Omega-3 polyunsaturated fatty acids (PUFAs) such as eiocosapentaenoic acid (EPA) and docosapentaenoic acid (DHA) are well known bioactive dietary compounds that are found particularly in marine-derived food sources such as e.g. fatty fish, seaweed, shellfish, microalgae and krill. Since 1970, regular consumption of fish (preferably fatty fish) has been stated to have several positive effects on cardiovascular health[1–5]. The American Heart Association dietary guidelines for healthy individuals proposes consumption of at least two servings of fish per week which should yield an intake of approximately 400-500 mg EPA and DHA, and they recommend an even higher intake of omega-3 fatty acids to patients with documented coronary heart disease[6, 7]. However, in large parts of the world the consumption of fish is considered to be inadequate and fish oil (FO) from anchovy, sprat, herring and salmon as a source of DHA and EPA has become widely used as a dietary supplement. The reported health benefits of FO have however led to an increased demand that may endanger natural resources of fish, and krill oil (KO) has recently emerged on the market as an alternative source of omega-3 PUFAs. Most FO on the market today have their omega-3 PUFAs incorporated into (triacylglycerols) TAG or in ethyl esters. However in KO are the majority of these omega-3 PUFAs esterified in phospholipids (PLs)[8–11]. KO has been stated to be a safe source of EPA and DHA that like other marine-based oils is able to efficiently raise the plasma levels of EPA and DHA[12–16]. However, the structural differences in the PUFA-rich lipid molecules may affect the distribution in cellular lipid fractions and tissue uptake and thereby promote different regulatory effects on lipid homeostasis[17]. KO also contains astaxanthin, which due to its anti-oxidative effect, might enhance the stability of the omega-3 PUFAs in the oil and thereby preserve them from lipid oxidation[10].

Intake of EPA and DHA has been shown to improve cardiovascular health by regulating lipid and glucose metabolism by acting as ligands for several nuclear transcription factors (e.g. peroxisome proliferator-activated receptors (PPARs) -α, -β/δ, and -γ and sterol regulatory element-binding protein 1 (SREBP-1))[18–20]. EPA and DHA also have anti-inflammatory effects due to their conversion to less inflammatory signaling molecules at the expense of production of pro-inflammatory molecules from arachidonic acid (for reviews see[21, 22]).

The liver is a central metabolic organ that regulates both circulating blood lipids and glucose levels by catabolism as well as synthesis of lipids and carbohydrates. Marine-derived omega-3 PUFAs have previously been shown to modulate the gene transcription profile in liver to enhance lipid degradation and decrease VLDL secretion (for review, see[23]). Recently, also KO was shown to modulate the transcriptional profile in mouse and rat liver and to affect plasma and liver lipids in mice[24–28].

Gene expression is regulated in the intestine in response to different metabolic conditions in order to cope with changes in nutrient load and content, to signal satiety and other stimuli to the rest of the body and to keep the intestinal defense barrier against pathogens intact. In addition, the intestine contributes to the plasma lipoprotein profile by absorbing lipids for chylomicron synthesis and being responsible for a significant part of the HDL production in the body[29, 30]. To the best of our knowledge, so far no study has addressed the effects of KO on regulation of gene expression in the small intestine, although KO was recently shown to attenuate inflammation and oxidative stress in colon in an experimental rat model of ulcerative colitis[31].

The aim of this study was to compare the effects of two of the major sources of omega-3 supplements on the market today, FO and KO, when supplemented to a Western-like high-fat diet.

Equal amount of FO and KO (6% by weight) were added to a high fat diet, and the effects on plasma and liver lipids and gene regulation in liver and intestine were measured. In spite of the markedly lower omega-3 PUFA content in KO, both dietary supplementations raised the content of omega-3 PUFAs in plasma as well as in liver phospholipids to a similar extent. However, FO more efficiently lowered plasma lipids and this decrease was associated with accumulation of lipids in liver. In contrast, KO was less efficient in lowering plasma lipids with less, if any, sign of TAG accumulation in the liver. These different effects by FO and KO can at least in part be ascribed to differential regulation of gene expression in liver and intestine and different effects on VLDL secretion.

Omega-3 PUFA supplementation and adequate intake of dietary omega-3 PUFAs is stated to have numerous beneficial health promoting effects including TAG lowering in plasma in humans, which have also been found in numerous animal studies[23]. The main outcome was the demonstration of a different metabolic regulation by KO and FO. FO significantly decreased several plasma lipid parameters (TAG, PLs and cholesterol) compared to HF, which was associated with lipid accumulation in liver. However, the post hoc test did not demonstrate any significant changes in total plasma TAG, PL or cholesterol in the KO supplemented group compared to HF, while plasma NEFA was significantly reduced. In addition, there were no significant differences between the two marine oil groups except for significantly lower VLDL cholesterol content in the FO group.

FO and KO differently regulated expression of genes involved in lipid degradation and synthesis. While FO provoked a strong PPARα activation like response in liver and intestine, these effects were weak, or absent, in the KO group, which instead mainly decreased the expression of genes involved in cholesterol and fatty acid synthesis.

The major factors contributing to these differences are likely the different content and structure of the omega-3 PUFAs in FO and KO. In this study we intentionally supplemented the high fat diet with equal amounts of oil (6% w/w) to compare the effects of commercially available omega-3 supplements. Although the omega-3 content is about double in FO, the concentration of omega-3 PUFAs in plasma was very similar between the two marine oil groups suggesting a higher bioavailability of KO, or that the content of omega-3 PUFAs in KO in our study is enough to obtain maximal plasma concentrations of omega-3 fatty acids. Incorporation of omega-3 PUFAs in PLs has been suggested to give these lipids a higher bioavailability[14, 42, 43], in fact, PLs enriched from KO was shown to be even more efficient in lowering plasma cholesterol than KO itself when given to rats fed a high cholesterol diet[44].

Unfortunately, in the present study was not enough plasma to analyze the fatty acid composition in TAG and PLs separately so it is not clear whether the similar levels of omega-3 PUFAs measured in plasma of the two groups are due to a specific enrichment in the PL fraction of the KO feed mice, which may be the case. However, the ratio of EPA/DHA (EPA > DHA) in total plasma was very similar in the FO and KO groups and rather closely resembled the ratio in the diet, indicating that these fatty acids are not modified during uptake in the intestine and transport in plasma. It is however evident that the omega-3 fatty acids are redistributed between the TAG and PL fractions following uptake and incorporation into liver lipids and that EPA is further metabolized to DHA, thereby markedly increasing the DHA/EPA ratio. DHA is mainly esterified into PLs in liver of the HF mice (constituting >90% of the omega-3 fatty acids in this fraction) and surprisingly reaching almost 50% of the DHA levels in the FO and KO groups. FO decreased total omega-6 fatty acids in total plasma lipids and liver PLs, while KO mainly lowered longer chain omega-6 fatty acid species in plasma and liver PLs. Both treatments thereby significantly increased the omega-3/omega-6 ratio in plasma compared to HF alone. Since KO specifically lowered the arachidonic acid content in the liver PL fraction and in plasma more efficiently than FO, KO may have a more potent anti-inflammatory effect. Such an effect by KO was seen in another FO and KO comparing supplementation study on rheumatoid arthritis with balanced amounts of DHA and EPA[45]. The decreased arachidonic acid levels are also likely to decrease the amounts of 2-arachidonoylglycerol (2-AG), a potent signaling lipid, as previously described[17]. Krill powder, which is also rich in omega-3 containing PLs, was shown to decrease another arachidonic acid derived endocannabinoid, anandamide (N-arachidonoylethanolamide, AEA), and its related metabolites palmitoylethanolamide (PEA) and oleoylethanolamide (OEA) in plasma of obese men[46]. KO has further been shown to be protective in a rat model of inflammatory bowel disease in which KO seemed to both act as an anti-inflammatory as well as anti-oxidative agent in the colon of these animals[31].

The plasma TAG lowering effect of EPA and DHA have been proposed to be at least partly mediated via PPARα, similar to fibrates, by stimulating the β-oxidation systems that would drain the liver of fatty acids and thereby decrease the production of VLDL particles. Our data, however, show a 3-4 fold TAG accumulation in livers from FO supplemented mice in spite of an apparent PPARα activation in both liver and intestine. Similar, but less pronounced, changes were seen also for fatty acids in plasma and liver PLs. The apparent ‘transfer’ of PLs from plasma to liver may reflect increased membrane synthesis, e.g. proliferation of peroxisomes (at least for the FO group) as indicated by increased expression of peroxisomal enzymes and mitochondria as indicated by slightly increased citrate synthase expression (see Figure 2). Another possibility is that liver PLs increase as a result of decreased HDL production that may at least in part explain the decrease in plasma HDL. Since fatty acid composition and amounts are similar in plasma and liver PLs in FO and KO supplemented mice, it is likely that the stronger PPARα activation seen in liver and intestine is mediated via omega-3 PUFAs of the TAG fraction of FO supplemented mice.

The increase in the fatty acid content of liver TAG found in the FO group indicates that the observed upregulation of genes coding for catabolic pathways (e.g. the upregulation of genes coding for fatty acid oxidation and the increased ACOX activity and possibly elevated activity of CPT1) is apparently of minor importance in determining liver lipid levels in these animals. It should be noted that PPARα activation by e.g. fenofibrate also increases liver expression of lipogenic genes a with concomitant increase in fatty acid synthesis and TAG accumulation[47]. Thus, our data are consistent with PPARα activation by FO, although weaker than fibrate treatment, and hence increased lipid synthesis and possibly decreased VLDL secretion may at least partially explain the observed TAG accumulation in liver. Also, a recent study demonstrated that DHA attenuates postprandial hyperlipidemia in a PPARα dependent manner by activation of fatty acid oxidation genes in the intestine leading to decreased TAG and apoB secretion. However, PPARα-independent pathways that reduce the assembly and secretion of VLDL particles also appear to be involved[23, 48]. Some data indicate that FO decreases VLDL production mainly due to decreased plasma NEFA, which is suggested to be the main source of fatty acids for VLDL synthesis[23]. However only KO significantly decreased fasting plasma NEFA in our study but did not significantly change plasma TAG, suggesting additional mechanisms being involved. EPA is reported to inhibit DGAT activity in liver[49, 50], and also negatively influence the assembly and secretion of VLDL via incorporation into choline and ethanolamine PLs, which due to the higher total content of EPA in FO could be an additional possible explanation for the different effects on plasma TAG by FO and KO in our study[51]. This may also explain why similar levels of omega-3 PUFAs were found in plasma, since the TAG pool entering the VLDL fraction might be affected by the high levels of EPA, and that EPA may be shunted into the cytosolic TAG-pool for storage rather than entering VLDL in the FO group, while the PL pool (which contained similar amounts of omega-3 PUFAs in both groups in liver) that enters the different lipoprotein fractions, e.g. HDL, may not be as affected[51]. We did not assess the severity of the lipid accumulation in liver in this study and therefore we can not predict the health effects associated with these changes and effects of long-term feeding of these omega-3 supplements.

KO feeding reduced the expression and the activity of fatty acid synthase (Fas), which is in line with previous findings that KO supplementation decreases the mRNA expression of this protein more efficiently than FO[25, 27, 28]. Expression of Acacb (acetyl-CoA carboxylase 2) was also decreased in the KO group compared to the FO group, which rather showed an increased expression of this gene. Acetyl-CoA carboxylase 2 is associated with mitochondria and generates malonyl-CoA, which is a key regulator of energy homeostasis[52], and deletion of Acacb in mice promotes fatty acid oxidation and increased energy expenditure[53, 54]. Therefore, in conjunction with an apparent downregulation of Fas (and possibly Acaca) and thereby fatty acid synthesis in the KO group, this may provide a possible alternative mechanism by which KO maintains lower liver TAG levels compared to FO.

KO also reduced the mRNA expression of the first segments of the cholesterol/isoprenoid synthesis pathway, including the rate-limiting enzyme in cholesterol synthesis HMG-CoA reductase. However, the reduction in expression seen in our current study did not reflect in significant changes in plasma cholesterol (-16%), and liver cholesterol was even slightly increased as in the FO group. This was also seen in a low fat/omega-3 PUFA balanced fish and krill oil based study in which KO caused a similar downregulation in expression of genes involved in gluconeogenesis and cholesterol and fatty acid synthesis without changing plasma lipids[24]. Similar amounts of KO supplementation have previously been shown to decrease plasma and liver cholesterol in C56BL/6 mice on a HF diet, however in this study the mice were fed a different high fat diet containing buttermilk and 0.15% cholesterol ± KO for 8 weeks which may explain the different results[26]. More experiments are needed to elucidate the discrepancy between gene expression and plasma/liver cholesterol levels, e.g. measurement of total body cholesterol pool and bile acid and cholesterol synthesis and excretion.

In the present study we have also investigated the effects of FO and KO on gene expression of proteins involved in lipid metabolism in the small intestine. A similar trend in gene regulation as in the liver could be seen in the intestine of these animals. The expression of genes involved in fatty acid modulation and transport, such as Acot1, Cd36, I-fabp and L-fabp were all upregulated by the FO-containing diet, which is in line with previous findings of PPARα activation in the intestine during increased ligand availability[55]. KO on the other hand did not promote the same response to these PPARα driven genes. The increased expression of fatty acid transporters in the intestine of the FO fed group would potentially lead to an enhanced chylomicron production due to increased uptake of fatty acids, however FO has previously been shown to cause PPARα activation and thereby increased β-oxidation in murine intestine[56], which may balance an increased uptake of fatty acids. In a recent study, fenofibrate was shown to reduce blood TAG content in the postprandial state in part due to decreased dietary fat absorption and increased β-oxidation of fatty acids in spite of upregulation of chylomicron synthesizing genes and fatty acid transporters[57]. It is possible that the different effects of FO and KO may in part be due to differences in PPARα activation and thereby oxidative degradation of fatty acids in the small intestine, although the quantitative importance of fatty acid oxidation in the intestine is likely to be small. The only significant changes found in the KO group in genes coding for proteins in the chylomicron synthesis pathway was a downregulation of Mttp in the first intestinal segment by KO compared to HF diet, and a significant difference between FO and KO in the expression of Acat2 (in segment 3) and Dgat1 (in segments 2 and 3). Feces were not collected for absorption studies or lipid analysis in our study, therefore it is not clear whether omega-3 PUFAs per se, or omega-3 PUFAs in TAG versus PLs, affect dietary lipid absorption in the small intestine. Some data indicate that FO may reduce lipid uptake in the intestine, or at least delay efflux into the circulation by a transient accumulation of lipids in the enterocytes[58]. Whether KO affects lipid uptake in the intestine remains to be studied.

Both FO and KO raised plasma levels of omega-3 PUFAs to the same extent in spite of a markedly lower omega-3 PUFA content in the KO diet. FO lowered plasma TAG, PL and cholesterol and KO lowered NEFA in comparison to the control group. The two omega-3 fatty acid supplementations also promoted different gene expression profiles in liver and intestine with FO causing an apparent PPARα response by increasing the expression of genes coding for proteins in the two β-oxidation systems and other lipid metabolizing genes. In contrast, KO supplementation rather acted as a negative regulator of endogenous cholesterol and fatty acid synthesis, at least at the mRNA level. The stronger plasma lipid lowering effect with FO can be partly explained by increased lipid accumulation, mainly as TAG, in liver in spite of increased PPARα activation that may not compensate for decreased VLDL secretion. The physiological/pathological implications of the liver lipid accumulation by FO are not clear and may depend on diet composition and dose of FO and should be evaluated.