsmokers for at least 6 months before the study. The studies were reviewed and approved by the ethics review committee and medical product agency and isotope committee of each participating institution and were conducted in conformance with applicable requirements regarding ethics committee review, informed consent, and other statutes or regulations regarding the protection of the rights and welfare of human subjects participating in biomedical research. Study Design Both studies were of an independent-group, single-blind, randomized, placebo-controlled design. The first study was conducted at Uppsala University and evaluated four oral doses of aprepitant (10, 30, 100, and 300 mg; Table 1) and placebo tablets administered once daily for 14 consecutive days in 12 subjects. The second study was conducted at the University of Turku and evaluated oral aprepitant 30 mg and placebo tablets administered once daily for 14 consecutive days in four subjects. (It should be noted that the tablet formulation of aprepitant used in these studies was different from the currently marketed capsule formulation.) Analyses focused on establishing the relationship between plasma concentration of aprepitant and central NK1 receptor occupancy after 14 days of chronic dosing, because it was expected that after this time the brain-to-plasma equilibrium would be relatively independent of oral dosage form. Each subject in each study was randomized to one dose group or to placebo. All doses in both studies were given within 30 min after a light breakfast. After 14 doses, the administration of aprepitant or placebo was discontinued and the PET studies conducted 24 hours later. The PET studies were conducted 24 hours after the last dose to simulate trough occupancies that might be expected with a daily dosing regime. No nonstudy medication was allowed throughout the studies. PET Tracer Synthesis The NK1 selective PET tracer [18F]SPA-RQ (“1” in Figure 1) was prepared by alkylation of a tert-butoxycarbonyl (BOC)-protected precursor (“3” in Figure 1) via alkylation of the phenolate anion with [18F]bromofluoromethane followed by acid-catalyzed removal of the protecting group and purification by high-performance liquid chromatography (HPLC). After removal of the HPLC eluent by evaporation on a roto-evaporator, the tracer was formulated (sterile 5% D-glucose solution containing 1% ethanol with pH-adjusted to 6.5–7.6 with .1 mol/L phosphate buffer) then sterilized by filtration through a .2-␮m filter into a sterilized injection vial. The details and results of the [18F]SPA-RQ production are presented in full detail elsewhere (Solin et al, in press).

Figure 1. Route of radiochemical synthesis and structure of [18F]SPA-RQ. h-IC50, binding affinity pM; IC50, concentration producing 50% inhibition of the binding of 125I substance P at human NK1 receptors; Log P, the octanol:aqueous buffer pH 7.4 partition coefficient; [18F]SPA-RQ, substance P antagonist receptor quantifier.

www.elsevier.com/locate/biopsych

M. Bergstrom et al In both studies, specific activity ranged from 20 to 3681 GBq/ ␮mol (average ⫽ 1207 ⫾ 894), and radiochemical purity was greater than 95%. The tracer was used within 60 min of final formulation, and there was no change in radiochemical purity during that time. PET Imaging and Analysis Positron emission tomography scans for study 1 were performed at the Uppsala University PET Center with a GE 4096 whole-body PET camera (General Electric Medical Systems, Milwaukee, Wisconsin), which simultaneously generates 15 tomographic slices with a slice separation of 6.5 mm, an in-plane resolution of approximately 5 mm, and an axial coverage of 10 cm. Before each injection of radioactivity, a 10-min transmission scan was performed with a rotating 68Ge-rod for attenuation correction. Each subject was then given an IV dose of [18F]SPARQ (approximately 130 –200 MBq with specific radioactivities of at least 20 GBq/␮mol) by rapid bolus injection of the tracer followed by approximately 10 mL saline to flush the IV lines. Brain images were acquired for 90 min with 1– 6-min frames, followed by additional dynamic scans consisting of 10-min frames from 120 to 160 min and then 190 to 240 min. For study 2, which was conducted at the Turku PET Center, the scanning procedures were quite similar, except for the following: a GE Advance whole-body camera (GE Medical Systems) was used, which simultaneously generates 35 transverse tomographic slices with a slice separation of 4.25 mm, an in-plane resolution of approximately 4 mm, and an axial coverage of 15.2 cm. Brain images were acquired for 87 min, including 3 ⫻ 1-min, 4 ⫻ 3-min, and 12 ⫻ 6-min frames, plus additional scans consisting of 10-min frames from 120 to 180 min, 210 to 270 min, and 300 to 360 min. At each center, two PET studies were conducted on each subject: the first within 2 weeks before the first dose of aprepitant and the second approximately 24 hours (⫾ 3 hours) after the last dose of aprepitant (i.e., on day 15). After each PET study, images were reconstructed and realigned to correct for positioning between baseline and posttreatment scans and for movement during scans. For study 1, regions of interest were drawn for various regions of the brain on summed PET images (0 –240 min) then applied to each frame to provide average radioactivity concentrations (standardized uptake values [SUVs] and time–activity curves [TACs]). For study 2, regions of interest were drawn on magnetic resonance images resliced according to PET slices (Hietala et al 1995) and transferred to the PET images. Regions of interest were delineated on three or four consecutive slices. Regional time– radioactivity curves were generated with in-house software. The primary analysis of each PET scan focused on the striatum as a surrogate for all NK1 receptor– containing brain regions. The striatum (caudate and putamen) was chosen because it has the highest concentration of NK1 receptors in humans (Carberlotto et al 2003), is relatively large, and provides the largest specific signal with this tracer in vivo. The cerebellum was chosen as a reference because the concentration of NK1 receptors in this tissue is very low (Carberlotto et al 2003). In addition, postmortem autoradiographic studies with [18F]SPA-RQ showed no displaceable binding radioactivity in the cerebellum at a concentration of a potent NK1 antagonist (GR203040) that blocked binding in the caudate, putamen, and occipital cortex almost completely (Hietala et al, unpublished data). It is also noteworthy that the current results show a posttreatment/pretreatment ratio of .99 ⫾ .14 for the cerebellum (190 –240 min

BIOL PSYCHIATRY 2004;55:1007–1012 1009 postinjection), which indicates that, if present, the density of NK1 receptors in the cerebellum is below detection limits for this tracer. The frontal cortex and the medial and lateral temporal cortices were also defined as regions of interest, but data for these regions were not used in the present analysis. The displacement of tracer binding by aprepitant in the striatum (percent occupancy) was determined by comparing the predose and postdose PET scans for each subject according to the linear method (Wong et al 1986). In this method, the rate of increase (slope) of the striatum/cerebellum SUV ratio over time provides an index that is inversely proportional to free receptor concentration. Thus, as occupancy increases, this slope decreases. The SUV (radioactivity concentration in tissue [Bq/mL] ⫻ weight [kg]/dose [Bq] ⫻ 1000 mL/kg) is a unitless parameter that is used to correct for differences in injected dose and body weight of the subject, making it possible to compare between studies and subjects. The slopes of the TACs for striatum and cerebellum (in SUV units) were obtained independently by computing a linear regression over the range from .1 to 1.5 hours after dosing the tracer, for both the baseline and postdose scans. The “linear” method of analysis is reported here because it is the method that was used to generate the PET data set that was submitted to the U.S. Food and Drug Administration Advisory Committee review of aprepitant for the prevention of acute and delayed chemotherapy-induced nausea and vomiting (March 6, 2003; https:// www.fda.gov.ohrms/dockets/ac/03/slides/3928s1.htm). Aprepitant Plasma Concentrations Blood was collected for measurement of plasma aprepitant concentrations on day 1 and at pre- and postscan on day 15 (approximately 24 hours after the last dose on day 14). Blood samples were collected directly into tubes containing ethylenediaminetetraacetic acid as an anticoagulant. The samples were processed immediately by centrifugation at 0 –5°C. After centrifugation, the plasma was immediately separated and stored frozen at ⫺20°C. Long-term stability studies established that aprepitant was stable in plasma at ⫺20°C for at least 10 months (Merck & Co. Inc.: data on file). All samples from the PET studies were analyzed within this stability window. The analytic method for the determination of aprepitant concentrations was based on a liquid–liquid extraction of analyte and internal standard from basified biological matrix. The plasma sample was extracted with methyl-t-butyl ether. The analytes were chromatographically separated by HPLC and detected by triple quadrupole atmospheric pressure chemical ionization mass spectrometry. The mobile phase for HPLC was 50/50 vol/vol percent acetonitrile: water containing 10 mmol/L ammonium acetate, .1% formic acid. The HPLC analytic column was BDS-Hypersil (Keystone Scientific Operations, Bellefonte, Pennsylvania), C8-3 ␮m, 50 ⫻ 4.6 mm. The mobile phase flow rate was 1.0 mL/min. Run time was 6 min. Representative retention times were 3.4 min for drug and 3.1 min for internal standard. The mass spectrometer was operated in multiple reaction monitoring mode at unit resolution monitoring the precursor3product ion combinations of 5353277 and 5033259 for aprepitant and internal standard, respectively. The lower limit of quantification for plasma assay was 10 ng/mL. The linear range for the plasma assay was 10 –5000 ng/mL. Safety Safety and tolerability were assessed by vital signs, 12-lead electrocardiograms, physical examinations, and laboratory safety evaluations performed at prestudy, day 1, and poststudy. Study www.elsevier.com/locate/biopsych

1010 BIOL PSYCHIATRY 2004;55:1007–1012

M. Bergstrom et al

subjects were also monitored for clinical adverse experiences throughout the study. Study Design These studies were designed to examine the relationship between the plasma aprepitant concentrations, spanning a range that would be achieved at clinically efficacious doses, and the corresponding NK1 receptor occupancy. The relationships between aprepitant dose and striatal NK1 receptor occupancy, and between aprepitant plasma concentration (average of pre- and postscan concentrations on day 15) and striatal NK1 receptor occupancy, were examined with data combined from both study centers, in Uppsala and Turku. Spearman’s rank-order correlation coefficient, r, was used to measure the association between each aprepitant dose (mg) and plasma concentration (ng/mL) with striatal NK1 receptor occupancy (percent) because prior PET data were not available and it could not be assumed that the data sampled were from a bivariate normal population. An additional post hoc analysis was performed to investigate the relationship between plasma concentration of aprepitant and NK1 receptor occupancy. The NK1 receptor occupancy and plasma concentration data were modeled with a form of the Hill equation (Kenakin 1998) under the assumptions that the maximum occupancy of NK1 receptors by aprepitant is 100%, the minimum occupancy is 0%, and the slope factor gamma (␥) is 1. Confining the slope factor to 1 assumes a single binding site. This reduces the more general Hill equation from:

E 共 Occ 兲 ⫽

x

␥

100x ␥ ⫹ 共 Occ 50 兲 ␥

to

E 共 Occ 兲 ⫽

100x . x ⫹ 共 Occ 50 兲

Iterative least-squares minimization with the Newton-Raphson algorithm was used to estimate the plasma concentration of aprepitant that results in occupancy of 50% of the NK1 receptors (Occ50). A subsequent analysis was similarly performed but additionally estimated the slope factor. The placebo subjects were not included in any of the analyses.

Figure 2. Positron emission tomography (PET) image from a subject who received aprepitant 100 mg. Predose (top) and postdose (bottom) PET scans at the level of the cerebellum (left) and striatum (right). Subject number 01, estimated occupancy ⫽ 94%.

therefore combined to produce a single striatal SUV. Similarly, the results for the individual cerebral cortices were combined to produce a single cerebral cortical SUV. The TAC showed that the mean SUV reached a stable maximum level at 2 hours after administration of [18F]SPA-RQ. The highest level of [18F]SPA-RQ uptake was in the striatum, and the lowest level was in the cerebellum. The uptake in the cortex was intermediate to these two areas. Because the striatum produced the strongest signal (Figure 2), this area provided the greatest sensitivity to differentiate levels of occupancy achieved by different doses of aprepitant. The ratio of striatal/cerebellar

Results Demographics In the first dose-ranging study, 12 healthy, white, male subjects (mean age ⫽ 26 years [range: 23– 41], mean height ⫽ 180 cm [range: 167–191], mean weight ⫽ 75 kg [range: 66 –92]) were randomized to receive either aprepitant 10 mg/day (n ⫽ 2), 30 mg/day (n ⫽ 3), 100 mg/day (n ⫽ 3), 300 mg/day (n ⫽ 2), or placebo (n ⫽ 2). In the second study, four healthy, white, male subjects (mean age ⫽ 24 years [range: 23–25], mean height ⫽ 180 cm [range: 173–186], mean weight ⫽ 77 kg [range: 72– 83]) were randomized to receive either aprepitant 30 mg/day (n ⫽ 3) or placebo (n ⫽ 1). Data from all 16 subjects were included in the analyses of pharmacokinetic and PET data. Image Analysis The SUV was calculated for each subject at each time point after injection of [18F]SPA-RQ for the prespecified regions of interest: caudate, putamen, frontal cortex, medial and lateral temporal cerebral cortices, and cerebellum. The SUVs in the caudate and putamen were similar to each other and were www.elsevier.com/locate/biopsych

Figure 3. Estimated relationship between plasma concentration of aprepitant and occupancy of striatal NK1 receptors. Curve depicted is based on fit of the data to the Hill equation (slope ⫽ 1). NK1, neurokinin 1.

M. Bergstrom et al [18F]SPA-RQ uptake was therefore used to calculate NK1 receptor occupancy. Relationship of Dose and Plasma Concentration to Striatal NK1 Receptor Occupancy Striatal NK1 receptor occupancy by aprepitant increased in both a dose- and concentration-dependent fashion (Table 1). The correlation between dose of aprepitant and NK1 receptor occupancy was estimated to be .94 (p ⬍ .001; 95% confidence interval [CI] ⫽ .86 –1.00) and between plasma concentration of aprepitant and occupancy was .97 (p ⬍ .001; 95% CI ⫽ .94 –1.00). High (ⱖ90%) NK1 receptor occupancy was achieved in all five subjects who received at least 100 mg/day aprepitant. The three subjects given placebo showed essentially no striatal NK1 receptor occupancy. The occupancies in the placebo-treated subjects were ⫺1%, 9%, and 11%, respectively), which indicates the test-to-retest variability with [18F]SPA-RQ for the present study and analysis methods. The post hoc Hill equation analysis (minimum ⫽ 0%, maximum ⫽ 100%, slope factor ⫽ 1) indicated a good fit to the data (Figure 3). On the basis of this analysis, it was estimated that aprepitant plasma concentrations of approximately 10 ng/mL and approximately 100 ng/mL produce NK1 receptor occupancies in the brain of 50% and 90%, respectively. Further modeling that included estimating the slope factor provided corresponding estimates of approximately 10 ng/mL and approximately 127 ng/mL. These are similar to the estimates obtained when the slope factor was assumed to be 1. The Hill slope factor estimate was .86, which supports the assumption that aprepitant interacts with a single binding site in the brain. Figure 2 shows pre- and posttreatment PET scans at the level of the cerebellum and the striatum for one subject administered aprepitant 100 mg/day. At baseline, [18F]SPA-RQ uptake in the cerebellum was very low, consistent with the near absence of NK1 receptors in this area (Carberlotto et al 2003). In contrast, at baseline, the striatum (and cortical areas) showed high levels of tracer uptake (Figure 2, top). Posttreatment images after administration of aprepitant 100 mg/day showed markedly reduced binding in the striatum and cortex, indicative of high (approximately 94% [median]) levels of NK1 receptor occupancy, with nonspecific binding in the cerebellum at a similarly low level to the baseline scan. Safety All subjects who were dosed were included in the assessment of safety and tolerability for each study. No subjects discontinued from the study, and there were no serious clinical, laboratory, or other adverse experiences. All reported adverse experiences were mild or moderate in intensity, short-lasting, and typical of those reported by healthy subjects in phase I studies.

Discussion These are the first studies to demonstrate a clinically useful method of assessing NK1 receptors and their occupancy by an NK1 receptor antagonist in the human brain in vivo. We used the high affinity and specificity of a novel PET ligand, [18F]SPA-RQ (Hargreaves 2002; Solin et al, in press), to assess the strength of the relationship between both dose and plasma concentration of aprepitant (a high-affinity, reversibly binding NK1 receptor antagonist) with receptor occupancy, and then to estimate the conjectured nonlinear logistic function relationship between plasma concentration and receptor occupancy, as described by a Hill equation. Greater than 90% NK1 receptor occupancy in the

BIOL PSYCHIATRY 2004;55:1007–1012 1011 striatum of healthy subjects occurred after daily administration of aprepitant doses of 100 mg or 300 mg for 2 weeks. Similarly, it was observed that plasma aprepitant concentrations greater than 100 ng/mL result in at least 90% occupancy. Based on plasma concentrations achieved with the capsule formulation of aprepitant used to treat chemotherapy-induced nausea and vomiting (approximately 500 ng/mL at 24 hours after each dose), a high level of occupancy (⬎90%) is expected to be achieved with the recommended and highly effective antiemetic regimen (125 mg on day 1 of chemotherapy, followed by 80 mg per day on the next 2 days). On the other hand, on the basis of plasma concentrations, a less effective antiemetic regimen of 40 mg on day 1 followed by 25 mg on days 2 through 5 that was evaluated in a phase II study (Chawla et al 2003) would be expected to result in approximately 80% receptor occupancy. Thus, maximal antiemetic efficacy of aprepitant is predicted to be associated with greater than 90% occupancy of brain NK1 receptors by aprepitant. A previous phase II study showed that a tablet dose of 300 mg/day of aprepitant was a well-tolerated and effective antidepressant. It improved symptoms of depression to a similar extent as the selective serotonin receptor inhibitor paroxetine (20 mg/day) and to a significantly greater extent than placebo (Kramer et al 1998). The trough plasma concentrations of aprepitant predicted to be achieved with this dose regimen at steady state (approximately 2000 ng/mL) would also be expected to produce greater than 90% receptor occupancy; however, recent results from phase III clinical trials indicate that aprepitant is not effective for the treatment of depression (Merck & Co. Inc., November 12, 2003: public announcement; http://pubaff. merck.com/media/111203.html). Striatal receptor occupancy by aprepitant was correlated with both the dose and plasma concentration of drug administered. Analysis of occupancy in smaller brain regions, such as those thought to be involved in emesis and depression, is somewhat confounded by lower tracer signal and the anatomic resolution of the PET image. Subsequent detailed secondary analysis of the three subjects dosed with 30 mg aprepitant and one subject dosed with placebo showed parallel displacement of [18F]SPARQ in the striatum, brain stem, and amygdala, thereby indicating that the striatum was likely to be a suitable surrogate for brain regions that could be involved in the therapeutic effects of aprepitant. The present study indicates that at the doses tested, a Hill equation model fits the data reasonably well, which suggests that NK1 receptor occupancy might be predicted from the plasma concentration of aprepitant. The model also indicated that approximately 50%–90% receptor occupancy was achieved over an approximately 10-fold range of plasma total drug concentration. This study involved only a small number of subjects and a limited range of doses and plasma concentrations; however, there was a good correlation between degree of receptor occupancy and plasma concentrations over the range achieved by clinically effective doses of aprepitant for the prevention of acute and delayed chemotherapy-induced nausea and vomiting. The description of this relationship is valuable to the development of aprepitant for CNS indications, because it can be used to speed and improve objective decision making in proof-of-concept clinical trials by helping to guide dose selection. This is especially valuable for clinical trials of long duration, in which errors in dose selection can have a major impact by prolonging drug development timelines, as well as in trials that produce negative results, because the PET occupancy data can confirm that www.elsevier.com/locate/biopsych

1012 BIOL PSYCHIATRY 2004;55:1007–1012 targeted occupancies were achieved and that the efficacy trial adequately tested clinical hypotheses. Indeed, the PET data with [18F]SPA-RQ were central to the decision-making processes to terminate the phase III trials of aprepitant for the treatment of depression because the doses used and plasma levels that were achieved in the trials were likely to have produced very high receptor occupancies that effectively blocked more than 95% of CNS NK1 receptors around the clock. This suggests that the right dose and dosing interval were selected for phase III trials. Thus, the concept of NK1 antagonism for the treatment of depression remains unproven as it was not supported by our studies with aprepitant. [18F]SPA-RQ will be a useful tool to continue to explore the pharmacology of NK1 antagonists as we search for other clinical utilities for this class of compounds. These studies were funded by Merck Research Laboratories, West Point, Pennsylvania. We thank the staff of the Positron Emission Tomography Laboratories at Uppsala and Turku for their help; Dr. Kevin Gingrich for oversight of some of these studies; and Dr. Christopher Lines for assistance with preparing the manuscript. Arai H, Emson PC (1986): Regional distribution of neuropeptide K and other tachykinins (neurokinin A, neurokinin B and substance P) in rat central nervous system. Brain Res 399:240 –249. Bergstro¨m M, Fasth K-J, Kilpatrick G, Ward P, Cable KM, Wipperman MD, et al (2000): Brain uptake and receptor binding of two (11C)-labeled selective high affinity NK1-antagonists, GR203040 and GR205171—PET studies in rhesus monkey. Neuropharmacology 39:664 –670. Brodin E, A˜gran SO, Theodorsson-Norheim E (1987): Effects of subchronic treatment with imipramine, zimelidine and alaproclate on regional tissue levels of substance P and neurokinin A/neurokinin B-like immunoreactivity in the brain and spinal cord of the rat. Neuropharmacology 26:581–590. Campos D, Pereira JR, Reinhardt RR, Carracedo C, Poli S, Vogel C, et al (2001): Prevention of cisplatin-induced emesis by the oral neurokinin-1 antagonist, MK-869, in combination with granisetron and dexamethasone or with dexamethasone alone. J Clin Oncol 19:1759 –1767. Carberlotto L, Hurd YL, Murdock P, Wahlin JP, Melotto S, Corsi M, et al (2003): Neurokinin 1 receptor and relative abundance of the short and long isoforms in the human brain. Eur J Neurosci 17:1736 –1746. Chawla SP, Grunberg SM, Gralla RJ, Hesketh PJ, Rittenberg C, Elmer ME, et al (2003): Establishing the dose of the oral NK1 antagonist aprepitant for the prevention of chemotherapy induced nausea and vomiting. Cancer 97:2290 –2300. Duffy RA, Varty GB, Morgan CA, Lachowicz JE (2002): Correlation of neurokinin (NK) 1 receptor occupancy in gerbil striatum with behavioral effects of NK1 antagonists. J Pharmacol Exp Ther 301:536 –542. Hargreaves R (2002): Imaging substance P receptors (NK1) in the living

www.elsevier.com/locate/biopsych

M. Bergstrom et al human brain using positron emission tomography. J Clin Psychiatry 63(suppl 11):18 –24. Hesketh PJ, Van Belle S, Aapro M, Tattersall FD, Naylor RJ, Hargreaves R, et al (2003): Differential involvement of neurotransmitters through the time course of cisplatin-induced emesis as revealed by therapy with specific receptor antagonists. Eur J Cancer 39:1074 –1080. Hietala J, Syva¨lahti E, Vuorio K, Ra¨kko¨la¨inen V, Bergman J, Haaparanta M, et al (1995): Presynaptic dopamine function in striatum of neuroleptic-naive schizophrenic patients. Lancet 346:1130 –1131. Ho¨kfelt T, Johansson O, Holets V, Meister B, Melander T (1987): Distribution of neuropeptides with special reference to their coexistence with classical transmitters. In: Meltzer HY, editor. Psychopharmacology: The Third Generation of Progress. New York: Raven Press, 401–416. Huskey SW, Dean BJ, Bakhtiar R, Sanchez RI, Tattersall FD, Rycroft W, et al (2003): Brain penetration of aprepitant, a substance P receptor antagonist, in ferrets. Drug Metab Dispos 31:785–791. Kenakin TP (1998): Pharmacologic Analysis of Drug-Receptor Interactions, 3rd ed. New York: Wiley-Interscience. Kramer MS, Cutler N, Feighner J, Shrivastava R, Carman J, Sramek JJ, et al (1998): Distinct mechanism for antidepressant activity by blockade of central substance P receptors. Science 281:1640 –1645. Mantyh PW, Hunt SP, Maggio JE (1984): Substance P receptors: Localization by light microscopic autoradiography in rat brain using [3H]SP as the radioligand. Brain Res 307:147–165. Navari RM, Reinhardt RR, Gralla RJ, Kris MG, Hesketh PJ, Khojasteh A, et al (1999): Reduction of cisplatin-induced emesis by a selective neurokinin1-receptor antagonist. N Engl J Med 340:190 –195. Patel L, Lindley C (2003): Aprepitant—a novel NK1 receptor antagonist. Expert Opin Pharmacother 4:2279 –2296. Pioro EP, Mai JM, Cuello AC (1990): Distribution of substance P-and enkephalin-immunoreactive neurons and fibers. In: Paxinos G, editor. The Human Nervous System. New York: Academic Press, 1051–1094. Poli-Bigelli S, Rodrigues-Pereira J, Carides AD, Guoguang-Ma J, Eldridge K, Hipple A, et al (2003): Addition of the oral NK1 antagonist aprepitant to standard antiemetic therapy improves control of chemotherapy induced nausea and vomiting: Results from a randomized, double-blind, placebo controlled trial in Latin America. Cancer 97:3090 –3098. Rupniak NM, Tattersall FD, Williams AR, Rycroft W, Carlson EJ, Cascieri MA, et al (1997): In vitro and in vivo predictors of the anti-emetic activity of tachykinin NK1 receptor antagonists. Eur J Pharmacol 326:201–209. Solin O, Eskola O, Hamill T, Bergman J, Lehikoinen P, Gro¨nroos T, et al (in press): Synthesis of and characterisation of a potent, selective, radiolabeled substance-P antagonist (SPA) for NK1 receptor quantitation: ([18F]SPA-RQ). Mol Imaging Biol. Talbot P, Laruelle M (2002): The role of in vivo molecular imaging with PET and SPECT in the elucidation of psychiatric drug action and new drug development. Eur Neuropsychopharmacol 12:503–511. Van Belle S, Lichinitser MR, Navari RM, Garin AM, Decramer ML, Riviere A, et al (2002): Prevention of cisplatin-induced acute and delayed emesis by the selective neurokinin-1 antagonist, L-758,298 and MK-869. Cancer 94:3032–3041. Wong DF, Gjedde A, Wagner HN Jr (1986): Quantification of neuroreceptors in the living human brain. I. Irreversible binding of ligands. J Cereb Blood Flow Metab 6:137–146.