Menhaden Oil Decreases High-Fat Diet–Induced Markers of Hepatic Damage, Steatosis, Inflammation, and Fibrosis in Obese Ldlr^−/− Mice^,,

Abstract

The frequency of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) has increased in parallel with obesity in the United States. NASH is progressive and characterized by hepatic damage, inflammation, fibrosis, and oxidative stress. Because C20–22 (n-3) PUFA are established regulators of lipid metabolism and inflammation, we tested the hypothesis that C20–22 (n-3) PUFA in menhaden oil (MO) prevent high-fat (HF) diet–induced fatty liver disease in mice. Wild-type (WT) and Ldlr^−/− C57BL/6J mice were fed the following diets for 12 wk: nonpurified (NP), HF with lard (60% of energy from fat), HF–high-cholesterol with olive oil (HFHC-OO; 54.4% of energy from fat, 0.5% cholesterol), or HFHC-OO supplemented with MO (HFHC-MO). When compared with the NP diet, the HF and HFHC-OO diets induced hepatosteatosis and hepatic damage [elevated plasma alanine aminotransferase (ALT) and aspartate aminotransferases] and elevated hepatic expression of markers of inflammation (monocyte chemoattractant protein-1), fibrosis (procollagen 1α1), and oxidative stress (heme oxygenase-1) (P ≤ 0.05). Hepatic damage (i.e., ALT) correlated (r = 0.74, P < 0.05) with quantitatively higher (>140%, P < 0.05) hepatic cholesterol in Ldlr^−/− mice fed the HFHC-OO diet than WT mice fed the HF or HFHC-OO diets. Plasma and hepatic markers of liver damage, steatosis, inflammation, and fibrosis, but not oxidative stress, were lower in WT and Ldlr^−/− mice fed the HFHC-MO diet compared with the HFHC-OO diet (P < 0.05). In conclusion, MO [C20–22 (n-3) PUFA at 2% of energy] decreases many, but not all, HF diet–induced markers of fatty liver disease in mice.

Introduction

Nonalcoholic fatty liver disease (NAFLD)⁷ has increased in parallel with central obesity, and its prevalence is anticipated to continue to increase (1, 2). NAFLD is now the most common cause of liver disease in developed countries (3) and is defined as excessive lipid accumulation in the liver, i.e., hepatosteatosis (4, 5). The American Liver Foundation estimates that ~25% of the U.S. population has NAFLD and 75% of obese and 100% of morbidly obese individuals have NAFLD. NAFLD is the hepatic manifestation of metabolic syndrome (MetS) (4); MetS risk factors include obesity, elevated plasma TG and LDL cholesterol, reduced HDL cholesterol, high blood pressure, and fasting hyperglycemia (5).

NAFLD ranges in severity from simple fatty liver (steatosis) to nonalcoholic steatohepatitis (NASH) (6). Simple steatosis is relatively benign until it progresses to NASH, which is characterized by hepatic injury (hepatocyte ballooning and cell death), increased blood levels of hepatic enzymes [alanine aminotransferase (ALT)], hepatic inflammation, oxidative stress, and fibrosis (1, 2, 7). Approximately 30–40% of individuals with simple steatosis progress to NASH (8), and NASH can progress to cirrhosis (8), which is a major risk factor for hepatocellular carcinoma (2). In the “2 hit hypothesis” for NASH (9), the first hit involves chronic hepatosteatosis as TG and cholesterol (free cholesterol and cholesterol esters) accumulation. Excessive hepatic lipid sensitizes hepatocytes to the second hit, which is manifested by increased inflammation derived from resident (Kupffer cells) and recruited macrophages, induction of oxidative stress, activation of stellate cells, and fibrosis (2, 10).

Although the management of lifestyle (diet and exercise) is one approach to control the onset and progression of NAFLD, the best strategy for managing NAFLD has yet to be defined (11). On the basis of the well-established effects of C20–22 (n-3) PUFA to regulate hepatic lipid metabolism, dyslipidemia, and inflammation (12–14), we tested the hypothesis that dietary (n-3) PUFA prevents high-fat (HF) diet–induced fatty liver disease in mice. Recent clinical studies have indicated that dietary (n-3) PUFA have the potential to reduce hepatic lipid content in children and adults (15–19). Our studies, however, go beyond the analysis of hepatic lipids and examine the capacity of (n-3) PUFA to regulate markers of NASH, such as hepatic damage, inflammation, oxidative stress, and fibrosis. We used wild-type (WT) and Ldlr^−/− mice and 3 HF diets: HF lard (HF diet; 60% of energy from fat), which induces obesity and insulin resistance (20, 21), HF–high cholesterol with olive oil (HFHC-OO diet; 54.4% of energy from fat, 0.5% cholesterol), which induces fatty liver and oxidative stress (22), and the HFHC-OO diet supplemented with menhaden oil (MO) (HFHC-MO) a rich source of EPA [20:5(n-3)] and DHA [22:6(n-3)]. EPA and DHA in the HFHC-MO diet were 2% of total energy, a level comparable to that used to treat dyslipidemia (23). Our studies established that dietary C20–22 (n-3) PUFA have the capacity to regulate some, but not all, HF diet–induced markers of NASH.

Materials and Methods

Animals and diets.

All procedures for the use and care of animals for laboratory research were approved by the Institutional Animal Care and Use Committee at Oregon State University. Male WT and Ldlr^−/− mice (on the C57BL/6J background; Jackson Laboratories) at 2 mo of age consumed one of the following diets ad libitum for 12 wk: 1) Purina chow 5001 [nonpurified (NP); 13.5% of energy from fat and 58.0% of energy from carbohydrates]; 2) HF [60% of energy from fat; D12492; Research Diets]; 3) HFHC-OO [54.4% of energy from fat, 0.5% cholesterol (weight%); D08010702; Research Diets]; or 4) HFHC-MO [54.4% of energy from fat, 0.5% cholesterol (weight%); D08010703; Research Diets] (Supplemental Table 1). The HF, HFHC-OO, and HFHC-MO diets were described previously (21, 22). Fat energy density in the HFHC-OO and HFHC-MO diets was identical (54.4% energy from fat). C20–22 (n-3) PUFA in the HFHC-MO diet represented ~2% of total energy (Supplemental Table 1). All diets were stored frozen (−20°C) until used to feed the mice; diets were replenished every other day in an effort to reduce the formation of oxidation products.

The study was carried out twice with 8 mice/diet group in each study. Energy intake was monitored every other day, and body weight was monitored weekly. At the end of the 12-wk feeding period, all mice were feed-deprived overnight (1800 to 0800 the next day); then half of the mice were refed their diets for 4 h (0800 to 1200). Feed-deprived and refed mice were killed (isoflurane anesthesia and exsanguination) at 0800 and 1200, respectively, for the collection of blood and liver. Blood was collected in tubes containing EDTA; plasma was collected by centrifugation. Livers were removed, weighed, and rapidly frozen in liquid nitrogen. Plasma, blood cells, and liver were stored frozen (−80°C) until used for specific assays.

Our studies used WT and Ldlr^−/− mice fed the NP diet as controls. Because there was no significant difference (Student’s t test) in any variable measured between WT and Ldlr^−/− mice fed the NP diet, NP-fed Ldlr^−/− mice served as the control group for all studies described below. The following 6 group designations are used to describe the study: Ldlr^−/− mice fed NP (NP), WT mice fed HF (WT-HF), WT mice fed HFHC-OO (WT-HFHC-OO), WT mice fed HFHC-MO (WT-HFHC-MO), Ldlr^−/− mice fed HFHC-OO (Ldlr^−/− HFHC-OO), and Ldlr^−/− mice fed HFHC-MO (Ldlr^−/− HFHC-MO). During the course of our studies, we found that markers of NASH were induced to higher levels in Ldlr^−/− mice compared with in WT mice. As such, much of the data presented below describe the capacity of MO to regulate NASH markers in Ldlr^−/− mice. A comparison of the effects of the 4 diets on inflammation and fibrosis markers in WT and Ldlr^−/− mice is shown in Supplemental Table 3.

Measurement of plasma markers.

Plasma glucose (Autokit Glucose), TG (L-type TG H triglyceride), nonesterified fatty acids (NEFA; NEFA-C), and cholesterol (Cholesterol E) were measured with the use of kits from Wako. Plasma β-hydroxybutyrate was measured with the use of a kit (β-hydroxybutyrate Liquicolor) from Stanbio. Plasma apo B (ApoB K-Assay) and apo CIII (ApoCIII K-Assay) were measured by immunoturbidimetric assay from Kamiya Biomedical. Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were measured by using kits from Thermo Electron.

Measurement of urinary isoprostanes.

After 11 wk of being fed the NP or HFHC diets, Ldlr^−/− mice were placed in metabolic cages for 24-h urine collection. Collected urine was stored at <80°C until extracted for F2- and F3-isoprostanes and F4-neuroprostanes analysis. Results were normalized to urinary creatinine as described (22, 24).

Lipid extraction and analyses.

Total lipids were extracted from liver and plasma; FAME were prepared and quantified by gas chromatography (20). To measure hepatic total TG and cholesterol, extracted hepatic lipids were dried, dissolved in 10% Triton X-100 (Fisher Scientific), and assayed for TG and cholesterol content using the L-type TG H triglyceride and total cholesterol assay kits from Wako as described above (20). Hepatic protein was measured by using the Quick Start Bradford Reagent (Bio-Rad) and bovine serum albumin (Sigma) as the standard (20).

RNA extraction and qRT-PCR.

Total RNA was extracted from liver, and specific mRNA were quantified by qRT-PCR (20, 25). Specific primers for each gene were described previously (20, 25) or are listed in Supplemental Table 2. Cyclophilin was used as the internal control for all genes.

Immunoblotting.

Hepatic nuclear protein extracts were prepared by using both protease (Roche Diagnostics) and phosphatase inhibitors (1 mmol/L β-glycerol phosphate, 2.5 mmol/L Na-pyrophosphate, 1 mmol/L Na₃VO₄) (20). Proteins (25–100 μg) were separated electrophoretically and transferred to nitrocellulose for immunoblotting. TBP (TATA-binding protein) served as a loading control for all immunoblot studies. Antibodies used in these studies were described previously (20); antibodies for NF-κB-p50 and NF-κB-p65 were purchased from Santa Cruz Biotechnology and Cell Signaling, respectively.

Statistical analysis.

We used 1- and 2-way ANOVA and Tukey’s honestly significant difference post hoc analysis to establish significant differences. One-way ANOVA was used to detect dietary effects when only Ldlr^−/− mice were included in the analysis. Two-way ANOVA was used to detect diet-gene interactions, when both WT and Ldlr^−/− mice were included in the analysis and for the analysis of refeeding effects (Fig. 1A). Data were analyzed for homogeneous variances by using the Levene test. If unequal variances were detected, the data were log-transformed. ANOVA was performed on both transformed and untransformed data. Untransformed data are presented for interpretation purposes. The Student’s t test was used when only 2 groups were compared. P < 0.05 was considered different. The correlation analysis used linear regression analysis. Statistical analysis was performed with VassarStats (http://vassarstats.net/) and Statgraphics (StatPoint Technologies, Inc.). All values reported are means ± SD.

FIGURE 1.

The feed-deprived feeding response of hepatic nuclear SREBP1, ChREBP, and MLX (A) and hepatic lipogenic gene expression (B) in Ldlr^−/− mice fed the HFHC-OO or HFHC-MO diets for 12 wk. (A) After 12 wk of being fed these diets, the mice were deprived of food overnight. At 0800 the next day, one-half of the mice in each group were fed their diets, i.e., HFHC-OO or HFHC-MO. Feed-deprived and fed mice were killed at 0800 and 1200, respectively, for the quantitation of hepatic nuclear SREBP1, ChREBP, and MLX. Mean (±SD) results are expressed as nuclear protein abundance, fold of feed-deprived HFHC-OO; n = 8. Data were analyzed by 2-way ANOVA. Labeled means without a common letter differ, P < 0.05. (B) Hepatic RNA from the fed mice described above was used to quantify Acc1, Fasn, Scd1, and Elovl6 mRNA by qRT-PCR. Mean (±SD) results are expressed as mRNA abundance, fold of HFHC-OO; n = 8. Data were analyzed by Student’s t test. Labeled means without a common letter differ, P < 0.05. ACC1, acetyl CoA carboxylase-1; ChREBP, carbohydrate response element binding protein; Elov16, fatty acid elongase-6; FASN, fatty acid synthase; HFHC-MO, high-fat–high-cholesterol with olive oil supplemented with menhadin oil; HFHC-OO, high-fat–high-cholesterol with olive oil; MLX, max-like protein X; SCD1, stearoyl CoA desaturase-1; SREBP-1, sterol regulatory element binding protein-1.

Results

Body weight and plasma markers.

After 12 wk of being fed the NP and test diets, the body weights of WT mice fed the HF or HFHC diets and Ldlr^−/− mice fed the HFHC diets were >39% higher than those of the NP group (P < 0.05) (Table 1). Body weights of WT and Ldlr^−/− mice were not different between the groups fed the HF, HFHC-OO, or HFHC-MO diets. Plasma markers were measured in feed-deprived mice. Diet and genotype did not significantly affect plasma glucose. Plasma TG were >190% higher in the Ldlr^−/− HFHC-OO group compared with the NP group (P < 0.05). Plasma TG in the Ldlr^−/− HFHC-MO group were ~50% lower than in the Ldlr^−/− HFHC-OO group (P < 0.05) and not different from the NP group. Analysis of plasma cholesterol showed a diet effect (P < 0.05); plasma cholesterol in all mice fed the HFHC-OO diet was >60% higher than in all mice fed the HFHC-MO diet. There was also a genotype effect (P < 0.05): plasma cholesterol in all Ldlr^−/− mice was >330% higher than in all WT mice. No significant diet-by-genotype interaction for plasma cholesterol was observed. Plasma apo B and apo CIII paralleled diet-induced changes in plasma TG and cholesterol in the Ldlr^−/− HFHC-OO and Ldlr^−/− HFHC-MO groups. When compared with the NP group, plasma NEFA was nearly 90% higher in the Ldlr^−/− HFHC-OO group (P < 0.05). Finally, plasma β-hydroxybutyrate was >85% higher in the Ldlr^−/− HFHC-OO and Ldlr^−/− HFHC-MO groups when compared with the NP group (P < 0.05).

TABLE 1.

Phenotypic comparison of WT and Ldlr^−/− mice fed the NP, HF, HFHC-OO, or HFHC-MO diets for 12 wk¹

	WT				Ldlr−/−
	NP2	HF	HFHC-OO	HFHC-MO	HFHC-OO	HFHC-MO
Body weight,3 g	28 ± 2a	45 ± 4b	43 ± 1b	39 ± 2b	40 ± 5b	40 ± 5b
Plasma variables4
Glucose, mmol/L	6.0 ± 1.0	9.3 ± 2.7	8.3 ± 2.8	5.6 ± 0.7	11.9 ± 4.6	11.0 ± 4.6
TG,56 mg/dL	120 ± 44a	90 ± 16a	66 ± 27a	106 ± 25a	352 ± 141b	185 ± 82a
Cholesterol,35 mg/dL	52 ± 21a	108 ± 46a	138 ± 66b	81 ± 6a	957 ± 252b	599 ± 145b
Apo B,56 mg/dL	11 ± 2a	14 ± 2a	15 ± 4a	14 ± 2a	160 ± 70b	50 ± 21a
Apo CIII,56 mg/dL	13 ± 3a	13 ± 7a	10 ± 4a	8 ± 2a	21 ± 6b	11 ± 6a
NEFA,56 mEq/ml	0.8 ± 0.2a	0.7 ± 0.2a	0.6 ± 0.1a	0.7 ± 0.3a	1.5 ± 0.6b	0.9 ± 0.4a
β-Hydroxybutyrate,5 mmol/L	1.4 ± 0.3a	1.3 ± 0.4a	1.5 ± 0.3a	1.7 ± 0.2a	2.9 ± 0.8b	2.6 ± 0.5b
ALT,356 U/L	4 ± 1a	19 ± 10b	20 ± 11b	8 ± 2b	38 ± 9c	21 ± 16b,c
AST,356 U/L	10 ± 2a	18 ± 2a	30 ± 12b	13 ± 2a	56 ± 1c	27 ± 12b
Liver variables
Weight, g	1.0 ± 0.1	1.6 ± 0.5	1.9 ± 0.5	1.2 ± 0.1	2.0 ± 0.6	1.6 ± 0.4
% Body weight3	3.6 ± 0.1a	3.4 ± 1.0a	4.5 ± 1.2b	3.0 ± 0.2a	4.8 ± 0.9b	3.9 ± 0.7a
TG,37 μg/mg	51 ± 23a	157 ± 88b	141 ± 63b	72 ± 15a	129 ± 13b	64 ± 26a
Cholesterol,3568 μg/mg	7 ± 1a	6 ± 2a	8 ± 1a	8 ± 1a	17 ± 3b	10 ± 5a

¹Values are means ± SD; = 8/treatment group. Labeled means in a row with superscripts without a common letter differ, P < 0.05. ALT, alanine aminotransferase; AST, aspartate aminotransferase; HF, high-fat; HFHC-MO, high-fat–high-cholesterol with olive oil supplemented with menhadin oil; HFHC-OO, high-fat–high-cholesterol with olive oil; NEFA, nonesterified fatty acid; NP, nonpurified; WT, wild-type.

²NP ^−/− mice were used as the control group.

³Two-way ANOVA diet effect, < 0.05.

⁴To convert from mg/dL to g/L, multiply by 0.0113 for TG and 0.02586 for cholesterol. To convert from mg/dL to g/L, multiply by 0.01 for apo B and apo CIII. To convert from mEq/mL to mEq/L, multiply by 1000 for NEFA.

⁵Two-way ANOVA genotype effect, < 0.05.

⁶Two-way ANOVA diet × gene interaction effect, < 0.05.

⁷Data are expressed as g TG/mg protein (liver); to convert from μg/mg to mmol/g, multiply by 0.00113.

⁸Data are expressed as g cholesterol/mg protein (liver); to convert from μg/mg to mmol/g, multiply by 0.002586.

Hepatic damage.

Plasma ALT and AST are markers of hepatic damage; these markers were measured in feed-deprived mice. When compared with the NP group, ALT was elevated ⩾100% in the WT and Ldlr^−/− mice fed the HF, HFHC-OO, or HFHC-MO diets (P < 0.05), whereas AST was elevated ⩾30% in the WT-HFHC-OO, Ldlr^−/− HFHC-OO, and Ldlr^−/− HFHC-MO groups (P < 0.05) (Table 1). Plasma ALT and AST in the Ldlr^−/− HFHC-OO group were ⩾90% and ⩾46% higher, respectively, than in the WT-HFHC-OO group (P < 0.05). Plasma AST was lower (P < 0.05) and ALT tended to be lower (P = 0.12) in the Ldlr^−/− HFHC-MO group by ∼50% compared with the Ldlr^−/− HFHC-OO group.

Hepatic lipids.

Hepatic TG in the WT-HF, WT-HFHC-OO, and Ldlr^−/− HFHC-OO groups were ⩾150% higher than in the NP group (P < 0.05) (Table 1). This level of TG accumulation is comparable to that in WT mice fed the HF diet (20, 26, 27). Hepatic TG in the WT-HFHC-MO and Ldlr^−/− HFHC-MO groups were not different from the NP group. Hepatic cholesterol was >140% higher only in the Ldlr^−/− HFHC-OO group when compared with the NP group (P < 0.05) (Table 1). Supplementing the HFHC-OO diet with C20–22 (n-3) PUFA (i.e., the HFHC-MO diet) prevented the HFHC-OO diet–induced accumulation of hepatic TG and cholesterol in Ldlr^−/− mice.

The level of plasma ALT correlated with hepatic cholesterol (r = 0.74, P < 0.05), but not with hepatic TG, in WT and Ldlr^−/− mice. This outcome agrees with recent studies linking hepatic cholesterol to NASH (26–28). Because hepatic damage, as reflected by plasma ALT, was quantitatively greater by 90% in the Ldlr^−/− HFHC-OO group compared with the WT-HFHC-OO group (P < 0.05) (Table 1), the studies described below examined the effects of the HFHC-OO and HFHC-MO diets on hepatic fatty acids and markers of NASH in Ldlr^−/− mice (i.e., oxidative stress, inflammation, and fibrosis).

Hepatic fatty acids.

Hepatic fatty acid composition is significantly affected by diet. Because the HFHC-OO and HFHC-MO diets had similar affects on hepatic fatty acid profiles in WT and Ldlr^−/− mice, only the Ldlr^−/− data are shown (Table 2). Oleic acid [18:1(n-9)] is the predominant fatty acid in the HFHC-OO diet (Supplemental Table 1); the mole% of hepatic oleic acid was >260% and >150% higher in the Ldlr^−/− HFHC-OO and Ldlr^−/− HFHC-MO groups, respectively, when compared with the NP group (P < 0.05). Hepatic oleic acid in the Ldlr^−/− HFHC-MO group was >30% lower than in the Ldlr^−/− HFHC-OO group (P < 0.05). Hepatic 20:4(n-6) was 48% and 68% lower in the Ldlr^−/− HFHC-OO and Ldlr^−/− HFHC-MO groups, respectively, when compared with the NP group (P < 0.05). Hepatic C20–22 (n-3) PUFA were ~70% lower in the Ldlr^−/− HFHC-OO group compared with the NP group (P < 0.05), but was >200% higher in the Ldlr^−/− HFHC-MO group (P < 0.05). The ratio of 20:4(n-6) to 18:2(n-6) or 20:3(n-6) to 18:2(n-6) was the same in both the NP and HFHC-OO groups, but was ⩾50% lower in the Ldlr^−/− HFHC-MO group (P < 0.05) (Table 2). Similar diet effects were seen in plasma C20–22 (n-3) PUFA and the ratio of 20:4(n-6) to 18:2(n-6) (Supplemental Fig. 1) as well as in liver and plasma of WT mice (not shown).

TABLE 2.

Hepatic fatty acid composition in Ldlr^−/− mice fed the NP, HFHC-OO, or HFHC-MO diets for 12 wk¹

	NP	HFHC-OO	HFHC-MO
Fatty acids, mol%
16:0	26.9 ± 1.0a	18.0 ± 1.9b	14.0 ± 4.7b
16:1(n-7)	2.2 ± 0.7	3.4 ± 0.9	2.2 ± 1.9
18:0	11.7 ± 2.7a	2.4 ± 0.5b	4.0 ± 2.3b
18:1(n-9)	15.2 ± 1.8a	55.0 ± 4.3b	37.6 ± 3.7c
18:1(n-7)	3.1 ± 0.3a	7.3 ± 1.2b	5.1 ± 1.1c
18:2(n-6)	24.7 ± 2.8a	14.2 ± 1.6b	16.4 ± 3.0b
18:3(n-3)	0.8 ± 0.2	0.3 ± 0.1	0.3 ± 0.1
20:0	0.4 ± 0.1	0.3 ± 0.1	0.6 ± 0.1
20:2(n-6)	0.2 ± 0.1a	1.8 ± 0.1b	0.8 ± 0.2a
20:3(n-6)	0.7 ± 0.2	0.6 ± 0.1	0.4 ± 0.1
20:4(n-6)	6.2 ± 0.1a	3.2 ± 0.7b	2.0 ± 1.1b
20:5(n-3)	0.6 ± 0.1a	0.2 ± 0.1a	7.9 ± 2.3b
22:5(n-6)	0.2 ± 0.1	0.8 ± 0.2	0.6 ± 0.2
22:5(n-3)	0.6 ± 0.1a	0.2 ± 0.1a	2.0 ± 0.3b
22:6(n-3)	5.1 ± 0.9a	1.4 ± 0.3b	9.6 ± 3.8c
Σ C20–222
C20–22(n-3)	6.3 ± 0.6a	1.8 ± 0.3b	19.5 ± 3.3c
C20–22(n-6)	7.3 ± 1.3a	6.4 ± 0.8a	3.7 ± 1.1b
Fatty acid ratios
20:4(n-6) to 18:2(n-6)	0.26 ± 0.07a	0.27 ± 0.08a	0.14 ± 0.09b
20:3(n-6) to 18:2(n-6)	0.04 ± 0.02a	0.05 ± 0.01a	0.02 ± 0.01b

¹Values are means ± SD and are representative of 2 separate studies; = 4/treatment group. Labeled means in a row with superscripts without a common letter differ, P < 0.05. HFHC-MO, high-fat–high-cholesterol with olive oil supplemented with menhadin oil; HFHC-OO, high-fat–high-cholesterol with olive oil; NP, nonpurified.

²Sum C20-22 refers to the sum of all the 20 to 22 carbon fatty acids analyzed.

Whole-body and hepatic oxidative stress.

F2- and F3-isoprostanes and F4-neuroprostanes are nonenzymatic oxidation products of 20:4(n-6), 20:5(n-3), and 22:6(n-3), respectively (22, 24). Because urinary isoprostanes are a measure of whole-body oxidative stress (29), we analyzed urinary isoprostanes (F2 and F3) and neuroprostanes (F4) in the NP, Ldlr^−/− HFHC-OO, and Ldlr^−/− HFHC-MO groups (22, 24) (Table 3). F2- and F3-isoprostanes and F4-neuroprostanes in urine from the NP group were comparable to previous reports (30). F2-isoprostanes were not elevated in urine of the Ldlr^−/− HFHC-OO group, but increased by >110% in the Ldlr^−/− HFHC-MO compared with the NP group (P < 0.05). Although urinary F3-isoprostanes and F4-neuroprostanes were not significantly affected by the HFHC-OO diet, these oxidized lipids were >490% and >40% higher, respectively, in the urine of the Ldlr^−/− HFHC-MO group when compared with the NP and Ldlr^−/− HFHC-OO groups (P < 0.05). Supplementing the HFHC-OO diet with C20–22 (n-3) PUFA (i.e., the HFHC-MO diet) increased lipid peroxidation as measured by urinary F2- and F3-isoprostanes and F4-neuroprostanes.

TABLE 3.

Urinary iso- and neuroprostanes in Ldlr^−/− mice fed the NP, HFHC-OO, or HFHC-MO diets for 12 wk¹

	NP	HFHC-OO	HFHC-MO
	ng/mg creatinine
F2-isoprostane	2.2 ± 0.1a	2.6 ± 1.1a	4.7 ± 1.2b
F3-isoprostane	2.1 ± 0.1 a	0.9 ± 0.6a	12.5 ± 8.2b
F4-neuroprostane	70.7 ± 5.7a	45.5 ± 12.8a	110.3 ± 57.6b

¹Values are means ± SD of 4 pooled samples/diet group; samples were derived from 2 separate studies. Labeled means in a row with superscripts without a common letter differ, < 0.05. HFHC-MO, high-fat–high-cholesterol with olive oil supplemented with menhadin oil; HFHC-OO, high-fat–high-cholesterol with olive oil; NP, nonpurified.

To assess hepatic oxidative stress, we measured hemeoxygenase-1 (Hmox1) expression (Table 4). Hmox1 is induced in NAFLD in response to oxidative stress (31–36); it is also cytoprotective (31–33, 37). Hepatic Hmox1 mRNA was induced >300% and >200% (P < 0.05) in the Ldlr^−/− HFHC-OO and Ldlr^−/− HFHC-MO groups, respectively, when compared with the NP group. Supplementing the HFHC-OO diet with MO did not lower HFHC-OO–induced hepatic oxidative stress (i.e., Hmox1 expression) in Ldlr^−/− mice.

TABLE 4.

Hepatic gene expression in Ldlr^−/− mice fed the NP, HFHC-OO, or HFHC-MO diets for 12 wk¹

	NP	HFHC-OO	HFHC-MO
Oxidative stress	Fold of NP
Hmox1	1.0 ± 0.3 a	4.2 ± 1.0 b	3.2 ± 1.1 b
PUFA synthesis
Elovl2	1.0 ± 0.4a	2.2 ± 0.9b	2.4 ± 1.1b
Elovl5	1.0 ± 0.2 a	0.8 ± 0.1a	0.3 ± 0.1b
Fads1	1.0 ± 0.1 a	0.8 ± 0.3a	0.2 ± 0.1b
Fads2	1.0 ± 0.1	1.1 ± 0.2	0.7 ± 0.3
Cholesterol metabolism
HMG CoA red	1.0 ± 0.3	1.0 ± 0.2	1.2 ± 0.4
HMG CoA syn 1	1.0 ± 0.2	1.0 ± 0.1	0.9 ± 0.2
Cyp7A1	1.0 ± 0.3	1.2 ± 0.2	2.1 ± 1.3
Abca1	1.0 ± 0.2 a	4.7 ± 0.3b	4.0 ± 0.4b
Abcg5	1.0 ± 0.1a	3.2 ± 0.3b	4.0 ± 0.5b
Acat1	1.0 ± 0.4a	4.8 ± 1.5b	3.0 ± 1.1c
Acat2	1.0 ± 0.1a	2.0 ± 0.5b	2.3 ± 0.7b
Inflammation
Mcp1	1.0 ± 0.5a	22.8 ± 10.4b	4.6 ± 2.8a
Cd68	1.0 ± 0.5a	7.7 ± 2.5b	4.1 ± 2.5c
Clec4f	1.0 ± 0.5a	6.3 ± 1.2b	4.7 ± 1.3b
Fibrosis
Tgfβ1	1.0 ± 0.2a	3.0 ± 0.9b	2.1 ± 0.4b
Procol1α1	1.0 ± 0.7a	43.0 ± 24.0b	20.1 ± 14.0c

¹Values are means ± SD and are fold of NP; = 8. Labeled means in a row with superscripts without a common letter differ, P < 0.05. HFHC-MO, high-fat–high-cholesterol with olive oil supplemented with menhadin oil; HFHC-OO, high-fat–high-cholesterol with olive oil; NP, nonpurified.

Lipid-regulated hepatic transcription factors and gene expression.

Transcription factors controlling de novo lipogenesis (DNL), MUFA, and TG synthesis include sterol regulatory element binding protein-1 (SREBP1), carbohydrate response element binding protein (ChREBP), max-like protein X (MLX), liver X receptor (LXR), peroxisome proliferator-activated receptor α (PPARα), and hepatic nuclear factor 4α (12). We examined the effect of feed-deprivation and refeeding on the nuclear abundance of these transcription factors in Ldlr^−/− mice maintained on the HFHC-OO or HFHC-MO diets (Fig. 1). The nuclear abundance of SREBP1 and ChREBP, but not MLX, was increased by ⩾250% (P < 0.05) in Ldlr^−/− mice 4 h after refeeding the HFHC-OO group. The nuclear abundance of SREBP1 in the Ldlr^−/− HFHC-MO group increased by 100% 4 h after refeeding (P < 0.05). This level of induction, however, is 60% lower when compared with the Ldlr^−/− HFHC-OO group (P < 0.05) (Fig. 1A). Because the induction of Srebp-1 mRNA abundance was not different between the Ldlr^−/− HFHC-OO and Ldlr^−/− HFHC-MO groups (Supplemental Fig. 2), MO suppression of hepatic SREBP1 nuclear abundance involved a post-translational mechanism (38). The refeeding response of ChREBP was not affected by MO. The nuclear abundance of other transcription factors controlling hepatic metabolism (i.e., LXRα, PPARα, HNF4α, and forkhead box protein O1,) remained unaffected by diet (not shown).

We next measured the abundance of Acc1, Fasn, Scd1, and Elovl6 mRNA in livers of the refed mice described above. Of these transcripts, hepatic abundance of Acc1 and Fasn was suppressed by 50% (P < 0.05) in the Ldlr^−/− HFHC-MO group compared with the Ldlr^−/− HFHC-OO group (Fig. 1B). PPARα regulates the expression of multiple genes involved in fatty acid oxidation (12), including acyl CoA oxidase (Aox), CYP450-4A10 (Cyp4A10), acyl-CoA thioesterase-1 (Acot1), carnitine palmitoyl transferase 1 (Cpt1), and Cpt2. None of these transcripts were affected by the HFHC-OO or HFHC-MO diets when compared with the NP diet (not shown).

Enzymes involved in converting essential fatty acids [18:2(n-6) and 18:3(n-3)] to C20–22 (n-6) and (n-3) PUFA include fatty acid elongases (Elovl2 and Elovl5) and desaturases (Fads1 and Fads2). Expression of these enzymes is controlled, at least in part, by SREBP1 and PPARα (12, 39–41). Hepatic Elovl2 was induced 120% (P < 0.05) in the Ldlr^−/− HFHC-OO and Ldlr^−/− HFHC-MO groups when compared with the NP group (Table 4). Hepatic Elovl5 and Fads1 mRNA were reduced by 70–80% (P < 0.05) in the Ldlr^−/− HFHC-MO group compared with the NP or Ldlr^−/− HFHC-OO groups. Hepatic Fads2 was not affected by diet. Changes in Elovl5 and Fads1 expression paralleled the decline in the hepatic 20:4/18:2 ratio in Ldlr^−/− mice fed the HFHC-MO diet (Table 2).

Because the combination of the HFHC-OO diet and the Ldlr^−/− genotype significantly increased hepatic cholesterol (Table 1), we examined pathways controlling hepatic cholesterol content (Table 4). The mRNA abundance of SREBP2 target genes (Hmg CoA red and Hmg CoA syn1) was not affected. Of the LXR target genes examined, only the ABC transporters (Abca1, Abcg5) were induced ⩾200% in livers of Ldlr^−/− mice fed the HFHC-OO and HFHC-MO diets when compared with the NP group (P < 0.05). Acat1 is expressed in Kupffer cells (42), whereas Acat2 is expressed in hepatocytes (43). The expression of hepatic Acat1 and Acat2 mRNA was induced 380% and 100%, respectively, by the HFHC-OO diet when compared with the NP diet. Hepatic Acat1 mRNA, however, was 38% lower in the Ldlr^−/− HFHC-MO group when compared with the Ldlr^−/− HFHC-OO group (P < 0.05) (Table 4).

Hepatic inflammation.

We next examined markers of inflammation. Monocyte chemoattractant protein-1 (Mcp1) mRNA, an early marker of inflammation (44), was induced ⩾600% in the WT-HF and WT-HFHC-OO groups (P < 0.05) (Supplemental Table 3) and >2000% in the Ldlr^−/− HFHC-OO group (P < 0.05) when compared with the NP group (Table 4). Furthermore, Mcp1 mRNA was >170% higher in the Ldlr^−/− HFHC-OO group compared with the WT-HFHC-OO group (P < 0.05). In contrast to the HFHC-OO diet, the HFHC-MO diet did not induce hepatic Mcp1 mRNA in either genotype (Table 4, Supplemental Table 3).

Cd68 and Clec4f are cell surface markers of monocytes and resident macrophages (Kupffer cells), respectively. Hepatic Cd68 mRNA was induced 170% in the WT-HFHC-OO group, >600% in the Ldlr^−/−HFHC-OO group, and >300% (P < 0.05) in the Ldlr^−/− HFHC-MO group when compared with the NP group (all P < 0.05) (Table 4, Supplemental Table 3). A comparison of the Ldlr^−/− HFHC-OO and the Ldlr^−/− HFHC-MO groups had a ~50% reduction in Cd68 mRNA (P < 0.05) (Table 4). Clec4f mRNA was elevated by >500% and >370% in the Ldlr^−/− HFHC-OO and Ldlr^−/−HFHC-MO groups, respectively, compared with the NP group (P < 0.05). Hepatic Clec4f was not induced by either HF or HFHC-OO diet in WT mice (Supplemental Table 3).

NF-κB-p50 and NF-κB-p65 are transcription factors that control the expression of multiple genes involved in inflammation (45). We examined the hepatic nuclear abundance of NF-κB-p50 and NF-κB-p65 in Ldlr^−/− mice fed the NP, HFHC-OO, and HFHC-MO diets. The nuclear content of NF-κB-p50 and NF-κB-p65 was elevated ⩾100% and ~200%, respectively in mice fed the HFHC-OO diet compared with the NP diet (both P < 0.05). While mice fed the NP or HFHC-MO diets had comparable levels of nuclear NF-κB-p50, levels of NF-κ-B-p65 in the HFHC-MO group remained elevated compared with the NP group, similar to the HFHC-OO fed group (Fig. 2).

FIGURE 2.

Hepatic nuclear abundance of NF-κB-p50 and NF-κB-p65 in Ldlr^−/− mice fed the NP, HFHC-OO, or HFHC-MO diets for 12 wk. Ldlr^−/− mice were deprived of food overnight before hepatic nuclear extracts were prepared and used to quantify the nuclear abundance of NF-κB-p50 and NF-κB-p65. Mean (±SD) results are expressed as nuclear NF-κB, fold change of NP; n = 4. Labeled means without a common letter differ, P < 0.05. HFHC-MO, high-fat–high-cholesterol with olive oil supplemented with menhadin oil; HFHC-OO, high-fat–high-cholesterol with olive oil; NP, nonpurified.

Hepatic fibrosis.

Markers of fibrosis examined in this study include TGF-β1 (Tgfβ1) and procollagen 1α1 (Procol1α1) (46–49). These proteins are expressed in Kupffer and stellate cells but not in hepatic parenchymal cells. Tgfβ1 mRNA was not induced in the WT-HF or WT-HFHC-OO groups compared with the NP group (Supplemental Table 3) but was induced 200% and 110% in Ldlr^−/−mice fed the HFHC-OO and HFHC-MO diets, respectively (P < 0.05) (Table 4). Procol1α1 mRNA was induced 1400%, 4200%, and 1900% in the WT-HFHC-OO, Ldlr^−/− HFHC-OO, and Ldlr^−/− HFHC-MO groups, respectively (P < 0.05) (Supplemental Table 3; Table 4). Procol1α1 mRNA abundance in the WT-HFHC-MO and NP groups was not different. Procol1α1 mRNA abundance in the Ldlr^−/− HFHC-MO group, however, was >50% lower when compared with the Ldlr^−/− HFHC-OO group (P < 0.05).

Discussion

We examined the effect of 3 HF diets on the development of fatty liver disease in WT and Ldlr^−/− mice. WT mice fed the HF and HFHC-OO diets and Ldlr^−/− mice fed the HFHC-OO diet for 12 wk had increased body weight (⩾39%) and multiple markers of dyslipidemia and NASH, including hepatic damage (plasma ALT and AST), hepatosteatosis, oxidative stress [urinary isoprostanes and hepatic Hmox1 (Ldlr^−/− only)], inflammation (mRNA encoding Mcp1, Cd68, Clec4f), and fibrosis (Tgfβ1 and Procol1α1) (Tables 1, 3, and 4; Supplemental Table 3). Although hepatic TG content increased equally in the WT-HF, WT-HFHC-OO, and Ldlr^−/− HFHC-OO groups, the excess accumulation of hepatic cholesterol in Ldlr^−/− mice correlated with hepatic damage (plasma ALT vs. hepatic cholesterol; r = 0.74, P < 0.05) and the induction of markers of inflammation and fibrosis. Thus, the combination of diet lipid composition and genotype contributed to the severity of fatty liver disease in mice.

Because fatty liver disease was more robust in Ldlr^−/− mice fed the HFHC-OO diet than in WT mice fed the same diet, most of our studies focused on determining if MO supplementation of the HFHC-OO diet (i.e., the HFHC-MO diet) could prevent diet-induced fatty liver disease in Ldlr^−/− mice. As expected, feeding Ldlr^−/− mice the HFHC-MO diet significantly elevated hepatic and plasma C20–22 (n-3) PUFA (Table 2, Supplemental Fig.1) and reduced the expression of many NASH markers when compared with Ldlr^−/− mice fed the HFHC-OO diet (Tables 1 and 4). The exception was oxidative stress.

Both Ldlr^−/− HFHC-OO and Ldlr^−/− HFHC-MO groups expressed the hepatic oxidative stress marker Hmox1 at levels >300% above those seen in the NP group (Table 4). Moreover, Ldlr^−/− mice fed the HFHC-MO diet had increased urinary isoprostanes (Table 3). Hmox1 expression is controlled by nuclear factor E2-related factor-2 (NRF2), a transcription factor regulated by reactive oxygen species (ROS) (35). NRF2 binds antioxidant response elements in the Hmox1 promoter as well as other promoters responsive to ROS (e.g., Gstα1). Induction of hepatic Hmox1 by the HFHC-OO and HFHC-MO diets is an indirect marker of elevated hepatic ROS.

NAFLD patients have elevated oxidative stress (31–36), and dietary supplement approaches that counter oxidative stress have been tested as a potential NAFLD therapy (50, 51). Our finding that dietary C20–22 (n-3) PUFA increased urinary isoprostanes may raise concerns for their use in NAFLD therapy (Table 3) (52). Interestingly, urinary F2-isoprostane levels were not different in the NP and HFHC-OO groups, but increased > 100% in the HFHC-MO group. One explanation for this outcome is the impact of dietary fat composition on tissue (n-6) and (n-3) PUFA content. In contrast to the NP diet, the HFHC-OO diet contains high levels of SFA and MUFA, relative to (n-6) PUFA (Supplemental Table 1). Hepatic 20:4(n-6), the substrate for F2-isoprostane, was lower in the HFHC-OO group than in the NP group (Table 2). Thus, low tissue 20:4(n-6) content may not favor increased F2-isoprostane formation, even when hepatic ROS is elevated. Moreover, the HFHC-OO diet did not suppress the expression of hepatic enzymes involved in PUFA synthesis (Table 4). The hepatic and plasma ratio of 20:4(n-6) to 18:2(n-6) also indicated that (n-6)PUFA synthesis was not suppressed by the HFHC-OO diet (Table 2, Supplemental Fig. 1). One explanation for the decline in hepatic 20:4(n-6) in the Ldlr^−/− HFHC-OO group is that synthesized 20:4(n-6) is exported via VLDL and stored in extrahepatic tissues. Additional studies are required to define the impact of HF diets on hepatic and whole-body 20:4(n-6) metabolism.

Although the HFHC-OO and HFHC-MO diets have the same SFA, MUFA, and (n-6) PUFA content, they differ in (n-3) PUFA content; the HFHC-MO diet increased urinary isoprostanes (Table 3). In contrast to the HFHC-OO group, low hepatic 20:4(n-6) in the HFHC-MO group (Table 2) can be explained by low expression of enzymes involved in PUFA synthesis and a 50% reduction in the 20:4(n-6) to 18:2(n-6) ratio (Tables 1, 2;Fig. 1B). Moreover, the HFHC-OO and HFHC-MO diets induced hepatic oxidative stress (i.e., Hmox1) (Table 4). In livers with elevated C20–22 (n-3) PUFA, F3-isoprostane and F4-neuroprostane content was elevated (22). Increased production of isoprostanes from C20–22 (n-3) PUFA may stimulate F2-isoprostane formation (22, 52). F2-isoprostanes activate thromboxane and prostaglandin F2α receptors; they activate platelets and promote smooth muscle cell–mediated vasoconstriction (52, 53), whereas isoprostanes generated from (n-3) PUFA do not (52). By analogy to the proinflammatory (series 2) and antiinflammatory (series 3) cyclooxygenase-derived eicosanoids (54), F3-isoprostanes may be antiinflammatory. If so, some effects of C20–22 (n-3) PUFA on inflammatory markers reported in Table 4 and Figure 2 may be attributed to the action of F3-isoprostanes and F4-neuroprostanes. Clearly, more study is required to establish whether isoprostanes generated from (n-3) PUFA are beneficial.

Because (n-3) PUFA inhibit hepatic fatty acid synthesis and induce fatty acid oxidation (12, 55), we hypothesized that increasing hepatic (n-3) PUFA would suppress hepatic accumulation of TG and prevent the induction of all markers linked to fatty liver disease. Major regulators of DNL include SREBP1 and the ChREBP/MLX heterodimer (56, 57); these transcription factors play a role in hepatosteatosis (2, 12, 38, 58, 59). The nuclear abundance of SREBP1, but not the ChREBP/MLX heterodimer, was vulnerable to C20–22 (n-3) PUFA regulation in obese Ldlr^−/− mice (Fig. 1A). The suppression of mRNA encoding Acc1 and Fasn parallels the suppression of hepatic nuclear abundance of SREBP1 (Fig. 1B). Although C20–22 (n-3) PUFA suppression of nuclear SREBP1 accounts for some reduction of hepatic TG, other mechanisms are likely involved. In humans with NAFLD, fatty acids entering hepatic TG are derived from multiple sources, including DNL, portal circulation, and NEFA mobilized from adipose tissue (10, 60, 61). In feed-deprived mice, plasma NEFA were reduced by 40% in the Ldlr^−/− HFHC-MO group compared with the Ldlr^−/− HFHC-OO group (Table 1). Thus, (n-3) PUFA control of lipid mobilization from adipose tissue and delivery of NEFA to the liver may contribute to the low hepatic TG in the Ldlr^−/− HFHC-MO group.

The combination of HFHC-OO and the Ldlr^−/− genotype was required to increase hepatic cholesterol content by >140% (P < 0.05) (Table 1) and induced high levels of expression of inflammation and fibrosis markers (Table 4). Several recent reports have linked elevated hepatic cholesterol to the induction of NASH markers (26–28, 62). In those studies, increased hepatic cholesterol activated SREBP2, upregulated the LDL receptor, reduced biotransformation of bile acids, and suppressed canalicular pathways for cholesterol and bile acid export (28, 62). In our study, however, the LDL receptor is absent, and SREBP2 target genes and the expression of CYP7A1, the rate-limiting step in bile acid synthesis, were not affected by the HFHC-OO or HFHC-MO diets (Table 4). As such, alternative mechanisms are needed to explain our observation. Because both HFHC-OO and HFHC-MO diets induced Abca1 and Abcg5, cholesterol efflux may not be affected. The very high plasma TG and cholesterol in feed-deprived Ldlr^−/− mice (Table 1) likely promotes LDL uptake via hepatic scavenger receptors. Using the same diet and mice as reported herein, Saraswathi et al. (22) found that dietary C20–22 (n-3) PUFA decreased hepatic cholesterol, but increased cholesterol storage in adipose tissue. Ldlr^−/− mice fed the HFHC-MO diet had significantly lower apo B, apo CIII (a lipoprotein lipase inhibitor), and plasma TG and cholesterol when compared with the HFHC-OO group (Table 1). Therefore, part of the control of hepatic TG and cholesterol may involve VLDL assembly and secretion as well as VLDL/LDL clearance by nonhepatic tissues. More studies are required to verify this mechanism.

A key finding of our study was the effect of the HFHC-OO and HFHC-MO diets on markers of NASH in nonparenchymal hepatic cells. Cd68 and Clec4f are expressed in macrophage and Kupffer cells (27), whereas Tgfβ1 is expressed predominantly in Kupffer cells and stellate cells, but not in parenchymal cells. Procol1α1 is expressed in stellate cells (46). Feeding WT and Ldlr^−/− mice the HFHC-OO diet induced expression of mRNA and nuclear proteins linked to inflammation (Mcp1, Cd68, Clec4f, and nuclear NF-κB-p50 and -p65) and fibrosis (Tgfβ1, Procol1α1) (Fig. 2, Table 4, Supplemental Table 3). More important, supplementing the diet with MO (HFHC-MO) lowered the levels of most of these markers (except for Clec4f and Tgfβ1) when compared with HFHC-OO–fed mice.

Although we have a good understanding of C20–22 (n-3) PUFA control of hepatic parenchymal cell gene expression (12), much less is known about (n-3) PUFA regulation of Kupffer and stellate cells. Mcp1 is an early marker of inflammation and is regulated by NF-κB (44, 63). In the Ldlr^−/− HFHC-OO group, the nuclear abundance of NF-κB-p50 and NF-κB-p65 was significantly induced when compared with the NP group (Fig. 2). In mice fed the HFHC-MO diet, however, the nuclear abundance of NF-κB-p50 was not induced when compared with the NP group (Fig. 2). Changes in Mcp1 mRNA paralleled changes in hepatic nuclear NF-κB-p50 content. This outcome suggests that genes requiring the classical p50/p65 heterodimer or the p50/50 homodimer for transcriptional activation will be attenuated by C20–22 (n-3) PUFA, whereas inflammatory genes that are regulated by p65 heterodimerization with other partners, such as c/EBPβ, will be less sensitive to C20–22 (n-3) PUFA control. In addition to NF-κB, other regulatory factors control the expression of hepatic inflammatory and fibrotic markers, including TGF-β, cJun, TNF-α, and the Toll-like receptors 2 and 4 (26, 64). Procol1α1 expression in stellate cells is induced by TGF-β1 through a H₂O₂-c/EBPβ-dependent mechanism (48, 49). The HFHC-MO diet had a modest effect on Tgfβ1 expression in Ldlr^−/− mice but significantly attenuated Procol1α1 expression in the WT and Ldlr^−/− HFHC-MO groups when compared with the WT and Ldlr^−/− HFHC-OO groups (Table 4, Supplemental Table 3). A systematic analysis of C20–22 (n-3) PUFA effects on hepatic target genes controlled by these regulatory factors will clarify the breadth of C20–22 (n-3) PUFA effects on hepatic inflammation and fibrosis markers.

In summary, we used an established model of HF- diet–induced fatty liver disease in mice (22) and assessed the capacity of dietary C20–22 (n-3) PUFA in MO to prevent fatty liver disease in WT and Ldlr^−/− mice. Ldlr^−/− mice developed more severe fatty liver disease than did WT mice when fed the HFHC-OO diet. Mice fed the HFHC-MO diet, which was supplemented with a physiologically relevant level (2% of total energy) of C20–22 (n-3) PUFA, had lower plasma and hepatic lipids (TG and cholesterol), hepatic damage, inflammation, and fibrosis than did mice fed the HFHC-OO diet. The HFHC-MO diet, however, did not prevent the induction of other HF diet–induced markers of fatty liver disease or MetS, including obesity, hyperglycemia, and hepatic or whole-body oxidative stress. Thus, the C20–22 (n-3) PUFA in MO reduced many, but not all, metabolic abnormalities associated with HF diet–induced fatty liver disease in mice.