Studies evaluating the whole body homeostasis of polyunsaturates (PUFA) in a variety of nutritional and metabolic states demonstrate that β-oxidation as a fuel is the quantitatively dominant route of use of both linoleate and α-linolenate (1,2). In addition to complete oxidation to CO2, studies using isotopically labeled linoleate and α-linolenate show that a considerable amount of partially oxidized carbon from these two fatty acids is incorporated into newly synthesized lipids rather than being fully β-oxidized to CO2. This “carbon recycling” pathway has been most commonly studied using labeled α-linolenate injected or gavaged into suckling rats (37), but parallel results have been reported in models varying from astrocytes in culture (8), to pregnant rats (1,8a), infant rhesus monkeys (9), and humans (10).

Three features of carbon recycling from α-linolenate are consistently observed: 1) it occurs in several different organs at levels exceeding the desaturation chain elongation of the parent to the long chain n-3 PUFA by five- to 200-fold (7), 2) it is particularly evident in the neonatal period, (37,9) and 3) it occurs prominently whether or not the main longer chain PUFA are present in the diet (7,9,10) or culture medium (8). These features suggest that carbon recycling is a quantitatively important pathway in the metabolism of α-linolenate, but the function of this pathway remains enigmatic.

Linoleate recycling is largely unaffected even by extreme deficiency of all dietary n-6 PUFA (11), but whether α-linolenate recycling is affected by deficient intake of all n-3 PUFA has not yet been established. The objective of the present study was therefore to quantify the magnitude of carbon recycling of [U-13C]-α-linolenate in rat pups that were born to dams that were extremely deficient in n-3 PUFA. Maternal consumption of a diet maximally deficient in n-3 PUFA was necessary to reduce the tissue content of docosahexaenoate in the suckling rat pups as much as possible. It was hypothesized that studying carbon recycling from α-linolenate in rat pups made n-3 PUFA deficient would clarify whether this pathway was simply an “overflow” pathway occurring because of abundant availability of preformed docosahexaenoate (in milk or from neonatal synthesis) or was integral to neonatal metabolism of α-linolenate independent of n-3 PUFA status. We provided an abundant source of dietary docosahexaenoate both to mimic rat milk and to have a clear contrast in docosahexaenoate supply versus the deficient group.

Among the n-3 PUFA, it is principally docosahexaenoate that is important for normal neonatal development (12,13). Accordingly, our measure of the magnitude of carbon recycling from α-linolenate was appearance of 13C in docosahexaenoate compared with 13C recovered in newly synthesized lipids in the liver or brain of the suckling rats.


The animal handling protocol was approved by the NIAAA Animal Care and Use Committee. Three-week-old female Long-Evans hooded rats were obtained from a commercial supplier (Charles River, Portage, MI) and maintained under conventional conditions at 23 ± 1°C, with a 12-h light-dark cycle. Drinking water and custom-pelleted diets were provided on an ad libitum basis. Groups of 60 females each were immediately placed on either a control diet containing adequate levels of n-3 PUFA or a diet maximally depleted of n-3 PUFA (Table 1). A large number of dams were used as this was part of a larger experiment (14). At 11 wk of age, the dams were paired with 12-wk-old Long-Evans proven male breeders for a period of 1 wk. During this period and after mating, the females continued on their respective diets. Dams were housed three/cage until near their time of delivery at which time they were individually housed.

Table 1 Composition of the control and n-3 PUFA–deficient (Deficient) diets

When the second-generation female pups were 10 d old, one pup from each of 10 separate litters was taken randomly from dams on each of the two diets. The pups were gavaged with a single oral dose of 1 mg [U-13C]-α-linolenate dissolved in 100 μL olive oil. The pups were returned to their respective dams and then killed 24 h later by decapitation. The brain and liver were rapidly excised and stored at −80°C until analyzed. Use of a similar tracer protocol in suckling rats previously revealed clear evidence of carbon recycling from [U-13C]-α-linolenate (6,7). Two undosed rat pups from dams on each of the two diets served as controls for background 13C enrichment.

Total lipids in the brain and liver of each rat pup were extracted into chloroform:methanol (2:1, vol:vol) as previously described (6). Two aliquots of each lipid extract were prepared, one containing heptadecanoic acid as internal standard for fatty acid quantification and one without the internal standard for 13C analysis. Following saponification of the total lipid extracts, fatty acid methyl esters were prepared using BF3-methanol and analyzed by capillary gas chromatography (GC) (6). Total sterols (mainly cholesterol) from the brain were extracted and derivatized for analysis by GC (15). 13C enrichment in individual fatty acids in the liver and brain and in brain cholesterol was determined by GC-combustion isotope ratio mass spectrometry as previously described (15). The fatty acid and sterol data are given as mg/g liver or brain. The 13C enrichment data are given as ng 13C/whole liver or brain (6,10). The data are expressed as mean ± SEM for n = 10 samples/group and were compared statistically by t test.


Compared with controls, there was 97% less docosahexaenoate (mg/g) in the liver (Table 2) of the n-3 PUFA–deficient group. In the n-3 PUFA–deficient rat pups,13C enrichment in liver α-linolenate was 67% lower and was not reliably detectable in liver eicosapentaenoate (Table 2). On the other hand, the n-3 PUFA–deficient group had 2.8 times more 13C enrichment in liver n-3 docosapentaenoate, but, due to wide variation in the data, there was similar 13C enrichment in liver docosahexaenoate as in the controls.

Table 2 Fatty acid content and distribution of 13C in the liver of the control and n-3 PUFA–deficient (Deficient) rats 24 h after dosing with [U-13C]-α-linolenate

In the liver of the control rats, about twice as much 13C was recycled from 13C-α-linolenate into the sum of the principal saturated and monounsaturated fatty acids as was incorporated into docosahexaenoate. Almost half of the recycled 13C in liver lipids of controls was found in palmitate. In the liver of the n-3 PUFA–deficient group, carbon recycling from 13C-α-linolenate into other fatty acids was 45% higher (not significant), with the main increase occurring in palmitate (p < 0.05; Table 2). In the liver, the ratio of 13C incorporated into docosahexaenoate versus recycled into other fatty acids was not significantly different between the two groups.

Compared with controls, there was 82% less docosahexaenoate in the brain of the n-3 PUFA–deficient group (Table 3). α-Linolenate and eicosapentaenoate were not reliably detected in the brain of either group so 13C enrichment data were not available for these two n-3 PUFA. In the controls, four times more 13C was present in brain docosahexaenoate as in brain n-3 docosapentaenoate. In the n-3 PUFA–deficient group, 13C enrichment was 3.5-fold higher in brain n-3 docosapentaenoate but, for docosahexaenoate, did not change significantly from control values (Table 3).

Table 3 Lipid content and distribution of 13C in the brain of the control and n-3 PUFA–deficient (Deficient) rats 24 h after dosing with [U-13C]-α-linolenate

Brain sterol content was the same (about 7.5 mg/g) in both groups and was included in the calculation of the sum of carbon recycling for the brain. In the brain lipids of the controls, 31% of the 13C recycled from α-linolenate was in sterols, 55% was in saturates (myristate, palmitate, and stearate), and 14% was in oleate. There was 49% less 13C recycling in the n-3 PUFA–deficient group than in the controls (Table 3), with a greater difference in 13C labeling of sterols (−71%) than fatty acids (−43%). The lack of change in 13C incorporation into brain docosahexaenoate and the lower recycling of 13C from α-linolenate into other brain lipids combined to reduce the excess of recycling versus docosahexaenoate synthesis from 7.5-fold in the controls to 2.8-fold in the n-3 PUFA–deficient group (−63%; p < 0.05).


The markedly reduced brain and liver content of n-3 PUFA in the n-3 PUFA–deficient rat pups was expected from our previous work with this model (16,17) and demonstrated that severe deficiency of n-3 PUFA had been achieved. Our goal was to investigate whether carbon recycling of α-linolenate was an apparent “overflow” pathway somehow linked to disposal of excess α-linolenate when docosahexaenoate was abundantly available during the suckling period or whether, as with n-6 PUFA deficiency, carbon recycling of α-linolenate seems integral to its metabolism because it occurs even when tissue levels of longer chain PUFA are extremely depleted (11).

The ratio of 13C recycling into other tissue lipids versus its incorporation into docosahexaenoate is a quantitative measure of carbon recycling from α-linolenate. Among the various n-3 PUFA that could have been chosen, we focused this ratio on docosahexaenoate because of its central and irreplaceable role in the membrane phospholipids of many cell types, especially neurons. Accordingly, a high ratio of carbon recycling from α-linolenate versus incorporation into docosahexaenoate implies that docosahexaenoate synthesis is a lower priority than overflow and vice versa. As in previous studies (37), the present results show that suckling rat pups receiving a normal supply of n-3 PUFA (the controls) recycle significantly more 13C from α-linolenate into newly synthesized lipids than they put into liver or brain docosahexaenoate. Under the conditions of this study, the recycling/docosahexaenoate ratio was 2.4 in the liver and 7.5 in the brain of the controls, but varies with the time point after dosing the tracer and with the organ or species in question (7,10). In our model, severe n-3 PUFA deficiency curtailed carbon recycling from α-linolenate, but, even in the deficient rat pups, this pathway was clearly still active and matched (liver) or exceeded (brain) 13C incorporation from α-linolenate into docosahexaenoate.

Under the present conditions, carbon recycling from α-linolenate equals or exceeds conversion to docosahexaenoate even when 1) tissue docosahexaenoate is very low, 2) docosahexaenoate synthesis is typically stimulated (n-3 PUFA deficiency), and 3) the need for brain docosahexaenoate is probably higher than at any other time in the life cycle. Given analogous data from humans (10) and other models (8,8a,9), our present results add to the emerging body of evidence demonstrating that carbon recycling is an integral part of metabolism of α-linolenate and, as with carbon recycling from linoleate (11), is not simply an overflow pathway arising from an abundance or excess of docosahexaenoate.

Comparing carbon recycling from α-linolenate to synthesis of docosahexaenoate indicates little about the relative importance of either pathway. However, both pathways consume carbon from α-linolenate so this comparison does provide a reference point from which the impact of dietary or metabolic manipulation on relative synthesis of docosahexaenoate can be evaluated. The brain has a high requirement for docosahexaenoate but not for other n-3 PUFA, so the fact that 13C incorporation into lipid products of recycling from α-linolenate normally exceeds by several fold α-linolenate conversion to docosahexaenoate supports other studies showing that incorporation of preformed (consumed) rather than endogenously synthesized docosahexaenoate is likely to be an important way for the brain to obtain docosahexaenoate (18,19). Why carbon recycling occurs so actively in the suckling period and in the face of high demand for docosahexaenoate is still unclear and will require further investigation.