Dietary factors alter hepatic innate immune system in mice with nonalcoholic fatty liver disease


  • See Editorial on Page 755

  • Potential conflict of interest: Nothing to report.


Dietary factors promote obesity and obesity-related disorders, such as fatty liver disease. Natural killer T (NKT) cells are components of the innate immune system that regulate proinflammatory (Th-1) and anti-inflammatory (Th-2) immune responses. Previously, we noted that NKT cells are selectively reduced in the fatty livers of obese, leptin-deficient ob/ob mice and demonstrated that this promotes proinflammatory polarization of hepatic cytokine production, exacerbating lipopolysaccharide (LPS) liver injury in these animals. In the current study, we show that hepatic NKT cells are also depleted by diets that induce obesity and fatty livers in wild-type mice, promoting Th-1 polarization of hepatic cytokine production and sensitization to LPS liver injury despite persistent leptin. Adult male C57BL6 mice fed diets containing high amounts of either fat or sucrose, or combined high-fat, high-sucrose, develop increased hepatic NKT cell apoptosis and reduced liver NKT cells. The hepatic lymphocytes are more Th-1 polarized with increased intracellular interferon gamma and tumor necrosis factor alpha. Mice fed high-fat diets also exhibit more liver injury, reflected by 2-fold greater serum alanine aminotransferase (ALT) than control animals after receiving LPS. In conclusion, when otherwise normal mice are fed with high-fat or sucrose diet, they become obese, develop fatty livers, and acquire hepatic innate immune system abnormalities, including increased NKT cell apoptosis. The latter reduces liver NKT cell populations and promotes excessive hepatic production of Th-1 cytokines that promote hepatic inflammation. These diet-induced alterations in the hepatic innate immune system may contribute to obesity-related liver disease. (HEPATOLOGY 2005;42:880–885.)

Of various environmental factors that might contribute to the rising incidence of obesity-related diseases, changes in dietary habits merit particular consideration because diets that are enriched in certain macronutrients (e.g., polyunsaturated fats, fructose, or sucrose) induce both obesity and insulin resistance in rats and mice.1, 2 There is also evidence that changes in these dietary factors influence insulin sensitivity in humans,3 and this may explain why Westernized diets are associated with an increased prevalence of type 2 diabetes. Because the incidence of various obesity-related diseases, including type 2 diabetes, fatty liver, glomerulonephritis, cataracts, neuropathy, and epithelial neoplasm, are increased in “over-nourished” rodents,1 these models are used to study the pathogenesis of such diet-induced diseases.

Obesity is also strongly associated with nonalcoholic fatty liver disease (NAFLD). Fatty livers are unusually susceptible to injury induced by a secondary inflammatory stress, including that evoked by exposure to endogenous, intestine-derived lipopolysaccarhide (LPS).4 However, the mechanisms underlying the susceptibility of fatty liver are not well understood. Our previous studies have shown that in genetically obese, leptin-deficient ob/ob mice,5 natural killer T cells (NKT cells) play important roles in fatty liver vulnerability to LPS.6, 7

NKT cells are components of the innate immune system. They express both T cell surface marker (e.g., CD3) and NK cell surface marker (i.e., NK1.1). These cells originate in the thymus but predominately accumulate in the liver, where they regulate local proinflammatory (Th-1) and anti-inflammatory (Th-2) cytokine production by other mononuclear cells.8 In leptin-deficient ob/ob mice, the hepatic NKT cells are depleted.6 The depletion of hepatic NKT cells leads to increased local production of proinflammatory (Th-1) cytokines.7 These cytokines mediate fatty liver vulnerability to LPS in these animals.7 However, compared with obese humans who have increased levels of leptin, ob/ob mice are genetically leptin-deficient and leptin itself is now known to have potent immunomodulatory actions.9 Therefore, it is questionable whether our previous findings in ob/ob mice are relevant to human fatty liver disease.

To address this issue, we studied another animal model of NAFLD. In the new model, obesity and fatty liver were induced by feeding diets enriched with fat and/or sucrose. Animals with diet-induced obesity have elevated leptin levels, similar to obese humans. Our objectives were to determine whether immune system alterations (i.e., hepatic NKT cell depletion, Th-1 polarization of hepatic cytokine-producing cells, and enhanced sensitivity to LPS-induced liver injury) that occur in leptin-deficient mice also occur in mice with diet-induced NAFLD and increased leptin levels.


NAFLD, nonalcoholic fatty liver disease, LPS, lipopolysaccharide; NKT cells, natural killer T cells; HMNC, hepatic mononuclear cell; TNF-α, tumor necrosis factor alpha; IFN-γ, interferon gamma; IL, interleukin.

Materials and Methods

Animal Experiments.

Adult (aged 6–8 weeks) male wild-type C57BL-6 mice were purchased from Jackson Laboratories (Bar Harbor, ME). The mice were fed with commercial diets with different nutrition contents (Table 1) for 4 to 12 weeks. All mice were maintained in a temperature- and light-controlled facility, and permitted ad libitum consumption of water and pellet chow. For LPS experiments, the mice were injected intraperitoneally with a single dose of Escherichia coli LPS (Sigma Chemical, St. Louis, MO) (50 μg/mouse) and then killed after 0, 1.5, or 6 hours to obtain serum and liver tissue. All animal experiments fulfilled National Institutes of Health and Johns Hopkins University criteria for the humane treatment of laboratory animals.

Table 1. Diet Composition
  1. NOTE. Percentage of calorie source from fat, sucrose, carbohydrates (other than sucrose), and protein in normal, high-fat, high-sucrose, and high-fat/high-sucrose diet.

Total carbohydrate59%25%24%9%

Hepatic Mononuclear Cell Isolation and Cell Surface Labeling.

Hepatic mononuclear cells (HMNC) were isolated and labeled using a minor modification of the method we described previously.6 Anti-mouse NK1.1-PE, CD3-FITC, CD4-APC, CD8-PerCP, Annexin-V-PE, and 7-AAD were obtained from Pharmingen (San Diego, CA). After surface labeling, HMNC were evaluated by flow cytometry (Becton Dickinson, Palo Alto, CA). Data are analyzed by Cell Quest software (Becton Dickinson).

Liver Mononuclear Cell Intracellular Cytokines Labeling.

After isolation, HMNC were incubated with phorbol 1,2-myristate 1,3-acetate (50 ng/mL, Sigma), ionomycin (500 ng/mL, Sigma), and GolgiPlug (1 μL/mL, Pharmingen). Cells were labeled with surface antibody as described previously and then permeabilized with Cytoperm/Cytofix (Pharmingen) according to the manufacturer's instruction. After permeabilization, cells were further labeled for intracellular cytokines such as anti-mouse tumor necrosis factor alpha (TNF-α) and interferon gamma (IFN-γ, Pharmingen). After incubation, cells were evaluated by flow cytometry. Data are analyzed as described in previous paragraph.

RNA Isolation and Ribonuclease Protection Assays.

As we have described.10 the method of Chomzynski and Sacchi11 was used to isolate total hepatic RNA. To evaluate cytokine gene expression, ribonuclease protection assays were performed with commercial kits (Pharmingen) that contained probes for groups of cytokine mRNAs, as well as internal standards for 2 different constitutively expressed mRNAs (L-32 and glyceraldehyde-3-phosphate dehydrogenase). Twenty micrograms total RNA was used per assay. After separation on 5% acrylamide gels, the mRNA hybridization signals were quantified by phosphoimager analysis and then visualized by exposing the dried gel to x-ray film. Specific mRNA species were identified by comparing their position on the gel with that of a concurrent standard according to the manufacturer's instructions.

Serum Alanine Aminotransferase and IFN-γ Levels.

Alanine aminotransferase levels were measured by a multi-channel autoanalyzer in the Clinical Chemistry Laboratory of the Johns Hopkins University, Department of Comparative Medicine. Interferon gamma levels were measured with in vivo capture assay (Pharmingen), using standards supplied by the manufacturer.

Statistical Analysis.

All values are expressed as mean ± SD. The group means were compared by ANOVA, using Microsoft Excel (Microsoft, Redmond, WA).


High-Fat Diets Induce Weight Gain and Fatty Liver in Wild-Type Mice.

Our previous study showed hepatic NKT cell depletion in leptin-deficient ob/ob mice, which leads to proinflammatory (Th-1) cytokine polarization and increased susceptibility to LPS-induced liver injury. In this study, we hypothesize that similar mechanisms occur in wild-type mice with diet-induced fatty livers, despite high leptin levels. To test this hypothesis, adult male C57BL6 mice were fed diets containing high amounts of fat (Table 1). Controls received an isocaloric volume of similar diet that contained less fat. As expected, mice fed the high-fat diets gained significantly more weight than control mice that were fed normal diets (Fig. 1A). The mice fed with high-fat diet also developed hepatic steatosis (Fig. 1C).

Figure 1.

High-fat diets induce weight gain and hepatic steatosis. (A) The weight of C57BL6 mice fed with high-fat diet. *P < .05 high-fat diet versus normal diet (n = 32). (B) Hematoxylin-eosin stain of liver histology from mice fed with normal diet (original magnification ×160). (C) Hematoxylin-eosin stain of liver histology from mice fed with high-fat diet (original magnification ×160).

High-Fat Diets Cause Selective Hepatic NKT Cell Depletion in Wild-Type Mice.

We then evaluated hepatic mononuclear cell subtypes using cell surface markers and flow cytometry. Mice fed high-fat diets had significantly fewer hepatic NKT cells (Fig. 2A). The decrease in liver NKT cells was predominately attributable to reduced numbers of CD4+ NKT cells. In contrast, mice that ate the normal diet and mice that consumed the high-fat diet had similar numbers of splenic NKT cells (Fig. 2B), indicating that diet-induced NKT cell depletion is tissue-specific. Preliminary studies suggest that hepatic NKT cell numbers remain constant before high fat-fed mice develop significant steatosis after consuming the high-fat diet for 1 week (data not shown). However, more studies are needed to better understand the temporal relationship between development of steatosis and NKT cell depletion. Next, we fed another group of wild-type mice high-sucrose diets, which also induce fatty liver and insulin resistance, and noted similar hepatic NKT cell depletion (Fig. 3). Interestingly, subsequent studies with combined high-fat and high-sucrose diets demonstrated that the latter diets cause even further reduction of hepatic NKT cells, suggesting an additive detrimental effect of high dietary fat and sucrose on this cell population (Fig. 3).

Figure 2.

Effect of high-fat diets on hepatic and splenic NKT cells. (A) Wild-type C57BL6 mice were fed with high-fat diet. Total liver mononuclear cells were harvested and analyzed by flow cytometry. Total liver mononuclear cell number was similar between mice fed with high-fat and normal diets. Therefore, the percentage of NKT cells reflects the number of NKT cells in the liver. *P < .05 normal diet versus high-fat diet (n = 16). (B) Total splenic mononuclear cells were isolated from wild-type C57BL6 mice fed with high-fat diet and analyzed by flow cytometry (n = 16). NKT, natural killer T.

Figure 3.

Effect of dietary factors on hepatic NKT cell content. Wild-type C57BL6 mice were fed either with high-fat, high-sucrose, or combination of high-fat/high-sucrose diet (see Table 1 for dietary contents). At the end of the treatment period, liver mononuclear cells were harvested and analyzed by flow cytometry (n = 16). NKT, natural killer T.

High-Fat Diets Increase Hepatic Proinflammatory Cytokine Production.

Because liver NKT cells play an important role in regulating cytokine production by other liver mononuclear cells, we next evaluated the effects of diet on liver mononuclear cell cytokine profiles. Compared with control mice, hepatic total mononuclear cells, T cells, and NKT cells from high-fat–fed mice produce significantly more proinflammatory cytokines, TNF-α and IFN-γ (Fig. 4). In addition, serum levels of IFN-γ are significantly higher in mice fed with high-fat diet compared with levels in mice fed with normal diet. This indicates that more proinflammatory cytokines are released, leading to a systemic proinflammatory state.

Figure 4.

Effect of high-fat diets on hepatic cytokine profiles and serum IFN-γ. The intracellular cytokines (TNF-α and IFN-γ) from hepatic total mononuclear cell T cell (TMNC) were measured by flow cytometry. (A) Representative flow cytometry data from normal diet– and high-fat diet–fed mice shows different cell populations with positive IFN-γ staining. x = gated as TMNC; y = gated as NKT cells (NKT cells usually are CD3low); z = gated as CD3+ T cells. (B) Mean (±SD) results of duplicate experiments of intracellular IFN-γ staining are graphed (n = 8). (C) Mean (±SD) results of duplicate experiments of intracellular TNF-α staining are graphed (n = 8). (D) Serum levels of IFN-γ in high-fat diet and normal diet mice were measured by in vivo capture assays (n = 4). *P < .05 normal diet versus high-fat diet. **P < .01 normal diet versus high-fat diet. IFN-γ, interferon gamma; TNF-α, tumor necrosis factor alpha; NKT, natural killer T.

High-Fat Diets Increase Hepatic Susceptibility to LPS-Induced Injury.

To determine whether these differences in hepatic cytokine production influence sensitivity to LPS-mediated liver injury, control and high-fat/high-sucrose diet-fed mice were treated with a small dose of LPS. Six hours after LPS treatment, serum alanine aminotransferase levels were twice as high in mice fed high-fat/high-sucrose diets as in mice that were fed the normal diets (Fig. 5). The histology showed that high-fat–fed mice had more inflammation and necrosis in the liver after LPS treatment (data not shown). These findings are consistent with the concept that diet-induced increases in proinflammatory cytokine production sensitize mice to LPS hepatotoxicity.

Figure 5.

Effect of high-fat diets on liver injury. Serum ALT levels, a marker of hepatic injury, in mice fed with high-fat diet after LPS (n = 8). ALT, alanine aminotransferase; LPS, lipopolysaccharide.

High-Fat Diets Increase Hepatic Interleukin-12, a Cytokine That Reduces NKT Cell Viability.

Our subsequent efforts were directed at determining mechanisms by which dietary factors might reduce liver NKT cells. The differentiation and viability of NKT cells themselves are regulated by cytokines. In particular, interleukin-12 (IL-12) and IL-18 significantly reduce, and IL-15 increases, NKT cell viability. Therefore, increased IL-12 and IL-18 or reduced IL-15 decrease liver NKT cells. Previous studies from our laboratory have demonstrated increased IL-12 and IL-18, and decreased IL-15, in leptin-deficient ob/ob mice with NAFLD and reduced hepatic NKT cells. Therefore, in the current study, we determined the effect of diet on liver expression of these cytokines. Hepatic production of IL-12, a cytokine that reduces NKT cell viability, is increased approximately 2-fold by high-fat diets (Fig. 6). However, unlike leptin-deficient ob/ob mice, hepatic IL-15 and IL-18 expression was not changed in high-fat diet–fed mice (Fig. 6).

Figure 6.

Effect of high-fat diets on hepatic expression of cytokines that regulate NKT cells. IL-12, IL-15, IL-18, and GAPDH expression were evaluated by ribonuclease protection assays. Representative autoradiographs are shown. GAPDH was used as an internal standard. Because of relatively weak expression of IL-12, IL-15, and IL-18, prolonged exposure of the autoradiographs is required, causing the GAPDH signal to appear over-exposed. To accurately measure signal intensity, phosphoimager analysis was used. Mean (SD) results from duplicate phosphoimager analysis experiments are graphed. P value indicates difference of IL-12 between high-fat diet–fed mice and normal diet–fed mice. There is no statistically significant difference in IL-15 or IL-18 expression between high-fat diet–fed mice and normal diet–fed mice (n = 6). NKT, natural killer T; IL, interleukin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

High-Fat Diets Increase Hepatic NKT Cell Apoptosis.

As mentioned earlier, IL-12 has been shown to reduce NKT cell viability. To determine whether NKT cell viability was reduced in our mice with diet-related increases in IL-12, we compared NKT cell death in mice fed normal or high-fat diets. In these studies, isolated liver mononuclear cells were incubated concurrently with annexin-V and 7-AAD to evaluate both apoptosis and necrosis. NKT cell apoptosis (as identified as annexin-V–positive and 7-AAD–negative NKT cells) was significantly increased in mice that were fed high-fat diets (Fig. 7).

Figure 7.

Effect of high-fat diets on hepatic NKT cell apoptosis. Hepatic mononuclear cells were isolated from the mice fed with high-fat diet and normal diet. Flow cytometry was used to identify different mononuclear cell subpopulations, and NKT cell apoptosis was measured by annexin-V staining with concurrent incubation of 7-AAD to assess cell necrosis. Apoptotic NKT cells are defined as annexin-V+/7-AAD of NKT cells (n = 8). NKT, natural killer T.


In liver, the innate immune system is predominately responsible for balancing the production of various pro- and anti-inflammatory cytokines. This function is accomplished by populations of cells that share features of both classical T cells and natural killer (NK) cells.12, 13 On activation, NKT cells respond with vigorous cytokine production. They are capable of producing both Th1-type cytokines, including IFN-γ and TNF, and Th2-type cytokines, such as IL-4 and IL-13.14, 15 Our previous studies in leptin-deficient ob/ob mice show that although these mice have normal numbers of splenic NKT cells, their livers are selectively deficient in certain NKT cell populations.6 Hepatic NKT cell depletion is associated with reduced hepatic expression of anti-inflammatory (Th-2) cytokines,6, 16 increased hepatic expression and activity of proinflammatory (Th-1) cytokines,7, 16 hepatic insulin resistance, and nonalcoholic steatohepatitis. In the current study, we show that similar defects develop in wild-type, leptin-sufficient mice when they are fed diets enriched with nutrients (fat, sucrose) that promote obesity. This finding suggests a novel mechanism that links dietary factors to the induction of an obesity-related chronic inflammatory state that promotes chronic liver disease.

Of various environmental factors that might contribute to the rising incidence of obesity-related diseases, changes in dietary habits merit particular consideration because diets that are enriched in certain macronutrients (e.g., fats, fructose, or sucrose) induce both obesity and insulin resistance in rodents and human.1–3 Growing evidence has pointed to a correlative and causative relation between inflammation and insulin resistance. Indeed, population-based studies of obese humans link insulin resistance to systemic inflammation.17, 18 Key factors such as TNF-α and IKKβ, which regulate inflammatory signaling, are now known to mediate diet-induced insulin resistance.19, 20 Taken together, these observations suggest that dietary factors may initiate signals that promote inflammation and insulin resistance. The current studies indicate a potential mechanism for this in the liver, namely, diet-induced depletion of the hepatic NKT cells that balance local production of pro- and anti-inflammatory cytokines. Age, sex, and other genetic factors all influence susceptibility to diet-induced metabolic diseases. We used adult C57BL6 mice in this initial study because previous studies have shown this strain is particularly susceptible to diet-induced fatty liver and insulin resistance.21 Further evaluation of other mouse strains, and different age groups and sexes, will be necessary to clarify whether any of these factors modulate susceptibility to diet-related changes in hepatic NKT cells. In addition to inducing NKT cell apoptosis (Fig. 7), dietary factors might reduce liver NKT cells by interfering with mechanisms that mediate the hepatic homing of extrahepatic NKT cells. NKT cells originate in the thymus and migrate to the liver, where they accumulate.8 Because the mechanisms that regulate this process are uncertain, we have not investigated diet effects on NKT cell migration. However, because it is known that local cytokines, such as IL-12, IL-18, and IL-15, play critical roles in maintaining hepatic NKT cell homeostasis,22, 23 we assessed the effect of diet on the local concentration of these NKT cell viability factors. We found significant increases in both IL-12 (Fig. 6), an inducer of NKT cell apoptosis,16 and NKT cell apoptosis (Fig. 7) in the livers of mice fed with high-fat diets Therefore, increased IL-12 likely contributes to depletion of hepatic NKT cells in these animals. Lipids presented by CD1 protein also regulate NKT cell viability.24 The possibility that this process might be disturbed in steatotic livers is particularly intriguing. A recent study suggests that a lysosomal glycosphingolipid, isoglobotrihexosylceramide (iGb3), is the endogenous ligand for the CD1 molecule.25 Although we do not yet know whether dietary factors alter the hepatic expression of CD1 or its putative ligand, iGb3, mice with diet-induced NAFLD provide clinically relevant models that can be used to study this and other mechanisms that regulate the hepatic innate immune system in obesity. Given growing evidence that links NAFLD with the insulin resistance (metabolic) syndrome, future research in this area has broad relevance.

NKT cells are a diverse group of cells that are composed of different subsets, each with unique phenotypical markers and functional characters.12, 13 Previous studies with diseased human tissue show there appear to be fewer classic invariant NKT cells in human liver than in mouse liver.26, 27 Recently, using healthy donor liver tissue and classic NKT surface markers, Morsy and colleagues showed that there are comparable amounts of classic NKT cells in human and mouse livers.28 Because there are also CD1-restricted hepatic NKT cells that lack classic NK1.1 surface markers,29 and our approach did not evaluate these, future studies with tetramers are needed to evaluate the impact of diet on this NKT cell population. Nevertheless, our findings are important because they clearly demonstrate significant dietary effects on “classic” NKT cells and cytokine production by other liver mononuclear cells.