Dietary fatty acids modulate antigen presentation to hepatic NKT cells in nonalcoholic fatty liver disease.

Dietary fatty acids are major contributors to the development and progression of insulin resistance and nonalcoholic fatty liver disease (NAFLD). Dietary fatty acids also alter hepatic NKT cells that are activated by antigens presented by CD1d. In the current study, we examine the mechanism of dietary fatty acid induced hepatic NKT cell deficiency and its causal relationship to insulin resistance and NAFLD. We discover that dietary saturated fatty acids (SFA) or monounsaturated fatty acids (MUFA), but not polyunsaturated fatty acids (PUFA), cause hepatic NKT cell depletion with increased apoptosis. Dietary SFA or MUFA also impair hepatocyte presentation of endogenous, but not exogenous, antigen to NKT cells, indicating alterations of the endogenous antigen processing or presenting pathway. In vitro treatment of normal hepatocytes with fatty acids also demonstrates impaired ability of CD1d to present endogenous antigen by dietary fatty acids. Furthermore, dietary SFA and MUFA activate the NFkappaB signaling pathway and lead to insulin resistance and hepatic steatosis. In conclusion, both dietary SFA and MUFA alter endogenous antigen presentation to hepatic NKT cells and contribute to NKT cell depletion, leading to further activation of inflammatory signaling, insulin resistance, and hepatic steatosis.

. Dietary fatty acids are also the major contributing factor to metabolic syndrome, coincidence conditions of obesity, insulin resistance, hypertension, dyslipidemia, and fatty liver diseases ( 6 ). Dietary SFAs induce both obesity and insulin resistance in experimental animals and humans ( 7,8 ). These observations provide the basis for the hypotheses that altered immune function by dietary fatty acids may contribute to the development of metabolic syndrome. In fact, infl ammatory conditions are well linked to insulin resistance, obesity, and fatty liver diseases ( 9,10 ). Human population studies have also linked insulin resistance to systemic infl ammation ( 11,12 ). In liver, natural killer T (NKT) cells are the key mediator between infl ammatory condition and hepatic insulin resistance and steatosis ( 6,13 ).
NKT cells are a group of "unconventional" T cells that express markers characteristic for both natural killer (NK) cells and T cells. They are most abundant in the liver and can balance the production of pro-infl ammatory and anti-infl ammatory cytokines ( 14,15 ). Our previous studies have shown that high-fat diets induce hepatic NKT cell depletion and lead to local and systematic infl ammatory conditions that contribute to insulin resistance and fatty liver diseases ( 6 ). However, there is little knowledge about the mechanism by which dietary fatty acids regulate hepatic NKT cells. An earlier study has shown that imbalance in lipid metabolism can alter lipid antigen presentation to NKT cells ( 16 ).
In the current study, we evaluated the effect of different kinds of dietary fatty acids on hepatic NKT cells. Importantly, we investigated mechanisms underlying the immuno-modulatory function of dietary fatty acids and their role in the infl ammatory process, insulin resistance, Abstract Dietary fatty acids are major contributors to the development and progression of insulin resistance and nonalcoholic fatty liver disease (NAFLD). Dietary fatty acids also alter hepatic NKT cells that are activated by antigens presented by CD1d. In the current study, we examine the mechanism of dietary fatty acid induced hepatic NKT cell defi ciency and its causal relationship to insulin resistance and NAFLD. We discover that dietary saturated fatty acids (SFA) or monounsaturated fatty acids (MUFA), but not polyunsaturated fatty acids (PUFA), cause hepatic NKT cell depletion with increased apoptosis. Dietary SFA or MUFA also impair hepatocyte presentation of endogenous, but not exogenous, antigen to NKT cells, indicating alterations of the endogenous antigen processing or presenting pathway. In vitro treatment of normal hepatocytes with fatty acids also demonstrates impaired ability of CD1d to present endogenous antigen by dietary fatty acids. Furthermore, dietary SFA and MUFA activate the NF B signaling pathway and lead to insulin resistance and hepatic steatosis. In conclusion, both dietary SFA and MUFA alter endogenous antigen presentation to hepatic NKT cells and contribute to NKT cell depletion, leading to further activation of infl ammatory signaling, insulin resistance, and hepatic steatosis. Supplementary key words infl ammation • obesity • insulin resistance Dietary fatty acids have been shown to infl uence im mune function ( 1 ). Polyunsaturated fatty acids (PUFA) can modulate cytokine production, lymphocyte proliferation, and apoptosis ( 2,3 ). Saturated fatty acids (SFA) affect MHC I-mediated antigen presentation, impair immune response, and increase host susceptibility to infection speed sorter (Becton Dickenson). Hepatocytes were isolated by in situ perfusion of the liver then cultured on collagen I-coated plates as described previously ( 17 ).

In vitro NKT cell activation and proliferation assays
V ␣ 14 + mouse CD1d-specifi c NKT hybridoma cells (DN32.D3) were cocultured with a mouse fi broblast cell line (LCD1dwt) that has been stably transfected with CD1d and naturally expresses an endogenous NKT cell ligand ( 18 ), in the presence or absence of FFAs, such as palmitoleic acid (PA), oleic acid (OA), and docosahexaenoic acid (DHA) (Sigma-Aldrich, St Louis, MO). IL-2 released from NKT cells to the media were determined by commercially available, mouse-compatible ELISAs (R and D systems, Inc., Minneapolis, MN) with recombinant murine cytokines as standards. For the NKT cell proliferation assay, freshly isolated primary liver NKT cells were incubated with CD1d-Ig based artificial antigen presenting cells (aAPC) as previously described ( 19 ) in the presence or absence of FFAs for 12 days. The number of viable NKT cells was determined by trypan blue exclusion.

Hepatocytes antigen presentation to NKT cells
After isolation, some hepatocytes were loaded with ␣ -GalCer (100ng/ml, ALEXIS Biochemical, San Diego, CA) for 4 h and washed extensively before coculture with NKT hybridoma cells. Other untreated hepatocytes were used to determine and hepatic steatosis. Results from our current study provide a better understanding of the relation between nutrition and immune function.

Hepatocyte and NKT cell isolation
Hepatic mononuclear cells (HMNC) were isolated as previously described ( 6 ). HMNCs were then labeled with CD1d tetramers (NIH tetramer facility) loaded with a ligand (PBS-57, an analog of ␣ -galactosylceramide (GalCer) or anti-mouse fl uorescent antibodies against CD8, CD4, CD3, and NK1.1 (Pharmingen, San Diego, CA). After surface labeling, HMNCs were analyzed by fl ow cytometry (Becton Dickinson, Palo Alto, CA), and NKT cells (NK1.1 + , CD3 + ) were isolated using a FACSVantage SE high To maintain the same adequate amount of dietary EFA among all high-fat diets, a small amount (5.3%) of corn oil was added to SFA and PUFA diets. Each type of high-fat diet has the same total fat content (35.85%) and EFA composition (3.5%). Abbreviations: EFA, essential fatty acid; PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid. ( Fig. 2B ). As previous published ( 6 ), the total HMNC content is not infl uenced by high fat diets. Therefore the percentage of hepatic NKT cells collates their number in the liver. Hepatic NKT cells produce both the pro-infl ammatory cytokine, IFN-␥ , and the anti-infl ammatory cytokine, IL-4. We examined the cytokine profi le of remaining hepatic NKT cells after different dietary fatty acid treatment. Intracellular cytokine staining was performed in conjunction with cell surface marker labeling. Dietary MUFA causes a signifi cant decrease in IL-4 expression, while dietary PUFA causes a signifi cant decrease in IFN ␥ expression in NKT cells ( Fig. 2C, D ). Therefore, dietary fatty acids not only alter the hepatic NKT cell content but also impact the NKT cell cytokine profi les.

Dietary fatty acids impair hepatocyte presentation of endogenous antigen to NKT cells
NKT cells recognize endogenous and exogenous lipid antigens presented by the MHC class I like molecule, CD1d. It has previously been shown that hepatic CD1d down-regulation contributes to NKT cell depletion ( 22 ), thus we evaluated the hepatic expression of CD1d in animals fed different fatty acid-enriched diets. Overall, diets of high fatty acid content led to lower hepatic CD1d expression ( Fig. 3A ). However, none of the fatty acid enriched diets caused a statistically signifi cant reduction of CD1d expression in hepatocytes (data not shown). We isolated hepatocytes from animals treated with different fatty acid diets and incubated these hepatocytes with the NKT cell hybridoma, DN32.D3. Hepatocytes from SFA and MUFA fed animals show a decreased ability to stimulate NKT cells, as refl ected by decreased IL-2 production ( Fig. 3B ). When these hepatocytes were loaded with an exogenous ligand, ␣ -GalCer, they show a similar ability to stimulate NKT cells regardless of whether they were from normal diet or high fat diet fed mice ( Fig. 3C ). The stimulation of NKT cells by hepatocytes is CD1d dependent because there was no effect by hepatocytes isolated from CD1d knockout (CD1dko) mice ( Fig. 3B ). These results indicate that diets enriched by SFA or MUFA impair hepatocyte presentation of endogenous antigen to NKT cells by either inhibiting the endogenous ligand or by interfering with the interaction between the endogenous ligand and CD1d. To further confi rm the effect of fatty acids on hepatocyte presentation of endogenous ligand to NKT cells, we isolated normal hepatocytes from wild-type animals fed normal diet and treated them in vitro with PA, OA or DHA to represent the treatment of SFA, MUFA or PUFA, respectively. In vitro fatty acid treatment has little effect on hepatocyte expression of CD1d ( Fig. 4A ). However, in vitro SFA (PA) treatment does reduce the ability of hepatocytes to stimulate NKT cells ( Fig. 4B ), which can be restored by loading exogenous ligand ␣ -GalCer ( Fig. 4C ). Our studies show that both in vivo and in vitro fatty acid treatment reduced the ability of hepatocytes to stimulate NKT cells.

Dietary fatty acids induce infl ammatory signaling
As mentioned earlier, hepatic NKT cells can secrete both pro-infl ammatory cytokines (IFN-␥ ) and anti-infl ammatory their endogenous lipid antigen presentation to NKT cells. IL-2 production from NKT hybridoma cells indicated NKT cell activation by hepatocytes. In separate experiments, freshly isolated hepatocytes from normal diet fed mice were first incubated with individual FFAs for 24 h. After extensive washing, these hepatocytes were then cocultured with NKT hybridoma cells.

Glucose tolerance test (GTT), liver histology and triglyceride content, RNA isolation, hepatic gene expression, Western blot, and DNA binding activity of NF-B
All these experiments were performed as previously described( 13 ).

Statistical analysis
All values are expressed as mean ± SE. Treatment related differences were evaluated by ANOVA. The paired-individual means were compared by t -test .

Dietary saturated and monounsaturated, but not polyunsaturated, fatty acids induce insulin resistance, hepatic steatosis, and NKT cell depletion
Our previous study showed that a high fat diet with mixed fatty acids (serving as the control high fat diet in the current study, HFC) causes obesity, hepatic steatosis and NKT cell depletion ( 6 ). Other studies have produced confl icting results concerning which kind of fatty acids induce obesity and insulin resistance ( 20,21 ). In order to better understand which dietary fatty acid contributes to high fat diet-induced metabolic syndrome, we fed wildtype C57BL6 mice with custom made diets enriched with saturated fatty acids (SFA), or monounsaturated fatty acids (MUFA), or polyunsaturated fatty acids (PUFA) for 12 weeks. Each diet has a similar energy composition of fats, carbohydrates and proteins, but differs in the predominant fatty acid added ( Table 1 ). The results show that SFA and MUFA diets, but not the PUFA diet, induce greater weight gain in wild-type mice, despite a similar total caloric intake among these animals ( Fig. 1A ). Mice fed SFA and MUFA diets also developed insulin resistance as refl ected by glucose tolerance tests ( Fig. 1B ). Since nonalcoholic fatty liver disease (NAFLD) is closely associated with obesity and insulin resistance, we evaluated the liver from these mice for NAFLD. Mice fed SFA and MUFA diets had signifi cant hepatic steatosis as seen in liver histology ( Fig. 1C ) as well as signifi cantly increased hepatic triglyceride contents ( Fig. 1D ). Therefore, it appears that different dietary fatty acids impact differently on the high fat diet-induced obesity, insulin resistance and fatty liver disease. Next, we tried to identify the mechanism of which dietary fatty acids exert their effects.
It has been shown that the development of obesity, insulin resistance and hepatic steatosis is associated with selective hepatic NKT cell depletion. We examined hepatic NKT cell contents in animals fed different high fatty acid diets. Dietary SFA and MUFA, but not PUFA induced signifi cant hepatic NKT cell depletion ( Fig. 2A , B ). Hepatic T cells are not affected by any kind of dietary fatty acids have signifi cantly increased hepatic IKK-␤ activity, as indicated by an increased ratio of phospho-I B ␣ to total-I B ␣ ( Fig. 5B ). We further evaluated NF B binding activities, the downstream signal for I B ␣ , with similar fi ndings. SFA and MUFA diets increased NF B binding activity, although only the MUFA diet group reached statistical signifi cance ( Fig. 5C ). Because adipokines made by fat tissue also contribute greatly to the regulation of infl ammation and insulin resistance, we evaluated the serum levels of leptin and adiponectin. SFA and MUFA diets signifi cantly increased the serum level of leptin ( Fig. 5D ). This was consistent with the increased obesity of the animals on SFA or MUFA diets. No differences in the adiponectin level were observed among diet groups ( Fig. 5D ) .

DISCUSSION
High dietary fat intake is known to be associated with obesity, insulin resistance, and NAFLD ( 6,23,24 ). However, the impact of dietary fat composition on the formation of cytokines (IL-4). Also, they play an important role in regulating hepatic cytokine production and balancing infl ammatory response. Hepatic NKT cell depletion is associated with excessive pro-infl ammatory cytokine production and activation of infl ammatory signaling ( 13 ). To evaluate the effect of fatty acid diets on hepatic infl ammatory signaling, we examined the hepatic cytokine expression profi le. SFA and MUFA, but not PUFA, diets signifi cantly increased the expression of IFN-␥ and decreased the expression of IL-4 ( Fig. 5A ), indicating a pro-infl ammatory hepatic cytokine profi le was induced by SFA and MUFA diets. These results represent the cytokine profi le in the whole liver tissue, which is different from the cytokine profi le of NKT cells shown in Fig. 2 . In the liver, there are many other cells that also produce IFN-␥ , such as NK cells. The cytokine profi le of the whole liver is the sum of all cells, including NKT cells. IKK-␤ , which phosphorylates I B ␣ , is a key factor that mediates diet-induced obesity and insulin resistance ( 9 ). Therefore, the ratio of phospho-I B ␣ to total-I B ␣ represents IKK-␤ activity. SFA and MUFA diet-fed mice Fig. 1. SFA and MUFA diets, but not PUFA diet, cause obesity, insulin resistance, and fatty liver disease. Wild-type C57BL6 mice were fed diets enriched with certain fatty acids (see Table 1) for 12 weeks. A: animal weight, (B) glucose tolerance tests, (C) liver histology, and (D) hepatic triglyceride contents were performed. (* P < 0.05 versus ND) (n = 4/group). HFC, high fat control diet; MUFA, monounsaturated fatty acid; ND, normal diet; PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid.
due to the diffi culty in performing dose-saturation studies of dietary fatty acid composition, we cannot draw a defi nitive conclusion on these synergistic effects.
The current study confi rms the previous fi nding that hepatic NKT cells are key contributors of high fat dietinduced hepatic steatosis ( 6,13 ). SFA-or MUFA-induced fatty liver is associated with depletion of hepatic NKT cells. Dietary fatty acids also cause a signifi cant alteration of NKT cell cytokine profi le. Because NKT cells have many subgroups and each subgroup has a unique profi le of cytokine production ( 25 ), these fi ndings could result from modifi cations of NKT cell subgroups by different NAFLD is still poorly understood. In some studies, the dietary intake of high amounts of saturated fatty acids and low amounts of polyunsaturated fatty acids is linked to patients with NAFLD ( 20 ). Another study showed a high intake of n-6 fatty acids in patients with nonalcoholic steatohepatitis (NASH) ( 21 ). In our current study, we have shown that dietary saturated and monounsaturated fatty acids play a much more signifi cant role in the pathogenesis of NAFLD than dietary polyunsaturated fatty acids. It also appears that a high-fat diet with mixed fatty acids (i.e., the high fat control diet in this study) may have synergistic effects on the formation of obesity, insulin resistance, and NAFLD. However, Fig. 2. SFA and MUFA diets selectively depleted hepatic NKT cells and alter NKT cell cytokine profi le. Wild-type C57BL6 mice were fed diets enriched with certain fatty acids (see Table 1) for 12 weeks. Hepatic mononuclear cells (HMNC) were isolated and NKT cells were identifi ed with CD1d tetramer loaded with a ligand and colabeled with IFN-␥ , IL-4, and CD3. The total HMNC content was not infl uenced by high fat diets. Therefore, the percentage of hepatic NKT cells refl ected their number in the liver. MHC class I-like molecule CD1d, which is noncovalently associated with ␤ 2-microglobulin ( 27 ). Previous study shows that diets containing high saturated or unsaturated fatty acids result in similar alterations in the phospholipid class distribution and fatty acid composition in the liver ( 28 ). It is known that lipid accumulation due to dysregulation of lipid metabolism can cause defective presentation of lipid antigens to NKT cells and impacts NKT cell accumulation ( 16 ). Little is known, however, about the effect of dietary fatty acids on CD1d-mediated NKT cell activation.
dietary fatty acids. We are currently investigating the effect of dietary fatty acids on NKT cell subgroups.
The ability of certain fatty acids to infl uence the function of different components of the immune system has been recognized for many years ( 26 ). SFA and MUFA infl uence major histocompatibility complex (MHC) I-mediated antigen presentation ( 5 ). SFA causes a signifi cant reduction in MHC class I surface expression and inhibits in the rate of antigen presenting cell (APC)-T cell conjugation. NKT cells recognize a group of glycolipid antigens bound to the  Table 1  However, VanderLaan et al. have shown that dendritic cells preincubated with serum from high-fat diet-fed mice can stimulate NKT cells ( 34 ). This difference could be due to the difference between hepatocytes and dendritic cells in response to high-fat diets. In fact, our study also found a signifi cant difference between hepatocytes and Kupffer cells in their ability of presenting endogenous antigen to NKT cells (see supplementary data), but Kupffer cells are equally effective in presenting exogenous antigen ( ␣ -GalCer) to NKT cells. Whether dietary fatty acids enhance or reduce the ability of Kupffer cells to present endogenous antigen is worth future investigation.
Previous studies have also shown that CD1d-mediated antigen presentation is regulated by both mitogen-activated protein kinases (MAPK) and protein kinase C (PKC) ␦ activities ( 35,36 ). Inhibition of PKC ␦ signifi cantly impairs antigen presentation by CD1d to NKT cells ( 35 ). In addition, it is known that dietary fatty acids regulate PKC activity ( 37 ). Compared with PUFA (DHA), SFA (PA) results in much less PKC activity in human hepatoma cells. It is possible that the change of dietary fatty acid composition alters PKC activity and impairs antigen presentation by CD1d to NKT cells. We are currently examining this hypothesis.

CONCLUSION
We have shown that high dietary SFA and MUFA, but not PUFA, cause hepatic NKT cell depletion, thus contributing to the formation of insulin resistance and hepatic steatosis. Dietary fatty acid-induced hepatic NKT cell depletion is In the current study, we have shown for the fi rst time that high-fat diets impair the ability of CD1d to present endogenous antigen to NKT cells. This likely contributes to the mechanism of dietary fatty acid-induced hepatic NKT cell depletion. The exact endogenous ligand for NKT cells is still in debate ( 29,30 ). Previous studies show that high-fat diets alter the composition of liver ceramides, which are likely the antigens for NKT cells ( 31 ). It will be interesting to investigate which candidate endogenous lipid has been altered by SFA and MUFA diets.
NKT cells also recognize ␣ -galactosylceramide ( ␣ -GalCer), a marine sponge-derived glycolipid presented by CD1d ( 32 ). In the liver, hepatocytes are the main source of CD1d expression and antigen presentation to NKT cells ( 33 ). A previous study showed that endoplasmic reticulum stress and lipid accumulation in hepatocytes of leptin-defi cient ob/ob mice decrease hepatocyte CD1d expression and NKT cell activation ( 22 ). In the current study, we demonstrate that dietary fatty acids cause a slight, but not statistically signifi cant, decrease of CD1d expression in hepatocytes. More importantly, both dietary fatty acids in vivo and fatty acid treatment in vitro signifi cantly reduced the ability of hepatocytes to present endogenous antigen to NKT cells. The ability of hepatocytes to present an exogenous antigen ( ␣ -GalCer) was not affected by either high fatty acid diet treatment in vivo or fatty acid treatment in vitro. This leads us to conclude that high dietary fatty acids either reduce the endogenous antigen for NKT cells or interfere with the interaction between the endogenous antigen and CD1d.  5. The effects of dietary fatty acids on hepatic cytokine expression, infl ammation signaling and serum adipokines. Wild-type C57BL6 mice were fed diets enriched with certain fatty acids (see Table 1) for 12 weeks. A: Hepatic IFN ␥ and IL-4 expressions were determined by quantitative real time PCR and normalized with GAPDH expression. B: The expression of total-I B ␣ in whole liver extract was measured by Western blot. The phospho-I B ␣ expression was then measured by stripping and reprobing the same membrane. Representative autoradiographs are shown. Mean (± SD) results from triplicate experiments are graphed. C: ELISA-based NF-B binding activity assay. D: Serum levels of leptin and adiponectin. (* P < 0.05; ** P < 0.01 versus ND) (n = 4/group). IFN, interferon; IL, interleukin; ND, normal diet. most likely the result of impaired CD1d-dependent endogenous antigen presentation by hepatocytes to NKT cells. Further studies to identify the endogenous antigen that responds to dietary fatty acid alteration will help to elucidate the pathogenesis of diet-induced obesity, insulin resistance, and NAFLD, and may have profound implications in identifying targets for therapy in obesity-related diseases.