Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The original D₁ and D₂ pharmacological subclasses of dopamine receptors (Kebabian and Calne, 1979) is now known in mammals to consist of two (Garau et al., 1978) D₁-like receptors (D_1A/D₁ and D_1B/D₅) and products of three D₂-like genes that result in four receptors: D_2long (D_2L) and D_2short (D_2S), D₃, and D₄. D₁-like receptors preferentially recognize 1-phenyl-tetrahydrobenzazepines (e.g., SCH23390) over benzamides (e.g., sulpiride), whereas the D₂-like receptors have the opposite pharmacological specificity. D₁- and D₂-like receptors have been defined traditionally by their opposing effects on the enzyme adenylate cyclase, with D₁ receptors positively coupled to this enzyme, whereas D₂ receptors are either negatively coupled or uncoupled to this effector. More recently, the actions of dopamine D₁- and D₂-like receptors on signaling systems other than adenylate cyclase have been confirmed in a variety of systems, including coupling to G protein inwardly rectifying potassium channels, phosphatidylinositol hydrolysis, and voltage-activated calcium channels (Jaber et al., 1996).

Dopamine D₂-like receptors have been a primary target of drug development efforts for treating disorders of dopamine neurotransmission such as Parkinson's disease and schizophrenia. The functional effects of dopamine D₂ receptor ligands in brain arise from their interactions with dopamine D₂ autoreceptors expressed on dopamine neurons and with D₂ heteroreceptors on target cells. D₂ autoreceptors regulate the firing of dopamine neurons, as well as the synthesis and release of dopamine. In the striatum, the major dopamine terminal field in mammalian brain, D₂ heteroreceptors include those present on giant cholinergic interneurons that regulate the release of acetylcholine, as well as D₂ receptors expressed on medium spiny -aminobutyric acidergic afferent neurons (Le Moine and Bloch, 1991; Le Moine et al., 1991; Weiner et al., 1991). In anterior pituitary, D₂ heteroreceptors regulate the release of prolactin (Caron et al., 1978). Although there has been controversy about how the various molecular isoforms of D₂-like receptors contribute to the myriad of functional events mediated by these receptors in brain and neuroendocrine tissues, it is clear that functional diversity is not defined solely, or even primarily, by molecular diversity. For example, D_2L, D_2S, and D₃ receptors are expressed as both autoreceptors and heteroreceptors (Bouthenet et al., 1991; Snyder et al., 1991). Although it has been suggested that D₃ receptors have a primary role as autoreceptors (Meller et al., 1993; Aretha et al., 1995; Kreiss et al., 1995), this notion has not been supported by more recent studies in which D₂ or D₃ receptors have been ablated using transgenic methods (Koeltzow et al., 1998; L'hirondel et al., 1998).

Dihydrexidine (DHX) and related members of the hexahydrobenzo[a]phenanthridine class have been shown in previous publications to comprise a particularly interesting class of dopamine agonists (Lovenberg et al., 1989; Brewster et al., 1990, 1995; Mottola et al., 1992). In addition to its properties as a high-affinity, full D₁ receptor agonist, DHX also has affinity for D₂ receptors similar to the prototypical D₂ receptor agonist quinpirole. Although there has been extensive characterization of the D₁ properties of DHX (Watts et al., 1993b, 1995), there is only preliminary data about the functional properties of DHX at D₂ receptors. These latter data suggest, however, that DHX has typical D₂ agonist properties, binding to both high- and low-affinity [³H]spiperone-labeled sites, inhibiting prolactin release in vivo, and also inhibiting D₁-stimulated cAMP efflux much like dopamine or other D₂ agonists (Mottola et al., 1992).

Together, it appears that members of the benzo[a]phenanthridine series might have particular pharmacological utility by virtue of their being high-affinity full agonists for both D₁- and D₂-like receptors. Moreover, Brewster et al. (1990) have shown that the D₁/D₂ selectivity of this series of rigid dopamine analogs can be "tailored" (e.g., N-n-propyl-DHX has twice the D₂ affinity of DHX, yet has much lower D₁ affinity). The present studies were designed to provide a full pharmacological characterization of two key members of this series, dihydrexidine and N-n-propyl-dihydrexidine (PrDHX) (Fig. 1) with functions in native tissues that are generally accepted as being mediated by D₂-like receptors. Contrary to the initial hypothesis, the current data lead to the surprising conclusion that DHX can activate D₂ receptors on striatal target cells to produce maximal inhibition of adenylate cyclase, while having essentially no effects at D₂ autoreceptors coupled to dopamine release, synthesis, or cell firing (despite binding to these receptors with equal affinity). Understanding the mechanisms underlying such "functional selectivity" may provide novel directions for drug design.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Structures of DHX and its N-n-propyl analog (PrDHX).

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials

(6a,12b)-trans-10,11-Dihydroxy-5,6,6a,7,8,12b-hexahydrobenzo- [a]phenanthridine (DHX), N-methyl-DHX, and PrDHX were synthesized as described previously (Brewster et al., 1990). [³H]Spiperone was purchased from Amersham Biosciences (Piscataway, NJ), and ¹²⁵I-epidepride was synthesized by Dr. Chester A. Mathis (University of Pittsburgh, Pittsburgh, PA), using a published protocol (Bishop et al., 1991). Apomorphine and U86170E were from Pharmacia & Upjohn, Inc. (Kalamazoo, MI). ¹²⁵Iodine and [³H]NPA were purchased from PerkinElmer Life Sciences (Boston, MA). SCH23390 was a gift from Schering Plough (Kenilworth, NJ) or was purchased from Sigma/RBI (Natick, MA), as was SKF38393, R-(+)- and S-(+)-NPA, R-(+)- and S-()-3-PPP, and haloperidol. Quinpirole (LY171555) was a gift from Lilly Research Laboratories (Indianapolis, IN). Spiperone, ketanserin, and domperidone were gifts from Janssen Pharmaceutica (Titusville, NJ). Remoxipride was a gift from Astra Arcus AB (Södertälje, Sweden). EDTA, isobutylmethylxanthine, GppNHp, dopamine, ()-sulpiride, reserpine, and chlorpromazine were purchased from Sigma-Aldrich (St. Louis, MO). Tris-HCl and HEPES were purchased from Research Organics (Cleveland, OH). The complexing agent 1,4,7,10,13,16-hexaoxacyclooctadecane (18-crown-6) was purchased from Fluka Chemical (St. Louis, MO). Anti-cAMP primary antibody for cAMP assays was obtained from Dr. Gary Brooker (Georgetown University School of Medicine, Washington, DC).

Animals and Tissue Preparation

All experiments were carried out in accordance with the Guide for the Care and Use of Laboratory Animals as adopted by the National Institutes of Health. Male Sprague-Dawley rats (Charles River Laboratories, Raleigh, NC) were housed under a 12-h light/dark cycle and given food and water ad libitum. In all experiments, rats weighing 250 to 400 g were sacrificed by decapitation. Some rats were treated with 5 mg/kg reserpine 20 h before euthanasia. Except where indicated, brains were immediately removed then rinsed and chilled in ice-cold 0.9% (w/v) sodium chloride solution. The striatum was dissected from two 1.2-mm coronal slices that were made with the aid of a cold dissecting block. The tissue was either used immediately as indicated, or was stored at 70°C for later use.

Radioligand Receptor Binding

Membrane Preparation. Radioligand binding followed published methods (Schulz et al., 1985) with minor modifications. Frozen rat striata were homogenized in ice-cold 50 mM HEPES, 4 mM MgCl₂ buffer, pH 7.4, at 25°C. This buffer was used throughout the binding experiments, except that for experiments with GppNHp or [³H]NPA, the buffer was modified to include 1 mM EDTA and 0.002% ascorbic acid. Homogenized tissue was centrifuged at 27,000g for 10 min. The pellet was again homogenized (five strokes by hand) with resuspension and centrifugation as described above. The final pellet was suspended at a concentration of approximately 2.0 mg wet weight/ml of binding buffer. In some cases, membranes were preincubated for 30 min at room temperature with 100 µM GppNHp before performing the radioligand binding experiments. The final tissue concentration per tube in all binding assays was 1 mg/ml. Protein levels were later estimated using the Folin reagent method (Lowry et al., 1951).

[³H]Spiperone-Labeled D₂ Receptors. Ten to 15 concentrations of unlabeled competitor were assessed at D₂ receptors labeled with 0.07 nM [³H]spiperone, as described previously (Mottola et al., 1992).

[³H]SCH23390-Labeled D₁ Receptors. Competition binding assays at D₁ receptors were performed as described previously (Watts et al., 1995). D₁ receptors were labeled with 0.3 nM [³H]SCH23390. Nonspecific binding was defined by 1 µM unlabeled SCH23390.

[³H]NPA-labeled D₂ Receptors. D₂ receptors in their high-affinity state were labeled with 0.05 nM [³H]NPA. Previous saturation binding experiments conducted with [³H]NPA in this laboratory had determined that [³H]NPA binds to striatal D₂ receptors with a K_D of ca. 0.1 nM. The B_max for [³H]NPA binding was found to be approximately 200 fmol/mg of protein, roughly one-half that of [³H]spiperone, suggesting that, under our experimental conditions, half of the striatal D₂ receptors were in the high-affinity state. Competition binding of [³H]NPA-labeled sites was determined as described for [³H]spiperone, with the exception that the incubation period was 30 min.

Receptor and Other Data Analyses

Data for the receptor competition binding experiments were analyzed by nonlinear regression (Prism; GraphPad Software, San Diego, CA) to provide estimates of receptor affinity (IC₅₀) and indirect Hill coefficients (n_H). Apparent affinity (K_0.5) was determined from IC₅₀ values by adjusting for ligand concentration using the Cheng-Prusoff equation for bimolecular interactions. Other selected data were analyzed using a one-way analysis of variance, followed by Tukey-Kramer multiple comparison post hoc tests using InStat version 3.05 (GraphPad Software).

DHX Accumulation in Brain

Sample Preparation. Rats were dosed with 3.0 mg/kg DHX or PrDHX s.c. dissolved in 0.1% ascorbic acid. Rats administered DHX were euthanized by decapitation at 10, 20, 40, or 120 min later. Rats dosed with PrDHX were euthanized at a single time point, 20 min after drug administration. The brains were removed quickly, and the striata excised as described above. Whole cerebella were gently peeled off the back of the brain and brain stem. Wet weights were obtained for each brain region dissected, the brain tissue placed in plastic vials, and stored at 70°C. On the day of the assay, brain nuclei were sonified using a sonicating tip (Branson Ultrasonics, Danbury, CT) in 1 ml of mobile phase containing a fixed amount of the internal standard N-methyl-DHX. After centrifugation (15,000g for 15 min), 600 µl of the supernatant was added to tubes containing 400 µl of 1 mM Tris buffer, pH 8.6, and 30 mg of alumina (Bioanalytical Systems, West Lafayette, IN). After shaking for 5 min, the alumina was separated by brief centrifugation and the fluid aspirated. The alumina was then washed twice by adding 1.5 ml of double-distilled H₂O and shaking for 2 min. Catecholamines were eluted from alumina by addition of 400 µl of perchloric acid (0.05 M) and shaking for 5 min. The eluate was prepared for HPLC by passing it through a 20-µm syringe-type filter (Gelman Instrument Co., Ann Arbor, MI). Separation by HPLC was performed as described below using 100-µl sample injections. Drug concentrations in cerebellum and striatal tissue were determined at each sacrifice time point. Comparisons were made of the relative drug concentrations found in the two brain regions to assess high-affinity accumulation of DHX and PrDHX. Data in each brain area were expressed as picograms per milligram of tissue.

Separation. To determine the time course of DHX accumulation in brain after s.c. dosing, we used ion-pair chromatography followed by electrochemical detection similar to the analysis of catecholamines and metabolites in brain tissue (Kilts et al., 1981). The mobile phase used to separate DHX and endogenous catechols consisted of 50 mM Na₂PO₄, 30 mM citric acid, 4 mM 18-crown-6, 86 µM sodium octyl sulfate, 0.1 mM disodium EDTA, and 11.5% acetonitrile (v/v). Diethylamine (0.1% v/v) was added to minimize secondary interactions between DHX and unreacted silanol residues. The mobile phase was adjusted to pH 4.5 before the addition of acetonitrile. The ion-pair reagent 18-crown-6 was added to the mobile phase to retain selectively the primary monoamines dopamine, DOPAC, and norepinephrine, but not the late-eluting secondary amine DHX. Dihydroxybenzoic acid was used as the internal standard with this mobile phase. When quantification of only DHX was desired, the 18-crown-6 and octyl sodium sulfate were removed and the acetonitrile concentration increased to 15%. Isocratic mobile phase flow at 0.75 ml/min was accomplished with an HPLC pump (Anspec Analytical, Geneva, IL). An Ultramex C₁₈ stainless steel column (150 × 4.6-mm i.d.) packed with 3-µm microparticulate silica (Phenomenex, Torrance, CA) was used for all separations. Electrochemical detection was accomplished with a detector (model 400; Princeton Applied Research, Princeton, NJ). The potential of the glassy-carbon working electrode was set at 0.6 V versus an Ag/AgCl reference electrode.

Prolactin Release Assay

To measure the inhibition of prolactin release after activation of D₂ dopamine receptors, male rats (350-400 g) were given i.p. injections of the serotonin precursor 5-HTP (30 mg/kg) to stimulate an increase in serum prolactin concentration. The 5-HTP pretreatment group consisted of four rats for each drug condition, whereas the control group (saline vehicle without 5-HTP) consisted of two rats. Five minutes after the 5-HTP injection, the rats were injected i.p. with DHX (10 mg/kg), quinpirole (10 mg/kg), PrDHX (5 mg/kg), or 0.1% ascorbic acid vehicle. Rats were placed singly in plastic cages containing bedding for 25 min, after which they were sacrificed by decapitation. Trunk blood was collected in chilled plastic centrifuge tubes, and allowed to clot slowly at 4°C. Serum was collected by centrifugation at 1000g for 15 min at 4°C and stored at 20°C until the time of the assay. Serum prolactin was assayed by radioimmunoassay (RIA) with reagents and procedures supplied by the Hormone Distribution Program of the National Institute of Diabetes and Digestive and Kidney Diseases.

cAMP Efflux from Striatal Slices

cAMP efflux was measured using fresh striatal tissue obtained as described above and sliced further in two directions (90° apart) using a McIlwain tissue chopper (400-µm setting). As they were being collected, slices were transferred to a test tube containing chilled, oxygenated Krebs' buffer (121 mM NaCl, 4.84 mM KCl, 1.89 mM CaCl₂, 1.16 mM MgSO₄, 1.17 mM KH₂PO₄, 25 mM NaHCO₃, 0.06 mM ascorbic acid, and 10 mM D-glucose, pH 7.35). Slices were subsequently washed with three 5-ml aliquots of Krebs' buffer and allowed to settle; slices (35 µl) were transferred to each well of the superfusion apparatus, and washed at 0.1 ml/min for 90 min. The perfusing Krebs' buffer (37°C), but not drug-containing solutions, was oxygenated with 95% O₂, 5% CO₂. Baseline aliquots of perfusate were collected for the first three 15-min intervals, after which perfusion aliquots were collected in four 10-min intervals. Differences in amounts of tissue per well were normalized by permitting each superfusion chamber to serve as its own control. The amount of drug-induced cAMP efflux was expressed as percentage of the basal efflux. Duplicate samples were analyzed by RIA (vide infra).

Adenylate Cyclase Assays in Rat Striatal Homogenates

Adenylate cyclase assays in rat striatal homogenates were performed as described by Watts et al. (1993b). A 20-µl aliquot of a 2.5-mg/ml homogenate solution was added to the reaction tubes. Baseline values of cAMP were subtracted from the total amount of cAMP produced for each drug condition. Data for each drug were expressed initially as picomoles of cAMP per milligram per minute, and converted to the percentage of stimulation produced by 5 µM forskolin.

Radioimmunoassay of cAMP

The concentration of cAMP in samples from superfusion or striatal homogenates was assayed by RIA of acetylated cAMP (Watts et al., 1993a). Normalized dose-response curves were analyzed by nonlinear regression using an algorithm for sigmoid curves using Prism (GraphPad Software). For each curve, the program provided point estimates of both the EC₅₀ and the maximal effect, either inhibition or stimulation produced (i.e., top or bottom plateau of sigmoid curve).

Tyrosine Hydroxylase Activity in Striatal Slices

Activity of tyrosine hydroxylase, the rate-limiting step in dopamine biosynthesis, was assessed by measuring the formation of [¹⁴C]carbon dioxide (¹⁴CO₂) evolved during the synthesis of dopamine from [1-¹⁴C]L-tyrosine. Aliquots (ca. 7.5 mg of tissue) of a 25-mg/ml suspension were added to tubes containing test ligands. Values of ¹⁴C cpm/30 min/assay tube, less a blank value (tissue omitted) were determined. Basal (control) tyrosine hydroxylase activity previously was found to be ca. 16.9 ± 1.2 pmol of ¹⁴CO₂/30 min/mg of protein. Data (as percentage of corresponding basal control values) were pooled and expressed as mean ± S.E.M.

In Vitro Voltammetry: Dopamine Release from Striatal Slices

Slices (350 µm in thickness) were prepared from the striatum of male Sprague-Dawley rats using a Lancer vibratome as described previously (Kelly and Wightman, 1987). The slices were submerged in a Scottish-type chamber, and perfused at 1 ml/min using a flow system equipped with a two-position, six-port valve. The perfusion medium, preheated to 32°C and saturated with 95% O₂, 5% CO₂, consisted of Krebs' buffer containing 126 mM NaCl, 2.5 mM KCl, 1.2 mM NaH₂PO₄, 2.4 mM CaCl₂, 1.2 mM MgCl₂, 25 mM NaHCO₃, and 11 mM D-glucose, pH 7.4. Nafion-coated carbon-fiber microelectrodes were prepared as described previously (Kristensen et al., 1986). Fast-scan cyclic voltammetry used a commercially available potentiostat (EI-400; Ensman Instrumentation, Bloomington, IN). A sodium-saturated calomel electrode was used as a reference electrode, and all voltages are reported versus this standard. The electrode potential was linearly scanned from 400 to 900 mV and back to 400 mV at 300 V/s every 100 ms. This voltage range encompasses the oxidation of dopamine (at 500-700 mV) and the reduction of the quinone form of dopamine (at 0 to 200 mV). The current in the voltage range of the oxidation of DA is proportional to the dopamine concentration at the electrode. To obtain DA concentration versus time plots, the current at the oxidation potential for DA, obtained for each potential scan, is integrated and converted to concentration based on postcalibration of the electrode. Thus, the time plots have 100-ms resolution in the time axis. The time response of the electrode is determined in a flow stream and was always <200 ms to reach 50% of maximum response.

Electrical stimulation was accomplished with a twisted wire, bipolar electrode with 100-µm tips (Plastics One, Roanoke, VA) separated by approximately 500 µm. Biphasic stimulation pulses, 2 ms in width each phase, were generated by a locally constructed stimulator and optically isolated by Neuro Log System stimulus isolators (Medical Systems, Greenvale, NY). Ten 300-µA pulses at 20 Hz were used. The stimulating electrodes were mounted on a micromanipulator and positioned at the surface of the slice near its center. The working electrode, mounted on a separate micropositioner, was lowered to 75 µm below the surface of the slice at a position 100 to 200 µm from the center of the stimulating electrode pair. Electrode placements were made with the aid of a stereomicroscope (Bausch & Lomb, Rochester, NY). Evoked DA overflow was first determined by obtaining responses to electrical stimulation with the slices perfused with physiological buffer. Then, physiological buffer containing drug was applied for 20 min and the responses recorded again. Data were initially expressed as current obtained after release, subtracting background current obtained before release. Values for dopamine D₂ agonists were then converted to a percentage of electrically stimulated release. Data are reported as mean ± S.E.M. of three to six slices, typically from different animals.

Cerebral Microdialysis in Freely Moving Animals

Guide cannulae (CMA/12; Bioanalytical Systems) were implanted with the aid of stereotaxic instruments (David Kopf Instruments, Tujunga, CA) at the dorsal surface of the body of the caudate putamen in rats anesthetized with pentobarbital (35 mg/kg i.p.) and ketamine (90 mg/kg i.m.). Relative to bregma, the coordinates were 1 mm anterior, 3 mm lateral, and 4 mm below skull surface. The cannula was cemented in place using cranioplastic cement and stainless steel mounting screws (Plastics One). The animals were housed individually for recovery. Rats with implanted guide cannulae were transferred to an isolated room 7 days later, placed in experimental cages (CMA/120 microdialysis system for awake animals; Bioanalytical Systems), and probes inserted (CMA/12, 4-mm fiber length, concentric design; Bioanalytical Systems). Probes were perfused at 2 µl/min (model 22 syringe pump; Harvard Apparatus, South Natick, MA) with a physiological Ringer's solution [120 mM NaCl, 1.2 mM CaCl₂, 5 mM KCl, 1.2 mM MgCl₂, and 0.15% (v/v) phosphate-buffered saline, pH 7.4]. Twelve to 18 h after probe insertion, samples were collected with the aid of a refrigerated fraction collector (CMA/170 refrigerated fraction collector; Bioanalytical Systems) at 20-min intervals. After collection of six to eight baseline samples, drugs were administered by s.c. injection or through inclusion in the dialyzing solution. Postdrug samples were collected for a 2- to 3-h period. Samples for dopamine overflow studies were collected in tubes containing a small volume of the HPLC mobile phase described below, and containing isoproterenol as an internal standard. Samples were frozen immediately and analyzed for dopamine, DOPAC, HVA, and 5-HIAA by using reverse phase HPLC-EC.

Quantitative determinations for dopamine and metabolites were made using the same stationary phase and detector as described above for the DHX accumulation studies. The electrode potential was set at +0.75 V versus a Ag/AgCl reference electrode. The mobile phase consisted of 0.05 M Na₂HPO₄, 0.03 M citric acid, 0.1 mM disodium EDTA, 2 mM sodium octyl sulfate, and 20% methanol, with a final pH of 3.4 and a flow rate of 0.80 ml/min. The column temperature was maintained at 39°C. Standard curves for the quantification of dopamine, DOPAC, and HVA were prepared by analyzing a series of solutions containing a fixed amount of isoproterenol and varying amounts of each compound. The HPLC data were collected and digitized by Analytic 900 series modules and Turbochrom software (PerkinElmer Instruments, Norwalk, CT) by measuring the peak height for each compound relative to the internal standard. Absolute values for each analyte were then normalized relative to the baseline values obtained before drug administration.

Electrophysiology: Firing of Substantia Nigra Dopamine Neurons

Electrophysiological studies were performed in accordance with the methods described by Piercey et al. (1997). Briefly, adult rats were anesthetized with chloral hydrate (400 mg/kg i.p.) and the femoral artery and vein were cannulated for drug administration and blood pressure monitoring, respectively. Drugs, measured as their salt, were dissolved in distilled water and injected intravenously in volumes of 0.15 ml or less. Stereotaxic coordinates were used for electrode placement in the substantia nigra pars compacta. Glass microelectrodes filled with pontamine blue in 2 M NaCl (4-10-M impedance) were used for extracellular recordings. Dopaminergic neurons were identified by their long-duration action potentials (>2.5 ms), shape, and firing rates (Bunney et al., 1991). Changes in neuronal firing rates were monitored as described previously (Piercey et al., 1997). Dose-response studies were based on cumulative dosing schedules. Injections were timed 1 to 2 min apart to allow maximal response, as well as sufficient accumulation of drug to approximate concentrations expected with single bolus injections.

D₂ Receptor Autoradiography

Tissue Processing. D₂ receptor autoradiography buffer contained 50 mM HEPES, 120 mM NaCl, 5 mM KCl, 2 mM CaCl₂, and 4 mM MgCl₂, pH 7.4. D₂ receptors were labeled with 0.03 nM ¹²⁵I-epidepride, with nonspecific binding defined by 1 µM domperidone. Twelve consecutive sections were incubated with one of 12 concentrations of dihydrexidine (0.01 nM to 100 µM). Ketanserin (400 nM) was included in all mailers to mask binding of the radioligand to 5-hydroxytryptamine₂ sites. Slides were incubated at room temperature for 4 h. Dried sections were then apposed to Kodak X-OMAT RP film, along with brain paste standards containing known amounts of radioactivity (Amersham Biosciences). Autoradiograph cassettes were stored in the dark at room temperature then the films developed after appropriate exposure duration (usually ca. 2-3 days). Films of substantia nigra sections were exposed for substantially longer periods due to much lower levels of D₂ binding compared with caudate nucleus.

Image Analysis. Films were developed and MCID image processing system (Imaging Research, St. Catherine's, Ontario, Canada) was used to construct a standard curve relating radioactivity to film density, and to quantify the amount of radioactivity bound to tissue. Film images of nonspecific binding were not visible (<5% of total binding); thus, nonspecific binding was not subtracted from either total binding or competition points. Prism (GraphPad Software) was used to obtain curve-fitting parameters (K_0.5 and n_H) from nonlinear regression analyses of competition curves.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Competition Binding in Rat Striatum. DHX has affinity for D₂ receptors labeled with [³H]spiperone (Fig. 2; Table 1), as well as for sites labeled with the D₁ receptor antagonist [³H]SCH23390 (Table 1). Compared with DHX, the analog PrDHX has 100-fold lower affinity for [³H]SCH23390 sites, and somewhat higher affinity for [³H]spiperone sites. Thus, DHX displays approximately 10-fold D₁/D₂ receptor selectivity, whereas PrDHX displays the reverse selectivity. The affinities of both DHX and PrDHX for [³H]spiperone sites are similar to that of the prototypical D₂ receptor agonist quinpirole. Also similar to quinpirole, the slopes of the competition curves of DHX and N-propyl-DHX for [³H]spiperone-labeled D₂ sites are shallow, whereas the competition curve for the antagonist chlorpromazine displays normal steepness. Preincubation of striatal membranes with 100 µM GppNHp produces 2- to 3-fold rightward shifts in the competition curves for DHX and PrDHX, although no curve steepening is detected (Fig. 2, A and B); the antagonist chlorpromazine is unaffected by GppNHp. The high-affinity interaction of DHX and PrDHX with D₂ receptors is evidenced also by the low nanomolar K_0.5 of both compounds for D₂ receptor sites labeled with the agonist [³H]NPA (Table 1; Fig. 2C).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Competition of DHX and PrDHX for dopamine D2 receptor sites in rat striatum. A, competition by DHX in the presence or absence of GppNHp. B, competition by PrDHX in the presence or absence of GppNHp. C, competition of DHX and PrDHX for [3H]NPA binding sites. The curves shown are from representative assays conducted in triplicate. Assays were repeated at least three times. Parameter estimates for receptor affinity (K0.5) and indirect Hill slope (nH) are shown in Table 1.

Demonstration of Bioavailability of DHX and PrDHX. Darney et al. (1991) and Smith et al. (1997) reported that administration of DHX and PrDHX by s.c. injection produces acute effects on behaviors that are known to be modulated by dopamine receptors in the central nervous system. We developed a sensitive HPLC-EC detection method to allow direct demonstration of the time course and extent of accumulation of DHX and PrDHX in dopamine terminal areas. These data provide a basis for evaluating the effects of these compounds in whole animal D₂ receptor functional assays that entail peripheral drug administration.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Time course of the accumulation of DHX in brain. The concentration of DHX was determined in striatum and cerebellum at different intervals after administration of 3 mg/kg DHX s.c. to adult rats. Each data point represents the mean ± S.E.M. of three to four rats.

Prolactin Release. One well characterized functional effect of D₂ receptor activation is the inhibition of prolactin release in vivo. Dopamine and D₂ agonists inhibit prolactin release, whereas D₂ receptor antagonists increase prolactin release by inhibiting the effects of endogenous dopamine. Prolactin release was stimulated by the serotonin precursor 5-HTP, and rats were subsequently treated with the test drugs. DHX (10 mg/kg) significantly inhibited the release of prolactin, as did quinpirole (10 mg/kg; Fig. 4), consistent with our previous report (Mottola et al., 1992). At a dose of 5 mg/kg, PrDHX inhibited the release of prolactin nearly as well as DHX, although its response is somewhat less than that of quinpirole.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Effects of DHX and PrDHX on serum prolactin levels. Rats were pretreated with 5-HTP (30 mg/kg i.p.) to increase serum prolactin then administered either DHX (10 mg/kg i.p.), quinpirole (Quin) (10 mg/kg i.p.), or PrDHX (5 mg/kg). Data are expressed as a percentage of the prolactin measured in control subjects receiving 5-HTP alone. Values for 5-HTP pretreatment represent the means ± S.E.M. of four rats per condition. Serum prolactin in saline-pretreated control rats (n = 2) was 34 ± 4 ng/ml. , p < 0.05.

cAMP Efflux from Striatal Slices. As shown in Fig. 5, in striatal minces, DHX or SKF38393 stimulates cAMP efflux. The net effect for DHX and dopamine has been hypothesized to represent a balance between D₁ receptor-mediated stimulation and D₂ receptor-mediated inhibition. Consistent with this, when D₂ receptors are blocked with the selective D₂ antagonist domperidone (10 µM), the total amount of cAMP in the superfusate is increased. In contrast, the addition of domperidone does not affect cAMP efflux induced by the selective D₁ agonist SKF38393. These data are consistent with the idea that DHX activates striatal D₂ receptors that are negatively coupled to adenylate cyclase, and replicate a previous study with DHX (Mottola et al., 1992) in which the D₂ antagonist sulpiride was used. Thus, the net effect of DHX or DA on cAMP efflux can be interpreted as being a balance between D₁ receptor stimulation and D₂ receptor inhibition due to the mixed agonist properties, whereas SKF38393 acts only at D₁ receptors in this preparation.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Effects of DHX on cAMP efflux from rat striatal slices in the presence or absence of the D2 receptor antagonist domperidone. Slices were prepared from rat striatum and perfused with physiological buffer containing DHX (1 µM) or SKF38393 (1 µM). These data represent means ± S.E.M. from at least three assays conducted in duplicate. *, significantly different from control P < 0.001; #, significantly different from DHX alone, P < 0.05.

Inhibition of Forskolin-Stimulated cAMP Synthesis in Striatal Membranes. The preceding cAMP efflux data with ex vivo superfused striatal slices indicate that DHX has agonist properties at D₂ receptors, but this technique has practical limitations on the number of samples that can be studied in a single assay. For this reason, further studies evaluating D₂ agonist inhibition of adenylate cyclase were performed in striatal membranes. The levels of cAMP were elevated artificially using 5 µM forskolin to permit easier evaluation of the effects of activation of D₂ receptors. This concentration of forskolin produces a 10- to 15-fold increase in basal levels of cAMP under our assay conditions.

View larger version (30K): [in this window] [in a new window]   Fig. 6.   Effects of dopamine receptor agonists on forskolin-stimulated cAMP accumulation in rat striatal homogenates. Forskolin (5 µM) was used to stimulate cAMP synthesis. Under these conditions, the effects of mixed dopamine receptor agonists on cAMP accumulation reflect a combination of stimulatory and inhibitory effects through dopamine D1 and D2 receptors, respectively. The effects of test compounds were assessed in the presence or absence of the D1 receptor antagonist SCH23390 (SCH) and/or the D2 receptor antagonist sulpiride (sulp) to evaluate the contribution of each receptor subtype. Data are expressed as percentage of the level of cAMP production stimulated by 5 µM forskolin alone. Each value represents the mean ± S.E.M. of duplicate samples from a single representative assay. Data were analyzed by nonlinear regression with a sigmoid equation using Prism. Basal values of cAMP were 145 ± 12 fmol/mg/min, whereas values for stimulation with 5 µM forskolin were 1790 ± 53 fmol/mg/min. Assays for each test compound were repeated at least three times. Quin, quinpirole.

Effects on Dopamine Synthesis. Although cAMP efflux in striatum is largely mediated by D₂-like receptors on target cells, there are D₂-like receptors on processes of dopamine neurons that control dopamine biosynthesis via modulation of tyrosine hydroxylase (TH), the rate-limiting enzyme of the dopamine synthetic pathway. Dopamine and D₂ agonists activate these autoreceptors, resulting in a decrease in the activity of TH. As shown in Fig. 7, a typical D₂ receptor agonist such as quinpirole inhibited TH activity in a manner that is reversed by spiperone, but not SCH23390. In contrast, although DHX and PrDHX (1 µM) both induced decreases in TH activity (Fig. 7), these effects were not reversed by spiperone. It thus appears that DHX and other members of the benzo[a]phenanthridine class inhibit the enzyme by a mechanism that is not receptor-mediated. Table 3 illustrates data in support of the hypothesis that any catechol may suppress tyrosine hydroxylase activity by a non-D₂ receptor-mediated mechanism (Nagatsu et al., 1964). Although both (R)-NPA and (S)-NPA inhibit TH, their effects are reversed by D₂ antagonists. Conversely, several catechol-containing drugs (the selective D₁ partial agonist SKF38393, as well as DHX or PrDHX) cause TH inhibition that is not reversed by D₂ antagonists. Moreover, the noncatechol partial D₁ agonist CY 208,243 has no effect, despite having a backbone similar to DHX.

View larger version (37K): [in this window] [in a new window]   Fig. 7.   Effects of dopamine receptor agonists on dopamine synthesis in rat striatal slices. TH activity was assessed via formation of 14CO2 from [1-14C]L-tyrosine. DHX, PrDHX, and quinpirole were used at 1 µM, whereas spiperone and SCH23390 were used at 0.5 µM. Data are expressed as a percentage of basal control values (16.9 pmol of 14CO2/30 min/mg of protein) and represent means ± S.E.M. from at least three independent samples. All values except for quinpirole + spiperone were significantly different from control (P < 0.001).

View this table: [in this window] [in a new window]   TABLE 3 Effects of DHX, PrDHX, and reference agonists on dopamine synthesis in rat striatal slices from normal rats TH activity was evaluated by measuring the formation of 14CO2 evolved during the synthesis of dopamine from [1  14C]L-tyrosine. Data were pooled (n  3) and expressed as mean ± S.E.M. percentage of corresponding basal control values (16.9 pmol 14CO2/30 min/mg protein).

View this table: [in this window] [in a new window]   TABLE 4 Effects of DHX and PrDHX on dopamine synthesis in reserpine-treated rats Data were pooled (n  3) and expressed as mean ± S.E.M. percentage of corresponding basal control values (16.9 pmol of 14CO2/30 min/mg protein).

Effects on Dopamine Release. To assess whether DHX (or the N-propyl analog) affects the release-modulating D₂ autoreceptors, fast-scan cyclic voltammetry was used to measure dopamine released from striatal slices after electrical stimulation. This technique is well suited for assessing autoreceptor effects on dopamine release because of the superior time resolution of the sampling (Kennedy et al., 1992). As shown in Fig. 8, neither DHX nor N-propyl-DHX significantly altered dopamine overflow after stimulation of striatal slices. In contrast, quinpirole inhibited dopamine release by greater than 50%, an effect that was blocked by the D₂ receptor antagonist sulpiride (data not shown). Thus, these benzo[a]phenanthridine analogs do not appear to activate D₂ autoreceptors controlling dopamine release.

View larger version (42K): [in this window] [in a new window]   Fig. 8.   Effects of DHX and N-propyl-DHX on stimulated dopamine release from rat striatal slices. Background-subtracted fast-scan cyclic voltammetry was used to monitor and measure dopamine overflow after electrical stimulation. DHX (1 and 10 µM) and PrDHX (1 and 10 µM) were compared with the D2 agonist quinpirole (Quin) (1 µM). Results are expressed as means ± S.E.M. of percentage of pulse-stimulated dopamine release (control). Data were obtained from at least three separate experiments for each condition, except 10 µM PrDHX (n = 1). , P < 0.05.

View larger version (26K): [in this window] [in a new window]   Fig. 9.   Effects of DHX on basal dopamine overflow in the striatum of freely moving rats. Agonists were administered by s.c. injection (A and B) or by inclusion in the dialysate (C). Samples were collected at 20-min intervals and analyzed for dopamine content by HPLC-EC. Data are expressed as a percentage of the three basal samples collected before drug administration. The mean ± S.E.M. for basal dopamine were 17.0 ± 1.8 pg/sample. This figure is from one of four experiments with essentially identical results. Quin, quinpirole; Hal, haloperidol.

Effects on Dopamine Neuron Firing. Two separate experiments were performed to evaluate the ability of DHX to activate D₂ autoreceptors coupled to dopamine cell firing. The first experiment (Fig. 10A) demonstrated that, as expected, the D₂ agonists quinpirole and U-86170, as well as the mixed agonist apomorphine (Fig. 10B), inhibit the firing of substantia nigra pars compacta dopamine neurons via activation of somatodendritic autoreceptors. Conversely, DHX has no such effect at i.v. doses up to 1 mg/kg. In the second study (Fig. 10C), the total cumulative dose of DHX was increased to 3 mg/kg. Again, no effects on firing were detected. In two of the four cells in which a full DHX dose-response curve was performed, apomorphine was given after the last dose of DHX. In the other two cells, quinpirole was given after the DHX. In all cases, these same cells then responded to the latter administration of these typical dopamine agonists (Fig. 10B).

View larger version (22K): [in this window] [in a new window]   Fig. 10.   Effect of DHX on firing of substantia nigra neurons. A, dose-dependent inhibition of single cell firing resulted from administration of quinpirole or U-86170, but not from administration of dihydrexidine or the selective partial D1 agonist SKF38393. B, dihydrexidine did not attenuate the effects of apomorphine administered immediately after the highest dose (ordinate, firing rate; abscissa, time; arrows, i.v. administration of noncumulative drug dose). C, in a separate experiment, higher doses of dihydrexidine still failed to cause significant inhibition of cell firing (n = 4 cells).

Competition Binding in Substantia Nigra. One hypothesis to explain the lack of functional presynaptic or autoreceptor effects of DHX or PrDHX is that these ligands do not bind to those D₂-like receptors that mediate these functions. In the striatum, the levels of D₂-like autoreceptors are much lower than postsynaptic D₂ receptors (ca. 10% or less). Because there are no ligands selective for only one of these populations, it is impossible to assess a selective binding mode for DHX in this brain region. On the other hand, D₂-like dopamine receptors in the substantia nigra are located exclusively on DA soma or dendrites (i.e., there are no postsynaptic D₂ receptors in this region; Morelli et al., 1988). Thus, D₂ competition assays using receptor autoradiography were performed to compare the binding of DHX to D₂ receptors in the striatum with that in the substantia nigra. As shown in Fig. 11, DHX competed similarly for ¹²⁵I-epidepride-labeled D₂ sites in the substantia nigra and striatum of the same animal. The affinity of DHX was essentially identical in both brain regions (K_0.5 values: striatum = 470 nM, substantia nigra = 340 nM). In three replicate experiments, there also was no significant difference in the slope of the competition curves.

View larger version (15K): [in this window] [in a new window]   Fig. 11.   Competition of DHX for D2 receptors in the substantia nigra and striatum. Nonspecific binding was defined with 1 µM domperidone. Data represent the mean ± S.E.M. for four (s. nigra) or five (striatum) series of consecutive sections from a single animal. The experiment was repeated three times with comparable results.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Although the binding characteristics of DHX and PrDHX suggested they would be typical agonists (Mottola et al., 1992), the current work demonstrates unprecedented functional properties in native tissue (Table 5). DHX has clear agonist effects at some functions mediated by D₂-like receptors (e.g., inhibition of adenylate cyclase or prolactin secretion), yet binds to other receptors of the same type where it apparently has antagonist, or no, functional actions (e.g., regulation of dopamine synthesis, cell firing, and release). A similar functional pattern was found with PrDHX, suggesting that at least some other hexahydrobenzo[a]phenanthridines share these novel functional properties. These data suggest that some drugs may, after binding to a single receptor isoform, cause grossly disparate functional effects, a phenomenon we have labeled functional selectivity.

Non-D₂ Dopamine Receptor Mechanisms and Functional Selectivity. There are a variety of mechanisms that could falsely lead to this notion of functional selectivity. A failure to observe effects of DHX might be due to pharmacokinetic mechanisms (e.g., failure of DHX to reach appropriate receptors). The current data, however, show that parenteral administration of DHX or PrDHX leads to accumulation of the drugs in dopamine terminal areas, and Darney et al. (1991) and Smith et al. (1997) show that both DHX and PrDHX cause behavioral effects. Similarly, the autoradiographic studies demonstrate that DHX can access and bind to D₂-like receptors in brain slices even though it has little intrinsic activity neurochemically, even when given directly through a microdialysis probe. Thus, functional selectivity cannot be explained by differences in distribution or metabolism.

Functional Selectivity and Dopamine D₂-Like Receptors. Another possibility is that the current data are a consequence of low DHX intrinsic efficacy at all D₂-like receptors. Yet, the literature data suggest that there is higher efficiency of dopamine autoreceptor coupling due to greater autoreceptor receptor reserve when measured in vivo (Meller et al., 1987; Yokoo et al., 1988; Cox and Waszczak, 1989). Autoreceptor inhibition of DA cell firing usually is sensitive to low intrinsic activity partial agonists; DHX did not cause such activation. Conversely, DHX was a full agonist at striatal D₂ receptors coupled to adenylate cyclase where both DHX potency and intrinsic activity were indistinguishable from quinpirole. Agonists of known low intrinsic efficacy [e.g., S-()-3-PPP and S-()-NPA] are unable to elicit maximal inhibition of adenylate cyclase. Thus, differences in receptor reserve cannot explain these data.

Cellular and Molecular Mechanisms. In addition to their expression on presynaptic terminals of nigrostriatal afferents, D₂-like receptors in striatum are expressed on giant cholinergic interneurons, medium spiny striatal efferent neurons, and possibly on the terminals of corticostriatal glutamatergic afferents (Gerfen et al., 1990; Weiner et al., 1991). The major source of D₂ inhibition of adenylate cyclase probably arises from D₂ heteroreceptors on cell bodies within the striatum, because this inhibition is lost when striatal cell bodies are ablated with kainic acid (Memo et al., 1986). This form of inhibition also can be reproduced entirely in cultures of striatal neurons (Onali et al., 1985). If one assumed that inhibition of adenylate cyclase in striatum is due to activation of D₂ receptors on target striatal cells then one explanation of the available data is that DHX discriminates functionally between D₂-like autoreceptors and heteroreceptors. Yet, a recent electrophysiological study (Ruskin et al., 1998) demonstrated that DHX did not activate D₂-like heteroreceptors on globus pallidus neurons that mediate firing rate increases. Thus, although the present data seem to suggest auto- versus heteroreceptor functional selectivity, DHX also may exhibit selectivity among other postsynaptic D₂ receptor functions, such as those controlling acetylcholine release (Stoof and Kebabian, 1982).