www.elsevier.com/locate/jpedsurg

Journal of Pediatric Surgery (2008) 43, 2010–2015

Reduction of hepatocellular injury after common bile ductligation using omega-3 fatty acidsSang Lee a, Sendia Kima, Hau D. Le a, Jonathan Meisel a, Robbert A.M. Strijbosch a,Vania Noseb, Mark Puder a,⁎

aDepartment of Surgery, The Vascular Biology Program, Children's Hospital Boston, Harvard Medical School,Boston, MA 02115, USAbDepartment of Pathology, Brigham and Women's Hospital, Harvard Medical School Boston, MA 02115, USA

Received 26 March 2008; revised 18 May 2008; accepted 20 May 2008

0d

Key words:Omegaven;Omega-3 fatty acids;Common bile ductligation;

Menhaden

AbstractPurpose: Bile duct obstruction and subsequent cholestasis produces hepatocellular injury and aninflammatory response. Fatty acid constitution of cell membranes plays a major role in the inflammatorycascade. Omega-3 fatty acids are antiinflammatory. We proposed that omega-3 fatty acidsupplementation would reduce hepatocellular damage and cell death in a model of murine commonbile duct ligation.Methods: Mice underwent bile duct ligation and were administered either control soy diet (omega-6) orMenhaden diet (omega-3), and parameters of liver injury were measured at postoperative days 1, 4, and8. Serum was analyzed for liver function tests. Liver tissue was scored for histologic necrosis andinflammation, and apoptosis was qualitatively measured.Results: At day 8, comparing control and Menhaden, liver function tests were not significantly different.The H&E slides were analyzed and scored. At day 4, the mean necrosis scores for the Menhaden-fedgroup was 0.01 ± 0.028 and 0.46 ± 0.108 for the soy-fed group (P = .001) and at day 8, 0.420 ± 0.107 and1.22 ± 0.132 (P b .001). The mean portal inflammation score for day 4 Menhaden-fed and soy-fed micewas 1.40 ± 0.245 for both groups (P = 1.00) and for day 8, 1.80 ± 0.200 and 2.80 ± 0.200 (P = .008). Atday 1, the median terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling scores of theMenhaden vs soy group were 6.0 and 0.0 (P b .001); day 4, 24.0 and 3.0 (P b .001); and day 8, 0.0 and 3.0(P b .001), respectively.Conclusion: Although there appears to be a trend toward biochemical protection and a marked reductionof necrosis and inflammation, there was no significant liver function test difference between control andMenhaden groups. Considering our data of blunted histologic hepatotoxicity with omega-3 fatty acidsupplementation, we hypothesize that this may be a method of reducing long-term complications of liverinjury secondary to diseases of cholestasis such as biliary atresia, namely fibrosis and cirrhosis.© 2008 Elsevier Inc. All rights reserved.

⁎ Corresponding author. Tel.: +1 617 355 7103; fax: +1 617 730 0298.E-mail address: mark.puder@childrens.harvard.edu (M. Puder).

022-3468/$ – see front matter © 2008 Elsevier Inc. All rights reserved.oi:10.1016/j.jpedsurg.2008.05.030

Bile duct obstruction and subsequent cholestasis produces

hepatocellular injury and an inflammatory response. Varyingdegrees of these processes occur in patients with chronicliver disease. Cholestasis results in poor bile secretion and

Table 1 Fatty acid content of Menhaden and soybean oil

Fat composition (%) Menhaden oil Soybean oil

Linoleic 1.5 54.8α-Linolenic 1.6 7.8Arachidonic acid 0.9 -EPA 15.5 -DHA 9.1 -Oleic 11.4 22.7Palmitic 17.1 10.2Stearic 2.8 4.5

DHA indicates docosahexaenoic acid.

2011Hepatocellular injury reduction using omega-3 fatty acids

subsequently, poor lipid digestion and absorption of fat-soluble vitamins. In addition, intraluminal flora is alteredwith increased bacterial translocation, portal bacteremia, andendotoxemia [1,2]. The accumulation of toxic bile saltswithin the liver produces apoptosis and necrosis ofhepatocytes. In turn, cell turnover may lead to scarring,fibrosis, and eventually, cirrhosis and liver failure [3]. Manyliver diseases, such as biliary atresia, primary biliarycirrhosis, hepatitis B, alcoholic hepatitis, and autoimmunehepatitis, demonstrate a cholestatic pathophysiology [4-7].

Hepatocellular damage from common bile duct ligationoccurs via 3 different mechanisms as follows: tumor necrosisfactor α (TNF-α), Fas ligand (FasL), and oncotic necrosis.Tumor necrosis factor α is an inflammatory cytokine that hasbeen implicated as a mediator of liver injury and apoptosis.Fas ligand is a membrane protein that is also a mediator ofapoptosis via trimerization with the Fas receptor (FasR) andsubsequent activation of caspase-8, which appears to beinvolved in cholestatic liver injury (ie, common bile ductligation) [8]. Fas ligand has also been implicated in cirrhosis,secondary to cholestasis and viral infection [9]. Oncoticnecrosis is characterized by cellular swelling and rupture inresponse to adenosine triphosphate (ATP) depletion [10].

Our previous work with parental nutrition–associatedliver disease (PNALD) encouraged us to examine thepossibility that omega-3 fatty acids may be effective againstliver damage secondary to cholestasis [11-14]. Parenteralnutrition–associated liver disease is characterized by aspectrum of fatty liver disease, inflammation, bile ductproliferation, fibrosis, and cirrhosis. Biochemically, thedisease is in part because of TNF-α–mediated inflammationand FasL-mediated apoptosis in animal models [15,16]. Inour initial studies, we hypothesized that the development ofPNALD may be dependent on the type of fat administered(soybean oil–based vs fish oil–based) and that omega-3 fattyacids would prevent or reduce de novo lipogenesis andsubsequent liver injury. In this model, mice were treated withoral PN for 19 days. This energy load was similar to theestablished dietary energy needs of the mouse [17].Experimental groups were supplemented with Intralipid20% (Baxter, Deerfield, Ill), which is primarily plant-derived(omega-6), and fish oil (omega-3). Mice that received PNwithout fat supplementation developed fatty liver injury.Mice receiving intravenous Intralipid had even more severeliver changes. Both groups of animals developed markedhepatic steatosis with macrovesicular fatty infiltration andsignificant elevations in spectroscopic liver fat content andserum transaminase levels. However, animals receiving fishoil had completely normal livers on histologic examination,and spectroscopy revealed normal liver fat content. Inaddition, there was no fatty acid deficiency in either group asdetermined by serum fatty acid analysis.

A hepatoprotective effect of omega-3 fatty acids has beendemonstrated in a study by Schmocker et al [18], in whichFat1 mice (genetically engineered to endogenously convertomega-6 fatty acids to omega-3 fatty acids) were subjected to

the D-galactosamine/lipopolysaccharide (D-GalN/LPS)model of acute hepatitis. Study findings indicated lowerserum alanine aminotransferase (ALT) levels, increasedserum and tissue TNF-α levels, and decreased histologicfindings of liver damage in the Fat1 mice as compared towild-type litter mates 6 hours after D-GaIN/LPS injection.

Essential fatty acids (EFA) are termed as such becausethey cannot be synthesized by the human body and thusmust be derived from exogenous sources. There are 2 EFAgroups—omega-6 and omega-3. The biologically activeomega-6 fatty acid is arachidonic acid (AA). The downstreamomega-3 products are eicosapentaenoic acid (EPA) anddocosahexaenoic acid. All of these fatty acids play animportant role in cell membrane composition, thus influen-cing fluidity and cell surface biochemical signaling, as well asserving as natural ligands for certain nuclear receptors thataffect gene expression. In addition, AA and EPA areimportant eicosanoid and prostanoid precursors. Arachidonicacid products are proinflammatory mediators, whereas EPAproducts are antiinflammatory or rather less proinflammatory[19-22]. This makes their interaction critical, especiallyconsidering that AA and EPA are competitive substrates.

Our hypothesis is that omega-3 fatty acid supplementationreduces hepatocellular damage and cell death in a model ofmurine common bile duct ligation, an animal model forcholestatic diseases such as biliary atresia, and to character-ize the mechanism by which this may occur.

1. Materials and methods

1.1. Murine common bile duct ligation

Twenty disease-free 6 to 8-week-old male C57BL/6 mice(Jackson Laboratories, Bar Harbor, Me) were divided into 2groups and fed either 5% soybean oil or 10% Menhaden oilin a purified rodent diet (Dyets Incorporated, Bethlehem, Pa)(Table 1) for 3 weeks before surgery. The animals, in groupsof 5, were housed in a barrier room and were acclimated totheir environment for at least 72 hours before the initiation ofeach experiment. Common bile duct ligation was performedunder 3% isoflurane anesthesia. A midline laparotomy,extending from xyphoid to pubis, was made. The duodenum

2012 S. Lee et al.

was retracted, and the common bile duct was dissected freefrom surrounding structures. The common bile duct was thendouble ligated with 5-0 silk suture, and the peritoneum andskin was closed with 4-0 polypropylene suture. Animalswere then administered a 5-mL normal saline bolussubcutaneously and allowed to recover under warminglamps. Sham-operated mice received a midline laparotomywith isolation of the common bile duct without ligation. Theanimals were killed on postoperative days 1, 4, and 8. Allanimal experiments were done in accordance with NationalInstitutes of Health guidelines as dictated by the Animal CareFacility at Children's Hospital Boston (Boston, Mass).

Upon euthanasia, mice were weighed and anesthetizedwith 300 μL of 2.5% Avertin (Tribromoethanol, Sigma-Aldrich Corporation, St Louis, Mo) by intraperitonealinjection. Approximately 400 μL of blood was collectedfrom each mouse via retroorbital puncture. The specimenswere then placed into serum separator tubes (BectonDickinson, Franklin Lakes, NJ) and centrifuged at 4°C at8000 revolutions per minute for 10 minutes to separateserum. Serum was frozen at −80°C and delivered to theClinical Laboratory at Children's Hospital Boston formeasurement of aspartate aminotransferase (AST), ALT,alkaline phosphatase (AP), and total (TB) and direct bilirubin(DB) levels. A liver tissue sample was then immersed in 10%formalin at 4°C overnight. The next day, the sample waswashed in phosphate-buffered saline (PBS) 3 times andimmersed in 70% ethanol (EtOH) in PBS. The liver sampleswere then sectioned (5 μmol/L) in paraffin before immuno-histochemical analysis. Sections were stained for H&E andterminal deoxynucleotidyl transferase biotin-dUTP nick endlabeling (TUNEL). Another tissue sample was snap frozen indry ice for further analysis.

1.2. Histologic scoring of liver injury

The livers were scored for areas of necrosis and portalinflammation using components of the modified histologicalactivity index (HAI) grading [23]. Focal (spotty) lyticnecrosis and focal inflammation was scored on a scale of0 to 4 as follows: 0 = none; 1 = 1 focus or less per 10×objective; 2 = 2 to 4 foci per 10× objective; 3 = 5 to 10 foci per10× objective; and 4 = more than 10 foci per 10× objective.Portal inflammation was also scored from 0 to 4 as follows:0 = none; 1 = mild, some, or all portal areas; 2 = moderate,some, or all portal areas; and 3 to 4 = marked, all portal areas.To determine the extent of necrosis, 10 fields were scored byan observer blinded to the treatment groups for each slide.The score for portal inflammation was obtained based on thehistologic observation of the entire slide.

1.3. Immunohistochemistry

Paraffin liver sections were rehydrated with sequentialxylene and EtOH washes. Sections were subjected to xyleneimmersion for 2 minutes twice, 100% EtOH for 3 minutes,

95% EtOH for 3 minutes, and then 70% EtOH for 3 minutes.Next, epitope retrieval was achieved by incubating sectionswith proteinase K (20 μg/mL) in a humidified chamber at37°C for 20 minutes, followed by a cool down period to roomtemperature for 20 minutes. Sections were then washed 3times with PBS + 0.3% Triton X-100 (Sigma, St Louis, MO).Serum blocking was then performed by incubating sections in5% normal goat serum (Jackson ImmunoResearch, WestGrove, Pa) in PBS + 0.3% Triton X-100 + 0.01% bovineserum albumin + 0.01% Thimerosal at 37°C for 30 minutes.The TUNEL staining protocol was performed using theDeadEnd Fluorometric TUNEL System (Promega, Madison,Wis), according to manufacturer's instructions. Sections thenunderwent three 5-minute washes with PBS + 0.1% Triton X-100 + bovine serum albumin (5 mg/mL) and then one washwith PBS. Hoechst 33258 (1 μg/mL) nuclear stain (BisBen-zimidine, Sigma, St Louis, Mo) was then applied to eachsection for 5 seconds then aspirated. Two drops of Gel MountAqueous Mounting Medium (Sigma-Aldrich Corp, St Louis,Mo) was applied followed by a cover slip. Clear nail polishwas applied to the edges of the coverslip and allowed to dry.

1.4. Confocal microscopy

All TUNEL stained sections were analyzed using theLeica TCS SP2 confocal microscope and its associatedsoftware (Leica Microsystems, Bannockbum, IL). Imageswere optimized for Z-position, gain, O/U flow, and laserintensity. Ten high-powered fields (63×) were selected atrandom and counted by 2 blinded observers for positiveTUNEL staining.

1.5. Statistical analysis

Comparison of the means between the 2 experimentalgroups was made using 2-tailed independent t tests. Allstatistical tests were performed using SigmaStat software(SPSS, Chicago, Ill). Values are listed as mean ± SEM.

2. Results

2.1. Morbidity and mortality

All animals survived. Three mice in the control soy-fedgroup and 1 mouse in the Menhaden group acquired bileleaks and were excluded from the study. Both groups showedappropriate growth during the 3-week period, with no animallosing weight and each animal gaining an average of 2 g(data not shown).

2.2. Liver function tests

At day 8, there were no significant differences in liverfunction tests (ALT, AST, AP, TB, DB) between soy and

Fig. 1 H&E sections of soy-treated (control) and Menhaden-treated mice after common bile duct ligation. Postoperative day 4.

Fig. 2 TUNEL sections of soy-treated (control) and Menhaden-treated mice after common bile duct ligation. Positively stained (apoptotic)nuclei appear green. Postoperative days 1, 4, and 8.

2013Hepatocellular injury reduction using omega-3 fatty acids

2014 S. Lee et al.

Menhaden groups after common bile duct ligation. Specifi-cally comparing the soy and Menhaden groups, AST levelswere 1141.3 and 877 (P = .334), ALT levels were 757.7 and641.7 (P = .438), AP levels were 505.9 and 423.3 (P = .272),TB levels were 18.7 and 17.8 (P = .784), and DB levels were19.2 and 20.9 (P = .821), respectively.

2.3. Histologic scoring

The H&E stained slides were analyzed and scored by ablinded observer. At day 4, the mean necrosis scores for theMenhaden-fed and soy-fed groups were 0.01 ± 0.028 and0.46 ± 0.108, respectively (P = .001) (Fig. 1). At day 8, themean necrosis score for the Menhaden-fed and soy-fedgroups were 0.420 ± 0.107 and 1.22 ± 0.132, respectively(P b .001). The mean portal inflammation score for day 4Menhaden-fed and soy-fed mice was 1.40 ± 0.245 for bothgroups (P = 1.00). However, on day 8, the mean portalinflammation score was 1.80 ± 0.200 for the Menhaden-fedmice and 2.80 ± 0.200 for the soy-fed mice (P = .008).

2.4. Immunohistochemistry

The TUNEL stained slides were analyzed and positivecells counted by a blinded observer. At day 1, the medianTUNEL scores of the Menhaden vs soy group were 6.0and 0.0 (P b .001), respectively. At day 4, the medianTUNEL score of the Menhaden vs soy group was 24.0 and3.0 (P b .001), respectively. However, at day 8, the medianscores were 0.0 and 3.0 for the Menhaden and soy groups(P b .001), respectively (Fig. 2).

3. Discussion

Murine common bile duct ligation is a standard modelof liver injury, most analogous to biliary atresia, which ischaracterized by hepatocyte apoptosis, necrosis, lympho-cytic infiltration, and later, hepatic fibrosis. As ananatomical physiologic equivalent to cholestasis, it pro-vides a manner in which biliary atresia, among otherhepatic diseases, may be mimicked. Based upon ourprevious experience with a different murine liver diseasemodel (PNALD), we hypothesized that “pretreatment” withomega-3 fatty acids may reduce hepatotoxicity resultingfrom common bile duct ligation. Replacement of conven-tional soy-based diets with Menhaden-oil supplementeddiets results in a shift in cell membrane composition fromomega-6 predominant to omega-3 predominant, thuscreating a less inflammatory milieu wherein hepatic injurymay be blunted.

All animals survived the procedure, proving the feasi-bility of common bile duct ligation in a murine model.Postoperative complications, as expected, included bile leak,observed upon necropsy. This occurred at an acceptable rateof 10% and prompted exclusion from the study. Mice in both

groups (control and Menhaden) experienced appropriategrowth, ensuring that the fish oil supplement was of requisitenutritional content to prevent malnutrition. We have alsodemonstrated in our laboratory that fish oil–supplementedfood provides sufficient fatty acids to prevent essential fattyacid deficiency, lending further support and evidence to thenormal growth observed in this experiment. Although theanimals were allowed to feed ad libitum, we have performedpair feeding experiments with these diets in recent studies,balancing energy expenditure and fat energy, and we havedemonstrated that Menhaden energy expenditure and energyefficiency was comparable if not better than normal chow-fed animals [24].

Although there appears to be a trend toward biochemicalprotection in the Menhaden-fed group, as measured byserum liver function tests, no significant difference betweengroups was detected. Both groups demonstrated transami-nitis and cholestasis at day 8, after common bile ductligation. The greatest differences were seen in the transami-nases with reduction of AST levels from 1141.3 to 877 andALT levels from 757.7 and 641.7 in the Menhaden groups.However, bilirubin levels, which are highly reflective ofcholestasis, as expected because of common bile ductligation, showed virtually no effect with TB levels of 18.7and 17.8 and DB levels of 19.2 and 20.9, in control andMenhaden groups, respectively. The disparity betweenhistologic improvement and biochemical markers over thetime span of the current study may be because of asynchronybetween the Menhaden-fed and soy-fed groups. The liverfunction tests were obtained after the acute necrosis andapoptosis was complete and may have missed an earlierelevation of enzymes. In addition, it is possible that theincreased apoptotic rate observed in the Menhaden groupcould be contributing to transaminitis.

The histologic data suggest an apparent omega-3 fattyacid–mediated protection against necrosis by H&E analysis/scoring and increased susceptibility to apoptosis by TUNELimmunohistochemical staining. Temporal analysis of apop-tosis appears to display an apoptotic wave (TUNEL-positive) at day 4 in Menhaden-treated animals with fullrecovery by day 8. This correlates with our clinicalobservations where we have observed a transient increasein direct bilirubin levels at the onset of omega-3 fatty acidtreatment followed by normalization in infants withPNALD, possibly representing the mechanism and paradigmby which this occurs.

Other research has demonstrated that the omega-3 fattyacid EPA induces apoptosis in hepatocelluar carcinoma cellsin vitro. Reversal of this apoptotic effect with FasL antisenseoligonucleotides indicates that the mechanism is Fas-mediated [25]. In vivo research with mice further indicatesa link between omega-3 fat consumption and induction ofFas-mediated apoptosis [26]. In this way, the wave ofapoptosis that was observed in the Menhaden-fed group atdays 1 and 4 in the present study was not surprising.Resolution of apoptosis was observed at postoperative day 8,

2015Hepatocellular injury reduction using omega-3 fatty acids

indicating that the effect was acute and not persistent.Furthermore, the Menhaden-fed group exhibited signifi-cantly decreased necrosis on H&E staining as compared tothe soy-fed group, suggesting a hepatoprotective capacity foromega-3 fatty acids.

Considering our data of blunted histologic hepatotoxicitywith omega-3 fatty acid supplementation, we hypothesizethat this may be a method of preventing long-termcomplications of liver injury secondary to diseases ofcholestasis such as biliary atresia, namely fibrosis andcirrhosis. In light of the severity and prevalence of theseconditions, further pursuit of omega-3 fatty acids as apossible treatment modality is warranted.

References

[1] Deitch EA, Ma L, Ma WJ, et al. Inhibition of endotoxin-inducedbacterial translocation in mice. J Clin Invest 1989;84(1):36-42.

[2] Gehring S, Dickson EM, San Martin ME, et al. Kupffer cells abrogatecholestatic liver injury in mice. Gastroenterology 2006;130(3):810-22.

[3] Schmucker DL, Ohta M, Kanai S, et al. Hepatic injury induced by bilesalts: correlation between biochemical and morphological events.Hepatology 1990;12(5):1216-21.

[4] Fox CK, Furtwaengler A, Nepomuceno RR, et al. Apoptotic pathwaysin primary biliary cirrhosis and autoimmune hepatitis. Liver 2001;21(4):272-9.

[5] Galle PR, Krammer PH. CD95-induced apoptosis in human liverdisease. Semin Liver Dis 1998;18(2):141-51.

[6] Ando K, Guidotti LG, Wirth S, et al. Class I-restricted cytotoxicT lymphocytes are directly cytopathic for their target cells in vivo.J Immunol 1994;152(7):3245-53.

[7] Galle PR, Hofmann WJ, Walczak H, et al. Involvement of the CD95(APO-1/Fas) receptor and ligand in liver damage. J Exp Med 1995;182(5):1223-30.

[8] Faubion WA, Guicciardi ME, Miyoshi H, et al. Toxic bile salts inducerodent hepatocyte apoptosis via direct activation of Fas. J Clin Invest1999;103(1):137-45.

[9] Ariki N, Morimoto Y, Yagi T, et al. Activated T cells and solublemolecules in the portal venous blood of patients with cholestatic andhepatitis C virus-positive liver cirrhosis. Possible promotion of Fas/FasL-mediated apoptosis in the bile-duct cells and hepatocyte injury.J Int Med Res 2003;31(3):170-80.

[10] Malhi H, Gores GJ, Lemasters JJ. Apoptosis and necrosis in the liver:a tale of two deaths. Hepatology 2006;43(2 Suppl 1):S31-S44.

[11] Alwayn IP, Gura K, Nose V, et al. Omega-3 fatty acid supplementationprevents hepatic steatosis in a murine model of nonalcoholic fatty liverdisease. Pediatr Res 2005;57(3):445-52.

[12] Gura KM, Duggan CP, Collier SB, et al. Reversal of parenteralnutrition-associated liver disease in two infants with short bowelsyndrome using parenteral fish oil: implications for future manage-ment. Pediatrics 2006;118(1):e197-e201.

[13] Javid PJ, Greene AK, Garza J, et al. The route of lipid administrationaffects parenteral nutrition-induced hepatic steatosis in a mouse model.J Pediatr Surg 2005;40(9):1446-53.

[14] Gura KM, Lee S, Valim C, et al. Safety and efficacy of a fish-oil-basedfat emulsion in the treatment of parenteral nutrition-associated liverdisease. Pediatrics 2008;121(3):e678-86.

[15] Tazuke Y, Drongowski RA, Btaiche I, et al. Effects of lipidadministration on liver apoptotic signals in a mouse model of totalparenteral nutrition (TPN). Pediatr Surg Int 2004;20(4):224-8.

[16] Wang H, Khaoustov VI, Krishnan B, et al. Total parenteral nutritioninduces liver steatosis and apoptosis in neonatal piglets. J Nutr2006;136(10):2547-52.

[17] Van Aerde JE, Duerksen DR, Gramlich L, et al. Intravenous fish oilemulsion attenuates total parenteral nutrition-induced cholestasis innewborn piglets. Pediatr Res 1999;45(2):202-8.

[18] Schmocker C, Weylandt KH, Kahlke L, et al. Omega-3 fatty acidsalleviate chemically induced acute hepatitis by suppression ofcytokines. Hepatology 2007;45(4):864-9.

[19] Leaf A, Weber PC. Cardiovascular effects of n-3 fatty acids. N Engl JMed 1988;318(9):549-57.

[20] Schmidt EB, Dyerberg J. Omega-3 fatty acids. Current status incardiovascular medicine. Drugs 1994;47(3):405-24.

[21] Grimminger F, Wahn H, Mayer K. Impact of arachidonic versuseicosapentaenoic acid on exotonin-induced lung vascular leakage:relation to 4-series versus 5-series leukotriene generation. Am J RespirCrit Care Med 1997;155(2):513-9.

[22] Lee TH, Hoover RL, Williams JD, et al. Effect of dietary enrichmentwith eicosapentaenoic and docosahexaenoic acids on in vitroneutrophil and monocyte leukotriene generation and neutrophilfunction. N Engl J Med 1985;312(19):1217-24.

[23] Ishak K, Baptista A, Bianchi L, et al. Histological grading and stagingof chronic hepatitis. J Hepatol 1995;22(6):696-9.

[24] Strijbosch RA, Lee S, Arsenault DA, et al. Fish oil prevents essentialfatty acid deficiency and enhances growth: clinical and biochemicalimplications. Metabolism 2008;57(5):698-707.

[25] Chi TY, Chen GG, Lai PB. Eicosapentaenoic acid induces Fas-mediated apoptosis through a p53-dependent pathway in hepatomacells. Cancer J 2004;10(3):190-200.

[26] Avula CP, Zaman AK, Lawrence R, et al. Induction of apoptosis andapoptotic mediators in Balb/C splenic lymphocytes by dietary n-3 andn-6 fatty acids. Lipids 1999;34(9):921-7.