Molecular Psychiatry (2002) 7, 113–117  2002 Nature Publishing Group All rights reserved 1359-4184/02 $15.00 www.nature.com/mp ORIGINAL RESEARCH ARTICLE Role of bombesin (BN)-like /receptors in emotional behavior by comparison of three strains of BN-like receptor knockout mice K Yamada1*, Y Santo-Yamada1,2,*, E Wada1,3 and K Wada1

1Department of Degenerative Neurological Diseases, National Institute of Neuroscience, National Center of Neurology and Psychiatry, 4–1-1 Ogawahigashi, Kodaira-City, Tokyo 187–8502, Japan; 2Department of Mental Retardation and Birth Defect Research, National Institute of Neuroscience, National Center of Neurology and Psychiatry, 4–1-1 Ogawahigashi, Kodaira-City, Tokyo 187–8502, Japan; 3Japan Science and Technology Corporation, 4–1-8 Honmachi, Kawaguchi-City, Saitama 332–0012, Japan

Keywords: -releasing peptide receptor (GRP-R); peptides and their receptors are widely distributed in bombesin receptor subtype-3 (BRS-3); recep- the mammalian CNS and modulate many aspects of tor (NMB-R); light-dark (L-D) box test; elevated plus maze behavior such as spontaneous activity and feeding test; anxiety; emotion; knockout mice behavior, as well as learning and memory.6–9 We pre- Bombesin (BN)-like peptides are involved in the regu- viously produced three strains of BN-like peptide lation of a wide variety of behaviors, such as spon- receptor knockout mice using a gene targeting strategy. taneous activity and feeding. We assessed the role of The resulting GRP-R-, NMB-R- and BRS-3-deficient BN-like peptides/receptors in emotional and/or anxiety- mice were used in studies designed to clarify and dis- related behavior using three strains of knockout mice, tinguish the functional properties of BN-like peptides. each deficient in a single BN-like peptide receptor These knockout mice exhibit unique phenotypes, such (gastrin-releasing peptide receptor, bombesin receptor as elevated spontaneous activity and non-aggressive subtype-3, or neuromedin B receptor). Two representa- 10–12 tive behavioral paradigms, the light-dark (L-D) box test social behaviors in GRP-R-deficient mice. BRS-3- and the elevated plus maze test, were chosen for this deficient mice exhibit moderate hyperphagia and obes- purpose. In these two tests, the level of anxiety can be ity,13 altered taste-related behaviors,14 and decreased measured as the preference for exploring the light box, social behavior.15 However, no behavioral phenotype or the length of time spent in the open arms, respect- has yet been identified for NMB-R-deficient mice.16 ively. By conventional parameters, the only significant The behavioral analyses conducted thus far indicate finding was that BRS-3-deficient mice exhibited a longer that BN-like peptides/receptors may play important duration of remaining in the open arms compared to the roles in modulating emotional behavior including wild-type cohort (P Ͻ 0.01). However, analyses of risk some anxiety-related behavior in mammals. Therefore, assessment behavior revealed that BRS-3-deficient mice exhibited increased ‘stretched attend posture’ in the present study, we assessed and compared behavior (P Ͻ 0.01, compared to wild-type mice in both properties of emotion and anxiety in our three strains the L-D box and elevated plus maze tests) while NMB-R- of BN-like peptide receptor-deficient mice using the L- deficient mice exhibited decreased behavior (P Ͻ 0.05, D box test as well as the elevated plus maze test.17–20 compared to wild-type mice in both tests). These results In the L-D box test, rats and mice usually avoid remain- suggest that BN-like peptides/receptors may play a role ing in the light compartment. Therefore, the level of in modulating emotion including some forms of anxiety anxiety can be measured as the latency in moving into (eg, risk assessment behavior). Further, we found that the light-compartment and/or the duration of staying in the type of emotional behavior to which each of the the light-compartment. Similarly, in the elevated plus peptide/receptor pathways contributes can be clearly maze test, the number of entrances into and/or the specified. Molecular Psychiatry (2002) 7, 113–117. DOI: 10.1038/ length of time spent in the ‘open arms’ are used as indi- sj/mp/4000974 ces of anxiety. Furthermore, we also assessed some ethological indices (eg risk assessment behavior to Bombesin (BN) was originally purified from the skin of assess anxiety, grooming behavior to assess nonspecific the European frog (Bombina bombina),1 while the BN- emotional state, and so on), in light of a recent proposal like peptides gastrin-releasing peptide (GRP) and neur- that suggests these indices should also be considered omedin B (NMB) were purified from mammalian in such studies to support the classical methods.21 tissues.2,3 These peptides exert their effects by binding Table 1 shows the results of the light-dark box test to G-protein coupled receptors such as the GRP recep- and the elevated plus maze test. In the L-D box test, tor (GRP-R), the NMB receptor (NMB-R), and the BN there were no significant differences between mutant receptor subtype-3 (BRS-3) for which no specific mice and their wild-type littermates with respect to the endogenous ligand(s) has yet been identified.4,5 BN-like conventional parameters such as dark-light (D-L) lat- Anxiety in BN-like peptide receptor knockout mice K Yamada et al 114 Table 1 Summary of the light-dark (L-D) box test and the elevated plus maze test

Indices/subjects GRP-R BRS-3 NMB-R

KO wild KO wild KO wild

L-D box test Dark-light latency (s) 149.8 ± 17.6 114.1 ± 22.6 201.2 ± 36.5 140.9 ± 39.6 200.2 ± 27.2 209.7 ± 37.3 Number of across the 5.43 ± 1.02 7.43 ± 1.69 5.50 ± 1.96 8.0 ± 2.41 3.29 ± 1.01 4.33 ± 2.17 compartments Duration of staying in the light 50.5 ± 12.0 49.4 ± 13.0 43.2 ± 15.5 79.1 ± 22.9 41.7 ± 13.4 34.0 ± 17.9 compartment (s) Risk assessment behavior 12.9 ± 1.26 10.1 ± 1.68 12.9 ± 1.53** 7.20 ± 0.88 5.00 ± 1.20* 10.8 ± 2.12

Elevated plus maze test No. of open arm entries 3.50 ± 0.60 3.57 ± 0.37 7.63 ± 0.75 5.88 ± 0.58 1.50 ± 0.27 1.00 ± 0.33 No. of closed arm entries 7.50 ± 0.69 8.14 ± 1.02 13.5 ± 1.05 12.4 ± 1.31 7.25 ± 1.49 5.13 ± 1.27 % Open arm entries 30.9 ± 3.92 30.2 ± 3.47 36.2 ± 3.07 32.4 ± 2.21 16.8 ± 3.78 13.7 ± 4.66 Time spent in open arm (s) 56.9 ± 16.4 39.6 ± 6.19 61.8 ± 3.79** 37.2 ± 4.96 16.2 ± 4.38 14.1 ± 8.83 % Time spent in open arm 19.0 ± 5.48 13.2 ± 2.06 20.6 ± 1.26** 12.4 ± 1.65 5.40 ± 1.46 4.71 ± 2.94 Time spent in closed arm (s) 169.2 ± 16.4 177.5 ± 16.4 149.6 ± 9.20 157.7 ± 15.4 237.8 ± 12.5 255.7 ± 13.7 % Time spent in closed arm 56.4 ± 5.45 59.2 ± 5.47 49.9 ± 3.07 52.6 ± 5.14 79.3 ± 4.17 85.2 ± 4.57 Stretched attend posture 0.20 ± 0.92 5.43 ± 0.81 7.25 ± 0.88** 4.50 ± 0.60 5.50 ± 1.11* 8.63 ± 0.94

KO = GRP-R-, BRS-3- or NMB-R-deficient (knockout) mice. Wild = wild-type littermates of each group of knockout mice. With respect to ethological indices, only the score of stretched attend posture is presented since significant changes were evident only in this behavior. Data represent means ± SEM. Asterisks represent statistical differences between knockout mice and wild- type littermates: **P Ͻ 0.01; *P Ͻ 0.05.

ency and the duration of staying in the light compart- assessment behavior or other behaviors in the elevated ment. However, in the analysis of risk assessment plus maze test. behavior (stretched attend posture), BRS-3-deficient In summary, in the present study we examined the mice exhibited such behavior significantly more fre- properties of emotion and/or anxiety in three strains of quently than wild-type littermates (U = 6.5, P Ͻ 0.01, BN-like peptide receptor knockout mice using the L-D two-tailed Mann–Whitney’s U-test). In contrast, NMB- box test and the elevated plus maze test. In the L-D box R-deficient mice exhibited a significant decrease in risk test, the conventional parameters were not significantly assessment behavior relative to the wild-type cohort different between the knockout mice and their wild- (U = 7, P Ͻ 0.05, two-tailed Mann–Whitney’s U-test). type counterparts. Similarly, in the elevated plus maze However, as with the conventional parameters, there test only BRS-3-deficient mice exhibited some decrease was no significant difference between GRP-R-deficient in anxiety as measured by conventional parameters. mice and their wild-type cohort in risk assessment These results suggest that GRP-R- and NMB-R-deficient behavior. mice are not abnormally anxious, and that BRS-3- Similar to the results of the L-D box test, the analyses deficient mice exhibit reduced properties of anxiety. of risk assessment behavior from the elevated plus BRS-3-deficient and NMB-R-deficient mice exhibited maze test revealed that BRS-3-deficient mice displayed altered risk assessment behavior in both the L-D and increased frequency of the stretched attend posture elevated plus maze tests. BRS-3-deficient mice exhib- relative to the wild-type cohort (U = 10, P Ͻ 0.01, two- ited an increase in the frequency of the stretched attend tailed Mann–Whitney’s U-test), while NMB-R-deficient posture, whereas NMB-R-deficient mice exhibited mice displayed a relative decrease in the frequency of decreased frequency for this behavior (see Table 1). this posture (U = 13.5, P Ͻ 0.05, two-tailed Mann– Although risk assessment behavior has yet to gain Whitney’s U-test). Furthermore, in the conventional broad consent as a behavioral index of anxiety, there parameters, BRS-3-deficient mice exhibited a longer are reports that this behavior may reflect some duration of remaining in the open arms (Time Spent emotional state of animals.21,22 Therefore, our results in Open Arm, and % Time Spent in Open Arm) com- suggest that the NMB/NMB-R pathway modulates pared to the wild-type cohort (U = 2, P Ͻ 0.01; U = 2, some forms of emotion (perhaps including anxiety) in P Ͻ 0.01, two-tailed Mann–Whitney’s U-test, a manner opposite to that of the BRS-3 pathway. In respectively). However, neither BRS-3- nor NMB-R- contrast to the results for these two knockout strains, deficient mice exhibited a significant difference in GRP-R-deficient mice showed no significant difference ethological indices other than risk assessment behavior in either risk assessment behavior or conventional anx- (eg, grooming behavior). Again, as in the light-dark box iety parameters compared to their wild-type cohort. test, GRP-R-deficient mice showed no altered risk Therefore, the GRP/GRP-R pathway may not impact this type of anxiety and/or emotional behavior.

Molecular Psychiatry Anxiety in BN-like peptide receptor knockout mice K Yamada et al 115 Previously, we reported the social response in these system.27 Our results indicate that the BN-like three strains of BN-like peptide receptor deficient mice peptide/receptor pathway may be involved in modulat- using an ordinal resident-intruder test paradigm. GRP- ing emotion in animals, and that each of the three path- R-deficient mice exhibited increased social ways we examined may differentially affect emotion. response,11,12 whereas BRS-3-deficient mice exhibited a There is a great need to clarify the neural and psycho- decrease in such behavior.15 In contrast to these results, logical mechanisms of anxiety and emotional behavior, NMB-R-deficient mice did not show altered social and to establish therapeutics for these abnormalities. response.16 Table 2 summarizes non-aggressive social Our present data indicate that BN-like peptide/receptor behavior, anxiety-related behavior, and risk assessment systems may represent risk factors for psychiatric dis- behavior in the three strains of BN-like peptide recep- ease. Further behavioral, biochemical and molecular tor knockout mice. The frequency of non-aggressive biological analyses are required to clarify the role of social behavior in a social interaction situation may the BN-like peptide/receptor systems in the neuro- reflect a level of emotion other than anxiety as meas- transmission aspects of emotion. ured by the L-D box and the elevated plus maze tests. In contrast to our present results, GRP-R-deficient mice Methods and BRS-3-deficient mice exhibited altered behavior to the intruder in the social environment,11,12,15 suggest- Housing and experimental conditions ing that the GRP/GRP-R pathway and BRS-3 pathway Animals were housed and all experiments were con- act in opposition with respect to modulating emotional ducted in a temperature- and humidity-controlled behavior aroused by social stimuli. environment (23 ± 2°C, 60 ± 5%). The light-dark cycle Pharmacological studies report that exogenously was 12 h:12 h (lights on at 0800). Subject mice were administered BN-like peptides may affect an animal’s housed individually in ordinary plastic caging (JCL emotional state (eg, an increase in scratching/grooming Inc, Japan, CL-0103–1pc, 190 × ෂ260 × ෂ125(H) mm, behavior).23,24 However, no change in these behaviors with wood shavings) for one month before experimen- was observed in our previous studies.10–16 Here, we tation. Ordinary laboratory chow (JCL Inc, CE-2, also assessed grooming as well as other behaviors using 342.2 kcal per 100 g) and water were freely available to ethological indices. Again, as in the open field test mice except during experimental trials. All experi- and/or the social interaction test situation, none of ments were performed during the light cycle between knockout mice exhibited changes in these behaviors 1300 and 1700. using the elevated plus maze test. This discrepancy between the pharmacological studies and knockout Animals mice studies suggests that the pharmacological effects All subject mice were bred in our laboratory. Male of BN-like peptides are not consistent with the physio- mice were used as subjects to maintain consistency logical function of these peptides. Moreover, the with previous studies that used male subjects to assess absence of change in these behaviors may provide risk assessment behavior.21,22 As the GRP-R and the additional evidence in support of the idea that the type BRS-3 genes are located on chromosome X, male GRP- of emotional behavior to which each of the R-deficient mice, BRS-3-deficient mice and their wild- peptide/receptor pathways contributes can be speci- type littermates were obtained by mating a GRP-R (or fied. BRS-3) heterozygous female (+/−) with a male Studies report that BN-like peptide/receptor path- C57BL/6J (JCL Inc). On the other hand, since the NMB- ways may modulate neurotransmission related to emo- R gene is located on an autosome, male and female tion. The NMB/NMB-R system and GRP/GRP-R system NMB-R heterozygotes (+/−) were mated to obtain male are involved in the regulation of the 5-HT system NMB-R-deficient mice and wild-type littermates. The and/or stress response.25,26 Further, the GRP/GRP-R generation, the number, the age, and the body weights system may interact with the GABAA/benzodiazepine of the experimental animals are summarized in Table 3. All animal experiments were performed in strict Table 2 Summary of social behavior, anxiety-related accordance with the guidelines of the National Insti- behavior and risk assessment behavior in three strains of BN- tute of Neuroscience, National Center of Neurology and like peptide receptor knockout mice Psychiatry (Japan), and were approved by the Animal Investigation Committee of the Institute. GRP-R BRS-3 NMB-R Apparatus Non-aggressive social ↑10 ↓15 –16 A two-compartment light-dark (L-D) box (Muromachi behavior ↓ Kikai, Japan) was used to measure exploratory Anxiety-related – – × × behavior behaviors. The light compartment (9 9.5 14 (H) cm) Risk assessment – ↑↓consisted of white Plexiglas walls and a clear Plexiglas behavior (stretched ceiling, while both the walls and ceiling of the dark attend posture) compartment (14 × 10.5 × 14 (H) cm) were made of black Plexiglas. A 5 × 5 cm sliding door connected the ↑: Increased; ↓: decreased; –: unchanged. two compartments. The floor of both two compart-

Molecular Psychiatry Anxiety in BN-like peptide receptor knockout mice K Yamada et al 116 Table 3 Subjects open arm. A standard test of 5-min duration was employed and the maze was thoroughly cleaned Subjects/test L-D box test Elevated plus maze between tests with damp towels and 70% alcohol to test remove any residual odor of the prior mice due to urine, feces, or sebum. Behavior was measured both GRP-R-deficient and wild-type mice (F12*) 5 months of age manually during the experiment and later from video- GRP-R-deficient n = 7 (28.9 ± 1.1 g) n = 10 (29.0 ± 0.58 g) tape (the video camera was positioned ෂ1 m above the mice maze at an angle of ෂ50° relative to the platform and = ± = ± wild-type mice n 7 (28.9 0.46 g) n 7 (28.6 0.45 g) closed arms; and the camera was controlled manually). Behavioral indices used in this study were modified BRS-3-deficient and wild-type mice (F13) 3 months of age from Dalivi & Rodgers21 and consisted of the frequency BRS-3-deficient n = 8 (27.2 ± 0.39 g) n = 8 (26.3 ± 0.43 g) mice and the ratio (%) of open and closed arm entries (arm wild-type mice n = 8 (25.3 ± 0.56 g) n = 8 (24.5 ± 0.59 g) entry defined as all four paws into an arm), and the duration of staying in the open and closed arms of the NMB-R-deficient and wild-type mice (F9 × F9) 6 months of maze. In addition to these parameters, the following age ethologically derived measures were recorded and ana- NMB-R-deficient n = 7 (34.9 ± 2.0 g) n = 8 (34.3 ± 1.1 g) lyzed: the frequency of rearing and leaning on the wall mice (mice stand on hind paws with/without leaning on the = ± = ± wild-type mice n 6 (37.6 2.3 g) n 8 (33.5 1.5 g) wall of the maze); the frequency of stretched attend posture (exploratory posture where the mouse ± Mean SEM; *backcrossed 12 times to C57BL/6J. stretches forward and then retracts to its original pos- ition without moving forward); the frequency of ments was constructed of stainless steel rods parallel to sniffing (olfactory exploration of the maze floor and the sliding door (5 mm diameter, spaced 10 mm apart). walls, and air sampling); the frequency of grooming The elevated plus maze (Muromachi Kikai, Japan) (washing face with forepaws, scratching/grooming its was constructed of brown smoke Plexiglas and con- fur with hind paws); the frequency of flatback sisted of a central 9 × 9-cm platform and four arms radi- approach behavior (exploratory locomotion where the ating from that platform in the shape of a plus. Two animal stretches to its full length and cautiously moves opposite arms were open (30 × 6 cm; length × width) forward); the frequency of step back behavior (mouse and the other two arms (30 × 6 cm) were opposite each moves forward, then steps back to the previous pos- other and enclosed by 15-cm high walls (including the ition without turning its body); and the duration of end of the arm) with an open top. The maze floor was immobility (no visible exploration, locomotion, and covered with black rubber and a slight raised edge sniffing, but not including ‘freezing’ in which animals (5 mm) was placed around the perimeter of the open become rigid). arms to provide additional grip for the animals. The The frequency of open and closed arm entry and the entire apparatus was elevated to a height of 40 cm time spent in open and closed arms were measured above floor level. manually, and the other behaviors were scored by VTR analysis. Videotapes were analyzed at least three times Light-dark box test by two highly trained observers who were unaware of The L-D box test was performed on mice following each animal’s experimental designation. Median scores acclimation to individual housing. Mice were placed were used for statistical analyses. individually into the dark compartment of the L-D box, and after the ceiling was shut, the sliding door was opened. Then mice were left to explore the L-D box Acknowledgments for 5 min during which the following parameters were We thank Dr H Ogura (Eisai Co Ltd) and Dr S Okuyama measured: latency to enter the light compartment, (Taisho Pharmaceutical Co Ltd) for conducting the pre- number of entrances into and duration of staying in liminary experiments. This work was supported in part the light compartment, and risk assessment behavior by research grants from The Ministry of Education, (stretched attend posture). Risk assessment behavior Culture, Sports, Science and Technology, The Ministry was defined as the exploratory posture where the of Health, Labour and Welfare, and Japan Science and mouse stretches its body and enters its nose and/or Technology Corporation. head into the light compartment, and then retracts to the dark compartment without its forepaws stepping into the light compartment. After each trial, the appar- References atus was cleaned with 70% alcohol to remove any 1 Anastasi A, Erspamer V, Bucchi M. Isolation and structure of bom- residual odor of the prior mice due to urine, feces, or besin and alytensin, two analogous active peptides from the skin sebum. of the European amphibians Bombina and Alytes. Experientia 1971; 27: 166–167. 2 McDonald TJ, Jornvall H, Nilsson G, Vagne M, Ghatei M, Bloom Elevated plus maze test SR et al. Characterization of a gastrin releasing peptide from por- To initiate the elevated plus maze test, an animal was cine non-antral gastric tissue. Biochem Biophys Res Commun 1979; placed on the central platform of the maze facing an 90:227–233.

Molecular Psychiatry Anxiety in BN-like peptide receptor knockout mice K Yamada et al 117 3 Minamino N, Kangawa K, Matsuo H. Neuromedin B: a novel bom- receptors revealed by the analysis of mice lacking neuromedin B besin-like peptide identified in porcine spinal cord. Biochem receptor. J Neurosci 1999; 19:948–954. Biophys Res Commun 1983; 114:541–548. 17 Crawley JN. Behavioral phenotyping of transgenic and knockout 4 Ohiki-Hamazaki H, Wada E, Matsui K, Wada K. Cloning and mice: experimental design and evaluation of general health, sen- expression of the bombesin receptor genes in the mouse. Brain Res sory functions, motor abilities, and specific behavioral tests. Brain 1997: 762: 165–172. Res 1999; 835:18–26. 5 Wada E, Wray S, Key S, Battey J. Comparison of gene expression 18 Crestani F, Lorez M, Baer K, Essrich C, Benke D, Laurent JP et al.

for two distinct bombesin receptor subtypes in rat central nervous Decreased GABAA-receptor clustering results in enhanced anxiety system. Mol Cell Neurosci 1992; 3:446–460. and a bias for threat cues. Nature Neurosci 1996; 2: 833–839. 6 Flood JF, Morley JE. Effects of bombesin and gastrin-releasing pep- 19 Ferrari PF, Palanza P, Parmigiani S, Rodgers RJ. Interindividual tide on memory processing. Brain Res 1988; 460:314–322. variability in Swiss male mice: relationship between social factors, 7 Flynn FW. Effects of fourth ventricle bombesin injection on meal- aggression, and anxiety. Physiol Behav 1998; 63: 821–827. related parameters and grooming behavior. Peptide 1991; 12: 20 Rodgers RJ, Johnson NTJ, Champion AJ, Mills S. Modulation of 761–765. plus-maze behavior in mice by the preferential D3-receptor agonist 8 Kirkham TC, Perez S, Gibbs J. Prefeeding potentiates anorectic 7-OH-DPAT. Pharmacol Biochem Behav 1996; 54:79–84. actions of neuromedin B and gastrin-releasing peptide. Physiol 21 Dalvi A, Rodgers RJ. GABAergic influences on plus maze behavior Behav 1993; 54:467–470. in mice. Psychopharmacology 1996; 128:380–397. 9 Santo-Yamada Y, Yamada K, Wada K. Post-training administration 22 Grewal SS, Shepherd JK, Bill DJ, Fletcher A, Dourish CT. of gastrin-releasing peptide (GRP) improves memory loss in scopol- Behavioral and pharmacological characterisation of the canopy amine- and hypoxia-induced amnesic mice. Physiol Behav 2001; stretched attend posture test as a model of anxiety in mice and rats. 74:139–143. Psychopharmacology 1997; 133:29–38. 10 Wada E, Watase K, Yamada K, Ogura H, Yamano M, Inomata Y et 23 Gibbs J, Kulkosky PJ, Smith GP. Effects of peripheral and central al. Generation and characterization of mice lacking gastrin-releas- bombesin on feeding behavior of rats. Peptides 1981; 2 Supple 2: 179–183. ing peptide receptor. Biochem Biophys Res Commun 1997; 239: 24 Masui A, Kato N, Itoshima T, Tsunashima K, Nakajima T, Yanai- 28–33. hara N. Scratching behavior induced by bombesin-related peptides. 11 Wada K, Wada E, Watase K, Yamada K, Ohki-Hamazaki H. Bom- Comparison of bombesin, gastrin-releasing peptide and phyllolitor- besin, obesity, and social behavior. Mol Psychiatry 1998; 3:204– ins. Eur J Pharmacol 1993; 238:297–301. 206. 25 Pinnock RD, Woodruff GN. Bombesin excites a subpopulation of 12 Yamada K, Wada E, Wada K. Male mice lacking gastrin-releasing 5-hydroxytryptamine-sensitive neurons in the rat dorsal raphe peptide receptor (GRP-R) display elevated preference for conspe- nucleus in vitro. J Physiol 1991; 440:55–65. cific odors and increased social investigatory behaviors. Brain Res 26 Merali Z, McIntosh J, Kent P, Michaud D, Anisman H. Aversive 2000; 870:20–26. and appetitive events evoke the release of corticotropin-releasing 13 Ohki-Hamazaki H, Watase K, Yamamoto K, Ogura H, Yamano M, hormone and bombesin-like peptides at the central nucleus of the Yamada K et al. Mice lacking bombesin receptor subtype-3 develop amygdala. J Neurosci 1998; 18: 4758–4766. metabolic defects and obesity. Nature 1997; 390:165–169. 27 Yamada K, Wada E, Wada K. Female gastrin-releasing peptide 14 Yamada K, Wada E, Imaki J, Ohki-Hamazaki H, Wada K. Hyperres- receptor (GRP-R)-deficient mice exhibit altered social preference ponsiveness to palatable and aversive taste stimuli in genetically for male conspecifics: implications for GRP/GRP-R modulation of obese (bombesin receptor subtype-3 deficient) mice. Physiol Behav GABAergic function. Brain Res 2001; 894:281–287. 1999; 66:863–867. 15 Yamada K, Ohki-Hamazaki H, Wada K. Differential effects of social Correspondence: K Yamada, PhD, Department of Degenerative Neuro- isolation upon body weight, food consumption, and responsiveness logical Diseases, National Institute of Neuroscience, National Center to novel and social environment in bombesin receptor subtype-3 of Neurology and Psychiatry, 4–1-1 Ogawahigashi, Kodaira, Tokyo (BRS-3) deficient mice. Physiol Behav 2000; 68: 555–561. 187–8502, Japan. E-mail: yamadaȰncnp.go.jp 16 Ohki-Hamazaki H, Sakai Y, Kamata K, Ogura H, Okuyama S, Wat- *These two authors contributed equally to this work. ase K et al. Functional properties of two bombesin-like peptide Received 20 July 2001; accepted 23 July 2001

Molecular Psychiatry