Molecular Psychiatry (2009) 14, 291–307 & 2009 Nature Publishing Group All rights reserved 1359-4184/09 $32.00 www.nature.com/mp ORIGINAL ARTICLE Chronic stimulation of corticotropin-releasing factor receptor 1 enhances the anxiogenic response of the cholecystokinin system T Sherrin1,2,3,5, C Todorovic1,2,5, T Zeyda1,2, CH Tan3, PWT Hon3, Y-Z Zhu3,4 and J Spiess1,2 1Specialized Neuroscience Research Program 2, John A Burns School of Medicine, University of Hawaii, Honolulu, HI, USA; 2Molecular Neuroendocrinology Laboratory, Max Planck Institute of Experimental Medicine, John A Burns School of Medicine, University of Hawaii, Honolulu, HI, USA; 3Department of Pharmacology, Faculty of Medicine, National University of Singapore, Singapore and 4Department of Pharmacology, School of Pharmacy, Fudan University, Shanghai, China

Corticotropin-releasing factor (CRF) and cholecystokinin (CCK), two highly colocalized neuropeptides, have been linked to the etiology of stress-related anxiety disorders. Recent evidence points to the possibility that some of the anxiogenic effects of the central CCK system take place through interplay with the CRF system. The aim of the present study was to

examine the effects of chronic, mild activation of CRF receptor 1 (CRF1) on the central CCK system of the C57BL/6J mouse. As shown by in situ hybridization, real-Time PCR and immunohistochemistry, 5 days of intracerebroventricular (i.c.v.) injections of a subeffective

dose (2.3 pmol) of cortagine, a CRF1-selective agonist, resulted in an increase in CCK mRNA levels and CCK2 receptor immunoreactivity in several brain regions, such as amygdala and hippocampus, known to be involved in the regulation of anxiety. Mice with elevated endogenous central CCK tone exhibited significantly higher anxiety-like behaviors in the open-field task and elevated plus maze, and enhanced conditioned fear. These behavioral

changes were reversed by i.c.v. administration of the CCK2-selective antagonist LY225910, after 5 days of priming with cortagine. Under the same conditions, the intraperitoneal

administration of the CRF1 antagonist antalarmin was ineffective. This result indicated that once the CCK system was sensitized by prior CRF1 activation, it exhibited its anxiogenic effects, without influence by CRF1, possibly because of its observed downregulation. In sum, our results provide a novel model for the interaction of the CRF and CCK systems contributing to the development of hypersensitive emotional circuitry. Molecular Psychiatry (2009) 14, 291–307; doi:10.1038/sj.mp.4002121; published online 15 January 2008

Keywords: CRF; CRF1; CCK; CCK2; fear conditioning; anxiety

Introduction mostly found in the CNS, whereas CCK1 is highly abundant in the gastrointestinal tract, but has also Several neurotransmitters and neuropeptides have been identified in the CNS. The CCK precursor is also been implicated in the etiology of stress-related cleaved into other biologically active fragments such anxiety disorders. Among them are cholecystokinin as cholecystokinin tetrapeptide (CCK4) and pentagas- (CCK) and corticotropin-releasing factor (CRF).1,2 The trin (CCK5).4–6 Exogenous administration of CCK4 or neuropeptide CCK originally discovered as a gut CCK5 provoke panic attacks in man.1,6 Patients with peptide has been identified in the mammalian brain.3 panic disorder are hypersensitive to CCK receptor The most abundant form, the C-terminally sulfated 2 stimulation in comparison with healthy volunteers octapeptide fragment of CCK, exerts its action through and patients with other anxiety disorders. They also two G-protein coupled receptor subtypes CCK recep- appear to differ from healthy subjects in CCK tor 1 (CCK ) and CCK receptor 2 (CCK ). CCK is 1 2 2 processing and thus final CCK forms.1,6 The anxio- genic potential of the CCK system in animal models of Correspondence: Dr C Todorovic, Specialized Neuroscience anxiety is also well documented. CCK2 receptor Research Project 2, John A Burns School of Medicine, 651, Ilalo Street, Honolulu, HI 96813, USA. agonists increase anxiety of mice and rats in the E-mail: [email protected] and elevated plus-maze test (EPM) and light–dark com- Dr Y-Z Zhu, School of Pharmacy and Institute of Biomedical partment test (LDT).4,5,7 Rats selected on the basis of Sciences, Fudan University, 138 Yi Xue Yuan Rd, Shanghai high anxiety-like behaviors in the EPM display an 200032, China. 8 increased density of CCK2. Mice lacking CCK2 show E-mail: [email protected] 9 5These authors share equal contribution to this work. reduced anxiety in EPM and LDT. Received 1 January 2007; revised 17 October 2007; accepted 26 CRF, a 41-residue neuropeptide mediates many November 2007; published online 15 January 2008 behavioral responses to stress.10,11 It exhibits its CRF and CCK: anxiety and fear conditioning T Sherrin et al 292 actions through two distinctly distributed G protein- Cannulation and administration of drugs coupled CRF receptor subtypes, CRF receptor 1 The injection system (C235; Plastics One, Roanoke, 12,13 (CRF1) and CRF receptor 2 (CRF2). CRF can be VA, USA) consisted of a double-guided cannula, anxiogenic or anxiolytic. Mice with targeted deletion dummy and a cap. The cannulae were implanted 14,15 16,17 of either the CRF1 gene or the CRF2 gene, under 1.2% avertin anesthesia (0.4 ml per mouse) into display reduced or heightened anxiety-like behavior, both lateral brain ventricles (anteroposterior (AP): respectively. These data suggest that activation of 0.5 mm, lateral 1 mm, depth 2 mm) and affixed to the CRF1 increases, and activation of CRF2 predominantly skull by dental cement. The animals were allowed to attenuates anxiety. A series of studies provides recover for 7–10 days before the experiments started. important behavioral models that link the CRF system On the day of the first experiment, the mice were in conjunction with the CCK central system to the exposed to a light isoflurane anesthesia, the cap and 18,19 etiology of anxiety disorders. It is shown that the dummy were removed and peptide solutions were repeated infusion of a subthreshold dose of the CRF delivered through an injector linked to two Hamilton receptor agonist urocortin 1 (Ucn1) into the basolat- microsyringes with plastic tubing. The peptides were eral nucleus of the amygdala (BLA) produces long- administered bilaterally by a microinjector (CMA/ lasting anxiogenic effects, as indicated by EPM and Microdialysis, Solna, Sweden) over a 30 s period so 19 social interaction test of anxiety. Furthermore, that 0.25 ml volume was injected to each side. The repetitive administration of subthreshold doses of cannula placement was verified for each mouse Ucn1 results in a ‘priming’ phenomenon. Once immediately after the behavior experiments by histo- primed, these animals exhibit a strong increase in logical examination of the brains following methylene blood pressure, respiratory rate and heart rate in blue injection (0.25 ml per site). Only data obtained response to intravenous administration of sodium from mice with correctly inserted cannulae were 19 lactate, a panicogenic agent in humans. included in the statistical analysis. Cortagine was Thus, in the present study, utilizing the ‘priming’ synthesized as described previously.20 It was initially procedure, we tested the possibility that the CRF dissolved in 10 mM acetic acid and diluted 1:2 with system is capable of elevating the endogenous CCK sterile saline (0.9% NaCl). The CCK2-selective agonist tone, in key brain regions, such as amygdala and CCK4 (Bachem, Heidelberg, Germany) and the anta- hippocampus, known to be involved in the regulation gonist 2-(2-(5-bromo-1H-indol-3-yl)–ethyl)-3-(3-(1- of innate and conditioned anxiety responses. We methylethoxy)-phenyl)-4-(3H)-quinazolinone-(LY225910) report that repeated stimulation of CRF1 resulted in (Tocris Cookson Ltd, Bristol, UK) were first dissolved an increase in CCK mRNA and CCK2 immunoreacti- in saline (0.9% NaCl) to a concentration of 100 mM. vity in the BLA and dentate gyrus, as well as in the For cannulae injection the stock solutions were elevated anxiety-related behaviors in CCK2-depen- further diluted in saline to their final concentrations. dent manner. Together, our data provide a plausible The CRF1-selective antagonist antalarmin (Ant; pathological model of the chronic, but mild stimula- N-Butyl-N-ethyl-[2,5,6,-trimethyl-7-(2,4,6-trimethyl- tion of the CRF system, as it may occur under stress phenyl)-7H-pyrrolo[2,3-d]pyrimidin-4-yl]amine) was pur- conditions, enhancing the sensitivity of the CCK chased from Sigma (St Louis, MO, USA) and system through a feed forward mechanism. dissolved in a solution of 10% cremaphor (Sigma), 5% ethanol, and 85% sterile saline with a final pH of 6.0. The vehicle was made as above with the omission Materials and methods of Ant. The number of mice in behavioral experi- ments per group was 7–12. The number of mice in Animals each experiment is specified in the corresponding C57BL/6J mice aged 9 weeks (Centre D’Elevage, Le figure legends. Genest St Isle, France) were used. The mice were individually housed in standard Macrolon cages (Type 2: 22 Â 16 Â 13 cm) and maintained on a 12 h light/dark cycle (lights on at 0700 hours) with food Acute administration of the drugs and water freely available. The experiments were For acute treatments, mice received injections of performed in the Max Planck Institute for Experi- CCK4, cortagine or vehicle, bilaterally into the lateral mental Medicine and in the John A Burns School of ventricles, as described above. After 30 min, the mice Medicine, University of Hawaii. All experimental were exposed to training for fear conditioning or were procedures were in compliance with the European behaviorally tested in the post-shock freezing, or EPM Council Directive (86/609/EEC) by permission of the and open-field test (OFT). Antagonists or vehicle, Animal Section Law enforced by the District Govern- were injected 30 min before the agonists. The choice ment of Braunschweig, Lower Saxony, Germany. of behavioral paradigms, was based on the fact that Experiments performed in Honolulu were approved they are the most frequently used, and the most by the University of Hawaii Animal Care Committee thoroughly validated in the assessment of the roles in accordance with National Institute of Health of the central CRF and CCK systems in the regulation guidelines. All efforts were made to minimize animal of innate anxiety (EPM, OFT) and learned fear (post- suffering. shock freezing, fear conditioning).21,22

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 293 Chronic administration of the drugs 100–500 lux). In the chamber, a high-frequency For chronic treatment with cortagine, the C57BL/6J loudspeaker (KT-25-DT; Conrad, Hirschau, Germany) mice were divided into two groups on experimental provided constant background noise (white noise, day 1. Group 1 received a subeffective dose of 10 ng 68 dB sound pressure level (SPL)). After a 3 min (2.3 pmol) cortagine i.c.v., whereas Group 2 was adjustment period to the chamber, an electric foot- administered vehicle (saline). Cortagine was selected shock (2 s, 0.7 mA, constant current) was adminis- in view of its ability to mediate CRF1-specific anxiety tered. The relative duration (%) of freezing defined as in EPM, without CRF2-mediated side effects. It was a cessation of all body movements except those determined that 25–30 ng of the peptide was the required for respiration and heart beat, was deter- smallest dose eliciting significant anxiogenic effects. mined for 5 min immediately after the final shock in Under these conditions, 10 ng of cortagine was tested, the chamber where the footshock was given. It was but was subeffective.20 On experimental days 1–4 recorded by two unbiased observers in 10 s intervals. both groups were injected with cortagine or saline, Locomotor activities and the percentage of the respectively, and returned to their cages after each explored area were automatically monitored by an injection. On day 5 both groups were further infrared beam system and analyzed using software subdivided after receiving a fifth injection of either developed in collaboration with TSE (Bad Homburg, saline or cortagine. Thirty minutes later, the groups Germany). either received saline, a subeffective dose of CCK4 (10 ng; 15.8 pmol), or 250 ng (0.5 nmol) of the Fear conditioning

CCK2-selective antagonist LY225910 (LY; 250 ng; The training (conditioning) consisted of a single trial. 495 pmol). All doses represented the total amount The mice were exposed to the conditioning context applied per mouse in a volume of 0.5 ml. Half of the (180 s) identical to the one described under ‘Post- total amount was infused (i.c.v) with a microinjector shock freezing.’ This exposure was followed by a (CMA/Microdialysis), over a period of 30 s, into each footshock (unconditioned stimulus (US), 0.7 mA, 2 s, side. On the basis of the sequence of injections, the constant current) delivered through a stainless steel groups were called Saline–Saline (Sal–Sal), Saline– grid floor. The mice were removed from the fear CCK4 (Sal–CCK4), Saline–LY (Sal–LY), Cortagine– conditioning box 30 s after shock termination to avoid Saline (Corta–Sal), Cortagine–CCK4 (Corta–CCK4) an aversive association with the handling procedure. and Cortagine–LY (Corta–LY; Figure 3f, protocol The fear-conditioning chamber was thoroughly scheme). The LY225910 dose used was selected on cleaned with 70% ethanol before each animal was the basis of dose response studies in rats.23 The tested. Memory tests were performed 24 h after fear efficacy of the dose used to block of CCK4-mediated conditioning. Contextual memory was tested in the behavioral effects was confirmed in a control experi- fear-conditioning box for 180 s without US presenta- ment with mice. Thirty minutes after the injection of tion, with the presence of background noise. Freezing the second compound (CCK4, LY225910 or saline), as an index of long-term memory, was recorded as the mice were subjected to EPM, OFT, post-shock described under Post-shock freezing, above. The freezing test or training for fear conditioning. The activity burst produced by the electric shock was inhibitor dose applied prevented the CCK4 actions automatically detected by an infrared beam system (Figure 2). and analyzed by software developed in collaboration

In a related experiment, the ability of Ant, a CRF1- with TSE. specific antagonist,24 to prevent behavioral effects mediated by CRF1, was tested. The mice were divided Elevated plus maze into two groups on experimental day 1, and treated The behavior of mice was recorded by a video camera for 4 days with cortagine or saline. On day 5, the connected to a PC and analyzed by TSE software groups were treated intraperitoneally (i.p.) with Ant (VideoMot 2). The time spent, distance traveled and or vehicle, 30 min before both groups received a fifth number of entries into the open arms, closed arms and and final injection of saline or cortagine (protocol center were recorded for 5 min. The light intensity in scheme, Figures 8e and i). Totally, four groups were the plus-maze was 650 lux in the open arms and generated: Vehicle–Saline (Veh–Sal), Vehicle–Corta- center and 350 lux in the closed arms. Selective gine (Veh–Corta), Ant–Saline (Ant–Sal) and Ant- change in the preference for the open arms, as Cortagine (Ant–Corta). The Ant dose used measured by percentage of the time spent in the open (10 mg kgÀ1) was selected on the basis of its ability arms and number of entries into the open arms of the to prevent behavioral effects of cortagine. Thirty plus maze, was interpreted as a measure of anxiety. minutes after the fifth injection of saline or cortagine, The distance traveled (cm) was taken as measure for the mice were subjected to EPM, OFT, post-shock locomotor activity. freezing test or training for fear conditioning. Open-field task Post-shock freezing Mice were placed in an open field (50 Â 50 cm2) The mice were placed in a Plexiglas chamber protected with 10 cm high opaque walls. The animals (36 Â 21 Â 20 cm3) with a metal floor grid and were placed in the center of the open field and then overhead houselights (12 V, 10 W halogen lamp, allowed to explore it for 5 min. The field was divided

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 294 into 16 outer squares (facing the wall), 8 inner squares (three for each group) sections, these data were and a central square, (10 cm  10 cm each) as pre- collapsed. The final statistical analysis for each brain viously described.17 Time spent, distance traveled region combined 3–5 sections from each treatment and number of visits in the inner, outer and center group, depending on the availability of comparable region were recorded by a video camera connected to sections. a PC and analyzed by TSE software (VideoMot 2). The

light intensity was 650 lux in the center and 470 lux in Immunohistochemical analysis of CCK2 the peripheral zones of the open field. In the first 4 days of the experiment, two groups of mice were treated as described under In-situ hybridi- In-situ hybridization of mouse CCK mRNA zation of mouse CCK mRNA. The Corta group, Sal C57BL/6J mice were divided into two groups. One group, and naive control group each consisted of four group received a subeffective dose of 10 ng cortagine20 mice. Thirty minutes after the last injection on day 5, i.c.v. (Corta group), and another group was adminis- the mice were anesthetized with a 1.2% avertin tered vehicle (saline; Sal group) each day for 4 days. solution injected i.p., and transcardially perfused On day 5, the mice were decapitated 30 min after with ice-cold PBS (pH 7.4) followed by 4% PFA (pH receiving the last injection of either cortagine 7.4, 150 ml per mouse). The brains were postfixed for or saline (cortagine-injected, saline-injected mice, 48 h in the same fixative (4% PFA) and then respectively). As a second control group naive- immersed for 24 h each in 10, 20 and 30% sucrose untreated mice (N) were employed. Brains were in PBS. The brains were cryosectioned into 50 mm removed, and frozen (Figure 5a). Coronal brain slices. Elimination of endogenous peroxidase activity

sections (20 mm thick) from two mice and sagittal was accomplished using 1% H2O2 in methanol for brain sections from three mice collected from each 15 min, followed by rinses in PBS containing 0.2% experimental group were obtained with a microtome, Triton-X 100 (TBS). Five percent goat serum and 0.3% thaw-mounted onto SuperFrost Plus slides (Fischer, TBS in 0.01 M PBS was used for preincubation. Pittsburgh, PA, USA), fixed in 4% paraformaldehyde Sections (50 mm thick) were then incubated (24 h) in phosphate buffer (PFA), followed by washing in with primary antibody (1:200) in TBS containing 5%

phosphate-buffered saline (PBS), and dehydration in goat serum. The CCK2 antibody was purchased from a graded series of ethanol. An oligonucleotide probe Lifespan Biosciences (Woburn, MA, USA; now MBL complementary to mouse CCK mRNA (50-CTAGGAC Labs; cat no: LS-A1282). Subsequently, the sections TGCCATCACCACGCACAGACATACGCCGCTCTT-30) were incubated with biotinylated anti-rabbit antibody was end labeled at 37 1C for 10 min using terminal (1:500) followed by the avidin–biotin complex (Vector deoxynucleotidyl transferase kit (Roche, Mannheim, ABC kit). For visualization, diaminobenzidine (DAB) Germany). The sections were incubated overnight in was used as chromogen (Sigma tablet set). The hybridization buffer (50% formamide, 10% dextran sections were mounted on glass slides, dehydrated, sulfate, 50 mM DTT, 0.3 M NaCl, 30 mM Tris HCl (pH and covered with Eukitt. The staining of the brain 7.4), 4 mM EDTA (pH 8), IX Denhardts solution, sections was quantified using the NIH Image software 0.5 mg mlÀ1 salmon sperm DNA and 0.5 mg mlÀ1 Poly as described previously.26 Briefly, the same threshold A DNA). Post-hybridization washes were carried out was applied for each section, and the number of

in 1 Â standard sodium citrate (SSC; 150 mM NaCl, CCK2-containing cells was counted automatically. 30 mM Na citrate, pH 7.2) for 2 Â 2 min at room The counting was performed in an area of the same temperature, followed by a stringent wash at 57 1C shape and size for each brain region. Cells were in 1 Â SSC for 45 min. The slides were subsequently counted individually and monitored as number of washed in progressively diluted SSC buffers and then CCK-containing cells per 0.1 mm2 area. dehydrated in aqueous ethanol (50–100%). For immunofluorescence analysis, a similar proto- The slides were air dried and exposed to Kodak col as the one used for perfusion and postfixation of Biomax MR–Roentgen film for 24 h. The images were mice brains was followed. The naive control group then transferred to the AIDA densitometric analysis was omitted, since this group did not differ from the software (version 3.2). For the purpose of quantifying saline-treated group in the pilot experiment. The mRNA levels, a 1.5 mm2 area was defined and placed cortagine- and saline-treated groups consisted of four in five different positions over the pyramidal cell mice, each. Double immunofluorescent labeling was

layer of the CA1 and CA3 regions. The mean value performed for CRF and CRF1 proteins following the over those five measures was included in the manufacturer’s protocol (Vector Labs, Burlingame, statistical data analysis. A second area of 3 mm2 was CA, USA). Briefly, sections were washed with TBS used to analyze the other brain regions in the same before overnight incubation at 4 1C with primary CRF way, and specific signal density was determined antibody (1:5000) (rabbit anti-CRF antibody, Peninsula relative to neutral background density. Equivalent Laboratories Inc., Belmont, CA, USA), followed by planes of coronal and sagittal sections25 were used for incubation with biotinylated anti-rabbit secondary comparison of animal groups. Values are expressed in antibody (1:4000; Vector Labs). The sections were arbitrary units (mean integral density/area back- washed, Fluorescin avidin DCS (Vector Labs) was ground). As the arbitrary values did not differ from applied, rendering green fluorescence. Following

those of coronal (two for each group) and sagittal several washes in TBS, staining for CRF1 (1:400; goat

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 295 anti-CRF1 (C-16), Santa Cruz Inc., Santa Cruz, CA, synthesis was carried out using SuperScript III First- USA) was performed, using a similar procedure as Strand Synthesis System for RT–PCR (Invitrogen, that for CRF. CRF1 was visualized by incubation with Karlsruhe, Germany), following instructions provided biotinylated anti-goat secondary antibody (1:400) by the manufacturer. Primer pairs for RT–PCR were (Rockland Immunochemicals Inc, Gilbertsville, PA, generated to amplify 180–200 bp fragments of the USA), followed by Texas red avidin DCS (Vector genes of interest using the web-based application Labs). The sections were mounted on Superfrost Plus Primer 3. The following primers were used: CCK, slides (Menzel Gla˚ser, Braunschweig, Germany) using 50-actgctagcgcgatacatcc-30,50-attcgtagtcctcggcactg-30, 0 0 0 Vectashield mounting medium (Vector Labs) to pre- CCK2,5-caccctttatgcggtgatct-3 ,5-atgaatgtgcccatga 0 0 0 0 vent photo bleaching. Single immunofluorescent ggtt-3 , CRF2,5-gatcagcttccacagcatca-3 ,5-tcgttctcct 0 staining was carried out for CCK2 (1:400; rabbit anti- cccttctcct-3 , CRF (PPM04632A) and CRF1 CCK2 (H85), Santa Cruz Inc.) and CRF2 (1:400) (PPM04312A) (real-time PCR primers, Superarray proteins (goat anti-CRF2 (LS-A3570), LifeSpan Bio- Biosciences). b-Actin was used as control. cDNA sciences) using the same protocol as that used for synthesized from 250 ng of mRNA was amplified with CRF, except for the antibodies applied, followed by the LightCycler FastStart DNA Master SYBR Green I fluorescent labeling using Texas red Avidin DCS and kit (Roche, Indianapolis, IN, USA). The PCR was Fluorescein avidin DCS (Vector Labs), respectively. carried out in a LightCycler 2.0 (Roche, Indianapolis, An Axioimager Zeiss microscope and Zeiss Axio USA). Real-time results were analyzed using the Cam HRM camera were used for the observations Roche LightCycler software (Roche LightCycler, and photography of the slides, respectively. Equally version 2.1) with automatic computation of baseline sized images from different treatment groups were and threshold fluorescence levels. Results of real-time analyzed using AIDA imaging software. For this PCR data were represented as Ct values, with Ct quantification, a defined area was placed over the defined as the threshold cycle of PCR at which the regions of interest and specific signal density was amplified product was first detected. Gene expression determined. To account for differences in background Ct values below those of endogenous controls were staining intensity, two background intensity measure- discarded from the analysis, and outlier removal was ments lacking immunofluorescent profiles, were performed in cases where the standard deviation of Ct taken from each section. The mean of these two values exceeded 0.3 cycles. Melting curve analysis measurements constituted the background intensity. showed a single component confirming the formation The background intensity was then subtracted from of a specific PCR product. All real-time reactions were the measured specific signal density to provide a final carried out in triplicate. The data were analyzed using immunofluorescence intensity value. Values are equation in which mean fold change in target gene ÀDDCt expressed in arbitrary units (quantum level/area; expression = 2 . DDCt was calculated as (Mean Ct, QL mmÀ2). target – Mean Ct, b-actin) Â Corta group À (Mean Ct, target À Mean Ct, b-actin) Â Sal group. Corta group Laser-capture microdissection and real-time represented the mean expression of the target gene in quantitative PCR the BLA or DG after 5 days of five injections of

Quantitative gene expression data of CCK, CCK2, CRF, cortagine, and Sal group represented the mean CRF1 and CRF2 were obtained with RT–PCR of total expression of the target gene after five injections of RNA extracted from the BLA and dentate gyrus of the Saline, normalized to b-actin as an endogenous cortagine (Corta) and saline-treated (Sal) groups (three control. The variance estimated from the replicate Ct mice per group) by laser-capture microdissection values was carried through to the final calculation of (LCM). Thirty minutes after receiving a fifth injection relative quantities using standard propagation of error of saline or cortagine, the mice were rapidly decapi- methods. In the final calculation, the error was ÀDDCt tated, the brains were extracted, snap frozen, and estimated by evaluating the 2 term using DDCt± s.d. sectioned coronally at a thickness of 20 mmona (Table 1).27 cryostat. Glass slides carrying brain sections were loaded in a version of PixCell II called Autopix Statistics Microdissection Microscope (CFG, University of Statistical evaluation (StatView 5.0.1 software; SAS Albany, NY, USA), and cells from the dentate gyrus Institute, Cary, NC, USA) was performed by one-way and BLA were collected in microcaps using the 60 Â analysis of variance (ANOVA), with Bonferroni/Dunn objective. The laser spot size was 25 mm in diameter. test applied, post hoc, for individual between-group The laser power and pulse varied between 60–100 mW comparisons at the P < 0.05 level of significance. Data and 1600–2500 msec, respectively. Pictures before and are expressed as mean±s.e.m. after laser capture were taken. The cells were processed immediately after microdissection. RNA Results was extracted using the Qiagen RNeasy microkit (Qiagen, Valencia, CA, USA), and genomic DNA was Dose-dependent modulation of anxiety-related removed with integrated on-column DNase digestion behaviors by CCK4 using RNAse-free DnaseI (Qiagen). RNA quality was In our initial experiments, the dose-dependent effect of verified with the Agilent 2100 Bioanalyzer. cDNA CCK4 on post-shock freezing and context-dependent

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al

296 t

C fear conditioning was established by intracerebroven- DD s.d.), 0.02 0.01 tricular (i.c.v.) injection of C57BL6/J mice 30 min ± ± ± before training with different CCK4 doses (Figure 1a). ntrols in Corta

(M Statistical analysis (one-way ANOVA followed by the t Bonferroni/Dunn test) revealed that CCK4 dose-depen- .C -Actin

b dently impaired immediate post-shock freezing 0.04 27.5 0.02 27.6

valuating the 2 (F(4,35) = 10.89; P < 0.05; Figure 1b), without affecting ± ± context-dependent long-term freezing (F(4,35) = 1.84;

ine. P > 0.05; Figure 1c). Immediate post-shock freezing was applied as a measure of short-term conditioned 1.07 27.5 0.15 27.4 fear, while freezing on the test day, 24 h later, was used ± ± 2 0.08 32.08 0.12 32.59 ± ± 0.11 33.68 0.19 35.01 ± ± CCK CCK 0.18 34.54 0.09 31.99 ± ± 0.16 36.59 ± 2 CRF 0.12 38.56 0.22 40 34.18 ± ± 0.09 37.45 0.12 39.2 ± ± minus the standard deviation. The asymmetric distribution is a consequence of converting the results of t C 1 DD 0.15 38.35 ±

Figure 1 Activation of cholecystokinin receptor 2 (CCK2) exhibits differential effects on anxiety-like behaviors and 0.04 36.02 conditioned fear. Scheme of behavior protocol (a and d). ± , fold change. t

C I.c.v. administration of 100 ng (158 pmol), 50 ng (79 pmol),

DD 25 ng CCK4 (39.5 pmol), but not of 10 ng (15.8 pmol) CCK4 significantly reduced post-shock freezing (b), without CRF CRF affecting fear conditioning to context (c), used as indicators s.d.; 2 0.12 37.94 0.15 40 40 39.8 of short- and long-term conditioned fear, respectively ± ± ± 0.20 (0.20–0.24)1.17 (0.98–1.29) 0.21 (0.17–0.27) 0.95 (0.85–1.07) 0.62 (0.56–0.76) 0.48 (0.46–0.53) 4.92 (4.25–6.45) 4.29 (3.23–4.78) 5.78 (5.35–6.99) 3.14 (2.44–4.37) Sal Corta Sal Corta Sal(8 Corta mice Sal per group). Corta I.c.v. Sal administration Corta Sal of CCK4 (doses

plus the standard deviation and used as above), produced increased anxiety levels as t

C indicated by the time spent on the open arms (e) and

DD number of entries into the open arms (f) in the elevated plus

Gene expression of CRF and CCK systems as determined with real-time PCR is altered in the cortagine-treated mice in comparison with saline-treated co maze. Total distance traveled did not differ significantly s.d.) 39.89 s.d.) 35.60 ±

± among the groups (data not shown). CCK4 was injected t t C C

(M 30 min before testing (7–8 mice per group). Statistically t DD DD Ct (M C 2 2 significant differences: Bonferroni/Dunn test, *P < 0.05 Dentate gyrus Table 1 Abbreviations: BLA, basolateral nucleusValues > of 1 are the indicative amygdala; ofterm CCK, an increase, cholecystokinin; using and CRF, corticotropin-releasing < 1 factor; a decrease, Corta, in cortagine; gene Sal, expression relative sal to the saline-treated control. The error term was estimated by e an exponential process into a linear comparison of amounts. The number of mice per group was three. Real-time reactions were carried out in triplicate the dentate gyrus and BLA mean threshold cycle BLA relative to control (saline).

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 297 as indicator of long-term conditioned fear. It should be revealed that 495 pmol of LY225910 completely noted that immediate post-shock freezing can also be blocked the effects of CCK4 on anxiety-like behaviors regarded as a measure of general anxiety if the context in EPM (Bonferroni/Dunn test, P > 0.05 vs Sal–Sal). used does not represent a specific stimulus.28 The To determine if LY225910 had a behavioral effect lowest dose of CCK4 peptide per mouse that produced when injected alone, mice were acutely administered significant impairment on post-shock freezing was 495 pmol LY225910 (LY–Sal group), a dose found 25 ng CCK4 (39.5 pmol; Bonferroni/Dunn test, P <0.05 to block anxiogenic effects of CCK4. Then the mice vs saline). In view of the specificity of CCK4 selectively were examined in the EPM test. It was found that 29 binding to CCK2, it was concluded that the post- 495 pmol LY225910, injected 30 min prior to the shock freezing deficit was mediated by CCK2. EPM test did not have a significant effect on the We also investigated dose-dependent effects of anxiety measures of the EPM behavior when admin- CCK4 on anxiety by monitoring the EPM behavior of istered alone (Bonferroni/Dunn test, P > 0.05 vs Sal– C57BL/6J mice after i.c.v. injection of the peptide Sal for time spent in the open arms, number of entries (Figures 1d–f). Administration of 100 ng (158 pmol), into the open arms of EPM, and locomotor effect; 50 ng (79 pmol), 25 ng (39.5 pmol) CCK4, but not of Figures 2a–c). 10 ng (15.8 pmol; Bonferroni/Dunn test, P < 0.05 vs saline) CCK4 i.c.v., 30 min before testing in EPM, Changes in anxiety levels after repeated stimulation of exerted a profound anxiogenic effect as indicated by a CRF1 reflect changes of endogenous tone and decreased time spent in the open arms (F(4,32) = 34.32; sensitivity of the CCK system P < 0.05; Figure 1e) and number of open arm entries Previous studies demonstrated that repeated adminis-

(F(4,32) = 16.07; P < 0.05; Figure 1f). CCK4 did not tration of the CRF receptor subtype-nonspecific (F(4,32) = 1.11; P > 0.05) elicit changes in locomotor agonist Ucn1 into the BLA of the rat resulted in a activity as indicated by the total distance traveled in long-lasting increase in anxiety-like behaviors.18,19 On EPM (data not shown). the basis of these findings we hypothesized that

To determine whether the dose of the CCK2- repeated stimulation of CRF1 may result in hyper- selective antagonist LY225910 (495 pmol), established excitability of anxiety circuits through either an as effective in previous studies in rats,23 is capable of elevated endogenous CCK tone or reduced thresholds inhibiting the anxiogenic effects of an acute CCK4 for activation of the CCK system.29,30 Therefore six application in mice, we injected 495 pmol of groups were generated (Sal–Sal, Sal–CCK4, Sal–LY, LY225910, i.c.v., 30 min prior to administration of Corta–Sal, Corta–CCK4, Corta–LY; see Materials and 100 ng CCK4 (158 pmol). The mice were tested 30 min methods and Figure 3f, protocol scheme). With the afterwards in EPM. As expected, CCK4 (158 pmol), Corta–Sal group the possibility was tested that decreased the time spent in the open arms (F(3,34) = repeated stimulation of CRF1 resulted in increased 11.85; P < 0.05; Bonferroni/Dunn test, P < 0.05 vs anxiety-like behaviors. With the Corta–CCK4 group Sal–Sal; Figure 2a) and number of entries the possibility was tested whether repeated stimula-

(F(3,34) = 4.95; P < 0.05; Bonferroni/Dunn test, P < 0.05 tion of CRF1 led to a changed threshold for activation vs Sal–Sal; Figure 2b) into the open arms of EPM, of the CCK system, while with the Corta–LY group the without eliciting significant locomotor effects possibility was tested that increased anxiety-like

(F(3,34) = 0.22; P > 0.05; Figure 2c). Post hoc analysis behaviors were CCK2 dependent.

Figure 2 The cholecystokinin receptor 2 (CCK2)-selective antagonist LY225910 prevents CCK4-induced changes in anxiety- like behaviors. LY225910 (495 pmol) was injected i.c.v. 30 min prior to administration of 100 ng CCK4 (158 pmol). The mice were tested 30 min afterwards in the elevated-plus maze. The anxiogenic effects of 100 ng CCK4 (158 pmol) were fully prevented by LY225910 (495 pmol) as measured by time spent in the open arms (a) and number of entries into the open arms (b) of the elevated plus maze. The total distance traveled did not differ significantly among the groups (c). LY225910 (495 pmol), when injected alone, 60 min before the elevated-plus maze, did not have any effects on the measures of the elevated-plus maze behavior (a–c) (9–10 mice per group). Statistically significant differences: Bonferroni/Dunn test, *P < 0.05 relative to control (Sal–Sal).

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 298

Figure 3 Repeated stimulation of corticotropin-releasing factor receptor 1 (CRF1) leads to cholecystokinin receptor 2

(CCK2)-dependent changes in anxiety-like behaviors in the elevated-plus maze and open-field tasks. A subeffective dose of

10 ng (2.3 pmol) of the CRF1-specific agonist cortagine was administered i.c.v. once daily for 5 days. On day 5, 30 min after the last injection, the cortagine-treated mice were subdivided into three groups and administered with saline (Corta–Sal

group), a subeffective dose of 10 ng (15.8 pmol) CCK4 (Corta–CCK4 group) or 250 ng (495 pmol) of the CCK2-selective antagonist LY225910 (Corta–LY group). All peptides were administered in a volume of 0.25 ml per side. Mice treated with saline instead of cortagine were subdivided in the same manner (see scheme of behavioral protocol (f)). Thirty minutes after the last injection, separate mouse groups were subjected either to the elevated plus-maze test (EPM) or open-field test (OFT). The Corta–Sal and Corta–CCK4 groups displayed increased anxiety-like behavior as indicated by the time spent on the open arms (a), number of entries into the open arms (b) in the elevated-plus maze. The increase of anxiety produced by repeated

stimulation of CRF1 was prevented by administration of the CCK2-selective antagonist LY225910 (495 pmol; Corta–LY group; a, b). The Corta–LY group showed an increase in the total distance traveled (cm; c) in the elevated-plus maze. In addition, the Corta–Sal and Corta–CCK4 groups spent significantly more time in peripheral areas (d), and less in central areas (e)in the open-field task. Peripheral and central activities of the Corta–LY group showed the opposite pattern (10–11 mice per group). Statistically significant differences: Bonferroni/Dunn test, *P < 0.05 relative to control (Sal–Sal); #P < 0.05 relative to Corta–Sal group. We found that the Corta–Sal and Corta–CCK4 P < 0.05 vs Sal–Sal; Figure 3e) and exhibited more

groups displayed significantly enhanced anxiety-like peripheral region activity (F(5,54) = 5.68; P < 0.05; behavior as indicated by decreased time spent Bonferroni/Dunn test, P < 0.05 vs Sal–Sal) than the

(F(5,57) = 5.12; P < 0.05; Bonferroni/Dunn test, P < 0.05 Sal–Sal group (Figure 3d). These data strongly vs Sal–Sal; Figure 3a) and number of entries suggested that the Corta–Sal and Corta–CCK4 groups

(F(5,57) = 4.2; P < 0.05; Bonferroni/Dunn test, P < 0.05 showed a more anxious response than the Sal–Sal vs Sal–Sal; Figure 3b) into the open arms of EPM. control group in these tests. The Corta–LY group did Importantly, this increase in anxiety disappeared in not reduce the central activity or increase the the Corta–LY group (Bonferroni/Dunn test, P > 0.05 vs peripheral activity as observed with the Corta–Sal Sal–Sal; Figures 3a and b). The locomotor activity of and Corta–CCK4 groups (Bonferroni/Dunn test, Corta–Sal and Corta–CCK4 mice was not affected in P < 0.05 vs Sal–Sal; Figures 3d and e). The overall

the EPM test (F(5,57) = 1.18; P > 0.05; Bonferroni/Dunn results from EPM and OFT suggested that chronic test, P > 0.05 vs Sal–Sal; Figure 3c). Interestingly, stimulation of CRF1 led to a CCK2-dependent increase Corta–CCK4 mice also displayed decreased time of anxiety measures, mediated by endogenous release spent into the open arms of EPM (Bonferroni/Dunn of CCK, in the exploratory models of anxiety. test, #P < 0.05 vs Corta–Sal; Figure 3a) when compared to Corta–Sal mice. This result was suggestive that Changes in fear conditioning after repeated stimulation

repetitive stimulation of CRF1 enhanced the sensiti- of CRF1 reflects changes of the endogenous CCK system vity of the CCK system to facilitate anxiety-like In addition, we used the same procedure to investigate behavior. post-shock freezing, a putative measure of short-term

Furthermore, the Corta–Sal and Corta–CCK4 groups conditioned fear. Post-shock freezing was determined spent significantly less time in the central areas of the immediately after administration of the electric shock.

OF (F(5,54) = 5.83; P < 0.05; Bonferroni/Dunn test, Long-term freezing, indicating long-term memory, was

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 299 measured 24 h after training (Figure 4c, schematic expression of the CCK gene in the brain regions that protocol). The Corta–Sal and Corta–CCK4 mice have been implicated in CCK-induced anxiogenic 4,5 showed significant impairment of post-shock freezing behavior. Repeated stimulation of CRF1 (Corta (F(5,54) = 4.74; P < 0.05; Bonferroni/Dunn test, P < 0.05 group; Figure 5a) via the brain ventricles evoked a vs Sal–Sal; Figure 4a) and enhancement of long-term marked rise in CCK mRNA levels in the frontal cortex freezing (F(5,54) = 8.17; P < 0.05; Bonferroni/Dunn test, (FCtx; F(2,12) = 8.71; P < 0.05), ventral part of the CA1 P < 0.05 vs Sal–Sal; Figure 4b). As in the anxiety region of the hippocampus (F(2,11) = 74.48; P < 0.05), models described, both impairment of contextual dentate gyrus (F(2,11) = 7.21; P < 0.05), lateral nucleus of post-shock freezing and enhancement of long-term amygdala (LA) and BLA (F(2,10) = 83.55; P < 0.05), freezing were prevented by the CCK2-specific antago- lateral posterior (LP; F(2,12) = 19.66; P < 0.05), medial nist LY225910, as demonstrated with the behavior of geniculate (MGN; F(2,12) = 27.33; P < 0.05) and ventral the Corta–LY group (Bonferroni/Dunn test, P > 0.05 vs posteromedial (VPM; F(2,12) = 18.13; P < 0.05) nuclei of Sal–Sal). In view of the observed effects of the the thalamus (Bonferroni/Dunn test, P < 0.05 vs Sal selective CCK2 antagonist LY225910, it was suggested group; Figure 5b). Micrographs of the in situ hybridi- that these actions were mediated by the endogenous zations using an antisense probe for CCK mRNA are CCK system. shown on sagittal (lateral 2.04 mm bregma; Figure 5c), We also considered the possibility that the ‘priming’ and coronal sections (À1.58 mm bregma; Figure 5d). procedure altered the responsiveness of the mice to Other areas rich in CCK-containing neurons that have the footshock serving as US. This consideration been implicated in regulation of anxiety-like beha- implied that the enhancement of contextual condi- viors,31,32 like the cortical, basal and medial nuclei of tioned fear, observed in the Corta–Sal group, might be amygdala, bed nucleus of the stria terminalis, later- due to higher sensitivity to the footshock. ‘Priming’ odorsal and paraventricular nuclei of the thalamus, with cortagine did not change the footshock reactivity cingulate and piriform cortical areas, did not differ during training for fear conditioning (F(1,18) = 0.33; significantly in CCK mRNA levels (data not shown). P > 0.05 vs Sal–Sal; Figure 4c). It was concluded that, the conditioned, but not the unconditioned fear Increase of CCK protein levels in the BLA and dentate response was affected in the Corta–Sal group. 2 gyrus after repeated stimulation of CRF1 We also addressed the question whether the increased Elevated CCK mRNA levels in the brain regions CCK mRNA levels, possibly causing an increase in the involved in CCK-induced anxiety after repeated secretion of CCK-like peptides, were associated with stimulation of CRF1 changes in the production of the CCK2 receptor We directly tested the hypothesis that the observed protein (Figure 6a). An immunhistochemical analysis behavioral effects resulted from changes in the showed that CCK2 immunoreactive cells were densely

Figure 4 Repeated stimulation of corticotropin-releasing factor receptor 1 (CRF1) leads to cholecystokinin receptor 2

(CCK2)-dependent changes in short- and long-term conditioned contextual fear. All groups of mice were treated as described in Figure 3. Five days of repeated i.c.v. administration of a subeffective dose (10 ng; 2.3 pmol) of the CRF1-specific agonist cortagine into the Corta–Sal and Corta–CCK4 groups of mice resulted in significant impairment of contextual post-shock freezing (a), and enhancement of long-term freezing (b). Both impairment of contextual post-shock freezing and enhancement of long-term freezing were blocked by the CCK2-selective antagonist LY225910 (495 pmol) (Corta–LY group; a, b). Footshock reactivity during fear conditioning training did not significantly differ between the Sal–Sal and Corta–Sal groups (c) (mice per 10–12 per group). Scheme of the behavioral protocol (d). LTM, long-term memory; PSF, post-shock freezing; UR, unconditioned response (Bonferroni/Dunn test, *P > 0.05 vs Sal–Sal).

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 300

Figure 5 Repeated stimulation of corticotropin-releasing factor receptor 1 (CRF1) increases cholecystokinin (CCK) mRNA in discrete brain regions. The experimental procedure is shown under (a). Densitometric analysis of CCK mRNA of cortagine (Corta group) and saline-treated (Sal group) mice. These two groups were generated as described under (a). A significant upregulation of CCK mRNA was observed in various brain regions 5 days after daily repeated administration of (2.3 pmol) 10 ng cortagine. Frontal Cortex (FCtx); dentate gyrus (DG); the ventral part of CA1 of the hippocampus (vCA1); lateral (LA) and basolateral (BLA) nuclei of amygdala; lateral posterior (LP), medial geniculate (MGN) and ventral posteromedial thalamic (VPM) nuclei (b). Representative micrographs of the in situ hybridizations using an antisense probe for CCK mRNA are shown on sagittal (lateral 2.04 bregma) panel (c) and coronal sections (À1.58 mm bregma) panel (d). The dashed line on the sagittal section indicates the plane of the coronal section below. Right, cortagine-treated mouse. Left, saline-treated mouse (c and d). Scale bar, 1 mm. Each group, 3–5 counted sections. Statistically significant differences: *P < 0.05, Bonferroni/Dunn test, relative to control (saline-treated mice).

concentrated in the forebrain, most notably in the Thalamic nuclei, hippocampal and cortical regions regions of the hippocampus and the cortex, and with increased levels of CCK mRNA after chronic sparsely distributed in the thalamic region (Figures 6c cortagine treatment, showed very small or no differ-

and d). The CCK2 protein production in the dentate ences in the production of CCK2 receptor protein after gyrus subregion of the hippocampus (F(2,9) = 5.91; the same treatment. Together, the data from in situ P < 0.05) and BLA (F(2,9) = 4.95; P < 0.05) of the Corta hybridization and immunohistochemical staining group was significantly higher than in the Sal group experiments supported the hypothesis that the den- (Bonferroni/Dunn test, P < 0.05) (Figures 6b–d). A tate gyrus and BLA were the major candidate sites similar tendency was found in cortex and LA. where changes in endogenous CCK tone could

However, no statistically significant difference bet- mediate the anxiogenic effects of chronic CRF1 ween the groups was found (Bonferroni/Dunn test, stimulation. P > 0.05 vs Sal group; Figures 6b–d). Quantitative

analysis of the optical density of CCK2 in the dentate Changes in anxiety levels and fear conditioning after gyrus (F(1,6) = 6.12; P < 0.05; P < 0.05 vs Saline; Figures repeated stimulation of CRF1 are not prevented by 7a and c) and BLA (F(1,6) = 12.34; P < 0.05; P < 0.05 blockade of CRF1 vs Saline; Figures 7b and c) obtained after fluores- It had to be considered that the 5-day treatment with a cence immunohistochemistry confirmed the results subthreshold dose of cortagine resulted in a changed obtained after staining with DAB. CRF system responsiveness that, together with Interestingly, overlapping distributions of upregu- changes in the endogenous CCK system, could be

lated CCK mRNA and increased CCK2 immunoreac- responsible for the observed anxiety and conditioned tivity were found only in the dentate gyrus, a fear enhancement. The anxiety-like behavior follow- subregion of the hippocampus and in the BLA. ing the administration of CRF into the brain ventricles,

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 301

Figure 6 Repeated stimulation of corticotropin-releasing factor receptor 1 (CRF1) increases cholecystokinin receptor 2

(CCK2) protein in the dentate gyrus and BLA. The experimental procedure is described under (a). Representative photomicrographs and a summarized diagram showing the number of CCK2-labeled cells following chronic (5 days) cortagine (Corta group) or saline (Sal group) administration, are presented. A significant increase was observed in the granular layer of the dentate gyrus and BLA 5 days after repeated administration of (2.3 pmol) 10 ng cortagine (b). Brain regions shown: Corpus callosum (cc); Cortex (Ctx); granular layer of dentate gyrus (grDG); polymorphic layer of dentate gyrus (poDG); CA1 subregion of hippocampus (CA1); lateral (LA) and basolateral nuclei of amygdala (BLA); habenula (Hb); caudate putamen (CPu); lateral posterior (LP) and ventral posterior thalamic (VPM) nuclei. Coronal sections from the Sal group, panel (c), and from the Corta group, panel (d). Scale bar, 220 mm. Each group, 4 mice. Statistically significant differences: *P < 0.05, Bonferroni/Dunn test, relative to control (saline-treated mice).

Figure 7 Immunofluorescent labeling for cholecystokinin receptor 2 (CCK2) protein indicates its upregulation in the dentate gyrus and BLA. The experimental procedure is the same as described in Figure 6a. Conventional fluorescent photomicrographs illustrating changes in the density of CCK2-labeled (red) cells (see white arrows) following chronic (5 days) cortagine (Corta group) or saline (Sal group) administration are presented. A significant increase was observed in the dentate gyrus (a) and in the BLA (b) 5 days after repeated administration of (2.3 pmol) 10 ng cortagine. Brain regions shown: granular layer of dentate gyrus (grDG); polymorphic layer of dentate gyrus (poDG) and basolateral nuclei of amygdala (BLA). Coronal sections from the Sal group (a and b, left), and from the Corta group (a and b, right). Scale bar, 200 mm(a) and 400 mm (b). The background intensity was subtracted from the calculated intensity of the selected region and resulting signal intensity was divided by the area. Mean fluorescence is expressed in arbitrary units (quantum level/area; QL per mm2) (c). Each group, 4 mice. Statistically significant differences: *P < 0.05, Bonferronni/Dunn test, relative to control (saline- treated mice).

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al

302 20 appears to be mediated through CRF1, while CRF2 arm; Figure 8c); (F(5,51) = 5.76; P < 0.05, for distance; activation, outside the lateral septum, appears to Figure 8d; Bonferroni/Dunn test, Ant10-Corta vs Sal– mediate anxiolytic behavior.17 First, we tested the Sal; P > 0.05, Ant20-Corta vs Sal–Sal; P > 0.05). Post

dose of Ant, a CRF1-selective antagonist, that is hoc analysis, also revealed that administration of 10 or capable of blocking the anxiogenic effects of 100 ng 20 mg kgÀ1 Ant alone, i.p., 30 min before testing in cortagine (23 pmol). Two different doses (10 or EPM did not affect anxiety-like behaviors in the EPM 20 mg kgÀ1) of Ant were injected, i.p., 30 min before test (Bonferroni/Dunn test, Ant10-Sal vs Sal–Sal; administration of 100 ng of cortagine. The mice were P > 0.05 for all three behavioral measures in EPM; examined in the EPM test 30 min later (Figure 8a, Ant20-Sal vs Sal–Sal; P > 0.05 for all three behavioral protocol scheme). Pre-treatment with 10 or 20 mg kgÀ1 measures in EPM, Figures 8b–d).

Ant prevented cortagine-induced anxiety (F(5,51) = 5.48; The involvement of CRF1 receptor in the ‘priming’ P < 0.05, for time spent in the open arm; Figure 8b); process was established with the 5-day injection

(F(5,51) = 5.17; P < 0.05, for number of visits in the open protocol as described in Materials and methods, with

Figure 8 The corticotropin-releasing factor receptor 1 (CRF1) antagonist differentially affects anxiety-related behaviors after acute and chronic treatment with cortagine. Mice were injected intraperitonally (i.p.) with 10 or 20 mg kgÀ1 antalarmin (Ant),

a CRF1-selective antagonist, 30 min prior to i.c.v. administration of 100 ng cortagine (23 pmol) and tested for anxiety-like behavior in the elevated-plus maze 30 min later (a). Under these conditions, both doses of Ant completely blocked cortagine- induced decrease in the time spent in the open arms (b), number of entries into the open arms (c) and distance traveled in the elevated plus maze (d). When 10 or 20 mg kgÀ1 Ant were injected alone, 60 min before the elevated-plus maze test, no effects on the anxiety-like behavior in the plus maze were observed (b, c). Ant (20 mg kgÀ1) produced an increase in the total distance traveled (cm) in the elevated-plus maze (d). The number of mice was 9–10 per group. Five days (schemes of

behavioral protocol; e, i) of repeated i.c.v. administration of a subeffective dose (10 ng; 2.3 pmol) of the CRF1-specific agonist cortagine into the mice of the Veh–Corta group resulted in a significant increase of anxiety-like behavior as indicated by the time spent on the open arms (f), and number of entries into the open arms (g) in the elevated-plus maze. The increase of À1 anxiety produced by repeated stimulation of CRF1 was not prevented by pre-treatment (i.p.) with Ant (10 ng kg ) 30 min before the fifth injection cortagine on the day 5 (Ant–Corta group, f, g). The groups did not show significant differences in the locomotor activity in the elevated plus maze (h). Similarly, i.p. pretreatment with Ant (10 ng kgÀ1) did not prevent impairment of contextual post-shock freezing (j) and enhancement of long-term freezing (k), observed after repeated

stimulation of CRF1 (Ant–Corta group).EPM, elevated plus maze; LTM, long-term memory; PSF, post-shock freezing. Number of mice was 9–10 per group. Statistically significant differences: *P < 0.05,Bonferroni/Dunn test, relative to control (Veh–Sal mice).

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 303 the exception that on the fifth day, Ant (10 ng kgÀ1)or vehicle, were injected i.p. 30 min prior to the fifth injection of the dose of 2.3 pmol cortagine (Figures 8e and i, protocol schemes). The experiment included four groups: Veh–Sal, Veh–Corta, Ant–Sal and Ant– Corta. The Veh–Corta group exhibited significantly increased anxiety-like behavior as indicated by decreased time spent (F(3,33) = 7.72; P < 0.05; Bonferroni/ Dunn test, P < 0.05 vs Veh–Sal; Figure 8f) and number of entries (F(3,33) = 2.26; P < 0.05; Bonferroni/ Dunn test, P < 0.05 vs Veh–Sal; Figure 8g) into the open arms of EPM. Importantly, these increases in anxiety were not blocked in the Ant–Corta group (Bonferroni/Dunn test, P < 0.05 vs Veh–Sal; Figures 8f and g). The locomotor activity was not affected in the

EPM test (F(3,33) = 1.18; P > 0.05; Bonferroni/Dunn test, P > 0.05 vs Veh–Sal; Figure 8h). These data suggested that stimulation of CRF1 elevated the endogenous CCK tone and increased anxiety-like behaviors in a

CCK2-dependent fashion. CRF1 itself, however, was not directly involved, under these conditions, in the regulation of the same behaviors. Similar to the results obtained in the EPM test, À1 antagonism of CRF1 by 10 ng kg of Ant did not block the cortagine ‘primed’ decrease (F(3,34) =4.42; P < 0.05; Bonferroni/Dunn test, Ant–Corta vs Sal–

Corta; P < 0.05; Figure 8j) or increase (F(3,34) =6.38; P < 0.05; Bonferroni/Dunn test, Ant–Corta vs Sal– Corta; P < 0.05; Figure 8k) of post-shock or long-term freezing, respectively.

Repeated stimulation of CRF1 leads to downregulation Figure 9 Immunofluorescent labeling for corticotropin- of the CRF1 protein level in BLA releasing factor (CRF), CRF1 and CRF2, shows a down- In addition, we examined whether repeated stimula- regulation of CRF1 in the BLA. Mice were injected with saline (a–d, left panel) or cortagine (a–d, right panel). The tion of CRF1 resulted in changed expression patterns of CRF and its receptor subtypes in the BLA, central experimental procedure was the same as described on nucleus of amygdala (CeA), dentate gyrus and CA1 Figure 6a. A double-immunofluoresence technique was applied for CRF and CRF1. Representative conventional subregions of the hippocampus. Five days of corta- fluorescent photomicrographs illustrating changes in the gine treatment, resulted in a significantly decreased number of CRF-, CRF1- and CRF2-labeled cells following production of CRF1 protein in the BLA (F(1,6) = 9.86; chronic (5 days) cortagine (Corta group) or saline (Sal group) P < 0.05; P < 0.05 vs Saline; Figures 9b and c). No administration are presented. A significant downregulation statistically significant changes in CRF1 protein levels of CRF1 receptors was observed in the BLA after 5 days of were found in the CeA, CA1 and dentate gyrus repeated administration of (2.3 pmol) 10 ng cortagine. CRF1 subregions of the hippocampus (Figures 9b, c and immunoreactivity (red; see yellow arrow; b). There was no 10b; P > 0.05 vs Saline). In contrast to the decreased change in the CRF and CRF2 immunoreactivity in the BLA or the central amygdalar nucleus (CeA). CRF immunoreac- production of CRF1, chronic cortagine treatment did not lead to changes in the production of CRF tivity (green; see white arrowhead; (a), CRF2 immunoreac- tivity in the amydala (green; see yellow arrowhead; d). peptide and CRF2 protein in the BLA, CeA or CA1 and Superimposition of images in (a, b, left), (a, b, right), dentate gyrus subregions of the hippocampus (Figures respectively, are shown in (c). White arrow indicates the

9a, c, d, 10a and c; P > 0.05 vs Saline), as observed by spots of colocalization of CRF and CRF1. Brain regions densitometric analysis of CRF, CRF1 and CRF2 shown: BLA, basolateral nuclei of amygdala; CeA, central immunoreactivity (Figure 11). nuclei of amygdala; scale Bar, 400 mm.

CCK, CCK2, CRF, CRF1 and CRF2 genes are differentially expressed after cortagine ‘priming’ as indicated by each gene of interest as well as for the reference gene RT–PCR of microdissected tissue (b-actin). Results for the target genes were normalized LCM in combination with RT–PCR was employed to to the reference gene. Comparison of the expression of check the expression of CCK, CCK2 CRF, CRF1 and the target genes in the dentate gyrus and BLA of the CRF2 genes in the dentate gyrus and BLA of the Sal- and Corta groups was determined with the delta– 27 saline-treated (Sal group) or cortagine-treated (Corta delta (DD)Ct method. The fold changes of the genes group) mice. Resulting Ct values were recorded for expressions is presented in Table 1. The RT–PCR

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 304

Figure 11 Repeated stimulation of corticotropin-releasing

factor receptor 1 (CRF1) leads to a decrease of the CRF1 production in the BLA. Quantitative densitometric analysis

of mean immunofluorescence of CRF, CRF1 and CRF2 in the BLA and dentate gyrus, as shown in Figures 9 and 10. The Figure 10 Immunofluorescent labeling for corticotropin- background intensity was subtracted from the calculated

releasing factor (CRF), CRF1 and CRF2 indicates no intensity of the selected region, and resulting signal hippocampal changes in response to the chronic cortagine intensity was divided by the area. Mean fluorescence is treatment. Mice were injected with saline (a–c, left) or expressed in arbitrary units (quantum level/area; QL mmÀ2). cortagine (a–c, right). The experimental procedure is the The number of mice per group was 4. Statistically same as described in Figure 6a. Representative photomicro- significant differences: *P < 0.05, Bonferroni/Dunn test,

graphs illustrating changes in the number of CRF-, CRF1- relative to control (saline-treated mice).

and CRF2-labeled cells following chronic (5 days) cortagine (Corta group) or saline (Sal group) administration are

presented. No significant changes in CRF, CRF1 and CRF2 immunoreactivity were observed in the dentate gyrus and tone, known to be involved in the modulation of CA1 subregions of hippocampus after 5 days of repeated anxiety disorders.5,6 Mice with elevated CCK tone administration of (2.3 pmol) 10 ng cortagine. CRF immuno- showed an anxiogenic behavioral profile in EPM and reactivity (green; see red arrow; a); CRF1 immunoreactivity OFT, and displayed impairment of post-shock freez- (red; see white arrow; b); CRF2 immunoreactivity (green; see ing, together with enhancement of long-term con- yellow arrow). CA1, CA1 subregion of hippocampus; CA3, CA3 subregion of hippocampus; grDG, granular layer of textual conditioned fear. Selective blockade of CCK2 dentate gyrus; poDG, polymorphic layer of dentate gyrus; with LY225910 prevented these behavioral effects. bar, 200 mm(a and b); bar, 400 mm(c). Repeated CRF1 stimulation, which resembled the repeated exposure to a mild, chronic stressor, leading to an increased CCK tone, might be a model for anxiety disorder. This view is supported by the analysis provided corroborative evidence for upregu- findings that the best predictor of a CCK-induced

lation of the CCK and CCK2 mRNA of the Corta group, panic attack in humans, in laboratory provocation in both dentate gyrus and BLA. Interestingly, RT–PCR experiments, is a preexisting state of anxiety pro- 1,30,33 data indicated moderate downregulation for CRF2 in voked by recent stressful life events. From an the dentate gyrus, as well as strong downregulation ethological point of view, it would make sense that

for both CRF and CRF1 mRNA in the BLA of the Corta the emotional state of generalized anxiety lowers the group. Only, the latter result was confirmed by threshold for the elicitation of an anxiety response in immunofluorescence analysis of their respective the event that danger is detected. protein levels. Several lines of research in rodents have already indicated that the basal CCK tone in the brain may play an important role in enhancing anxiogenic Discussion factors. It has been observed that CCK2 agonists In this study, we examined the effects of chronic produce pronounced anxiogenic effects only in 34 administration of the newly developed CRF1 agonist stressed but not in unstressed animals. In addition, cortagine20 on the CCK system. For this purpose we after exposure to an acute stressor, both CCK

employed a previously described ‘priming’ para- immunoreactivity and CCK2 density in the hippo- digm.18,19 Evidence was provided that 5-day i.c.v. campus are increased.7,35 It is conceivable that an administration of a subthreshold dose of cortagine, is elevated CCK tone was associated with these treat- capable of elevating the endogenous central CCK ments. Moreover, the level of anxiety in EPM is

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 305 correlated with an increased CCK tone, as indicated is suggested that ‘cortagine-treated’ mice represent a by a higher CCK2 binding activity in the hippocampus model of anxiety, which also includes a heightened and cortex in anxious compared to nonanxious responsiveness to negative associations. Moreover, 8 controls, and upregulation of CCK2 immunoreacti- both acute injection of CCK4 and the cortagine- vity in the basolateral amygdala of rats with elevated induced elevated CCK tone resulted in reduced anxiety.36 In addition, anxiety responses after cat post-shock freezing, regarded as a measure of short- exposure among various rat strains can be attributed term conditioned fear. It is possible that activation of 37 to a correlated cortical production of CCK2. CCK2-induced changes in the anxiety-related reacti- Recently, it was observed that mice with inducible vity and processing of a novel conditioning context, 52 overexpression of forebrain CCK2 exhibit a signifi- which led to impaired post-shock freezing. cantly increased CCK2 binding capacity and anxiety- Within the amygdala, acute stress and anxiety-like like behaviors in the open-field and social interaction behaviors are associated with an increase in the tests.38 These findings indicate that the anxiety level production and release of CRF acting through 53–55 may be dynamically regulated by the stress-related CRF1. Also, acute i.c.v. administration of CRF CCK tone. results in a modest increase of CRF1 mRNA expres- Importantly, behavioral differences between ‘corta- sion in the BLA.56 We have found that chronic gine-stimulated’ and control animals were paralleled administration of cortagine downregulated CRF1 by changes in CCK mRNA levels and CCK2 immuno- protein in the BLA, without concomitant changes in reactivity in brain regions demonstrated to be involved CRF and CRF2 production as indicated by immuno- in the regulation of anxiety-like behaviors.39,40 The reactivity. It is tempting to speculate, that the role of areas with overlapping elevation of CCK mRNA and CRF1 in regulating BLA neurons may be gradually CCK2 immunoreactivity were the dentate gyrus sub- reduced when the system is adapted, which suggests region of the hippocampus and BLA. We hypothesize that CRF1 may have a potent influence on adjusting that these areas could be the major sites of behavioral behavioral responses to acute stress conditions, but effects observed after chronic stimulation of CRF1.In not during chronic stress. In line with this reasoning, support, it is well established that chronic stimula- administration of the CRF1-selective antagonist, Ant, tion of BLA CRF receptors with a subthreshold dose was not capable of blocking the behavioral effects of of Ucn1 induces a type of plasticity called ‘prim- chronic stimulation of CRF1. It is, therefore, suggested 18,19 ing’. Once ‘primed,’ rats exhibit enhanced respon- that chronic stimulation of CRF1 enhanced the siveness that persists for 30 days to panicogenic sensitivity of the CCK system to promote changes in sodium lactate and N-methyl-D-aspartate (NMDA) anxiety-like behaviors. However, once such mecha- receptor-dependent anxiogenic responses in the nism was set into the motion, CRF1 was probably no EPM and social interaction tests.18,19 In addition, longer directly involved in the regulation of anxiety, using in situ hybridization and quantitative receptor possibly because of its observed downregulation. autoradiographical techniques CCK mRNA, CCK2 CRF might exhibit positive effects on CCK tone protein and mRNA have been localized within directly or indirectly. A direct action of CRF may 41,42 BLA, and injection of nonselective CCK agonists have evolved through CRF1 receptors that are known into BLA, produce anxiogenic responses of both rats to be present in BLA, and different subregions of the and mice in EPM.42,43 There are no previous studies hippocampus, cortex or thalamus.12,13 Animal studies elucidating the role of dentate gyrus in CRF or CCK- indicate that CRF may play a role in the anxiogenic57 related anxiety. However, dentate gyrus is well and endocrine58 actions of CCK. A recent in vitro connected with BLA44,45 and seems to be an important study in which AtT-20 cells, a cellular model of structure in gating the formation of memory that corticotrophs, were used, demonstrated that CRF is renders emotional significance for the organism.46–49 capable of strongly inducing the expression of the 59 In our study, acute injections of CCK4 increased CCK gene in a CRF1-dependent manner. Indirectly, anxiety-like behaviors in the EPM and OFT tests. CRF interacts with several neurotransmitter systems, Thereby, the well established role of the CCK system including the serotonergic,60 dopaminergic61 and in exploratory models of anxiety4–6 was confirmed. GABAergic systems.62 Similarly, CCK is colocalized Interestingly, the same treatment resulted in impair- with dopamine,63 serotonin64 and GABA65 in different ment of post-shock freezing, without affecting long- brain areas, and CCK actions on anxiety regulation term contextual fear memory. Other studies can be mediated through interaction with these 4 performed with CCK2 antagonists or CCK2-deficient neurotransmitter systems. It is conceivable that mice in fear conditioning models provide ambiguous stimulating actions of CRF in these areas of coloca- results. It has been reported that CCK2 antagonists lization induce differential synthesis and/or release of exhibit either a decreasing50 or no5 effect on fear CCK. It is suggested that such actions may represent a conditioning. Similarly, fear conditioning was not mechanism through which CCK has modulatory 51 4,5 changed in CCK2-deficient mice. Chronic injection influence on anxiety-related behaviors. Such inter- of cortagine did not only result in an increase of action with CCK would present a novel element in the nonassociative forms of anxiety, but also increased CRF function, which goes beyond the early phase of the ability to learn and retain negative associations in the stress response,66,67 and involves long-lasting context-dependent fear conditioning. On this basis, it changes in other neuropeptide systems.

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 306 The demonstration of the role of chronic CRF1 specific subcortical nuclei in rat brain: comparison with CRF1 stimulation in anxiety in this study not only further receptor mRNA expression. J Neurosci 1995; 15: 6340–6350. confirms ‘priming’ as a valid pharmacological anxiety 13 Van Pett K, Viau V, Bittencourt JC, Chan RK, Li HY, Arias C et al. 18,19 Distribution of mRNAs encoding CRF receptors in brain and model in the mouse, but also reveals its role in pituitary of rat and mouse. J Comp Neurol 2000; 428: 191–212. profound sensitization of the central CCK system, 14 Timpl P, Spanagel R, Sillaber I, Kresse A, Reul JM, Stalla GK et al. thus projecting a novel molecular mechanism for the Impaired stress response and reduced anxiety in mice lacking a expression of anxiety. This feed-forward sensitiza- functional corticotropin-releasing hormone receptor 1. Nat Genet tion, in addition to contributing to an increased 1998; 19: 162–166. 15 Smith GW, Aubry JM, Dellu F, Contarino A, Bilezikjian LM, anxiety, also changes the aversive learning capacity. Gold LH et al. Corticotropin releasing factor receptor 1-deficient These changes could conceivably trigger the deve- mice display decreased anxiety, impaired stress response, lopment of different psychopathological forms. There- and aberrant neuroendocrine development. Neuron 1998; 20: 1093–1102. fore, we propose that CRF1 priming and consequent elevation of the CCK tone represent an important 16 Bale TL, Contarino A, Smit GW, Chan R, Gold LH, Sawchenko PE et al. Mice deficient for corticotropin-releasing hormone receptor- mechanism by which concurrent changes in two 2 display anxiety-like behaviour and are hypersensitive to stress. neuropeptide systems can translate the effects of mild Nat Genet 2000; 24: 410–414. and chronic stress into psychopathological responses. 17 Kishimoto T, Radulovic J, Radulovic M, Lin CR, Schrick C, Hooshmand F et al. Deletion of crhr2 reveals an anxiolytic role for If so, nonpeptidic CRF1 and CCK2 antagonists could corticotropin-releasing hormone receptor-2. Nat Genet 2000; 24: have major effects in preventing the spiral of stress- 415–419. induced sensitization of the emotional circuitry. 18 Sajdyk TJ, Schober DA, Gehlert DR, Shekhar A. Role of corticotropin-releasing factor and urocortin within the basolateral amygdala of rats in anxiety and panic responses. Behav Brain Res Acknowledgments 1999; 100: 207–215. 19 Rainnie DG, Bergeron R, Sajdyk TJ, Patil M, Gehlert DR, Shekhar A. We thank Dr Martin Rayner for helpful discussion Corticotrophin releasing factor-induced synaptic plasticity in the during preparation of the manuscript. This work was amygdala translates stress into emotional disorders. J Neurosci supported by Max Planck Society, NIH grant 2004; 24: 3471–3479. 20 Tezval H, Jahn O, Todorovic C, Sasse A, Eckart K, Spiess J. 5U54NS039406-08 and NMRC grant (NMRC/0754/ Cortagine, a specific agonist of corticotropin-releasing factor 2003). receptor subtype 1, is anxiogenic and antidepressive in the mouse model. Proc Natl Acad Sci USA 2004; 101: 9468–9473. 21 Griebel G. Is there a future for neuropeptide receptor ligands in the References treatment of anxiety disorders? Pharmacol Ther 1999; 82: 1–61. 22 Gordon JA, Hen R. Genetic approaches to the study of anxiety. 1 Bradwejn J. Neurobiological investigations into the role of Annu Rev Neurosci 2004; 27: 193–222. cholecystokinin in panic disorder. J Psychiatry Neurosci 1993; 23 Bertoglio LJ, Zangrossi Jr H. Involvement of dorsolateral peri- 18: 178–188. aqueductal gray cholecystokinin-2 receptors in the regulation of a 2 Smoller JW, Yamaki LH, Fagerness JA, Biederman J, Racette S, panic-related behavior in rats. Neurosci Biobehav Rev 2005; 29: Laird NM et al. The corticotropin-releasing hormone gene and 1361–1373. behavioral inhibition in children at risk for panic disorder. Biol 24 Webster EL, Lewis DB, Torpy DJ, Zachman EK, Rice KC, Chrousos Psychiatry 2005; 57: 1485–1492. GP. In vivo and in vitro characterization of antalarmin, 3 Vanderhaeghen JJ, Signeau JC, Gepts W. New peptide in the a nonpeptide corticotropin-releasing hormone (CRH) receptor vertebrate CNS reacting with antigastrin antibodies. Nature 1975; antagonist: suppression of pituitary ACTH release and peripheral 257: 604–605. inflammation. Endocrinology 1996; 137: 5747–5750. 4 Rotzinger S, Vaccarino FJ. Cholecystokinin receptor subtypes: role 25 Franklin KBJ, Paxinos G. The Mouse Brain in Stereotaxic in the modulation of anxiety-related and reward-related beha- Coordinates. Academic Press: San Diego, CA, 2001. viours in animal models. J Psychiatry Neurosci 2003; 28: 171–181. 26 Radulovic J, Kammermeier J, Spiess J. Relationship between fos 5 Dauge V, Lena I. CCK in anxiety and cognitive processes. Neurosci production and classical fear conditioning: effects of novelty, Biobehav Rev 1998; 22: 815–825. latent inhibition, and unconditioned stimulus preexposure. 6 Bradwejn J, Koszycki D. Cholecystokinin and panic disorder: past J Neurosci 1998; 18: 7452–7461. and future clinical research strategies. Scand J Clin Lab Invest 27 Livak KJ, Schmittgen TD. Analysis of relative gene expression data Suppl 2001; 234: 19–27. using real-time quantitative PCR and the 2(Àdelta delta C(T)). 7 Harro J, Lofberg C, Rehfeld JF, Oreland L. Cholecystokinin Methods 2001; 25: 402–408. peptides and receptors in the rat brain during stress. Naunyn 28 Fanselow MS. Contextual fear, gestalt memories, and the hippo- Schmiedebergs Arch Pharmacol 1996; 354: 59–66. campus. Behav Brain Res 2000; 110: 73–81. 8 Koks S, Vasar E, Soosaar A, Lang A, Volke V, Voikar V et al. 29 Noble F, Roques BP. CCK-B receptor: chemistry, molecular biology, Relation of exploratory behavior of rats in elevated plus-maze to biochemistry and pharmacology. Prog Neurobiol 1999; 58: brain receptor binding properties and serum growth hormone 349–379. levels. Eur Neuropsychopharmacol 1997; 7: 289–294. 30 Barlow DH. Cognitive-behavioral therapy for panic disorder: 9 Horinouchi Y, Akiyoshi J, Nagata A, Matsushita H, Tsutsumi T, current status. J Clin Psychiatry 1997; 58: 32–36. Isogawa K et al. Reduced anxious behavior in mice lacking 31 Meziane H, Devigne C, Tramu G, Soumireu-Mourat B. Distribution the CCK2 receptor gene. Eur Neuropsychopharmacol 2004; 14: of cholecystokinin immunoreactivity in the BALB/c mouse 157–161. forebrain: an immunocytochemical study. J chem Neuronat 1997; 10 Spiess J, Rivier J, Rivier C, Vale W. Primary structure of 12: 191–209. corticotropin-releasing factor from ovine hypothalamus. Proc Natl 32 Cain BM, Connolly K, Blum A, Vishnuvardhan D, Marchand JE, Acad Sci USA 1981; 78: 6517–6521. Beinfeld MC. Distribution and colocalization of cholecystokinin 11 Vale W, Spiess J, Rivier C, Rivier J. Characterization of a 41-residue with the prohormone convertase enzymes PC1, PC2, and PC5 in ovine hypothalamic peptide that stimulates secretion of cortico- rat brain. J Comp Neurol 2003; 467: 307–325. tropin and b-endorphin. Science 1981; 213: 1394–1397. 33 Bouton ME, Mineka S, Barlow DH. A modern learning theory 12 Chalmers DT, Lovenberg TW, De Souza EB. Localization of novel perspective on the etiology of panic disorder. Psychol Rev 2001; corticotropin-releasing factor receptor (CRF2) mRNA expression to 108: 4–32.

Molecular Psychiatry CRF and CCK: anxiety and fear conditioning T Sherrin et al 307 34 Cohen H, Kaplan Z, Kotler M. CCK-antagonists in a rat exposed to receptor gene knockout mice in the plus-maze. Neurosci Lett 2002; acute stress: implication for anxiety associated with post- 335: 115–118. traumatic stress disorder. Depress Anxiety 1999; 10: 8–17. 53 Swiergiel AH, Takahashi LK, Kalin NH. Attenuation of stress- 35 Siegel RA, Duker EM, Pahnke U, Wuttke W. Stress-induced induced behavior by antagonism of corticotropin-releasing factor changes in cholecystokinin and substance P concentrations in receptors in the central amygdala in the rat. Brain Res 1993; 623: discrete regions of the rat hypothalamus. Neuroendocrinology 229–234. 1987; 46: 75–81. 54 Merlo Pich E, Lorang M, Yeganeh M, Rodriguez de Fonseca F, 36 Nelovkov A, Areda T, Innos J, Koks S, Vasar E. Rats displaying Raber J, Koob GF et al. Increase of extracellular corticotropin- distinct exploratory activity also have different expression releasing factor-like immunoreactivity levels in the amygdala of patterns of gamma-aminobutyric acid- and cholecystokinin- awake rats during restraint stress and ethanol withdrawal as related genes in brain regions. Brain Res 2006; 1100: 21–31. measured by microdialysis. J Neurosci 1995; 15: 5439–5447. 37 Farook JM, Zhu YZ, Wang H, Moochhala S, Lee L, Wong PT. Strain 55 Sajdyk TJ, Gehlert DR. Astressin, a corticotropin releasing factor differences in freezing behavior of PVG hooded and Sprague- antagonist, reverses the anxiogenic effects of urocortin when Dawley rats: differential cortical expression of cholecystokinin2 administered into the basolateral amygdala. Brain Res 2000; 877: receptors. Neuroreport 2001; 27: 17–20. 226–234. 38 Chen Q, Nakajima A, Meacham C, Tang YP. Elevated cholecysto- 56 Mansi JA, Rivest S, Drolet G. Regulation of corticotropin-releasing kininergic tone constitutes an important molecular/neuronal factor type 1 (CRF1) receptor messenger ribonucleic acid in the mechanism for the expression of anxiety in the mouse. Proc Natl paraventricular nucleus of rat hypothalamus by exogenous CRF. Acad Sci USA 2006; 103: 3881–3886. Endocrinology 1996; 137: 4619–4629. 39 Davis M. Are different parts of the extended amygdala involved in 57 Biro E, Sarnyai Z, Penke B, Szabo G, Telegdy G. Role of fear versus anxiety? Biol Psychiatry 1998; 44: 1239–1247. endogenous corticotropin-releasing factor in mediation of neuro- 40 LeDoux J. The emotional brain, fear, and the amygdala. Cell Mol endocrine and behavioral responses to cholecystokinin octapep- Neurobiol 2003; 23: 727–738. tide sulfate ester in rats. Neuroendocrinology 1993; 57: 340–345. 41 Zarbin MA, Innis RB, Wamsley JK, Snyder SH, Kuhar MJ. 58 Kamilaris TC, Johnson EO, Calogero AE, Kalogeras KT, Bernardini R, Autoradiographic localization of cholecystokinin receptors in Chrousos GP et al. Cholecystokinin-octapeptide stimulates hypo- rodent brain. J Neurosci 1983; 3: 877–906. thalamic-pituitary-adrenal function in rats: role of corticotropin- 42 Moran TH, Schwartz GJ. Neurobiology of cholecystokinin. Crit Rev releasing hormone. Endocrinology 1992; 130: 1764–1774. Neurobiol 1994; 9: 1–28. 59 Peeters PJ, Gohlmann HW, Van den Wyngaert I, Swagemakers SM, 43 Harro J, Pold M, Vasar E. Anxiogenic-like action of caerulein, a Bijnens L, Kass SU et al. Transcriptional response to corticotropin- CCK-8 receptor agonist, in the mouse: influence of acute and releasing factor in AtT-20 cells. Mol Pharmacol 2004; 66: subchronic diazepam treatment. Naunyn Schmiedebergs Arch 1083–1092. Pharmacol 1990; 341: 62–67. 60 Linthorst AC, Reul JM. Inflammation and brain function under 44 Pikkarainen M, Ronkko S, Savander V, Insausti R, Pitkanen A. basal conditions and during long-term elevation of brain Projections from the lateral, basal, and accessory basal nuclei of corticotropin-releasing hormone levels. Adv Exp Med Biol 1999; the amygdala to the hippocampal formation in rat. J Comp Neurol 461: 129–152. 1999; 403: 229–260. 61 Matsuzaki I, Takamatsu Y, Moroji T. The effects of intracerebro- 45 Petrovich GD, Canteras NS, Swanson LW. Combinatorial amygda- ventricularly injected corticotropin-releasing factor (CRF) on the lar inputs to hippocampal domains and hypothalamic behavior central nervous system: behavioural and biochemical studies. systems. Brain Res Brain Res Rev 2001; 38: 247–289. Neuropeptides 1989; 13: 147–155. 46 Izquierdo I, Medina JH. Role of the amygdala, hippocampus and 62 Skelton KH, Nemeroff CB, Knight DL, Owens MJ. Chronic entorhinal cortex in memory consolidation and expression. Braz J administration of the triazolobenzodiazepine alprazolam produces Med Biol Res 1993; 26: 573–589. opposite effects on corticotropin-releasing factor and urocortin 47 Kim JJ, Rison RA, Fanselow MS. Effects of amygdala, hippocam- neuronal systems. J Neurosci 2000; 20: 1240–1248. pus, and periaqueductal gray lesions on short- and long-term 63 Hokfelt T, Rehfeld JF, Skirboll L, Ivemark B, Goldstein M, Markey K. contextual fear. Behav Neurosci 1993; 107: 1093–1098. Evidence for coexistence of dopamine and CCK in meso-limbic 48 Yaniv D, Vouimba RM, Diamond DM, Richter-Levin G. Simulta- neurones. Nature 1980; 12: 476–478. neous induction of long-term potentiation in the hippocampus 64 van der Kooy D, Hunt SP, Steinbusch HW, Verhofstad AA. Separate and the amygdala by entorhinal cortex activation: mechanistic and populations of cholecystokinin and 5-hydroxytryptamine- temporal profiles. Neuroscience 2003; 120: 1125–1135. containing neuronal cells in the rat dorsal raphe, and their 49 Maren S, Quirk GJ. Neuronal signalling of fear memory. Nat Rev contribution to the ascending raphe projections. Neurosci Lett Neurosci 2004; 5: 844–852. 1981; 26: 25–30. 50 Tsutsumi T, Akiyoshi J, Isogawa K, Kohno Y, Hikichi T, Nagayama H. 65 Hendry SH, Jones EG, DeFelipe J, Schmechel D, Brandon C, Suppression of conditioned fear by administration of CCKB receptor Emson PC. Neuropeptide-containing neurons of the cerebral antagonist PD135158. Neuropeptides 1999; 33: 483–486. cortex are also GABAergic. Proc Natl Acad Sci USA 1984; 81: 51 Raud S, Innos J, Abramov U, Reimets A, Koks S, Soosaar A et al. 6526–6530. Targeted invalidation of CCK2 receptor gene induces anxiolytic- 66 Owens MJ, Nemeroff CB. Physiology and pharmacology of like action in light-dark exploration, but not in fear conditioning corticotropin-releasing factor. Pharmacol Rev 1991; 43: 425–473. test. Psychopharmacology (Berl) 2005; 181: 347–357. 67 Koob GF, Heinrichs SC. A role for corticotropin releasing factor 52 Miyasaka K, Kobayashi S, Ohta M, Kanai S, Yoshida Y, Nagata A and urocortin in behavioral responses to stressors. Brain Res 1999; et al. Anxiety-related behaviors in cholecystokinin-A, B, and AB 848: 141–152.

Molecular Psychiatry