Maternal dietary imbalance between omega-6 and omega-3 polyunsaturated fatty acids impairs neocortical development via epoxy metabolites

Abstract

Omega-6 (n-6) and omega-3 (n-3) polyunsaturated fatty acids (PUFAs) are essential nutrients. Although several studies have suggested that a balanced dietary n-6:n-3 ratio is essential for brain development, the underlying cellular and molecular mechanism is poorly understood. Here, we found that feeding pregnant mice an n-6 excess/n-3 deficient diet, which reflects modern human diets, impairsed neocortical neurogenesis in the offspring. This impaired neurodevelopment occurs through a precocious fate transition of neural stem cells from the neurogenic to gliogenic lineage. A comprehensive mediator lipidomics screen revealed key mediators, epoxy metabolites, which were confirmed functionally using a neurosphere assay. Importantly, although the offspring were raised on a well-balanced n-6:n-3 diet, they exhibited increased anxiety-related behavior in adulthood. These findings provide compelling evidence that excess maternal consumption of n-6 PUFAs combined with insufficient intake of n-3 PUFAs causes abnormal brain development that can have long-lasting effects on the offspring's mental state. Stem Cells 2016;34:470–482

Significance Statement

Balanced maternal intake of omega-6 and omega-3 polyunsaturated fatty acids is believed to be necessary for normal brain development; however, the precise neurodevelopmental outcome, the underlying mechanism, and neurodevelopmental impact on emotion in adulthood were poorly understood. Here we found that feeding pregnant mice an omega-6 rich/omega-3 poor diet impaired neocortical neurogenesis in the offspring. We also performed the first comprehensive quantification of lipid metabolites in the developing brain, and identified epoxy metabolites as key fate regulators of neural stem cells through neurosphere assays. We also found that these offspring exhibited more anxious behavior. This report underscores the dangers associated with the current dietary situation in many countries.

Introduction

Given the current global situation in which non-communicable disorders—which are often driven by dietary behavior—account for the vast majority of disability and early-onset morbidity and mortality, understanding how our diet affects our health and drives pathogenesis has become increasingly important. The “developmental origins of health and disease” hypothesis, which posits that nutrition in early life stages determines the risk of onset of various diseases in adulthood 1, 2, has been supported by many studies (for review, see McMillen et al. 3). Thus, evaluating the impact of food—in particular, food changes in modern society—on organogenesis is important from a public health perspective. Because neurogenesis in the brain is largely completed in the embryonic period 4, impaired neurogenesis due to poor maternal nutrition might lead to long-lasting functional defects in the offspring.

The mammalian brain contains extremely high numbers of neurons and glial cells, both of which arise from embryonic neural stem cells (NSCs). In early developmental stages, NSCs primarily give rise to neurons; in contrast, in late developmental stages, NSCs primarily give rise to astrocytes, with the peak occurring just following birth. Thus, a neurogenic-to-gliogenic fate transition occurs in NSCs, and this transition has been examined in mice 5. This transition can be studied easily in vitro, as cultured NSCs can undergo the fate transition. Moreover, most neurons and astrocytes that constitute the adult nervous system originated from NSCs, reflecting the important role of NSCs in brain development and function in adulthood.

Polyunsaturated fatty acids (PUFAs) are essential nutrients that are used to synthesize cellular structures and to produce biologically active substances. PUFAs are categorized primarily as omega-6 (n-6) and omega-3 (n-3) PUFAs; arachidonic acid (AA) and docosahexaenoic acid (DHA) are the principal n-6 and n-3 PUFAs, respectively, in the brain gray matter 6. Although AA and DHA are produced from linoleic acid (LA) and α-linolenic acid (ALA), respectively, these precursors are not synthesized de novo in mammals; thus, n-6 and n-3 PUFAs must be obtained from dietary sources. The proposed optimum dietary ratio of n-6:n-3 PUFAs is 1-4:1, which is considered to reflect competition between n-6 and n-3 PUFAs for synthesis, metabolism, transport, and incorporation into the cell membrane. However, in many developed and developing countries, the increased intake of seed oils (which are rich in LA) and/or the insufficient consumption of fish (which is rich in n-3 PUFAs) have led to a significant increase in this ratio, reaching as high as 25:1 7. Importantly, the intake of both AA and DHA is suggested to be necessary for normal brain development 8-12, and several government health agencies recommend consuming less amounts of n-6 PUFAs and more n-3 PUFAs (reviewed by Harris et al. 13); however, precisely how consuming an n-6 excess/n-3 deficient (n-6^ex/n-3^def) diet affects brain development remains poorly understood.

Here, we report that maternal consumption of an n-6^ex/n-3^def diet impairs formation of the neocortical neuronal layer in the developing offspring. These neurodevelopmental defects are due to a precocious neurogenic-to-gliogenic fate transition in NSCs. A comprehensive mediator lipidomics screen revealed that the amounts of n-6 and n-3 PUFA epoxy metabolites are significantly affected in the developing brain, and these metabolites actually regulate the fate transition. Furthermore, we examined the neurodevelopmental consequences of maternal intake of the n-6^ex/n-3^def diet on the offspring's brain function, and we found that the offspring's anxiety-related behaviors were increased, even though the offspring were raised on a nutritionally optimized diets. Our results raise important concerns regarding the increased consumption of an n-6^ex/n-3^def diet, particularly during pregnancy.

Materials and Methods

Animals

Wild-type C57BL/6N (WT) mice were obtained from Clea Japan (Tokyo, Japan, http://www.clea-japan.com/en/index.html). The WT and Fat-1 mice 14 were housed at the Tohoku University School of Medicine under a standard 12-hour light/12-hour dark schedule (lights on at 8:00 a.m.). Food and water were available ad libitum. E0.5 was defined as midday on the day in which a vaginal plug was observed. All animal experiments were performed in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and were approved by our university's committee for animal experimentation (MED#2013-114).

Diets

The mice were fed one of two semi-purified diets containing 16% energy from fat 9. In brief, the control diet contained 3.1% and 1.2% energy in the form of LA and ALA, respectively; the n-6^ex/n-3^def diet contained 11.8% and 0.3% energy in the form of LA and ALA, respectively. The percent energy from LA in the n-6^ex/n-3^def diet was within the range of most human diets 9. Virgin female mice were fed their respective diet beginning two weeks prior to mating and were maintained on their respective diet through gestation and the first 10 days of lactation (i.e., until their offspring reached P10). Thereafter, the maternal mice were fed the AIN-93G diet 15 (Oriental Yeast, Tokyo, Japan, http://www.oyc.co.jp/en/index.html). After weaning, the offspring were group-housed at two to three mice per cage and fed the AIN-93M diet 15 (Oriental Yeast). For additional information, see Supporting Information Figures S1 and S2.

Fatty Acid Analysis of Brain Lipid Fractions

Fatty acid analysis was performed as described previously 16. The procedural details are provided in the Supporting Information Materials and Methods.

Immunohistochemistry

Immunohistochemistry was performed as described in the Supporting Information Materials and Methods.

Neural Stem Cell Cultures

Neurospheres were generated from the embryonic forebrain as described previously 17, with minor modifications. The procedural details are provided in the Supporting Information Materials and Methods.

Immunocytochemistry

Immunocytochemistry was performed as described previously 18, with minor modifications. The procedural details are provided in the Supporting Information Materials and Methods.

Mediator Lipidomics

A liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based mediator lipidomics analysis was performed as described previously 19. The procedural details are provided in the Supporting Information Materials and Methods.

Real-Time RT-PCR

Total RNA was extracted from the neocortex of P10 pups using the RNeasy kit (Qiagen, Hilden, Germany, https://www.qiagen.com/jp/) in accordance with the manufacturer's instructions. Reverse transcription (RT) was performed using 2 μg of total RNA, oligo d(T)_12-18 primers (Invitrogen, Carlsbad, CA, https://www.thermofisher.com/jp/en/home/brands/invitrogen.html), and SuperScript III Reverse Transcriptase (Invitrogen). The resulting cDNA was subjected to real-time PCR in a Mastercycler ep realplex cycler (Eppendorf, Hamburg, Germany, https://www.eppendorf.com/JP-ja/) with FastStart Essential DNA Green Master (Roche, Indianapolis, IN, http://www.roche.com/index.htm). The relative amount of mRNA was calculated using Gapdh mRNA as an invariant control. Primer sequences are provided in the Supporting Information Materials and Methods.

Behavioral Tests

Mice at 13–15 weeks of age were handled on a regular basis and subjected to behavioral testing. Each test was performed in the light cycle on separate days. For the open field test, the apparatus consisted of a transparent floor (50 cm × 50 cm) with 30-cm-high black walls. The floor was divided into the center zone (the central 30 cm × 30 cm) and the peripheral zone (the other area). Each mouse was placed in the center of the open field and allowed to explore freely for 10 minutes. A video tracking system (ANY-maze; Stoelting, Wood Dale, IL, http://www.stoeltingco.com) was used to record the precise trajectory of each mouse. The time spent in the center zone, the number of entries into the center zone, and the total distance traveled were measured. For the elevated plus maze test, the apparatus consisted of a plus-shaped maze elevated 47 cm above the floor, containing two opposing open arms (30 cm × 5 cm each) and two arms of the same dimensions enclosed by 30-cm-high walls with an open top. In addition, a 1-cm-high clear Plexiglas edge surrounded the open arms to prevent falls. Each mouse was placed in the center of the maze facing the open arm and allowed to explore freely for 5 minutes. A video tracking system (ANY-maze; Stoelting) was used to record the precise trajectory of each mouse, and the times spent in the open and closed arms were measured.

Statistical Analysis

All statistical analyses were performed using Excel Statistics software (Social Survey Research Information, Tokyo, Japan, https://software.ssri.co.jp/index.html). Differences with p < 0.05 were considered to be statistically significant.

Results

Nutritional Imbalance of PUFAs Impairs Neocortical Development

To evaluate the consequences of consuming a diet with improperly balanced PUFA composition, we fed WT female mice either a control diet or an n-6^ex/n-3^def diet 9 (Supporting Information Fig. S1). We then measured the levels of fatty acids in the brains of the developing offspring. We found that the diet had no effect on the levels of saturated or monounsaturated fatty acids in the offspring at embryonic day 14.5 (E14.5) (Fig. 1A–1E); however, the offspring of the mice that were fed the n-6^ex/n-3^def diet had increased levels of n-6 PUFAs and decreased levels of n-3 PUFAs compared with the offspring of mice fed the control diet (Fig. 1F, 1J). In particular, both AA and docosapentaenoic acid (n-6) were increased (Fig. 1G–1I), whereas DHA was decreased (Fig. 1K–1M). We found the same PUFA profile in the brains of postnatal day 10 (P10) offspring (data not shown). These results suggest that the maternal intake of the n-6^ex/n-3^def diet causes an imbalance between n-6 and n-3 PUFAs in the developing offspring's brain, and this imbalance persists into the postnatal period.

Figure 1

Maternal consumption of the n-6^ex/n-3^def diet alters the relative levels of n-6 and n-3 polyunsaturated fatty acids (PUFAs) in the offspring's brain at E14.5. (A–E): Maternal diet does not affect the levels of total saturated fatty acids (A), palmitic acid (B), stearic acid (C), total monounsaturated fatty acids (D), or oleic acid (E). (F–I): Maternal intake of the n-6^ex/n-3^def diet increases the levels of total n-6 PUFAs (F), arachidonic acid (H), and docosapentaenoic acid (n-6) (I), but not the level of linoleic acid (G). (J–M): Conversely, maternal intake of the n-6^ex/n-3^def diet decreases the levels of total n-3 PUFAs (J) and docosahexaenoic acid (M), but not the levels of α-linolenic acid (K) or eicosapentaenoic acid (L). n = 5 animals per group. **, p < 0.01 (Tukey-Kramer's test). Levels of fatty acids are expressed as the percentage of total lipids. All data are expressed as the mean ± SEM. Abbreviations: AA, arachidonic acid; ALA, α-linolenic acid; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; EPA, eicosapentaenoic acid; LA, linoleic acid; MUFA, monounsaturated fatty acid; PA, palmitic acid; PUFA, polyunsaturated fatty acid; OA, oleic acid; SA, stearic acid; SFA, saturated fatty acid.

Next, we investigated the developmental consequences of this n-6:n-3 imbalance by examining neocortical development in the offspring. We found that the rostral neocortex was significantly thinner in the n-6^ex/n-3^def offspring (Fig. 2A). We then investigated neocortical structure further using a variety of molecular markers. We found no difference in the Pax6⁺ germinal layer (Fig. 2B), indicating that the total number of NSCs was not affected by the n-6:n-3 imbalance; in contrast, both the βIII tubulin⁺ immature neuronal layer and the Tbr1⁺ mature neuronal layer were thinner in the n-6^ex/n-3^def offspring (Fig. 2C, 2D). Taken together, these results demonstrate that the n-6:n-3 imbalance impairs development of the neuronal layer in the rostral neocortex of developing offspring. With respect to the middle neocortex, the maternal diet had no effect on total thickness, the Pax6⁺ layer, or the βIII tubulin⁺ layer (Fig. 2E–2G); however, the Tbr1⁺ layer was thinner in the n-6^ex/n-3^def offspring (Fig. 2H). Finally, the maternal diet had no significant effect on any of the layers in the caudal neocortex (Fig. 2I–2L). Thus, at E14.5, the caudal neocortex appears to be spared with respect to the effects of the n-6^ex/n-3^def maternal diet; this is likely due to the relatively slow development of the caudal neocortex 4, 20 (see also Fig. 6).

Figure 2

Maternal consumption of the n-6^ex/n-3^def diet reduces the thickness of the neuronal layers in neocortical regions at E14.5. Rostral (A–D), middle (E–H), and caudal (I–L) immunofluorescence (left panels), thickness, and cell density (right panels) in the indicated layers of E14.5 offspring of WT female mice fed the control diet or the n-6^ex/n-3^def diet. In the indicated layers, the cells were stained with DAPI (A, E, I), Pax6 (B, F, J), βIII tubulin (C, G, K), or Tbr1 (D, H, L). Rostral (A′–D′), middle (E′–H′), and caudal (I′–L′) thickness and cell density in the indicated layers of E14.5 offspring of female Fat-1 mice fed the n-6^ex/n-3^def diet. In the indicated layers, the cells were stained with DAPI (A′, E′, I′), Pax6 (B′, F′, J′), βIII tubulin (C′, G′, K′), or Tbr1 (D′, H′, L′). n = 5 cortices per WT group; n = 4 cortices (rostral and middle levels) or 5 cortices (caudal level) per Fat-1 group. *, p < 0.05; **, p < 0.01 (Tukey-Kramer's test). Scale bar = 100 μm. Thickness is expressed in μm, and density is expressed in cells per 5.0 × 10³ μm² (B, F, J, B′, F′, J′) or cells per 3.0 × 10³ μm² (D, H, L, D′, H′, L′). All summary data are expressed as the mean ± SEM. Abbreviation: DAPI, 4′,6-diamidino-2-phenylindole.

A traditional approach to modify the dietary n-6:n-3 ratio is to use diets with different oil sources; however, such different diets contain different amounts of other nutrients, including cholesterol and antioxidants, which are known to affect NSCs 21. To examine whether different nutrient contents among diets has a confounding effect on neocortical development, we re-evaluated the consequences of n-6:n-3 imbalance in combination with a genetic approach. Although WT mice cannot convert n-6 PUFAs to n-3 PUFAs, Fat-1 transgenic mice, which carry a transgene that encodes the C. elegans n-3 desaturase Fat-1, can convert n-6 PUFAs to n-3 PUFAs 14. Thus, Fat-1 mice can modify their n-6:n-3 ratio directly, regardless of diets. First, to characterize the fatty acid composition, we fed female Fat-1 mice either the control diet or the n-6^ex/n-3^def diet, then analyzed their offspring (Supporting Information Fig. S2A). We found that the brain of the offspring of the Fat-1 mice that were fed the n-6^ex/n-3^def diet had n-6 and n-3 PUFA levels that were similar to the offspring of WT mice that were fed the control diet; in contrast, the brain of the offspring of Fat-1 mice that were fed the control diet had decreased levels of n-6 PUFAs and increased levels of n-3 PUFAs (Supporting Information Fig. S3). Thus, the offspring of Fat-1 mice that were fed the n-6^ex/n-3^def diet can serve as an additional control for evaluating the effects of n-6:n-3 imbalance (Supporting Information Fig. S2B). Consistent with this notion, we found that the 4′,6-diamidino-2-phenylindole (DAPI)⁺, Pax6⁺, βIII tubulin⁺, and Tbr1⁺ layers was normal in the offspring of the Fat-1 mice that were fed the n-6^ex/n-3^def diet (Fig. 2A′–2L′). Moreover, we found no histological difference between the two control groups (i.e., between the offspring of the WT mice fed the control diet and the Fat-1 mice fed the n-6^ex/n-3^def diet; Fig. 2A′–2L′). These results suggest that the impaired development of the neuronal layer in WT mice fed the n-6^ex/n-3^def diet (Fig. 2A–2L) was caused by a change in the n-6:n-3 ratio, rather than by any confounding dietary factors.

Neurogenic-to-Gliogenic Fate Transition of NSCs Is Accelerated in the Developing n-6^ex/n-3^def Brain

We next investigated the mechanism by which feeding pregnant mice the n-6^ex/n-3^def diet impairs neocortical development in the offspring. We found no difference between the two dietary groups with respect to the number of Tbr2⁺ intermediate progenitor cells (IPCs) (Fig. 3A) or the number of active Caspase 3⁺ cells (Supporting Information Fig. S4). Thus, the abnormal neocortical development in the n-6^ex/n-3^def offspring was not due to impaired generation of IPCs or increased cell death.

Figure 3

Maternal consumption of the n-6^ex/n-3^def diet drives NSCs down the gliogenic lineage. (A, C, E, F): Immunostaining (left or upper panels) and summary graphs (right or lower panels) of the Tbr2⁺ layer (A) (n = 5 cortices per group), phosphohistone-3 (PH3)⁺ cells (C) (n = 7 cortices per group), and ALDH1L1⁺ cells (E, F) (n = 6 cortices per group) of offspring at the indicated developmental stage. In (A), thickness and density are expressed as micrometers and cells per 4.0 × 10³ μm², respectively. In (F), no ALDH1L1⁺ astrocyte is observed in the neocortex although there are ALDH1L1⁺ astrocytes in the lateral ganglionic eminence. (B): The percentage of βIII tubulin⁺ neurons is decreased in NSCs cultured from E14.5 n-6^ex/n-3^def embryos (n = 4 dishes per group). (D): The percentage of GFAP⁺ astrocytes is increased in NSCs cultured from E14.5 n-6^ex/n-3^def embryos (n = 4 dishes per group). *, p < 0.05; **, p < 0.01 (unpaired Student's t-test). Scale bars = (A, C, E) 50 μm; (B, D) 40 μm; (F) 100 μm. All summary data are expressed as the mean ± SEM. Abbreviations: AP, apical progenitor; GFAP, glial fibrillary acidic protein; IPC, intermediate progenitor cell.

Next, we performed an in vitro analysis to examine whether abnormal neuronal differentiation might underlie the impaired neocortical development observed in the n-6^ex/n-3^def offspring. Using the neurosphere assay, we found that the cells derived from n-6^ex/n-3^def NSCs contained fewer βIII tubulin⁺ neurons (Fig. 3B). Therefore, we speculated that the n-6:n-3 imbalance in the maternal diet leads to an inhibition of neuronal differentiation of NSCs in the offspring (i.e., NSCs underwent less asymmetrical cell division). However, when we histologically analyzed the neocortex, we found that the maternal diet had no effect on the number of Phosphohistone-3⁺ mitotic NSCs or IPCs (Fig. 3C), indicating that the rate of cell division was not affected. This finding led us to hypothesize that neurogenic NSCs are precociously fated to become gliogenic NSCs as a result of the n-6:n-3 imbalance condition in the offspring.

The most direct means to test this hypothesis would be to measure the number of gliogenic NSCs in the offspring; however, selective markers that can distinguish between neurogenic and gliogenic NSCs are not available. Therefore, we examined the in vitro differentiation of NSCs into astrocytes (using glial fibrillary acidic protein [GFAP], a well-established marker of astrocytes), thereby providing an indirect measure of the number of gliogenic NSCs in the brain. Consistent with our hypothesis, we found that the proportion of GFAP⁺ astrocytes was significantly higher among the cultured cells derived from n-6^ex/n-3^def NSCs (Fig. 3D). We further confirmed that the expression level of ALDH1L1, a marker for mature astrocytes 22, was actually increased in the E16.5 n-6^ex/n-3^def offspring (Fig. 3E), although no ALDH1L1⁺ astrocytes were detected in the neocortex at E14.5 (Fig. 3F). Thus, maternal consumption of the n-6^ex/n-3^def diet results in significantly more gliogenic NSCs and astrocytes in the offspring's developing neocortex.

Epoxy Metabolites Regulate the Neurogenic-to-Gliogenic Fate Transition of NSCs

Next, we determined which molecular mechanisms induce the precocious gliogenic fate shift in NSCs. Both n-6 and n-3 PUFAs can be metabolized into various lipid mediators with potent bioactive properties 19, 23, 24, and some of these mediators—including prostaglandins (PGs) and leukotrienes (LTs)—can regulate the proliferation and differentiation of NSCs 25-28; however, the profile of lipid mediators in the embryonic brain has not been studied previously. To identify which lipid mediators are altered in the n-6^ex/n-3^def offspring, we screened the lipid mediator profiles using LC-MS/MS 19, and we selected the mediators of AA, eicosapentaenoic acid, or DHA that exceeded 1 ng in either the control samples or the n-6^ex/n-3^def samples. We found seven lipid mediators that met this criterion (Fig. 4; Supporting Information Fig. S5). Surprisingly, six of these seven lipid mediators—four of which are epoxyeicosatrienoic acids (EETs) and two of which are epoxydocosapentaenoic acids (EpDPEs)—are epoxy metabolites of AA or DHA, and all six mediators are produced via the same metabolic pathway, namely cytochrome P450 monooxygenases (P450s) 29. The seventh mediator, 12-hydroxyeicosatetraenoic acid (12-HETE), is a hydroxy metabolite of AA; however, the level of 12-HETE was decreased in the n-6^ex/n-3^def samples (Fig. 4). Based on these results, we speculate that the maternal n-6^ex/n-3^def diet increases the levels of EETs and decreases the levels of EpDPEs in the offspring, and one or both of these effects promotes the gliogenic fate transition of NSCs in the developing offspring's brain.

Figure 4

Comprehensive quantification of AA- and docosahexaenoic acid (DHA)-derived mediators measured using LC-MS/MS and multiple reaction monitoring. AA- and DHA-derived mediators in the neocortex of E14.5 offspring in female mice fed either the control diet or the n-6^ex/n-3^def diet. n = 3 animals per group. Each lipid mediator is expressed as ng per brain. All data are expressed as the mean ± SEM. Mediators that exceeded 1 ng either in the control samples or the n-6^ex/n-3^def samples are boxed. Abbreviations: COX, cyclooxygenase; DHT, dihydroxyeicosatrienoic acid; diHDoHE, dihydroxydocosahexaenoic acid; diHDoPE, dihydroxydocosapentaenoic acid; diHETE, dihydroxyeicosatetraenoic acid; EET, epoxyeicosatrienoic acid; EpDPE, epoxydocosapentaenoic acid; HETE, hydroxyeicosatetraenoic acid; HDoHE, hydroxydocosahexaenoic acid; HHTrE, hydroxyheptadecatrienoic acid; LOX, lipoxygenase; LT, leukotriene; LX, lipoxin; N.D., not detected; NPD1, Neuroprotectin D1; oxo-ETE, oxo-eicosatetraenoic acid; P450, cytochrome P450 monooxygenase; PG, prostaglandin; Rv, Resolvin; Tx, thromboxane.

Therefore, we examined the effect of these epoxy metabolites on the gliogenic fate transition of NSCs using the neurosphere assay. We focused our analysis on 16,17-EpDPE and 11,12-EET, as these two epoxy metabolites differed the greatest between the control and n-6^ex/n-3^def samples (specifically, 16,17-EpDPE and 11,12-EET had the largest decrease and the largest increase, respectively, in the n-6^ex/n-3^def samples) (Figs. 4, 5A). Applying either 16,17-EpDPE or 11,12-EET at a final concentration of 10⁻⁵ M inhibited the formation of neurospheres (Fig. 5B), prompting us to measure their effect at lower concentrations, ranging from 10⁻⁸ M to 10⁻⁶ M. At 10⁻⁸ M, 16,17-EpDPE increased the relative number of βIII tubulin⁺ neurons (Fig. 5C–5E). In contrast, 11,12-EET at 10⁻⁶ M decreased the proportion of βIII tubulin⁺ neurons and increased the proportion of GFAP⁺ astrocytes (Fig. 5C, 5F, 5G). These results suggest that low levels of 16,17-EpDPE drive the neurogenic potential of NSCs, whereas high levels of 11,12-EET drive the gliogenic fate transition of NSCs. Taken together, our data support a model in which an optimal ratio of n-6:n-3 PUFAs contributes to the neurogenic potential of cortical NSCs via EpDPEs, whereas consuming n-6^ex/n-3^def diets drives cortical NSCs down a gliogenic lineage via EETs.

Figure 5

Functional analysis of epoxy metabolites. (A): The chemical structures of 16,17-epoxydocosapentaenoic acid (16,17-EpDPE) and 11,12-epoxyeicosatrienoic acid (11,12-EET). (B): Summary of the neurosphere formation. Neurosphere formation was measured in the presence of the indicated concentrations of 16,17-EpDPE (left) or 11,12-EET (right). n = 3 animals per group. *, p < 0.05 (Dunnett's test). (C–G): Epoxy metabolites regulate the gliogenic fate transition of NSCs. NSCs were treated with the indicated concentration of 16,17-EpDPE or 11,12-EET, then stained for βIII tubulin and glial fibrillary acidic protein (nuclei were counterstained with DAPI). n ≥ 3 dishes. *, p < 0.05; **, p < 0.01 (Dunnett's test). Scale bar = 40 μm. All summary data are expressed as the mean ± SEM. Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; EET, epoxyeicosatrienoic acid; EpDPE, epoxydocosapentaenoic acid; GFAP, glial fibrillary acidic protein.

Impaired Neuronal Layer Formation in the Postnatal Brain

We next investigated whether maternal consumption of the n-6^ex/n-3^def diet has postnatal effects on the neocortex by examining P10 offspring; P10 was chosen because the neocortical neuronal layer has been established by this age. The maternal diet had no effect on the Foxp2⁺ layer VI in the primary motor cortex (Fig. 6A) or the primary somatosensory cortex (Fig. 6D); in contrast, the primary visual cortex was significantly thinner in the n-6^ex/n-3^def offspring (Fig. 6G). In the n-6^ex/n-3^def offspring, the Ctip2⁺ layer V in the primary motor cortex was also thinner (Fig. 6B), and the density of the primary visual cortex was decreased (Fig. 6H); in contrast, the Ctip2⁺ layer V in the primary somatosensory cortex was not affected (Fig. 6E). Finally, the upper layers (specifically, Cux1⁺ layers IV and II/III) were thinner in all three cortices in the n-6^ex/n-3^def offspring (Fig. 6C, 6F, 6I). Thus, taken together, these results indicate that the impaired neurogenesis observed at E14.5 (Fig. 2) is a persistent phenotype, and the n-6^ex/n-3^def condition has a severe effect on the formation of the upper layers in the neocortex.

Figure 6

Maternal consumption of the n-6^ex/n-3^def diet affects the formation of specific neuronal layers in the offspring's neocortex. Immunofluorescence (left panels), thickness, and cell density (right panels) in the primary motor cortex (A–C), primary somatosensory cortex (D–F), and primary visual cortex (G–I) in P10 offspring of female mice fed the control diet or the n-6^ex/n-3^def diet. In the indicated cortices, the cells were stained for Foxp2 (A, D, G), Ctip2 (B, E, H), or Cux1 (C, F, I). n ≥ 3 cortices per group. (J): The neocortex of P10 offspring contains similar numbers of ALDH1L1⁺ astrocytes (upper panels). Relative expression of the indicated genes was also measured using quantitative RT-PCR (lower panels, n = 4 cortices per group). (K): The neocortex of P1 offspring contains similar numbers of proliferating cell nuclear antigen (PCNA)⁺ ALDH1L1⁺ cells (arrowheads) (n = 4 cortices per group). *, p < 0.05, **, p < 0.01 (unpaired Student's t-test). Scale bars = (A–I) and (J): 50 μm; (K): 20 μm. Thickness and density are expressed as μm and cells per 4.0 x10⁴ μm², respectively. All summary data are expressed as the mean ± SEM.

Next, we examined the effect of maternal n-6^ex/n-3^def diet on astrogenesis in the postnatal (P10) neocortex. Surprisingly, the expression of ALDH1L1—as well as several other astrocytic markers—was similar between the two dietary groups (Fig. 6J). A previous study found that astrocytes proliferate during the early postnatal period, and this proliferation is a major source of astrocytes in the postnatal brain 30. We therefore asked whether the n-6:n-3 imbalance induced by the n-6^ex/n-3^def maternal diet might have affected the proliferation of astrocytes in the developing offspring. Interestingly, maternal diet had no effect on the number of proliferating astrocytes (i.e., positive for both proliferating cell nuclear antigen and ALDH1L1) in neonatal (P1) offspring (Fig. 6K). Thus, we conclude that the transient increase in gliogenic NSCs and astrocytes in the embryonic brain of n-6^ex/n-3^def offspring does not induce excess astrogenesis in the postnatal neocortex, given that astrocyte proliferation is intact in the postnatal brain.

Maternal Consumption of the n-6^ex/n-3^def Diet Increases Anxiety-Related Behavior in the Offspring

To examine long-term functional consequences of the maternal n-6:n-3 imbalance, we looked for evidence of behavioral defects in the offspring. In humans, reduced intake of n-3 PUFAs has been associated with an increased risk of anxiety disorders 31, 32, and animal studies have shown that increased n-6 PUFAs and decreased n-3 PUFAs in the brain increases several anxiety-related behaviors 33-35. Thus, to address the neurodevelopmental impact of the maternal n-6:n-3 imbalance on the offspring's mental state, we fed postpartum mothers and their offspring a balanced (AIN-93) diets 15 after the offspring reached P10, and we performed behavioral experiments on the offspring when they reached adulthood (Supporting Information Fig. S1).

We first examined locomotor activity using the open field test. We found no significant difference between the two groups with respect to distance traveled (Fig. 7A). Moreover, body weight, which could affect locomotion, did not differ significantly between groups (Fig. 7B). Thus, maternal consumption of the n-6^ex/n-3^def diet does not affect the offspring's basal locomotor activity.

Figure 7

Maternal consumption of the n-6^ex/n-3^def diet increases anxiety-related behavior in the offspring. (A): Distance traveled in the open field test. Unpaired Student's t-test was used. (B): Body weight of the adult offspring. Unpaired Student's t-test was used. (C): Time spent in the center zone (left graph) and the number of entries into the center zone (right graph) in the open field test. *, p < 0.05, unpaired Welch's t-test (for time in the center zone) or unpaired Student's t-test (for entries into the center zone). (D): Time spent in the open (left graph) or closed (right graph) arms in the elevated plus maze test. *, p < 0.05 (unpaired Welch's t-test). n = 19 animals per group. All data are expressed as the mean ± SEM.

We next examined anxiety-related behavior using the open field and elevated plus maze tests. In the open field test, the offspring in the n-6^ex/n-3^def group spent less time in the center zone and had fewer entries into the center zone (Fig. 7C). Moreover, in the elevated plus maze test, the n-6^ex/n-3^def offspring spent relatively less time in the open arms and more time in the closed arms (Fig. 7D). These results demonstrate that maternal consumption of the n-6^ex/n-3^def diet causes abnormal anxiety-related behavior in the offspring.

Discussion

Understanding how diet affects health and drives disease is an issue of growing importance, particularly given the changing composition of dietary PUFAs in both developing and developed countries. Here, we used a combination of lipidomics, transgenics, and neurobiology to investigate how inducing an imbalance in the ratio between dietary n-6 and n-3 PUFAs affects neural development at the molecular and cellular levels. We found that feeding pregnant mice the n-6^ex/n-3^def diet causes impaired neuronal layer formation in the neocortex of the offspring, and this effect is due to a precocious gliogenic fate transition of NSCs in the brain. In addition, LC-MS/MS analyses revealed that the ratio of n-6:n-3 epoxy metabolites is significantly altered in the embryonic brain, and these changes affect the gliogenic properties of NSCs. Finally, inducing a nutritional imbalance in critical developmental periods has long-term functional consequences, manifesting as increased anxiety-related behavior in the adult offspring. Thus, our data provide the molecular and cellular bases of neurodevelopmental abnormalities caused by the n-6^ex/n-3^def diet, and our results reveal a link between abnormal maternal PUFA consumption and subsequent anxiety-related behavior in the offspring (Supporting Information Fig. S6). These results have wide-reaching implications with respect to public health.

Consuming a diet containing a well-balanced ratio of n-6 and n-3 PUFAs is believed to be important for normal brain development; this notion is based on several studies that evaluated the effects of maternal intake of n-6^ex/n-3^def diets on brain function in children (reviewed by Ryan et al. 36). Coti Bertrand et al. 9 reported the first evidence showing the effect of consuming an n-6^ex/n-3^def diet on brain development; using Nissl staining, they found that maternal intake of the n-6^ex/n-3^def diet reduced neocortical thickness in embryonic rats. In our study, we examined in further detail the neurodevelopmental consequences of the maternal n-6^ex/n-3^def diet by focusing on both embryonic and postnatal neocortical structures using molecular markers of cortical layers (Figs. 2, 6). Previous studies reported that AA and DHA affect cellular proliferation and differentiation of NSCs in vitro 10, 11, 37, 38. Here, we also identified a key cellular phenotype caused by an n-6:n-3 imbalance, namely the precocious gliogenic fate transition of NSCs (Fig. 3). Taken together, our results and previously reported data 9 strongly support the notion that maternal consumption of a well-balanced diet is important for normal brain development.

Several metabolites of AA, including 15-deoxy-PGJ₂ 26, LTB₄ 25, and lipoxin (LX) A₄ 25, have been reported to regulate the proliferation and differentiation of NSCs in vitro; however, the neurodevelopmental significance of these metabolites had been difficult to measure due to the lack of data describing the general profile of mediators in the embryonic brain. In our study, we performed the first profiling study of mediators of n-6 and n-3 PUFAs in the embryonic brain, and we found that epoxy metabolites of these PUFAs are predominant among all lipid mediators (Fig. 4). We also found that these epoxy metabolites appear to function as regulators of the neurogenic-to-gliogenic fate transition of NSCs (Fig. 5); nevertheless, the physiological roles of these epoxy metabolites in the brain is poorly understood, with the exception of mediating the release of neurohormones 39, 40. We suggest that the low level of 16,17-EpDPE followed by feeding a well-balanced diet enhances the neurogenic potential of NSCs, and that the high level of 11,12-EET caused by feeding an n-6:n-3 unbalanced diet precociously bring out the gliogenic potential of NSCs. Evaluation of the combined effect of these metabolites will further support our model. At the same time, our data cannot exclude an in vivo role for other metabolites in regulating the fate of NSCs, particularly given that several lipid mediators have biological actions even at picogram levels 41, 42. A comprehensive examination of the effects of lipid mediators on the fate regulation of NSCs will help elucidate the complex processes by which lipid molecules regulate brain development.

Epoxy metabolites of n-6 and n-3 PUFAs are produced by an enzymatic reaction mediated by P450s 29. P450s are encoded by the CYP genes, which are comprised of 14 families in mammals 29. With respect to the epoxygenase activity of P450s, the CYP2C and CYP2J isoforms are the most commonly studied, and their expression has been demonstrated in various brain regions in rodents 43. However, the expression patterns of CYP genes have not been studied thoroughly. In addition, which types of P450s recognize AA and DHA remain unclear, although we hypothesize that AA and DHA compete for the same P450 enzyme. It is possible that AA and DHA are epoxygenated by different P450 enzymes. Furthermore, P450s are expressed in the brain at 1/30th level than the expression in the liver 44, suggesting that the transport of epoxy metabolites from the liver should also be considered. Examining the expression patterns and substrate specificity of P450s will be essential for creating a precise model to describe how n-6 and n-3 PUFAs—and their epoxy metabolites—regulate brain development.

In humans, an excess and/or insufficient intake of a variety of nutrients during pregnancy is believed to be related to an increased risk of the offspring's anxiety disorders 45, suggesting that the “developmental origins of health and disease” hypothesis is relevant to mental health. Here, we provide direct evidence that poorly balanced maternal intake of n-6 and n-3 PUFAs is linked to anxious behavior in the offspring; specifically, consumption of the n-6^ex/n-3^def diet from E0 through P10 increased anxiety-related behavior, even though the offspring consumed nutritionally optimized diets from P10 through adulthood (Fig. 7). These data help define the window of susceptibility with respect to the effect of maternal n-6 and n-3 intake on the offspring's behavior; this window is consistent with a previous report that maternal consumption of an n-3 deficient diet from E0 through P21 increased anxiety-related behavior in the offspring at adulthood 34. Interestingly, in rodents, the n-6^ex/n-3^def condition leads to increase anxiety-related behavior in adult animals, but not in adolescents 46. Taken together, the results suggest that adult-onset anxious behavior due to the maternal n-6^ex/n-3^def condition arises from impaired neurodevelopment in the offspring at an early stage of life.

Several studies have provided clear evidence that patients with depression have decreased neocortical volume (reviewed by Duman and Aghajanian 47). Moreover, a recent report showed that the neocortex is involved in the pathophysiology of anxiety disorders secondary to n-6^ex/n-3^def conditions 34. Thus, in our study, impaired neocortical development due to the maternal consumption of the n-6^ex/n-3^def diet might contribute—at least in part—to the etiology of anxiety disorders. However, we currently do not have sufficient evidence to determine whether n-6^ex/n-3^def diet-induced developmental abnormalities in the neocortex are the direct cause of the offspring's subsequent anxious behavior. Furthermore, one study found that long-term consumption of an n-3 deficient diet (from 3 weeks of age through 18 weeks of age) caused a depressive phenotype in rats 48, suggesting that postnatal diet also plays a role. Determining the extent to which neonatal n-6^ex/n-3^def conditions increase anxiety-related behavior in adulthood requires further study.

The maternal consumption of a high-fat diet is known to inhibit neurogenesis in the offspring's brain 49 and to enhance the offspring's susceptibility to anxiety 50; thus, the amount of fatty acids consumed during pregnancy is clearly important for the offspring's brain development and future behavior. In our study, we used isocaloric diets 9 and confirmed no difference between dietary conditions with respect to several obesity-related metabolic parameters in the maternal mice (Supporting Information Fig. S7); nevertheless, the offspring in the n-6^ex/n-3^def group had impaired neocortical development and increased anxiety at adulthood. These data demonstrate that the n-6:n-3 imbalance causes both developmental and late-onset behavioral abnormalities, even when the total consumption of fatty acids is similar between groups. This notion was supported by our use of Fat-1 mice, in which the n-6:n-3 ratio was directly affected, thereby excluding any confounding dietary factors. The increased ratio of n-6:n-3 PUFAs led to an imbalance in the ratio of n-6:n-3 epoxy metabolites and induced a precocious gliogenic fate transition in NSCs in the developing brain. Thus, these data provide compelling evidence that both total fat intake as well as the types of lipid molecules consumed is important for normal brain development and for preventing anxiety disorders in later life.

Conclusions

Maternal consumption of the n-6^ex/n-3^def diet impaired neocortical neurogenesis in the offspring. Epoxy metabolites of these PUFAs regulated the neurogenic to gliogenic fate transition of NSCs. We also found that these offspring exhibited more anxious behavior in adulthood. These data raise important concerns regarding the increased consumption of such a modern diet during pregnancy.

Acknowledgments

We thank Drs. C. Yokoyama and F. Jacka for critically reading the manuscript; Drs. T. Sugiyama, R.A. Dyer, and R. Kimura, and Ms. S. Makino, Y. Chiba, and T. Takasugi for technical assistance; Ms. K. Ueda, J. Yoon, E. Otsuki, and A. Ogasawara and Mr. K. Koike for animal care; and all members of our laboratory for helpful discussions. This work was supported by a Research Fellowship of the Japan Society for the Promotion of Science for Young Scientists from the Japan Science and Technology Agency (to N.S.), by a Scientific Research Promotion Grant from the Mishima Kaiun Memorial Foundation (to N.S.), by a Grant-in-Aid for Scientific Research (B) from MEXT (Grant 21300115; to N.O.), and by a Research Grant from the Asahi Glass Foundation (to N.O.).

Author Contributions

N.S.: conception and design, financial support, provision of study material or patients, collection and/or assembly of data, data analysis and interpretation, manuscript writing; T.K., J.X.K., and S.M.I.: provision of study material or patients, data analysis and interpretation; H.T. and M.A.: collection and/or assembly of data, data analysis and interpretation; E.K.: provision of study material or patients; K.Y., H. Kaw., T.Y., H.A., H. Kat., and H.S.: data analysis and interpretation; N.O.: financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.

Disclosure of Potential Conflicts of Interest

H.T., H. Kaw., and H.S. are employees of Suntory Wellness Ltd., a company that markets health-food products; their employment did not affect their compliance with all relevant scientific and ethical guidelines. All other authors indicate no potential conflicts of interest.