Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Corticotropin releasing factor (CRF), a 41-amino acid peptide, plays a pivotal role in the behavioral, endocrine, immune, and autonomic responses of the body to stress (Owens and Nemeroff, 1991). In addition to the hypothalamic paraventricular nucleus where it was originally identified, CRF is also widely distributed across brain regions (Chalmers et al., 1996; Heinrichs and De Souza, 1999; Gilligan et al., 2000a). The physiological functions of CRF are mediated by at least two G-protein coupled receptors, CRF₁ and CRF₂ (including splice variants CRF₂, CRF₂, CRF₂), both of which are linked to adenylyl cyclase activation but have distinct brain distributions. CRF₁ receptors are widespread in the cortex, limbic system, cerebellum, and pituitary, whereas CRF₂ receptors are dominant in subcortical areas including the lateral septum (CRF₂), ventromedial hypothalamus (CRF₂), and choroid plexus (CRF₂) (De Souza, 1987; Chalmers et al., 1995; Primus et al., 1997; Rominger et al., 1998).

Increasing evidence suggests that the CRF system is involved in pathophysiology of anxiety disorders (Heinrichs and De Souza, 1999; Gilligan et al., 2000a). Intracerebroventricular administration of CRF induces stress behaviors, whereas application of the peptide antagonist, -helical CRF, diminishes CRF-elicited as well as stress-elicited behavioral manifestations of anxiety (Korte et al., 1994). In transgenic animals, overexpression of CRF peptide produces increased levels of anxiety, whereas knockout of CRF₁ receptors decreases stress responses (Stenzel-Poore et al., 1996; Timpl et al., 1998). Clinically, patients diagnosed with anxiety-related disorders such as post-traumatic stress disorder (Bremner et al., 1997), obsessive compulsive disorder (Altemus et al., 1994), and anorexia nervosa (Kaye et al., 1987) exhibit increased levels of CRF in the cerebrospinal fluid (CSF).

Evidence from recent studies on nonpeptide CRF₁ antagonists supports the contention that CRF₁ antagonists may have a use in the treatment of anxiety disorders (Gilligan et al., 2000a; Takahashi, 2001). For example, the prototype CRF₁ antagonist, CP-154,526, attenuates social isolation-evoked stress responses in rat pups (Kehne et al., 2000), whereas its analog antalarmin decreases behavioral responses caused by social stress in primates (Habib et al., 2000). Another nonpeptide CRF₁ antagonist, R121919, is efficacious in several stress tests including elevated plus-maze, defensive withdrawal, and defensive burying (Keck et al., 2001; Heinrichs et al., 2002). In addition, the same compound showed anxiolytic and antidepressant efficacy in a small, open-labeled clinical study (Zobel et al., 2000). Likewise, several other CRF₁ antagonists including CRA1000 and CRA1001 (Okuyama et al., 1999), DMP695 (Bakthavatchalam et al., 1998; Millan et al., 2001), DMP696 (McElroy et al., 2002), and DMP904 (Gilligan et al., 2000b; Takahashi et al., 2001) have shown anxiolytic profiles in preclinical animal models. For peripherally applied nonpeptide CRF₁ antagonists to be effective, these compounds should have good plasma drug exposure, blood-brain barrier penetration, and brain CRF₁ receptor occupancy. Indeed, several recent studies show that CP-154,526 and R121919 effectively cross the blood-brain barrier and occupy central CRF₁ receptors (Arborelius et al., 2000; Keck et al., 2001; Heinrichs et al., 2002; Keller et al., 2002). In the study by Heinrichs et al. (2002), a dose-dependent relationship was demonstrated between behavioral efficacy and CRF₁ receptor occupancy by R121919. A dose of R121919 that achieves minimal efficacy also occupies ~50% CRF₁ receptors in the brain. Thus, it is of importance to test if this occupancy-efficacy relationship is applicable to other nonpeptide CRF₁ antagonists. In addition, it is important to further understand the relationship between plasma and brain drug exposure and receptor occupancy and behavioral efficacy of nonpeptide CRF₁ antagonists.

DMP696 (see Fig. 1 for chemical structure) is a potent and selective CRF₁ receptor antagonist (IC₅₀, 2-5 nM) (He et al., 2000; Zhang et al., 2003). It blocks CRF-induced adenylate cyclase activity in rat cortical membranes and inhibits adrenocorticotropic release from cultured pituitary cells (He et al., 2000). In a rat defensive withdrawal test of anxiety, acute oral administration of DMP696 (3 mg/kg) reduces the latency for a rat to exit from an isolated box and reverses the stress-induced increases in plasma corticosterone levels without any effect on locomotor activity or motor coordination (He et al., 2000; McElroy et al., 2002). In addition, at a higher concentration (30 mg/kg), DMP696 attenuates the enhanced stress response caused by maternal separation in rats (Maciag et al., 2002).

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Chemical structure of DMP696 [4-(1,3-dimethoxyprop-2-ylamine)-2,7-dimethyl-8-(2,4-dichlorophenyl)-pyrazolo[1,5-a]-1,3,5-triazine].

The present study was designated to investigate 1) in vivo receptor occupancy of DMP696 in relation to the behavioral efficacy, and 2) the relationship between receptor occupancy and plasma and brain exposure of DMP696. The receptor occupancy of DMP696 was measured using ex vivo binding autoradiography. The anxiolytic effects were determined using the defensive withdrawal test. Plasma and brain drug exposure was measured by a liquid chromatography/tandem mass spectrometric method in the same animals in which receptor occupancy was studied. Part of this study was previously presented in an abstract form (Hill et al., 2001).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. [¹²⁵I]Sauvagine and [¹²⁵I]ovine CRF (oCRF) were purchased from PerkinElmer Life Sciences (Boston, MA). Urocortin II was purchased from American Peptide Co. (Sunnyvale, CA), and -helical CRF_9-41 was purchased from Peninsula Labs (Bemont, CA). DMP696 and CP-154,526 were synthesized by the Chemical and Physical Sciences Department (Bristol-Myers Squibb Company, Wilmington, DE), and antisauvagine-30 was prepared by the Applied Biotechnology Group (Bristol-Myers Squibb Company). For in vitro application, nonpeptide compounds were dissolved in dimethyl sulfoxide and diluted in assay buffers. For in vivo application, compounds were prepared as suspensions in an aqueous vehicle of 0.25% methocel (methyl cellulose, Type AL5c; Dow Chemicals, Midland, MI). Stock suspensions were bead-milled overnight using three layers of 4-mm glass beads. Compounds were administered orally by gavage (p.o.) in a volume of 2 ml/kg b.wt. Doses of all drugs were calculated and are expressed in terms of the free base weight.

General Procedures. Male Sprague-Dawley rats (200-300 g b.wt.) were purchased from Charles River Laboratories (Wilmington, MA). The rats were double housed in shoebox cages (except those used in the defensive withdrawal test) in a colony room maintained at constant temperature (21-22°C) and humidity (50 ± 10%). The room was illuminated 12 h/day (lights on at 600 AM). The rats had ad libitum access to food and water throughout the study. For in vivo studies, the rats were fasted overnight before oral administration of drugs, and all experiments were conducted between 6:00 AM and 1:00 PM. All experimental procedures were performed according to protocols approved by the Animal Care and Use Committee of the Bristol-Myers Squibb Company and the published guidelines in the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

In Vitro Binding Autoradiography. Naive rats (without treatment) were sacrificed by decapitation, and the brain and pituitary were collected, embedded in M-1 embedding matrix (Thermo Shandon, Pittsburgh, PA), and frozen in ice-chilled 2-methylbutane (Alfa Aesar, Ward Hill, MA). The brain and pituitary tissues were cut into 20-µm sections on a Cryostat, and sections were mounted on superfrost slides (VWR International, Wilmington, DE) and stored at 70°C until use.

Ex Vivo Binding Autoradiography. For dose-related studies, rats were orally administrated with DMP696 at various doses (0.3, 1.0, 3.0, 10, and 30 mg/kg) or CP-154,526 (1.0, 3.0, 10, and 30 mg/kg) or vehicle and subsequently sacrificed at 90 min postdose. The 90-min postdose survival time was chosen as orally dosed DMP696 reaches the maximal concentration in plasma at the time point (see time course study results). For time course studies, rats were dosed with 10 mg/kg DMP696, and subsequently sacrificed at times from 10 min up to 22 h, and the brain and pituitary were then collected. The forebrain, the upper brainstem including the rostral portion of the cerebellum, and the pituitary were sectioned in a Cryostat (20 µM). The remaining brainstem and cerebellum tissues were used for measuring drug concentrations in the brain tissues. In most cases, the trunk blood samples were collected immediately after decapitation, and the plasma was separated by centrifugation for assessment of drug concentration.

Measurement of DMP696 Concentrations in Plasma, Brain, and Cerebrospinal Fluid. DMP696 concentrations in the plasma and brain were measured using a liquid chromatography/tandem mass spectrometric method (LC/MS/MS). Briefly, the 0.1-ml sample (plasma or homogenized brain tissue), 50 µl of 200 nM internal standard solution, and 0.1 ml of 0.1 M Na₂CO₃ were mixed followed by the addition of 1.0 ml of 1:1 methyl-t-butyl ether/EtOAc. Samples were vortexed, centrifuged, and the organic layer was transferred and evaporated until dry under nitrogen at 60°C. Residues were reconstituted with 0.1 ml of H₂O/CH₃CN/HCOOH (50/50/0.1, v/v/v). High-performance liquid chromatography separation was achieved using an acetonitrile (0.1% formic acid)/water (0.1% formic acid) gradient on a Zorbax, SB-C₁₈ column (2 × 50 mm, 5 µm) at a flow rate of 200 µl/min with an analysis time of 5 min. Detection was performed in the positive, multiple reaction monitoring mode using a Quattro Ultima with an EI source as the LC/MS/MS interface.

Defensive Withdrawal Test. The defensive withdrawal procedure as described by McElroy et al. (2002) was used. The testing apparatus consisted of an opaque Plexiglas open field (106-cm length × 92-cm width × 50-cm height), containing a cylindrical galvanized chamber (14-cm length, 10-cm diameter) that was positioned lengthwise against one wall, with the open end 40 cm from the corner. The open field was illuminated by a 60-W incandescent bulb, and illumination was titrated by a Powerstat transformer to a 23-lux reading at the entrance to the cylinder. Rats were habituated to handling by gently stroking their dorsal surface for approximately 1 min daily for 5 to 6 consecutive days before testing. DMP696 and vehicle (0.25% methocel) was orally dosed 60 min before behavioral testing. To initiate testing, the rat was placed within the cylinder, which was then secured to the floor. Behavior was assessed for 15 min by a trained observer (unaware of treatment assignment) by a video monitor in an adjacent room. The latency to exit the chamber, defined by the placement of all four paws into the open field, was recorded (in seconds). The Plexiglas chamber and the cylinder were cleaned with 1.0% glacial acetic acid between animals to prevent olfactory cues from influencing the behavior of subsequently tested animals.

Data Analysis. Digital images were generated with a Cyclone storage phosphor-imaging system and analyzed using OptiQuant Analysis software (PerkinElmer Life Sciences). Radioligand binding density in a defined brain region (according to the rat atlas by Paxinos and Watson, 1998) was measured as digital light units per millimeter squared. For in vitro studies, specific binding in a defined brain region was calculated by subtracting the value of the nonspecific binding density from that of the total binding density measured in the corresponding brain region and normalized using nondrug-treated sections as control. For ex vivo ligand binding studies, the percentage of specific binding in drug-treated rats was calculated as the following: percent specific binding = (specific binding in drug-treated minus nonspecific binding in vehicle-treated)/(specific binding in vehicle-treated-nonspecific binding in vehicle-treated) × 100%. The percentage of specific binding in a drug-treated condition is inversely proportional to the percent inhibition or percent receptor occupancy by the drug.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of DMP696 on [¹²⁵I]oCRF and [¹²⁵I]Sauvagine Binding: In Vitro Studies. The overall distribution pattern of [¹²⁵I]oCRF and [¹²⁵I]sauvagine binding corresponded well with that reported previously in the same species (Aguilera et al., 1987; De Souza, 1987; Primus et al., 1997; Rominger et al., 1998). CRF₁ binding sites labeled by both [¹²⁵I]oCRF and [¹²⁵I]sauvagine were dominant in the cerebral cortex, the subcortical limbic system, the cerebellar cortex, and the anterior pituitary, whereas CRF₂ binding sites labeled only by [¹²⁵I]sauvagine were concentrated in the lateral septal nucleus, the medial nucleus of the amygdala, the ventromedial nucleus of the hypothalamus, and the choroid plexus. Figure 2 illustrates [¹²⁵I]sauvagine and [¹²⁵I]oCRF binding in a representative forebrain level in the absence and presence of various blocking ligands. [¹²⁵I]Sauvagine binding sites in the cortical regions and anterior pituitary were displaceable with 1 µM DMP696 (Fig. 2B), CP-154,526 (Fig. 2C), or -helical CRF(Fig. 2D), a nonselective CRF receptor antagonist, but not by 20 nM antisauvagine-30 (Fig. 2E), a selective CRF₂ receptor antagonist, or urocortin II (Fig. 2F), a selective CRF₂ agonist, indicating that these binding sites represent CRF₁ receptors. Dense [¹²⁵I]sauvagine binding sites in the lateral septal nucleus and choroid plexus were displaceable with -helical CRF (Fig. 2D), antisauvagine-30 (Fig. 2E), or urocortin II (Fig. 2F), but not DPC696 (Fig. 2B) or CP-154,526 (Fig. 2C), indicating that they represent CRF₂ receptor binding sites. [¹²⁵I]oCRF binding to the cortical regions and anterior pituitary was completely displaceable with DMP696 (Fig. 2GH).

View larger version (103K): [in this window] [in a new window]   Fig. 2.   Representative autoradiograms of coronal forebrain sections showing [125I]sauvagine binding in the absence (A) and presence of 1 µM DMP696 (B), 1 µM CP-154,526 (C), 1 µM -helical CRF (D), 30 nM antisauvagine-30 (E), or 30 nM urocortin II (F), and 125I-oCRF binding in the absence (G) and presence (H) of 1 µM DMP696. FC, frontal cortex; PC, parietal cortex; LS, lateral septum; AP, anterior pituitary; ChP, choroid plexus; CPu, caudate-putamen. Scale bar = 2 mm.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Comparison of concentration-dependent effects of DMP696 on [125I]oCRF and [125I]sauvagine binding in several brain regions, anterior pituitary, and choroid plexus. Data are the mean ± S.E.M. of the percentage of specific binding (n = 4). A, [125I]oCRF binding; B and C, [125I]sauvagine binding.

Effect of DPC696 on [¹²⁵I]Sauvagine Binding: Ex Vivo Studies. Given the comparability between [¹²⁵I]sauvagine and [¹²⁵I]oCRF binding to CRF₁ receptor sites, we used [¹²⁵I]sauvagine exclusively for ex vivo studies so that in vivo effects of DMP696 on both CRF₁ and CRF₂ receptors could be simultaneously monitored. To test if in vitro processing (primarily incubation time) affects DMP696 receptor occupancy values, two experiments were performed to examine the dissociation kinetic profile of DMP696. In the first experiment, brain sections from rats dosed with 10 mg/kg DMP696 were incubated with [¹²⁵I]sauvagine for various times, from 10 up to 240 min, and the effect of the incubation time on DMP696 inhibition of [¹²⁵I]sauvagine binding in the parietal cortex was calculated (Fig. 4A). Ten minutes after incubation, 82% of [¹²⁵I]sauvagine binding was inhibited. The inhibition increased to 91% at 20 min and maintained around 90% up to 60 min after incubation. Further increasing the incubation time resulted in a gradual decline of the inhibition (down to 26% at 4 h). In the second experiment, brain sections from naive rats were preincubated with DMP696 at 1 or 10 nM for 2 h before incubation with [¹²⁵I]sauvagine, from 10 up to 240 min. As shown in Fig. 4B, between 40 to 60 min after incubation, [¹²⁵I]sauvagine binding was inhibited by about 50 and 90% in the sections preincubated with 1 and 10 nM DMP696, respectively. The percent inhibition is consistent with theoretically calculated values based on the IC₅₀ (0.9 nM for the parietal cortex) of DMP696 at both concentrations (50% inhibition at 1 × IC₅₀ and 90% inhibition at 10 × IC₅₀). The percent inhibition from incubation times shorter or longer than 40 to 60 min deviated from these values with increased variability. Based on the results from these two tests, we chose a 40-min incubation time for all of our ex vivo binding studies.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Effects of [125I]sauvagine incubation time on DMP696 inhibition of specific binding. A, ex vivo binding: brain sections collected from rats (n = 3) orally dosed with 10 mg/kg DMP696. B, in vitro binding: naive rat (n = 3) brain sections preincubated with 1 or 10 nM DMP696 for 2 h before incubation with [125I]sauvagine. Data are the mean ± S.E.M. for three binding tests.

View larger version (97K): [in this window] [in a new window]   Fig. 5.   Representative autoradiograms of coronal forebrain sections showing dose-dependent inhibition of [125I]sauvagine binding by orally dosed DMP696. Note that DMP696 did not affect [125I]sauvagine binding in the choroid plexus. FC, frontal cortex; PC, parietal cortex; LS, lateral septum; AP, anterior pituitary. Scale bar = 2 mm.

View larger version (33K): [in this window] [in a new window]   Fig. 6.   Orally dosed DMP696 (A, B, C, D, E) and CP-154,526 (F) produced dose-dependent inhibition of [125I]sauvagine binding in several brain regions and anterior pituitary.

Time Course of Receptor Occupancy and Drug Concentrations in Plasma and Brain. The time course of the percent inhibition of [¹²⁵I]sauvagine binding by DMP696, drug plasma-free concentrations, and brain concentrations of DMP696, following a oral dose of 10 mg/kg of the drug, were examined. Plasma protein binding of DMP696 in rats is 98.5% and, consequently, the plasma-free drug concentrations were calculated by multiplying the plasma-free fraction (1.5%) by the total plasma concentrations. For easy comparison, percent inhibition was expressed as receptor occupancy of CRF₁ receptors (Fig. 7). Each time point was collected from pooled data of three to five rats. Overall, the time course of the receptor occupancy correlated well with that of the plasma-free concentrations and the brain concentrations of DMP696 (Fig. 7). The drug concentrations in the brain were over 150-fold higher than the plasma-free levels, however. At 40 min after dosing, the plasma-free level of DMP696 reached 5 nM, total brain concentration was over 80 nM, and the receptor occupancy was over 60%. The receptor occupancy and the drug concentration in plasma and brain peaked at 90 min postdosing [95%, 9.8 nM (free plasma) and 1547 nM (brain), respectively] and declined afterward. By 22 h after dosing, when the plasma-free drug level dropped to below 1 nM, there was apparently no DMP696 occupancy in the brain.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Time course of CRF1 receptor occupancy, drug plasma-free concentrations and brain tissue concentrations following a single oral dose of DMP696 (10 mg/kg). CRF1 receptor occupancy was defined by the percent inhibition of specific [125I]sauvagine in the parietal cortex. The data represent the mean ± S.E.M. for four rats at each time point.

Correlation of Receptor Occupancy with Free Plasma Concentrations. The relationship between CRF₁ receptor occupancy and free concentrations of DMP696 in the plasma was examined using pooled data from the ex vivo studies described above (i.e., Fig. 6). Figure 8 is a plot of % specific binding versus free plasma concentrations of DMP696. The data were fitted using an inhibitory E_max model with the % specific binding as the observed effect (Winnonlin Version 3.3; Pharmasight 2001, MountainView, CA). The kinetic-modeling analysis of the data yielded an in vivo IC₅₀ value of 1.2 nM, which is consistent with the in vitro IC₅₀ value (0.9 nM; both the in vivo and in vitro values were measured from the same brain region, parietal cortex with the same ligand, [¹²⁵I]sauvagine as ligand).

View larger version (17K): [in this window] [in a new window]   Fig. 8.   Relationship between % specific binding in the parietal cortex and DMP696 plasma-free concentrations. The data were fitted using an inhibitory Emax model with a predicted in vivo IC50 of 1.2 nM.

DMP696 Concentrations in the Plasma Versus CSF. The purpose of this experiment was to compare drug concentrations in CSF versus plasma after an oral dose of 10 mg/kg DMP696. The data were averaged from seven rats for each time point. One hour postdosing, the plasma-free concentration was 8.7 ± 2.4 nM, and the CSF concentration was 14 ± 4.3 nM. Three hours after dosing, the concentration was slightly decreased in both plasma (8.0 ± 1.4 nM) and CSF (13 ± 2.5 nM). This result indicates the consistency of the plasma-free levels of DMP696 with the CSF concentrations.

Defensive Withdrawal Test. In this experiment, rats were subjected to the defensive withdrawal test following varying oral doses of DMP696 (i.e., 0, 1, 3, 10, 30, and 90 mg/kg). Figure 9 shows exit latencies for each dose at 60 min after oral administration of DMP696. At 3 mg/kg, there was a substantial, albeit not statistically significant, decrease in the exit latency (48% from the vehicle level). At 10 mg/kg, the reduction was greater (62%) and statistically significant. Behavioral efficacy appears to plateau at 10 mg/kg since higher doses (30 and 90 mg/kg) did not produce significantly lower exit latencies.

View larger version (17K): [in this window] [in a new window]   Fig. 9.   Effect of orally dosed DMP696 (1-90 mg/kg) on the exit latency in a defense withdrawal test of anxiety. Exit latency is defined as the time taken for an animal to emerge from a darkened chamber into a novel environment. Total test time is 900 s. Data presented are the mean ± S.E.M. for eight animals per group; #, p < 0.05 [compared with the vehicle (0.25% methocel) group].

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present study demonstrates that DMP696 selectively and dose dependently occupies CRF₁ receptors in the brain with no apparent regional differences. DMP696 administrated at 3 mg/kg occupied just over 50% CRF₁ receptors, a dose equivalent to that produced a minimal anxiolytic effect (~50% reduction in the exit latency). Examination of the relationship between receptor occupancy and plasma-free concentrations yields an in vivo IC₅₀ of 1.22 nM, which is similar to the in vitro IC₅₀ (0.9 nM) of the compound. Furthermore, this study shows a parallel time course of receptor occupancy, plasma-free concentrations and brain tissue concentrations of DMP696.

The binding affinity of DMP696 for CRF₁ receptors measured using the brain section binding autoradiography is in accordance with the previously studies using homogenized cell membrane assays (He et al., 2000; Zhang et al., 2003). For example, Zhang et al. (2001) revealed an IC₅₀ value of 1.96 and 5.2 nM for DMP696 to inhibit CRF₁ receptors in the cortex and pituitary, respectively. These values are consistent with those observed in the present study (cortex, 0.9 nM; pituitary, 2.7 nM), suggesting comparability between autoradiographic and homogenized assays for assessing CRF₁ antagonist affinity. In combination with storage phosphor-imaging techniques, the brain section binding assay allows evaluation of drug potency in anatomically defined brain structures, with significantly improved throughput compared with conventional film autoradiography.

There was no significant regional difference of DMP696 binding affinity to CRF₁ receptors in the brain. In contrast, DMP696 was apparently less potent in the pituitary compared with the brain, as described previously (Zhang et al., 2003). The functional significance of the affinity difference between the brain and pituitary is not known. Nevertheless, DMP696 appears to have equal efficacy in producing anxiolytic effects in the defensive withdrawal test and in reversing stress-elicited corticosterone release (McElroy et al., 2002). In addition, our ex vivo binding data did not reveal any significant differences of in vivo receptor occupancy between the pituitary and brain structures of the compound.

Ex vivo binding autoradiography allows in vitro measurement of a receptor population occupied by drugs administrated in vivo. This method has been used for assessing in vivo binding profiles of a variety of drugs including CRF₁ antagonists (Arborelius et al., 2000; Keck et al., 2001; Heinrichs et al., 2002). The inherent limitation of the method is the requirement of in vitro processing of tissue sections, which gives rise to the probability of dissociation of receptor-bound drugs from their binding sites, causing underestimation of in vivo occupancy. One recent ex vivo binding study has demonstrated that peripherally administrated raclopride, an antipsychotic, dissociates swiftly from the binding sites during a 30-min incubation and causes over 70% underestimation of the receptor occupancy (Kapur et al., 2001). To avoid potential underestimation of DMP696 occupancy, we studied the dissociation time course of DMP696 during in vitro processing. In the first test on brain sections from rats dosed with 10 mg/kg DMP696, the receptor occupancy was around 90% 20 to 60 min after the beginning of radioligand incubation, and decreased significantly afterward. In the second test, 40 to 60 min after the beginning of radioligand incubation of naive sections, which were preincubated with 1 or 10 nM DMP696, ~50% or ~90% receptors were occupied. The occupancy values at these two concentrations are consistent with calculated based on the IC₅₀, e.g., 1 nM DMP696 occupying ~50% receptors and 10 nM occupying ~90%. Data from these tests suggest that a limited incubation time (40 min or less) is required to minimize underestimation of DMP696 occupancy.

Literature that suggested shortening in vitro incubation time "as much as possible" may not necessarily be optimal for every drug of interest (Kapur et al., 2001; Langlois et al., 2001). For DMP696, an incubation time less than 40 min caused decreased receptor occupancy with increased variability. Conceivably, in such a short incubation time, radioligands do not have time to reach equilibrium. In this study, a 40-min incubation may be long enough to allow radioligand binding to approach equilibrium, but not too long to cause significant dissociation of DMP696. Thus, caution should be taken in selection of an incubation time for ex vivo measurement of receptor occupancy. An in vitro dissociation time course study of drugs of interest may be warranted before full-scale ex vivo measurement.

Using the ex vivo binding, we measured systemically receptor occupancy of DMP696 in the brain. Orally administrated DMP696 dose dependently occupied CRF₁, but not CRF₂ receptors, with no significant regional differences in the brain. On average, at 1 mg/kg, DMP696 produced ~40% occupancy of CRF₁ receptors in the brain, and the percentage increased to ~60% at 3 mg/kg. A further dose increase to 10 and 30 mg/kg resulted in receptor occupancy over 80 and 90%, respectively. The receptor occupancy appears related to the behavioral effect. In the defensive withdrawal test, 3 mg/kg (but not 1 mg/kg DMP696) produced 48% reduction of the exit latency, whereas 1 mg/kg DMP696 was ineffective. The statistical insignificance at 3 mg/kg was likely due to a greater variability in this study, as the same dose repeatedly shows significant efficacy in the same test from our previous studies (He et al., 2000; McElroy et al., 2002). Therefore, 3 mg/kg DMP696 appears to be the minimal oral dose for producing anxiolytic effects. At 10 mg/kg, DMP696 produced a greater, but apparently saturating, effect as further increasing the dose to 30 or 90 mg/kg produced no significant increase in the efficacy, consistent with the results of previous studies (He et al., 2000; McElroy et al., 2002). Taken together, these data suggest that in vivo receptor occupancy of DMP696 is dose-dependent and is closely related with the anxiolytic efficacy of the compound. The dose of DMP696 that occupied ~50% CRF₁ receptors in the brain is identical to the minimally effective dose in the defensive withdrawal test. These data suggest that blockade of at least 50% of CRF₁ receptors is a requisite for anxiolytic effects of DMP696 in the defensive withdrawal model of anxiety. Further studies are necessary to confirm that this relationship holds for other models of anxiety and for potential anxiolytic effects in humans.

A close correlation of receptor occupancy with behavioral effects has also been observed for several other CRF₁ antagonists. R121919 occupied central CRF₁ receptors in a dose-dependent manner (Keck et al., 2001; Heinrichs et al., 2002). An oral dose (2.5 mg/kg) of R121919, which produced a minimal anxiolytic effect occupied 50% CRF₁ receptors (Heinrichs et al., 2002). CP-154,526 was not effective in a defensive withdrawal test until a dose over 35 mg/kg (Arborelius et al., 2000), at which the compound occupied over 50% brain CRF₁ receptor (Arborelius et al., 2000; Fig. 5 of this study). In addition, we have observed the consistency of doses occupying over 50% receptors with doses effective in behavioral tests for a number of in-house CRF₁ antagonists (Y.-W. Li and J. McElroy, unpublished observations).

We explored further the relationship between drug exposure and receptor occupancy following DMP696 dosing. An understanding of the relationship is important in that it could potentially serve as a guide for the selection of a dose regime in clinical studies. The parallel decline of unbound (free) plasma concentrations and total brain concentrations of DMP696 in relation to receptor occupancy from our in vivo time course study (Fig. 7) suggests that DMP696 rapidly equilibrates with CRF₁ receptors in the brain.

For DMP696, a highly lipophilic compound (ClogP 4.97), passive diffusion is likely to be the main route of its entry into the brain (Pardridge, 1998). Unbound drug concentrations in plasma is an important factor, which governs the extent of brain distribution of compounds that enter via passive diffusion (Sawchuk and Yang, 1999). Thus, unbound rather than total plasma concentrations of DMP696 were used to examine the relationship between receptor occupancy and plasma exposure. Analysis of the brain receptor occupancy and the free plasma concentrations yielded an estimated in vivo IC₅₀ value of 1.22 nM. Interestingly, the in vivo IC₅₀ value is remarkably close to the in vitro IC₅₀ of DMP696 observed throughout in vitro binding assays from this and a previous study (Zhang et al., 2003). This suggests that the availability of DMP696 for CRF₁ receptors in the brain is similar to free concentrations observed in plasma. We have found that free plasma levels of CRF₁ antagonists are an important factor in the availability of these compounds to the receptors in the brain. Previously examined CRF₁ antagonists with plasma protein binding >99.8% showed no appreciable receptor occupancy despite having excellent plasma exposures of total drug (H. Wong and Y.-W. Li, unpublished observations). Compared with the plasma, brain had remarkably high concentrations of DMP696. The peak concentration in the brain at 90 min postdosing was 1547 nM, over 150-fold higher than the unbound plasma level. The accumulation of DMP696 in the brain is likely the consequence of the lipid-enriched brain tissue functioning as a "sink" for the highly lipophilic compound. It is therefore conceivable that brain concentrations of compounds like DMP696 may not necessarily be meaningful in predicting their occupancy of targeted receptors.

In conclusion, the results from this study support the hypothesis that the anxiolytic effect of DMP696 is mediated by acting on brain CRF₁ receptors and suggest that at least 50% receptor occupancy is needed for efficacy. The similarity in the in vivo and in vitro IC₅₀ values for DMP696 suggests that plasma-free concentrations of DMP696 are important for the entry of the compound into the brain and binding to CRF₁ receptors.