Heart arachidonic acid is uniquely sensitive to dietary arachidonic acid and docosahexaenoic acid content in domestic piglets

Abstract

This study determined the sensitivity of heart and brain arachidonic acid (ARA) and docosahexaenoic acid (DHA) to the dietary ARA level in a dose-response design with constant, high DHA in neonatal piglets. On day 3 of age, pigs were assigned to 1 of 6 dietary formulas varying in ARA/DHA as follows (% fatty acid, FA/FA): (A1) 0.1/1.0; (A2) 0.53/1.0; (A3-D3) 0.69/1.0; (A4) 1.1/1.0; (D2) 0.67/0.62; (D1) 0.66/0.33. At necropsy (day 28) higher levels of dietary ARA were associated with increased heart and liver ARA, while brain ARA remained unaffected. Dietary ARA had no effect on tissue DHA accretion. Heart was particularly sensitive, with pigs in the intermediate groups having different ARA (A2, 18.6 ± 0.7%; A3, 19.4 ± 1.0%) and a 0.17% increase in dietary ARA resulted in a 0.84% increase in heart ARA. Further investigations are warranted to determine the clinical significance of heart ARA status in developing neonates.

Introduction

Arachidonic acid (20:4n-6; ARA) and docosahexaenoic acid (22:6n-3; DHA) are two long chain polyunsaturated fatty acids (LCPUFA; ≥2 double bonds, ≥20 carbon) that contribute to normal growth and development during the perinatal period. Both are natural components of breastmilk and since 2002, these LCPUFA have been added to infant formulas in the U.S. at target levels of 0.66% and 0.33% total fatty acids (FA), respectively (1). Mean worldwide levels of ARA and DHA in breastmilk are 0.47 ± 0.13% total FA (range: 0.24 – 1.0% FA) and 0.32 ± 0.22% FA (range: 0.06 – 1.4% FA), respectively, with individual variability in DHA attributed largely to maternal diet (2). Inevitably, the wide variation in breastmilk LCPUFA as well as the relative inaccessibility to human infant tissue presents a major challenge for optimizing levels of ARA and DHA in formula to support neonatal development.

It is well-established that the addition of ARA and DHA to formula enhances blood LCPUFA levels equivalent to those of breastfed infants and improves visual acuity and cognitive performance compared with infants fed LCPUFA-free formula (3–6). Functional outcome studies with infants provide clear evidence for the addition of DHA at 0.32% total FA (7,8), while higher DHA levels may offer further benefit for neural development, especially in target populations (9–11). Optimal levels of ARA remain to be determined at any DHA intake and are based largely on mean global levels in breastmilk (1). ARA comprises approximately 10 – 12% total FA in human infant central nervous tissue (e.g. cerebral cortex and retina) and appears to be influenced to a greater extent by postnatal age than dietary ARA supply (12). Studies with animal models suggest a relative resistance of brain and retina ARA to dietary ARA intake, possibly as a mechanism for regulating its potent bioactivity (13,14). ARA levels in the heart, liver and blood-borne pools vary with ARA intake and may compete with n-3 LCPUFA for tissue incorporation (3,15–22), especially in the liver (19,22). Further, in some regions of the neonatal brain, ARA declines when DHA intake is increased (22).

The sensitivity of heart ARA to dietary ARA content has been reported a few times but has not been systematically studied (16–18,20). In pigs (20), dietary ARA and DHA were maintained at a constant 2:1 ratio and total levels varied up to 5 times the target levels in formula. While liver ARA increased with only the highest ARA intakes, heart ARA responded in a direct dose-response manner to each incremental increase in dietary ARA. Further, heart ARA more than doubled with the highest ARA level compared with pigs fed LCPUFA-free formula.

We tested the hypothesis that heart and brain ARA is sensitive to dietary ARA in a dose-response design with DHA constant at a high physiological level. We examined four levels of ARA ranging from 0.09 to 1.0% total FA against a background of 1.0% DHA. Intermediate levels were 0.53 and 0.67% total FA, where 0.53% is slightly above the worldwide ARA mean in human breastmilk and 0.67% is commonly used in infant formulas, though it is among the higher values reported for human breastmilk (2). A DHA dose-response was also included to determine the influence of dietary DHA level on tissue ARA incorporation when ARA levels are comparable to those added in conventional formula (0.66% ARA). The range of DHA content in human milk is 0.06 to 1.4% (2). A randomized, maternal-reared (MR) reference group was included for comparison. The results of these experiments indicate that the heart and liver were sensitive to dietary ARA level, but the brain showed little response.

Materials and Methods

Animals and diets

All procedures involving animals were approved by the Institutional Animal Care and Use Committee at Cornell University. On day 3 of life, domestic piglets were matched for weight and gender and assigned to one of six milk replacer diets (n = 8 per diet). Piglets in the formula-reared (FR) groups were then transferred to the Large Animal Research and Teaching Unit, where they were housed individually in raised metal cages and maintained on a 16/8 hr light/dark cycle. The seventh group was MR and remained with the dam at the swine facility for the duration of the study. Diet intake and growth performance have been reported elsewhere (23).

Milk replacer formula consisting of 60% experimental diet (Research Diets, Inc., New Brunswick, NJ) and 40% Birthright baby pig milk replacer (Ralco Nutrition, Inc., Marshall, MN) was fed to FR piglets on days 3 – 28 of age. The ingredient composition and nutrient analysis of the base experimental diet are presented in Table 1. A detailed nutrient composition and analysis of the Birthright baby pig milk replacer has been reported previously (24). As fed, the milk replacer formula had a caloric density of 0.7 kcal/mL and a nutrient composition as follows (% wt/wt): protein 27.0, fat 22.1, carbohydrate 39.5, fiber 0.7 and ash 10.7. The FA compositions of the milk replacer formulas and sow milk (day 14 in lactation) are presented in Table 2. Diets varied in the ratio of ARA/DHA as follows (% FA/FA): (A1) 0.1/1.0; (A2) 0.53/1.0; (A3-D3) 0.69/1.0; (A4) 1.1/1.0; (D2) 0.67/0.62; (D1) 0.66/0.33 (conventional infant formula). Diets A1 – A4 are a dose-response for ARA against DHA constant at 1.0%, while Diets D1 – D3 are a dose-response for DHA against ARA constant at 0.67%. The sow diets were a conventional pig ration consisting of corn and soybean as 96% of the total dry weight, and the PUFA content was comprised of 18:2n-6 (53.9 ± 0.2% FA) and 18:3n-3 (3.4 ± 0.1% FA). The excess 18:2 and low 18:3 apparently drove the very low sow milk DHA (0.01 ± 0.01% FA), while ARA was high compared to the human milk range (0.74 ± 0.02% FA).

TABLE 1.

Nutrient composition and analysis of experimental diets

Composition	Content
Ingredient, g/kg dry
Dried skim milk	550
Oil blend	239
Calcium sodium	96
Mineral mix1	48
Vitamin mix2	48
Xanthan gum	14
Methionine, DL	5
Nutrient analysis, % kcal
Protein	24.3
Carbohydrate	28.7
Fat	47
Caloric density (kcal/g)	4.67

¹The mineral mix contained (g/kg): 257.8 sucrose, 615.8 dibasic calcium phosphate, 106.7 sodium chloride, 11.8 ferric citrate, 4.0 zinc carbonate, 3.4 magnesium oxide, 0.4 cupric carbonate, 0.2 manganous carbonate, 0.1 potassium iodate, 0.01 sodium selenite.

²The vitamin mix contained (g/kg): 910.1 sucrose, 75.0 choline bitartrate, 6.2 vitamin E acetate (50%), 3.4 vitamin A acetate (500,000 IU/g), 1.5 vitamin D3 (100,000 IU/g), 1.3 vitamin B12 (0.1% mannitol), 0.9 pantothenic acid (d, calcium), 0.9 niacin, 0.4 biotin (1%), 0.1 riboflavin, 0.1 thiamine HCl 0.03 menadione sodium bisulfite, 0.01 folic acid, 0.01 pyridoxine HCl.

TABLE 2.

FA composition of milk replacer formulas and sow milk¹

Diet	A1	A2	A3-D3	A4	D2	D1	MR
	% total FA
ARA	0.09	0.53	0.69	1.06	0.67	0.66	0.74 ± 0.02
DHA	1.00	1.02	1.001	1.04	0.62	0.33	0.01 ± 0.01
ARA/DHA	0.1	0.5	0.7	1.0	1.1	2.0	74.0
Σ SFA + MUFA	79.86	79.94	79.43	79.18	80.02	80.02	83.94 ± 3.16
18:2n-6	17.25	16.83	17.11	16.93	16.9	17.19	12.05 ± 2.98
Σ n-6	17.5	17.52	17.95	18.15	17.76	18	15.21 ± 3.08
18:3n-3	1.53	1.41	1.48	1.47	1.48	1.53	0.57 ± 0.06
Σ n-3	2.53	2.43	2.49	2.51	2.1	1.86	0.71 ± 0.04
18:2n-6/18:3n-3	11.3	12	11.6	11.5	11.4	11.3	21.1

¹Milk from two sows taken on day 14 of lactation; MR piglets remained with the sow for the duration of the study. The PUFA content of sow diets during gestation and lactation consisted of 53.90 ± 0.17% 18:2n-6 and 3.44 ± 0.04% 18:3n-3.

Sampling

On day 28 of age, piglets were killed via an intravenous injection of Fatal Plus (1 mL/4.54 kg body weight; Vortech Pharmaceuticals, Dearborn, MI, USA) followed by exsanguination. Necropsy was performed by a team to facilitate rapid removal and weighing of organs, and all samples were flash frozen in liquid nitrogen within 10 min of cessation of heart beat. The eyes and brain were removed by the attending veterinarian using bone cutters and surgical scissors. Retinas were submerged in physiologic saline prior to flash freezing. Cerebral cortex and cerebellum samples were harvested from the left hemisphere prior to submerging individual hemispheres in liquid nitrogen. Internal brain regions were sampled from partially thawed left hemispheres at a later time. Tissues were stored at −80°C until FA analysis.

Lipid analysis

FAME were prepared from ~50 mg of tissue or dry diet and 50 μL sow milk according to the one-step hydrolysis, extraction and methylation procedure (25) with previous modifications (26). Retinas, which were suspended in physiological saline at necropsy, were first lyophilized using a Savant SpeedVac Concentrator (Thermo Fisher Scientific, Waltham, MA) to prevent excess water molecules from hindering the methylation reaction. FAME were quantified on a 5890 Series II gas chromatograph (Hewlett-Packard, Palo Alto, CA) equipped with a BPX70 fused silica column (25 m × 0.22 mm i.d. × 0.25 m film; SGE Inc, Austin, TX). FAME were structurally identified by covalent-adduct chemical ionization tandem mass spectrometry on a Saturn 2000 ion trap mass spectrometer (Varian, Inc., Walnut Creek, CA) (27). An equal weight FAME mixture was used daily to verify response factors.

Statistical analysis

Values are reported as means ± SD. Statistical analysis was carried out using the Fit Model platform of JMP (2008 SAS Institute, 8.0) to fit mixed models and tested the hypothesis that all mean treatment values were not equal. Fixed effects were diet, gender, day 3 of age body weight and the full factorial of interactions. The random effect of litter was included in the model. Interaction effects were considered significant at P < 0.10 and fixed effects at P < 0.05. For each parameter analyzed, effects not considered significant were removed from the final model in a stepwise manner. Significance of pairwise comparisons was determined using the Student's t-test. Linear regression analysis was performed to determine the dose-response of tissue LCPUFA to the dietary ARA and DHA level. Figures are presented with best-fit curves to summarize the relationship between LCPUFA intake and tissue composition.

Results

Intake and growth

Formula intake and growth performance have been reported elsewhere (23). Formula intake was similar among all FR groups and averaged 51.6 ± 0.1 L across the entire study. Mean body weight for the FR groups on day 28 of age was 10.3 ± 0.6 kg, while MR piglets weighed significantly less with a mean weight of 7.9 ± 0.9 kg. Relative organ weights were similar for the FR groups. Absolute brain weights were similar among all piglets.

LCPUFA accretion in liver and heart

Table 3 presents selected LCPUFA from the heart and liver of pigs fed varying levels of ARA and DHA from day 3 – 28 of age. A detailed FA profile of these tissues is presented in Supplemental Tables 1 – 2. Heart ARA increased with increasing dietary ARA. Heart ARA levels were higher in the A3 group compared with Diet A2 (P = 0.04) such that a 0.17% increase in dietary ARA corresponded to an increase of 0.84% in heart ARA between the two treatments. Mean heart ARA was 40% higher in the A4 group compared with A1 pigs. The A3 and A4 diets were not significantly different from the MR heart ARA levels and the mean value (20.3% FA) was only reached at an ARA intake of 1.06% FA in the A4 diet. Liver ARA increased with increasing dietary ARA, although the greatest increase in liver ARA was observed between Diets A1 and A2. The low level of ARA intake in the A1 diet did not support the same level of liver ARA as that observed in the MR piglets.

TABLE 3.

Selected LCPUFA of heart and liver from piglets fed varying levels of ARA and DHA from days 3 – 28 of age¹

Diet ARA/DHA2	A1 0.09/1.00	A2 0.53/1.02	A3-D3 0.69/1.01	A4 1.06/1.04	D2 0.67/0.62	D1 0.66/0.33	MR 0.74/0.01	P
Heart	% total FA
ARA	14.76 ± 1.07a	18.57 ± 0.67b	19.41 ± 1.00c	20.90 ± 0.93d	19.98 ± 1.38cd	19.09 ± 1.88bc	20.26 ± 1.20cd	<.0001
22:4n-6	0.63 ± 0.09a	0.71 ± 0.09a	0.71 ± 0.06ab	0.82 ± 0.09bc	0.88 ± 0.13cd	0.96 ± 0.12d	1.33 ± 0.13e	<.0001
22:5n-6	0.00 ± 0.01a	0.00 ± 0.00a	0.00 ± 0.00a	0.00 ± 0.00a	0.00 ± 0.00a	0.02 ± 0.03a	0.08 ± 0.01b	<.0001
20:5n-3	0.75 ± 0.09a	0.54 ± 0.09b	0.49 ± 0.04b	0.41 ± 0.04c	0.42 ± 0.04c	0.32 ± 0.03d	0.28 ± 0.06d	<.0001
22:5n-3	0.59 ± 0.07a	0.59 ± 0.08a	0.57 ± 0.04a	0.57 ± 0.04a	0.71 ± 0.07b	0.83 ± 0.13c	1.24 ± 0.07d	<.0001
DHA	4.81 ± 0.24a	5.20 ± 0.49b	4.78 ± 0.43a	5.01 ± 0.25ab	4.20 ± 0.64c	3.00 ± 0.50d	0.75 ± 0.08e	<.0001
Σ 22C LCPUFA3	6.04 ± 0.25ab	6.50 ± 0.51c	6.06 ± 0.40abc	6.40 ± 0.23ac	5.78 ± 0.72b	4.81 ± 0.59d	3.40 ± 0.15e	<.0001
Liver
ARA	13.92 ± 1.49a	16.14 ± 0.87b	16.12 ± 1.10bc	16.88 ± 1.50c	17.36 ± 0.63c	17.87 ± 1.46d	17 ± 1.36bcd	<.0001
22:4n-6	1.08 ± 0.29a	1.15 ± 0.22ab	1.27 ± 0.22b	1.36 ± 0.22bc	1.28 ± 0.20bc	1.50 ± 0.10c	2.27 ± 0.35d	<.0001
22:5n-6	0.03 ± 0.06a	0.06 ± 0.05a	0.04 ± 0.04a	0.06 ± 0.07a	0.2 ± 0.09b	0.37 ± 0.11c	1.89 ± 0.53d	<.0001
20:5n-3	0.61 ± 0.14a	0.42 ± 0.15bc	0.45 ± 0.10b	0.36 ± 0.11cd	0.34 ± 0.05cd	0.31 ± 0.04d	0.18 ± 0.06e	<.0001
22:5n-3	0.94 ± 0.11a	0.91 ± 0.09a	0.94 ± 0.12a	0.94 ± 0.14a	1.01 ± 0.07a	1.11 ± 0.16b	1.35 ± 0.20c	<.0001
DHA	10.77 ± 1.77a	11.43 ± 1.22a	10.49 ± 1.41a	10.45 ± 1.79a	9.65 ± 1.01b	7.52 ± 0.69c	2.37 ± 0.30d	<.0001
Σ 22C LCPUFA	12.82 ± 1.86a	13.55 ± 1.26a	12.74 ± 1.44ab	12.81 ± 1.90ab	12.14 ± 1.21b	10.51 ± 0.81c	7.88 ± 1.22d	<.0001

¹Values represent mean ± SD, n = 8 per diet. P < 0.05 indicates that all means were not equal; NS, not significant. Pairwise comparisons determined using Student's t-test; means not sharing a common superscript were significantly different (P < 0.05).

²ARA/DHA, dietary content of ARA/DHA (% FA/FA).

³Σ 22C LCPUFA = Σ 22:4n-6, 22:5n-6, 22:5n-3, DHA.

DHA accretion in the liver reflected dietary intake with A1 – A4 pigs having the highest liver levels, D1 and D2 pigs having intermediary levels and MR pigs having the lowest liver DHA (Table 3). Similarly, DHA levels in the heart reflected dietary DHA content, with A1 – A4 pigs having the highest, D1 and D2 pigs having intermediary and MR pigs having the lowest heart DHA levels (P < 0.0001); although for Diets A1 – A4, A2 pigs had higher heart DHA levels than the A1 and A3 groups (P < 0.05). MR pigs had a strikingly low heart DHA content (0.75 ± 0.08%) that was 6-fold lower than DHA levels in the A2 group, reflecting the low DHA content of the sow milk.

In both the heart and liver, the sum of 22-carbon LCPUFA reflected tissue DHA content and was lowest in D1 and MR pigs (P < 0.0001; Table 3). Compared with the D3 group, D1 pigs had lower liver DHA and higher liver levels of 22:4n-6, 22:5n-6 and 22:5n-3 (P < 0.001). In the heart, however, D1 pigs showed increases in 22:4n-6 and 22:5n-3 compared with D3 pigs (P < 0.0001), while trace levels of 22:5n-6 were detected in all samples.

LCPUFA accretion in neural tissues

Tables 4 & 5 present selected LCPUFA from neural tissue of pigs fed varying levels of ARA and DHA from day 3 – 28 of age. A detailed FA composition of each neural tissue is presented in Supplemental Tables 3 – 8. ARA levels in the cerebral cortex, hippocampus, globus pallidus and inferior colliculus were similar across all dietary treatments. ARA levels in the cerebellum were significantly influenced by diet (P = 0.007) but showed no distinct pattern associated with ARA or DHA intake. A1 pigs had the lowest retinal ARA levels of the FR groups, although differences among the means only reached significance for the A4, D1 and D2 groups (P < 0.05).

TABLE 4.

Selected LCPUFA of cerebral cortex, cerebellum and retina from piglets fed varying levels of ARA and DHA from days 3 – 28 of age¹

Diet ARA/DHA2	A1 0.09/1.00	A2 0.53/1.02	A3-D3 0.69/1.01	A4 1.06/1.04	D2 0.67/0.62	D1 0.66/0.33	MR 0.74/0.01	P
Cerebral cortex	% total FA
ARA	11.95 ± 0.29	11.82 ± 0.60	11.80 ± 0.35	12.04 ± 0.63	11.93 ± 0.40	11.77 ± 0.68	12.33 ± 0.34	NS
22:4n-6	4.03 ± 0.52	4.06 ± 0.16	4.09 ± 0.35	4.21 ± 0.60	4.21 ± 0.32	4.19 ± 0.26	4.81 ± 0.30	NS
22:5n-6	4.49 ± 0.40a	4.44 ± 0.47a	4.68 ± 0.29a	4.61 ± 0.56a	5.52 ± 0.64b	5.49 ± 0.25b	8.26 ± 0.18c	<.0001
22:5n-3	0.20 ± 0.02a	0.19 ± 0.02a	0.19 ± 0.04a	0.19 ± 0.03a	0.20 ± 0.02a	0.25 ± 0.05b	0.32 ± 0.03c	<.0001
DHA	12.10 ± 0.83a	11.91 ± 0.74ab	11.98 ± 0.69ab	11.90 ± 0.69ab	11.30 ± 0.70b	10.56 ± 0.67c	6.88 ± 0.55d	<.0001
Σ 22C LCPUFA3	20.82 ± 0.81	20.60 ± 0.61	20.94 ± 1.01	20.91 ± 0.99	21.23 ± 0.88	20.49 ± 0.79	20.27 ± 0.72	NS
Cerebellum
ARA	10.40 ± 0.30a	10.76 ± 0.26ab	11.03 ± 0.31bc	10.44 ± 0.44a	10.95 ± 0.62bc	10.78 ± 0.43ab	11.48 ± 0.73c	0.007
22:4n-6	3.29 ± 0.41a	3.23 ± 0.38a	3.39 ± 0.32ab	3.43 ± 0.32ab	3.55 ± 0.19b	3.97 ± 0.25c	4.66 ± 0.31d	<.0001
22:5n-6	2.12 ± 0.26ab	2.19 ± 0.34a	2.18 ± 0.19ab	2.17 ± 0.26a	2.70 ± 0.37bc	3.34 ± 1.09c	4.82 ± 0.23d	<.0001
22:5n-3	0.35 ± 0.06ab	0.31 ± 0.05bc	0.32 ± 0.05bc	0.32 ± 0.05c	0.31 ± 0.02bc	0.36 ± 0.05a	0.47 ± 0.04d	<.0001
DHA	11.29 ± 1.61a	11.08 ± 1.68a	11.29 ± 1.19a	11.19 ± 1.44a	10.37 ± 0.75ab	9.83 ± 0.76b	6.10 ± 0.28c	<.0001
Σ 22C LCPUFA	17.47 ± 1.79	17.19 ± 2.23	17.56 ± 1.51	17.47 ± 1.54	17.36 ± 0.70	17.88 ± 1.37	16.53 ± 0.30	NS
Retina
ARA	9.48 ± 0.23a	9.70 ± 0.61ab	9.66 ± 0.32ab	9.81 ± 0.31b	9.96 ± 0.37b	9.95 ± 0.25b	10.74 ± 0.50c	0.0004
22:4n-6	1.66 ± 0.15a	1.72 ± 0.12ab	1.77 ± 0.12b	1.81 ± 0.17b	1.98 ± 0.19c	2.13 ± 0.12d	2.98 ± 0.17e	<.0001
22:5n-6	1.48 ± 0.30a	1.48 ± 0.32a	1.45 ± 0.16a	1.55 ± 0.30a	2.03 ± 0.64b	2.29 ± 0.26c	5.24 ± 0.39d	<.0001
22:5n-3	0.64 ± 0.04a	0.60 ± 0.05b	0.63 ± 0.06ab	0.60 ± 0.03b	0.55 ± 0.03c	0.60 ± 0.05b	0.60 ± 0.05abc	0.0001
DHA	20.42 ± 1.20ab	20.70 ± 0.54a	20.07 ± 1.47ab	20.49 ± 1.31ab	19.50 ± 1.50bc	18.68 ± 0.96c	14.16 ± 1.37d	<.0001
Σ 22C LCPUFA	24.29 ± 1.34	24.61 ± 0.49	24.03 ± 1.49	24.55 ± 1.48	24.20 ± 1.31	23.81 ± 0.90	23.10 ± 0.98	NS

¹Values represent mean ± SD, n = 8 per diet. P < 0.05 indicates that all means were not equal; NS, not significant. Pairwise comparisons determined using Student's t-test; means not sharing a common superscript were significantly different (P < 0.05).

²ARA/DHA, dietary content of ARA/DHA (% FA/FA).

³Σ 22C LCPUFA = Σ 22:3(n-9), 22:4n-6, 22:5n-6, 22:5n-3, DHA.

TABLE 5.

Selected LCPUFA of hippocampus, globus pallidus and inferior colliculus from piglets fed varying levels of ARA and DHA from days 3 – 28 of age¹

Diet ARA/DHA2	A1 0.09/1.00	A2 0.53/1.02	A3–D3 0.69/1.01	A4 1.06/1.04	D2 0.67/0.62	D1 0.66/0.33	MR 0.74/0.01	P
Hippocampus	% total FA
ARA	10.40 ± 0.89	10.78 ± 0.93	10.57 ± 1.09	10.61 ± 1.57	11.10 ± 0.97	10.85 ± 1.23	11.77 ± 0.96	NS
22:4n-6	4.57 ± 0.36	4.68 ± 0.43	4.63 ± 0.57	4.72 ± 0.63	4.78 ± 0.44	5.03 ± 0.51	5.37 ± 0.43	NS
22:5n-6	2.89 ± 0.22a	2.96 ± 0.39a	3.06 ± 0.39ab	2.68 ± 0.67b	3.71 ± 0.40c	3.31 ± 0.53ac	5.58 ± 0.60d	<.0001
22:5n-3	0.24 ± 0.03ab	0.22 ± 0.05bc	0.22 ± 0.02bc	0.20 ± 0.05bc	0.19 ± 0.02c	0.24 ± 0.04ab	0.27 ± 0.04a	0.009
DHA	9.89 ± 0.26a	9.74 ± 1.43a	9.66 ± 0.68ab	9.63 ± 1.73a	9.44 ± 1.03a	8.27 ± 1.48b	6.12 ± 0.68c	0.0002
Σ 22C LCPUFA3	17.20 ± 1.41	17.60 ± 1.87	17.13 ± 1.99	17.23 ± 2.89	18.12 ± 1.44	16.84 ± 2.23	17.34 ± 0.64	NS
Globus pallidus
ARA	10.33 ± 0.68	9.33 ± 1.28	10.07 ± 0.44	9.92 ± 0.76	10.02 ± 0.88	10.09 ± 0.96	10.73 ± 0.82	NS
22:4n-6	3.81 ± 0.32ab	3.59 ± 0.73b	4.25 ± 0.51bc	3.98 ± 0.87ab	3.72 ± 0.37ab	4.04 ± 0.58ab	4.78 ± 0.65c	0.02
22:5n-6	2.93 ± 0.37a	2.24 ± 0.77b	2.81 ± 0.40ab	2.64 ± 0.58ab	3.16 ± 0.73a	3.04 ± 0.78a	5.40 ± 0.92c	<.0001
22:5n-3	0.15 ± 0.12	0.12 ± 0.04	0.13 ± 0.03	0.13 ± 0.02	0.11 ± 0.04	0.14 ± 0.06	0.19 ± 0.06	NS
DHA	9.45 ± 0.97a	8.09 ± 2.61ab	8.80 ± 0.94ab	8.18 ± 1.36ab	7.91 ± 1.95b	7.66 ± 1.68bc	6.08 ± 0.35c	0.01
Σ 22C LCPUFA	16.74 ± 1.12	14.49 ± 3.72	16.44 ± 1.37	15.41 ± 2.20	15.37 ± 2.68	15.34 ± 2.50	16.92 ± 1.43	NS
Inferior colliculus
ARA	8.58 ± 0.35	8.55 ± 0.75	8.85 ± 0.84	8.61 ± 0.52	8.63 ± 0.50	8.97 ± 0.73	9.45 ± 1.25	NS
22:4n-6	3.54 ± 0.21a	3.36 ± 0.19a	3.55 ± 0.31a	3.56 ± 0.29a	3.42 ± 0.30a	3.70 ± 0.35a	4.68 ± 1.16b	<.0001
22:5n-6	1.58 ± 0.49ab	1.50 ± 0.35a	1.60 ± 0.45ab	1.62 ± 0.52ab	1.93 ± 0.53b	1.91 ± 0.32b	3.76 ± 0.45c	<.0001
22:5n-3	0.19 ± 0.02ab	0.18 ± 0.04abc	0.19 ± 0.04ab	0.15 ± 0.02c	0.16 ± 0.03bc	0.18 ± 0.03bc	0.21 ± 0.06a	0.02
DHA	10.71 ± 0.94a	10.32 ± 1.88a	10.84 ± 1.50a	9.58 ± 1.21a	10.18 ± 1.27a	10.53 ± 0.85a	6.27 ± 1.14b	<.0001
Σ 22C LCPUFA	16.35 ± 1.10	15.66 ± 2.18	16.46 ± 2.03	15.28 ± 1.60	16.02 ± 1.46	16.61 ± 1.23	15.40 ± 1.74	NS

¹Values represent mean ± SD, n = 8 per diet. P < 0.05 indicates that all means were not equal; NS, not significant. Pairwise comparisons determined using Student's t-test; means not sharing a common superscript were significantly different (P < 0.05).

²ARA/DHA, dietary content of ARA/DHA (% FA/FA).

³Σ 22C LCPUFA = Σ 22:3(n-9), 22:4n-6, 22:5n-6, 22:5n-3, DHA.

Pigs fed 1.0% DHA had significantly higher DHA levels in the cerebral cortex, cerebellum and retina than the D1 and MR groups, while D2 pigs showed intermediary levels (P < 0.0001; Table 4 & 5). DHA accretion in the hippocampus, globus pallidus and inferior colliculus was greatest for pigs fed 1.0% DHA and lowest for the MR group (P < 0.05). For these tissues, DHA levels were 17%, 15% and 3% higher, respectively, in the D3 compared with D1 pigs, although differences were not statistically significant.

The sum of the 22-carbon LCPUFA, corresponding to the sum of 22:3(n-9), 22:4n-6, 22:5n-6, 22:5n-3 and DHA, was similar across all dietary treatments in all neural tissues reported (Tables 4 & 5). For these tissues, the magnitude of the increase in non-DHA 22-carbon LCPUFA was greatest for 22:5n-6. Differences in the composition of 22-carbon LCPUFA between Diets D1 and D3 were significant and apparent, with D1 pigs having higher 22:5n-6 in cerebral cortex, cerebellum and retina (P < 0.01), and higher 22:4n-6 in cerebellum and retina than D3 pigs (P < 0.001).

Tissue ARA and DHA dose-response to diet

Figure 1 summarizes the linear regression analysis for the tissue ARA dose-response to the dietary ARA level (Diets A1 – A4). ARA levels in the heart and liver were highly responsive to the dietary ARA content (P < 0.0001), whereas ARA levels in the cerebral cortex and retina were not responsive to ARA intake. Regression analysis for the tissue DHA dose-response to dietary DHA content (Diets D1 – D3) is presented in Figure 2. Cerebral cortex, retina, liver and heart DHA levels increased with dietary DHA content (P < 0.05). For heart and liver, however, differences were most pronounced between Diets D1 and D3, as increasing the DHA content from 0.62 to 1.0% total FA (Diets D2 vs. D3) did not significantly increase heart or liver DHA levels. Cerebellum and hippocampus DHA levels also increased with dietary DHA content (P < 0.05), while globus pallidus and inferior colliculus DHA levels did not correlate with dietary DHA content (Supplemental Figure 1).

FIGURE 1.

ARA tissue dose-response to dietary ARA with constant DHA (1.0% total FA). Dietary ARA content for A1, A2, A3 and A4 was 0.09, 0.53, 0.69 and 1.06% total FA, respectively. (A) Heart and (b) Liver responded significantly (P < 0.0001) while (C) Brain (cerebral cortex) and (D) Retina did not. Each point represents one animal (n = 32 per panel). P < 0.05 indicates significant correlation between dietary ARA content and tissue ARA level; NS, not significant. Student's t-test used to determine significance of pairwise comparisons of mean tissue ARA content; means not sharing a common letter were significantly different (P < 0.05).

FIGURE 2.

DHA tissue dose-response to dietary DHA with constant ARA (0.67% total FA). Dietary DHA content for D1, D2 and D3 was 0.32, 0.62 and 1.01% total FA, respectively. Tissue DHA rises with dietary DHA in all tissues presented, (A) Heart, (B) Liver, (C) Brain (Cerebral cortex) and (D) Retina. Each point represents one animal (n = 24 per panel). P < 0.05 indicates significant correlation between dietary DHA content and tissue DHA level; NS, not significant. Student's t-test used to determine significance of pairwise comparisons of mean tissue DHA content; means not sharing a common letter were significantly different (P < 0.05).

ARA and DHA competition for tissue incorporation

Dietary ARA level had no effect on DHA accretion in the heart, liver or neural tissues (P > 0.05) (Figure 3, Supplemental Figure 2). Conversely, there was an apparent competition between liver ARA and diet-derived DHA (Diets D1 – D3), with D3 pigs showing lower liver ARA levels than D1 pigs (P = 0.004) (Figure 4). This competition was not detected in the heart, brain or retina, indicating that the dietary DHA content has no effect on ARA levels in these tissues (Figure 3).

FIGURE 3.

DHA tissue dose-response to dietary ARA content with constant DHA (1.0% total FA). Dietary ARA content for A1, A2, A3 and A4 was 0.09, 0.53, 0.69 and 1.06% total FA, respectively. Increasing ARA had no effect on DHA levels in (A) Heart, (B) Liver, (C) Brain (Cerebral cortex) or (D) Retina. Each point represents one animal (n = 24 per panel). P < 0.05 indicates significant correlation between dietary ARA content and tissue DHA level; NS, not significant. Student's t-test used to determine significance of pairwise comparisons of mean tissue DHA content; means not sharing a common letter were significantly different (P < 0.05).

FIGURE 4.

ARA tissue dose-response to dietary DHA content with constant ARA (0.67% total FA). Dietary DHA content for D1, D2 and D3 was 0.32, 0.62 and 1.01% total FA, respectively. Increasing DHA reduced ARA levels in (B) Liver but not (A) Heart, (C) Brain (Cerebral cortex) or (D) Retina. Each point represents one animal (n = 24 per panel). P < 0.05 indicates significant correlation between dietary DHA content and tissue ARA level; NS, not significant. Student's t-test used to determine significance of pairwise comparisons of mean tissue ARA content; means not sharing a common letter were significantly different (P < 0.05).

Discussion

The neonatal pig serves as a practical biomedical model of human infant development. Pigs are similar to humans in their metabolic responses, genetics of the FA desaturases and rates of perinatal brain growth (28,29). Their rapid postnatal growth enables facile collection of data relevant to humans over a relatively short study duration. The work presented here employed the neonatal pig model and clearly demonstrated that the heart and liver were sensitive to dietary ARA level.

The sensitivity of heart ARA to the dietary ARA content has been reported from time to time in pigs (18,20) and rats (16,17). However, the underlying metabolic mechanisms are unknown and the clinical significance largely unstudied. In the present study, heart ARA levels reflected dietary ARA content and were significantly different for pigs in all ARA groups including those with intermediate intakes, indicating that heart ARA is limited by intake, at least at high infant DHA intake levels. These observations are particularly relevant to ARA recommendations for infant formula as 0.53% (A2) is slightly above the worldwide ARA mean in human breastmilk and 0.67% (A3) is the level currently added in formula and is near the higher end of breastmilk values reported (2). ARA serves several known functions in cardiac physiology, such as providing precursor for eicosanoid signaling molecules and acting directly to modulate voltage-gated ion channel activity and cellular excitability (30). The degree to which heart ARA status influences cardiac physiology in the developing neonate, as well as the potential long term implications, remains to be determined. Nonetheless, our findings suggest that the level of ARA in formula may not yet be fully optimized.

Biosynthesis of ARA and DHA from the nutritionally essential linoleic and alpha-linolenic acids occurs through a series of elongation and desaturation reactions catalyzed by a common set of enzymes (31). These reactions occur primarily in the liver. Accordingly, liver lipids typically reflect dietary FA content, but the lipids are also influenced by the expression and regulation of the lipid-related enzymes (31,32). The brain and heart also express the FA desaturases and elongases (Elovl-1, -5 and -6) for synthesizing ARA and DHA, but these organs are not generally considered major sites of LCPUFA biosynthesis (33). Indeed, previous findings and the data presented here strongly suggest that dietary DHA content is important for the developing brain and that dietary ARA content is important for the developing heart. Emerging evidence points to a number of important roles for ARA related to brain function (34–37). In addition, considerable amounts of preformed ARA is elongated into 22:4n-6 (11 – 16% of preformed ARA) and 22:5n-6 (2 – 4% of preformed ARA) during early brain development in non-human primates (38). Some studies have indicated that ARA supplementation leads to increased neurogenesis and prepulse inhibition (39), alters brain catecholamines (40) and supports more normative values of brain N-arachidonylglycerol and ARA species of monoacylglycerols (41).

Studies with humans indicate that the rate of postnatal endogenous DHA synthesis is insufficient to support maximal DHA accretion in neural tissues (42,43). Furthermore, supplementation with ARA and DHA improves visual and cognitive performance compared with infants fed formula lacking preformed LCPUFA (3,4,6,44,45). In addition, positive correlations between LCPUFA status and visual maturation in term infants, and ARA and growth during the first year after preterm delivery, have been reported (46,47). Collectively, these observations serve as the basis for the addition of ARA and DHA to formula.

In the present study, the FA composition of the sow ration consisted of more than 50% linoleic acid with 3% alpha-linolenic acid. This composition led to a remarkably low DHA content in sow milk (2). The choice to include an MR reference group thus provided a unique opportunity to investigate the contribution of sow milk, essentially devoid of DHA but not ARA, to the LCPUFA status of the developing offspring. This treatment is representative of an extreme case of DHA-deficient breastmilk that is reflective of few human populations worldwide (2,48). Because of this, we advise caution in extending results observed with the MR group, in particular comparing MR pigs with the FR groups fed higher levels of DHA, to breastfeeding practices in humans.

In the present study, higher levels of dietary ARA were associated with increased heart and liver ARA, while brain ARA was constant and unaffected by ARA intake. Similarly, the neural tissue of the retina exhibited only a modest trend for an increase in ARA with increasing dietary ARA that was not supported by the linear regression analysis of the ARA dose-response data. The sensitivity of the heart to the dietary ARA level therefore suggests that the ARA content is influenced predominantly by preferential FA uptake and incorporation. Further, these data indicate that the rate of endogenous ARA biosynthesis may not be sufficient to match ARA accretion in the growing piglet heart. Several studies have noted a sensitivity of the nervous system and other tissues to ARA supplementation against a relatively high linoleic acid background. For example, ARA supplementation increased ARA in rat hippocampus (39) and increased in piglet striatum phosphatidylserine (40). Whelan et al. (49) noted ARA increases in many internal organ phospholipids in Syrian hamsters after an ARA supplementation relative to a background diet containing 30% linoleic acid. In humans, moderate consumption of ARA-rich foods (e.g. tropical fish, kangaroo meat) was shown to increase ARA, but not LA levels in plasma phospholipids (50,51).

The competition between n-3 and n-6 for tissue incorporation and the resulting effect on LCPUFA biosynthesis is well-studied. The relative availability of linoleic and alpha-linolenic acids, as well as their LCPUFA derivatives, drives the biosynthesis of n-3 and n-6 LCPUFA and modulates tissue levels of the two FA families in a reciprocal manner (19,52,53). Over the range of dietary ARA levels in the present study, tissue DHA was constant and unaffected by diet. Conversely, we observed a specific reduction in liver ARA with the highest DHA intakes, consistent with previous reports in pigs (18) and baboon neonates (22) consuming preformed DHA. These observations reinforce the concept that ARA and DHA in the liver are closely linked to diet because of the liver's central role in LCPUFA biosynthesis.

The sensitivity of the central nervous tissue DHA level has been previously investigated in numerous studies with term and preterm neonatal non-human primates (e.g. (21,22,54)). The brain consistently shows increased region-specific DHA accretion related to level and duration of preformed DHA feeding, and premature birth generally has a significant negative impact on DHA levels in this tissue type. In the present study, we observed higher levels of DHA in all tissues with increasing DHA content. Pigs in the high DHA group showed significantly higher levels of DHA in the cerebral cortex, retina and cerebellum compared with pigs in the low DHA group. Increases in mean DHA were 13%, 7% and 15%, respectively, for these neural tissues. Mean DHA levels in the hippocampus and globus pallidus were 17% and 15% higher in the high DHA group compared with the low DHA group; while these differences were not statistically significant, they may be of biological importance. The relative insensitivities of basal ganglia, limbic and midbrain regions to postnatal dietary DHA consumption, compared with the cerebral cortex, are consistent with data from baboons (22) and suggest that these deep brain regions may be influenced to a greater extent by the DHA status of the mother during gestation. Thus, these data support the recent dietary recommendations on total fat and fatty acids by the FAO/WHO for the provision of 200 mg/day DHA in the diets of pregnant and lactating women (55). Currently, there have only been a handful of investigations which have focused on the effects of ARA supplementation by women during pregnancy and lactation on neurodevelopment and other outcomes (56–58). Clearly, more work is needed in this area before it becomes possible to provide recommendations for ARA supplementation by expecting or nursing women.

In this study, pigs that exhibited the lowest neural tissue DHA also showed enriched tissue 22:4n-6, 22:5n-6 and 22:5n-3, with the greatest increases occurring in 22:5n-6. The increase in tissue 22:5n-6 in animals fed n-3-inadequate diets is well-characterized (59–61) and presumably occurs as an alternate substrate source for DHA-dependent functions in neural tissue. Recent examination of the frontal cortex from rat pups fed an n-3-deficient diet revealed a highly specific substitution of 22:5n-6, but not 22:4n-6 or ARA, for DHA in brain phosphatidylethanolamine and phosphatidylserine (62). Tight regulation of the “reciprocal replacement” of DHA and 22:5n-6 (59) among brain phospholipid molecular species illustrates the essential structure-function relationship between LCPUFA and outcomes such as vision or cognition. Functionally, low DHA levels in neural tissue cannot be fully compensated by increased 22:5n-6 and manifests as poorer electroretinogram performance (63) and spatial learning (64), and there is some evidence that high DHA enhances motor development (65). In the present study, the compositional difference in neural tissue 22-carbon LCPUFA between diets D1 and D3 is of particular interest. Pigs in the low DHA group had significantly lower DHA and higher 22:5n-6 in cerebral cortex, cerebellum and retina. DHA at 0.32% total FA is the level currently added to many infant formulas. Higher levels may provide further benefit for visual acuity and cognition, especially in target populations with endogenously low DHA status (5,9,11).

Changes in tissue FA composition may influence physiological function and alter safety outcomes and measures of overall development. In a separate report, we show that the dietary ARA level (0.09 – 1.0% total FA) had a negligible effect on safety-related outcomes, specifically clinical chemistries, hematology and immune function parameters in the subset of the present piglets fed 1.0% DHA (23). Likewise, there were no differences in growth among any of the dietary treatment groups and no differences in general behavior reported by daily caretakers.

In summary, results from the present study demonstrate a tissue-specific responsiveness to the dietary ARA level in rapidly growing neonatal pigs. Whereas brain ARA remained unaffected by diet ARA, heart and liver ARA levels were responsive. Dietary ARA level had no effect on DHA accretion in heart, liver or neural tissue. The magnification of heart ARA levels with increasing dietary ARA content suggests that the growing heart specifically incorporates and concentrates ARA at a rate that may exceed the biosynthetic capacity for ARA. Further investigations are warranted to determine the biological implication of heart ARA status in relation to an optimal balance of dietary n-6/n-3 for use in infant formulas.

Abbreviations

ARA

arachidonic acid

DHA

docosahexaenoic acid

FA

fatty acid

FAME

FA methyl ester

FR

formula-reared

LCPUFA

long chain polyunsaturated fatty acid

MR

maternal-reared