Omega-3 fatty acids alleviate chemically induced acute hepatitis by suppression of cytokines


  • Potential conflict of interest: Nothing to report.


Cytokines such as tumor necrosis factor alpha (TNF-α) are key factors in liver inflammation. Supplementation with essential omega-3 polyunsaturated fatty acids (n-3 PUFA) has been demonstrated to lower TNF-α and IL-1 production in mononuclear cells. An inflammation-dampening effect has been observed with increased omega-3 fatty acid supplementation in several inflammatory diseases. In this study, we used the transgenic fat-1 mouse, expressing a Caenorhabditis elegans desaturase endogenously forming n-3 PUFA from n-6 PUFA, to analyze the effect of an increased n-3 PUFA tissue status in the macrophage-dependent acute D-galactosamine/lipopolysaccaride (D-GalN/LPS) hepatitis model. We show less severe inflammatory liver injury in fat-1 mice with a balanced n-6/n-3 PUFA ratio as evidenced by reduced serum alanine aminotransferase levels and less severe histological liver damage. This decreased inflammatory response was associated with decreased plasma TNF-α levels and with reduced hepatic gene expression of TNF-α, IL-1β, IFN-γ and IL-6 in fat-1 mice, leading to a decreased rate of apoptosis in livers from fat-1 animals, as measured by DAPI-staining. Conclusion: The results of this study offer evidence for an inflammation dampening effect of omega-3 polyunsaturated fatty acids in the context of liver inflammation. (HEPATOLOGY 2007;45:864–869.)

Acute hepatitis due to viral, toxic, or autoimmune pathogenesis is characterized by an activation of macrophages and T cells with an increased production of cytokines that leads to parenchymal liver damage and liver dysfunction. An increase in tumor necrosis factor alpha (TNF-α) is one of the early events in liver inflammation. TNF-α has been implicated in liver damage in alcoholic hepatitis and steatohepatitis,1 and a study examining TNF-α blockade has shown improvement in alcoholic hepatitis.2

Growing evidence indicates that omega-3 polyunsaturated fatty acids (n-3 PUFA) and their specific lipid mediators can reduce the activity of inflammatory processes.3 High concentrations of n-3 PUFA reduce the lipopolysaccharide (LPS)–dependent activation of nuclear factor kappaB (NF-κB) in monocytes,4 which leads to a decreased production of TNF-α.5 Similar effects also have been observed in human mononuclear cells.6 Recent studies implicated newly identified n-3 PUFA–derived lipid mediators such as resolvin E1 in these processes.7, 8 Our recent results in the macrophage-dependent dextrane sodium sulfate colitis model showed suppression of experimental colitis by increased omega-3 fatty acid tissue content and formation of n-3 PUFA–derived lipid mediators.9 These observations suggest that omega-3 fatty acids might lower inflammation susceptibility in general. Hence, they could also dampen the inflammatory response in liver tissue, probably by regulating Kupffer cell activation and suppressing cytokine production.

We therefore evaluated the role of high n-3 PUFA content in the pathogenesis of D-galactosamine/lipopolysaccaride (D-GalN/LPS)–induced hepatitis in transgenic fat-1 mice. These mice carry the fat-1 gene from the roundworm Caenorhabditis elegans and are able to convert omega-6 into omega-3 fatty acids.10 Therefore, these mice have an endogenously elevated n-3 PUFA tissue content as compared with their wild-type littermates, even when maintained on a low n-3 PUFA diet. This is in contrast to feeding procedures that may bring in confounding factors as different diets need to be fed. Use of the transgenic fat-1 mouse model eliminates confounding factors of diet because only one diet is needed as the genetic approach using the fat-1 gene modifies the n-6/n-3 fatty acid ratio (converts n-6 to n-3) endogenously.

D-GaIN/LPS hepatitis is a well-established model for macrophage-dependent liver injury in mice.11D-GalN is known as a specific hepatotoxic transcriptional inhibitor that sensitizes the liver toward LPS treatment in subtoxic amounts12 and leads to an acute cytokine-dependent liver inflammation.13D-GaIN/LPS hepatic injury is T cell–independent: LPS binds to Kupffer cells and activates them, resulting in a liberation of large amounts of cytokines, such as TNF-α, IL-1, and IL-6.14–16 TNF-α is a critical mediator of apoptotic liver damage in this model.17 Interferon gamma (IFN-γ), produced by activated natural killer cells, is also a pro-inflammatory cytokine involved in this model.18

The results presented here show that transgenic fat-1 mice with a balanced n-6/n-3 PUFA ratio developed less severe D-GalN/LPS–induced inflammatory liver damage than wild-type mice, as evidenced by decreased production of pro-inflammatory cytokines and significantly lower serum alanine aminotransferase as well as less severe liver pathology.


AA, arachidonic acid; D-GalN, D-galactosamine hydrochloride; EPA, eicosapentaenoic acid; HAI, hepatitis activity index; IFN-γ, interferon gamma; LPS, lipopolysaccaride; n-3-PUFA, omega-3 polyunsaturated fatty acids; NF-κB, nuclear factor kappaB; RT-PCR, reverse transcription PCR; TNF-α, tumor necrosis factor alpha; IL-1β, interleukin 1 beta; IL-6, interleukin 6.

Materials and Methods


Transgenic fat-1 C57BL/6 mice were generated as previously described.10 Only heterozygous mice were used in this study. Mice were fed with an identical diet rich in n-6 fatty acids and low in n-3 fatty acids (10% safflower oil). All experiments were done with 7-week-old to 10-week-old male mice weighing 23 to 30 g. Mice were phenotyped according to the n-6/n-3 PUFA ratio in their ears determined by gas chromatography. Transgenic animals with an n-6 PUFA arachidonic acid to n-3 PUFA eicosapentaenoic acid (AA/EPA) ratio of approximately 1 in ear tissue were chosen for this study to reflect a balanced n-6/n-3 ratio in the experimental group (AA/EPA = 0.85 ± 0.1, n = 5). All animals received care according to institutional guidelines, and all experiments were approved by the MGH Subcommittee on Research Animal Care.

Reagents and Treatments.

D-GaIN was purchased from Sigma Chemical Co. (St. Louis, MO) and injected intraperitoneally at a dose of 700 mg/kg body weight dissolved in sterile phosphate-buffered saline. LPS from Salmonella abortus equi was purchased from Sigma Chemical Co. and administered intraperitoneally at a concentration of 2 μg/kg body weight together with D-GaIN. As a control, wild-type mice and fat-1 mice were injected with 200 μl sterile phosphate-buffered saline intraperitoneally.

Measurement of Liver Damage.

Six hours after D-GaIN/LPS injection, mice were anesthetized with isoflurane and killed by cardiopuncture to maximize blood yield for further study. Blood samples were allowed to clot and subsequently centrifuged at 1,000 g for 15 minutes. The serum was collected and stored at −80°C. Serum ALT activities were measured using a commercial kit purchased from BioTron Diagnostics (Hemet, CA) as specified by the manufacturer.

Determination of Plasma TNF-α.

To measure TNF-α concentrations in blood plasma, blood was withdrawn from the mice tail vein 90 minutes afterD-GaIN/LPS injection and centrifuged at 1,000 g. The plasma was collected and stored at −80°C until TNF plasma concentration was determined using the TNF-α OptEIA ELISA kit from BD Pharmingen (Franklin Lakes, NJ) according to the manufacturer's instructions.

Analysis of Cytokine mRNA Expression.

To analyze cytokine mRNA transcription levels in the liver, mice were killed 6 hours after D-GaIN/LPS injection and the liver was removed immediately. To determine intrahepatic TNF-α, IFN-γ, IL-1β, and IL-6 mRNA levels, total liver RNA was isolated by using the guanidine isothiocyanate based TRIzol solution (Invitrogen, Carlsbad, CA) according to the manufacturer's specifications. RNA concentration and purity was assessed spectrophotometrically at 260 nm in relation to the absorbance at 280 nm. Liver RNA was transcribed into cDNA using random primers and a reverse transcriptase system (Promega, Madison, WI).

Real-time reverse-transcription PCR (RT-PCR) was performed using SYBR Green Mastermix (Applied Biosystems, Foster City, CA) and an ABI Prism 7000 Light Cycler according to the manufacturer's instructions. All samples were processed in triplicate, and GAPDH was used as a housekeeping gene to normalize mRNA levels. Primers used in amplifications were as follows: mouse GAPDH, 5′- ATGGACTGTGGTCATGAGCC-3′ and 5′-ATTGTCAGCAATG CATCCTG-3′; mouse IL-1β, 5′-GCAACTGTTCCTGAACTCAACT-3′ and 5′- ATCTTTTGGGGTCCGTCAACT-3′; mouse IL-6, 5′- TTGGTCCTTAGCCACTCCTTC-3′ and 5′- TAGTCCTTCCTACCCCAATTTC-3′; mouse TNF-α, 5′-GCTACGACGTGGGCTACA-3′ and 5′-CCCTCACACTCAGATCATCTTCT-3′; mouse IFN-γ, 5′-AGAAATAGTTGAGGAGACAGAAAT-3′ and 5′-TTAGATGCATCAACCAAAGAAGTA-3′.

Analysis of Fatty Acids Composition.

To determine the fatty acid profile in liver tissue, samples were prepared as described.19 Fatty acid methyl esters were analyzed by gas chromatography using a fully automated HP 5890 system equipped with a flame-ionization detector. Peaks of resolved fatty acids were identified by comparison with fatty acid standards (Nu-chek-Prep, Elysian, MN), and area percentage for all resolved peaks was analyzed using a PerkinElmer M1 integrator.

Concentrations of 18:2 (n-6), 20:4 (n-6), 22:4 (n-6), and 22:5 (n-6) were taken together and divided by the sum of 18:3 (n-3), 20:5 (n-3), 22:5 (n-3), and 22:6 (n-3) to determine the ratio of n-3 and n-6 fatty acids in liver tissue.

For phenotyping, ear tissue of every mouse was cut and samples were measured by gas chromatography, and the concentration of 20:4 (n-6) was divided by the concentration of 20:5 (n-3). For our study, we only used fat-1 mice with a 20:4 (n-6) / 20:5 (n-3) ratio of approximately 1 (AA/EPA = 0.85 ± 0.1, n = 5).


Liver samples were fixed in 10% buffered formalin and embedded in paraffin, sliced, and stained with hematoxylin-eosin.

Histological examination was performed in a blinded manner and graded by an experienced pathologist (H.L.). To evaluate the degree of liver damage, we used the modified hepatitis activity index (HAI) following Ishak et al.20 and Knodell et al.21 The modified HAI is a combined score that assesses liver inflammation, necrosis, and fibrosis. Because of the short period of hepatitis induction, not leading to significant degrees of fibrosis, we excluded the fibrosis score from the HAI analysis used here.


To analyze the typical morphological signs of apoptosis, paraffin-embedded semithin sections were deparaffinized and stained with DAPI (4′,6-diamine-2-phenylindole dihydrochloride) in an aqueous dilution of 1:10,000 for 5 to 10 minutes. DAPI-stained tissue was examined with a fluorescence microscope and photographed. For quantification of apoptosis, 300 nuclei were randomly viewed, and apoptosis was counted by 2 independent persons.

Statistical Analysis.

Statistical analysis was performed with Prism 3.02v Software (GraphPad, San Diego, CA). All data in this study are expressed as mean ± SEM. The significance of differences was tested by 2-tailed t test. A P value ≤ 0.05 was considered significant.

Results and Discussion

Although liver tissue of wild-type mice showed high levels of n-6 PUFA with a very low n-3 PUFA content (ratio n-6/n-3 PUFA 64.61 ± 16.78, n= 5), the transgenic fat-1 mice with a balanced AA/EPA ratio had significant endogenous amounts of n-3 PUFA in their liver tissue (ratio n-6/n-3 PUFA 5.90 ± 0.21, n = 5) (Table 1). The most notable differences between the two groups were found in α-linolenic acid (18:3 n-3), eicosapentaenoic acid (20:5 n-3), docosapentaenoic acid (22:5 n-3), and docosahexaenoic acid (22:6 n-3). Slightly lower levels of arachidonic acid were seen in the liver tissue of fat-1 animals as compared with the wild-type mice; however, the difference was not statistically significant.

Table 1. Composition of Polyunsaturated Fatty Acids in the Liver as Analyzed by Gas Chromatography
Fatty AcidsPercent of Total Fatty Acids
Wild-Type MiceFat-1 Mice
  • NOTE. Fat-1 mice showed increased amounts of n-3 PUFA.

  • Abbreviation: ND, not detectable.

  • ***

    P < 0.001 versus wild-type mice; n = 5.

  • **

    P < 0.01 versus wild-type mice; n = 5.

  • *

    P < 0.05 versus wild-type mice; n = 5.

18:2 n-627.01 ± 0.8128.63 ± 1.47
18:3 n-3ND0.06 ± 0.02*
20:4 n-614.30 ± 1.8411.12 ± 0.82
20:5 n-3ND0.29 ± 0.03*
22:4 n-60.92 ± 0.080.49 ± 0.11*
22:5 n-63.39 ± 0.510.67 ± 0.19**
22:5 n-3ND0.45 ± 0.04*
22:6 n-30.92 ± 0.256.17 ± 0.31***
Total n-643.46 ± 1.5440.90 ± 1.20
Total n-30.92 ± 0.256.97 ± 0.32***
Ratio n-6/n-364.61 ± 16.785.90 ± 0.21**

Administration of LPS to D-GaIN–sensitized mice induced severe hepatic damage as detected by increased serum ALT at 6 hours after injection. Serum ALT activities were found to be significantly lower in the fat -1 group than in wild-type mice (Fig. 1). All mice treated with D-GaIN/LPS showed histopathological signs of acute hepatitis 6 hours after challenge, reflected by necrosis, apoptosis, inflammatory cell infiltrate, and hemorrhage. Histological examinations of liver sections showed severe confluent and focal necrosis, apoptosis, and focal inflammation in wild-type mice. Liver damage and histological changes were found to be significantly less severe in fat-1 mice (Fig. 2). The scores of the HAI were significantly different between the wild-type and fat-1 group (P < 0.05 for the comparison between fat-1 mice versus wild-type mice; Table 2).

Figure 1.

Serum ALT levels in wild-type mice (4,773 ± 1,036 U/l, n = 5) and transgenic fat-1 mice (801.2 ± 154.3 U/l, n = 5) 6 hours after intraperitoneal D-GaIN/LPS injection. **P < 0.01 D-GaIN/LPS–treated fat-1 versus D-GaIN/LPS–treated wild-type mice.

Figure 2.

Decreased susceptibility to D-GaIN/LPS–induced acute liver injury in omega-3–rich fat-1 mice. Hematoxylin-eosin stains in 2 different magnifications are shown for the different groups. As compared with saline-treated control animals (A, B), wild-type mice challenged with D-GalN/LPS showed massive apoptosis and necrosis, in particular hyperchromatic and condensed hepatocyte nuclei (C, D). In contrast, histopathological signs of acute hepatitis are reduced in fat-1 mice treated with D-GalN/LPS (E, F).

Table 2. Histological Analysis of Liver Tissue 6 Hours After D-GaIN/LPS Treatment
Histological changeDegree of Liver Injury Determined by Using the Modified HAI
WT MiceFat-1 Mice
  • NOTE. To quantify histological changes, the modified hepatitis activity index was used including (A) periportal or periseptal interface hepatitis, (B) confluent necrosis, (C) focal lytic necrosis, apoptosis, and focal inflammation, (D) portal inflammation. Fibrosis was not scored. Significantly less hepatic damage was observed in transgenic fat-1 mice.

  • **

    P < 0.01 versus D-GaIN/LPS–treated wild-type mice, n = 5.

  • *

    P < 0.05 versus D-GaIN/LPS–treated wild-type mice, n = 5.

A0.4 ± 0.20.4 ± 0.2
B3.0 ± 0.31.2 ± 0.5*
C4.0 ± 0.02.6 ± 0.4**
D0.4 ± 0.20.6 ± 0.2
Total HAI/187.8 ± 0.44.8 ± 1.1*

Of particular interest was analyzing the pathways of cell death in this context, because the examination of particularly apoptotic activity in the inflamed hepatic tissue can contribute to the understanding of the damage processes in the liver. Previous results have indicated that TNF-α is a strong inductor of apoptosis in D-Gal–sensitized liver tissue.22 DAPI staining was used to detect apoptotic hepatocytes. Administration of D-GaIN/LPS to wild-type mice induced high numbers of apoptotic cells (109.3 ± 8.3 per 300 nuclei; n = 3), whereas fat-1 mice showed remarkably less hepatocellular apoptosis (41.5 ± 5.1 per 300 nuclei; n = 3), indicating less severe cellular damage in the fat-1 animals (Fig. 3).

Figure 3.

Analysis of hepatocellular apoptosis. (A) Fat-1 mice showed significantly less apoptotic hepatocytes 6 hours after D-GaIN/LPS treatment. **P < 0.01 versus D-GaIN/LPS–treated wild-type mice; n = 3. (B-D) Representative DAPI-stained liver tissue. (B) Control tissue, untreated wild-type mouse without apoptotic signs. (C) D-GaIN/LPS–treated wild-type mouse 6 hours after hepatitis induction: Note the typical condensed chromatin of apoptotic hepatocytes. (D) Liver tissue of a fat-1 mouse 6 hours after D-GaIN/LPS injection with less apoptotic hepatocytes. 100× magnification.

TNF-α has been shown to be a crucial pro-inflammatory mediator in acute liver inflammation.1, 13, 17 Plasma levels of TNF-α were therefore determined at 90 minutes after D-GaIN/LPS challenge and found to be significantly higher in wild-type mice (2,216 ± 684.6 pg/ml; n = 4) than in fat-1 mice (455.5 ± 145.1 pg/ml; n = 4). Furthermore, intrahepatic TNF-α expression, as measured by real-time RT-PCR 6 hours after hepatitis induction, was significantly different between wild-type and fat-1 animals (wild-type mice, 30.65 ± 6.09, -fold induction as compared with control animals, n = 5, fat-1 mice 13.31 ± 2.68, -fold induction, n = 5) (Fig. 4). Hence, plasma and intrahepatic TNF-α levels correlated with the severity of liver disease.

Figure 4.

Modulation of (A) plasma TNF-α levels and (B) hepatic TNF-α expression. Plasma TNF-α concentrations were measured by ELISA 90 minutes after D-GaIN/LPS treatment. Expression of intrahepatic TNF-α mRNA was measured by quantitative real-time RT-PCR, 6 hours after challenge. Fat-1 mice showed significantly reduced levels in both parameters as compared with the wild-type group. *P < 0.05 versus D-GaIN/LPS–treated wild-type mice; n = 4. #P < 0.05 versus D-GaIN/LPS–treated wild-type mice; n = 5.

These findings are consistent with previous studies in animal models and in humans, which showed that n-3 PUFA decreased TNF-α production.4–6 This could be due to n-3 PUFA–derived lipid mediators, the resolvins and protectins, which have been shown to be potent antiinflammatory mediators.7 A recent study analyzing the formation and the molecular effect of n-3 PUFA has identified a G protein-coupled receptor–specific pathway involved in NF-κB down-regulation by the n-3 PUFA–derived resolvin E1,8 which in turn could also down-regulate NF-κB–induced TNF-α formation.

We next examined the hepatic expression of the inflammatory cytokines IL-6, IFN-γ, and IL-1β by using quantitative real-time RT-PCR. The fat-1 group showed a significant reduction in IL-6 mRNA compared with the wild-type animals (Fig. 5A). The proinflammatory IL-6 is elevated in the D-GalN/LPS hepatitis model13; therefore, the significantly decreased expression of IL-6 mRNA in fat-1 mice could contribute to the reduced inflammatory response observed in the fat-1 mice. Higher expression of IFN-γ and IL-1β was also seen in the wild-type animals, demonstrating dampening of macrophage-associated pro-inflammatory cytokines in fat-1 mice (Fig. 5B,C). Particularly with regard to IFN-γ expression, examining underlying mechanisms further in the future will be interesting, because cytokines such as IL-18 and IL-12 have been implicated in the upstream, leading to expression of IFN-γ in macrophages.23

Figure 5.

Expression of different pro-inflammatory cytokines in the liver of D-GalN/LPS–challenged animals. (A) IL-6, (B) IL-1β, and (C) IFN-γ mRNA levels in the liver were measured by quantitative real-time RT-PCR 6 hours after D-GaIN/LPS treatment. **P < 0.01 versus D-GaIN/LPS–treated wild-type mice; n = 4. #P < 0.05 versus D-GaIN/LPS–treated wild-type mice; n = 5. *P < 0.05 versus D-GaIN/LPS–treated wild-type mice; n = 4 (wt) and n = 5 (fat-1).

The results presented here indicate that increasing the hepatic content of n-3 PUFA could decrease inflammatory activity in acute hepatitis. A limitation of our study might be the narrow range of the AA/EPA ratio chosen for the mice in the experimental group. Whether higher or lower n-6/n-3 PUFA ratios modify the inflammation dampening effect observed here is not clear.

The n-3 fatty acids might exert an anti-inflammatory effect via competitive inhibition of the n-6 PUFA–derived pro-inflammatory eicosanoids. However, in this study, we found only small and not significant differences in the content of arachidonic acid (AA, 20:4 n-6) in the liver tissue between fat-1 transgenic and wild-type mice. Indeed, we found higher levels of arachidonic acid as compared with the direct n-3 PUFA competitor eicosapentaenoic acid (EPA, 20:5 n-3) also in the fat-1 mice. A significant difference was seen in the amounts of n-3 fatty acids such as EPA and DHA, between fat-1 mice and their wild-type littermates. Therefore, based on the results of our previous study in a colitis model,9 lipid mediators formed from the n-3 PUFA may be responsible for the inflammation-dampening effect seen in this population.

Although the D-GalN/LPS model of acute hepatitis in mice is not directly comparable to liver inflammation in humans, it is a well established model of hepatitis, mirroring activation of macrophages and cytokine release, factors crucial also in human hepatitis of various causes.1, 13 Our results indicate a role for n-3 PUFA in alleviation of hepatic injury and inflammation. Indeed, a recent case report suggests that n-3 PUFA might be beneficial in infants with intestinal failure and parenteral nutrition–related liver disease.24 Future studies will be necessary to analyze in more detail the optimal fatty acid ratios and lipid mediators involved in n-3 PUFA–associated inflammation dampening in the liver and also to expand the data presented here into models of clinically important chronic hepatitis.


The authors thank I. Preuss, ZELMI, TU-Berlin, and Prof. Dr. A. Ding, Department of Physics, TU-Berlin, for help with microscopy.