Morinda citrifolia fruit reduces stress-induced impairment of cognitive function accompanied by vasculature improvement in mice

Physiology & Behavior 101 (2010) 211–217

Contents lists available at ScienceDirect

Physiology & Behavior

j ourna l homepage: www.e lsev ie r.com/ locate /phb

Morinda citrifolia fruit reduces stress-induced impairment of cognitive functionaccompanied by vasculature improvement in mice

Junko Muto a, Lee Hosung a, Akemi Uwaya b, Fumiyuki Isami b, Makoto Ohno a, Toshio Mikami c,⁎a Nippon Sport Science University Graduate School of Health and Sport Science, Japanb Tahitian Noni Inc., Japanc Department of Health and Sports Science, Nippon Medical School, Japan

⁎ Corresponding author. Tel./fax: +81 44 733 3719.E-mail address: [email protected] (T. Mikami).

0031-9384/$ – see front matter © 2010 Published by Edoi:10.1016/j.physbeh.2010.04.014

a b s t r a c t

Article history:Received 24 March 2010Received in revised form 7 April 2010Accepted 14 April 2010

Keywords:Morinda citrifoliaVitamin EHippocampusStressMemoryAngiogenesis

The purpose of this study was to investigate effects of Morinda citrifolia fruit juice, which is locally calledNoni, on stress-induced impairment of cognitive function. Male ICR mice were divided into four groups:Control (C mice), Restraint stress (RS mice), Restraint+Noni (Noni mice), and Restraint+vitamin E (VEmice). The RS, Noni, and VE mice were subjected to 8 h of chronic restraint stress (CRS) 6 days a week for6 weeks. During this period, the Noni and VE mice were given a diet supplemented with either Noni orvitamin E, respectively. At Week 5, the mice were subjected to the Morris water maze (MWM) test tomeasure cognitive function. At Week 7, mouse brains were isolated for immunohistochemical analysis withBrdU or CD31 antibody to assess the proliferation of new cells and blood vessel density in the dentate gyrusof the hippocampus. The time taken to reach the platform in the MWM test was shorter in the Noni micethan in the RS mice on Day 16. Malondialdehyde (MDA ) level of the Noni mice was significantly higher thanthat of the C mice; however no difference was found in MDA levels between the VE and C mice. Blood vesselarea was significantly lower in the R and VE mice than in the C mice; no difference was found between the Cand Noni mice. These findings suggest that the administration of Noni fruit juice protects brains from stress-induced impairment of cognitive function and that this protective effect may be related to improvement instress-induced decreases in blood vessel density in the hippocampal dentate gyrus.

lsevier Inc.

© 2010 Published by Elsevier Inc.

1. Introduction

Aging leads to deterioration of brain functions such as learning andmemory, and is accelerated by chronic stress, especially psychologicalstress [1]. Chronic immobilization stress impairs significantly spatialmemory in the Morris water maze (MWM) test and attenuateshippocampal long-term potentiation (LTP) in rodents [2]. It was reportedthat rats restrained for 21 days showedsignificantly longer escape latencythan control rats [3].

Stress-induced impairment of learning and memory is closelyrelated to the suppression of hippocampal neurogenesis. Chronicrestraint stress (CRS) leads to impairment of cognitive function andconcomitant suppression of neurogenesis in the hippocampal dentategyrus [4]. In addition, impairment of spatial memory was found to benegatively correlatedwith hippocampal neurogenesis [5]. Hippocampalneurogenesis is therefore regarded as a dominant factor in cognitivefunction.

One of the reasons that chronic stress suppresses hippocampalneurogenesis is that it increases oxidative stress. Repeated restraint

stress results in an increase in thiobarbituric acid reactive substance(TBARS) and glutathione peroxidase activity in rats [6]. We have shownalso an increase in Malondialdehyde (MDA) in the brains of miceexposed toCRS [4]. VitaminE is shown to improve cognitive impairmentin streptozotocin-induced diabetes via its antioxidative effect [7], and isshown also to improve aging-caused cognitive deficit through neuro-protecting effect and an antioxidant efficacy [8]. Research has focusedincreasingly on whether ingestion of antioxidants, especially plant-derived antioxidants, leads to reduction of oxidative stress andimprovement in impaired cognitive function. Van Praag et al. [9]showed that a plant-derived flavanol, epicatechin, enhanced retentionof spatial memory in combination with wheel-running exercise.Administration of strawberry extract offered protection against spatialmemory deficits induced by whole-body irradiation [10]. In addition,GliSODin, a superoxide dismutase (SOD) extracted from melons andcombined with gliadin, prevents stress-induced impairment of cogni-tive function and maintains neurogenesis in the hippocampus [11].

Another factor related to hippocampal neurogenesis is angiogen-esis in the brain. Vascular endothelial growth factor (VEGF) is aprotein that promotes angiogenesis through the tyrosine kinasereceptors on endothelial cells [12]. Several studies have indicated thatVEGF acts as a neurotrophic factor [13,14] and produces neurogeniceffects on neuronal progenitors [13,15].

Fig. 1. Experimental schedule. Mice were subjected to restraint stress and experimentsas illustrated above. MWM, and BrdU indicate Morris water maze test, and injection ofBrdU (5-bromo-2′-deoxyuridine), respectively.

212 J. Muto et al. / Physiology & Behavior 101 (2010) 211–217

Morinda citrifolia is a plant indigenous to the Pacific Islands, knownlocally as Noni, and is used by the Polynesians in folk medicine. Thejuice extracted from Noni fruit possesses antioxidative activity.Kamiya et al. [16] showed that methanol extracts and ethyl acetate-soluble phases from Noni fruit inhibited copper-induced LDL (Lowdensity lipoprotein) oxidation. Wang et al. [17] showed that extractsfrom Noni-induced dose-dependent inhibition of the oxidant sub-stances produced by lipid hydroperoxide or superoxide anion radicals.However, there have been no reports concerning the effects of Nonion stress-induced impairment of cognitive function, neurogenesis, orangiogenesis.

In this study, we investigated whether supplementation with Nonijuice could attenuate stress-induced impairment of cognitive func-tion, and then we investigated whether Noni-induced improvementof cognitive function was related to neurogenesis or angiogenesis. Thepresent findings suggest that administration of Noni juice attenuatesstress-induced impairment of cognitive function and that thisinhibitory effect is possibly attributable to Noni-induced improve-ment of angiogenesis in the hippocampal dentate gyrus.

2. Materials and methods

2.1. Physical restraint stress and diet

This study was approved by the Animal Care and Use Committee ofNipponMedical School.We obtained 40male ICRmice (Clea Japan, Inc.)aged 6 weeks and weighing 30–32 g. The mice were randomly dividedinto four groups consisting of 10 mice each: Control mice (C mice),Restraint-stressed mice (RS mice), Restraint-stressed mice supplemen-ted with Noni (Noni mice), and Restraint-stressed mice supplementedwith Vitamin E mice (VE mice). Each mouse with restraint stress wasplaced in a 3×3×7.5 cm stainless-steel cage, which completelyrestricted their movement, but allowed them to drink water ad libitum[4,11]. Immobilization stress was given 8 h per day (0900–1700 h) for6 days each week. Each immobilized mouse was housed individually ina small 10×10×10 cm compartment of a multicompartment cage forthe remaining time to reduce physical activities. Unstressed mice werehoused in standard-sized cages consisting five mice per cage, and theywere handled daily without stress.

The C and RS mice were fed with a standard animal powdered diet(Oriental Yeast Co., Ltd, Japan), which was mixed with hot water,kneaded, cut, and dried to make hard pellets. The Noni and VE micewere fed with this diet, supplemented with Noni Juice® and vitamin E,respectively. The vitamin E diet was prepared by supplementing thestandard powdered diet with alpha-tocopherol acetate at 80 mg per100 g of powdered diet to maintain an alpha-tocopherol intake of70 mg/kg/day according to Li et al. [18]. Tap water was given to the C,RS, and VE mice. The Noni diet was prepared as follows. Noni Juice®

(provided by Tahitian Noni Inc.) was centrifuged at 3000 rpm for5 min. The supernatant was diluted twenty-fold with tap water andfed to the Noni mice as drinking water. The precipitate waslyophilized, added to the standard powdered diet, and used to makepellets as described above. The volume of Noni Juice® ingested viafood and drink was adjusted to 10 ml/kg/day according to the methodof Ma et al. [19]. All the mice were allowed to eat and drink ad libitum.

Schedules for providing restraint stress and performing experi-ments are schematically illustrated in Fig. 1.

2.2. Spatial learning and memory

After 4 weeks of CRS, the spatial memory of the mice wasdetermined using the MWM test according to the method of Morriset al. with several modifications [20]. Briefly, the mice were trainedwith four trials per day for 5 days. A circular pool with a diameter of115 cm was filled with water 1.5 cm above a plastic platform to hideit. The water was made opaque with white nontoxic paint, and the

water temperature was set at 24 °C. A mouse was released into thepool facing the pool wall from four different starting points, randomlyselected each day. The time taken to reach the platform (escapelatency) was recorded for every trial. Each trial lasted either until themouse had found the platform or for a maximum of 60 s. In each trial,the mice were allowed to stay on the platform for 20 s at the end ofeach trial. Twenty-four hours after the last training session, a probetrial was performed, after removing the platform, to evaluatememory. The mice were released into the pool at the oppositeposition of the quadrant where the platformwas located, and the timespent in the target quadrant was recorded for 60 s. To determine long-term retention (memory), the MWM test was performed again on the15th and 16th days after the first day of the MWM test in accordancewith the previous study [21,22]. During the MWM test, the mice weresubjected to restraint stress for 4 h per day.

2.3. Sample collection

The day following completion of the MWM test, the mice wereanesthetized with pentobarbital and transcardially perfused with60 ml of saline via the left ventricle. Their brains were carefullyisolated and the hemispheres separated. The left hemisphere wasfixed in 4% paraformaldehyde in 0.1 M phosphate-buffered saline(PBS, 137 mM NaCl, 8.10 mM Na2HPO4, 2.68 mM KCl, 1.47 mMKH2PO4, pH7.4) overnight at room temperature. After being washedthree times with PBS, the brain was cut rostrally at bregma−1.30 mm, caudally at bregma −5.80 mm, and ventrally at 4.5 mm.The areas were serially sectioned rostrocaudally using a Leicavibratome (VT 1000 S, Leica Microsystems, Germany) at 40 µm andimmersed free-floating in PBS. Ninety-six well plates were used topreserve the order of the sessions in PBS at 4 °C. The right hemispherewas divided into the hippocampus, cerebral cortex, hypothalamus,and cerebellum. These samples were quickly frozen with liquidnitrogen and stored at −80 °C until analysis.

2.4. BrdU immunohistochemistry

To label the newly generated progenitor cells with 5-bromo-2′-deoxyuridine (BrdU), intraperitoneal injection of BrdU at a dose of50 mg/kg was performed on 5 consecutive days (Days 36–40) duringWeek 6.

BrdU-positive cells were identified immunohistochemically usingan M.O.M. immunodetection kit (Vector Laboratory, USA). Briefly, thesections were reacted with 3% hydrogen peroxide in methanol for30 min to block endogenous peroxidase. After being washedwith PBS,the sections were incubated with 2 M HCl at 37 °C for 30 min. After asecond washing with PBS, the sections were incubated with M.O.M.mouse IgG blocking reagent for 1 h. The sections were then washedonce more with PBS and incubated with monoclonal anti-BrdUantibody (1:200, BD Pharmingen, USA) in M.O.M. diluent (0.1 PBS; pH

213J. Muto et al. / Physiology & Behavior 101 (2010) 211–217

7.4, 0.5% Triton X-100, 8% protein concentrate), while gently shaking,for two nights at 4 °C. After a further washing with PBS, the sectionswere incubated with biotinylated anti-mouse IgG in M.O.M. diluent(1:250) for 2 h at room temperature. The sections were then washedagain with PBS and incubated with avidin–biotin–horseradishcomplex for 2 h at room temperature. Finally, the sections wereincubated with Vectastain ABC reagent (Vector Laboratories) for 1 hand developed using 0.67 mg/ml of 3,3′-diaminobendizine (DAB).

2.5. CD31 immunohistochemistry

To investigate vascularization, CD31 immunohistochemistry wasperformed. After incubating the sections with 3% hydrogen peroxidefor 30 min and normal rabbit serum for 1 h, they were incubated withrat monoclonal anti-mouse CD31 antibody (1:50, PD Pharmingen,USA) while gently shaking for two nights at 4 °C. After being washedwith PBS, the sections were incubated with goat anti-rabbitbiotinylated IgG (Vector Laboratories, USA, 1:100) for 1 h at roomtemperature. Then, after being again washed with PBS, the sectionswere treated in the same way as for the BrdU detection.

2.6. Quantification of BrdU-positive cells and blood vessel area

The sections reacted with BrdU or CD31 antibody were mounted,dehydrated, and coverslipped using Permount mounting medium.BrdU-positive cells were counted in the subgranular zone (SGZ) of thedentate gyrus using a Leica DM3000 microscope (Leica, Germany)with a 40× objective. To calculate the proportion of the surface areacovered with microvasculature, CD31-immunolabeled sections werephotographed using a Leica DFC290 digital camera connected to aLeica DM3000 microscope (10× objective). The photos were thentransferred to a Windows XP computer. The area covered with blood

Fig. 2. Body weight, food consumption, and water intake. Body weights of RS and Noni micelower in the RS, Noni and VE mice; however, no statistical analysis was carried out. Data ar

vessels and the total area of the hilus in the dentate gyrus was thencalculated using Image-J software. Vessel density was expressed as apercentage of the surface area covered with blood vessels in the totalarea of the hilus of the dentate gyrus.

2.7. ELISA for VEGF

The level of VEGF in the cerebral cortex was assayed with anenzyme-linked immunosorbent assay (ELISA) (Quantikine®, R&DSystems, Inc., USA) according to the manufacturer's instructions. Weused the hypothalamus for VEGF analysis because the hippocampusand cerebral cortex had already been used for other analyses.

2.8. MDA determination

To evaluate lipid peroxidation in the cerebral cortex, malondial-dehyde (MDA) was determined using a Bioxytech® MDA-586™ kit(OxisResearch, USA), according to the manufacturer's instructions.Briefly, the cerebral cortex was homogenized with PBS containing 1%butylated hydroxytoluene (BHT), and the homogenate was centri-fuged at 15,000 rpm for 10 min. The obtained supernatant was addedto N-methyl-2-phenylindole and incubated at 45 °C for 1 h. Aftercentrifugation, the absorbance of the supernatant was determined at586 nm using an absorption spectrometer.

3. Statistical analysis

All values are shown as the mean±standard error of the mean(SEM). Differences between groups were analyzed using one- or two-wayANOVA.When statistical differenceswere found, Fisher's PLSD posthoc test was performed. Statistical significance was accepted as pb0.05.

All the experiments were examined in blinded fashion.

were significantly lower than that in C mice. Food consumption and water intake weree means±SEM. (C: n=11, RS and Noni: n=10, VE: n=9).

Fig. 3. a. Escape latency of MWM test. On Day 5, C mice reached an invisible platform ina significantly shorter time than those in the other mice. On Day 16, escape latency wassignificantly shorter in the Noni mice than in RSmice. Day 5: * C versus RS, Noni, and VE,Day 16: # Noni versus RS, pb0.05. Data are means±SEM. (C: n=11, RS and Noni:n=10, VE: n=9). b. Probe test using MWM test. The Noni mice showed a significantlylonger stay in the platform quadrant compared with the RS mice (*:pb0.05) Data aremeans±SEM. (C: n=11, RS and Noni: n=10, VE: n=9).

214 J. Muto et al. / Physiology & Behavior 101 (2010) 211–217

4. Results

4.1. Weight and diet consumption

Chronic immobilization and feeding in the six-divided cage led tolower values of body weight (Fig. 2). Notably, the Noni mice showedsignificantly lower body weight than the C mice from week 2 to theend of the experimental period (pb0.05), while the RS mice alsoshowed significantly lower body weight than the C mice at Week 6(pb0.05). Although, body weight of the VE mice showed a lowertendency than that of the Cmice, there were no significant differencesbetween the C and VE mice. The stressed mice showed lowerconsumption of the diet and drink than the C mice; however, wecould not statistically analyze diet and drink consumption, sinceinformation regarding individual diet and drink consumption for the Cmice was not obtained due to gregarious raising. The decreases inbody weight suggest that the experimental conditions used in thisstudy were stressful.

4.2. Learning and memory

To examine whether CRS influences cognitive performance, wetested learning and memory using the MWM test. As for learning, theC mice showed a shortened escape latency day after day during the5 days of training. The RS, Noni, and VE mice were similar in theirtimes to escape by Day 4. On Day 5, the C mice showed a significantlyshorter latency than the other mice (pb0.05) (Fig. 3a).

To examine memory, the mice were subjected to a probe test onDay 6 and re-examination of escape latency on Days 15 and 16. In theprobe trial, the Noni mice showed a significantly longer stay in theplatform quadrant compared with the RSmice (pb0.05) (Fig. 3b). TheVE mice showed no significant differences between the other three;however, they displayed a similar beneficial effect to that fromNoni inthe probe test. In addition, the Noni mice showed a significantlyshorter latency than the RSmice on Day 16. No difference between theVE and C mice was observed, which is consistent with data ofNakajima et al. [11]. These findings suggested that administration ofNoni attenuated the impairment of cognitive function in the stressedmice.

4.3. Hippocampal neurogenesis

To examine whether the beneficial effects of Noni on learning andmemory are related to hippocampal neurogenesis, we measured thenumber of BrdU-positive cells in the hippocampal dentate gyrus.Contrary to our expectations, there were no significant differences inthe number of BrdU-positive cells among the four groups of mice(Data is not shown.). In the present study, since the last injection ofBrdU was performed 5 days before the dissection, the number ofBrdU-positive cells reflected the proliferation of new cells in the SGZ.Therefore, in the present study, CRS had no effect on cell proliferationin the SGZ of the hippocampal dentate gyrus.

4.4. Blood vessel density in the dentate gyrus

To examine whether Noni affects the vasculature in the dentategyrus, we measured the blood vessel density in the hilus of thedentate gyrus by immunohistochemical staining with CD31 antibody.The percentage of the area covered with blood vessels in the hilus wassignificantly lower in the RS and VE mice than in the C mice (pb0.05);and, there were no significant differences between the C and Nonimice (Fig. 4a). Although there was no significant difference betweenthe RS and Noni mice, the obtained results are considered to show thepossibility that Noni administration attenuated the reduction of bloodvessel density in the hilus of the dentate gyrus of stressedmice. Fig. 4bshows representative pictures of each group of mice.

4.5. Lipid peroxide in the cerebral cortex

To examine whether chronic stress increases lipid peroxidation inthe brain, MDA was determined in the cerebral cortex. The MDA levelin the cerebral cortex was significantly higher in the RS and Noni micethan in the C mice (pb0.05); however, the VE mice showed nosignificant increase compared with the C mice (Fig. 5). This findingsuggests that Noni administration does not prevent induction ofoxidative stress in the brains of stressed mice.

4.6. VEGF in the hypothalamus

To examine whether chronic stress influences VEGF content in thebrain, VEGF was measured using an ELISA kit. The VEGF content in thehypothalamus was significantly lower in the RS, Noni, and VE micethan in the C mice (pb0.05) (Fig. 6); however, there were nosignificant differences among the RS, Noni, and VE mice. This finding

Fig. 4. a. Percentage of vessel area stained with CD31. The percentage of DG area covered with blood vessels in the RS mice was significantly decreased than that in the control;however, there was no significant difference between the control and the Noni mice. * C versus RS and VE, *:pb0.05. Data are means±SEM. (C: n=11, RS and Noni: n=10, VE:n=9). b. Typical picture of CD31 staining of the DG. These are typical DG sections of the hippocampus stained with CD31. The vascular endothelium appears red.

215J. Muto et al. / Physiology & Behavior 101 (2010) 211–217

suggests that Noni administration does not prevent reduction of VEGFexpression in the brains of stressed mice.

5. Discussion

To investigate whether the juice of the M. citrifolia fruit has anyeffect on stress-induced impairment of learning and memory, themice were exposed to CRS, and memory was determined using theMWM test after administration of Noni juice. CRS seems to result instress because bodyweights in the stressedmicewere lower than thatof the C mice; however, not all the body weights of the stressedanimals were significantly lower than the control and the stress in thisstudy is considered to be slightly low. The reasons for it might includethat we selected 8-h restraint in accordance with our recent research

(not published); however, restraint time might not have been longenough to achieve sufficient stress. Restraint duration was 10 h in theresearch of Nagata et al. [4] and 12 h for Nakajima et al. [11]. Amongthe three groups of stressed mice, Noni mice had significantly lowerbody weight from Week 2 to Week 7 than the C mice. Based on theirlower body weight profile, Noni mice appear to undergo stress. Wehad estimated food intake per body weight to be 6 g/40 g, and waterintake to be 8 ml/40 g. In this research, percentages of food (g) andwater intake (ml) per body weight (g) in comparison with theestimated values were found to be 80% and 76% in the stressed mice.Total amounts of Noni and vitamin E actually administered werelower than estimated, and this might affect the effects of Noni andvitamin E. There are several studies showing that Noni has a beneficialeffect on diabetes. Horsfall et al. gave Noni juice in combination with

Fig. 5. MDA levels in the cerebral cortex. MDA levels in Noni mice were significantlyhigher than those of C or VE mice. The MDA levels in Noni mice did not significantlydiffer from those of the RS mice, and were significantly higher than in C mice. * C versusRS; # C versus Noni; & VE versus RS, pb0.05. Data are means±SEM. (C: n=11, RS andNoni: n=10, VE: n=9).

216 J. Muto et al. / Physiology & Behavior 101 (2010) 211–217

insulin to SD rats after induction of diabetes. Under these conditions,fasting blood sugar was found to be 8.0 mmol/L, which was lowerthan with insulin administration alone at 12.9 mmol/ml [23]. Anotherstudy reports Noni to be an agonist of the P2Y receptor, which is a G-Protein-Coupled receptor on beta cells, that triggers increasedsecretion of insulin by stimulating cAMP production [24]. Thesefindings suggest that Noni improves glucose metabolism, whichmight also affect lipid metabolism. Therefore, we supposed that Nonimight decrease body weight in Noni mice by inhibition of fataccumulation via the increased consumption of carbohydrate takenfrom food.

CRS led to suppression of spatial memory based on Day 5 values inthe MWM test and probe test, and Noni administration attenuatedthis suppression. Immunohistochemistry with BrdU antibodyrevealed no significant differences in the proliferation of new cellsin the dentate gyrus among each group of mice. However, staining of

Fig. 6. VEGF levels in the hypothalamus. VEGF levels in C mice were significantly higherthan those in the other mice. * C versus RS, Noni, and VE, respectively, pb0.05. Data aremeans±SEM. (C: n=11, RS and Noni: n=10, VE: n=9).

vascular endothelial cells using CD31 antibody revealed blood vesseldensity in the hilus of the dentate gyrus to be reduced by CRS, whereasadministration of Noni counteracted the decrease in blood vesseldensity. These findings suggest that administration of Noni attenuatepossibly stress-induced impairment of cognitive function, whichmight be attributed to improvement in blood vessel density causedby Noni administration.

The present findings show that CRS impairs cognitive function,which is in agreement with previous findings [4,11], although ourresult appears to reflect less stress. Starting points on Day 1 differedamong the 4 groups. Stressed mice (RS, Noni and VE) had longerescape latency than the C mice, possibly due to restraint stress. Thereason that escape latency on Day 2 of RS, Noni and VE decreased sorapidly could be that the values on Day 1 for RS, Noni and VE wererelatively long due to CRS. On Day 16, the escape latency of the RS,Noni, and VEmice were shorter than that on Day 15 except for C mice.This shortening tendency is consistent with previous researches[21,22]. However, on Day 16, the escape latency of the C mice becamesignificantly longer than that of the Noni mice. The rational for itremains unclear.

Contrary to our expectations, the VE mice showed no improvementin learning and memory, which might be due to lower amounts ofvitamin E than that of previous studies. The amounts of VE supple-mentation in previous researches were 100 mg/kg for 7 weeks [7] and200 mg/kg for 96 weeks [8], and in our research, it was 70 mg/kgfor 4 weeks. In addition to this, actual intake of vitamin E was 80%(56 mg/kg).

The stressed mice showed shorter escape latency than expected,and no significant difference was observed between the C and RS inthe probe test. The reasons for it being a weak stress might include thefollowing. 1) 8-h restraint might not have been long enough toachieve sufficient stress as mentioned above. 2) Exercise was notconcomitantly introduced. Antioxidative plant extracts are oftenstudied in combination with exercise to examine their effects oncognitive function [9].

A decrease in neurogenesis is thought to be one of the reasons thatchronic stress impairs cognitive function. Neurogenesis is divided intotwo stages: proliferation and survival of new cells. In this study, therewere no significant differences in cell proliferation among the groups.Previous studies have reported the effects of chronic stress onproliferation and survival. Pham et al. [25] showed that 3 weeks(6 h/day) of CRS resulted in decreases in both proliferation andsurvival. Luo et al. also showed that CRS lowered cell proliferation inthe hippocampus [26]. On the other hand, Torner et al. [27] showedthat 21 days of CRS (4 h of daily immobilization) led to a significantdecrease in hippocampal cell survival but not cell proliferation. Wetherefore assume that chronic stress decreases cell survival; however,whether chronic stress decreases cell proliferation is unclear.Regrettably, we did not determine the survival of new cells. Furtherstudy is needed to clarify whether the Noni-induced improvement incognitive function is attributable to an improvement in cell survival.

The most important factor that improved cognitive function afterNoni administration was considered to be the recovery of blood vesseldensity in the hilus of the dentate gyrus (Fig. 4ab). The presentfindings showed that chronic stress decreased the proportion of thesurface area covered with vasculature in the hilus, which is consistentwith a previous study [28], and disagreed with another finding thatshowed no change in blood vessel density after chronic unpredictablestress (CUS) [29]. This contradictionmight be caused by differences inthe stress conditions. We stressed mice by daily severe restriction ofphysical activity. There is no apparent difference between Noni andRS, but in the Noni mice, suppression of vessel areas, which wasobserved in the RS and VE mice, appears to be attenuated. We regardthis as meaningful. We used whole juice, not an extract, to model thereal situation of drinking Noni in daily life. This may be one reason forthe somewhat fuzzy results. Since the relation between blood vessel

217J. Muto et al. / Physiology & Behavior 101 (2010) 211–217

density and cognition function has been studied [9,28], this could beevidence of a relation, though it may be a weak effect.

VEGF is an important signaling protein involved in vasculogenesisand angiogenesis. Since Noni administration suppressed stress-induced decrease in blood vessel density, we hypothesized that itwould increase VEGF content in the brain. However, all stressed miceshowed a significantly lower VEGF content in the brain comparedwith the controls (Fig. 6), so we were unable to find any effect of Nonion VEGF. Another pathway by which Noni administrationmight causeimprovements in blood vessel density is via the catechins contained inNoni juice. Epicatechins, which are plant-derived flavanols, causesignificant vasculature increases in the dentate gyrus [9]. M. citrifoliafruit extracts contain high levels of catechins, at 53.7±5.7 mg/g [30].Noni administration might therefore improve the stress-inducedimpairment of blood vessel density via the catechins contained inNoni juice.

Oxidative stress has been suggested to cause the stress-inducedimpairment of neurogenesis. Chronic stress results in the presence oflipid peroxide in the brain [4,11]. We also found significant increasesin cerebral MDA in the RS and Nonimice in comparisonwith the C andVEmice (Fig. 5), so no inhibitory effect of Noni on lipid peroxidation inthe brain was observed. Kamiya et al. [16] reported that extracts fromM. citrifolia fruits possessed in vitro antioxidant capacity as assessedby LDL oxidation. On the other hand, Harada et al. [31] reported thatfree intake of 10% Noni juice for 7 days had a protective effect onneuronal damage after ischemic stress, while 3% Noni juice did notshow any such effect. In the present study, Noni juice was centrifuged,and the supernatant was diluted 20-fold, resulting in approximately5% Noni juice, which was administered to the mice. The precipitate ofcentrifuged Noni juice was added to the solid diet and administered tothe mice. We assumed there to be more antioxidant substances in thesupernatant of Noni juice than in the precipitate. Therefore, thevolume of Noni administered to the mice might not have beensufficient to prevent lipid peroxidation in the brains of stressed mice.Further studies are needed to compare the antioxidant contentbetween the supernatant and precipitate of Noni juice and clarify thispoint.

In summary, to investigate whether stress-induced reductions incognitive function were improved by the administration of Noni juice,we investigated cognitive function, neurogenesis, blood vesseldensity, VEGF content, and lipid peroxide levels in chronicallyrestraint stressed mice fed with a diet and drink containing Nonijuice. Noni administration was found to attenuate the stress-inducedimpairment of cognitive function. This beneficial effect may be partlymediated by the improvement in angiogenesis induced by Noni.

Acknowledgement

The authors would like to thankMr. Yoshiyuki Iyama for providinginstructions on calculating blood vessel density.

References

[1] Chida Y, Sudo N, Mori J, Kubo C. Social isolation stress impairs passive avoidancelearning in senescence-accelerated mouse (SAM). Brain Res 2006;1067(1):201–8.

[2] Radecki DT, Brown LM, Martinez J, Teyler TJ. BDNF protects against stress-inducedimpairments in spatial learning and memory and LTP. Hippocampus 2005;15(2):246–53.

[3] Abidin I, Yargicoglu P, Agar A, Gumuslu S, Aydin S, Ozturk O, et al. The effect ofchronic restraint stress on spatial learning and memory: relation to oxidant stress.Int J Neurosci 2004;114(5):683–99.

[4] Nagata K, Nakashima-Kamimura N, Mikami T, Ohsawa I, Ohta S. Consumption ofmolecular hydrogen prevents the stress-induced impairments in hippocampus-dependent learning tasks during chronic physical restraint in mice. Neuropsy-chopharmacology 2008;34(2):501–8.

[5] Nilsson M, Perfilieva E, Johansson U, Orwar O, Eriksson PS. Enriched environmentincreases neurogenesis in the adult rat dentate gyrus and improves spatialmemory. J Neurobiol 1999;39(4):569–78.

[6] Fontella FU, Siqueira IR, Vasconcellos AP, Tabajara AS, Netto CA, Dalmaz C.Repeated restraint stress induces oxidative damage in rat hippocampus.Neurochem Res 2005;30(1):105–11.

[7] Tuzcu M, Baydas G. Effect of melatonin and vitamin E on diabetes-inducedlearning and memory impairment in rats. Eur J Pharmacol 2006;537(1–3):106–10.

[8] Takatsu H, Owada K, Abe K, Nakano M, Urano S. Effect of vitamin E on learning andmemory deficit in aged rats. J Nutr Sci Vitaminol (Tokyo) 2009;55(5):389–93.

[9] van Praag H, Lucero MJ, Yeo GW, Stecker K, Heivand N, Zhao C, et al. Plant-derivedflavanol (−)epicatechin enhances angiogenesis and retention of spatial memoryin mice. J Neurosci 2007;27(22):5869–78.

[10] Shukitt-Hale B, Carey AN, Jenkins D, Rabin BM, Joseph JA. Beneficial effects of fruitextracts on neuronal function and behavior in a rodentmodel of accelerated aging.Neurobiol Aging 2007;28(8):1187–94.

[11] Nakajima S, Ohsawa I, Nagata K, Ohta S, Ohno M, Ijichi T, et al. Oralsupplementation with melon superoxide dismutase extract promotes antioxidantdefences in the brain and prevents stress-induced impairment of spatial memory.Behav Brain Res 2009;200(1):15–21.

[12] Cao L, Jiao X, Zuzga DS, Liu Y, Fong DM, Young D, et al. VEGF links hippocampalactivity with neurogenesis, learning and memory. Nat Genet 2004;36(8):827–35.

[13] Jin KL, Mao XO, Greenberg DA. Vascular endothelial growth factor: directneuroprotective effect in in vitro ischemia. Proc Natl Acad Sci U S A 2000;97(18):10242–7.

[14] Matsuzaki H, Tamatani M, Yamaguchi A, Namikawa K, Kiyama H, Vitek MP, et al.Vascular endothelial growth factor rescues hippocampal neurons from glutamate-induced toxicity: signal transduction cascades. FASEB J 2001;15(7):1218–20.

[15] Warner-Schmidt JL, Duman RS. VEGF is an essential mediator of the neurogenicand behavioral actions of antidepressants. Proc Natl Acad Sci U S A 2007;104(11):4647–52.

[16] Kamiya K, Tanaka Y, Endang H, Umar M, Satake T. Chemical constituents ofMorinda citrifolia fruits inhibit copper-induced low-density lipoprotein oxidation.J Agric Food Chem 2004;52(19):5843–8.

[17] WangMY, Su C. Cancer preventive effect ofMorinda citrifolia (Noni). Ann N Y AcadSci 2001;952:161–8.

[18] Li RK, Sole MJ, Mickle DA, Schimmer J, Goldstein D. Vitamin E and oxidative stressin the heart of the cardiomyopathic syrian hamster. Free Radic Biol Med 1998;24(2):252–8.

[19] Ma DL,West BJ, Su CX, Gao JH, Liu TZ, Liu YW. Evaluation of the ergogenic potentialof noni juice. Phytother Res 2007;21(11):1100–1.

[20] Morris RG, Garrud P, Rawlins JN, O'Keefe J. Place navigation impaired in rats withhippocampal lesions. Nature 1982;297(5868):681–3.

[21] Carro E, Trejo JL, Busiguina S, Torres-Aleman I. Circulating insulin-like growthfactor I mediates the protective effects of physical exercise against brain insults ofdifferent etiology and anatomy. J Neurosci 2001;21(15):5678–84.

[22] Carro E, Trejo JL, Spuch C, Bohl D, Heard JM, Torres-Aleman I. Blockade of theinsulin-like growth factor I receptor in the choroid plexus originates Alzheimer's-like neuropathology in rodents: new cues into the human disease? NeurobiolAging 2006;27(11):1618–31.

[23] Horsfall AU, Olabiyi O, Aiyegbusi A, Noronha CC, Okanlawon AO.Morinda citrifoliafruit juice augments insulin action in Sprague–Dawley rats with experimentallyinduced diabetes. Nig Q J Hosp Med 2008;18(3):162–5.

[24] Su CX, Jensen JC, Zhou BN. Morinda citrifolia (noni): its effect on insulin secretionby G-protein-coupled receptor systems; 2005.

[25] Pham K, Nacher J, Hof PR, McEwen BS. Repeated restraint stress suppressesneurogenesis and induces biphasic PSA-NCAM expression in the adult rat dentategyrus. Eur J Neurosci 2003;17(4):879–86.

[26] Luo C, Xu H, Li XM. Quetiapine reverses the suppression of hippocampalneurogenesis caused by repeated restraint stress. Brain Res 2005;1063(1):32–9.

[27] Torner L, Karg S, Blume A, Kandasamy M, Kuhn HG, Winkler J, et al. Prolactinprevents chronic stress-induced decrease of adult hippocampal neurogenesis andpromotes neuronal fate. J Neurosci 2009;29(6):1826–33.

[28] Czeh B, Abumaria N, Rygula R, Fuchs E. Quantitative changes in hippocampalmicrovasculature of chronically stressed rats: no effect of fluoxetine treatment.Hippocampus 2009.

[29] Heine VM, Zareno J, Maslam S, Joels M, Lucassen PJ. Chronic stress in the adultdentate gyrus reduces cell proliferation near the vasculature and VEGF and Flk-1protein expression. Eur J Neurosci 2005;21(5):1304–14.

[30] Pak-Dek MS, Abdul-Hamid A, Osman A, Soh CS. Inhibitory effect of Morindacitrifolia L on lipoprotein lipase activity. J Food Sci 2008;73(8):C595–8.

[31] Harada S, HamabeW, Kamiya K, Satake T, Tokuyama S. Protective effect ofMorindacitrifolia on the ischemic neuronal damage. Yakugaku Zasshi 2009;129(2):203–7.