Main

For a number of years, the accepted pathway for endogenous synthesis ofω3 and ω6 long chain, polyunsaturated fatty acids has been a series of alternating desaturation and elongation steps beginning withΔ6 desaturation of C18:3ω3 and C18:2ω6 and terminating with Δ4 desaturation of C22:5ω3 and C22:4ω6, respectively, to C22:6ω3 and C22:5ω6(Fig. 1). Recently, an alternative pathway independent ofΔ4 desaturation was demonstrated in isolated rat hepatocytes(1, 2) and human skin fibroblasts(3). This pathway consists of elongation of C22:5ω3 to C24:5ω3 and C22:4ω6 to C24:4ω6 followed byΔ6 desaturation of the C24 fatty acids and peroxisomalβ-oxidation of C24:6ω3 and C24:5ω6, respectively, to C22:6ω3 and C22:5ω6.

Figure 1
figure 1

Pathways of ω3 and ω6 fatty acid metabolism. The classical pathway involving Δ6, Δ5, and Δ4 desaturation followed by elongation is depicted by solid arrows. Putative alternative steps consistent with the data reported are depicted by broken arrows.

Full size image

If the alternative pathway, which includes two Δ6 desaturation steps, is operative in vivo, C18:3ω3 and C18:2ω6 will compete not only with each other but also with C24:5ω3 and C24:4ω6 for Δ6 desaturation. This highlights the potential importance of the dietary intake of both C18:3ω3 and C18:2ω6. For example, a high intake of C18:3ω3 relative to C18:2ω6 may result in successful competition with C18:2ω6 for the initial Δ6 desaturation step but, if the same Δ6 desaturase is involved, inhibit the final Δ6 desaturation step. Thus, delineation of the specific pathway(s) involved in endogenous synthesis of long chain, polyunsaturated fatty acids has important implications for the optimal C18:3ω3 and C18:2ω6 contents of infant formulas and may help resolve the current controversy of whether infant formulas should be supplemented with C22:6ω3 and/or other long chain polyunsaturatedω3 and ω6 fatty acids.

To date, in vivo utilization of the alternative pathway for endogenous synthesis of long chain polyunsaturated ω3 and ω6 fatty acids by either human adults or infants has not been demonstrated. The objective of the study reported here was to determine whether human infants use this alternative pathway for endogenous conversion of C18:3ω3 and C18:2ω6 to longer chain, more unsaturated products and if other alternative metabolic steps also are involved.

METHODS

Recently, we reported the effects of dietary C18:3ω3 intake on the desaturation and elongation of C18:2ω6 and C18:3ω3, respectively, to C20:4ω6 and C22:6ω3 by both term and preterm infants(4). Infants participating in these studies were selected randomly from a larger group of term(5, 6) and preterm infants(7) enrolled in a prospective, double-blind study of the biochemical and clinical effects of C18:3ω3 intake. Term infants were selected randomly to receive formulas with 0.4, 1.0, 1.8, or 3.2% of total fatty acids as C18:3ω3 from shortly after birth until 56-wk postmenstrual age; preterm infants received formulas with either 1.0 or 3.2% of total fatty acids as C18:3ω3 for the same period. The C18:2ω6 content of all formulas was approximately 16% of total fatty acids.

At 43- and 56-wk postmenstrual age, infants selected at the time of initial randomization to participate in the conversion studies were given a single dose of 15-25 mg/kg [U-13C]18:3ω3 and 25-40 mg/kg[U-13C]18:2ω6 (>95% purity; Martek Biosciences Corp., Columbia, MD) with a regular feeding, and blood samples were obtained 8, 12, and 24 h later for determination of the enrichment of the [M + 18] isotopomers of all detectable ω3 and ω6 fatty acids of the plasma phospholipid fraction. Plasma was separated, lipid was extracted with 4:1 (vol/vol) hexane/isopropanol(8), and the phospholipid fraction was separated by silica column gel chromatography(9). The fatty acids of this fraction were saponified, converted to pentafluorobenzyl esters(10), and eluted on capillary columns using a Hewlett-Packard 5890 gas chromatograph. Two columns (DB-Wax (60 m), J and W Scientific, Folsom CA, and SP 2380 (60 m), Supelco, Bellefonte, PA) were required to resolve all of the putative intermediates of the pathway. Mass spectrometry was performed on a Hewlett-Packard 5989A quadrupole mass spectrometer. Signals of the [M] and [M + 18] isotopomers of all detectableω6 and ω3 fatty acids were determined with selected ion monitoring under negative chemical ionization. Data are reported as tracer/tracee ratios.

Separate runs with up to 4 μL splitless injection conditions for the gas chromatograph and with high electron multiplier settings for the mass spectrometer were necessary to detect isotopic enrichments of several fatty acids, including the C24 fatty acids. Fatty acids were identified by comparison of retention times with those of fatty acid standards and by mass determination. C18, C20, and C22 fatty acid standards were purchased from Sigma Chemical Co. (St. Louis, MO) and Nu Chek Prep (Elysian, MN). Some C24 fatty acid standards were generously provided by Dr. Howard Sprecher, Ohio State University, Columbus, OH.

The fractional rates of conversion of C18:3ω to C22:6ω3 and C18:2ω6 to C20:4ω6 as well as the fractional rates of incorporation of C22:6ω3 and C20:4ω6 into the plasma phospholipid fraction by term infants fed C18:3ω3 intakes of 0.4, 1, and 3.2% of total fatty acids have been reported(4). We report here the tracer/tracee ratios of the [M + 18] isotopomers of all detectable intermediates of C18:3ω3 and C18:2ω6 metabolism at 43-wk postmenstrual age in both term (n = 10) and preterm (n = 6) infants who received the formula with 3.2% of fatty acids as C18:3ω3. These data are qualitative but specific and unambiguous. The presence in the phospholipid fraction of the [M + 18] isotopomer of any ω3 or ω6 fatty acids indicates that this fatty acid was synthesized endogenously from[U-13C]18:3ω3 or [U-13C]18:2ω6, respectively, and incorporated into plasma phospholipids.

All aspects of the study were approved by the Institutional Review Board for Human Subject Research for Baylor College of Medicine and Affiliated Hospitals. Written informed consent was obtained from at least one parent of all infants enrolled.

RESULTS

The mean tracer/tracee ratio (± SEM) of the [M + 18] isotopomers of all detectable plasma phospholipid ω3 and ω6 fatty acids of term(n = 10) and preterm (n = 6) infants 8, 12, and 24 h after administration of [U-13C]18:3ω3 and [U-13C]18:2ω6 are shown in Tables 1 and 2. Because different doses of both uniformly labeled precursors were administered, these data have been normalized to doses of 25 mg/kg of both [U-13C]18:3ω3 and[U-13C]18:2ω6. This was achieved by multiplying the observed tracer/tracee ratio by 25/the actual dose of [U-13C]18:3ω3 or[U-13C]18:2ω6 given. This changes neither the nature of the data nor the conclusions drawn from it but tends to eliminate the intersubject variability in tracer/tracee ratios that is secondary to the dose of labeled precursor administered.

Table 1 Normalized enrichment of the [M + 18] isotopomers* of ω3 and ω6 fatty acids in plasma phospholipids of term infants (n = 10) after administration of[U-13C]18:3ω3 and[U-13C]18:2ω6
Full size table
Table 2 Normalized enrichment of the [M + 18] isotopomers* of ω3 and ω6 fatty acids in plasma phospholipids of preterm infants (n = 6) after administration of [U-13C]18:3ω3 and[U-13C]18:2ω6
Full size table

[M + 18] isotopomers of C18:3ω3, C18:4ω3, C20:3ω3, C20:4ω3, C20:5ω3, C22:5ω3, C22:6ω3, C24:5ω3, and C24:6ω3 were detected in the plasma phospholipid fraction of all infants as were [M + 18] isotopomers of C18:2ω6, C18:3ω6, C20:2ω6, C20:3ω6, and C20:4ω6. [M + 18] isotopomers of C22 and C24 ω6 fatty acids were not detected in any infant. The [M + 18] isotopomer of C22:4ω3, presumably an elongation product of C20:4ω3, was detected in the plasma phospholipid fraction of six of the 10 term infants and in the plasma phospholipid fraction of all preterm infants. The [M + 18] isotopomer of C24:4ω3, presumably an elongation product of C22:4ω3, was detected in the plasma phospholipid fraction of two term infants; neither the[M] nor [M + 18] isotopomer of this fatty acid was detected in the plasma phospholipid fraction of the other eight term infants or the preterm infants.

Enrichment of the [M + 18] isotopomers of the precursor fatty acids, C18:3ω3 and C18:2ω6, generally was highest 8 h after administration and declined continuously thereafter, whereas enrichment of the[M + 18] isotopomer of most multiply desaturated and elongated products increased from 8 to 24 h. The time course of the change in isotopic enrichments of the [M + 18] isotopomers of all detectable ω3 andω6 fatty acids of a representative term infant is shown in Figure 2.

Figure 2
figure 2

Tracer ([M + 18])/tracee ([M]) ratios of ω3 (left panel) and ω6 (right panel) polyunsaturated fatty acids of a representative term infant 8, 12, and 24 h after administration of[U-13C]18:3ω3 and [U-13C]18:2ω6.

Full size image

Assuming that only desaturation and elongation steps are involved in conversion of C18:3ω3 and C18:2ω6 to longer chain, more unsaturated fatty acids, the detected [M + 18] isotopomers of ω3 andω6 fatty acids are most consistent with the pathways outlined in Figure 1.

DISCUSSION

The qualitative data reported above confirm recent reports that both term and preterm infants can synthesize C22:6ω3 and C20:4ω6 endogenously from C18:3ω3 and C18:2ω6(11–13). In addition, they provide the first indication that the alternative pathway for endogenous synthesis of C22:6ω3 described recently in isolated rat hepatocytes(1, 2) and human skin fibroblasts(3) is operative in term as well as preterm infants. Although these data do not constitute conclusive proof that the alternative pathway is operative, the presence in plasma of the [M + 18] isotopomers of C24:5ω3 and C24:6ω3 after administration of[U-13C]18:3ω3 coupled with the in vitro data constitutes a strong argument for its being involved.

The fact that [M + 18] isotopomers of C24 ω6 fatty acids were not detected does not mean necessarily that C24 fatty acids are not involved also in endogenous synthesis of C22:5ω6. Rather, the very low enrichment of[M + 18] C20:4ω6 in both term and preterm infants(Tables 1 and 2) suggest that the dose of[U-13C]18:2ω6 administered was insufficient to permit detection of [M + 18] isotopomers of either C22 or C24 ω6 fatty acids. Alternatively, because data beyond 24 h post-administration of the precursor are not available, the possibility that [M + 18] isotopomers of these fatty acids might have appeared beyond 24 h cannot be excluded. Further, because only the phospholipid fraction was analyzed, the presence of [M + 18] isotopomers of these fatty acids in other fractions also cannot be excluded. Certainly, difficulty in phospholipid synthesis and/or remodeling would limit(if not exclude) ability to detect these or other endogenously synthesized fatty acids using the approach described.

The extent to which the alternative pathway involving β-oxidation of C24:6ω3 to C22:6ω3 versus Δ4 desaturation of C22:5ω3 to C22:6ω3 is used in endogenous synthesis of C22:6ω3 is not clear. The data reported here simply demonstrate presence of [M + 18] isotopomers of C22:5ω3, C24:5ω3, C24:6ω3, and C22:6ω3. They do not permit definitive quantitation of the rates of conversion of C22:5ω3 to C24:6ω3 versus C22:6ω3 or of the rates of conversion of C22:5ω3 versus C24:6ω3 to C22:6ω3. However, it is important to note thatΔ4-desaturase activity has never been assayed directly; thus, the alternative pathway provides a mechanism for conversion of C22:5ω3 to C22:6ω3 without Δ4 desaturation.

The primary importance of the alternative pathway involving C24 fatty acids derives from the fact that not only C18:3ω3 and C18:2ω6 but also C24:5ω3 and, probably, C24:4ω6 undergo Δ6 desaturation. If only one enzyme is involved, the four fatty acids rather than only C18:3ω3 and C18:2ω6 compete for Δ6 desaturation. Because the concentrations of C18:3ω3 and, especially C18:2ω6, are much greater than those of C24:5ω3 and C24:4ω6, competition for desaturation of C24:5ω3 and C24:4ω3 is likely to be particularly powerful. Such competition provides a logical explanation for the frequently mentioned phenomenon of substrate inhibition of desaturation and elongation of C18:3ω3 and C18:2ω6. This also is a likely explanation for our observation that the plasma phospholipid content of C22:6ω3 was only minimally higher in infants fed 3.2 versus 0.4% C18:3ω3 despite markedly higher plasma phospholipid levels of C18:3ω3, C20:5ω3, and C22:5ω3(5). Further, observations by others suggest that an even higher intake of C18:3ω3 may actually inhibit conversion of C18:3ω3 to C22:6ω3(14, 15). This inhibitory effect of C18:3ω3 on endogenous synthesis of 22:6ω3 highlights the difficulty of achieving plasma and tissue lipid levels of C22:6ω3 similar to those of breast-fed infants by altering only the C18:3ω3 content and/or the C18:2ω6/C18:3ω3 ratio of the diet.

Involvement of the alternative pathway for C22:6ω3 synthesis also is a viable explanation for the low plasma and tissue lipid levels of C22:6ω3 in patients with a variety of peroxisomal disorders(16). Because the last step in C22:6ω3 synthesis by this pathway is thought to involve peroxisomal β-oxidation of C24:6ω3 to C22:6ω3, this step is likely to be blocked in patients with peroxisomal disorders. Indeed, Carnielli et al.(17) found that the plasma phospholipid fraction of two infants with Zellweger's syndrome contained barely detectable amounts of the[M + 18] isotopomer of C22:6ω3 after administration of[U-13C]18:3ω3.

The presence of [M + 18] isotopomers of C20:3ω3 and C20:2ω6 as well as C18:4ω3 and C18:3ω6 suggests that C18:3ω3 and C18:2ω6 are both elongated and desaturated (see Fig. 1). The elongation products of C18:3ω3 and C18:2ω6 have been detected previously in in vitro(18–21) as well as in vivo studies(12, 22, 23), including one(12) of two recent studies(12, 13) which used essentially the same method used here to demonstrate endogenous synthesis of C22:6ω3 and C20:4ω6 in infants. However, the importance of these elongation products remains to be established. In some studies, C20:3ω3 and C20:2ω6 appear to undergo Δ8 desaturation, respectively, to C20:4ω3 and C20:3ω6(18–20, 22); in others, these fatty acids appear to have been oxidized and/or retroconverted to the parent fatty acids which, in turn, were converted to longer chain, more unsaturated fatty acids via the classical pathway(21, 23). The data reported here are compatible with either or both of these possibilities.

Also of potential importance is the fact that the [M + 18] isotopomer of C22:4ω3 was detected in the plasma phospholipid fraction of all preterm infants and six of the term infants. This fatty acid, presumably the chain elongation product of C20:4ω3, could be desaturated to C22:5ω3(see Fig. 1); however, this would require Δ7 desaturation, which has not been described previously. Alternatively, C22:4ω3 could either be retroconverted by β-oxidation to C20:4ω3 or elongated to C24:4ω3. C20:4ω3, of course, could undergo Δ5 desaturation to C20:5ω3 and C24:4ω3 could undergo Δ9 desaturation to C24:5ω3, the C24 substrate ofΔ6-desaturase. Our data do not permit delineation of the specific steps involved but it is important to note that the [M + 18] isotopomer of C24:4ω3, the elongation product of C22:4ω3, was detected in the plasma phospholipid fraction of two infants. In the remaining infants, neither[M] nor [M + 18] C24:4ω3 was detected.

In toto, the qualitative data reported here represent the most thorough in vivo assessment of ω3 and ω6 fatty acid metabolism currently available. Although further work is required to confirm some of the putative alternative steps of the metabolic pathway outlined in Figure 1 and to quantify all of the steps, the data demonstrate conclusively that C24:5ω3 and C24:6ω3 synthesized endogenously from C18:3ω3 are present in the plasma phospholipid fraction of both term and preterm infants. This fact plus the in vitro demonstration in rat hepatocytes(1, 2) and human fibroblasts(3) that C24:6ω3 is converted to C22:6ω3 argues strongly that human infants use the same or a similar alternative pathway for conversion of C18:3ω3 to C22:6ω3. Although not demonstrated by the data reported here, it is very likely that similar steps are involved in conversion of C18:2ω6 to C22:5ω6. The data also suggest that additional alternative steps may be involved in endogenous synthesis of long chain polyunsaturated ω3 and ω6 fatty acids. Moreover, they provide important clues for the design of future studies to clarify and, perhaps, quantify specific steps of the pathway.