Abstract
The neurokinin 3 (NK3) receptor antagonists represent a novel class of pharmacological agents, which are currently under evaluation for the treatment of psychiatric disorders. An efficient brain penetration is one of the main prerequisites to further evaluate compounds displaying high potency to bind the NK3 receptor. The present report describes a method for determining the in vivo occupancy of central NK3 receptors after peripheral administration of drugs. An ex vivo measurement of NK3 receptor occupancy by quantitative autoradiography employing [3H]senktide as the radioligand has been developed. The speed of the method, which is usually considered low due to the time dedicated to film exposure (from weeks to months), has been considerably increased by the use of the β-imager. The high sensitivity of this new radioimager was used to visualize and quantitatively analyze the [3H]senktide binding sites in brain sections within hours. Using this method, we have demonstrated that the reference NK3 antagonist SR142801 dose dependently occupied the NK3 receptors in the gerbil brain after subcutaneous administration with an ED50 of 0.85 mg/kg. The less active enantiomer SR142806 occupied the NK3 receptors only by 25% at the highest used dose of 10 mg/kg. These values are in accordance with the reported behavioral effects of the compounds. Our results indicate that ex vivo receptor occupancy measurements can be dependently used to predict the central activity of NK3 antagonists. More generally, the combination of ex vivo receptor autoradiography with the β-imager detection constitutes a new and fast method to evaluate the brain penetration of drug candidates.
The three main tachykinins, substance P, neurokinin A (NKA), and neurokinin B (NKB), constitute a family of neuropeptides interacting with three distinct neurokinin (NK) receptors (for a review, see Maggi, 1995). Substance P binds preferentially to the NK1 receptor, NKA to the NK2 receptor, and NKB to the NK3 receptor. The NK3 receptor is characterized by a predominant expression in the brain, and numerous data suggest its involvement in the modulation of central monoaminergic systems (Alonso et al., 1996; Jung et al., 1996; Marco et al., 1998). These properties make the NK3 receptor a potential target for central nervous system (CNS) disorders such as psychosis, anxiety, and depression. Structurally different nonpeptide NK3 antagonists have been synthesized (for a review, see Giardina et al., 2000). All these compounds display a high affinity for the human NK3 receptor, but one absolute requirement to select them for further development in psychiatric indications is their ability to penetrate the CNS.
The observation of changes in the behavior of animals is the standard method to initially evaluate the central activity of drugs. However, no modulation of normal behavior could be observed after administration of NK3 antagonists, including the reference compound SR142801 (Jung et al., 1996; Ribeiro et al., 1999). In fact, the central activity of NK3 antagonist has been first demonstrated by their ability to antagonize the behavioral responses induced by the selective NK3 agonist senktide (Emonds-Alt et al., 1995; Jung et al., 1996; Sarau et al., 1997, 2000; Ribeiro et al., 1999). It is well known that the nonpeptide NK3 antagonists have a higher affinity for the human, the guinea pig, and the gerbil than for the rat and the mouse NK3 receptors (Chung et al., 1995). Due to this species difference, the potency of compounds could be underestimated in mice and rats, the two species mostly used for psychopharmacological evaluation. Another complicating factor is the difference in senktide-induced behavior depending on the species used. Indeed, the principal effect of senktide in gerbils is a reduction of the locomotor activity, whereas in rats and mice the peptide induces mainly head twitches (Jung et al., 1996; Sarau et al., 2000). Both effects seem to be mediated by the NK3 receptor since NK3 antagonists dose dependently inhibit them. These differences in senktide-induced behavior could be explained by our recent study showing marked differences in the localization of NK3 receptors between the rat and the gerbil (Langlois et al., 2001). Altogether, these differences between species make the choice of an in vivo animal model to screen centrally active NK3 antagonists difficult.
The CNS penetration of drugs can be also directly evaluated by measuring the brain concentration of compounds after peripheral administration. This method has been used to demonstrate the central activity of the NK3 antagonists SB-222200 and SB-223412 (Sarau et al., 1997, 2000) and has revealed a superior CNS penetration of the first compound. However, this technique requires development of a specific analytical method for each compound, which is very laborious. Another possibility is to measure directly the occupancy of NK3 receptors by compounds in the brain of treated animals. We have explored this alternative by developing an ex vivo measurement of NK3 receptor occupancy by quantitative autoradiography using the radioligand [3H]senktide. Receptor autoradiography experiments with tritiated ligands have been always considered as a lengthy process, due to the length of time needed for exposure on film. Consequently, ex vivo autoradiographic measurement has never been used as a screening method for determining the CNS penetration of compounds. To circumvent this inconvenience, we have used a new device, the β-imager, instead of autoradiographic films for quantifying radioactivity. This radioimager is based on a gaseous detector of β-particles developed by Charpak and collaborators (1989). The high sensitivity of the β-imager allowed us to detect specific [3H]senktide binding sites on brain sections in only a few hours, which would be an acceptable time frame even for fast screening of compounds.
To validate our method, we have demonstrated that the reference NK3 antagonist SR142801 dose dependently occupied the central NK3 receptors in gerbils, a species whose NK3 receptors are pharmacologically similar to that of humans. Our study provides the first direct demonstration that SR142801 penetrates efficiently the brain where it would exert its pharmacological action by blocking the central NK3 receptors.
Materials and Methods
Drug Treatment and Tissue Preparation.
Male Mongolian gerbils (40–60 g) were treated by subcutaneous injections of saline or test compounds at four dosages ranging from 0.16 to 10 mg/kg. In complementary experiments, for determining the occupancy of NK3 receptors by SR142801 in different brain areas, gerbils were treated with a single dose of 1 mg/kg. In every case, three animals were used per dose of compound. The animals were killed by decapitation 1 h after drug administration. Brains were immediately removed from the skull and rapidly frozen in dry ice-cooled 2-methylbutane (−40°C). Coronal sections (20 μm thick) were cut using a Reichert Jung 2800R cryostat-microtome (Cambridge Instruments, Cambridge, UK) and thaw-mounted on silanized microscope slides (Star Frost, Knittel Gläser, Germany). The sections were stored at −20°C until use.
Ex Vivo [3H]Senktide Binding in Brain Sections.
After thawing, sections were dried under a cold stream of air. The sections were not washed prior to incubation, to avoid dissociation of the drug-receptor complex. Three adjacent brain slices from the same animal have been collected per slide. Two brain slices were used to measure the total binding, and the third one was evaluated for nonspecific binding. The use of silanized slides allows the co-incubation of the two conditions on the same slide without mixing. Total binding was measured by incubating sections with 3 nM [3H]senktide (63.5 Ci/mmol) in Tris-HCl buffer (50 mM, pH 7.4) containing 3 mM MnCl2, 0.02% (w/v) bovine serum albumin, 40 μg/ml bacitracin, 2 μg/ml chymostatin, 4 μg/ml leupeptin (total volume 400 μl). Nonspecific binding was measured in the adjacent section by the addition of 10 μM SR142801 (total volume 200 μl) to the incubation medium. Incubation was restricted to 10 min at room temperature to minimize dissociation of the drug from the receptor. To stop the incubation, the slides were washed (4 × 1 min) in Tris-HCl buffer, pH 7.4, at 4°C followed by a rapid dip in cold distilled water and drying under a stream of cold air.
To evaluate the [3H]senktide binding in equilibrium conditions, brain sections of control gerbils were incubated according to a standard protocol (Dam et al., 1990). Briefly, sections were preincubated (3 × 5 min) in 50 mM Tris-HCl, pH 7.4, at room temperature and incubated for 90 min in the incubation medium described above. Washing and drying steps were the same as for the ex vivo experiment.
Quantitative Analysis.
Slides were made conductive by disposing a copper foil tape (3M, Diegem, Belgium) on the free side. They were placed in the gas chamber (mixture of argon and triethylamine) of the β-imager (BioSpace, Paris, France). Each β-particle emitted from samples generated a light spot 1 mm in diameter, which was detected by a charged coupled device camera. The coordinates of the center of gravity of each light spot were calculated and visualized on a monitor (Charpak et al., 1989). Depending on the conditions, data from brain sections were collected during 8, 12, or 16 h. The levels of bound radioactivity in the brain areas were directly determined by counting the number of β-particles emerging from the delineated area by using the Beta vision program (BioSpace). Consequently, the radioligand binding signal was expressed in counts per minute per square millimeter.
Sections were then exposed to 3H-Hyperfilm (Amersham Pharmacia Biotech UK, Ltd., Buckinghamshire, UK) for 8 to 12 weeks. After development of films, autoradiograms were analyzed and quantified using a MicroComputer Imaging Device M1 image analysis system (Imaging Research, St. Catharines, Ontario, Canada). Optical densities in the anatomical region of interest were transformed into levels of bound radioactivity after calibration of the image analyzer with gray values generated by the coexposed3H-microscales standards (Amersham Pharmacia Biotech). The radioligand binding signal was expressed in femtomoles per milligram of tissue equivalent (fmol/mteq).
Ex vivo receptor labeling by [3H]senktide in cingulate cortex of drug-treated gerbils was expressed as the percentage of receptor labeling in corresponding brain areas of saline-treated animals. Since only unoccupied receptors remain available for the radioligand, ex vivo receptor labeling is inversely proportional to the receptor occupancy by the in vivo administered drug. Percentages of receptor occupancy by the drug administered to the animal correspond to 100% minus the percentage of receptor labeling in the treated animal. The percentage of receptor occupancy was plotted against dosage, and the sigmoidal log dose-effect curve of best fit was calculated by nonlinear regression analysis, using the GraphPad Prism program (San Diego, CA). From these dose-response curves, the ED50 values were calculated with their 95% confidence limits.
Chemicals.
[3H]Senktide was obtained from PerkinElmer Life Science Products (Boston, MA). Bovine serum albumin and chymostatin were purchased from Sigma (St. Louis, MO). Bacitracin and leupeptin were obtained from Boehringer Ingelheim GmbH (Ingelheim, Germany). SR142801 and SR142806 have been synthesized in house.
Results
NK3 Receptor Labeling by [3H]Senktide in Gerbil Forebrain Sections.
Autoradiograms showing total binding after long (90 min) and short (10 min) incubation times with [3H]senktide are shown in Fig.1. Consecutive slides from the same saline-treated gerbil were used to compare the level of specific binding in both conditions. After 90 min of incubation and 8 weeks of exposure on films, [3H]senktide labeled mainly the mid-cortical layers in the gerbil forebrain. The addition in the incubation medium of 10 μM SR142801 reduced the signal to the film background (not shown). The highest densities of specific binding sites were found in the cingulate cortex (100.2 ± 13 fmol/mteq, mean ± S.E.M., n = 3). A minimum of 12 weeks of exposure on film was necessary to detect quantifiable binding sites in gerbil brain sections after a 10-min incubation with [3H]senktide. In this condition, the level of specific binding in the cingulate cortex was equal to 7.7 ± 0.5 fmol/mteq (mean ± S.E.M., n = 9), which represents less than 10% of the specific binding measured at the equilibrium.
Comparison of β-Imager Detection with Film Autoradiography.
The β-imager is a real-time imaging system displaying the status of the acquired image on line. Total and nonspecific binding were measured on adjacent brain sections of saline-treated gerbils after a 10-min incubation with [3H]senktide. The slides were placed in the β-imager for 12 h. As shown in Fig.2, the distribution of β-particles emerging from gerbil brain sections could be assessed already after 2 h. The contrast of the image increased with the time and reached its maximum at 12 h. Comparison of the digital image at 4 h with the film autoradiogram obtained after 12 weeks (∼2000 h) showed that the β-imager is at least 500 times more sensitive and faster. Whereas the brain section incubated in the presence of 10 μM SR142801 (nonspecific binding) was hardly distinguishable from the background on the film, it could be visualized on the monitor of the β-imager after 4 h of data acquisition. The anatomical resolution of the digital image is not comparable with film autoradiograms but by far sufficient enough for quantifying the specific signal in the cingulate cortex (0.163 ± 0.02 cpm/mm2, mean ± S.E.M.,n = 9).
NK3 Receptor Occupancy by the Selective NK3 Antagonist SR142801.
SR142801, its less active (R)-enantiomer SR142806 and the racemic mixture were subcutaneously administered to gerbils at four different dosages ranging from 0.16 to 10 mg/kg (three animals per dose). The animals were sacrificed 1 h after administration. Including the saline-treated animals, 15 animals were used per compound in a typical experiment. Because a maximum of 15 slides can be loaded at one time in the gas chamber of the β-imager, the occupancy of NK3 receptors by one compound can be measured during a single acquisition using five doses in triplicate. Figure3A shows the [3H]senktide binding (total and nonspecific) on brain sections of gerbils treated with increasing doses of SR142801 and SR142806. The data of both compounds were individually collected during 8 h. This time was mainly chosen for a practical reason, permitting measurement of two different treatments per 24 h. Moreover, the signal in the cingulate cortex was strong enough to be quantified after 8 h. Juxtaposition of both images shows that SR142801 was much more potent than SR142806 in occupying the central NK3 receptors. At 2.5 and 10 mg/kg SR142801, [3H]senktide could no longer label the cortical layers, whereas at the same doses of SR142806, there was still a clear labeling corresponding to the unoccupied NK3 receptors.
Individual values and mean curves illustrating the occupancy of NK3 receptors by SR142801, SR142806, and the racemic mixture are shown in Fig. 3B. Occupancy of NK3 receptors by SR142801 was detectable for dosages ≥0.63 mg/kg and was total at the highest dose of 10 mg/kg. The ED50 value (95% confidence limit) generated from the curve was 0.85 mg/kg (0.63–1.16). The percentage of occupancy of NK3 receptors by the less active enantiomer SR142806 at the highest dose of 10 mg/kg was 25 ± 9% (mean ± S.E.M.,n = 3). Occupancy of NK3 receptors by the racemic mixture required a 2- to 3-fold higher dose (ED50= 2.18 mg/kg, 1.57–3.03) than with SR142801.
After the determination of the ED50 of SR142801 in the cingulate cortex, gerbils were administered with a single dose of the compound (1 mg/kg, 1 h, s.c.) to determine whether a similar level of NK3 receptor occupancy could be observed in different brain areas. This dose has been chosen since it occupies slightly more than 50% NK3 receptors but still allows visualization of the brain areas in the drug-treated animals. Acquisition time with the β-imager has been extended to 16 h to quantify the specific signal in brain areas displaying low levels of NK3 receptors such as the hypothalamus. Figure 4 shows that the NK3 receptors are occupied by 69% in the cingulate cortex, by 61% in the parietal cortex, and by 66% in the hypothalamus indicating that SR142801 displays a similar occupancy of NK3 receptors independent of the brain area quantified.
Discussion
The present report describes a new method for fast evaluation of the central potency of NK3 receptor antagonists. An ex vivo autoradiographic protocol has been developed to evaluate the occupancy of NK3 receptors in the gerbil brain by drug candidates. A major improvement of the presented approach is the use of a novel read-out technology, which overcomes the conventional time-consuming exposure on autoradiographic film. Due to the high sensitivity and the rapidity of the employed β-imager, this protocol has the potential to be used for general screening of centrally active compounds, as exemplified here using specific NK3 antagonists.
In ex vivo receptor binding experiments, the unlabeled drug is administered peripherally to the animal; thereafter, the animal is sacrificed and the brain processed for in vitro receptor labeling. The advantage of labeling the receptor on tissue sections rather than in tissue homogenates has been already demonstrated (Schotte et al., 1989,1993). In that manner, the dissociation of the drug-receptor complex formed in vivo can be kept minimal by immediate freezing of the brains, omitting preincubations of the sections, and by using short incubations with the radioligand. The short incubation time is a critical step of ex vivo autoradiographic protocols, particularly when the method is used for evaluation of compounds with unknown properties (as in screening mode). Since the binding kinetic properties of the administered compounds are generally not known at early stages of the selection process, it cannot be predicted how long the drug-receptor complex would stay stable during the incubation with the radioligand. Therefore, to be able to compare the ability of structurally different compounds for occupying central receptors, the incubation time with the radioligand has to be minimized as much as possible. The standard incubation time that we are generally using in our laboratory for ex vivo autoradiography experiments is 10 min (Schotte et al., 1996). This is significantly shorter than the incubation time applied in general receptor autoradiography protocols and results, consequently, in a reduction of the intensity of the autoradiographic signal. For example, the optimal time of incubation with the NK3 agonist [3H]senktide to study the distribution of NK3 receptors in brain sections is 90 min (Dam et al., 1990; Langlois et al., 2001). As illustrated in Fig. 1, the reduction of the incubation time to 10 min caused a large decrease of the specific [3H]senktide binding (from 100.2–7.7 fmol/mteq in the cingulate cortex). To compensate for this signal diminution, the exposure time on film has to be prolonged. In our view, the time dedicated to the exposure on film (from several weeks to several months) (Schotte et al., 1996) was the major drawback of conventional ex vivo autoradiographic approaches in the past. The slow throughput of the method (and the long waiting time for the results) was largely incompatible with its use in earlier stages of drug development.
In recent years, radioimagers have been developed as alternatives to film autoradiography. The introduction of storage phosphor imaging systems was the first major improvement in radioimaging; the exposure time of samples being thereafter counted in days in place of weeks or even months. With the novel β-imager, the time unit of exposure has been even further reduced. Indeed, our study showed that a very low specific signal (<10 fmol/mteq) could be detected within 2 h only (see Fig. 2). In contrast, 12 weeks exposure on film were necessary to visualize the cortical layers on the same sample labeled with [3H]senktide. For information, the slide shown on Fig. 2 has been also exposed to an imaging plate (Fuji Photo Film Co., Ltd., Tokyo, Japan) to compare the sensitivity and the speed of a storage phosphor imaging system (Fuji BAS 2000, Raytest Benelux B.V., Tilburg, Netherlands) with those of the β-imager. A minimum exposure of 1 week before scanning the plate was necessary to obtain a signal equivalent (not shown) to an 8-h acquisition with the β-imager. It was not the aim of the present publication to review in detail the performances of both systems. Rather we would like to focus on the new possibilities offered by the new β-imaging technology. One of these is the ability to detect very low levels of binding sites in a few hours only. This characteristic is particularly important for the fast and accurate quantification of ex vivo autoradiography experiments as exemplified here in the study performed with [3H]senktide. The requirement for such studies is the ability to quantify the dose-dependent occupancy of the NK3 receptor facing an already low maximal signal. The digital image (Fig.3A) shows that the receptor labeling was dose dependently and fully inhibited by the in vivo administration of SR142801 but only partially inhibited by SR142806. Taking the low specific signal (from 7.7 to 0 fmol/mteq) of our assay into consideration, a precise ranking of compounds according to their degree of NK3 receptor occupancy might be put into question. The plotting of each individual value (Fig. 3B) demonstrates that we could precisely measure the level of receptor occupancy with a minimum of intragroup variability. Also, the possibility of differentiating the active enantiomer SR142801 from the racemate mixture illustrates the high dependence and sensitivity of our assay. Considering that SR142806 displayed a 70-fold lower affinity than SR142801 for the gerbil NK3 receptor in radioligand binding assay (Emonds-Alt et al., 1995), the dose of the racemate should be theoretically twice the dose of SR142801 for occupying an equivalent level of NK3 receptors. The present data confirms this assumption since the calculated ED50 value of SR142801 and the racemate are equal to 0.85 and 2.18 mg/kg, respectively.
The above ED50 values have been determined in the cingulate cortex. This is the gerbil brain area showing the highest density of NK3 receptors and consequently the largest signal to quantify a dose-dependent occupancy of NK3 receptors. The NK3 receptors are also present in subcortical areas but with a significantly lower level of expression (for more details, see our recent publication on the detailed distribution of NK3 receptors in the gerbil brain,Langlois et al., 2001), which makes an accurate quantification more difficult; particularly after a short-time incubation with [3H]senktide. Nevertheless, we evaluated the occupancy of NK3 receptors by 1 mg/kg of SR142801 in three different brain areas: the cingulate cortex, the parietal cortex, and the hypothalamus. Figure 4 demonstrates that SR142801 occupies the NK3 receptors (by 61–69%) similarly throughout the brain, indicating that the cingulate cortex is a representative area for evaluating the occupancy of NK3 receptor in the whole brain. This experiment illustrates also the linearity of our assay, since we found the same level of receptor occupancy independent of the density of NK3 receptors and the brain region quantified.
However, the main criterion to be used for validating ex-vivo studies is the correlation found between the dose of compound occupying the receptors in the brain and the dose active in centrally mediated pharmacological tests. In our assay, the ED50value of SR142801 for occupying the NK3 receptors in gerbil brain is equal to 0.85 mg/kg (1 h, s.c.). Published data indicate that SR142801 inhibited turning behavior induced by intrastriatal injection of senktide in gerbils with an ID50 of 0.58 mg/kg (30 min, i.p., Emonds-Alt et al., 1995). SR142801 antagonized also the reduction of exploratory activity in gerbils induced by i.c.v. injection of senktide with an ID50 of 1.92 mg/kg (30 min, i.p.; Jung et al., 1996). The inactive enantiomer SR142806 was always used as a negative control in these behavioral experiments. Our data confirm the relevance of this control since at the highest dose of 10 mg/kg, only 25% of receptors were occupied by SR142806. This favorable comparison between ex vivo autoradiographic and in vivo behavioral data makes us confident about the use of NK3 receptor occupancy measurement for screening of centrally active NK3 antagonists.
In practice the compounds would be first tested at one dose and depending on the initial scores, selected ones would undergo a full dose occupancy study. By this strategy, we hope to be able to rapidly select better CNS penetrating NK3 antagonists than the reference compound SR142801. As judiciously notified by Giardina and collaborators (2000), since the first description of SR142801 as a selective and potent nonpeptide NK3 antagonist, few new chemical entities have been identified. In our view, difficult in vivo evaluations due to the differences in receptor pharmacology and the absence of convincing data on the presence of NK3 receptors in primate brain have slowed down the interest for developing such compounds. Recently, several reports have demonstrated the presence of NK3 receptors in human brain (Koutcherov et al., 2000; Tooney et al., 2000). Our own recent report has shown that the NK3 receptor is localized in discrete areas of the monkey brain (Langlois et al., 2000), indicating that primates could be used in preclinical investigation of NK3 antagonists. Before starting such investigation, one way to evaluate the potential central activity of new synthesized NK3 antagonists would be to measure their NK3 receptor occupancy with the improved method described in the present report.
Footnotes
- Abbreviations:
- NK
- neurokinin
- CNS
- central nervous system
- fmol/mteq
- femtomole per milligram of tissue equivalent
- Received March 19, 2001.
- Accepted June 26, 2001.
- The American Society for Pharmacology and Experimental Therapeutics