Arachidonic acid and docosahexaenoic acid are derived from the essential fatty acids linoleic acid [18:2(n-6)] and -linolenic acid [18:3(n-3)], respectively. Arachidonic (AA)³ and docosahexaenoic (DHA) acids are found in large concentrations in brain lipids. Both fatty acids are incorporated into membrane phospholipids and are essential for development and function of brain (Bazan 1990, Bourre et al. 1989, Clandinin et al. 1980a and 1980b, Moore et al. 1990, Neuringer et al. 1986). The optimum level of these components of neural tissue has not been determined. The capacity of the human fetus or premature neonate to synthesize 20:4(n-6) and 22:6(n-3) from linoleic and -linolenic acids, respectively, has not been conclusively established. The capacity to elongate fatty acid chains and desaturate essential fatty acid precursors is a concern for the premature or low-birthweight infant born with low fat stores, and subsequently minimal stores of essential fatty acids.

Infant formula as a substitute for human breast milk has become an acceptable alternative for feeding preterm infants. Changes to the composition of infant formula designed to optimize brain development by addition of 20:4(n-6) and 22:6(n-3) have been proposed (Clandinin et al. 1981 and 1982) and recommended (Clandinin et al. 1989, Koletzko et al. 1987). In Canada and Australia, the 18:2(n-6)/18:3(n-3) fatty acid ratio has been recommended to be within the range of 4:1 to 10:1 (in Europe, 5:1 to 15:1). This range is similar to that found in human breast milk (Clandinin et al. 1982, Jensen 1989), which also contains 20:4(n-6) and 22:6(n-3).

The relationship between diet and brain development has generally focused on undernutrition, malnutrition or essential fatty acid deficiency. Studies examining alteration of (n-6) and (n-3) fatty acid composition have been primarily limited to analyzing the brain as a whole or to investigating the response at only one time period. Much of the research examining effects of diet on brain cells has focused on neuronal function. Much less is known about glial cell responsiveness to alterations in nutrient supply in the absence of malnutrition (Greenwood and Craig 1987). Studies investigating regional variations in function or composition have focused primarily on the cerebrum, brainstem or cerebellum. Few studies have examined the hippocampus despite the relative ease of removal and its postnatal period for completion of development.

The rat model was chosen for this experiment because, like the preterm infant, much of rat brain development occurs postnatally. The cerebellum of both rats and humans is immature at birth (Vitiello et al. 1989) in comparison to the more advanced stage of cerebellum development in chickens and guinea pigs at birth. Cellular events that occur postnatally in rats are representative of brain growth events that could occur in an infant born prematurely (Morgane et al. 1993). Thus, the effect of diet on fatty acid accretion in neonatal rat brain may be indicative of effects on developing brain in the preterm infant.

The purpose of this experiment was to determine in developing rat brain whether accretion of 20:4(n-6) and 22:6(n-3) differs between cell types and brain regions in relation to postnatal timing of cellular and regional development. The impact on rat brain development of varying dietary 18:2(n-6)/18:3(n-3) fatty acid ratios and addition of 20:4(n-6) and 22:6(n-3), alone or in combination, was also investigated.

MATERIALS AND METHODS

Semipurified experimental diets (Clandinin and Yamashiro 1980) contained 200 g fat/kg diet and varied in fat composition (Table 1). Dietary fats were formulated based on the fat composition of an existing infant formula providing an (n-6)/(n-3) ratio of 7.3:1 (Wyeth-Ayerst Research, Philadelphia, PA). Five experimental diets were formulated by addition of triglycerides to alter the composition of this basic fat formulation (Table 1). The 20:4(n-6) and 22:6(n-3) triglycerides utilized were obtained from single cell oils.

Table 1. Fatty acid composition of diets fed to lactating rat dams and weaned pups.1 [View Table]

Rat pups were separated according to gender prior to decapitation. The number of brains pooled for one sample varied with animal age. From newborn pups, five brains were pooled per sample. At both one and two weeks of age, three brains were pooled per sample. For three-week-old and for six-week-old pups, two brains and one brain were used per sample, respectively. One entire litter was killed at the same age. Excised brains were placed in ice-cold sucrose solution (0.32 mol/L). Meningeal membranes were removed, and dissection of frontal, cerebellar and hippocampal regions was performed on ice. Stomach contents of two rats from each litter were also removed and analyzed for fatty acid composition.

Gel electrophoresis and immunoblotting were carried out to ensure purity of cell fractions (Harlow and Lane 1988). Proteins from isolated brain cell types were comparatively analyzed by gel electrophoresis with glial fibrillary acidic protein (GFAP) and neurofilaments as standards. Glial and neuronal fractions isolated were not cross-contaminated with protein characteristic of other cell types.

RESULTS

Cellular preparations contained only minor cross contamination from cell membrane fragments and microvessels as determined by microscopic examination. The presence of glial fibrillary acidic protein only in glial samples, and the presence of neurofilament only in neuronal samples was verified by gel electrophoresis and immunoblotting.

Table 2. Effects of diets differing in (n-6) and (n-3) fatty acids on arachidonic acid (AA) and docosahexaenoic acid (DHA) levels from stomach contents of one-week-old rat pups1 [View Table]

Analysis of stomach contents from rat pups has been shown to reflect the maternal milk composition (Lien et al. 1994). Analysis of stomach contents from rat pups in this study reflected the fatty acid profile of diets consumed by the dams. For example, stomach contents from two-week-old rat pups whose dams were fed the diet containing 10 g AA and 7 g DHA/kg, showed an AA content of 1.2 g /100 g fatty acids and a DHA content of 0.9 g/100 g fatty acids. Table 2 lists DHA and AA from stomach contents of rats at one week of age.

Table 3. Effects of diets varying in (n-6) and (n-3) fatty acids on 20:4(n-6) in phosphatidylethanolamine (PE) and phosphatidylcholine (PC) in cerebellar neuronal cells of Sprague Dawley rats from birth to six weeks of age1 [View Table]

Glial cells of brain regions examined exhibited differences in accretion of 22:6(n-3) over time. For example, in cerebellar and frontal regions, accretion of 22:6(n-3) was apparent up to six weeks of age (Fig. 2). In the hippocampus, the maximal level of 22:6(n-3) was attained at two weeks of age except in rats fed H-22:6(n-3) (Fig. 2C). In these rats, accretion of 22:6(n-3) in the hippocampus continued up to three weeks of age. Feeding diets containing an 18:2(n-6)/18:3(n-3) ratio of 4:1 or L-22:6(n-3) (Fig. 2B and F, respectively) produced the greatest relative percent of 22:6(n-3) at all ages in all three brain regions compared to rats fed diets supplying a 18:2(n-6)/18:3(n-3) fatty acid ratio of 7.3:1 (Fig. 2A), the same diet containing 10 g 20:4(n-6)/kg (Fig. 2D) or the same diet containing 22:6(n-3) and 20:4(n-6) (Fig. 2E). Feeding the L-22:6(n-3) diet produced higher 22:6(n-3) content in glial cells compared to feeding the H-22:6(n-3) diet (Fig. 2F vs. C). However, the age and regions in which this occurred differed somewhat from neuronal cells. The effect of the L-22:6(n-3) diet treatment on 22:6(n-3) in glial cells was observed in the cerebellum and hippocampus at one week of age and at one, two and six weeks of age in the frontal region (Fig. 2F). Feeding 20:4(n-6) resulted in significant accretion of 20:4(n-6) in each brain region compared to feeding the same diet lacking 20:4(n-6), i.e., the diet providing a 18:2(n-6)/18:3(n-3) ratio of 7.3:1 (Table 3). Feeding the diet containing both 22:6(n-3) and 20:4(n-6) increased 20:4(n-6) levels but only in the hippocampus (data not shown).

In glial cells, few differences due to diet were observed for accretion of 22:6(n-3) (Fig. 4). In addition to the greater 22:6(n-3) content noted in the hippocampus at three weeks of age (Fig. 4B), the 22:6(n-3) content was greater at six weeks of age only in cerebellum (Fig. 4B). The higher level of 22:6(n-3) in cerebellum at six weeks was observed in rats fed the diet containing a 18:2(n-6)/18:3(n-3) fatty acid ratio of 4:1 (Fig. 4B). Similar increase in 22:6(n-3) was observed in the cerebellum of rats fed L-22:6(n-3) (Fig. 4F). Accretion of 20:4(n-6) in phosphatidylcholine of glial cells (Fig. 5) was maximal in the hippocampus at three weeks of age when rats were fed both 22:6(n-3) and 20:4(n-6) (Fig. 5E). 20:4(n-6) level in all other diets was maximal prior to three weeks of age.

DISCUSSION

Effects of diet on phosphatidylethanolamine and on phosphatidylcholine were apparent (Fig. 1 and 2, Fig. 3 and 4, respectively). The predominant fatty acids vary between phosphoglycerides (Sun and Foudin 1985), thus, supplementation with 20:4(n-6) and 22:6(n-3) will likely affect phosphoglyceride fractions differently. The pools and domains of phosphoglycerides, the preference for specific fatty acids and transport mechanisms for fatty acids and phosphoglycerides may play an important role (Zevenbergen and Houtsmuller 1989). Compared to phosphatidylcholine, changes were more pronounced in ethanolamine phosphoglycerides. The predominant changes observed in phosphatidylethanolamine fatty acids are perhaps to be expected (Alling et al. 1974, Selivonchick and Roots 1979) since ethanolamine phosphoglycerides are a quantitatively larger pool of unsaturated phospholipid in cell membranes of brain (Martinez 1989) and contain the highest content of longer-chain polyunsaturated fatty acids (Mead 1975).

The hypothesis that the diet containing an 18:2(n-6)/18:3(n-3) fatty acid ratio of 4:1 would reduce 20:4(n-6) content was verified for glial phosphatidylcholine. For other cell types and regional comparisons (Fig. 5 and Table 3), a lower 18:2(n-6)/18:3(n-3) ratio may be required to reduce 20:4(n-6) levels in brain although a 18:2(n-6)/18:3(n-3) ratio of 4:1 when fed to piglets reduced 20:4(n-6) in liver and plasma phospholipids (Rioux and Innis 1992). The amount of dietary 18:3(n-3) may need to be almost equal to or exceed the level of 18:2(n-6) before enzymatic conversion of 18:2(n-6) to 20:4(n-6) is inhibited (Mohrhauer and Holman 1963).

Feeding the diet containing 22:6(n-3) and 20:4(n-6) was hypothesized to increase levels of 20:4(n-6) and 22:6(n-3) in brain phospholipids in rats after weaning. In both phosphatidylethanolamine and phosphatidylcholine, feeding this diet increased content of 20:4(n-6) (Fig. 5F and Table 3), and levels of 20:4(n-6) did not differ from the 20:4(n-6) content observed in rats fed the same diet containing 20:4(n-6) without 22:6(n-3). Thus, addition of 22:6(n-3) to a diet containing 20:4(n-6) did not interfere with incorporation of 20:4(n-6), and, for most comparisons, was as effective at raising 20:4(n-6) levels as feeding 20:4(n-6) without 22:6(n-3). It is possible that increase in only 20:4(n-6) in phosphatidylethanolamine of rats fed 22:6(n-3) and 20:4(n-6) may result from dominance of (n-6) fatty acids in early life, particularly in phosphatidylethanolamine. Martinez et al. (1974) observed phosphatidylethanolamine and phosphatidylcholine to be the major phosphoglycerides in the earliest stages of life and noted that (n-6) fatty acids were predominant in the youngest brains while (n-3) fatty acids increased with increasing age.

Mature brain possesses the necessary pathways to convert 18:3(n-3) to 22:6(n-3) (Bourre et al. 1990b, Cook 1978, Dhopeshwarkar and Subramanian 1976). When rats were fed increased 18:3(n-3), an increase in the level of 22:6(n-3) was noted only in phosphatidylethanolamine (Fig. 1B and 2B), not in phosphatidylcholine (Fig. 3B and 4B). It appears that accretion of 22:6(n-3) in all phosphoglycerides is better supported when 22:6(n-3) is supplied directly in the diet. This result is in agreement with that reported previously (Anderson et al. 1990, Sinclair 1975).

In phosphatidylethanolamine, an increase in both 22:6(n-3) and 20:4(n-6) occurred in rats fed 22:6(n-3) (Fig. 1C, 2C and Table 3) compared to rats fed the same diet without these fatty acids. This suggests that supplying 22:6(n-3) is sufficient to raise 22:6(n-3) levels in phosphatidylethanolamine yet does not cause inhibition of conversion of 18:2(n-6) to 20:4(n-6). In brain phosphatidylcholine, feeding 22:6(n-3) increased the level of 22:6(n-3) in the cerebellum and frontal region but only after weaning (Fig. 3C and 4C).

The composition of fatty acids in phosphoglycerides is distinct and exhibits change with age (Alling et al. 1974, Martinez and Ballabriga, 1987). Independent response to age for 20:5(n-3), 18:2(n-6) and 18:3(n-3) may indicate that brain maintains minimal levels of these fatty acids which are attained at an early age. Lack of deposition of 18:2(n-6) and low levels of 18:3(n-3) and 20:5(n-3) in brain phosphoglycerides has previously been reported (Bourre et al., 1988; Carlson et al., 1986, Clandinin et al. 1980a and 1980b, Cook 1982, Mohrhauer and Holman 1963). Desaturase activity has been reported to be age-related (Bourre et al. 1990a, Purvis et al. 1982, Ulmann et al. 1991). Some reports suggest that desaturase activity during initial postnatal brain growth is limited (Clandinin et al. 1980a and 1980b, Innis 1992, Sanders and Raha 1987), particularly in premature infants (Clandinin et al. 1982, Koletzko 1992). In the present study, desaturation and elongation was suggested by the level of 22:6(n-3) observed in the brain of rats fed the diet containing an 18:2(n-6)/18:3(n-3) ratio of 4:1. This increase was noted primarily in ethanolamine phosphoglycerides during the first two weeks of age. It is also apparent that the milk provided by the dam during this period provides some 22:6(n-3) (0.2%; data not shown) perhaps sufficient to support accretion of 22:6(n-3) in brain.

This study focused on comparing the accretion of 20:4(n-6) and 22:6(n-3) in brain cells and brain regions and on comparing the effects of varying (n-6)/(n-3) fatty acid ratio with or without 20:4(n-6) or 22:6(n-3) on brain fatty acid composition during development. The results suggest that amount and rate of accretion of 20:4(n-6) or 22:6(n-3) differs between brain regions and between brain cell types and indicate diet fat compositions within the recommended range of (n-6)/(n-3) ratios suggested for infant feeding significantly alters fatty acid content in developing rat brain. Differences in effect of diet on brain phosphoglycerides are apparent. Age, brain region, brain cell type and individual phosphoglycerides reflect variations in dietary fat intake to different degrees in developing brain.

Inclusion of long-chain polyunsaturated fatty acids is currently being implemented in infant formulas. It is clear that the balance between 20:4(n-6) and 22:6(n-3) in the diet is a powerful determinant of the level of these fatty acids in developing brain. The overall significance of this study is perhaps in the suggestion that addition of 20:4(n-6) or 22:6(n-3) alone may be inappropriate and balanced addition of both (n-6) and (n-3) long-chain polyunsaturated fatty acids [20:4(n-6) and 22:6(n-3)] seems to be required to achieve accretion of both (n-6) and (n-3) essential fatty acids. Further research is necessary to show if additional -linolenic acid will provide optimal (n-3) and (n-6) long-chain polyunsaturated fatty acids.