Introduction g2 subunit is determined by the presence or absence of an

Brain Research 887 2000 23–33 www.elsevier.com locate bres Research report Regulation of benzodiazepine receptor binding and GABA subunit A mRNA expression by punishment and acute alprazolam administration Min Liu, John R. Glowa Department of Pharmacology and Therapeutics , Louisiana State University Health Sciences Center in Shreveport, 1501 Kings Highway, P .O. Box 33932, Shreveport, LA 71130-3932, USA Accepted 12 September 2000 Abstract Quantitative autoradiography of benzodiazepine BZ receptors and competitive reverse transcription–polymerase chain reaction were used to characterize changes in BZ binding and GABA receptor subunit transcription levels associated with the anxiolytic effects of A alprazolam. Effects were assessed on punished and non-suppressed water consumption using a lick suppression Vogel paradigm. 3 Alprazolam had no effect on non-suppressed licking, [ H]Ro 15-1788 binding or receptor subunit transcript levels, compared to non-drug controls. When each fifth lick produced a shock 0–0.5 mA, responding was suppressed in an intensity-related manner. The highest 3 intensity significantly decreased licking 85, [ H]Ro 15-1788 binding 12 and a1 transcript levels 63 in the basolateral nucleus 3 of the amygdala, and [ H]Ro 15-1788 binding in the mediodorsal thalamic nucleus 15, compared to non-punished controls. Punishment increased the ratio of g2L S transcripts in the basolateral nucleus of the amygdala. Alprazolam blocked or reversed each of these effects. These results show that punishment has similar effects on BZ binding and GABA receptor subunit expression and that A alprazolam can block or reverse those effects. Such changes may be related to the anxiolytic effects of alprazolam.  2000 Elsevier Science B.V. All rights reserved. Theme : Neural basis of behavior Topic : Motivation and emotion Keywords : Benzodiazepine; Alprazolam; GABA receptor; Receptor subunit; Punished responding; Anxiety A

1. Introduction g2 subunit is determined by the presence or absence of an

eight amino acid sequence, respectively [38]. The g2L g-Aminobutyric acid GABA is the principal inhibitory subunit encodes a sequence that can be phosphorylated by 21 neurotransmitter in the vertebrate central nervous system. protein kinase C [37] and by a Ca calmodulin-depen- Its actions on GABA receptors mediate a bicuculline- dent protein kinase II [21]. The g2S subunit most likely A 2 3 sensitive Cl ion channel that is thought to play a major binds to the b subunit; site-directed mutagenesis and [ H] role in the neurochemistry of stress and anxiety. Native muscimol binding studies suggest two homologous do- GABA receptors are most likely heteropentameres, co- mains of this subunit are critical for activation of the A assembled from a variety of possible subunits e.g., a1–6, GABA receptor [5]. Receptors containing different A b1–3, g1–3, d, e, p and r1–3 [4,23,26]. While co- combinations are topographically distributed, predominant- expression of a-, b-, and g-subunits is a likely requirement ly in cortical, limbic and cerebellar structures. for a fully functional GABA receptor, alternative splice Benzodiazepines BZs pharmacologically modulate A variants further increase possible receptor configurations. GABA receptors through an allosteric binding site. They A For example, a ‘long g2L’ or ‘short g2S’ variant of the are the principal pharmacological means of treating anxiety and stress. BZ binding is thought to require both a and g subunits, and different a subunits can determine the Corresponding author. Tel.: 11-318-675-4803; fax: 11-318-675- affinity of different BZs for the GABA complex. While 7857. A E-mail address : jglowalsumc.edu J.R. Glowa. diazepam does not discriminate among GABA receptors A 0006-8993 00 – see front matter  2000 Elsevier Science B.V. All rights reserved. P I I : S 0 0 0 6 - 8 9 9 3 0 0 0 2 9 6 2 - 0 24 M composed of a1–3, or a5 subunits, the imidazoben- 2. Materials and methods zodiazepine Ro 15-4513 prefers receptors constructed from a4 or a6 subunits. g2 subunits can confer further spe- 2.1. Materials cificity for BZs. Diazepam-sensitive sites are primarily located in the cerebral cortex, cerebellum, amygdala, Alprazolam Upjohn Co., Kalamazoo, MI was dis- hippocampus, and hypothalamus. These sites have been solved in 10 alcohol, 10 alkamuls EL-620, and 80 3 further distinguished as BZ1 and BZ2, based on their saline vehicle. [ H]Ro 15-1788 87.0 Ci mmol was pharmacology [26]. purchased from New England Nuclear Boston, MA. The While the role of GABA and BZ receptors in stress is GABA receptor a1, b2, g2S and g2L subunit internal A well accepted, there is growing evidence to suggest that standards, cloned in pGEM 1 and containing targeted stress, and possibly anxiety, can affect GABA physiology. restriction enzyme cleavage sites Bgl II, were obtained For example, early reports showed that stressful situations from Dr Dennis R. Grayson. Sph I and Bgl II were can down-regulate GABA receptor density [2]. BZ bind- purchased from Promega Co. Madison, WI. Polymerase A ing can also undergo rapid changes after exposure to stress chain reaction PCR primer pairs were synthesized by [37]. This modulation can depend on several factors Life Technologies Gaithersburg, MD. Ready-To-Go RT– including the type of site, the nature of the stressful event, PCR beads, DNA and the gel band Purification kit were the duration of stress, as well as psychological factors such purchased from Pharmacia Biotech Piscataway, NJ. All as escapability or specificity of conditioning [8,9,40]. As other chemicals used in this study were obtained from few reports have explored whether anxiety-related situa- Sigma Chemical Co. St Louis, MO. tions affect binding through actions on specific GABA A subunits, the current study pursued this possibility using 2.2. Subjects the high-affinity triazolobenzodiazepine alprazolam [6]. As methods to assess such effects in humans are not yet Seventy-two male Sprague–Dawley rats, weighing 250– available, a preclinical model was used to study the 300 g, were used. The animals were kept individually anxiolytic effects of this drug. under standard controlled conditions temperature |228C; The preclinical anxiolytic effects of BZs are often humidity |55; normal phase 12-h light–dark cycle for at studied on punished suppressed responding, because least 1 week after arrival, and during the experiments. increases in punished responding can distinguish an an- Food and water were freely available prior to beginning xiolytic effect from other types of drug effects e.g., the experimental procedures. anticonvulsant, sedative, hypnotic, amnesic, appetite stimulating. The potencies with which different BZs 2.3. Procedure exhibit anti-punishment effects, as well as their affinities at BZ receptors, are well correlated with their clinical poten- Conflict testing was carried out according to the method cies [20]. Increases in punished responding can also be of Vogel [36], with minor modifications. Animals were obtained in humans [1], allowing for follow-up studies water-restricted for 36 h before the experiment. On the first when appropriate methods are developed. A number of two days, water-restricted subjects were placed in the studies in animals have shown that direct injections of BZs testing chamber 28320320 cm with a stainless steel grid or GABA receptor agonists into limbic sites can increase floor and a metal drinking spout connected to a constant A punished responding [17,30,33]. In addition, intra- current shock generator Model E53-13, Coulbourn Instru- amygdala injections of the BZ antagonist Ro 15-1788 can ments, Lehigh Valley, PA. Rats were allowed to find the block BZ-induced increases in punished responding [34]. drinking spout and consume water without shock for 10 These studies suggest a role for limbic sites in the min. The number of licks was detected through a drin- anxiolytic effects of BZs. A preliminary study found that kometer circuit. After each session, rats were allowed to the anxiolytic effects of alprazolam were also associated drink water for 60 min in their home cages but otherwise with changes in GABA receptor subunit composition in remained water-restricted. During subsequent experimental A the amygdala of the rat [19]. As such, the current study sessions the rats were placed into the apparatus and focussed on that structure. Quantitative receptor au- allowed access to water without shock for 5 s, then a shock toradiography was used to confirm the topographic dis- circuit was activated. A constant current shock generator tribution of BZ sites and to assess changes in binding in connected to the floor grids and the drinking spout. Shock response to punishment, alprazolam, and their combina- was delivered at each fifth lick, and the apparatus recorded tion. Then, a quantitative, competitive reverse transcrip- the number of shocks delivered. Session length was always tion–polymerase chain reaction RT–PCR assay was used 10 min. All experiments were carried out between 9.00 to assess the expression of genes associated with a1, b2, a.m. and 11.00 a.m. Initial experiments assessed the effects g2S and g2L subunits of the GABA receptor in the of shock intensity 0.0–0.5 mA and then at 0.5 mA, A amygdala under the same conditions. alprazolam dose 0.3–4.8 mg kg. For subsequent experi- M . Liu, J.R. Glowa Brain Research 887 2000 23 –33 25 3 ments, animals were divided into six groups n56 group: standards Amersham. The non-specific binding of [ H]Ro saline control without shock, vehicle-treated without 15-1788 was less than 5 of the total binding. The 3 shock, alprazolam-treated without shock, saline control analyses were performed on total [ H]Ro 15-1788 binding. with shock, vehicle-treated with shock, and alprazolam- treated with shock. A shock intensity of 0.5 mA was used. 2.5. Competitive RT–PCR: basolateral amygdaloid Saline, vehicle and alprazolam 1.2 mg kg were given i.p. nucleus in a volume of 1 ml kg body weight, 30 min before the test session. Immediately following the last test session, Competitive RT–PCR reactions using internal standards animals were removed to an adjacent room and decapitated specific for GABA receptor subunits were conducted A without anesthesia. Two separate experiments were con- according to methods described previously [12]. Plasmids ducted n536 each, one for the autoradiographic evalua- containing mutant cDNA specific for a1, b2, 2gS, and 3 tion of [ H]Ro 15-1788 binding and one for competitive 2gL subunits were transformed into JM109 competent RT–PCR. cells, grown in LB medium with ampicillin, and extracted by alkaline lysis plasmid Midipreps. For in vitro transcrip- 3 2.4. Autoradiographic evaluation of [ H]Ro 15-1788 tion, these plasmids were linearized by Sph I, and were binding in selected brain sites transcribed by T7 RNA polymerase according to the manufacturer’s Promega Co. instructions. Concentrations The rat brains were rapidly removed, frozen in dry of cRNA of each subunit were quantified at A by UV 260 ice-chilled isopentane, wrapped in aluminum foil, and spectrophotometry and used as an internal standard. stored at 2868C. Coronal sections 10 mm were cut from Tissue was micropunched from basolateral amygdaloid selected brain areas in a cryostat-microtome, maintained at nucleus the only region studied by using an 18-gauge 2198C, and mounted onto gelatin–chrome–alum-coated stainless-steel tube made from a hypodermic needle. RNA slides at room temperature. At each brain level studied was extracted and isolated using RNAgents Total Isolation 16.7 mm and 14.2 mm, interaural, two sections were System according to the manufacturer’s protocol Promega taken: one for total binding and one for non-specific Co., briefly described below. Tissue was homogenized in binding. The slides were dried in a stream of air for 1 h, denaturing solution 26 mM sodium citrate, pH 4.0, 0.5 packed into boxes containing silica gel and stored at N-lauryl sarcosine, 0.125 M b-mercaptoethanol, 4 M 2868C until required for binding. On the day of the aguanidine thiocyanate, mixed with sodium acetate 2 M, analysis, the slides were allowed to reach room tempera- pH 4.0 and phenol: chloroform: isoamyl alcohol ture in the boxes. The brain sections were subjected to 25:24:1, and centrifuged at 10,000 g for 20 min at 48C. osmotic shock to eliminate endogenous GABA [22] by a RNA was precipitated from the aqueous phase by the 3-min wash in room temperature distilled deionized water, addition of isopropanol, washed with 75 ethanol, and followed by two 5-min rinses in ice-cold buffer, then dried dissolved in DEPC-treated water. The yield of total RNA for 1 h in a stream of cold, dry air. was determined by measuring the absorbance at 260 280 3 BZ receptors were visualized using [ H]Ro 15-1788 nm. Removal of contaminating DNA was confirmed by under standard autoradiographic conditions [39]. Briefly, PCR analysis of total RNA samples without reverse slide-mounted tissue sections were incubated for 40 min at transcription. The RNA was stored at 2868C until the 3 48C with 4 nM [ H]Ro 15-1788 87 Ci mmol, New competitive RT–PCT assay was run. England Nuclear, Boston, MA in 0.17 M Tris–HCl pH Various amounts of cRNA prepared from the appropriate 7.4 at 48C. Non-specific binding was estimated with cold standard template were added to a constant amount of total 26 Ro 15-1788 10 M Ro 15-1788 in the cold incubation RNA 1 mg tube. Reverse transcription and PCR amplifi- solution. After incubation, the slides were washed for 2 cation were performed in a single tube, which contained a min in cold buffer to reduce non-specific binding. The Ready-To-Go RT–PCR bead. Each reaction 50 ml final slides were then dipped briefly in ice-cold distilled water volume contained 10 mM Tris–HCl pH 9.0, 60 mM and dried immediately under a stream of cool, dry air. KCl, 1.5 mM MgCl , 200 mM of each dNTP, Moloney 2 Slides were affixed to a mounting board, placed in X-ray Murine Leukemia Virus M-MuLV reverse transcriptase, cassettes with radioactive standards Amersham, Arlington 2.0 units of Taq DNA polymerase, ribonuclease inhibitor 3 Heights, IL and apposed to tritium-sensitive film H- porcine and stabilizers, including RNasa DNasa-free Hyperfilm; Amersham. After a 3-week exposure, the films BSA. Each reaction also contained 1 mM each of 59 sense were developed using D19, according to the manufactur- and 39 antisense subunit-selective primers, described in er’s instructions Kodak, Rochester, NY. Table 1. The reaction was overlaid with mineral oil and The autoradiograms were quantified using computer- placed in a Perkin–Elmer 2400 thermocycler at 428C for assisted microdensitometry. Optical density measurements 30 min. At the end of the reverse transcription reaction, the were determined with an image analyzer NIH Image 1.57 thermocycler was programmed to increase temperature to 3 and converted to fmol per mg of tissue by reference to H 958C for 5 min to inactivate the reverse transcriptase and 26 M Table 1 GABA receptor subunit-selective RT–PCR primers A Subunit Primer sequence 59 to 39 Position Product size a1 sense AGCTATACCCCTAAC 1178–1482 304 TTAGCCAGG a1 antisense AGAAAGCGATTCTC AGTGCAGAGG b2 sense TGAGATGGCCACAT 1201–1518 317 CAGAAGCAGT b2 antisense TCATGGGAGGCTGG AGTTTAGTTC g2S sense AAGAAAAACCCTGCCCCTACAATT 1156–1492 336 g2S antisense TTCGTGAGATTCAGCGAATAAGAC g2L sense CTTCTTCGGATGTTTTCCTTCAAG 1144–1534 390 g2L antisense CATAGGGTATTAGATCGTTGGACT to completely denature the template. The thermocycler was applied to the drinking tube, responding decreased P, then programmed for 32 cycles, with each cycle consisting 0.001 as a function of increasing shock intensity Fig. 1a, of a temperature of 958C for 45 s, 558C for 45 s, and 728C demonstrating that responding was punished. Intensities for 45 s, followed by a final elongation step 728C for 6 greater than 0.2 mA significantly decreased responding min. PCR products were digested overnight with 10 units P,0.001. When the shock intensity was 0.5 mA the Bgl II and separated on a 1.8 agarose gel in 0.53Tris intensity used for all subsequent studies the number of Borate EDTA buffer, stained with ethidium bromide and shocks was 28.863.94 for non-injection controls and photographed under UV illumination. Adjacent lanes with 30.062.88 for vehicle controls Fig. 1b. Fig. 1b also control PCR samples no template were always included shows that increasing doses up to 1.2 mg kg of al- to determine background. prazolam dose-dependently increased P,0.01 punished Data are presented as the log ratio of density of the responding. Doses higher than 1.2 mg kg increased amplified cRNA internal standards to the density of target punished responding less than 1.2 mg kg, possibly due to GABA subunit mRNA amplification product and plotted their mild ataxic effects. Although significant increases A against the log of known amounts of internal standard were obtained at all doses from 0.6–4.8 mg kg P, cRNA added to the test sample to generate a competitive 0.001, 1.2 mg kg was used for all subsequent studies PCR linear regression curve. The absolute amount of target because it produced maximal increases 208.8611.08 GABA subunit mRNA was calculated from the curve as shocks session. Fig. 1c compares the effects of saline, A the point of equivalency see arrows, Fig. 3, where the vehicle and alprazolam on punished and non-punished ratio of internal standard to target RNA was equal to 1. responding. The number of shocks taken under the punish- ment compared to the no punishment conditions was 2.6. Statistical analyses significantly different for both saline and vehicle. Al- prazolam significantly P,0.0001 increased punished All data behavioral, binding and RT–PCR were ana- responding compared to both saline and vehicle controls. lyzed using a two-way analysis of variance with Neuman– Keuls post hoc tests. Differences were considered signifi- cant if P-values less than 0.05 were obtained. All data are 3.2. Receptor autoradiography expressed as the mean6S.E.M. Data for licking are 3 characterized as the number of shocks per session, as five Fig. 2 shows that [ H]Ro 15-1788 binding was topog- licks always produced a shock, regardless of intensity. raphically distributed in the sections that were studied. Table 2 shows that among the 20 brain regions studied, the 3 greatest density of specific [ H]Ro 15-1788 binding was

3. Results found in the cerebral cortex. Moderately high levels of