Human brain development starts at the fifth postmenstrual week and continues after birth with most of the brain’s neurobiological processes fully developed by adolescence [1
]. There is a spurt in brain growth during the last trimester of pregnancy, with the mass of the brain approaching that of adult brain by the time the child is about three years old [3
]. The rapid growth of the brain during the last trimester requires a significant supply of long chain polyunsaturated fatty acids (PUFA), especially docosahexaenoic acid (DHA, C22:6, ω-3) and arachidonic acid (AA, C20:4, ω-6) which are the major PUFA components of brain lipids [4
DHA accumulates in brain gradually over the course of its development specifically from the third trimester onwards [5
]. Analysis of the phosphatidylethanolamine component of brain phospholipids revealed increases in ω-3 PUFA, contributed mainly by DHA, and in the ω-3:ω-6 PUFA ratio, in brains of children from six months to 8 years old compared to brains of zero to six month old infants. In turn, the latter had a higher ω-3 PUFA content and ω-3:ω-6 PUFA ratio than fetuses at 26 to 42 weeks of gestation [6
]. Similarly, when rat brains were examined at postnatal day (P) 8 (comparable to 36–40 weeks of gestation in humans in terms of brain maturation [7
]) and at embryonic days (E) 17 and E20, an increase in ω-3 PUFA was observed, with most of the increase being accounted for by DHA [8
]. A similar scenario was reported for developing piglets, with increased DHA content at term and 14 weeks postnatally compared to mid-gestation [9
Long chain ω-3 and ω-6 PUFAs can be endogenously synthesized from their precursors, alpha-linolenic acid (ALA) and linoleic acid (LA), respectively. However, these conversions are believed to be insufficient for the growing infant and must be supplemented by diet. Human breast milk contains about 13–22 wt% of its total fatty acid content as PUFA, with the DHA content varying widely depending on the mother’s diet [10
]. For example, in Japan, the average DHA content in breast milk is about 1% of total fatty acids while in Pakistan it is about 0.06% [14
]. Studies in primate models have shown that maternal and neonatal diets deficient in ω-3 PUFAs result in altered brain and retinal fatty acid composition and are associated with impaired neural and visual function [15
Fatty acid binding proteins (FABPs) are a family of 10 intracellular proteins that bind hydrophobic ligands including fatty acids [16
]. Different members of the FABP family are expressed in the brain where they participate in the intracellular trafficking of different fatty acids [16
]. Of these FABPs, brain fatty acid binding protein (B-FABP or FABP7) binds DHA with the highest affinity, although it can also bind other PUFAs such as AA [18
]. FABP7 has well-established roles in brain development and has also been shown to be a key determinant of malignant glioma growth properties and prognosis [16
]. It has been postulated that the relative levels of AA and DHA in brain may affect FABP7 expression [23
]. A recent study suggests a link between FABP7, DHA and gene expression [24
In this study, we explore the impact of DHA-rich maternal and weaning diets on brain fatty acid composition as well as FABP7 expression during the first six weeks of life. Although some reports have assessed the effect of feeding ω-3-rich diets to dams during pregnancy and lactation on brain fatty acid composition at embryonic and weanling stages [25
], to our knowledge, there has been no follow-up at later developmental stages. The brains of three-week and six-week old rats correspond to that of 2–3 year old and 12–18 year old humans, respectively [7
]. At these two developmental stages, the brains of rats and humans are thought to undergo fairly similar developmental processes [7
]. Our model should therefore provide relevant information on brain fatty acid needs in humans at these two stages.
DHA-induced improvement in brain function is believed to be due to its modulation of synaptic proteins and overall activity [31
]. Our results indicate that the developing brain readily incorporates DHA supplied during toddler/juvenile stages since the brains of six-week old rats are developmentally equivalent to 12–18 year-old human brains [32
]. Furthermore, early introduction of DHA (during lactation) maintains high DHA levels in the brain even after the pups are switched to low-DHA weaning diet. Importantly, boosting brain DHA levels is still achievable through direct dietary supply at weaning in cases where DHA was not provided during suckling. Finally, we report that DHA down-regulates the expression of FABP7, a key factor associated with neural proliferation and differentiation [33
]; however, this effect is only apparent later in brain development.
Diets were designed to provide an adequate supply of ω-6 (including AA) and ω-3 PUFA in the absence (Cnt) or presence of DHA (ω-3). Our dietary ratios of total ω-6 to total ω-3 PUFA are within the previously reported range in human breast milk samples, as well as milk fatty acid profiles of rats [10
]. We observed the following effects of DHA feeding on brain lipid composition at the toddler and juvenile stages: (i) increased brain total ω-3 PUFA content (especially DHA), (ii) reduced brain total ω-6 PUFAs, (iii) reduced ω-6:ω-3 ratio and AA:DHA ratio, and (iv) reduced brain adrenic (C22:4ω-6) and nervonic acid (C24:1ω-9) content.
It is well known that adult brains are more resistant than juvenile brains to diet-induced changes in fatty acid composition [36
]. Minor increases in brain DHA and total ω-3 PUFA content (~5%) with concomitant decreases in AA and total ω-6 PUFA (5–6%) have previously been reported in adult rats fed DHA-rich diet for eight weeks [37
]. These observations are consistent with our results in lactating dams that were fed a DHA-rich diet for three weeks. We did observe increases in LA (C18:2ω-6) and in the LA:AA ratio (C18:2ω-6:C20:4ω-6), along with a decrease in C22:4ω-6 (adrenic acid), in the brains of dams fed a DHA-rich diet. These changes likely indicate DHA-mediated inhibition of LA metabolism through ∆-6 and ∆-5 desaturases or inhibition of AA elongation, as previously reported [38
]. The lack of effect on AA levels is likely due to the adequate supply of AA in the DHA-rich (ω-3) diet.
In keeping with previous work, we observed more pronounced DHA-rich diet-induced changes in brain fatty acid composition at three weeks [26
]. Human infant brains have been reported to be similarly susceptible to changes in dietary fatty acids, with higher (39%) brain DHA accumulation in six-month old breast-fed infants compared to infants fed formula that did not contain AA or DHA [42
]. Our results indicate that the DHA-rich diet increases total ω-3 PUFA (mainly DHA) by 14.7% and decreases total ω-6 PUFA (mainly AA and C22:4ω-6) and ω-6:ω-3 ratio by 12.8% and 24%, respectively. Since the brain of a three-week old rat is at a comparable stage as that of a human toddler (2-3 years), our results suggest that DHA accretion in human brain may well extend beyond 6 months. As with the dams, decreases in LA metabolic products (C22:4ω-6 and AA), together with an increase in the LA:AA ratio, were observed at three weeks, indicating inhibition of ∆-6 and ∆-5 desaturases by DHA [43
At six weeks, changes in brain fatty acid composition were most marked in pups born to Cnt dams and fed a DHA-rich diet (Cnt/ω-3) or born to ω-3 dams and maintained on a DHA-rich diet (ω-3/ω-3). There is clear indication of inhibition of the ω-6 PUFA desaturation and elongation pathway in these pups. Unlike changes in ω-6 PUFA which were readily reversible, increases in the levels of DHA in six-week old pup brains were not reversed when DHA was discontinued. In fact, brain DHA levels in ω-3/Cnt pups showed increases that were equivalent to pups fed a DHA-rich diet for three weeks post-weaning (Cnt/ω-3) or those exposed to DHA from birth up to six postnatal weeks (ω-3/ω-3 group). In comparison, DHA-induced decreases in ω-6 PUFA were readily reversible and disappeared when DHA was discontinued (ω-3/Cnt group). It will be important to determine whether the effect of a maternal DHA-rich diet on pup brain DHA levels can be extended past six weeks.
Levels of nervonic acid (C24:1ω-9) in three-week and six-week pup brains were significantly reduced by increased levels of DHA in the diet. Nervonic acid is the major very long chain fatty acid found in sphingomyelin, one of the main components of myelin [47
]. Although studies show that there is postnatal accretion of nervonic acid in sphingomyelin, there are no systematic reports assessing the effect of a DHA-rich diet on myelination [6
]. Interestingly, a diet high in DHA results in longer latencies of the auditory startle response (a functional indicator of myelination) [50
]. In contrast to MUFA such as nervonic acid, PUFA content in myelin phospholipids is low, consisting of 1/6 to 1/3 of the PUFA content of gray matter phospholipid [48
]. Adrenic acid (C22:4ω-6) is a major PUFA of myelin [6
]. As our DHA-rich diet also significantly decreased adrenic acid levels in the brains of three-week and six-week pups, it will be important to carry out follow-up studies on the effect of a DHA-rich diet on the myelination of juvenile brain.
The mammalian brain has elevated levels of DHA and AA compared to other tissues, with the DHA:AA ratio increasing as a function of brain maturation [6
]. Analysis of human brain at different stages has revealed different ratios of DHA:AA in the different phospholipid classes, with phosphotidylserine having the highest DHA:AA ratio and phosphotidylcholine having the lowest DHA:AA ratio [6
]. The DHA:AA ratio in phosphotidylethanolamines changes over the course of brain maturation, from <1:1 to >1:1. It has been estimated that phosphotidylserine and phosphotidylethanolamine contain ~92% of the esterified DHA in total brain phospholipids of one-week old rat pups [8
Different regions of the brain as well as the different phospholipid classes show different susceptibilities to diet-induced changes in fatty acid composition. For example, the DHA-rich frontal cortex appears to be particularly sensitive to ω-3 PUFA deficiency [51
]. Phosphatidylethanolamines prepared from neuronal cells isolated from frontal cortex, cerebellum and hippocampus of pups whose dams were fed various diets during lactation showed differential accretion of DHA and AA over time depending on diet [53
]. For example, one-week to three-week old pups born from dams fed either a diet with an LA to ALA ratio of 4:1 or a DHA-supplemented (0.8 g/100 g fat) diet showed steady increases in DHA levels, especially in the cerebellum. Extending the dam diet to weaned pups for an additional three weeks resulted in further accretion of DHA in the cerebellum, but not in the frontal lobe or hippocampus [53
]. In comparison, DHA supplementation (0.8 g/100 g fat) had no effect on DHA levels in phosphotidylcholine in all three regions tested, although six-week old pups did show increased accretion of DHA in this phospholipid subclass in the frontal lobe and to a lesser extent in cerebellum [53
]. Thus, results from the phosphotidylethanolamine analysis are in general agreement with our results, with the exception that we did not observe a further increase in DHA levels in whole brain phospholipids at six weeks compared to three weeks in pups fed a continuous DHA-rich diet for six weeks.
We have previously noted associations between the AA:DHA ratio and FABP7 in normal brain and brain tumors [23
]. For example, FABP7 levels are high during normal brain development when the AA:DHA ratio is relatively high [8
]. FABP7 expression decreases from birth onwards, a period that coincides with high brain DHA accumulation and a lower AA:DHA ratio [9
]. Furthermore, the AA:DHA ratio in human malignant glioma tumors is increased compared to that of normal brain [54
], with FABP7 expression also up-regulated in these tumors [21
]. During brain development, FABP7 is expressed in radial glial cells, neural stem/progenitor cells that have self-renewal capacity and can differentiate into both neuronal and glial cells [19
]. Radial glial cells form the fiber network along which neurons migrate in developing brain. Although radial glial cells are primarily found in developing brain, these cells are also retained in the centers of the brain that undergo neurogenesis in the adult [57
]. In vitro
binding studies indicate that FABP7 has a special affinity for PUFAs, including DHA and AA [18
In this study, we tested the hypothesis that increased FABP7 levels are associated with a high AA:DHA ratio in a normally developing brain. Interestingly, we found that boosting brain DHA levels (thus decreasing the AA:DHA ratio) was associated with significant reductions in FABP7 levels at six weeks. The correlation between high levels of FABP7 and AA suggests a role for FABP7/AA in processes related to radial glial cell function such as formation of the fiber network that guide neuronal migration. Thus, there may be a reduced need for FABP7/AA-mediated events in the brains of six-week old pups exposed to a DHA-rich diet. In general agreement with our observation that changes in FABP7 levels were noted at six weeks but not at three weeks, Pelerin et al.
reported little if any change in FABP7
RNA levels in the cortex and microvessels of P14 pups whose dams were fed a DHA-supplemented diet [59
]. Brains from older pups were not analyzed by these investigators.
Studies involving humans and pigs have shown that brain DHA content increases postnatally (up to eight years and 14 weeks, respectively) while brain AA plateaus or decreases postnatally [6
]. Similar patterns have been observed in rats at ED17, ED20, and P8, with AA:DHA ratios of ~2 and ~1 observed in total brain lipids at E17 and P8, respectively [8
]. While we didn’t observe an increase in DHA content from three weeks to six weeks, there was a decrease in AA content during this period (by 19%) (Supplemental Figure S1
), resulting in an overall decrease in the AA:DHA ratio as the brain assumes higher levels of structural and functional maturation. Along with this change in the AA:DHA ratio, we observed a significant reduction (69%) in brain FABP7 levels from three weeks to six weeks. Thus, we propose that a DHA-rich diet during lactation and/or weaning may enhance or accelerate brain maturation, as suggested by the observed: (i) increase in ω-3 PUFA; (ii) decrease in ω-6 PUFA; (iii) decrease in ω-6:ω-3 PUFA (and AA:DHA) ratio; and (iv) decrease in FABP7 protein levels.