Patricia E. Wainwright

Dietary essential fatty acids and brain function: a developmental perspective on mechanisms.

Brain development is a complex interactive process in which early disruptive events can have long-lasting effects on later functional adaptation. It is a process that is dependent on the timely orchestration of external and internal inputs through sophisticated intra- and intercellular signalling pathways. Long-chain polyunsaturated fatty acids (LCPUFA), specifically arachidonic acid and docosahexaenoic acid (DHA), accrue rapidly in the grey matter of the brain during development, and brain fatty acid (FA) composition reflects dietary availability. Membrane lipid components can influence signal transduction cascades in various ways, which in the case of LCPUFA include the important regulatory functions mediated by the eicosanoids, and extend to long-term regulation through effects on gene transcription. Our work indicates that FA imbalance as well as specific FA deficiencies can affect development adversely, including the ability to respond to environmental stimulation. For example, although the impaired water-maze performance of mice fed a saturated-fat diet improved in response to early environmental enrichment, the brains of these animals showed less complex patterns of dendritic branching. Dietary n-3 FA deficiency influences specific neurotransmitter systems, particularly the dopamine systems of the frontal cortex. We showed that dietary deficiency of n-3 FA impaired the performance of rats on delayed matching-to-place in the water maze, a task of the type associated with prefrontal dopamine function. We did not, however, find an association over a wider range of brain DHA levels and performance on this task. Some, but not all, studies of human infants suggest that dietary DHA may play a role in cognitive development as well as in some neurodevelopmental disorders; this possibility has important implications for population health.

Do essential fatty acids play a role in brain and behavioral development

The membrane phospholipids of the brain contain high levels of polyunsaturated fatty acids (PUFA), particularly arachidonic acid, 20:4n-6 and docosahexaenoic acid, 22:6n-3. These long-chain PUFA are synthesized from their respective essential fatty acid (EFA) precursors, linoleic acid, 18:2n-6 and linolenic acid, 18:3n-3. Although the necessity of n-6 fatty acids for optimum growth has been established, a similar requirement for those of the n-3 family is less clear. The rapid accumulation of the long-chain n-3 PUFA in the brain during prenatal and preweaning development suggests that the provision of n-3 fatty acids to the developing brain may be necessary for normal growth and functional development. The intent of this review is to assess the experimental work which addresses this question, most of which has been conducted on rodents. The emphasis will be on studies which measure behavioral outcomes, and particular attention will be paid to methodological issues which affect the interpretation of these data. An integration of the research findings will be presented and discussed in light of possible implications for therapeutic interventions.

Effects of environmental enrichment on cortical depth and morris-maze performance in B6D2F2 mice exposed prenatally to ethanol

Pregnant mice were fed a liquid diet with 25% of the calories as ethanol from day 5 to 17 of gestation; controls received equivalent amounts of diet with maltose-dextrin substituted isocalorically for the ethanol. Two male weanlings from each litter were assigned randomly to an enriched or isolated environmental condition. After 6 weeks in these environments measures of brain growth were obtained, including thickness of frontal, parietal, and occipital cortex (study 1), or their behavioral capabilities were assessed in a Morris water maze (study 2). Ethanol decreased birth weight (both studies), postweaning body weight (study 2), and brain weight (study 1), while the enriched animals in both studies were heavier. Ethanol decreased the thickness of the occipital cortex only. All groups demonstrated learning by showing a decrease in latency to locate the hidden platform over the 5 days of testing; this was supported by their spending most time in the target quadrant during the probe trial. The latencies of the enriched animals were shorter than the isolated; covariance analysis indicated that this was not due solely to their faster swimming speed.

The role of n−3 essential fatty acids in brain and behavioral development: A cross-fostering study in the mouse

A cross-fostering design was used to examine the effects on brain and behavioral development in mice of pre-and/or postnatal dietary supplementation with n−3 fatty acids. Pregnant mice were fed either of two liquid diets, control (con) or experimental (exp). Each diet provided 3% of the calories in the form of n−6 fatty acids; the experimental diet was supplemented with an additional 1.5% from long chain n−3 fatty acids derived from fish oil. There were four treatment groups, with all pups fostered at birth. These groups were (prenatal diet/ postnatal diet): Group 1, exp/exp; Group 2, exp/con; Group 3, con/exp; Group 4, con/con; a fifth control group (unfostered) was fed lab chow (LC) throughout the study. Animals from the exp/exp and con/con groups were weaned onto lab chow for later behavioral assessment. Prenatal n−3 supplementation resulted in a small acceleration of behavioral development. The adult animals did not differ in visual discrimination learning nor did they differ in visual acuity. During development the fatty acid composition of the brain membrane phospholipids reflected closely that of the pre- and postnatal dietary conditions. Levels of 22∶5n−3 and 22∶6n−3 increased in the n−3 supplemented groups, accompanied by a decrease in levels of 22∶4n−6 and 22∶5n−6; the net effect of these changes was to increase the total levels of C22 fatty acids. While these results support considerable plasticity of the fatty acid composition of the developing brain with respect to the immediate dietary availability of n−3 compounds, they do not support long term effects on learning capacity of n−3 supplementation during the developmental period.

The Effects of dietary n−3/n−6 ratio on brain development in the mouse: a dose response study with long-chain n−3 fatty acids

This study examines the effects of the ratio of n−3/n−6 fatty acids (FA) on brain development in mice when longchain n−3 FA are supplied in the diet. From conception until 12 days after birth, B6D2F1 mice were fed liquid diets, each providing 10% of energy from olive oil, and a further 10% from different combinations of free FA concentrates derived from safflower oil (18∶2n−6), and fish oil (20∶5n−3 and 22∶6n−3). The range of dietary n−3/n−6 ratios was 0,025, 0.5, 1.0, 2.0, and 4.0, with an n−6 content of greater than 1.5% of energy in all diets, and similar levels of total polyunsaturated fatty acids (PUFA). In an additional group of ratio 0.5, 18∶2n−6 was partially replaced by its δ6 desaturation product, 18∶3n−6. Biochemical analyses were conducted on 12-day-old pup brains, as well as on samples of maternal milk. No obvious effects on overall pup growth and development were observed, apart from a smaller litter size at ratio 1. Co-variance analysis indicated that increasing the n−3/n−6 ratio was associated with slightly smaller brains, relative to body weight. We found that 18∶2n−6 and 20∶5n−3 were the predominant n−6 and n−3 FA in the milk; in the brain these were 20∶4n−6 and 22∶6n−3, respectively. Increasing dietary n−3/n−6 ratios generally resulted in an increase in n−3 FA, with a corresponding decrease in n−6 FA. The n−3/n−6 ratio of the milk lipids showed a strong linear relationship with the diet, but in the brain the rate of increase tended to decrease beyond 0.5 (phosphatidylcholine, PC) and 0.25 (phosphatidylethanolamine, PE), such that there was a significant quadratic contribution to the relationship. The partial replacement of dietary 18∶2n−6 with 18∶3n−6 raised levels of 20∶4n−6 in milk, brain PC, and brain PE. These results indicate that the n−3/n−6 ratio of the phospholipids in the developing mouse brain responds maximally to maternal dietary long-chain n−3/n−6 ratios of between 0.25 and 0.5.

An Imbalance of Dietary Essential Fatty Acids Retards Behavioral Development in Mice

This study investigated the effects of an imbalance of dietary essential fatty acids on behavioral development. Pregnant and lactating mice were fed a diet with a very low (n-6):(n-3) ratio, in which the (n-6) and (n-3) fatty acids were provided solely as linoleic acid [LA, 18:2 (n-6)] and very high levels of docosahexaenoic acid [DHA, 22:6 (n-3)], respectively. The development of the pups was compared with that of pups of similar age and body weight that had been undernourished by rearing in large litters. On the day of conception (Day 0), pregnant B6D2F1 mice were assigned randomly to one of four dietary groups. Two of these groups were fed lab chow, and after birth varied in terms of the number of pups per litter, large (LgLIT) = 12, and normal (NmLIT) = 6. The remaining two groups both had six pups per litter, but varied in dietary (n-6):(n-3) ratio, low (LoRAT = 0.32) and normal (NmRAT = 4.0). On Day 32 postconception both the LgLIT and the LoRAT groups had lower body weights and were behaviorally retarded relative to their respective NmLIT and NmRAT controls. Nonetheless, there was some sparing of function in both these groups, as they were behaviorally advanced relative to younger animals of a similar body weight. These findings show that the growth retardation seen in the offspring of dams fed a diet with a low (n-6):(n-3) ratio and very high levels of DHA is accompanied by behavioral retardation of a similar degree to that seen in malnourished pups.

Short‐Term Memory Impairment and Reduced Hippocampal c‐Fos Expression in an Animal Model of Fetal Alcohol Syndrome

BACKGROUND Previous work in our laboratory has shown that exposure to ethanol during the brain growth spurt impairs spatial short-term memory in rats on the delayed matching-to-place (DMP) version of the Morris water maze. The objectives of this study were to ascertain whether this impairment could: 1) be prevented by increasing the length of encoding time and 2) be related to hippocampal c-Fos expression. METHODS Using an artificial rearing model, male Long-Evans rats were fed 6.5 g/Kg/day of ethanol from postnatal days 6-9, with controls fed an isocaloric amount of maltose dextrin. As adults, rats in each treatment condition were trained and subsequently tested on either the DMP version of the Morris water maze, or on a random platform version (RAN) that incorporated the same performance requirements, but disallowed spatial learning. Brains were processed for c-Fos expression. RESULTS Ethanol-exposed rats showed longer search trials during training and took longer to learn the DMP task. When the delay between search and recall trials was increased from 60 sec to 120 min, the performance of ethanol-exposed rats was impaired compared with that of controls after a 10 sec, but not after a 45 sec, encoding time. Brain c-Fos expression was increased in hippocampus, prefrontal cortex and visual cortex in rats trained on the DMP compared to the RAN task. Furthermore, in the DMP-trained rats, hippocampal c-Fos expression was lower in ethanol-exposed rats. CONCLUSIONS These results suggest that the short-term memory impairment of ethanol-exposed rats 1) can be improved slightly by an increase in encoding time and 2) is related to a decrease in c-Fos expression in the hippocampus.

Effects of Dietary n-3 Fatty Acid Deficiency on Morris Water-Maze Performance and Amphetamine-Induced Conditioned Place Preference in Rats.

In these studies we examined whether dietary n-3 fatty acid (FA) deficiency in adult male rats was associated with effects on performance in the Morris water-maze and with the development of a conditioned place preference to low (0.5 mg/kg) and high (2.0 mg/kg) doses of amphetamine. The male rats used in these studies had been raised for two generations on n-3 deficient diets, which produced an n-6: n-3 FA ratio in brain lipids three times that of animals fed an n-3 adequate diet. Although the two groups did not differ on learning the position of the hidden platform in the Morris water-maze, the n-3 deficient rats did show deficits on a subsequent working memory version of this task, and swam longer distances to reach a visible platform. There were no differences between the groups on the development of a conditioned place preference although, during the initial conditioning cycle, the increase in activity in response to the high dose of amphetamine was apparent only in the n-3 deficient group. These findings provide preliminary support for effects of n-3 FA deficiency on working memory, but not on motivational processes as measured by response to a drug reward.

Spontaneously hypertensive, Wistar-Kyoto and Sprague-Dawley rats differ in performance on a win-shift task in the water radial arm maze

The spontaneously hypertensive rat (SHR) is a commonly used animal model of attention deficit hyperactivity disorder. Previous literature is inconclusive with respect to the exact nature of memory impairments in this strain. The objective of this study was to assess spatial memory as measured by performance of male SHR, Wistar-Kyoto (WKY) and Sprague-Dawley (SD) rats on a win-shift version of the water radial arm maze. On this task, all strains made more errors on Trial 4 when the mnemonic demand was highest, and showed a similar response when the delay was increased from 60s to 2h on Week 3. Both SHR and WKY rats made more reference memory errors than SD, however, SHR showed minimal improvement over weeks. The increase in errors may be due to a greater inclination of SHR and WKY to use a chaining strategy of entering consecutive arms than SD. Furthermore, the number of incomplete arm entries into reference memory arms decreased over weeks in WKY and SD, but increased in SHR, suggesting increased impulsivity of SHR at the later stages of testing. Although based on number of errors, the data indicate that SHR may have memory deficits, the data relating to arm entries suggest that the minimal improvement in SHR over weeks may have been due to greater impulsivity in the later weeks, rather than defective memory. Thus, these findings are consistent with SHR having impairments with selection of appropriate behavioural responses in a goal-directed task.

Nutrition and behaviour: the role of n -3 fatty acids in cognitive function

The lipid content of retina and grey matter in the mammalian brain is high in both arachidonic acid (AA, 20:4 n-6) and docosahexaenoic acid (DHA, 22:6 n-3). These longchain polyunsaturated fatty acids are derived from their respective dietary essential fatty acid precursors, linoleic acid (LA, 18:2n-6) anda-linolenic acid (LNA, 18:3n-3), through a series of desaturations and subsequent chain elongations. In addition to their role as integral structural components of cell membranes twenty C fatty acids such as AA make an important contribution to regulatory function by serving as precursors for the eicosanoids, including prostaglandins (for review see Wainwright, 1997). AA and DHA accrue rapidly in the human brain during the third trimester and the early postnatal period, when the rate of brain growth is maximal and therefore vulnerable to the effects of nutritional deficiencies. There is controversy at present over whether the infant formulas that contain only LA and LNA are sufficient for optimum brain development, or whether additional pre-formed AA and DHA, as found in human milk, are also necessary. A specific dietary deficiency of n-3 fatty acids during development results in characteristic changes in brain fatty acid composition that include a decrease in DHA and a reciprocal increase in 22:5 n-6. Whether 22:5n-6 serves as a functional substitute for DHA, or whether there are specific functions attributable to DHA is a question that remains unresolved. However, based mainly on studies that show that n-3 deficiency is associated with changes in the electroretinogram, as well as in some aspects of visual function in various species, including human infants (for review see Carlson & Neuringer, 1999), it has been suggested that DHA plays a unique role in the function of excitable membranes. There are several published studies in which human infant subjects have been randomly assigned to be fed on formulas supplemented with DHA or with both DHA and AA, and assessed on cognitively-related measures, including visual recognition memory as assessed by the Fagan Infantest. Although there were no effects on visual recognition, preterm infants fed on DHA-supplemented diets showed shorter look durations (Carlson & Werkman, 1996; Werkman & Carlson, 1996). Interestingly, this effect has also been reported in rhesus monkeys, and it has been suggested that the longer look durations associated with lower DHA status may be due to an inability to shift attention from a visual stimulus (Reisbick et al. 1997). A recent study has shown improved problem solving in 10month-old term infants fed on diets supplemented with AA and DHA compared with those fed on a control formula that was very low in the terms of n-3 fatty acid content (Willatts et al. 1998). In contrast, lower language scores have been reported in 14-month-old term infants fed on formulas supplemented with DHA (Scott et al. 1998). However, these effects appear to be transient, and the predictive validity of early language with respect to later cognitive function is controversial (Carlson & Neuringer, 1999). Also noteworthy with respect to the functional effects of DHA are reports of case studies where DHA supplementation has been shown to have a beneficial effect in patients with Zellwegger’s syndrome, which is a peroxisomal disorder associated with severe retardation (Martinez, 1996). Chronic dietary deficiency of LNA in animals has been associated not only with changes in retinal and visual function, but also with alterations in performance on various tests of learning and memory (for review see Reisbick & Neuringer, 1997; Wainwright, 1997; Carlson & Neuringer, 1999). Although an emphasis on learning is understandable in terms of a desire to identify nutritional factors that may have an impact on the development of human intelligence, there are various methodological reasons to be cautious in this regard. First, these findings are not consistent across laboratories. Second, some studies misinterpret main effects of diet on performance of such tasks. Specifically, learning is, by definition, change over time, and dietary-induced differences in learning ability can be inferred from different rates of change, i.e. diet × time interactions, not main effects. Third, it would seem that implicit in much of this work has been the assumption that learning and memory are unitary phenomena in terms of brain function. However, there is evidence for the involvement of different neural systems in different types of memory, and experimental manipulations that impair performance on one type of learning task may actually lead to an enhancement on another (Everitt & Robbins, 1997). Thus, it is important to consider the functional domain that is being tapped by a particular task, and to assess the outcome over a variety of tasks before drawing the conclusion that overall behavioural adaptation is better or worse. Furthermore, it is important to realise that performance on cognitive measures (learning, memory) may be confounded by alterations in non-cognitive functions (emotionality, arousal) or by inadequate sensory and motor skills. This may not be viewed as a problem if one subscribes to the argument that, since it is performance that ultimately counts, to be able to demonstrate an effect on performance is sufficient. However, if one’s objective is to elucidate mechanisms, it then becomes essential to include the control groups necessary to the identification of these potential confounds. For example, as discussed in the paper by Carrie ́ et al. (2000) in the current issue of this journal, behavioural differences in the first session of testing on some behavioural tasks, such as those requiring an active avoidance response, are more likely to be indicative of the effects of factors such as British Journal of Nutrition(2000),83, 337–339 337