(Japanese Herbal) Medicine, Koso-San (Xiang-Su-San), and Its Mode of Action Via the Hypothalamic–Pituitary–Adrenal Axis N

ARTICLE IN PRESS

Phytomedicine 13 (2006) 658–667 www.elsevier.de/phymed

Antidepressant-like activity of a Kampo (Japanese herbal) medicine, Koso-san (Xiang-Su-San), and its mode of action via the hypothalamic–pituitary–adrenal axis N. Itoa,b, T. Nagaia,c, T. Yabea,c, S. Nunomea,c, T. Hanawac, H. Yamadaa,b,c,Ã aKitasato Institute for Life Sciences, Kitasato University, 5-9-1 Shirokane, Minato-ku, Tokyo 108-8641, Japan bGraduate School of Pharmaceutical Sciences, Kitasato University, 5-9-1 Shirokane, Minato-ku, Tokyo 108-8641, Japan cOriental Medicine Research Center, The Kitasato Institute, 5-9-1 Shirokane, Minato-ku, Tokyo 108-8642, Japan

Abstract

Koso-san (Xiang-Su-San in Chinese), a Kampo (Japanese herbal) medicine, is used clinically in East Asia for the treatment of depression-like symptoms associated with the initial stage of the common cold, allergic urticaria due to food ingestion, irritable bowel syndrome, chronic fatigue syndrome, insomnia, and autonomic imbalance. However, the antidepressant-like activity of Koso-san has never been evaluated scientifically. In this study, ddY mice subjected to a combination of forced swimming and chronic mild stresses were termed depression-like model mice. The degree of the depression-like state was measured by the animal’s duration of immobility using the forced swimming test (FST). Oral administration of Koso-san (1.0 g/kg/body wt./day, 9 days) significantly shortened the duration of immobility of the depression-like model mice in the FST; however, locomotor activity was not affected. Hyperactivity of the hypothalamic–pituitary–adrenal (HPA) axis plays an important role in the pathophysiology of depression. Levels of corticotropin-releasing hormone mRNA expression in the hypothalamus and proopiomelanocortin mRNA expression in the pituitary were significantly increased, and glucocorticoid receptor protein expression in the hypothalamus paraventricular nucleus was downregulated in the depression-like model mice. However, Koso-san ameliorated these alterations to the normal conditions. The results of this study suggest that Koso-san shows the antidepressant-like effect through suppressing the hyperactivity of the HPA axis in depression-like model mice. r 2006 Elsevier GmbH. All rights reserved.

Keywords: Koso-san; Xiang-Su-San; Antidepressant; HPA axis; Forced swimming test; Chronic mild stress; Depression-like model mice

Introduction (Cyperi rhizoma, Perillae herba, Aurantii nobilis pericarpium, Glycyrrhizae radix and Zingiberis rhizoma) Koso-san, a Kampo (Japanese herbal) medicine and is used clinically to treat depression-like symptoms (Xiang-Su-San in Chinese), is composed of ﬁve herbs associated with the initial stage of the common cold, allergic urticaria due to ingestion of food, irritable bowel syndrome, chronic fatigue syndrome, insomnia and ÃCorresponding author. Kitasato Institute for Life Sciences, Kitasato University, 5-9-1 Shirokane, Minato-ku, Tokyo 108-8641, autonomic imbalance. In addition, it has been clinically Japan. Tel.: +81 3 5791 6364; fax: +81 3 5791 6365. suggested that Koso-san can alleviate depression, a E-mail address: [email protected] (H. Yamada). common side effect of interferon-a therapy for hepatitis

C infection (Hanawa, 1995). Pharmacological studies 1987) in the present study. The present study describes have reported that Koso-san is effective for globus the antidepressant-like effects of Koso-san using depres- pharyngeus (Motoo et al., 1999), skin injury induced by sion-like model mice, and its possible mechanism of X-rays (Wang et al., 1990) and secretion of bile (Sasaki action on the HPA axis. et al., 1989). However, no investigation of the possible mechanisms of action of Koso-san’s antidepressant-like effects has yet been reported. Materials and methods Clinical depression is an endemic disorder in modern societies. The incidence of human depression is world- Animals wide, with a prevalence of 4–16% in some developed countries (Wong and Licinio, 2001). Stressful lifestyles Seven-week-old male ddY mice (Japan SLC, Hama- and events often precede the onset of clinical depression matsu, Japan), weighing 35–40 g each at the beginning and stress-related neuroendocrine mechanisms are of the experiment, were used. The mice were housed proposed to be involved in the pathogenesis of affective under conditions of constant temperature (2372 1C) disorders (Checkley, 1992; Sapolsky, 1996). and humidity (55710%) with food and water available Impaired function of the hypothalamic–pituitary– ad libitum (unless specified otherwise), and a 12/12 h adrenal (HPA) axis has consistently been observed in light–dark cycle (8:00–20:00) during the stress proce- patients with clinical depression. Hyperactivation of the dure. All animal experiments were performed according HPA axis results in stimulating levels of corticotropin- to the Guidelines for Care and Use of Laboratory releasing hormone (CRH), adrenocorticotropic hor- Animals at Kitasato Institute and Kitasato University. mone (ACTH) and glucocorticoid in the hypothalamus, pituitary and adrenal cortex, respectively. Abnormalities of the HPA axis have been reported with levels returning Drugs to normal values after treatment with antidepressant The following drugs were used: desipramine hydro- drugs or electroconvulsive therapy (Gold et al., 1988; s Pariante and Miller, 2001; Stout et al., 2002). Gluco- chloride (Sigma, St. Louis, MO, USA), Toledomin corticoid receptors (GR) are associated with the (milnacipran hydrochloride) (Asahi Kasei Pharma Corp., Tokyo, Japan), Luvoxs (fluvoxamine maleate) (Solvay negative-feedback regulation of the HPA axis and s believed to play an important role in regulating the Seiyaku Co., Tokyo, Japan), Horizon (diazepam) response to stress when the endogenous level of (Yamanouchi Pharmaceutical Co. Ltd., Tokyo, Japan). glucocorticoid is high (Pariante and Miller, 2001). Component herbs of Koso-san (Table 1) were Treatment of transgenic mice with impared GR function purchased from Tsumura & Co., Tokyo, Japan (C. using tricyclic antidepressant drugs shows increases in rhizoma, P. herba, A. nobilis pericarpium and Z. hypothalamic GR mRNA expression and dexametha- rhizoma) and Uchida Wakan-yaku Co. Ltd., Tokyo, sone-binding activities, while decreasing levels of plasma Japan (G. radix). The Kampo formula was decocted ACTH and corticosterone, resulting in the disappear- with 600 ml of distilled water until the volume was ance of deficient behaviors (Montkowski et al., 1995; reduced by half, and the extract was filtered immediately Pepin et al., 1992). These findings suggest that normal- through paper in vacuo. The filtrate was lyophilized and ization of GR dysfunction may be closely associated the yield of Koso-san extract was ca. 28% of the herbal with the recovery of impaired neuroendocrine and mixture relative to dry weight. depressive moods. To create a depressive-like state using an animal Three-dimensional (3-D) HPLC analysis of model, ddY mice were subjected to chronic and Koso-san unpredictable stressors such as the forced swimming (FS) (Porsolt et al., 1977; Detke et al., 1997) and chronic The components of Koso-san extract were identified mild stress (CMS) (Solberg et al., 1999; Willner et al., using 3-D HPLC with an Agilent 1100 3-D HPLC

Table 1. The composition of Koso-san

Herb ingredients Weight (g)

Cyperi rhizoma (the rhizome of Cyperus rotundus L.) 4.0 Perillae herba (the leaf of Perilla frutescens Britton var. acuta Kudo) 2.0 Aurantii nobilis pericarpium (the fruit of Citrus unshiu Markovich) 3.0 Glycyrrhizae radix (the root of Glycyrrhiza uralensis Fisher) 2.0 Zingiberis rhizoma (the rhizome of Zingiber ofﬁcinale Roscoe) 0.5 ARTICLE IN PRESS 660 N. Ito et al. / Phytomedicine 13 (2006) 658–667

Fig. 1. The 3-D HPLC profile of Koso-san: (A) glycyrrhizin (Glycyrrhizae radix); (B) liquiritigenin (Glycyrrhizae radix); (C) rosmarinic acid (Perillae herba); (D) hesperidin (Aurantii nobilis pericarpium); (E) narirutin (Aurantii nobilis pericarpium); (F) liquiritin (Glycyrrhizae radix); and (G) caffeic acid (Perillae herba). system (Agilent Technologies, Tokyo, Japan) equipped et al., 1977; Detke et al., 1997) and CMS (Solberg with a photodiode-array detector. A TSK gel ODS-80Ts et al., 1999; Willner et al., 1987). Briefly, the mice column (4.6 250 mm) (Tosoh Corp., Tokyo, Japan) were placed individually into separate 51 glass beakers was used and kept at 40 1C. Elution of the mobile phase (height 27 cm, diameter 18 cm) filled with 4 l of was carried out by 10 mM phosphoric acid-acetonitrile water (2371 1C) for 15 min. Beakers were separated linear gradient (95:5–5:95, 60 min), and its flow rate was by non-transparent panels to prevent the mice from 0.8 ml/min. The following standard samples were used seeing each other. After 15 min in the water, mice for identification: rosmarinic acid (purchased from were removed and allowed to dry with a drier before Extrasynthe` se, Genay, France), caffeic acid (Extra- being returned to their home cages. Mice were synthe` se), luteolin (purchased from Roth, Karlsruhe, then separated into groups by measuring the duration Germany), luteolin-7-glucoside (Roth), l-menthol (pur- of immobility for the first 5 min of FS, in order to chased from Wako Pure Chemical Industries, Osaka, minimize the variation in immobility among the groups. Japan), apigenin (Roth), l-perillaldehyde (Extrasynth- A mouse was judged to be immobile when it ceased e` se), hesperidin (Extrasynthe` se), naringin (Extrasynth- struggling and remained floating motionless in the water e` se), naringenin (Roth), narirutin (Extrasynthe` se), making only those movements necessary to keep its glycyrrhizin (Wako), glycyrrhetic acid (Wako), liquiritin head above water. After 2 days, mice were exposed to (Wako), liquiritigenin (Extrasynthe` se), formononetin CMS, which consisted of three different stress situa- (Roth) and 6-gingerol (Wako). A 3D-HPLC profile of tions: tilting the cage twice to 301 from the horizontal the Koso-san extract dissolved in distilled water (5 mg/ (CMS 1), pouring 200 ml of water onto the sawdust ml) is shown in Fig. 1. In the present study, seven bedding of a cage (CMS 2), and shaking the cages at components (glycyrrhizin, liquiritigenin, rosmarinic 200 rpm with a Green S. Seriker II (Vision Scientific, acid, hesperidin, narirutin, liquiritin and caffeic acid) Kyunggi, Korea) (CMS 3). These stress situations were were identified. applied for 48, 24 and 24 h, respectively, with 24-h intervals (Scheme 1). Mice were then placed again Preparation of depression-like model mice into 5 l beakers with 4 l of water at 60 min after the final drug administration, and the total duration of immobi- The depression-like model mice were prepared lity during 5 min of forced swimming test (FST) was using modified FS twice with an 11-day interval (Porsolt observed. ARTICLE IN PRESS N. Ito et al. / Phytomedicine 13 (2006) 658–667 661

0 1 2 3 4 5 6 7 8 9 10 11 (days) lin. CRH, POMC and cyclophilin mRNA produced PCR products of 585, 304 and 396 bp, respectively. FS (15 min) CMS 1 CMS 2 CMS 3 CMS 1 1 h FST CRH, POMC or cyclophilin cDNA product was (5 min) ampliﬁed for 27, 16 or 18 cycles, each cycle consisting of denaturation at 96 1C for 30 s, annealing at 58, 67 or Drugs (500 µL/mouse/day, p. o.) 57 1C for 30 s, respectively, and extension at 72 1C for Scheme 1. Schedule for drug administration in the depression- 3 min. PCR products were analyzed on a 2% agarose gel like model mouse. containing 0.0025% ethidium bromide (Sigma).

Drug administration Measurement of corticosterone

All drugs were suspended in distilled water. Diazepam Blood samples were collected by decapitation of mice (0.5 mg/kg/body wt.), desipramine (30 mg/kg/body wt.), at 60 min after final drug administration and were fluvoxamine (60 mg/kg/body wt.), milnacipran (60 mg/ centrifuged at 4 1C. Sera were stored at 80 1C until kg/body wt.) and Koso-san (0.1, 0.5 and 1.0 g/kg body assayed for corticosterone. wt.) were administered orally to depression-like model Serum corticosterone was measured using a Corre- mice by intragastric gavage at a volume of 0.5 ml/mouse late-EIATM Corticosterone Enzyme Immunoassay Kit once daily on days 1, 2, 4, 5, 6, 7, 8, 9 and 11 (9 times) (Assay Designs, Ann Arbor, MI, USA) according to the (Scheme 1). manufacturer’s instructions. The sensitivity of the measurement was 26.99 pg/ml. The intra- and inter- Spontaneous locomotor activity test assay coefficients of variations were 8% and 13%, respectively. Spontaneous locomotor activities of mice were evaluated using the open field test. Briefly, mice were Immunohistochemical study placed individually in the center of an open field (40 40 20 cm), with the field divided into 25 equal Mice were quickly anesthetized with sodium amobarbi- squares (Sanki Kagaku Kogei, Kawasaki, Japan), and tal (Sigma) 5 min after the FST and perfused transcardially allowed to move freely for 15 min. The total number of with 0.9% saline, followed by 4% paraformaldehyde line crosses for the first 5 min was counted in order to (Wako) at twice the volume of body weight. minimize variation in the total number of line crosses Serial coronal brain sections of 50-mm thickness were among the groups. Eleven days later, mice were again cut with a vibratome. Free-floating sections were placed individually in the open field and the total preincubated for 2 h at 20 1C in 80% methanol/3% number of line crosses was counted during a 5-min H O to eliminate endogenous peroxidases. After interval, 60 min after the final drug administration. 2 2 washing in phosphate buffered saline (PBS), sections were incubated for 2 h at room temperature with 1% Reverse transcription-polymerase chain reaction bovine serum albumin (BSA) in PBS containing 0.3% (RT-PCR) Triton X-100 (PBS-T) and incubated overnight at 4 1C with rabbit antiserum against glucocorticoid receptor Total RNA was isolated from the hypothalamus or (GR) 1:1600 (Santa Cruz Biotechnology, Santa Cruz, pituitary of depression-like model mice by the acid CA, USA) in PBS-T containing 1% BSA. After washing guanidinium–phenol–chloroform method using ISO- in PBS-T, sections were incubated for 2 h at room GEN (Nippon gene, Toyama, Japan). Total RNA was temperature with biotinylated goat anti-rabbit IgG used as a template for cDNA synthesis by the first-strand (Vector Laboratories, Burlingame, CA, USA) 1:200 in cDNA synthesis kit (Amersham Biosciences, Piscataway, PBS-T containing 1% BSA, and washing PBS-T, NJ, USA). PCR amplification of this cDNA was followed by amplification for 2 h at room temperature performed using Taq DNA polymerase (Takara Bio, with an avidin–biotin complex using Vectastain Elite Otsu, Japan). Primers (Invitrogen, Carlsbad, CA, USA) ABC kit (Vector Laboratories) in PBS-T. Sections were used for PCR were as follows: 50 GCA TCC TGA GAG incubated with PBS containing 0.05% diaminobenzidine 0 0 AAG TCC CTC TG 3 (sense) and 5 GCC CTG GCC (Sigma) and 0.02% H2O2. After washing in PBS, ATT TCC AAG AC 30 (antisense) for CRH, 50 CGG sections were counterstained with 0.05% toluidine blue, CCC CAG GAA CAG CAG CAG T 30 (sense) and 50 coverslipped, dehydrated, and mounted in permount GGG CCC GTC GTC CTT CTC C 30 (antisense) for (Fisher Scientific International Inc., Fair Lawn, NJ, proopiomelanocortin (POMC), 50 GGT CAA CCC CAC USA). GR-positive cells were detected and processed CGT GTT CTT CGA 30 (sense) and 50 TTG CCA TCC using an Olympus BX 41 microscope (Olympus Corp., AGC CAT TCA GTC TTG 30 (antisense) for cyclophi- Tokyo, Japan). ARTICLE IN PRESS 662 N. Ito et al. / Phytomedicine 13 (2006) 658–667

Statistical analysis Table 2. Effect of Koso-san on the duration of immobility of depression-like model mice in the FST Results were expressed as means7SEM. Behavioral Drug Dose (g/kg) Duration of p-value data, CRH, POMC mRNA data and serum corticoster- immobility (s/ one data were analyzed using one-way analysis of 5 min) variance (ANOVA). Post hoc comparisons, if applicable, were carried out using the Scheffe´test. The p-values Water 184.0715.4 — lower than 0.05 (po0:05) were considered signiﬁcant. Koso-san 0.1 163.1710.5 0.9113 0.5 148.6720.6 0.5685 1.0 101.7714.0 0.0074

Milnacipran 0.06 47.179.2 o 0.0001 Results Drugs were administered orally to depression-like model mice once daily according to the schedule shown in Scheme 1. The duration of Effects of diazepam, desipramine, ﬂuvoxamine and immobility was measured during a 5 min of FST at 60 min after the milnacipran on the duration of immobility of ﬁnal drug administration. Each duration of immobility represents the depression-like model mice in the FST mean7SEM (n ¼ 9210).

Depression-like model mice were prepared by combination of FS with three different combined chronic mild Effects of Koso-san on expressions of CRH mRNA stresses (CMS) to create depressive-like states. To in the hypothalamus and POMC mRNA in the evaluate this animal model, a known minor tranquilizer pituitary of depression-like model mice and antidepressants were administered to determine the effects on the duration of immobility of mice in the FST. The mRNA expression of CRH and POMC was Mice administered minor tranquilizer, diazepam, versus measured in the hypothalamus and pituitary gland of water (control mice) showed no significant difference in the model mice using RT-PCR. Stress induced by the duration of immobility, using the FST as an index. combination of FS and CMS significantly enhanced By contrast, administration of the antidepressants, the CRH and POMC mRNA expressions, as compared desipramine (tricyclic antidepressant), fluvoxamine with control mice (p ¼ 0:0020, Fig. 2; p ¼ 0:0123, Fig. 3, (selective serotonin reuptake inhibitor) and milnacipran respectively). The increased expressions of CRH mRNA (serotonin-noradrenaline reuptake inhibitor) reduced in the hypothalamus and POMC mRNA in the pituitary the duration of immobility significantly as compared indicated hyperactivation of the HPA axis in the to the water-administered control (p ¼ 0:0028, 0:0211 depression-like model mice. Administration of Koso- and po0:0001, respectively; data not shown). san (0.1, 1.0 g/kg body wt.) significantly and dose- dependently reduced the expression of CRH mRNA in the hypothalamus as compared to the water-adminis- Effect of Koso-san on the duration of immobility of tered control (p ¼ 0:0178 and po0:0001) (Fig. 2). The depression-like model mice in the FST positive control, milnacipran (60 mg/kg/body wt.) also reduced the expression of CRH mRNA significantly Oral administration of Koso-san reduced the duration (po0:0001). Koso-san (1.0 g/kg body wt.) or milnaci- of immobility in a dose-dependent manner from 0.1 to pran also significantly reduced the expression of POMC 1.0 g/kg/body wt./day, and Koso-san administered at mRNA in the pituitary, as compared to the water- 1.0 g/kg/body wt./day significantly reduced the duration administered control (p ¼ 0:0407, 0:0178) (Fig. 3). of immobility as compared with the water administered ¼ control (p 0:0074) (Table 2). Milnacipran, a positive Effect of Koso-san on corticosterone levels in the control, also significantly reduced the duration of sera of depression-like model mice immobility as compared to control (po0:0001). Stress caused by the combination of FS and CMS resulted in significant increases of corticosterone levels, Effect of Koso-san on the spontaneous locomotor as compared with non-stressed mice (p ¼ 0:0029) (data activity of mice not shown). Milnacipran (60 mg/kg/body wt.) reduced the level significantly, as compared to the water- The total number of line crosses was not affected by administered control (p ¼ 0:0003). Administration of the administration of Koso-san (0.1, 1.0 g/kg body wt.) Koso-san (0.1, 1.0 g/kg body wt.) did not significantly or milnacipran (60 mg/kg/body wt.) (data not shown). reduce corticosterone levels. ARTICLE IN PRESS N. Ito et al. / Phytomedicine 13 (2006) 658–667 663

CRH

Cyclophilin

Non stress Water 0.1 1.0 Milnacipran

Koso-san (g/kg)

Stress

1.4

B P = 0.0020 P< 0.0001

1.2 P< 0.0001

P = 0.0178

1.0

0.8

0.6 CRH/cyclophilin

0.4

0.2

0 Non stress Water 0.1 1.0 Milnacipran Koso-san (g/kg)

Stress Fig. 2. Effect of Koso-san on CRH mRNA expression in the hypothalamus of depression-like model mice. Drugs were administered orally to depression-like model mice once daily (0.1 or 1.0 g/kg/body wt./day for Koso-san, 60 mg/kg/body wt./day for milnacipran) according to the schedule shown in Scheme 1. The brains were quickly removed after FST, and CRH mRNA expression in the hypothalamus was measured using RT-PCR: (A) agarose gel electrophoresis pattern of CRH and cyclophilin; and (B) results of statistical analysis of CRH/cyclophilin ratio. Each column represents the mean7SEM (n ¼ 425).

Effect of Koso-san on GR protein expression in the Koso-san (0.1, 1.0 g/kg body wt.) or milnacipran (60 mg/ hypothalamic paraventricular nucleus (PVN) of kg/body wt.) increased the expression of GR protein depression-like model mice (Fig. 4C–E, respectively) as compared to the water- administered control (Fig. 4B). Compared with non-stressed mice (Fig. 4A), stress caused by the combination of FS and CMS resulted in the decrease of GR protein expression in the PVN Discussion (Fig. 4B), indicating that GR protein expression in the PVN of depression-like model mice may be down- The purpose of this study was to determine the regulated by the addition of stresses. Administration of efﬁcacy of Koso-san using ddY mice subjected to ARTICLE IN PRESS 664 N. Ito et al. / Phytomedicine 13 (2006) 658–667

POMC

Cyclophilin

Non stress Water 0.1 1.0 Milnacipran

Koso-san (g/kg)

Stress

0.4 B P = 0.0123 P = 0.0178

P = 0.0407

0.3

0.2 POMC/cyclophilin

0.1

0 Non stress Water 0.1 1.0 Milnacipran Koso-san (g/kg)

Stress Fig. 3. Effect of Koso-san on POMC mRNA expression in the pituitary of depression-like model mice. Drugs were administered as described for Fig. 2. The pituitaries were removed after FST, and POMC mRNA expression in the pituitary was measured using RT-PCR: (A) agarose gel electrophoresis pattern of POMC and cyclophilin; and (B) results of statistical analysis of POMC/ cyclophilin ratio. Each column represents the mean7SEM (n ¼ 526).

various stresses as the depression-like animal model. amine and milnacipran were evaluated by subjecting FST is a behavioral despair procedure that has been mice to the FST and measuring the duration of used widely to predict the efficacy of antidepressants immobility. Significant reductions in the durations of (Porsolt et al., 1977; Detke et al., 1997). Chronic stress is immobility in the FST were observed with the oral thought to play an important role in the etiology of administration of desipramine, fluvoxamine and milna- depression. Several studies have demonstrated the cipran but not with diazepam (data not shown), usefulness of CMS to measure the effects of antidepres- suggesting that the present animal model of combined sants (Solberg et al., 1999; Willner et al., 1987). In this of FS and CMS responds to antidepressants. study, in order to take account of the depressive mood Clinical studies have shown that patients with major induced by different daily stresses, depression-like model depression have impaired function of the HPA axis states were prepared in mice subjected to a combination (Gold et al., 1988; Nemeroff, 1996) and, thus, persistent of FS and a series of daily CMS (Scheme 1). hyperactivity in the HPA axis may play a critical role in To examine the specificities of drug responses using the pathogenesis of endogeneous depression (Barden our animal model, the effects of the minor tranquilizer et al., 1995; Heuser, 1998; Holsboer and Barden, diazepam and the antidepressants desipramine, fluvox- 1996). Because the HPA axis was activated in the ARTICLE IN PRESS N. Ito et al. / Phytomedicine 13 (2006) 658–667 665

Fig. 4. Effect of Koso-san on GR protein expression in the hypothalamus of depression-like model mice. Drugs were administered as described for Fig. 2. Mice were killed under deep amobarbital anesthesia by transcardiac perfusion with 4% paraformaldehyde after FST and the brains were removed. Immunohistochemical analysis of GR protein was performed on serial 50 mm vibratome sections of the brains of mice without (A) or with stress administered water (B), Koso-san 0.1 g/kg/body wt./day (C), Koso-san 1.0 g/ kg/body wt./day (D) or milnacipran (E). GR-positive cells represented as brown spots were detected using Olympus BX 41 microscope. Scale bar equals 100 mm and applies to all sections. depression-like model mice used in the study, it afforded a fructus or Cinnamomi cortex, in which saponins or suitable model for evaluating antidepressant compounds. essential oil are known for sedative (Hong, 1987; A dose-dependent reduction in the duration of Harada et al., 1976) or antioxidant activity (Shimizu immobility was observed with a 1.0 g/kg body wt. dose et al., 1990), did not reduce the duration of immobility of Koso-san showing the greatest effect (Table 2). (Ho et al., unpublished). These results suggest that Psychostimulants such as caffeine decrease the duration Koso-san has the antidepressant-like property, but that of immobility in the FST, but in contrast to antidepres- all Kampo medicines do not necessarily reduce the sants, caffeine causes marked locomotor stimulation duration of immobility in depression-like model mice. indicating that the effects may be nonspeciﬁc (Porsolt Since sedatives usually exhibit the suppression of et al., 1977). Koso-san did not change the locomotor locomotor activities, sedative compounds do not neces- activities of the naı¨ve mice at the doses that produced sarily exhibit antidepressant-like activity. Although reduction in the duration of immobility (data not some antioxidants exhibit antidepressant-like effects shown). Other Kampo medicines, Boi-ogi-to and (Kolla et al., 2005), Kampo medicines containing Tokaku-joki-to, containing herbs such as Zizyphus C. cortex might contain low antioxidants. ARTICLE IN PRESS 666 N. Ito et al. / Phytomedicine 13 (2006) 658–667

It has been reported that the compounds (rosmarinic lead to more effective clinical treatments of depressive acid, caffeic acid and apigenin) in Perillae herba, one of states. the component herbs of Koso-san, showed antidepressant-like effects in mice subjected to the FST (Takeda et al., 2002a, b; Nakazawa et al., 2003). The 3-D HPLC Acknowledgements analysis showed that the water extract of Koso-san contained rosmarinic acid and caffeic acid, but not We would like to thank Dr. K. Yabusaki for critical apigenin (Fig. 1). Therefore, Perillae herba may play a discussions and reading of the manuscript. We wish to partial role in the antidepressant-like effect of Koso-san. thank Ms. Y. Arai, Ms. M. Emori and Ms. T. Narikawa However, it is not clear whether rosmarinic acid is an for their technical assistance and Mr. S. Kim (the active component due its low concentration in Koso-san pharmacy of Oriental Medicine Research Center, The and its very high dose-response antidepressant effect Kitasato Institute) for his advice on Kampo medicines. (Takeda et al., 2002b). A part of the present work was supported by a grant Hyperexpressions of CRH and POMC mRNA in from Tsumura & Co., Tokyo. model mice were reduced by the oral administration of Koso-san, suggesting that hyperactivation of the HPA axis was, in part, suppressed by the Koso-san in References depression-like model mice. Moreover, the antidepressant-like effect of Koso-san would be associated with Barden, N., Reul, J.M., Holsboer, F., 1995. Do antidepres- the recovery of the negative feedback against the sants stabilize mood through actions on the hypothalamic– hyperactivation of HPA axis, by up-regulation of GR pituitary–adrenocortical system? Trends Neurosci. 18, protein expression in the PVN. However, no significant 6–11. difference in GR protein expression was observed in the Checkley, S., 1992. Neuroendocrine mechanisms and the hippocampus of depression-like model mice following precipitation of depression by life events. Br. J. Psychiat. administration of Koso-san in the present study (data Suppl., 7–17. not shown). Therefore, GR protein in the PVN Detke, M.J., Johnson, J., Lucki, I., 1997. Acute and chronic may play a more important role in the negative feedback antidepressant drug treatment in the rat forced swimming of the HPA axis than in the hippocampus. Oral test model of depression. Exp. Clin. Psychopharmacol. 5, 107–112. administration of Koso-san did not significantly reduce Gold, P.W., Goodwin, F.K., Chrousos, G.P., 1988. Clinical serum corticosterone levels (data not shown), indicating and biochemical manifestations of depression. Relation to that Koso-san may not be effective in acute stress the neurobiology of stress (1). N. Engl. J. Med. 319, situations. It has been reported that a Kampo medicine, 348–353. Saiko-ka-ryukotsu-borei-to, prevented chronic stress- Hanawa, T., 1995. Koso-san and Hange-koboku-to. J. Kampo induced disruption of the negative feedback mechanism, Med. 42, 418–426. evaluated by a dexamethasone suppression test using Harada, M., Fujii, Y., Kamiya, J., 1976. Pharmacological blood samples 10 days (recovery period) after exposure studies on chinese cinnamon. III. Electroencephalographic to the last stress situation in order to avoid the acute studies of cinnamaldehyde in the rabbit. Chem. Pharm. influences of stressors in rats (Mizoguchi et al., 2002). Bull. 24, 1784–1788. Further studies on the influence of Koso-san on Heuser, I., 1998. Anna-Monika-Prize paper. The hypothalamic–pituitary–adrenal system in depression. Pharmacopsy- corticosterone levels by dexamethasone after the recov- chiatry 31, 10–13. ery period could help to further explain the modulation Holsboer, F., Barden, N., 1996. Antidepressants and hypo- of Koso-san against HPA axis. These studies are now thalamic–pituitary–adrenocortical regulation. Endocr. Rev. underway. 17, 187–205. This study represents the first report showing the Hong, G.X., 1987. Advances in research on the seed of efficacy of Koso-san as a possible clinically useful Suanzaoren. J. Trad. Chin. Med. 12, 51–53. antidepressant Kampo medicine, using depression-like Kolla, N., Wei, Z., Richardson, J.S., Li, X.M., 2005. mice as an animal model. The results of this study show Amitriptyline and fluoxetine protect PC 12 cells from cell that Koso-san exhibits antidepressant-like action when death induced by hydrogen peroxide. Rev. Psychiatr. administered to depression-like mice by suppressing the Neurosci. 30, 196–201. hyperactivity of the HPA axis. Further studies are Mizoguchi, K., Yuzurihara, M., Ishige, A., Sasaki, H., Tabira, T., 2002. Saiko-ka-ryukotsu-borei-to, an herbal medicine, warranted and underway to identify the active compo- prevents chronic stress-induced disruption of glucocorticoid nents found in Koso-san and to elucidate in more detail negative feedback in rats. Life Sci. 72, 67–77. the mechanisms and modes of action of each active Montkowski, A., Barden, N., Wotjak, C., Stec, I., Ganster, J., component with respect to antidepressant-like activities. Meaney, M., Engelmann, M., Reul, J.M., Landgraf, R., Studies on the synergistic effects of various combina- Holsboer, F., 1995. Long-term antidepressant treatment tions of the active components found in Koso-san may reduces behavioural deficits in transgenic mice with ARTICLE IN PRESS N. Ito et al. / Phytomedicine 13 (2006) 658–667 667

impaired glucocorticoid receptor function. J. Neuroendo- analgesic and anti-inflammatory properties. J. Trad. Med. crinol. 7, 841–845. 7, 54–60. Motoo, Y., Taga, H., Su, S.B., Sawabu, N., 1999. Effect of Solberg, L.C., Horton, T.H., Turek, F.W., 1999. Circadian Koso-san on globus pharyngeus. Am. J. Chin. Med. 27, rhythms and depression: effects of exercise in an animal 283–288. model. Am. J. Physiol. 276, 152–161. Nakazawa, T., Yasuda, T., Ueda, J., Ohsawa, K., 2003. Stout, S.C., Owens, M.J., Nemeroff, C.B., 2002. Regulation of Antidepressant-like effects of apigenin and 2,4,5-trimethoxy- corticotropin-releasing factor neuronal systems and hy- cinnamic acid from Perilla frutescens in the forced swimming pothalamic–pituitary–adrenal axis activity by stress and test. Biol. Pharm. Bull. 26, 474–480. chronic antidepressant treatment. J. Pharmacol. Exp. Ther. Nemeroff, C.B., 1996. The corticotropin-releasing factor 300, 1085–1092. (CRF) hypothesis of depression: new findings and new Takeda, H., Tsuji, M., Miyamoto, J., Matsumiya, T., 2002a. directions. Mol. Psychiat. 1, 336–342. Rosmarinic acid and caffeic acid reduce the defensive Pariante, C.M., Miller, A.H., 2001. Glucocorticoid receptors freezing behavior of mice exposed to conditioned fear in major depression: relevance to pathophysiology and stress. Psychopharmacology (Berl) 164, 233–235. treatment. Biol. Psychiat. 49, 391–404. Takeda, H., Tsuji, M., Matsumiya, T., Kubo, M., 2002b. Pepin, M.C., Pothier, F., Barden, N., 1992. Antidepressant Identification of rosmarinic acid as a novel antidepressive drug action in a transgenic mouse model of the endocrine substance in the leaves of Perilla frutescens Britton var. changes seen in depression. Mol. Pharmacol. 42, acuta Kudo (Perillae herba). Nihon Shinkei Seishin 991–995. Yakurigaku Zasshi 22, 15–22. Porsolt, R.D., Bertin, A., Jalfre, M., 1977. Behavioral despair Wang, C.M., Ohta, S., Shinoda, M., 1990. Studies on chemical in mice: a primary screening test for antidepressants. Arch. protectors against radiation. XXXII. Protective effects of Int. Pharmacodyn. Ther. 229, 327–336. methanol extracts of various Taiwan crude drugs on Sapolsky, R.M., 1996. Why stress is bad for your brain. radiation injuries. Yakugaku Zasshi 110, 885–889. Science 273, 749–750. Willner, P., Towell, A., Sampson, D., Sophokleous, S., Sasaki, T., Ohta, S., Kamogawa, A., Shinoda, M., 1989. Muscat, R., 1987. Reduction of sucrose preference by Choleretic effects of methanol extracts obtained from chronic unpredictable mild stress, and its restoration by a various Chinese traditional medicine. Yakugaku Zasshi tricyclic antidepressant. Psychopharmacology (Berl) 93, 109, 487–495. 358–364. Shimizu, H., Tsuchiya, K., Sakurai, H., 1990. Superoxide Wong, M.L., Licinio, J., 2001. Research and treatment dismutase-like activities of Kampo-formulations possessing approaches to depression. Nat. Rev. Neurosci. 2, 343–351.