Neuroscience Research 88 (2014) 58–66

Contents lists available at ScienceDirect

Neuroscience Research journal homepage: www.elsevier.com/locate/neures

Differential effect of arachidonic acid and docosahexaenoic acid on age-related decreases in hippocampal neurogenesis Hisanori Tokuda a , Masanori Kontani a , Hiroshi Kawashima a , Yoshinobu Kiso a , Hiroshi Shibata a , Noriko Osumi b,∗ a b

Institute for Health Care Science, Suntory Wellness Ltd., Osaka, Japan Department of Developmental Neuroscience, Tohoku University Graduate School of Medicine, Sendai, Japan

a r t i c l e

i n f o

Article history: Received 8 June 2014 Received in revised form 23 July 2014 Accepted 6 August 2014 Available online 19 August 2014 Keywords: Hippocampal neurogenesis Neural stem cell Arachidonic acid Docosahexaenoic acid Aging

a b s t r a c t Hippocampal neurogenesis affects learning and memory. We evaluated in rats effects of ingestion of arachidonic acid (ARA) and/or docosahexaenoic acid (DHA) on age-related decreases in proliferating neural stem/progenitor cells (NSPCs) or newborn neurons (NNs). Rats were fed with ARA- and/or DHA-containing diet from 2 to 18 months old and then sacrificed 1 day or 4 weeks after 5-bromo-2deoxyuridine (BrdU) injections at 2, 6 and 18 months. The numbers of NSPCs (SOX2+/BrdU+) and NNs (NeuN+/BrdU+) were determined immunohistochemically. The number of BrdU+ cells 1 day after BrdU injections decreased with age, but increased 65% after ARA ingestion compared to the control at 18 months. The SOX2+/BrdU+ cell ratio was unchanged by aging or ingestion of ARA or DHA. The number of NeuN+/BrdU+ cells 4 weeks after BrdU injections decreased with age, but increased 34% (yet not statistically significant) after DHA ingestion compared to the control at 18 months. These results indicate that ARA ingestion can ameliorate the age-related decrease in the number of NSPCs in rats. The functions of ARA and DHA in hippocampal neurogenesis appear to be different in aged rats; ARA may maintain an NSPC pool, whereas DHA may support NN production and/or survival. © 2014 Elsevier Ireland Ltd and the Japan Neuroscience Society. All rights reserved.

1. Introduction Arachidonic acid (ARA) and docosahexaenoic acid (DHA) are important components of brain phospholipids (PLs) that decrease with age (Söderberg et al., 1991; McGahon et al., 1997, 1999). Agerelated deficits in hippocampal functions, including learning and memory, have recently been reported to be ameliorated by ARA or DHA supplementation in aged rats (Okaichi et al., 2005; Kotani et al., 2003; Gamoh et al., 2001); however, the mechanisms involved remain unknown. Neurogenesis in the dentate gyrus (DG) of the hippocampus occurs throughout the lifetime of humans and rodents (Eriksson et al., 1998; van Praag et al., 2002) and is related to learning and memory in primates and rodents (Aizawa et al., 2009; Imayoshi et al., 2008). Newborn neurons (NNs) are generated from neural stem/progenitor cells (NSPCs) in the subgranular zone (SGZ) of the DG through differentiation and maturation. The numbers of both proliferating NSPCs and NNs in the DG decrease with age (Kuhn et al., 1996; Abrous et al., 2005). These findings suggest that

∗ Corresponding author. Tel.: +81 22 717 8201. E-mail address: [email protected] (N. Osumi).

maintaining the number of proliferating NSPCs will ameriorate the age-related decrease in NNs (Rao et al., 2005) and hippocampal function. Several previous reports have addressed the relationship between neurogenesis and ingestion of ARA and DHA. Maekawa et al. showed that ingestion of ARA during postnatal days 2–31 increases the number of NSPCs in neonatal rats (Maekawa et al., 2009). Administration of DHA for 7 weeks increases the number of NNs in aged n-3 polyunsaturated fatty acid (PUFA)-deficient rats in the third generation of diet-deficient breeding (Kawakita et al., 2006). In addition, DHA promotes neuronal differentiation from embryonic cerebral cortex in a neurosphere assay (Kawakita et al., 2006; Katakura et al., 2009). Recently, Sakayori et al., using a neurosphere assay, showed that ARA and DHA have different effects on maintenance and differentiation of NSPCs (Sakayori et al., 2011). However, the effects of ARA and DHA ingestion on the agerelated decline in neurogenesis in normal-aged rats remain unclear, although ARA and DHA may have different roles in neurogenesis. ARA and DHA are biosynthesized from linoleic acid (LA) and ␣-linolenic acid (ALA), respectively, in the body (Fig. 1). The conversion reactions from LA and ALA are catalyzed by the same enzymes common to the n-6 and n-3 series, and they compete with each other. The n-6/n-3 ratio is an important factor for brain functions

http://dx.doi.org/10.1016/j.neures.2014.08.002 0168-0102/© 2014 Elsevier Ireland Ltd and the Japan Neuroscience Society. All rights reserved.

H. Tokuda et al. / Neuroscience Research 88 (2014) 58–66

COOH

COOH

ALA

LA

COOH

COOH

ARA

DHA

n-6 series

n-3 series

Fig. 1. Biosynthetic pathways from linoleic acid and ␣-linolenic acid to ARA and DHA, respectively. LA: linoleic acid; ALA: ␣-linolenic acid; ARA: arachidonic acid; DHA: docosahexaenoic acid.

because this ratio in the diet affects cognitive functions in humans and other vertebrates (Loef and Walach, 2013). Therefore, to properly evaluate the effects of ARA and DHA on hippocampal neurogenesis, the effects of variation in the n-6/n-3 ratio should be excluded from the experimental conditions as much as possible. The purpose of this study was to evaluate the effects of ingestion of ARA and/or DHA on age-related decreases in NSPCs or NNs in rats. We used diets containing ARA and/or DHA with the same n-6/n-3 ratio. We observed age-related decreases in ARA and DHA in the hippocampus and in the number of hippocampal NSPCs, and these phenomena were ameliorated by ARA and/or DHA ingestion.

2. Materials and methods 2.1. Animals Male Fisher 344 rats, aged 7 weeks, were obtained from Oriental Bioservice Inc. (Kyoto, Japan). For 1 week prior to experimentation, the animals were fed a commercial diet (CRF-1, Oriental Yeast Co. Ltd., Tokyo, Japan) and tap water ad libitum and were housed at 25 ± 1 ◦ C and 60 ± 5% humidity under a 12-h light-dark cycle. Four (at 2 and 6 months) and 4–6 rats (at 18 months) in each group were examined. Experiments were approved by the Animal Care and Use Committee of Suntory Holdings Limited according to the guidelines for animal experiments prescribed by the Science Council of Japan on June 1, 2006.

2.2. Diets We prepared four different diets by altering the fatty acid composition of AIN-76A: the control, ARA(+), DHA(+), and ARA(+)DHA(+) diets (Table 1). The groups of rats were fed each diet as follows: Control diets (Cont), ARA(+) diets (ARA), DHA(+) diets (DHA), and ARA(+)DHA(+) diets (A + D). The diets were prepared to have almost the same n-6 PUFA: n-3 PUFA and PUFA: monounsaturated fatty acid: saturated fatty acid ratios (2:1 and 1:1:1, respectively). The diets were stored at 4 ◦ C and protected from light to prevent oxidation and denaturation. The rats were fed with these special diets from the age of 8 weeks until the end of the experimental period.

59

Table 1 Fatty acid composition of the modified AIN-76A diets used in this experiment. FA

Control

ARA(+)

DHA(+)

ARA(+)DHA(+)

Palmitic acid Stearic acid Oleic acid Linoleic acid ␣-Linolenic acid Arachidonic acid Eicosapentaenoic acid Docosahexaenoic acid Others Total polyunsaturated FA Total monounsaturated FA Total saturated FA n-6/n-3

26.7 4.4 32.1 22.3 11.1 0.0 0.0 0.0 3.5 33.7 32.6 31.3 2.0

26.0 4.8 29.9 18.2 11.6 4.0 0.0 0.0 5.7 35.0 30.4 32.1 2.0

26.0 4.1 31.6 22.2 7.2 0.0 0.0 3.9 5.1 33.7 32.2 30.4 2.0

26.1 4.5 29.6 17.6 7.2 3.9 0.0 3.9 7.2 34.1 30.3 31.9 2.1

FA: fatty acid. FA composition (%) of the total FA in each diet.

2.3. BrdU labeling assay Two-, 6-, and 18-month-old rats fed with the control, ARA(+), DHA(+), or ARA(+)DHA(+) diet received intraperitoneal injections of BrdU (Sigma., St. Louis, MO, USA) dissolved in 0.01 M phosphatebuffered saline (PBS) at 50 mg/kg body weight, four times a day, and were sacrificed 1 day or 4 weeks after the injections. 2.4. Immunohistochemistry To identify markers related to cell proliferation of NSPCs and NNs, rats were deeply anesthetized with sodium pentobarbital and then transcardially perfused with 4% paraformaldehyde and 0.5% picric acid in PBS. The brains were removed and further immersionfixed in the same fixative at 4 ◦ C for 24 h. Coronal sections of 14 ␮m thickness were prepared using a cryostat (CM1900, Leica, Wetzlar, Germany) and mounted onto slide glasses. The sections were washed with PBS containing 0.1% Triton X100 (PBS-Tx). For immunostaining of BrdU+ cells on day 1 after BrdU injections, the sections were boiled in 0.01 M citric acid for 2 min, incubated in 2 N HCl for 10 min at 37 ◦ C, and then neutralized with 0.1 M borate buffer (pH = 9.0) for 1 min, washed in PBS-Tx, and blocked in 2% normal goat serum. The sections were incubated at 4 ◦ C overnight with mouse IgG anti-BrdU antibodies (1:100) (Becton Dickinson, Franklin Lakes, NJ, USA). To visualize antigens, the sections were incubated with goat anti-mouse IgG Alexa594-conjugated antibodies (1:250) (Invitrogen, Life Technologies Corp., Carlsbad, CA, USA) at room temperature for 1 h. For nuclear counterstaining, the sections were incubated with 20 ␮g/mL 4 ,6-diamidino-2phenylindole (DAPI) (Invitrogen, Life Technologies Corp., Carlsbad, CA, USA) in PBS-Tx at room temperature for 5 min, and covered by coverslips with PermaFluorTM (Thermo Fisher Scientific Inc., Waltham, MA, USA). For double-immunostaining of SOX2+/BrdU+ cells on day 1 after BrdU injections, the sections were incubated with mouse monoclonal IgG2A anti-SOX2 antibodies (1:200) (R&D Systems, Minneapolis, MN, USA) at room temperature for 2 h. The sections were then incubated with goat anti-mouse IgG Alexa488conjugated antibodies (1:250) (Invitrogen, Life Technologies Corp., Carlsbad, CA, USA) at room temperature for 1 h. The sequential procedure for detecting BrdU was similar to that described above, except that rat anti-BrdU antibodies (1:125) (Oxford Biotechnology Ltd., Oxford, UK) and goat anti-rat IgG Alexa594-conjugated antibodies (1:250) (Invitrogen, Life Technologies Corp., Carlsbad, CA, USA) were used. For double-immunostaining of NeuN+/BrdU+ cells 4 weeks after BrdU injections, the sections were incubated with mouse monoclonal IgG anti-NeuN antibodies (1:200) (Chemicon, Billerica, MA, USA) at room temperature for 2 h. The sections were then incubated with goat anti-mouse IgG Alexa488-conjugated antibodies (1:250) (Invitrogen, Life Technologies Corp., Carlsbad,

60

H. Tokuda et al. / Neuroscience Research 88 (2014) 58–66

CA, USA) at room temperature for 1 h. The sequential procedure was the same as that described in double-immunostaining for SOX2+/BrdU+ cells. 2.5. Quantification For quantitative analysis of BrdU+ cells 1 day after the final BrdU injection, samples were obtained from the entire subgranular zone (SGZ, the two-cell thick region from the inner margin of the granule cell layer (GCL) (Rao et al., 2005)), in the dentate gyrus (DG) of the hippocampus of one hemisphere of each experimental animal. Every sixth section was used for counting, and the total number of positive cells was obtained by multiplying the counted value by 6. To evaluate SOX2+/BrdU+ ratio in each aged rat, we considered the age-related decrease of the number of BrdU+ cells. Thus, 6, 10, and 48 equivalent sections of the SGZ from animals at 2, 6, and 18 months were used to calculate the SOX2+/BrdU+ ratio. To determine the number of NeuN+/BrdU+ cells 4 weeks after BrdU injections, samples were obtained from the GCL and SGZ of the entire DG in the hippocampus of one hemisphere of each experimental animal. Every tenth section was used for counting, and the total number of positive cells was obtained by multiplying the counted value by 10. Fluorescent signals were detected using a confocal laser-scanning microscope (FV1000, Olympus, Tokyo, Japan) for counting SOX2+/BrdU+ cells and NeuN+/BrdU+ cells or a fluorescent microscope (DM6000, Leica, Wetzlar, Germany) for counting BrdU+ cells. Cell numbers were obtained by a researcher who was blind to sample groups.

rats. Tukey honestly significant difference (HSD) tests were used to compare each group at each time point. The relationships between the number of cells and the PUFA content of RBC PLs were evaluated by calculating Pearson’s correlation coefficients. A p-value less than 0.05 was considered statistically significant. 3. Results 3.1. ARA and DHA in hippocampal and red blood cell (RBC) PLs We found no significant differences in body weight changes during the experimental period in the groups of rats we studied. To assess age-related changes in the hippocampus, we analyzed the fatty acid composition of hippocampal PLs at 3 and 19 months of age (Table S1). In the control group, the ARA content slightly decreased from 13.5% at 3 months to 13.2% at 19 months (Fig. 2A). At 19 months, the ARA content was decreased in the group ingesting DHA, but remained similar to the 3-month control level in the ARA and ARA + DHA (A + D) groups. The DHA content also decreased from 15.0% at 3 months to 14.2% at 19 months in the control and ARA groups, but was similar to the level at 3 months in the DHA and A + D groups (Fig. 2B). The ARA and DHA contents of RBC PLs changed in response to dietary ARA and DHA (data not shown). The ARA contents of hippocampal and RBC PLs were positively correlated (r = 0.869, p < 0.001; Fig. 2C), and the same was true for the DHA contents (r = 0.745, p < 0.001; Fig. 2D). Supplementary Table S1 related to this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.neures.2014.08.002.

2.6. Fatty acid analysis Red blood cells and one hippocampus were prepared from the same animals used for the BrdU labeling assay. The blood samples for RBCs were collected from the hearts before the perfusion of the fixative. The brains were removed after the perfusion and further immersion-fixed at 4 ◦ C for 24 h. Although there is information that the amount of PUFA in the 3-day formalin-fixed brain is slightly different from that in the unfixed brain (Yamamoto et al., 1987), we had confirmed by comparing unfixed and fixed brains that the fixation condition in the present study has little effect on the hippocampal fatty acid composition before the study (data not shown). Hippocampal lipids were extracted and purified according to the method reported previously (Folch et al., 1957). RBC lipids were extracted with H2 O–isopropanol–chloroform (1:11:7, v/v/v) and then re-extracted and purified according to the same procedure (Folch et al., 1957). These purified lipids were separated into neutral lipids and PLs using thin-layer chromatography with silica gel 60 (Merck, Darmstadt, Germany). The solvent system consisted of hexane–diethyl ether (7:3, v/v). Fatty acid residues in the PL fractions were analyzed (Sakuradani et al., 1999). In brief, each fraction with an additional internal standard (pentadecanoic acid) was incubated in methanolic HCl at 50 ◦ C for 3 h for transmethylation of fatty acid residues to fatty acid methyl esters. Fatty acid methyl esters were extracted with n-hexane and analyzed with capillary gas–liquid chromatography using pentadecanoic acid as an internal standard. Analytical conditions were as follows: (1) apparatus: Agilent 6890 (Agilent, DE, USA); (2) column, SP2330 (30 m × 0.32 mm × 0.2 ␮m, SUPELCO., PA, USA); (3) carrier: He (30 cm/s); (4) column temperature: 180 ◦ C (2 min) followed by +2 ◦ C/min to 220 ◦ C. 2.7. Statistics Data are presented as the mean ± standard error (SE). Data were analyzed with one-way analysis of variance (ANOVA) followed by Dunnett’s tests for the comparison of age-related changes in control

3.2. Effects of ARA and/or DHA ingestion on the age-related decrease in hippocampal NSPC number Next we assessed the age-related change in the number of hippocampal NSPCs using a BrdU labeling assay (Fig. 3A). The number of BrdU+ cells in the SGZ of the hippocampal DG decreased with age, from 5630 ± 569 at 2 months to 3537 ± 158 at 6 months and 618 ± 120 at 18 months (Fig. 3B and C). We further evaluated the effect of dietary PUFAs. The number of BrdU+ cells in the ARA group was significantly larger compared to the number in the control group at 6 and 18 months (118% and 165%, respectively) (Fig. 3D). No significant increase in the number of BrdU+ cells was observed in the DHA and A + D groups, although a trend toward an increase was seen in the A + D rats at 18 months. At 18 months, the number of BrdU+ cells in the SGZ was positively correlated with the ARA content of RBC PLs (r = 0.643, p = 0.002; Fig. 3E), whereas no significant correlation was observed between the number of BrdU+ cells and the DHA content (r = -0.303, p = 0.182; Fig. 3F). To quantitate NSPCs, we determined the SOX2+/BrdU+ cell ratio because BrdU+ cells included not only NSPCs but also other proliferating cells. The percentage of SOX2+ cells compared to BrdU+ cells was unchanged from 2 to 18 months in the control group and similar tendency was detected in the other groups at 6 or 18 months (Table 2). Therefore, ingestion of ARA may positively relate with the maintenance and/or proliferation of NSPCs. 3.3. Effects of ARA and/or DHA ingestion on the age-related decrease in hippocampal NN number We next addressed the age-related change in the survival of hippocampal NNs. It is believed that hippocampal NNs become matured within 4 weeks and that a certain amount of NNs die during this process (Piatti et al., 2006; Kitamura et al., 2010). Using immunohistochemistry, cells that were positive for the neuron-specific marker NeuN and for BrdU were identified 4 weeks after BrdU injection (Fig. 4A). The number of NeuN+/BrdU+ cells

H. Tokuda et al. / Neuroscience Research 88 (2014) 58–66

61

Fig. 2. The effects of aging and ARA and DHA ingestion on the ARA and DHA content in RBCs and hippocampal PLs. (A and B) Contents of ARA and DHA in rats fed ARA and/or DHA diet (n = 4, Cont (3 months); n = 6, Cont (19 months); n = 5, ARA; n = 4, DHA; n = 5, A + D). Mean (SE). F(4, 19) = 15.1 p < 0.001, one-way ANOVA; p = 0.048 Cont (3 months) vs. Cont (19 months), p < 0.001 Cont (3 months) vs. DHA (19 months), Tukey HSD tests in (A). F(4, 19) = 6.05, p < 0.001, one-way ANOVA; p = 0.019 Cont (3 months) vs. Cont (19 months), p = 0.027 Cont (3 months) vs. ARA (19 months), Tukey HSD tests in (B). *p < 0.05 vs. Cont (3 months). (C) Correlations between the ARA contents of RBC PLs and hippocampal PLs. (D) Correlations between the DHA contents of RBC PLs and hippocampal PLs. Relationships were evaluated by calculating Pearson’s correlation coefficients.

in the hippocampal DG decreased with age, from 4810 ± 236 at 3 months to 1918 ± 139 at 7 months and 323 ± 29 at 19 months (Fig. 4B and C). The number of NeuN+/BrdU+ cells tended to increase with PUFA dietary treatment, but no significant difference was found at 7 or 19 months (Fig. 4D). The number of NeuN+/BrdU+ cells in the DHA group was 134% of that in the control group at 19 months. A positive correlation was found between the number of NeuN+/BrdU+ cells in the DG and the DHA content in RBC PLs (r = 0.527, p = 0.025; Fig. 4F), whereas no significant correlation with the ARA content was observed (r = -0.179, p = 0.477; Fig. 4E). Therefore, ingestion of DHA may positively be related with the survival of hippocampal NNs in aged rats.

4. Discussion ARA and DHA are important components of brain PLs and an age-related decrease in these fatty acids has been reported in the human or rodent hippocampus (Söderberg et al., 1991; McGahon et al., 1997, 1999). Results shown here are basically consistent

with those of previous studies, in which the concentrations of ARA and DHA in hippocampal PLs of 22-month-old rats are decreased and then improved by the ingestion of ARA and DHA for 8 weeks (McGahon et al., 1997, 1999). ARA and DHA are essential fatty acids that are synthesized in the body from dietary LA and ALA, respectively. This conversion capability decreases with age (Hrelia et al., 1989; Bordoni et al., 1988), and likely affects the amounts of ARA and DHA in the hippocampus of aged rats. In fact, our control group at 19 months failed to maintain previous levels of ARA and DHA in the hippocampus, although the contents of LA and ALA in the control diet were sufficient and higher than those in the ARA(+) and DHA(+) diets. These findings may suggest a significant possibility that direct ingestion of ARA and DHA may be more important than ingesting their precursors, LA and ALA, during aging. Hippocampal neurogenesis is related to learning and memory (Imayoshi et al., 2008), transition from hippocampus dependent to cortex dependent memory (Kitamura et al., 2009), and levels of anxiety (Santarelli et al., 2003), and is thought to be affected by two important factors, the number of NSPCs and their differentiation to mature neurons. NSPCs are important because they are regarded as

62

H. Tokuda et al. / Neuroscience Research 88 (2014) 58–66

Fig. 3. The effect of aging and ARA and DHA ingestion on the number of SGZ BrdU+ cells. (A) Experimental design for the BrdU assay 1 day after BrdU injection in PUFAadministered rats. (B and C) Age-related change in the number of SGZ BrdU+ cells in the Cont group. Blue, DAPI; Red, BrdU. Arrowheads indicate BrdU+ cells in SGZ. Scale bar = 100 ␮m. Mean (SE). F(2, 11) = 74.6, p < 0.001, one-way ANOVA; p = 0.002 Cont (2 months) vs. Cont (6 months), p < 0.001 Cont (2 months) vs. Cont (18 months), Dunnett’s tests. # p < 0.01 vs. 2 months, n = 4 (2 months); n = 4 (6 months); n = 6 (18 months). (D) The effect of ARA and DHA ingestion on the number of SGZ BrdU+ cells at 6 and 18 months. Mean (SE). F(3, 12) = 3.56, p = 0.048, one-way ANOVA for all groups at 6 months; p = 0.039 Cont vs. ARA. F(3, 17) = 3.59, p = 0.035, one-way ANOVA for all groups at 18 months; p = 0.047 Cont vs. ARA, Tukey HSD tests. *p < 0.05 vs. age-matched Cont, n = 4 in each group (6 months); n = 6, Cont (18 months); n = 6, ARA; n = 4, DHA; n = 5, A + D. (E and F) Correlations between the ARA and DHA content of RBC PLs and the number of SGZ BrdU+ cells at 18 months. Relationships were evaluated by calculating Pearson’s correlation coefficients. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

H. Tokuda et al. / Neuroscience Research 88 (2014) 58–66 Table 2 The effect of aging and ARA and DHA ingestion on the SOX2+/BrdU+ cell ratio (%).

2 mo 6 mo 18 mo

Cont

ARA

DHA

A+D

59.3 ± 1.7 59.0 ± 4.4 57.9 ± 3.0

– 71.3 ± 3.2 53.7 ± 3.0

– 65.7 ± 3.8 56.4 ± 5.7

– 57.9 ± 3.9 52.2 ± 3.7

Mean (SE). No significant difference was found in the Cont group. F(2, 11) = 0.056, p = 0.946, one-way ANOVA for the control groups at 2, 6 and 18 months, n = 4 (2 months); n = 4 (6 months); and n = 6 (18 months). No significant difference was found among the groups. F(3, 12) = 2.64, p = 0.097, one-way ANOVA for all groups at 6 months. F(3, 17) = 0.506, p = 0.683, one-way ANOVA for all groups at 18 months, n = 4 in each group (6 months); n = 6, Cont (18 months); n = 6, ARA; n = 4, DHA; n = 5, A + D.

the source of NNs. Previous studies have shown that the number of NSPCs decreases with age (Kuhn et al., 1996; Rao et al., 2005). In the present study, the number of BrdU+ cells, a marker of proliferating cells, in the SGZ 1 day after BrdU injections also decreased with age. The number of BrdU+ cells is considered to reflect the number of NSPCs, because the ratio of cells positive for SOX2, a marker for NSPCs (Steiner et al., 2006; Suh et al., 2007), to BrdU+ cells remained unchanged with aging. When the rats ingested ARA and/or DHA from 2 to 18 months of age, only ARA ingestion ameliorated the age-related decrease in the number of BrdU+ cells. This is consistent with our observation of a significant positive correlation between the number of BrdU+ cells and the ARA content in RBC PLs. In contrast, ingestion of DHA did not affect the number of BrdU+ cells, and no significant correlation was observed between the number of BrdU+ cells and the DHA content in RBC PLs. These results indicate that ARA may have a greater impact on the maintenance of the number of NSPCs in the hippocampus of aged rats than DHA. It is strange that the increase in the number of NSPCs was not statistically significant in the A + D group. It seems that DHA suppresses the effect of ARA on NSPCs, but this possibility may be unlikely because there was no negative correlation between DHA content and the number of NSPCs. The reason for this phenomenon is unclear at this point, and further studies are necessary to elucidate. Another important issue is the differentiation of NSPCs to mature neurons, which can be determined by the number of NNs. A previous study has clarified that ingestion of DHA for 7 weeks significantly increases the number of NNs by 1.6-fold in aged rats that are fed n-3 PUFA-deficient diets for three generations (Kawakita et al., 2006). In the present study, using normal wild-type rats, we also observed a positive correlation between the number of NeuN+/BrdU+ cells in the DG and the DHA content in RBC PLs. These results strongly suggests the importance of DHA in the differentiation of NSPCs to NNs. Consistently, the number of NeuN+/BrdU+ cells in the DHA group was 134% of that in the control group, although this increase was not statistically significant. On the other hand, ARA ingestion did not affect the number of NeuN+/BrdU+ cells. Such differential effects of DHA and ARA in hippocampal neurogenesis in aged rats have also been demonstrated in neurosphere assays using fetal cortical cells (Sakayori et al., 2011). All together, DHA has a greater impact on the differentiation of NSPCs to mature neurons in the hippocampus of aged rats than ARA. In other words, ARA may ameliorate the age-related decrease in the NSPC pool, whereas DHA may ameliorate the age-related decrease in NNs (Fig. 5). This is the first report indicating the difference between ARA and DHA on neurogenesis in aged rats. It is previously reported that hippocampal NSPCs became gliogenic with aging (van Praag et al., 2005; Okada et al., 2008), and that the effect of ARA and DHA on gliogenic NSPCs was distinctive in neurosphere assays (Sakayori et al., 2011). The effect of DHA on the neuronal differentiation from gliogenic NSPCs in the previous study is similar to that in the present study. However, there is a minor discrepancy in regard with the

63

maintenance of gliogenic NSPCs in the two studies. In the previous study, DHA, but not ARA, promoted the differentiation from gliogenic NSPCs to neurons and also increased the number of gliogenic NSPCs (Sakayori et al., 2011). This discrepancy in effects of ARA and DHA may be due to differences in the experimental design (in vitro or in vivo), the source of NSPCs (cerebral cortex or hippocampus) and the age of animal (embryonic or aged). The idea shown in Fig. 5 is also in agreement with previous findings regarding differential effect of ARA and DHA. ARA and some of their metabolites are involved in the proliferation of NSPCs. For example, mice deficient for diacylglycerol lipase have lower levels of ARA and 2-arachidonoyl-glycerol (2-AG), i.e., a metabolite of ARA and a ligand for cannabinoid receptor (CB) 1 and CB2, and show a significantly decreased number of BrdU+ cells in the hippocampal DG (Gao et al., 2010). In addition, agonists of these CB receptors promote the proliferation of NSPCs in a neurosphere assay (Molina-Holgado et al., 2007). Cyclooxygenase metabolites of ARA, such as prostaglandin E2 (Uchida et al., 2002) and prostaglandin J2 (Katura et al., 2010), also increase the proliferation of NSPCs. Therefore, ARA metabolites may have been involved in the effect of ARA ingestion by rats in the present study. In contrast, it seems that neuronal differentiation is more related to DHA. As mentioned above, DHA promotes the differentiation of NSPCs to neurons in neurosphere assays (Kawakita et al., 2006; Sakayori et al., 2011). Katakura et al. (2009) further show that DHA affects the expression of basic helix-loop-helix transcription factors such as Hes1, Mash1, Neurogenin1, and NeuroD, all of which control neural differentiation. Generally speaking, DHA seems to be crucial for the production of NNs. We could not exclude the possibility that ARA or DHA might maintain NSPCs or promote the neural differentiation via affecting fluidity of cell membrane in NSPCs and neurons because adequate membrane fluidity might be necessary for the cell division and the morphological change. Due to their multiple double bonds, ARA and DHA are considered to increase the fluidity in the cell membrane. Actually, ARA or DHA supplementation has improved the neural membrane fluidity in rodents (Fukaya et al., 2007; Suzuki et al., 1998). Taken together, it is currently unclear which of these candidate mechanisms may contribute to the differential actions of ARA and DHA. Further studies are necessary to clarify the mechanisms. Age-related decrease in neurogenesis is mainly due to decrease in the number of NSPCs. In our study, the number of NSPCs in the SGZ was >5000 at 2 months of age but reached to approximately 500–1000 (10–20% of the 2 months level) at 18 months of age in rats. This reduction rate with aging is similar to that in previous studies (Kuhn et al., 1996; Rao et al., 2005; Drapeau et al., 2003). In the present study, ARA and/or DHA ingestion increased the number of NSPCs by only several hundred, a smaller number compared with the large age-related decrease. However, an increase of several hundred NSPCs is considered meaningful because performance in the Morris water maze task by 20-month-old rats is well correlated with the number of hippocampal BrdU+ cells in the range of