* CorrespResearch,Nanjing 21Sciences, BNanjing 21E-mail: zh†These acorrespond

PHYTOTHERAPY RESEARCHPhytother. Res. 28: 1102–1106 (2014)Published online 13 December 2013 in Wiley Online Library(wileyonlinelibrary.com) DOI: 10.1002/ptr.5103

Copyright

SHORT COMMUNICATION

The Antidepressant Effects of Ginseng TotalSaponins in Male C57BL/6N Mice by EnhancingHippocampal Inhibitory Phosphorylation of GSK-3β

Lin Chen,1 Jianguo Dai,1 Zhongli Wang,1 Huiyu Zhang,1 Yufang Huang2*† and Yunan Zhao3*†1Basic Medical College, Nanjing University of Chinese Medicine, Nanjing 210023, China2Laboratory of Pathological Sciences, Basic Medical College, Nanjing University of Chinese Medicine, Nanjing 210023, China3Key Laboratory of Brain Research, Basic Medical College, Nanjing University of Chinese Medicine, Nanjing 210023, China

Ginseng total saponins (GTS) are principal bioactive ingredients ofPanax ginseng. In this study, we investigated theantidepressant effect of GTS on the corticosterone-induced mouse depression model and explored the underlyingmechanism. Corticosterone (20mgkg�1 d�1) was subcutaneously administered for 22 d to induce the model, beforedoses of GTS (12.5, 25, and 50mgkg�1 d�1) or fluoxetine (10mgkg�1 d�1) were subsequently given by gavage. Onday 20 and 21, depression-like behavior was observed via a forced swimming test and a tail suspension test respec-tively. At 6 h after the last dose of corticosterone (day 22), all mice were sacrificed followed by serum corticosteroneassays and Western blot analysis. The results showed that GTS (25 and 50mgkg�1 d�1) treatments relieveddepression-like behavior without altering the elevated serum corticosterone levels. Furthermore, GTS treatmentsraised the down-regulated levels of hippocampal glycogen synthase kinase-3β (GSK-3β) inhibitory phosphorylation.In contrast, fluoxetine (10mgkg�1 d�1) treatment reversed the increased corticosterone level and had no effect onthe decreasedGSK-3β inhibitory phosphorylation. These findings confirmed the antidepressant effect of GTS in thecorticosterone-induced mouse depression model. Enhancing GSK-3β inhibitory phosphorylation may be one of theunderlying mechanisms. Copyright © 2013 John Wiley & Sons, Ltd.

Keywords: ginseng total saponins; corticosterone; depression-like behavior; glycogen synthase kinase-3β; hippocampus.

INTRODUCTION

Ginseng total saponins (GTS) are the principalbioactive ingredients of Panax ginseng C.A. Meyer(Araliaceae) which has been used as a popular tonicherb for 2000 years in Far East countries. Recently,ginsenosides have been reported to have several benefi-cial functions in the brain, including antidepressanteffects (Dang et al., 2009; Jiang et al., 2012). Despitethe widely demonstrated activity of ginsenosides ondepression, the underlying mechanism remains to beelucidated. Some studies implied that the classicalmonoaminergic transmission hypothesis may notaccount for the antidepressant effects of ginsenosides(Jiang et al., 2012).Glycogen synthase kinase-3 (GSK-3) is an inhibitory

signaling molecule which regulates a large group oftranscription factors and transcriptional modulators. Itexists in two closely related isoforms, namely, GSK-3αand GSK-3β. The constitutively active GSK-3β is dem-onstrated to be involved in many intracellular signalingpathways related to neuroplasticity; however, its activityis inhibited by Ser9 phosphorylation. Previous studies

ondence to: Yunan Zhao, Ph.D., Key Laboratory of BrainBasic Medical College, Nanjing University of Chinese Medicine,0023, China; Prof. Yufang Huang, Laboratory of Pathologicalasic Medical College, Nanjing University of Chinese Medicine,0023, China.aoyunan-js@163.com; yufanghnj@163.comuthors contributed equally to the work and share theing-authorship.

© 2013 John Wiley & Sons, Ltd.

have shown that GSK-3β activity is significantlyincreased in the hippocampus of adult rats subjected tochronic mild stress (CMS) or patients with majordepressive disorders (Silva et al., 2008; Oh et al., 2010).Antidepressant-like behavior was likewise observed inGSK-3β haploinsufficient mice (O’Brien et al., 2004),as well as in CMS rats treated with GSK-3β inhibitors(Silva et al., 2008). In the present study, we observedthe antidepressant effects of GTS on the corticoste-rone-induced mouse model of depression, and exploredthe underlying mechanism with respect to GSK-3β.

MATERIALS AND METHODS

Preparation and quality assessment of GTS. Chineseginseng, specifically the root of Panax ginseng, waspurchased from Beijing Tong Ren Tang Group Co.,Ltd. (Beijing, China). Following the methodologyprovided by China Pharmacopoeia (2010), the air-driedginseng (150 g) was powdered and then decocted thricewith 1.2 L of deionized water (for 1 h during each decoc-tion). The decoction was collected and filtered. Then thefiltrate was pumped through a fixed-bed column(20mm×300mm) filled with 50 g of D101 macroporousadsorption resin (dry weight) at 50mLh�1. When theadsorption reached equilibrium, 10 bed volumes ofdistilled water were pumped through the column toremove the contaminants at a rate of 250mLh�1.

Received 29 September 2013Revised 18 November 2013

Accepted 23 November 2013

1103GINSENG TOTAL SAPONINS ON INHIBITORY PHOSPHORYLATION OF GSK-3β

Subsequently, five bed volumes of 60% aqueous ethanolwere used to elute the ginsenosides in isocratic mode ata flow rate of 60mLh�1. The eluent was collected anddried under vacuum at 60 °C to produce the purifiedGTS (3.709 g).For quality control, ultrahigh-performance liquid

chromatography with a charged aerosol detector wasapplied to quantify the marker components. The totalsaponin content was estimated using the colorimetricmethod with a vanillin–vitriol system. The total saponincontent of GTS was 66.8%±1.4%. The amount ofrepresentative ginsenosides Rg1, Re, and Rd was4.9%, 3.5%, and 1.5%, respectively.

Experimental design. A total of 72 male C57BL/6Nmice (weighing 18 g to 20 g, from the Laboratory Ani-mal Center of Nanjing Medical University, Nanjing,China) were allowed 1week to adapt to the laboratoryenvironment before the actual experiment. The micewere divided into six matched groups (n= 12 per group),and the weight gain among these groups were not statis-tically different at the start of the experiment. Groups ofsix animals were housed in each cage with a 12 h light/12 h dark cycle (lights on between 7:00 and 19:00), at aconstant room temperature of 22 ± 1 °C, with free accessto food and tap water. The animals were treatedaccording to the Guidelines of Accommodation andCare for Animals formulated by the Chinese Conven-tion for the Protection of Vertebrate Animals Used forExperimental and Other Scientific Purposes. Everyeffort was made to minimize the suffering and thenumber of animals used for the experiment.The six matched groups are as follow: (i) the control

group, (ii) the corticosterone group (CORT), (iii) the cor-ticosterone+ 10mgkg�1 d�1 fluoxetine (CORT+FLU)group, (iv) the corticosterone+ 50mgkg�1 d�1 GTS(CORT+GTSH) group, (v) the corticosterone+ 25mgkg�1 d�1 GTS (CORT+GTSM) group, and (vi) the corti-costerone+ 12.5mgkg�1 d�1 GTS (CORT+GTSL)group. The administered doses were based on previousreports that GTS and fluoxetine could effectively producephysiological and behavioral effects on rodents. Cortico-sterone (Sigma, 20mgkg�1) was administered subcutane-ously once daily for 22 d to induce the depression model.The control group received subcutaneous administrationof the vehicle. After corticosterone injection, GTS orfluoxetine was administered intragastrically once dailyto the respective groups, whereas the control and CORTgroups received distilled water.After treatment for 20 and 21 d, the depression-like

behavior was observed via a forced swimming test(FST) and a tail suspension test (TST), respectively. At6 h after the last dose of corticosterone, all mice werequickly sacrificed to obtain venous blood samples forserum corticosterone assays and hippocampal tissuesfor Western blot analysis. The methods of behavioraltests, corticosterone assays, and Western blot analysiswere similar to our previous report (Wang et al., 2011).The primary antibodies included the rabbit monoclonalanti-GSK-3β (1:1000, Cell signal, 9315S), the rabbitmonoclonal anti-phospho-GSK-3β (Ser9) (1:1000, Cellsignal, 9323S), and the mouse monoclonal anti-β-actin(1:2000, Sigma, A1978). The secondary antibodiesincluded the horseradish peroxidase-conjugated goat

Copyright © 2013 John Wiley & Sons, Ltd.

anti-mouse IgG (1:4000, GeneScript) and the goatanti-rabbit IgG (1:4000, GeneScript).

Statistical analysis. Data were expressed as the mean ±SEM, and analyzed using the Statistical Package forSocial Sciences computer program (version 13.0). Thestatistical significance of the results was determinedusing one-way ANOVA, followed by Tukey’s post hoctests. The significance level was set at p≤ 0.05 for allstatistical comparisons.

RESULTS

GTSH and GTSM relieved corticosterone-induceddepression-like behavior

As shown in Fig. 1, the immobility time differed signifi-cantly among the groups in both FST (F5, 66 = 33.359,p< 0.01) and TST (F5, 66 = 21.861, p< 0.01). The posthoc test revealed that the corticosterone injectionsinduced a marked increase in the immobility timeduring FST and TST, as compared with the controlgroup (p< 0.01). The aforementioned immobility timewas reversed by simultaneous treatment with fluoxetine(p< 0.01 vs. the CORT group; p= 0.830 and p= 0.988,respectively, vs. the control group), GTSH (p< 0.05and p< 0.01, respectively, vs. the CORT group; p< 0.01vs. the control group) or GTSM (p< 0.01 vs. the CORTgroup; p< 0.01 and p= 0.111, respectively, vs. the con-trol group). Fluoxetine performed better than all threedoses of GTS at reversing depression-like behavior,and the differences are statistically significant (p< 0.05).

GTS had no significant effect on normalizing theincreased serum corticosterone level

At 6h after the last dose of corticosterone, the serum corti-costerone level in the CORT group (125.76±12.16μgL�1)increased significantly compared with the control group(32.12±4.86μgL�1) (p< 0.01). The fluoxetine treatmentmarkedly reversed the increased serum corticosterone level(p< 0.01) while all three doses of GTS treatments had nostatistically significant effect (data not shown).

GTS raised the down-regulated levels of hippocampalGSK-3β inhibitory phosphorylation in thecorticosterone-induced mouse depression model

As seen in Fig. 2A, the hippocampal protein level ofGSK-3β increased in the CORT group than that in thecontrol group (p< 0.01). Compared with the CORTgroup, neither GTS nor fluoxetine influenced GSK-3βexpression. The post hoc test showed that therewere no significant differences between groups of theCORT, the fluoxetine-treated and the GTS-treated(F4, 55 = 1.409, p= 0.243). The ratio of phospho-GSK-3β(Ser9) and GSK-3β was statistically different amonggroups (F5, 66 = 37.507, p< 0.01; Fig. 2B). It drasticallydeclined in the CORT group compared with the control(p< 0.01), CORT+GTSH (p< 0.01), CORT+GTSM

Phytother. Res. 28: 1102–1106 (2014)

Figure 1. Effects of fluoxetine or different doses of ginseng total saponins on the depression-like behavior in the forced swimming test (A)and tail suspension test (B) in the corticosterone-induced mouse depression model. ## p < 0.01 vs. the control group; *p < 0.05 vs. theCORT group, **p < 0.01 vs. the CORT group; & p < 0.05 vs. the CORT + FLU group, && p < 0.01 vs. the CORT + FLU group.

1104 L. CHEN ET AL.

(p< 0.01), or CORT+GTSL (p< 0.05) group. However,no significant difference was seen between the CORT+FLU and the CORT group.

DISCUSSION

In this study, depression-like behavior was induced byrepeated corticosterone injections. This method has beenreported to mimic the behavioral and neurochemicalalterations associated with depression, and regarded as aconvenient and reliable depression model (Zhao et al.,2008). The present study demonstrated that adverseeffects of corticosterone on depression-like behavior canat least be partially removed by GTS treatment (25 and50mgkg�1 d�1), considering the ameliorated immobilitytime in FST and TST. Previous studies have revealed thatginsenosides exhibited antidepressant effects in normalmice or in CMS models (Dang et al., 2009; Jiang et al.,2012), and GTS (at 25, 50, or 100mgkg�1 d�1, for 7 days)did not statistically affect locomotor activity in normalmice (Dang et al., 2009). The present study furtherconfirms the antidepressant activity of GTS on the corti-costerone-induced mouse depression model.Results from the serum corticosterone assays showed

that fluoxetine, a classical antidepressant applied in this ex-periment as a positive control, could effectively depress the

Copyright © 2013 John Wiley & Sons, Ltd.

increased level of serum corticosterone while GTS had nosuch effect. Therefore, themechanism of GTSmay involvemodulating the central nervous system targets instead ofthe peripheral anti-glucocorticoids. These results aresimilar to our previous study on water-based ginsengextracts (Wang et al., 2011), but are inconsistent with otherstudies on ginsenosides that employed theCMSdepressionmodel (Jiang et al., 2012). This discrepancy may be causedby several parameters in the experimental design, includ-ing model, saponin, and dosage differences.

Our results of GSK-3β and phospho-GSK-3β (Ser9)on the corticosterone-induced mouse depression modelare consistent with previous studies on CMS rats anddepressive patients (Silva et al., 2008; Oh et al., 2010),which further confirms that insufficient GSK-3β inhibi-tion is a risk factor for developing depression. In thepresent study, all doses of GTS (12.5, 25, and 50mgkg�1

d�1) significantly reversed the down-regulated inhibi-tory phosphorylation of GSK-3β in the depressionmodel, thereby suggesting that GTS may inhibit GSK-3β. To the best of our knowledge, this is the first studyto examine hippocampal GSK-3β level and activity inthis depression model, as well as to reveal the effect ofGTS on GSK-3β. In addition, our results showed thatfluoxetine altered neither hippocampal GSK-3β levelnor its activity in this depression model. These findingsindicate that GTS and fluoxetine may share differentantidepressant mechanisms.

Phytother. Res. 28: 1102–1106 (2014)

Figure 2. Hippocampal GSK-3β and phospho-GSK-3β (Ser9) protein levels were determined by Western blot analysis. The values of GSK-3β(A) and phospho-GSK-3β (Ser9) (B) were normalized against the amount of β-actin and GSK-3β, respectively. ## p<0.01 vs. the controlgroup; *p<0.05 vs. the CORT group, **p <0.01 vs. the CORT group.

1105GINSENG TOTAL SAPONINS ON INHIBITORY PHOSPHORYLATION OF GSK-3β

In summary, this study confirms the antidepressant activi-ties of certain doses GTS in the corticosterone-inducedmouse model of depression and implies hippocampal GSK-3β inhibition may be one of the underlying mechanisms.

Acknowledgments

The study was financially supported by the National Natural ScienceFoundation of China (81303246), the Jiangsu Provincial Natural

Copyright © 2013 John Wiley & Sons, Ltd.

Science Foundation of China (BK2011815), the Natural Science Founda-tion for Colleges and Universities in Jiangsu Province (12KJB360008),‘QingLan’ project of the Jiangsu Provincial FrameworkTeacher SupportScheme, and a project funded by the Priority Academic ProgramDevel-opment of Jiangsu Higher Education Institutions.

Conflict of Interest

The authors have declared that there is no conflict of interest.

REFERENCES

China Pharmacopoeia, Volume 1. 2010. China Medical SciencePress: Beijing; 367.

Dang H, Chen Y, Liu X et al. 2009. Antidepressant effects of ginsengtotal saponins in the forced swimming test and chronic mildstress models of depression. Prog Neuropsychopharmacol BiolPsychiatry 33: 1417–1424.

Jiang B, Xiong Z, Yang J et al. 2012. Antidepressant-like effects ofginsenoside Rg1 are due to activation of the BDNF signalingpathway and neurogenesis in the hippocampus. Br JPharmacol 166: 1872–1887.

O’Brien WT, Harper AD, Jove F et al. 2004. Glycogen synthasekinase-3beta haploinsufficiency mimics the behavioral andmolecular effects of lithium. J Neurosci 24: 6791–6798.

Phytother. Res. 28: 1102–1106 (2014)

1106 L. CHEN ET AL.

Oh DH, Park YC, Kim SH. 2010. Increased glycogen synthasekinase-3beta mRNA level in the hippocampus of patients withmajor depression: a study using the stanley neuropathology con-sortium integrative database. Psychiatry Investig 7: 202–207.

Silva R, Mesquita AR, Bessa J et al. 2008. Lithium blocks stress-induced changes in depressive-like behavior and hippocampalcell fate: the role of glycogen-synthase-kinase-3beta. Neurosci152: 656–669.

Copyright © 2013 John Wiley & Sons, Ltd.

Wang Z, Dai J, Chen L, Huang Y, Zhao Y. 2011. Preventive actionof Panax ginseng roots in hypercortisolism-induced Impair-ment of hippocampal neurons in male C57BL/6N mice.Phytother Res 25: 1242–1245.

Zhao Y, Ma R, Shen J, Su H, Xing D, Du L. 2008. A mouse modelof depression induced by repeated corticosterone injections.Eur J Pharmacol 581: 113–120.

Phytother. Res. 28: 1102–1106 (2014)