Main

Infants who are not breast-fed depend on commercial formulas to meet essential fatty acid requirements. Although infant formulas supply both linoleic acid (LA) and α-linolenic acid, commercially available formulas in North America do not provide AA and DHA as does breast milk. Infants fed formula have lower plasma and erythrocyte AA and DHA than infants fed breast milk (1, 2). Although evidence in vivo using stable isotope tracers (3, 4) and in vitro using radioisotopes (4) suggests that the infant who is born prematurely can endogenously synthesize AA and DHA, it is uncertain whether endogenous synthesis can meet the requirements of tissue accretion. As brain and retinal tissue contain high quantities of AA and DHA, the response of brain and retinal function to formula supplemented with and without AA and DHA has been more intensively evaluated than in many other tissues (5, 6). The effect of these fatty acids on other organs and physiologic systems, such as calcium and bone metabolism (7, 8), have not been thoroughly investigated in growing humans or animal models. Such investigations are necessary to prove supplemented formulas safe and effective in supporting whole body growth and development.

Investigation of the response of bone to AA and DHA supplementation is important as AA is the precursor to PGE2 that is hypothesized to be one of the most potent stimulators of bone formation in vivo (9). Conversely, supplementation of infant formula with DHA alone or as fish oil could compromise endogenous synthesis of PGE2 (10) and thereby limit bone formation. Prematurely born infants already have compromised bone growth and mineralization compared with term-born children (11, 12). It is unknown whether synthesis of PGE2 is reduced in bone of premature infants as a consequence of limited AA.

Few studies have investigated the potential for long-chain fatty acids to influence bone metabolism during growth. Research in adult rodents fed diets containing varied amounts of fish oil indicates that a total n-6:n-3 ratio of 3:1 or 1:1 results in higher amounts of calcium in bone compared with control, but a ratio of 1:3 demonstrated no effect. Feeding the diet with an n-6:n-3 ratio of 3:1 also resulted in significantly greater absorption of dietary calcium compared with all other diets (7). In a subsequent analysis, feeding fish oil suppressed bone turnover as indicated by urinary pyridinium cross-links and total hydroxyproline (13). These studies suggest that either addition of AA and DHA using fish oil or low ratios of n-6:n-3 fatty acids may result in higher bone calcium through two potential routes: increased calcium absorption (7) and lower bone turnover (13). Although suppression of bone turnover has application to the aging human by potentially maintaining bone mass, suppressed bone turnover could have adverse consequences to developing bones in which high bone turnover is expected to enable remodeling.

Bone remodeling occurs rapidly in growing premature infants (12, 14). Whether total n-6:n-3 ratios or AA and DHA reduce bone turnover is unclear (7, 13). If addition of AA and DHA rather than the total n-6:n-3 ratio results in low bone turnover (13), it is possible that supplementing infant formula with AA and DHA could alter bone remodeling and growth. No studies have examined the response of bone to combined AA and DHA supplementation during periods of rapid growth or using amounts of AA and DHA and outcome measurements appropriate for human investigation. Research indicates that the piglet is responsive to manipulation of dietary fat (15) and that piglet bone responds to nutrition in a manner parallel to that of human infant bone (16). However, piglets double or triple their body weight over as little as 10 to 15 d (16), whereas the prematurely born infant does so over as much as 12 wk between birth and estimated term age (11). Thus, the primary objective of this research was to establish whether addition of AA and DHA to formula affects growth and bone mineralization in piglets after 14 d. A secondary objective was to establish whether liver fatty acid composition is related to that measured in bone.

METHODS

Animals and care.

Twelve male piglets were removed from the sow at 5 d of age and transported from the Glenlea Research Unit to the University of Manitoba. All procedures were in agreement with the Guide for the Care and Use of Experimental Animals (17) and approved by the University of Manitoba Protocol Management and Review Committee.

Upon arrival, piglets were randomized to receive either 14 d of standard diet or a fatty acid supplemented diet from d 10 to 24 of life. The first 5 d were used to teach the piglets to lap the liquid formula and to ensure that all piglets tolerated the formula at full strength without symptoms of diarrhea. The standard formula was suitable as a control diet because it supports rapid growth (16). In addition, it provides dietary fat (37 g/L) in similar quantities as that measured in formula designed for prematurely born infants (Enfalac Premature 20 with 34 g fat/L and Enfalac Premature Plus with 41 g fat/L, Meadjohnson Nutritionals, Canada). The amount of fat in the formula was lower than in sow's milk (62 g/L as measured in milk from the Glenlea Research Unit) but contained enough LA (11.5 to 11.8 g/L) to support growth. The standard formula had a total n-6:n-3 fatty acid ratio of 4.9:1.0, but the fatty acid supplemented formula contained 0.5% wt/wt of the fat as AA (RBD-ARASCO: 40.6% AA; Martek Biosciences Corp., Columbia, ML) and 0.1% wt/wt DHA (RBD-DHASCO: 40.0% DHA; Martek Biosciences Corp.), resulting in a total n-6:n-3 fatty acid ratio of 4.8:1.0. The fatty acid composition of the formulas is given in Table 1. The standard formula contained 55 g/L protein as skim milk powder and whey protein, 101 g/L carbohydrate, 37 g/L fat (70% canola and 30% corn oils), and 2.9 g/L calcium and 2.1 g/L phosphorous from calcium carbonate and minerals in skim milk powder. The formula energy (957 kcal/L) and LA (10.9% of energy) were within recommended limits for healthy growing piglets between 3 and 10 kg as set by the National Research Council (18). Formulas were isocaloric with equal amounts of fat and fed at 400 mL·kg−1·d−1 divided into equal amounts at 0900, 1300, 1700, and 2100 h for the duration of the study. Piglets were housed individually in stainless steel cages with ambient temperature maintained at 29 to 30°C.

Table 1 Fatty acid composition of standard and fatty acid supplemented piglet formula
Full size table

Growth.

Weight was measured to the nearest gram daily in the nonfed state by using a scale with an animal weighing program (Mettler-Toledo Inc., Hightstown, NJ), and average daily weight gain was calculated (g·kg−1·d−1). Snout-to-tail length was measured to the nearest millimeter by using a nonstretchable measuring tape at the end of study; pigs were anesthetized before measurement of calcium absorption.

Calcium absorption.

After 14 d of study, piglets were anesthetized using sodium pentobarbital (30 mg/kg); food was withheld for 12 h before anesthesia. After deep anesthesia was achieved, an incision was made in the abdomen, and the duodenum was located. Distal to the ligament of Treitz, a 10-cm section of duodenum was ligated at each end before infusion of a Tris-HEPES buffer (10 mmol; pH 7.4) according to the method described by Weiler et al. (16). In brief, the buffer contained 45Ca (18 MBq/L; Amersham Ltd., Mississauga, Canada), 100 mg of Poly R-478 (Sigma Chemical Co.-Aldrich Ltd.), and CaCl2 (2 mmol/L). Samples were obtained every 5 min over 30 min, after which the loop was excised and the remaining buffer collected. Samples were analyzed for 45Ca by scintillation counting (Model LS 6000TA; Beckman Instruments Inc., Fullerton, CA), Poly-R 478 measured by UV spectrometry as described by Stahl et al. (19), and absorption calculated as previously reported (16).

Tissue samples.

Immediately after the perfusion study, blood was taken into heparinized syringes followed by sodium pentobarbital overdose. Plasma was obtained by centrifugation at 2000 ×g for 20 min at 4°C. Plasma was centrifuged again to remove any remaining red blood cells and stored at −80°C until analysis. Liver, heart, spleen, and kidneys were excised and weighed to the nearest 0.1 g. Liver tissue was then flash-frozen in liquid nitrogen before storage at −80°C until analyzed for fatty acids. Urine was obtained using bladder puncture and stored at −20°C until analysis for minerals and PGE2.

Fatty acid analysis.

Total lipids in liver were extracted using 1:2 methanol to chloroform according to the method of Folch et al. (20). Before extraction of lipid from bone, 1 g of cortical bone free of periosteum and marrow was frozen in liquid nitrogen followed by crushing with pestle and mortar. The crushed bone was then homogenized with 1:2 methanol to chloroform for 10 min by using a Polytron (Brinkmann Instruments, Rexdale, Ontario, Canada) and then extracted according to Folch et al. (20). Lipid extracts from bone were separated into phospholipid and FFA by using thin-layer chromatography (Whatman K8 Silica Gel 80A) with a solvent phase of petroleum ether, ethyl ether, and acetic acid. Phospholipid and FFA were visualized using dichlorofluorescein and extracted using chloroform:methanol:acetic acid:water (50:339:1:10). Crude lipid extracts from liver tissue and phospholipid and FFA from bone were transmethylated in 1 mL of methanolic HCl (3N, Supleco Inc., Bellefonte, PA) at 100°C for 15 min. Fatty acid methyl esters were separated by gas-liquid chromatography (Hewlett Packard 5890A) and using hydrogen as the carrier gas. The gas chromatograph was equipped with a 30-m Supelcowax 10 column (Supelco Inc., Bellefonte, PA), an autosampler, 3392A integrator, and flame ionization detector. Samples were injected (0.5 μL) at an initial temperature of 175°C, and the oven temperature was increased to a final temperature of 235°C at a rate of 3°C per minute. Fatty acid methyl esters were identified by comparison with retention times of Supelco 37 component FAME mix (Supelco Inc.) and expressed as percent of total lipid (% wt/wt).

Biochemical assessment of bone metabolism.

Plasma osteocalcin, ICTP, and carboxy-terminal PICP were analyzed by RIA (INCstar Corp., Stillwater, MN). The average coefficient of variation (CV) for triplicate analysis of all samples was 4.7% for osteocalcin, 5.1% for ICTP, and 2.6% for PICP. Urinary and femur calcium and phosphorous were wet ashed in concentrated nitric acid for 24 h before dilution with deionized water (5% nitric acid) and measurement using a Varian Liberty 200 ICP emission optical spectrometer (Varian Canada, Mississauga, Canada). The CV for triplicate analysis of each sample was <1%. Urinary PGE2 was analyzed in diluted samples by using an ELISA (Cedarlane Laboratories, Hornby, Canada). The average CV for triplicate analysis of all samples was 5.5%. Calcium, phosphorous, and PGE2 were corrected to creatinine as determined by the Jaffe method (procedure no. 555; Sigma Chemical Co.-Aldrich Ltd., Oakville, Canada). The average CV for triplicate analysis of creatinine in all samples was 5.4%.

BMC and BMD.

After tissues were removed, the abdominal cavity was closed with suture to maintain tissue depth. Piglet carcasses were then transported to the dual energy x-ray absorptiometer (QDR4500W; Hologic Inc., Waltham, MA). Triplicate scans were completed to determine BMC and BMD of the whole body (software version V8.16a:5), lumbar 2–4 by using the low-density spine program, and left femur by using the subregion array hip program. All scans were performed with the piglet in the anteroposterior position with limbs extended. Whole body BMC was corrected to body length (g/cm) and weight (g/kg) to account for potential differences in size. The average CV was 3.8% for whole body, 3.2% for lumbar 2–4, and 1.2% for femur BMC. Femurs were then excised and freed of soft tissue for measurement of weight, length, and mineral content.

Statistical analysis.

All data are mean ± SD unless otherwise stated. Differences observed between groups were detected by t tests with values of p < 0.05 accepted as significantly different. Relationships between liver fatty acids with growth, BMC and BMD, bone fatty acids and urinary PGE2 were detected using 2-tailed Pearson correlation analysis.

RESULTS

Piglet characteristics at baseline and end of study are presented in Table 2. No differences in piglet weight were evident at baseline (standard 3.0 ± 0.3 versus fatty acid 2.7 ± 0.2 kg). Over the 14 d of study, average rate of weight gain was not significantly different between groups (standard 80.7 ± 0.8 versus fatty acid 83.7 ± 0.7 g·kg−1·d−1) but resulted in a cumulative and significant difference in absolute weight by d 14 (standard 7.9 ± 0.6 versus fatty acid 10.0 ± 0.9 kg, p < 0.05). Whole body length was not different (standard 61.1 ± 1.5 versus fatty acid 63.5 ± 3.7 cm), and no difference was observed in formula intake or organ weights (data not shown).

Table 2 Plasma and urine biochemistry after 14 d of standard or fatty acid supplemented formula Data are mean ± SD;n = 6 per group.
Full size table

After 14 d of feeding, whole body BMC was significantly greater in the fatty acid supplemented piglets (standard 112 ± 12 versus fatty acid 137 ± 13 g, p < 0.01), but no differences were observed in BMC of lumbar 2–4 (standard 1.04 ± 0.11 versus fatty acid 1.23 ± 0.21 g) or femur (standard 4.56 ± 0.25 versus fatty acid 5.01 ± 0.55 g). Whole body BMC remained significantly higher after correction with body length (standard 1.8 ± 0.2 versus fatty acid 2.2 ± 0.1 g/cm, p < 0.01) but not weight (standard 14.3 ± 1.4 versus fatty acid 15.3 ± 0.5 g/kg). Excised femurs, free of soft tissue, were similar between groups in weight (standard 36.3 ± 2.1 versus fatty acid 38.5 ± 3.9 g) and length (standard 8.8 ± 0.1 versus fatty acid 8.9 ± 0.4 cm). Calcium and phosphorous (mg/g bone) were not different between groups. Whole body, average lumbar 2–4, and femur BMD were significantly greater in the piglets fed the fatty acid supplemented diet (Fig. 1).

Figure 1
figure 1

Whole body, left femur, and average lumbar 2–4 BMD as measured by dual energy x-ray absorptiometry after 14 d of feeding formula supplemented with or without AA and DHA. Data are mean ± SD;n = 6 per group. ★p < 0.05.

Full size image

Plasma and urine biochemistry were similar between groups (Table 2) at the end of study with no differences observed in osteocalcin, PICP, or ICTP, indicating that osteoblast activity, collagen synthesis, and bone turnover were not affected by the dietary manipulation. Urinary PGE2 was not affected by dietary manipulation, although the measurement was highly variable. Urinary and plasma calcium and phosphorous were similar between groups. In addition, calcium absorption between groups was similar when corrected to intestinal weight (standard 12.7 ± 2.4 versus fatty acid 11.9 ± 4.5 mmol·30 min−1·g intestine−1).

Liver fatty acid content and bone phospholipid and FFA were similar between groups with the exception of LA (18:2 n-6) and AA (20:4 n-6) in liver total lipid and bone FFA. Liver LA was lower (p < 0.05) and AA was higher (p < 0.05) in the fatty acid supplemented group (Table 3). Liver LA and AA (% wt/wt total lipid) were negatively related (r = −0.91, p < 0.001). Similarly, bone free LA was lower (p < 0.05) and free AA higher (p < 0.05) in the supplemented group (Table 3). No evidence was found that addition of 0.1% DHA altered liver or bone fatty acid composition. Thus, correlation analysis was only carried out between liver LA and AA with outcome measurements of growth, BMC and BMD, bone free LA and AA, and urinary PGE2/creatinine. Liver AA was positively related (p < 0.05) to growth, free AA acid in bone, BMC of whole body, lumbar and femur, BMD of lumbar and femur regions and negatively related (p < 0.05) to free LA in bone (Table 4). Inverse relationships were observed when liver LA was substituted for liver AA as the independent variable (Table 4). Liver AA (r = 0.55, p < 0.01) but not LA (r = −0.52, p = 0.08) was related to urinary PGE2.

Table 3 Selected n-6 and n-3 fatty acids measured in liver and bone of piglets fed formula with and without supplemental AA and DHA Data are mean ± SD;n = 6 per group. * Significant difference between groups (p < 0.05) by t test.
Full size table
Table 4 Multiple correlation analysis between liver LA and AA with outcome measurements of growth, bone free LA and AA, and BMC and BMD Values are Pearson correlation coefficients;n = 12. *p < 0.05. **p < 0.01.
Full size table

DISCUSSION

Feeding formula supplemented with 0.5% AA (% wt/wt total fat) and 0.1% DHA results in greater whole body weight, BMC, and BMD in piglets after only 14 d compared with feeding a standard diet already proven to support rapid growth (16). Feeding infants formula supplemented with similar amounts of AA and DHA as used in the present study results in erythrocyte AA and DHA most similar to that observed in infants fed breast milk (1, 21). The goal of establishing the appropriate amount of AA and DHA to be added to infant formula is to mimic the developmental outcomes resulting from feeding breast milk. Although addition of AA and DHA to formula does support similar blood fatty acid profiles to those observed with breast milk feeding, the benefit to brain and eye development remains controversial (5, 6). The present study demonstrates a positive effect of dietary AA to elevate liver and bone AA in addition to supporting consistently higher BMD in all sites measured. The relationships observed between AA status and growth and with BMC and BMD justify further investigations to confirm the relationships followed by explanation of the underlying mechanisms.

This is the first report of greater whole body BMC and BMD in response to dietary AA and DHA in growing piglets. Investigations in adult rodents demonstrate increased femur weight and calcium content after feeding diets supplemented with fish oil or purified eicosapentaenoic acid (7, 22). Greater whole body BMC and BMD reported in the present study were observed despite the already high amounts of dietary calcium and phosphorous in the diet. However, femur and lumbar BMC were not different between groups. The discrepancy between the results of whole body and regional BMC could reflect alterations in bone architecture. Watkins et al. (23) observed changes in tibia formation rate and total bone and medullary cavity area without observing any change to tibia length, ash weight, or calcium content in chicks fed diets with varied fatty acid composition over 42 d. Similar whole body BMC/weight, femur weight, length, and calcium/phosphorous content between piglet groups corroborate the thesis that changes in bone architecture rather than increased amounts of calcium in bone resulted in the higher BMD. The greater whole body BMC and BMD could have resulted from combined effects on cortical and trabecular bone as well as different effects in woven versus lamellar bone (24) in addition to greater whole body growth.

Bone growth is a result of multiple factors including autocrine and paracrine regulation of bone metabolism. Alterations in bone AA contribute by effecting substrate for endogenous synthesis of PGE2. The fatty acid composition of dietary fat is reflected in bone (25, 26) and liver fatty acids (10). Although both liver and bone AA were elevated in the piglets after dietary supplementation with AA, urinary PGE2 was not different between feeding groups. It is possible that the small increase in dietary AA and the resulting AA status did not provide sufficient competition for cyclooxygenase between AA and eicosapentaenoic acid or that the study design was insufficient in length. When data from both groups were combined, liver AA was associated with urinary PGE2. The relationship with synthesis of PGE2 at the level of bone was not determined and, thus, requires further investigation to identify relationships among dietary AA, AA status, PGE2, and BMD.

The mechanisms responsible for higher BMD could not be explained by osteoblast (osteocalcin) or osteoclast (ICTP) activity as determined in the present study or by intestinal absorption of dietary calcium. Lower bone resorption, as indicated by urinary pyridinium after feeding fish oils to rodents (13), was not evident in this study, suggesting that low amounts of AA and DHA do not impair bone resorption and, thus, do not impair bone remodeling or growth. In addition and in contrast to Coetzer et al. (8), no difference in intestinal absorption was observed between formula groups in the present study. Coetzer et al. (8) measured calcium uptake in brush border and basolateral membrane vesicles that represent entry and extrusion of calcium in enterocytes but not intracellular transport. The lack of difference was likely due to the already high rate of calcium absorption observed in these and other (16) growing piglets.

The higher weight of the fatty acid supplemented group suggests that low amounts of AA (0.5%) and DHA (0.1%) do not compromise growth. Wainwright et al. (27) observed reduced growth of mouse pups after dietary supplementation with up to 9% DHA (wt/wt), regardless of dietary AA or total n-6:n-3 ratios. The amount of dietary DHA fed to the piglets was not great enough to significantly alter liver fatty acid composition, implying that the observations of higher weight, whole body BMC, and BMD at all sites measured were due to dietary AA. Each outcome variable for growth and bone was positively related to liver AA and negatively related to liver LA. This relationship confounds the relationship between higher AA status as indicated by liver AA and the resulting greater BMD. Heavier piglets, as a result of supplementing dietary AA, could have larger bones and thus higher BMD due to an effect of increased weight-bearing load or an artifact of dual energy x-ray absorptiometry measurements in which larger bones have artificially higher BMD. However, whole body BMC was higher relative to body length, an indicator of bone length, implying that the bones were either wider or that the mineral was more densely packed. In support of this, femur BMC, weight, and length were similar between groups, suggesting that the higher BMD resulted from changes in bone architecture rather than larger bone size. Inclusion of bone morphology in future studies will further classify higher BMD as greater mineralization of bone or simply larger bones with normal mineralization patterns.

Ratios of n-6:n-3 fatty acids in infant formula have not been established that result in accretion of AA and DHA in tissues and developmental outcomes parallel to those observed with breast milk (1). Feeding infants formulas with low amounts of purified AA and DHA results in similar erythrocyte membrane fatty acid profiles compared with those of infants fed breast milk regardless of term or preterm birth (1, 2, 28). As plasma and erythrocyte fatty acids are not predictive of fatty acid composition of brain (29) or perhaps of other tissues like bone, it is difficult to ascertain the necessity of supplementing infant formulas with AA and DHA. In the current study in piglets in which growth and BMD were elevated in response to dietary AA supplementation, it is unknown whether a higher BMD confers any developmental benefit or detriment. In addition, no reports of the effect of supplemental AA or DHA on bone growth and mineral content exist in human infants. Existing reports of human infants with no evident change to visual acuity despite normal blood fatty acid profiles (1, 5) indicate that supplementation of formulas with AA and DHA is complicated by variable responses accompanied by lack of true indicators of tissue accretion. Further investigation regarding the effects of AA and DHA on whole body growth and development is necessary to ensure that supplementation of formulas is effective in mimicking the developmental milestones experienced by infants fed breast milk.