Main

LCP and particularly AA (C20:4n-6) and DHA (C22:6n-3) are found in high concentrations in structural lipids of the CNS(1–3). DHA in particular increases markedly during brain growth(4, 5). LCP are biosynthesized in animal tissues by desaturation and elongation of the 18-carbon essential FA, LA (C18:2n-6) and LNA (C18:3n-3)(6). This process is thought to occur at a sufficient rate in healthy human adults. During fetal life LCP may predominantly derive from the mother via placental transfer. A facilitated placental transfer is supported by the observation that at birth cord plasma lipids contain higher LCP and lower parent essential FA levels than maternal plasma(7). After birth, the newborn infant, when fed a formula containing ample amounts of LA and LNA, but unlike breast milk devoid of LCP, exhibit rapidly declining plasma and red blood cell LCP concentrations(8). Breast milk-fed infants have higher plasma, red cells, and brain LCP concentrations than bottle-fed infants(9–13). A decreased DHA status in infancy is associated with poorer retinal development and visual function(13–16). Infants fed with DHA-supplemented, compared with standard, formulas have better neurodevelopmental quotients at 4 mo(17) and have higher cognitive scores at 12 mo(18).

The conversion of LA and LNA to LCP in mammals remains largely unexplored(19, 20), and data in human adults are limited(21–24). The consistent observations of declining LCP after birth pose the question as to whether the newborn infant and especially the small preterm infant is capable of biosynthesizing AA and DHA.

In this study, we investigated whether the preterm infant can synthesize LCP by administering carbon-13-labeled LA and LNA and measuring the carbon-13 enrichment of their metabolic products in plasma phospholipids with a GC-IRMS method(25).

METHODS

Five preterm infants with a birth weight between 835 and 1605 g were studied (Table 1). They received a standard preterm formula (Nenatal, Nutricia, Zoetermeer, The Netherlands), which contained 4.4 g of fat/100 mL, 12.5 and 1.2 weight% of total FA as LA and LNA, respectively. LCP were not added. Infants received total parenteral nutrition during the first 7 d of life; oral feeding was started after d 7 and progressively increased. The infants reached full feeds by about d 15. All infants were fed exclusively the study formula, and the isotope study was performed at approximately 1 mo of age (Table 1). Patients studied were all in stable clinical condition. The isotope study consisted in the simultaneous and continuous oral administration of [U-13C]LA and[U-13C]LNA (Martek, Columbia, MD) for 48 h.

Table 1 Clinical and experimental data
Full size table

Chemical purity was 96.7% for LA and 93.9% for LNA, with impurities being mainly the 16-carbon atom unsaturated and polyunsaturated FA. Isotopic enrichment was greater than 98% for both tracers.

The first patient received [U-13C]LA only, and the remaining four patients were given both tracers. The isotopes were mixed with the formulas and were added in amounts so that the isotope represented approximately 2.5% of the dietary LA and LNA. All infants received the formulas via continuous intragastric infusion (and thus the isotope) as is customary in our neonatal intensive care unit. Blood was obtained at baseline and at 12, 24, 48, 72, 96, and 168 h after the start of the isotope. Blood (0.5 mL) was collected in EDTA-containing Vacutainers and centrifuged at 4 °C for 10 min at 900× g. Plasma was stored with pyrogallol as an antioxidant at-20 °C until analysis. Plasma lipids were extracted with chloroform:methanol (2:1) after the addition of cholesteryl nonanoate and cholesteryl heptadecanoate, trinonanoin and triheptadecanoin, L-α-phosphatidylcholine dinonanoyl and L-α-phosphatidylcholine diheptadecanoyl, and nonanoic and heptadecanoic acid for the quantification of sterol esters, triglycerides, phospholipids, and FFA, respectively. Lipid classes were separated by thin layer chromatography developed with heptane/di-isopropyl ether/acetic acid (60/40/3, vol/vol) and visualized with 1,2-dichlorofluorescein by comparison with authentic standards. For the methylation of each lipid class we used 3M dry HCl in methanol. The separation and identification of FA methyl esters from the plasma lipid classes were performed by capillary GC (model 5890 II; Hewlett Packard, Amstelveen, The Netherlands) equipped with a fused silica column (Supelcowax 10, 60 m × 0.20-mm inside diameter, 0.2-μm film thickness; Supelco, Leusden, The Netherlands), a flame ionization detector (280 °C), and a split-splitless injector used in splitless mode. FA were identified by comparing retention times with known standards (Nu Chek Prep, Elysian, MN)(26). The 13C enrichment of the plasma phospholipid FA was measured by GC-IRMS(25). The GC-IRMS system used was a Isochrom III (VG Isotech, Middlewich, Cheshire, UK). Results obtained from the instrument are based on the 13C enrichment of the CO2 of the combusted sample and are expressed in APE compared with a reference CO2 gas of known mass 45/44 ratio. Results obtained from the patients are also given in APE that expresses an increase in enrichment above baseline after the isotope infusion. A calibration curve was constructed by adding different amounts of enriched LA and LNA to their naturally enriched counterparts. No attempt was made to correct the enrichments of the FA methyl esters for the changes in enrichment due to relatively minor contribution of the methyl group added during the transesterification of the FA (this is partially corrected during the calculation of the APE). Data inFigures 1 and2 are expressed as APE of the CO2 derived from the complete combustion of each FA. For the calculations described below we introduced the following correction factors. During the chain elongation reaction of the essential FA 2 carbon atoms are added at each elongation step. Under the assumption that the carbon atoms used for the chain elongation have a 13C content close to the natural abundance value, and that this value remains constant throughout the study period, we multiplied the measured 13C enrichment of the C20 FA by 1.1(10%) and of the C22 FA by 1.2 (20%). These corrected values may give better estimates of the synthesized LCP.

Figure 1
figure 1

13C enrichment of selected n-6 polyunsaturated FA in plasma phospholipids of preterm infants after [13C]LA and[13C]LNA. Time curves for LA (C18:2n-6), γ-LNA(C18:3n-6), dihomo-γ-linolenic acid (C20:3n-6), and AA (C20:4n-6) are depicted from left to right and top to bottom. Subject numbers are as in the tables and Figure 2.

Full size image
Figure 2
figure 2

13C enrichment of selected n-3 polyunsaturated FA in plasma phospholipids of preterm infants after [13C]LA and[13C]LNA. Time curves for LNA (C18:3n-3), EPA(C20:5n-3), docosapenaenoic acid (C22:5n-3), and DHA(C22:6n-3) are depicted from left to right and top to bottom. Subject numbers are as in the tables and Figure 1.

Full size image

We used the approach described by Emken et al.(23) for the calculation of the amount of AA and DHA produced. We calculated the area under the curve for both the APE values and the plasma concentration of the labeled FA in plasma phospholipids over time. This was done by assuming a linear decay for the enrichments of the FA after the 168-h time point.

The activity of the chain elongation and desaturation processes was estimated from the area under the curve of the APE values as well as from the maximal increase in enrichment. The amount of tracer found in phospholipid AA and DHA in comparison with the total dose of the tracer administered was used to estimate the amount of plasma phospholipid AA and DHA endogenously produced.

The project was approved by the local Ethical and Scientific Committees. The use of stable isotopes is generally accepted for use in infants(27). Written informed consent was obtained from both parents. Only infants whose mothers decided not to breast feed were approached and asked to participate in the study.

RESULTS

Total plasma phospholipids and their polyunsaturated FA contents did not change significantly during the 7-d study period (data not shown). The enrichments of selected plasma phospholipid n-6 and n-3 polyunsaturated FA are depicted in Figures 1 and2, respectively.

Both tracers were rapidly incorporated in plasma phospholipids, although neither LA nor LNA reached a plateau at the 48-h time point corresponding to the end of the isotope administration. The mean enrichment of phospholipid LA at 48 h was 1.5 APE that is 60% of the enrichment of the diet(Fig. 1). Enrichments of the metabolic products showed a time delay and progressively lower values for each metabolite. The enrichment curve of γ-LNA (C18:3n-6), was, however, relatively similar to that of its metabolic precursor LA (maxima at 48 h and rapid decline by 72 h) with a mean maximum of 0.9 APE (60% of the APE of LA at 48 h). Only in patient 2 the difference appeared more pronounced than in the other four patients(Fig. 1). Dihomo-γ-linolenic acid (C20:3n-6) peaked at around 72 h from the administration of the isotope and AA even later, probably between 72 and 96 h. LNA enrichment (Fig. 2, top left panel) also showed a very rapid increase during the administration of the label and a rapid decrease after stopping the labeled diet. The maxima of the APE value (mean 2.6 APE) was close to 100% of the enrichment in the diet. DHA enrichment peaked between 72 and 96 h, similar to that of AA.

The APE values at maxima were also rather similar between AA and DHA. The APE of all n-6 and n-3 FA increased markedly in all patients after the administration of the tracers, and this clearly shows that all infants were capable of actively synthesizing the long chain polyunsaturated FA from their dietary precursors. Remarkably, differences among patients were rather small.

The areas under the curve calculated in the APE values are given inTable 2. It is clear that each step from the parent compound (LA and LNA) to the end product (AA and DHA) caused a further reduction in enrichment. The enrichment of comparable products in the elongation/desaturation process, as AA and EPA, showed similar enrichments.

Table 2 Plasma phospholipid FA enrichment*in 5 infants receiving [U-13C]18:2n6 and[U-13C]18:3n3
Full size table

The dose of [13C]LA and [13C]LNA administered to the patients was 48.1 ± 2.8 and 4.5 ± 0.2 mg·kg-1·48 h, respectively. The total amount of 13C-labeled AA and DHA in plasma phospholipids calculated as the area under the curve of the concentrations of the labeled FA was 4.37 ± 1.52 and 1.02 ± 0.33 mg of13 C-labeled FA. These figures correspond to 6.05 ± 2.26% and 14.07 ± 4.20% of the total dose of 13C tracer given as LA and LNA, respectively.

DISCUSSION

We report a newly developed approach which enabled us to measure in vivo the biosynthesis of LCP with stable isotopes. The study shows that small preterm infants are capable of converting both LA and LNA into LCP. We were also able to measure the 13C enrichment of all major metabolites of the essential FA including C18:3n-6, which is the δ6 desaturase product of LA and thought to be the limiting step in EFA metabolism(9).

A second important technical achievement was the simultaneous measurement of both the n-6 and n-3 FA. After we studied the first patient with[13C]LA only, and saw no coelution in the GC-IRMS between different FA of the n-6 and n-3 families, we decided to add the second tracer. The simultaneous measurement of n-6 and n-3 FA metabolism represents an important advantage, because with this approach it will be possible to investigate in future studies, for example, whether dietary AA affects both the n-6 and the n-3 FA synthesis in the same individual.

The doses of 25 mg·kg-1·d-1 of [3C]LA and 2.5 mg·kg-1·d-1 of [13C]LNA were chosen because they produced 13C enrichments of about 2.5 APE for both LA and LNA that were present in the formula in a 10:1 ratio. Similar enrichments of the dietary parent essential FA were meant to simplify the interpretation of the data. Different enrichments of n-6 and n-3 FA, occupying analogous positions in the synthetic pathway, could then be interpreted in the view of different metabolic characteristics of the individual FA and not of the different degree of labeling in the diet. The dose of the tracers was also“relatively” high to maximize the chance of detection of the most diluted metabolites. In spite of similar enrichments in the diet, the enrichments in plasma phospholipid LA and LNA were markedly different. Phospholipid LNA reached the enrichment of the diet by 48 h, but LA was only 60% of the dietary value. Lower enrichments of LA could result either from a faster turnover, or from “dilution” in a larger body pool.

The major finding of this study is that the healthy preterm infant at approximately 1 mo of age can desaturate and elongate LA and LNA into n-6 and n-3 LCP, respectively. Several studies measuring plasma polyunsaturated FA levels indicate that the infants fed formulas devoid of LCP exhibit lower plasma levels than their human milk-fed counterparts(8). Decreasing plasma LCP concentrations are also described in preterm infants fed their own mother's milk(28). Whether this results from a complete inability by the small preterm infant to synthesize LCP(circulating LCP would then result from the recirculation of LCP acquired during fetal life) or whether the requirements are so high that the levels decrease despite active LCP synthesis was not known. Postmortem studies with liver microsomes from human neonates born close to the end of gestation demonstrate significant δ6 and δ5 desaturase activities(29). Apart from the diversities between the in vitro conditions and studies in healthy subjects in vivo as in the present study, it is worth mentioning, that the infants who died and were studied by Poisson et al.(29) were at least 5 wk more mature than our infants, who could thus exhibit more reduced δ6 and δ5 desaturase activities. Although our data are mainly qualitative in nature, in our view the isotopic profiles of LA and GLA(Fig. 1, top panels) do not show marked differences. These differences are not, at least, larger than those between other precursor molecules and their products (Fig. 1 andTable 2). This observation suggests that the δ6 desaturation may not be a rate-limiting step in our patients. An important question is whether the n-6 and n-3 pathways have an equal or different activity. When looking at the mean enrichment of EPA in plasma phopholipids at the 48-h time point (1.18 ± 0.14 APE) and comparing it with that of AA (0.17 ± 0.02 APE), one could conclude that the n-3 pathway is more active. This difference in enrichment between “corresponding metabolites,” although important, could, however, result from different sizes in body pools. We chose to express our data in APE and not in absolute plasma concentrations of the labeled FA, because FA enrichments may be a better probe of the intracellular metabolism and thus of the synthetic process. The lower plasma phospholipid concentrations of EPA in comparison with DHA, for example, could result from a lower affinity of EPA than DHA for incorporation during phospholipid synthesis or remodeling. According to this view, low plasma concentrations of[13C]EPA do not exclude that significant amounts of labeled EPA (which is not incorporated in phospholipids) are being converted into DHA in the cells.

Another approach to estimate the activity of the enzyme systems is to determine the production of the compound as indicated by the area under the curve. The area under the curve of the enrichments as shown inTable 2 may indicate that the chain elongation and desaturation processes are not markedly different for the n-3 and n-6 series up to the level of EPA and AA.

We are unable with the present data to reach a conclusion regarding the activity of the conversion in both series. To do so one should have data on the enrichment of the intracellular pools and the turnover of all individual FA. With the data presented we cannot also calculate the total amount of AA and DHA produced by the infants. Plasma phospholipid represents only a portion of the circulating lipids, and AA and DHA in plasma triglycerides or sterol esters might have different kinetics. Our data allow us to only estimate the amount of dietary LA and LNA converted into plasma phospholipid AA and DHA. We found that 6.05 ± 2.26% and 14.07 ± 4.20% of the total dose of[13C]LA and LNA were converted into plasma phospholipid AA and DHA, respectively. A different percentage of conversion of the parent FA into plasma phospholipid AA and DHA is not contradictory to the finding of similar APE areas between EPA and AA, as the percentage converted also depends on the total intake of the parent FA. Assuming that the same percentages of conversion are applicable to the unlabeled intake of the parent FA, one can calculate that, with a mean intake of 1480 mg of LA in 2 d (as it was the case in our infants), 90 mg·kg-1 of AA is produced in plasma phospholipids. In analogy for the n-3 FA with a mean dietary intake of 126 mg of LNA, 17.8 mg·kg-1 DHA is produced. As these estimates are based only on plasma phospholipids, the values will be higher if the other lipid classes are included. These figures on the other hand may be an overestimation because the continuous infusion of the tracers in our study design determines a “prolongation” of the production, therefore a“prolonged” half-life of these compounds.

We chose to study preterm infants receiving a formula with a 10:1 ratio of LA and LNA because this ratio is often found in human milk lipids. However, we cannot draw any conclusion on the adequacy of this ratio. New studies comparing different ratios should be performed.

The duration of our studies was far longer than any other published work, and we show that at 168 h the plasma phospholipid LCP were still highly enriched. Calculations in previous studies in adults, with blood sampling limited to 48 h after the isotope administration, underestimate the percentage of LCP biosynthesis(23). We now sample blood up to 10 d after administration of the isotope.

In conclusion, we have measured the simultaneous conversion of LA and LNA into their long chain derivatives in preterm infants. The studies were performed with stable isotopes. Our data demonstrate that the small preterm infant can effectively convert both parent essential FA into LCP at 1 mo of life. Studies will be possible with this methodology to study whether different clinical and pathologic conditions are associated with enhanced or reduced LCP synthesis. Appropriate isotopic modeling will be necessary to“quantify” these processes.