Abstract
We compared the neurokinin 1 receptor (NK1R) antagonists aprepitant, CP-99994 [(2S,3S)-3-(2-methoxybenzylamino)-2-phenylpiperidine], and ZD6021 [3-cyano-N-((2S)-2-(3,4-dichlorophenyl)-4-[4-[2-(methyl-(S)-sulfinyl)phenyl]piperidino]butyl)-N-methyl]napthamide]] with respect to receptor interactions and duration of efficacy in vivo. In Ca2+ mobilization assays (fluorometric imaging plate reader), antagonists were applied to human U373MG cells simultaneously with or 2.5 min before substance P (SP). In reversibility studies, antagonists were present for 30 min before washing, and responses to SP were repeatedly measured afterward. The compounds were administered i.p. to gerbils, and the gerbil foot tap (GFT) response was monitored at various time points. The NK1R receptor occupancy for aprepitant was determined in striatal regions. Levels of compound in brain and plasma were measured. Antagonists were equipotent at human NK1R and acted competitively with SP. After preincubation, aprepitant and ZD6021 attenuated the maximal responses, whereas CP-99994 only shifted the SP concentration-response curve to the right. The inhibitory effect of CP-99994 was over within 30 min, whereas for ZD6021, 50% inhibition still persisted after 60 min. Aprepitant produced maximal inhibition lasting at least 60 min. CP-99994 (3 μmol/kg) inhibited GFT by 100% 15 min after administration, but the effect declined rapidly together with brain levels thereafter. The efficacy of ZD6021 (10 μmol/kg) lasted 4 h and correlated well with brain levels. Aprepitant (3 μmol/kg) inhibited GFT and occupied striatal NK1R by 100% for >48 h despite that brain levels of compound were below the limit of detection after 24 h. Slow functional reversibility is associated with long-lasting in vivo efficacy of NK1R antagonists, whereas the efficacy of compounds with rapid reversibility is reflected by their pharmacokinetics.
The neurokinins substance P (SP), neurokinin (NK) A (NKA), and NKB belong to the tachykinin peptide family (Severini et al., 2002). The tachykinin receptors are divided into three subtypes: NK1R, NK2R, and NK3R. The rank order of potency of the endogenous tachykinins are: for NK1R, SP ≥ NKA > NKB; for NK2R, NKA > NKB > SP; and for NK3R, NKB > NKA > SP (for review, see Pennefather et al., 2004). Hemokinin-1 and endokinins A and B are relatively new mammalian members of the tachykinin family but appear to have similar receptor pharmacology as SP (Page, 2006). On the other hand, endokinins C and D have negligible affinity for known NK receptors (Page, 2006).
Preclinical research has implicated especially the NK1Ras being involved in several pathological disorders, including emesis, asthma, psychiatric disorders, gastrointestinal disorders, pain, migraine, inflammation, and urinary bladder disorders. This has led to the subsequent development of selective and potent NK1R antagonists (for recent review, see Quartara and Altamura, 2006). However, so far, only aprepitant has reached the market for treatment of chemotherapy-induced emesis.
To date, still little is known about the way antagonists interact with NK1R and, especially, about the mechanisms that govern the duration of their effects in vivo. The in vivo efficacy of an antagonist and its duration of action can sometimes be difficult to predict based only on potency values obtained by in vitro assays (Copeland et al., 2006). To provide more information, two additional approaches have often been used in in vitro pharmacological studies. One consists of exposing antagonist-pretreated tissue or cells to fresh medium and monitoring the restoration of receptor responsiveness to an agonist. Such functional washout experiments not only provide information about the functional dissociation rate of the antagonist-receptor complex but under appropriate conditions, also provide information about the likelihood of the liberated antagonist to undergo fast rebinding to receptors in the neighborhood of where they were released (Lüllmann et al., 1988; Fierens et al., 1999a; Chu et al., 2004).
The second approach consists of monitoring antagonists for their potential to be insurmountable, i.e., for their capability to decrease the maximal response that can be elicited by a subsequently added agonist (Vauquelin et al., 2002b). Although this approach has been most often used in “organ bath” experiments with intact tissues, it can also be successfully applied in intact cell-based experiments (Vauquelin et al., 2002a). As an illustration of this approach, in vitro assays with NK1R-expressing cells pointed at a causal link between the insurmountable behavior of the competitive NK1R-selective antagonists SR 140333 and aprepitant and their slow rate of dissociation from the receptor (Emonds-Alt et al., 1993; Hale et al., 1998). That slow dissociation may produce insurmountable inhibition can easily be explained by the fact that the antagonist fails to liberate all the receptor sites during the ensuing challenge with the agonist so that the measured response is suboptimal. On the other hand, the surmountable behavior of fast dissociating antagonists is likely to reflect a swift liberation of the receptors.
However, insurmountable antagonism can also be explained by noncompetitive interactions. This latter mechanism has been held responsible for the behavior of the NK1R-selective antagonist CP122,721 (McLean et al., 1996). These studies illustrate that still little is known about the way antagonists interact with NK1R.
In the present study, we compare three different NK receptor antagonists with respect to their functional interactions in vitro and how these interactions correlate to effect duration in vivo. The study has been performed in U373MG cells endogenously expressing the human NK1R (Eistetter et al., 1992). The experiments in vitro were designed to evaluate competitive and insurmountable NK1R interactions, and functional reversibility was tested after pretreatment of antagonist. Gerbils represent a species with similar NK1R pharmacology to man (Beresford et al., 1991; Engberg et al., 2007). We therefore investigated the pharmacokinetic/pharmacodynamic relationship of the compounds in vivo using the gerbil foot tap (GFT) assay, which is a model reflecting central NK1R activation (Bristow and Young, 1994). We also determined the degree of NK1R occupancy for aprepitant in gerbil striatum using autoradiography to verify the prolonged effect in vivo.
Materials and Methods
Chemicals. The selective NK1R antagonists CP-99994 (McLean et al., 1993) and aprepitant (Hale et al., 1998) and the pan-NK receptor antagonist ZD6021 (Bernstein et al., 2001) were synthesized at AstraZeneca (Mölndal, Sweden).
Cells. Human glioblastoma astrocytoma (U373MG) cells endogenously expressing NK1R were used (European collection of cultures 89081403; Sigma-Aldrich, St. Louis, MO). The cells were cultured in a humidified incubator under 5% CO2 in minimal essential medium with Earle's medium and GlutaMAX, 10% fetal bovine serum, 1% nonessential amino acids, and 1% minimal essential medium-sodium pyruvate. The cells were grown in T175 flasks and passaged when 70 to 80% confluency was achieved for up to a maximal of 20 passages. Before each experiment, U373MG cells were plated in black-walled/clear-bottomed 96-well plates (Costar 3904; Corning Life Sciences, Acton, MA) at 2.5 × 104 cells per well and grown for approximately 24 h in normal growth media in a 37°C CO2 incubator to achieve confluence.
Intracellular Measurements of Ca2+. U373MG cells, grown in 96-well plates, were loaded with the Ca2+ sensitive dye Fluo-4 (Teflabs 0152, Austin, TX) at 4 μM in a loading media consisting of Nut Mix F12 (HAM) with glutamax I, 22 mM HEPES, 2.5 mM probenecid (P-8761; Sigma-Aldrich), and 0.04% Pluronic F-127 (P-2443; Sigma-Aldrich) and kept dark for 30 min in a 37°C CO2 incubator. The cells were then washed three times in assay buffer, which consisted of Hanks' balanced salt solution containing 20 mM HEPES, 2.5 mM probenecid, and 0.1% bovine serum albumin (BSA), using a multichannel pipette leaving them in 150 μl at the end of the last wash. Serial dilutions of test compound in assay buffer (final dimethyl sulfoxide concentration kept below 1%) and/or agonist were automatically pipetted into each test well, and the peak fluorescence intensity was recorded (λex, 488 nm; λem, 540 nm) by the FLIPR charge-coupled device camera for approximately 2.5 min. The response was measured as the peak relative fluorescence after agonist addition. The potency of the antagonists used was determined using the same methodology but with Chinese hamster ovary-K1 cells transfected with human NK1R (Engberg et al., 2007).
Coincubation Experiments. To test for competitive interactions, a coincubation procedure was used by adding aprepitant, ZD6021, or CP-99994 (at final concentrations ranging from 40–640 nM) to the wells by the FLIPR automatic station simultaneously with increasing concentrations of SP.
Preincubation Experiments. To test for insurmountable interactions, a preincubation protocol was used by adding aprepitant, ZD6021, or CP-99994 (at final concentrations ranging from 1–40 nM) to the wells by the FLIPR automatic station 2.5 min before the addition of increasing concentrations of SP.
Reversibility of NK1R Antagonist Effect. U373MG cells, seeded in 96-well plates, were loaded with 4 μM Fluo-4 (see above) together with 10 nM aprepitant, ZD6021, CP-99994, or loading buffer (controls) and kept dark for 30 min in a 37°C CO2 incubator. The plates were then washed three times in assay buffer (see above), leaving the cells in 150 μl of assay buffer at the end of the last wash. The cells were then incubated for 1, 3, 10, 30, or 60 min at 37°C in a CO2 incubator before an SP solution (final concentration, 3 nM) was automatically pipetted.
Gerbil Foot Tap Experiments. Male Mongolian gerbils (60–80 g) were purchased from Charles River (Sulzfeld, Germany). On arrival, they were housed in groups of 10 in cages (height, 40 cm; width, 80 cm; length, 60 cm) containing an enriched environment including hay, plastic tubes, nesting material, and sand. Food and water were available ad libitum, and the cages were placed in temperature and humidity-controlled holding rooms. The animals were allowed at least 7 days to acclimatize to the housing conditions before experiments. All experiments were approved by the local animal ethical committee of Göteborg, Sweden.
Compounds and corresponding vehicles were administered under brief isoflurane (Forene; Abbott Scandinavia AB, Solna, Sweden) anesthesia. A dose of 3 μmol/kg aprepitant (dissolved in ethanol/solutol/saline, 5:5:90) or 10 μmol/kg ZD6021 (dissolved in 28% cyclodextrin) or 3 μmol/kg CP-99994 (dissolved in saline) or corresponding vehicle was administered i.p. at various time points before the experiment. At the indicated time point after compound administration, the animals were anesthetized (isoflurane), and a small incision was made in the skin over bregma. Ten picomoles of acetyl-[Arg6,Sar9,Met(O2)11]-SP6–11 (ASMSP), a selective agonist for NK1 receptors, was administered i.c.v. in a volume of 5 μl using a Hamilton syringe with a needle 4.5 mm long. The wound was clamped shut, and the animal was allowed to recover in a small plastic cage. The cage was placed on a piece of plastic tubing filled with water and connected to a computer via a pressure transducer. The number of taps produced by the animal was recorded for 6 min using customized computer software (PharmLab on-line 4.0; AstraZeneca). The average number of taps per minute during the middle 5 min was calculated (thus, the first and last 30 s were excluded). Ten picomoles of ASMSP typically evoked an average of 100 taps/min. Antagonist efficacy was expressed as percent inhibition in comparison with corresponding vehicle. After the experiment, the animals were sacrificed under anesthesia by exsanguination of the heart. Half of the brain together with plasma was removed to determine levels of compound. In aprepitant experiments, the other half of brain was used for autoradiography (see below).
Determination of Compound Concentrations in Brain and Plasma. The collected brains were thawed, and 3 ml of water/g brain tissue was added. The brain was homogenized by ultrasonication, and the brain homogenate and plasma samples were stored at –20°C until analysis. Brain homogenate and plasma samples (50 μl) were protein precipitated by the addition of 150 μl of acetonitrile containing 0.2% formic acid and internal standard. After vortexing, the samples were centrifuged for 20 min at 2900g and 4°C. The supernatant (75 μl) was diluted with 75 μl of 0.2% formic acid in water. Brain homogenate and plasma samples were analyzed by liquid chromatography (LC)-tandem mass spectrometry. An Agilent 1100 LC pump (Agilent Technologies, Waldbronn, Germany) was used with gradient elution using a flow rate at 0.6 ml/min. The mobile phase consisted of 2% acetonitrile and 0.2% formic acid in water (A) and 0.2% formic acid in acetonitrile (B). Separation was performed on a 30- × 2.1-mm C18 HyPURITY column with 5-μm particle size (Thermo Electron Corporation, Waltham, MA) using a linear gradient of 5 to 90% B for 2 min, held at 90% for 1 min, and returned to initial conditions in one step. The front was diverted to waste by using a six-port valve (VICI AG, Schenkon, Switzerland), and after 0.5 min, the effluent entered the mass spectrometer without splitting. Sample storage and injection was performed with a CTC HTS Pal autosampler (CTC Analytics, Zwingen, Switzerland). Detection was performed with positive electrospray ionization mode by multiple reaction monitoring using a Micromass Quattro LC triple quadrupole (Waters, Manchester, UK). Instrument control, data acquisition, and data evaluation were performed using Masslynx 4.0.
Autoradiography and Binding Experiments. Following a single dose of aprepitant (3 μmol/kg i.p., see GFT experiments above), the animals were sacrificed after various time points (0.5, 1, 2, 4, 8, 24, 48, and 72 h). The brains were rapidly removed and frozen on dry ice and stored in –80°C until further use. Sagittal frozen sections (16 μm) were sectioned in a cryostat at –15°C and thaw-mounted on SuperFrostPlus section slides (Menzel Gmbh and Co KG, Braunschweig, Germany) and stored at –80°C until use. Tissue sections were preincubated at room temperature for 5 min in 50 mM Tris-HCl containing 0.3% BSA. Sections were then incubated at room temperature in 50 mM Tris-HCl, pH 7.4, containing 0.1% BSA, 40 μg/ml bacitracin, 3 mM MnCl2, and Complete EDTA-free protease inhibitor cocktail tablets (Roche, Mannheim, Germany) for 60 min in the presence of 4 nM [3H]Sar,Met(O2)-substance P. CP-99994 (10 μM) was used to assess nonspecific binding. The sections were subsequently washed in 50 mM ice-cold Tris-HCl, pH 7.4, for 2 × 5 min, briefly dipped in ice-cold distilled water, and then dried. The sections were placed in hypercassettes and exposed for 4 days to imaging plates with 3H microscales (GE Healthcare, Little Chalfont, Buckinghamshire, UK) as standard. Imaging plates were scanned using a BAS-5000 Bio-Imaging Analyzer (Fuji Photo Film, Tokyo, Japan) and quantified using an image analysis software system (AIDA 4.10; Raytest, Straubenhardt, Germany) to measure optical densities. Ligand binding in the striatum was monitored since the NK1R is most abundant in this region in gerbil (Griffante et al., 2006), monkey (Bergström et al., 2000), and man (Hargreaves, 2002). Specific binding in the presence of [3H]Sar,Met(O2)-substance P was set to 100%, whereas nonspecific binding in the presence of 10 μM CP-99994 was set to 0%.
Data Analysis. Curve fitting, IC50, and EC50 estimations were carried out using a four-parameter logistic model in XLfit for Microsoft Excel (Microsoft, Redmond, WA). Data are expressed as mean values ± S.E.M.
Results
Effect of NK Receptor Antagonists on SP Concentration-Response Curves. In vitro antagonist interactions were monitored using U373MG cells that endogenously express the human NK1R. In coincubation experiments, all three antagonists produced a right-ward shift of the SP concentration-response curve (Fig. 1, A–C). Maximal responses to SP and Hill slopes remained the same, indicating a competitive interaction. The potency (pKB values) for aprepitant, ZD6021, and CP-99994 in Chinese hamster ovary-K1 cells were 8.7 ± 0.2, 8.7 ± 0.2, and 8.6 ± 0.4, respectively.
In preincubation experiments (antagonist added 2.5 min before SP), aprepitant suppressed the maximal response to SP in a concentration-dependent manner (Fig. 1, D–F). Preincubation with 1 nM aprepitant attenuated the maximal response to SP, whereas 10 nM virtually abolished the response (Fig. 1D). ZD6021 also suppressed the maximal response to SP in a concentration-dependent manner in preincubation experiments (Fig. 1E). When compared with the effect of aprepitant, the suppression by ZD6021 was less complete. In contrast, preincubation with all concentrations of CP-99994 produced a right-ward shift of the SP concentration-response curve with maintained maximal SP-evoked responses (Fig. 1F).
Reversibility of NK Receptor Antagonist Inhibition. Preincubation with 10 nM aprepitant produced long-lasting inhibition of SP-evoked responses (Fig. 2). The response to SP was not restored following 60-min washout of aprepitant. ZD6021 (10 nM) also produced time-dependent inhibition of SP-evoked responses resulting in ∼60% inhibition after 1 h. On the other hand, inhibition produced by preincubation of 10 nM CP-99994 was completely reversed within less than 30 min. The washout procedure per se did not affect the ability of SP to evoke increases in intracellular Ca2+.
Effect of NK Receptor Antagonists in Vivo. Aprepitant (3 μmol/kg i.p.) produced a long-lasting complete inhibition of the GFT response (Fig. 3A). Maximal inhibition was attained after 2 h, and brain levels peaked at this time point, reaching 450 nmol/kg. After 4 h, aprepitant levels in the brain started to decline; however, a full inhibitory response (100%) was maintained. At 48 h, levels of aprepitant in the brain were below the limit of quantification (10 nmol/kg), but a prominent inhibitory effect (80 ± 13%) was still present. The time-dependent inhibitory response elicited by aprepitant correlated extremely well with the degree of NK1R brain occupancy in autoradiography studies (Fig. 4). After 72 h, both the ASMSP-evoked GFT response and the NK1R occupancy by radioligand were restored (Fig. 4).
ZD6021 (10 μmol/kg i.p.) also inhibited GFT, with maximal effects (69 ± 11%) appearing after 1 h (Fig. 3B). The levels of ZD6021 peaked already at 30 min, reaching 123 ± 14 nmol/kg. The inhibitory effect and brain levels of ZD6021 slowly decreased after 1 h. At 8 h, levels of ZD6021 were below the level of quantification (10 nmol/kg), and the inhibitory effect had subsided.
Treatment with CP-99994 (3 μmol/kg i.p.) resulted in complete inhibition of GFT 15 min after treatment (Fig. 3C). The inhibitory effects were relatively short-lasting and reflected brain levels of compound that declined rapidly after 15 min. A summary of maximal compound levels detected in plasma and brain is shown in Table 1.
Discussion
The present study compares the in vitro NK1R interaction properties of the nonpeptide antagonists aprepitant, ZD6021, and CP-99994 with time-wise changes in blockade of NK1R function in vivo. Assays on intact U373MG cells that endogenously express human NK1R showed that all compounds are competitive antagonists with similar potency but that there is a marked difference in the duration of receptor blockade; i.e., aprepitant >> ZD6021 > CP-99994. The in vitro interaction properties of aprepitant correlate well with long-lasting functional GFT inhibition and in vivo NK1R occupancy in the gerbil CNS.
Earlier human and rabbit pulmonary artery relaxation and guinea pig ileum contraction studies already revealed that NK1R receptor antagonists like SR 140333, CP122,721, and MEN 11149 decrease the maximal response to substance P or related agonists (Emonds-Alt et al., 1993; Croci et al., 1995; Cirillo et al., 1998; Pedersen et al., 2000). As usual for such organ bath experiments (Leff and Martin, 1986), the tissues were preincubated with the antagonist before their challenge with agonist. Antagonists that inhibit the maximal response under such conditions are referred to as insurmountable (Gaddum et al., 1955; Vauquelin et al., 2002a,b). This type of antagonism can also be demonstrated to take place in cell lines provided that they are exposed to the antagonists before their challenge with an agonist (Fierens et al., 1999a). In the present study on human glioblastoma astrocytoma (U373MG) cells endogenously expressing NK1R, aprepitant produced a complete and ZD6021 a nearly complete decline in the substance P-mediated cytosolic Ca2+ transients. In contrast, CP-99994 acted surmountably, i.e., it only produced a right-ward shift of the substance P concentration-response curve.
Several models have been proposed to explain the operative mechanism of insurmountable antagonism. The most cited ones refer to noncompetitive interactions, including functional inhibition (i.e., blockade of an essential step in the agonist-induced chain of cellular events) and binding to an allosteric site at the receptor, as well as to competitive interactions (i.e., binding of the antagonist and agonist to at least partially overlapping sites at the receptor) but associated with slow antagonist dissociation (Vauquelin et al., 2002a,b). When the receptors are allowed to pre-equilibrate with the antagonist, these scenarios all lead to a reduction in receptor activity and are therefore difficult to resolve (Bond et al., 1989). In contrast, coincubation experiments allow a clearcut discrimination since, in that case, competitive antagonists no longer decrease the maximal agonist-evoked response, whereas noncompetitive antagonists still do (Fierens et al., 1999b). In such coincubation experiments, aprepitant, ZD6021, and CP-99994 only produced parallel right-ward shifts of the substance P concentration-response curves. This clearly points at the competitive nature of these antagonists and, hence, at a potential link between their degree of insurmountability and their dissociation rate from the receptor.
Antagonist dissociation from the NK1R was monitored by functional “washout” experiments involving preincubation of the U373MG cells with saturating concentrations of aprepitant, ZD6021, and CP-99994, washing and exposure to fresh medium for the indicated periods of time before measuring the maximal substance P-mediated calcium transients. In this experimental paradigm, the rate by which the response is restored depends on the dissociation rate of the preformed antagonist-receptor complexes (Vanderheyden et al., 2000). Agonists are well known to promote the internalization of NK1 receptors (and of GPCRs in general) via endocytotic processes. Among the several theories that have been put forward to explain insurmountable antagonism, it was proposed by Liu et al. (1992) that it could reflect the ability of such compounds to promote receptor internalization as well. This model was specifically proposed for AT1 receptor antagonists. However, subsequent confocal microscopic examinations revealed that nonpeptide antagonists did not affect the subcellular distribution of fluorescent AT1 receptor-green fluorescent protein conjugates (Hein et al., 1997; Le et al., 2005). Similar studies also indicate that nonpeptide NK1 receptor antagonists are unable to induce receptor internalization and even that they will prevent SP-induced NK1 receptor endocytosis and stress-induced NK1 receptor internalization in the basolateral amygdala (Southwell et al., 1996; Jenkinson et al., 1999; Smith et al., 1999). Accordingly, presently available experimental evidence does not support the potential link between insurmountable antagonism and receptor internalization as proposed by Liu et al. (1992). In agreement with the surmountable behavior of CP-99994, the response was rapidly restored to the control level (i.e., the level in nonpretreated cells) for the CP-99994-pretreated cells. The restoration was appreciably slower (reaching about 40% of the control level after 60 min) for ZD6021-pretreated cells, and even no restoration could be demonstrated within 60 min for the aprepitant-pretreated cells. These findings may explain the insurmountable behavior of aprepitant and ZD6021 in the preincubation experiments. Indeed, these antagonists should have been unable to liberate a substantial part of the NK1R during their subsequent challenge with substance P so that the maximally attainable response should be less than in the control situation, i.e., when all receptors are free at the moment of their challenge with agonist (Paton and Rang, 1966; Paton and Waud, 1967).
Interestingly, slow dissociation has previously also been observed for other insurmountable NK1R antagonists in organ bath washout experiments. The contractile response of SR 140333-pretreated guinea pig ileum to NK1R stimulation took more than an hour to recover half maximally (Emonds-Alt et al., 1993). Even slower recoveries of the response were recorded with FK888- and MEN 11149-pretreated guinea pig ilea (Cirillo et al., 1998). Because the slow dissociation of those antagonists offers a sufficient explanation for their insurmountable behavior, there is no strict necessity to invoke noncompetitive interactions.
Although aprepitant produced a full decline of the maximal response in preincubation experiments, increasing the ZD6021 concentration first decreased the maximal response to a limit and then only produced right-ward shifts of the substance P concentration-response curves. In the case of angiotensin AT1 receptors, such partial insurmountability was also observed for antagonists like irbesartan, valsartan, and EXP3174 (Fierens et al., 1999b; Verheijen et al., 2000). To explain this behavior, it was proposed that the antagonist-receptor complexes are able to adopt two distinct but inter-converting states: a fast reversible state (for the surmountable inhibition) and a slow reversible state (for insurmountable inhibition) (Fierens et al., 1999b; Vauquelin et al., 2001). Although still speculative at the present level of investigation, such a model could also provide a simple explanation for the partial insurmountable behavior of ZD6021.
The very slow dissociation of the aprepitant-NK1R complexes in the present intact cell-based experiments coincides with its long-lasting in vivo occupancy of central NK1R and its inhibitory effects in the GFT assay. In this respect, slow receptor dissociation has been proposed to contribute to the long-lasting clinical action of antagonists for angiotensin AT1 (Wienen et al., 1993; Aiyar et al., 1995; De Arriba et al., 1996; Unger, 1999), histamine H1 (Anthes et al., 2002), nicotinic (el-Bizri and Clarke, 1994), adrenergic α2A (Kukkonen et al., 1997), serotonergic 5-HT3 (Blower, 2003), and muscarinic M3 (Swinney, 2004) receptors. In this respect, recent simulation studies (Vauquelin and Van Liefde, 2006) reveal that, compared with a fast dissociating antagonist, prolonged in vivo receptor occupancy should take place when the antagonist-receptor complexes dissociate much slower than the antagonist gets eliminated. This implies that the duration of in vivo receptor protection by antagonists should not only depend on their rate of elimination via excretion and/or metabolism but also on the rate at which they dissociate from their receptor (Unger, 1999; Swinney, 2004). In line with this view, we show here that the sustained GFT-inhibiting efficacy of aprepitant reflect its in vivo NK1R occupancy in the CNS rather than compound levels of at the site of action. Long-lasting effects of aprepitant in GFT have been reported previously (Hale et al., 1998; Duffy et al., 2002), and the excellent correlation between the sustained GFT inhibition and central NK1R occupancy in the present study is also consistent with others (Duffy et al., 2002). The present study, however, extends these findings by demonstrating that prominent GFT-inhibiting efficacy of aprepitant persisted even when its CNS levels were below the limit of detection. This contrasts with the early phases of the treatment, where NK1R occupancy and inhibition of GFT by aprepitant closely followed its CNS levels until maximal inhibition was attained after about 2 h.
Elegant studies using positron emission topography have been performed with aprepitant in man (for summary, see Keller et al., 2006). Interestingly, the plasma levels required in man for 95% occupancy of central NK1R were approximately 1 μg/ml (approximately equivalent to 2 μM). In the current study, plasma levels peaked at 0.77 μM, suggesting that the dose used in gerbils is similar to clinically relevant doses in man, although potential species differences in protein binding and brain/plasma ratios need to be taken into account.
Compounds metabolized to pharmacologically active metabolites are also likely to prolong effect duration in vivo. In ferrets, administration of aprepitant results in formation of metabolites with affinity for NK1R (Huskey et al., 2003). However, the level of metabolites detected in ferret brain were much (>4-fold) lower than the parent compound aprepitant. In addition, the metabolites had weaker affinity for the NK1R (4–100-fold), suggesting that active metabolites do not play a role in mediating the pharmacological effects of aprepitant in vivo. To our knowledge, active metabolites of CP-99994 or ZD6021 have not been reported.
The pan NKR antagonist ZD6021 has been described to act as a competitive, surmountable antagonist at NK1R and NK2R in rabbit pulmonary arteries while having noncompetitive interactions at NK3R in guinea pig ileum (Rumsey et al., 2001). These findings clearly differ with the partially insurmountable effect of ZD6021 in the current study on human NK1R. This discrepancy could result from many causes, including species-related differences in receptor behavior as well as the much shorter challenge of the receptors with agonist before measuring the response in the present study. Furthermore, in contrast with partially insurmountable behavior and the relatively slow reversibility of ZD6021 antagonism in the present in vitro washout experiments, the inhibitory effects of ZD6021 on GFT corresponded well with the in vivo CNS levels of this antagonist. This could be related to an unfavorable ratio between the half-life of the ZD6021-NK1R complexes (about 1 h in the in vitro washout experiments) and the half-life of the compound in the CNS (approximately 3 h). Indeed, simulation studies (Vauquelin and Van Liefde, 2006) reveal that, even for slow dissociating antagonists, the in vivo receptor occupancy is mainly dictated by their rate of elimination if the half-life of the antagonistreceptor complexes is shorter. Because the surmountable antagonist CP-99994 dissociates even faster than ZD6021, it is thus no surprise that its inhibitory effect on GFT closely followed its in vivo CNS levels. Both reached a peak after 15 min and rapidly declined afterward.
Despite possessing similar potency at human NK1R in vitro, the brain levels required for efficacy in GFT differed somewhat between antagonists. This is not due to species-related differences in NK1R pharmacology because the pKB values for the antagonists at cloned gerbil NK1R were 8.8 for aprepitant (S. Engberg, I. Ahlstedt, B. von Mentzer, unpublished data) and 8.9 and 9.0 for ZD6021 and CP-99994, respectively (Engberg et al., 2007). These values correlate well when investigating antagonist potency at human NK1R (8.7, 8.7, and 8.6 for aprepitant, ZD6021, and CP-99994, respectively). In contrast, differences in efficacy in GFT in relation to brain levels may be explained by compound-dependent differences in protein binding, resulting in different levels of free antagonist in the CNS.
In conclusion, the present results comply with earlier simulation studies (Vauquelin and Van Liefde, 2006) by showing that the antagonist aprepitant exhibits very slow NK1R receptor dissociation in vitro and, likewise, produces long-lasting in vivo receptor blockade that cannot be explained by the time-wise decline of its free concentration. On the other hand, in compliance with their faster receptor dissociation in vitro, the in vivo effect duration of ZD6021 and CP-99994 is rather dictated by the pharmacokinetics of the compounds. The present findings also lend support to recent considerations (Copeland et al., 2006) about potential advantages of long receptor occupancy by a drug in terms of its pharmacological effect duration and the underlying need to allocate more attention to kinetic approaches in in vitro drug discovery studies.
Footnotes
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.107.124958.
-
ABBREVIATIONS: SP, substance P; NK, neurokinin; NK1R, NK1 receptor; SR 140333, nolpitantium; GFT, gerbil foot tap; CP-99994, (2S,3S)-3-(2-methoxybenzylamino)-2-phenylpiperidine; BSA, bovine serum albumin; FLIPR, fluorometric imaging plate reader; ASMSP, acetyl-[Arg6,Sar9,Met(O2)11]-SP6–11; LC, liquid chromatography; CNS, central nervous system; MEN 11149, (2-(2-naphthyl)-1-N-{(1R,2S)-2-N-[1(H)indol-3-yl-carbonyl]aminocyclohexanecarbonyl}-1-[N′-methyl-N′-(4-methylphenylacetyl)]diaminoethane; ZD6021, 3-cyano-N-((2S)-2-(3,4-dichlorophenyl)-4-[4-[2-(methyl-(S)-sulfinyl)phenyl]piperidino]butyl)-N-methyl]napthamide]; CP122,721, (+)-(2S,3S)-3-(2-methoxy-5-trifluoromethoxybenzyl)amino-2-phenylpiperidine.
- Received April 27, 2007.
- Accepted June 14, 2007.
- The American Society for Pharmacology and Experimental Therapeutics