The Influence of Dietary Docosahexaenoic Acid and Arachidonic Acid on Central Nervous System Polyunsaturated Fatty Acid Composition

Abstract

Numerous studies on perinatal long chain polyunsaturated fatty acid nutrition have clarified the influence of dietary docosahexaenoic acid (DHA) and arachidonic acid (ARA) on central nervous system PUFA concentrations. In humans, omnivorous primates, and piglets, DHA and ARA plasma and red blood cells concentrations rise with dietary preformed DHA and ARA. Brain and retina DHA are responsive to diet while ARA is not. DHA is at highest concentration cells and tissues associated with high energy consumption, consistent with high DHA levels in mitochondria and synaptosomes. DHA is a substrate for docosanoids, signaling compounds of intense current interest. The high concentration in tissues with high rates of oxidative metabolism may be explained by a critical role related to oxidative metabolism.

Neural tissue contains the highest concentrations of the most highly unsaturated fatty acids found in mammals. The aggressive increase in specific CNS long chain polyunsaturated fatty acids (LCP) starting about 25 weeks of gestation lead to studies of the perinatal ontogeny of LCP in the CNS [1-3], and to studies of consequences of abnormal supply of LCP [4, 5]. These in turn led to concerns for CNS development in the smallest preterm infants [6, 7], who do not benefit from placental transfer of LCP, particularly docosahexaenoic acid (DHA) and arachidonic acid (ARA), during the third trimester of gestation and until recently were fed on diets devoid of LCP. Though several tracer studies showed that even the smallest preterm infants synthesize DHA and ARA (e.g., [8]), and that primates synthesize DHA in utero [9], the capacity to synthesize DHA appears to be very low in humans at all life stages [10]. Studies of autopsy samples of infants dying due to non-neurological events showed that the concentration of CNS DHA is 10-30% lower in bottle fed term infants than in breastfed infants, depending on the region of the CNS[11, 12]. These studies were performed prior to the era when DHA was a component of infant formula, and were widely interpreted as suggestive of the hypothesis that dietary DHA promotes CNS DHA status. These studies along with early animal studies prompted intense interest as to whether DHA enhances functional development in human infants.

While the prevalence of DHA and ARA in the cerebral cortex has long been established, studies of these LCP in other regions of the CNS are scattered and few. Using conventional methods, we studied DHA and ARA of fatty acid profiles in the well-nourished breast-fed 4 week old baboon neonate to determine if DHA and ARA are at high concentrations outside the cerebral cortex [13]. We also took advantage of samples derived from a series of our infant formula studies [14-16] in neonate baboons, which focused on DHA/ARA supplementation and prematurity, to determine how the concentration of these LCP responds to these treatments.

Figure 1 is a graphical representation of the results of DHA and ARA analysis in 26 different regions of the 4 week old baboon CNS. DHA ranges from 4.5% (w/w) in the optic nerve to 15;8% (w/w) in the globus pallidus. A striking difference is found between white matter and gray matter when the various regions are arranged from lowest to highest. The white matter region with the greatest DHA level was the corpus callosum at 7%, while the gray matter region with the least DHA was the lateral geniculate at 11.2%. Figure 1B shows the analogous data for ARA, again arranged lowest to highest. The same general trend is found, where regions of predominantly white matter are poorer in ARA than gray matter. However, in contrast to DHA, ARA levels show no sharp break between white and gray matter. Because these data are derived from the very same chromatographic analyses, we are confident that this observation is unmistakably a property of the tissue. The origin of this difference is not known; one suggestion is that it is related to the very high level of ARA in vascular membranes throughout the CNS, which would serve as a depot for ARA. DHA is not concentrated in vascular membranes, but rather is at high concentration in synaptosomes and mitochondria. A separate observation to emerge from Figure 1 is that the range of DHA concentrations occurs over 3.5-fold, while that for ARA is about 2-fold.

Figure 1.

The levels of DHA and ARA in 4 week old baboon neonate CNS (n=7), arranged in order of concentration for each LCP (means±SD). A) DHA show a sharp demarcation between white matter (hatched bars) of lower concentration and gray matter (black bars) [spinal cord is a mixture of gray and white matter]. B) ARA concentrations are higher in gray matter, but no sharp demarcation is observed.

Figure 2 provides some data on one hypothesis as to the function of DHA. Data are extracted in part from a publication of others who measured the local cerebral metabolic rate for glucose (LCMRglc) by positron emission tomography (PET) for 3 juvenile, sedated, vervet monkeys [17]. Our DHA levels were plotted against the LCMRglc for CNS regions that were reported elsewhere [18]. The correlation coefficient is highly significant with or without the excluded point (cerebellum) and the relationship is remarkably tight for different studies in different species. These data are consistent with the hypothesis that DHA plays a key structural role in membranes associated with energy production. DHA concentrations tend to follow rates of oxidative respiration from tissue to tissue. They are notably higher still in the retina, another neural tissue that normally maintains even greater concentrations of DHA. A molecular mechanism directly involving DHA has not yet been elucidated.

Figure 2.

DHA levels vs local cerebral metabolic rate of glucose uptake (LCMRglc) using our measurements and literature values of metabolic rate measurements in sedated vervet monkeys (ref [17]). Excluding the cerebellum, the correlation coefficient is r²=0.68 (p=0.003) [18].

Effect of Supplementation and Prematurity

Previous studies in primates established that experimental omega-3 PUFA deficiency results in impaired retinal function[19], confirming and extending earlier data showing similar effects in rats[20]. These studies were extreme in that omega-3 PUFA were completely omitted from the diets of developing monkeys. These studies show that tissue composition in omega-3 deficiency changes systematically. As tissue DHA concentrations drop, the omega-6 analogue docosapentaenoic acid (DPA, 22:5n-6) rise and approximately reciprocally replace DHA. Functional deficits in retinal function, measured by electroretinogram, and visually evoked potentials, are directly related to this drop in tissue change. Measurements in human suggest that cognitive development is not optimal either.

In a series of studies in primates [14-16, 21-23], we investigated clinically relevant diets to further extend the findings of those studies and establish how CNS composition responds to diet and prematurity in perinatal baboons. We choose the baboon in part because it is an omnivorous primate with lipid metabolism similar to humans.

Five treatments were compared. The breastfed animals (B) were left with the mother to nurse normally. We prepared diets with (+) and without (−)DHA/ARA supplements. Term neonates (T) were delivered spontaneously, and preterm animals (P) were taken by Cesarian section at about 3.5 weeks before expected term delivery.

Figure 3 shows DHA levels for two CNS regions that are representative of others. In the frontal lobe, the B group DHA is significantly greater than all other treatments. Supplementation increases DHA above unsupplemented groups. DHA in all cerebral cortex lobes and the cerebellum behaved consistently. In contrast, in the globus pallidus and all other gray matter, supplementation did raise DHA to breastfed levels. Our most recent studies are consistent with these observations. Importantly, they extend them by showing that dietary DHA levels of 1% do raise cerebral cortex DHA beyond that which can be achieved by lower levels [24].

Figure 3.

The response of two representative regions of the 4 week old baboon CNS to prematurity and supplementation. Cerebral lobes are significantly richer in DHA for the breastfed (B) animals than all other treatments, whereas for globus and all other gray matter (except cerebellum), DHA/ARA supplemented groups were not significantly different than the breast fed group.

Another important observation from these studies is that neither supplementation nor prematurity had any significant effects on ARA levels in the CNS. In fact, the small trend observed was to paradoxically decrease ARA with supplementation[13]. Apparently, ARA is more tightly controlled than DHA, at least in the CNS. This seems reasonable given that ARA serves as a substrate for synthesis of an array of highly bioactive eicosanoids, though it should be noted that DHA is also coming to be understood as a substrate for bioactive compounds [25]. Alternatively, the pathway for ARA biosynthesis from its more abundant 18 carbon precursor linolenic acid (18:2n-6) may be more efficient than the pathway for DHA biosynthesis from alpha-linolenic acid (18:3n-3). Biosynthesis of ARA requires three steps (18:2→18:3→20:3→20:4) all considered to be located on the endoplasmic reticulum (ER), compared to seven steps for the widely accepted DHA biosynthetic pathway (18:3→18:4→20:4→20:5→22:5→24:5→24:6→22:6) , where the last step requires transport from the ER to the peroxisomes for beta-oxidation [26].

Conclusions

These observations point to several important characteristics of DHA as a unique component of mammalian tissue. DHA is widely distributed at high concentration in the CNS, and thus plays some critical role throughout the CNS. That role may be related to the rate of oxidative respiration. Because of DHA's widespread distribution, one of its roles is also likely to be central to oxidative respiration, rather than in support of some specific function that requires energy. Unlike ARA, DHA concentrations are vulnerable to limitations in the supply of precursor. For an essential component of tissue, this implies that the dietary supply of DHA has not been limiting until modern times. DHA in the diets of infants must be adequate or there is risk of suboptimal development.