The variety of commercially available n-3 PUFA sources raises the question of which is the most beneficial. In KO, PUFAs are mainly in PL form whereas in fish oils, PUFAs are mainly in TG form, leading to claims that n-3 LC-PUFAs are better absorbed as KO than fish oils . This study compared sources of n-3 PUFAs that differed in the amount, type, and/or structural form of n-3 PUFAs.
PUFA digestibility was influenced by the structural form. In our study, DHA content was ~2-4 times higher in KO compared to the other sources of n-3 PUFAs. However, apparent digestibility was greater (P = 0.009) in SO compared to KO-fed rats. This may be due to ~50% of DHA being in PL form in KO, whereas DHA was entirely in TG form in SO. DHA was also in TG form in the other fish oils (MO and TO), yet apparent DHA digestibility was not significantly different than KO-fed rats. In human studies, absorption efficiency of DHA from KO was similar to MO [27, 28]. In our study, higher DHA digestibility in SO, but not rats fed MO or TO may be due to differences in the positional distribution of DHA on the TG molecule. Christensen et al.  reported DHA on the sn-2 position resulted in higher absorption compared to the sn-1 and sn-3 position of the TG molecule. While, the fatty acid position on the lipid structure may be a factor affecting digestibility. Overall total lipid digestibility was lowest (P < 0.02) in KO-fed rats which may be accounted for by fatty acids in KO being associated with PL. Amate and Ramirez  observed reduced absorption of n-3 LC-PUFAs in rats fed PL-rich pig brain. In contrast, infants and pre-term infants fed LC-PUFA PL as egg lecithin improved fat absorption [31, 32]. In the brain, DHA are mainly associated with phosphatidylethanolamine and phosphatidylserine , whereas the predominant PL in egg is phosphatidylcholine . Further studies are needed to clarify whether different PLs influence fatty acid digestibility. This is important because the source of PUFAs that provides high digestibility also increases tissue deposition.
Brain PUFA Deposition
Epidemiological studies have linked low DHA to poor neural development in infants and to cognitive decline in the aging individuals . In animals, decreased DHA in the developing brain resulted in deficits in neurogenesis, neurotransmitters, visual function, and learning . In humans, PL contributes to ~25% of the dry weight of the brain . However, in our study DHA was better incorporated when consumed in TG than PL form. TO with the highest DHA in TG form resulted in the highest (P < 0.001) brain DHA deposition. KO with the highest DHA in PL form did not result in the highest brain deposition due to reduced DHA digestibility. The higher DHA digestibility associated with SO also resulted in the highest (P < 0.001) brain DHA deposition. Talahalli et al.  observed rats fed EPA+ DHA increased brain DHA, but produced only a small increase in EPA due to inefficient tissue uptake. Of the marine oils, TO with the lowest EPA content resulted in lower (P = 0.01) brain EPA deposition compared to rats fed MO or FO. Dietary FO contained no detectable EPA; therefore, EPA deposition in the brain tissue of FO-fed rats suggested that de novo synthesis of the n-3 LC-PUFAs occurred in the brain.
Brain metabolism, function, and structure also depend on adequate concentrations of ARA . In our study, the CO and FO diets had no detectable ARA, but the highest LA content. In turn, brain LA content was highest (P < 0.001) in CO and FO-fed rats. Efficient metabolism of LA to n-6 LC-PUFAs in the brain was indicated by ARA deposition in rats fed CO and FO. Brain ARA content in CO-fed rats was comparable to rats fed SO and TO with the highest dietary ARA-TG content. On the other hand, conversion of ALA to n-3 LC-PUFAs was less efficient. Feeding rats FO, but not CO, increased brain EPA. This may be due to higher ALA content in the FO (14.6 ± 2.1 mg/g) than the CO (0.1 ± 0.01 mg/g) diet. Brain DHA deposition in rats fed CO and FO diets containing no DHA indicated conversion of ALA to n-3 LC-PUFAs. Several studies reported increased brain DHA in rats fed ALA [39–42]. In our study, rats fed SO and TO with pre-formed DHA had significantly higher brain DHA deposition than FO or CO-fed rats having no dietary DHA. The results indicated greater brain incorporation of pre-formed DHA compared to conversion of ALA to n-3 LC-PUFAs. KO and MO also provided pre-form n-3 LC-PUFAs; however, DHA content provided by MO was low and KO had reduced DHA digestibility. Due to limited biosynthesis of DHA in the mammalian brain, DHA deposition in the brain relies on the diet and on the release of DHA synthesized from ALA in the liver [33, 43]. Therefore, the effect of feeding different sources of n-3 PUFA on liver fatty acid composition was evaluated.
Liver PUFA Deposition
The amount of dietary fatty acids affected liver deposition. Rats fed FO with the highest ALA content also had the highest (P < 0.001) liver ALA deposition. Rats fed SO with the highest dietary EPA-TG resulted in the highest (P < 0.001) liver EPA deposition. Of the oil sources, KO had the highest DHA content; however, liver DHA deposition was highest (P < 0.001) in rats fed TO and SO. TO provided the highest DHA in TG form and SO-fed rats showed greater DHA digestibility than KO. Liver DHA incorporation as TO was ~1.5 times greater in the form of TG than PL in TO and SO-fed rats, whereas DHA was equally incorporated as TG and PL in KO-fed rats. Song and Miyazawa  reported that rats fed DHA in PL form had lower liver DHA incorporation compared to DHA fed in TG form.
The liver is a major site of LC-PUFA biosynthesis. Therefore, n-3 LC-PUFA deposition is not only dependent on the intake of pre-formed EPA and DHA, but it also depends on intake of the precursor, ALA. Feeding rats CO containing ALA and no n-3 LC-PUFAs resulted in DHA, but no detectable EPA liver deposition. DHA has a structural role, whereas EPA is preferentially utilized for β-oxidation or eicosanoid synthesis . Similarly, FO contains no pre-formed n-3 LC-PUFA. Feeding rats FO with the highest ALA content resulted in detectable DHA as well as EPA deposition in the liver. Still, conversion of ALA to n-3 LC-PUFAs resulted in lower liver EPA and DHA deposition compared to consumption of SO and TO with pre-formed n-3 LC-PUFAs. Talahalli et al.  reported rats needed to consume 12.5 times more ALA to produce the same liver n-3 LC-PUFAs incorporation as rats consuming EPA and DHA. This may be due to preferential use of ALA in β-oxidation . Also, several enzymes required for synthesis of LC-PUFAs in the desaturation-elongation pathway are inefficient .
On the other hand, there was efficient liver conversion of LA to n-6 LC-PUFAs. This was indicated by similar liver ARA deposition in rats fed CO containing no detectable ARA compared to SO containing pre-formed ARA. Additionally, liver ARA deposition was significantly higher in CO than FO, KO or MO-fed rats. This may have occurred because FO had the highest ALA content followed by MO and KO. ALA and LA compete for the same enzymes to form their respective LC-PUFAs with ALA having higher affinity for the rate-limiting Δ-6 desaturase . This has important implications because decreased ARA reduces synthesis of pro-thrombotic and inflammatory 2-series eicosanoids by COX II. However, our study showed no significant differences in the 2-series eicosanoids, PGE2 and TXB2 metabolites among the diet groups. Feeding rats FO, which is high in ALA, decreased liver ARA, but this was not accompanied by increased EPA deposition. Conversely, feeding rats SO increased liver EPA, but this was not accompanied by decreased ARA deposition. Furthermore, measurement of urinary PGE2 and TXB2 metabolites provided an indicator of systemic rather than tissue change. Responses to n-3 PUFA intakes are not uniform among tissues. For example, liver fatty acid composition may vary depending on the uptake of chylomicron remnants from the circulation . Also, the liver not being a major storage organ is less responsive to the diet than adipose tissues . Recent research suggests changing the fatty acid profile of adipose tissue results in health benefits by altering lipid metabolism and adipokine secretion [48, 49]. Therefore, the effect of consuming different n-3 PUFA sources on adipose tissue PUFA composition was determined.
Adipose Tissues PUFA Deposition
As the chief site for lipid storage, the fatty acid profile of adipose tissue reflected the diet. Rats fed KO with the highest amount of EPA had the highest (P < 0.001) adipose tissue EPA deposition. Rats fed FO with the highest amount of ALA content had the highest (P < 0.001) adipose tissue ALA deposition. Talahalli et al  reported rats fed increasing doses of ALA resulted in a linear increase in ALA accumulation in the adipose tissue. In our study, rats fed FO increased n-3 PUFA deposition in the adipose due to the high dietary ALA content. Additionally, rats fed KO and MO increased adipose n-3 PUFA deposition. Others also reported feeding rats fish oils increased n-3 LC-PUFA incorporation in the adipose tissue [50–53]. Increasing adipose tissue n-3 PUFA incorporation has been observed to reduce adipose mass . However, in our study, there was no significant reduction in the adipose mass of rats fed different n-3 PUFA sources.
The adipose tissue is another site of LC-PUFA biosynthesis. Efficient conversion of LA to n-6 LC-PUFAs in gonadal adipose was indicated by rats fed FO and CO containing no ARA having similar tissue ARA deposition to rats fed MO, KO or SO containing pre-formed ARA. In retroperitoneal adipose, rats fed CO had similar tissue ARA to rat fed marine oils. FO also resulted in tissue ARA, but lower (P < 0.001) amounts than rats fed CO or marine oils. This may be due to FO having the highest ALA content. High ALA competitively inhibits LA, the precursor of ARA.
Conversion of ALA to n-3 LC-PUFAs was less efficient. In our study, there were no detectable EPA and DHA in the adipose of CO-fed rats. However, rats fed FO containing ALA, but no DHA, resulted in DHA deposition in gonadal and not in retroperitoneal adipose tissue. This suggested conversion of ALA to n-3 LC-PUFAs was less efficient in retroperitoneal adipose compared to gonadal adipose. Muhlhausler et al.  observed female rats fed n-3 PUFA had different degrees of responsiveness to PUFA deposition and in turn, tissue PUFAs induced changes in lipogenic gene expression. Furthermore, gene expression changes were more pronounced in the omental compared to retroperitoneal adipose tissue.
Rats fed FO and marine oils had an adipose tissue n-6/n-3 ratio of 1:1-1:2. An n-6/n-3 ratio of ~1:1 in tissues has been reported to reduce atherosclerosis due to the inhibition of systemic and vascular inflammation in apolipoprotein E-deficient mice . While the optimal n-6/n-3 ratio in tissues has not been defined, increasing tissue unsaturation has been considered to be health beneficial. However, the higher tissue unsaturation needs to be considered since a greater number of double bonds increases tissue susceptibility to lipid peroxidation.
DHA is particularly susceptible to lipid peroxidation due to its high degree of unsaturation . Our study showed rats fed TO, the highest source of DHA of the fish oils or SO with the high DHA digestibility resulted in lower (P = 0.005) serum TBARS compared to all diet groups, except FO-fed rats. However, there were no significant differences in RBC TBARS among the diet groups. Oxidative stress occurs when accumulation of oxidative products overwhelms the body's antioxidant capacity. TAC measures endogenous antioxidant, dietary antioxidant, and interactions between antioxidants . In our study, serum TAC was not significantly different among the diet groups. Circulating TBARS and TAC may not reflect the tissue concentration.
The liver is a primary target for oxidative stress-induced damage in oil-fed rats . Rats fed MO had the highest (P = 0.03) liver TBARS. The TBARS assay is accepted as an index of oxidative stress; however, this method quantifies MDA-like compounds and does not specifically measure lipid peroxidation . Furthermore, it is the imbalance of oxidants and antioxidants that leads to oxidative stress. Reena and Lokesh  reported that the lipid peroxides generated by feeding rats PUFAs was partly nullified by the capability of PUFAs to increased liver antioxidant enzymes, Cu/Zn SOD, Mn SOD, CAT, and GSH-Px. Extracellular and cytosolic forms of SOD depend on Cu/Zn and in the mitochondria SOD depends on Mn. SOD detoxifies superoxide radicals giving rise to hydrogen peroxide (H2O2). H2O2 is a potent free radical generator and can generate hydroxyl radicals, which induce lipid peroxidation of cell membranes. Therefore, to prevent accumulation of H2O2, it is important that enhanced SOD activity be accompanied by increased cellular CAT and GSH-Px . Ruiz-Gutierrez et al.  reported increased n-3 PUFAs enhanced the efficiency of the antioxidant defense system.
In our study, rats fed SO and TO had a P:S ratio of 3:1 compared the range of 2:1 to 1.2:1 in the other treatment groups. However, liver TAC was higher in MO (P = 0.03) than TO or SO-fed rats. According to Venkatraman et al. , different lipids had differential effects on the antioxidant defense system at the molecular level. Mice fed MO and KO had higher liver SOD, CAT, and GSH-Px mRNA expression than rats fed CO . In our study, there were no significant differences in SOD or CAT gene expression in rats fed different sources of n-3 PUFAs. Rats fed fish oils increased GSH-Px mRNA expression compared to CO or FO-fed rats, although this was not statistically significant. Demoz et al.  reported that mice provided high doses of EPA enhanced liver antioxidant enzyme activities. In our study, EPA was in TG form in SO, whereas in KO ~27% of EPA was in PL form. In turn, liver deposition of EPA as PL was ~50% in KO-fed rats compared to ~40% EPA the SO-fed rats. Additionally, KO also contains high amounts of the powerful antioxidant, astaxanthin [61, 62]. However, in the present study there were no significant differences in SOD, CAT or GSH-Px mRNA expression in rats fed KO compared to the other diet groups.