Kupffer cell and interleukin-12–dependent loss of natural killer T cells in hepatosteatosis


  • Potential conflict of interest: Nothing to report.


Hepatosteatosis is associated with increased expression of tumor necrosis factor alpha (TNF-α) and interleukin (IL)-12, major T helper (Th) 1 cytokines, and reduced hepatic natural killer T (NKT) cell numbers. The relationship between lipid accumulation, cytokine expression, and hepatic NKT cells is not known. This study was conducted to assess the role of IL-12 in the development of hepatic steatosis and its potential impact on liver NKT cells. Male C57Bl/6 wildtype (WT) and IL-12-deficient (IL-12−/−) mice were fed a choline-deficient diet (CDD) for 0, 10, or 20 weeks. CDD led to marked hepatosteatosis, reduced hepatic but not splenic NKT cell numbers and function, and increased hepatic expression of the Th1-type cytokines IL-12, interferon gamma (IFN-γ), and TNF-α in WT mice. The absence of IL-12 resulted in similar CDD-induced hepatosteatosis, but preserved hepatic NKT cells and significantly reduced hepatic IFN-γ and TNF-α expression. Treatment of CDD-fed mice with lipopolysaccharide led to a significant increase in hepatic IL-12 expression, and Kupffer cell (KC) depletion reduced liver IL-12 expression and restored NKT cells in CDD-induced fatty liver. Interestingly, KCs from CDD-fed mice failed to produce increased quantities of IL-12 upon activation in vitro when compared to similarly treated KCs from control fed mice, suggesting that secondary factors in vivo promote heightened IL-12 production. Finally, human livers with severe steatosis showed a substantial decrease in NKT cells. Conclusion: Hepatosteatosis reduces the numbers of hepatic NKT cells in a KC-and IL-12-dependent manner. Our results suggest a pivotal and multifunctional role of KC-derived IL-12 in the altered immune response in steatotic liver, a process that is likely active within human nonalcoholic fatty liver disease. (HEPATOLOGY 2010;51:130–141.)

Over the last decade the role of the liver as a major organ of the innate immune system with immunomodulatory functions has been increasingly recognized. The liver contains one of the largest populations of resident macrophages (Kupffer cells [KCs]), natural killer (NK) cells, and natural killer T (NKT) cells.1 They all are important mediators in the innate immune system, although NKT cells with phenotypic markers of both NK cells and T cells could represent a link between innate and adaptive immunity.2 NKT cells have attracted a great deal of attention during the last years, given their role in a variety of immunological responses including cancer, microbial infection, and autoimmunity.3–5 Most NKT cells recognize lipid antigens presented by the atypical major histocompatibility complex (MHC) class I-like molecule CD1d, expressed primarily on antigen-presenting cells including monocytes, macrophages, dendritic cells, and B cells.2, 6 Importantly, the functional properties attributed to NKT cells appear to be largely due to CD1d-restricted T cells.7 In addition, NKT cells are either CD4+ or CD4- CD8- in contrast to typical CD8+ Class I restricted T cells. Most notably, NKT cells express an extremely limited T-cell receptor (TCR) repertoire, as their TCR is composed almost exclusively of Vα14/Jα281 paired with Vα8, Vα7, or Vα2,8, 9 which bind lipids, glycolipids, or highly hydrophobic peptides presented by CD1d molecules.6, 10 CD1d-restricted T cells demonstrate potent production of both T-helper (Th)-1 associated cytokines like interleukin (IL)-12 and interferon (IFN) γ, and Th2 associated cytokines like IL-4 and IL-1011 and, therefore, are suitable to affect/control the local environment in an either pro-inflammatory or antiinflammatory manner.

Previously, we reported that hepatosteatosis induced by feeding choline-deficient diet (CDD) for 6 weeks was associated with a substantial reduction in resident NKT cell numbers, concomitant with increased Th1 cytokine production (i.e., tumor necrosis factor (TNF)-α, IL-12, and IFN-γ) but unchanged levels of Th2 cytokines (i.e., IL-4 and IL-10).12 Similar observations have been made in different models of obesity. Leptin-deficient ob/ob mice,13 insulin-resistant fa/fa Zucker rats,14 and diet-induced obesity12, 15 in rodents have shown that hepatosteatosis is associated with changes in local cytokine patterns, resembling a state of chronic hepatic inflammation with changes in hepatic lymphocyte subpopulations.12, 15, 16 Together these results suggest that the presence of fat in the liver alters the hepatic immune system, which likely contributes to their increased susceptibility to secondary insults. We recently showed that the predominance of Th1 cytokines was even more pronounced after T-cell activation in steatotic liver, associated with elevated signal transducer and activator of transcription (STAT) 4 and T-box transcription factor expressed in T cells (T-bet), crucial transcription factors for Th1 commitment.12

IL-12 was initially termed natural killer cell stimulatory factor because of its ability to stimulate NK cells,17 but it also was found to stimulate T-regulatory cells and T cells.18 IL-12 plays an essential role in the protective immune responses against intracellular pathogens by directing the development of Th1 reactions.19, 20 Different studies suggest a significant involvement of IL-12 in the process of hepatosteatosis as it influences the local Th1/Th2 balance and NKT cell activation and regulation, but the exact role of IL-12 in diet-induced pathogenesis of hepatosteatosis has not been explored.21, 22

The aim of the present study was to evaluate the role of IL-12 in hepatosteatosis and its impact on hepatic resident NKT cells, and further to determine whether humans suffering from hepatosteatosis show a similar reduction in hepatic NKT cells. To this end, using a CDD model of hepatosteatosis, we identified the importance of IL-12 production by Kupffer cells in the reduction of hepatic NKT cells and translated this observation of reduced NKT cell numbers to human liver samples with hepatosteatosis.


αGal, alpha galactosylceramide; CDD, choline deficient diet; CSD, choline sufficient diet; ELISA, enzyme linked immunosorbent assay; H&E, hematoxylin and eosin; HRP, horseradish peroxidase; IFN-γ, interferon gamma; IL, interleukin; KC, Kupffer cell; LPS, lipopolysaccharide; MHC, major histocompatability complex; NK, natural killer; NKT, natural killer T; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; RT, reverse transcription; SEM, standard error of the mean; STAT, signal transducer and activator of transcription; T-bet, T box transcription factor expressed in T cells; TCR, T-cell receptor; Th, T helper; TMB, tetramethylbenzidine; TNF-α, tumor necrosis factor alpha; T-TBS, tween-20 Tris-buffered saline; WT, wildtype.

Materials and Methods

Animals and Treatment.

All animals received humane care according to the criteria outlined in the “Guide for the Care and Use of Laboratory Animals.” The conduct of the study was approved by the University Institutional Animal Care and Use Committee. Male wild type (WT) C57BL/6 mice and IL-12 p40-deficient mice were obtained from the Jackson Laboratories (Bar Harbor, ME). Mice received a CDD (Dyets, Bethlehem, PA) for 0, 10, or 20 weeks, which results in hepatocellular lipid accumulation. For lipopolysaccharide (LPS) studies, WT mice fed CDD for 0 or 20 weeks were administered LPS (from E. coli, Sigma, St. Louis, MO) at a concentration of 2.5 mg/kg (or saline vehicle) by intraperitoneal injection 6 hours prior to sacrifice. For NKT cell activation, WT mice fed CDD for 0 or 10 weeks were administered alpha-galactosylceramide (αGal, Alexis Chemicals, Axxora, San Diego, CA) at a concentration of 200 ng/g by intravenous injection 3 hours prior to sacrifice. All animals were housed in pathogen-free barrier facilities accredited by the Association for Assessment and Accreditation of Laboratory Animal Care. All animals had free access to food and water. After 0, 10, or 20 weeks of feeding mice were sacrificed, serum and tissue collected, and samples stored at −80°C until used.

Depletion and Detection of Macrophages.

Multilamellar liposomes containing clodronate (a gift obtained from Roche Diagnostics, Mannheim, Germany) were prepared as described.23 Mice were inoculated intravenously with 100 μL of 0.5 mg/mL liposome encapsulated clodronate or phosphate-buffered saline (PBS) as control suspended in saline. Mice, after 10 weeks on CDD, received clodronate liposomes (or PBS liposomes as control) for the following 3 weeks; mice after 20 weeks on CDD received clodronate liposomes (or PBS liposomes as control) for the following 5 weeks every 4-5 days. During clodronate administration, mice were fed CDD continuously. Macrophages were stained immunohistochemically with F4/80 as previously described.24 Positive cells were counted from 10 high-powered fields from each animal.

KC Isolation and In Vitro Activation.

KCs were enriched from livers of 10-week choline-sufficient diet (CSD) or CDD-fed mice. The liver was perfused with collagenase (0.6 mg/mL Collagenase D, Roche Applied Sciences, Indianapolis, IN) followed by filtration through a 70-μm nylon filter. Macrophages were enriched by density gradient centrifugation on Histogenz (0.288 g/14 mL, Sigma). Cells at the interface were plated at 1 × 105 cells/well on a 48-well tissue culture plate in RPMI 1640 plus 10% fetal calf serum and antibiotics. After 16 hours the media was changed and cells treated with either saline or LPS (10 μg/mL) for 3 hours. Cells were then collected, RNA isolated, and IL-12 gene expression analyzed by real-time polymerase chain reaction (PCR).

Human Liver Samples and NKT Cell Staining.

Human liver tissue with different degrees of hepatic steatosis was obtained from subjects undergoing routine diagnostic liver biopsy under a protocol approved by the UNC Institutional Review Board or the University of Heidelberg Review Board and subjects gave written informed consent. Samples presenting with fibrosis, cirrhosis, or significant inflammatory cell infiltrate indicating hepatitis were excluded. Samples were divided into three groups, those showing lipid accumulation in less than 5% of hepatocytes (n = 5), those presenting with steatosis in 6%-32% of hepatocytes (n = 4), and those showing lipid accumulation in greater than 33% of hepatocytes per field (n = 3). For staining, formalin-fixed, paraffin-embedded sections were stained for NKT cells according to a modified protocol by Harada et al.25 In short, rehydrated samples were blocked with 10% goat serum, incubated with anti-CD57 (1:200, BD) overnight at 4°C, followed by antimouse IgG linked to alkaline phosphatase (1:100, 1 hour, Sigma). Liquid permanent red (Dako) stained subsequently CD57-positive cells red. For T-cell staining, slides were incubated overnight at 4°C with mouse anti-CD3 (Novacastra) followed by antimouse IgG horseradish peroxidase (HRP; 1:200, 1 hour, Amersham). For color development by HRP, the tetramethylbenzidine (TMB) Substrate Kit (Vector Laboratories) was used, which stained CD3-positive cells blue, and for alkaline phosphatase liquid permanent red (Dako) was used, which stained CD57+ cells red. Double positive-stained cells (purple) were identified as NKT cells. Supplemental Figure 2 shows individually stained sections as control for this procedure.

Other Methods.

The following methods are described in the supporting materials. Enzyme-linked immunosorbent assay, tissue pathology, triglyceride assay measurement, mononuclear cell isolation, and RNA isolation.


Data are presented as mean ± SEM. The statistical significance of difference between CDD-fed mice and control groups was determined by comparison of the mean using the independent samples t test. A P value of < 0.05 was selected as the level of significance. Statistical analyses were performed using SPSS 11.0 software (Chicago, IL).


CDD-Induced Hepatosteatosis Reduces Hepatic NKT Cell Number and Function.

Previous studies have shown reduced numbers of NKT cells in the presence of liver lipid accumulation.1 To determine if CDD-induced hepatosteatosis could significantly alter hepatic NKT cell numbers and/or function, WT mice were fed CDD or CSD as control for 10 weeks followed by administration of either αGal or vehicle. As shown in Fig. 1A, feeding WT mice CDD for 10 weeks led to a reduction in hepatic NKT cells as assessed by percentages of TCRβ and NK1.1 double positive cells and by TCRβ and CD1d-tetramer/PBS57, a reduction which was associated with significantly reduced IL4 production when compared to αGal-treated CSD controls (Fig. 1B). The use of TCRβ as a marker of NKT cells was confirmed in lieu of CD3 by co-staining liver mononuclear cells with both CD3 and TCRβ (Supplemental Fig. 1.) Together, these data confirm the reduction not only in number but also function of hepatic NKT cells.

Figure 1.

Hepatic NKT cell number and function are reduced within the steatotic liver. (A) Hepatic mononuclear cells were analyzed by flow cytometry for T cell receptor beta (TCRβ) and NK1.1 expression or TCRβ and binding to CD1d-tetramer loaded with PBS57 in mice fed CSD or CDD for 10 weeks. Representative dotplots presented. (B) Hepatic IL4 protein expression was measured by enzyme-linked immunosorbent assay (ELISA) 3 hours following intravenous αGal (200 ng/g) administration to 10-week CSD-fed or CDD fed mice; n = 6-9 mice per group. *P < 0.05 versus CSD-fed vehicle-treated mice, +P < 0.05 versus CSD-fed, αGal-treated mice.

CDD-Induced Hepatosteatosis Is Associated with a Significant Increase in Hepatic IL-12.

Hepatosteatosis is associated with increased Th1 cytokine expression.12, 15 In our previous study we reported, following 6 weeks of CDD feeding, which induces mild hepatic lipid accumulation, a strong increase in hepatic Th1 (IFN-γ, IL-12) cytokine production both in control treated and concanavalin A treated CDD-fed mice.12 Here we wished to better understand the effects of severe steatosis, a setting where hepatic NKT cell numbers and function are significantly reduced, on hepatic IL-12 expression. Hepatic IL-12 mRNA-levels were significantly elevated following 10 weeks of CDD feeding compared to 0-week feeding (1.09 ± 0.17 fold change in control versus 7.8 ± 1.09 fold change after 10 weeks CDD, P < 0.05) an elevation which was sustained through 20 weeks of feeding (10.6 ± 2.8 fold change, P < 0.05) (Fig. 2A).

Figure 2.

Increased IL-12 production does not influence the progression of hepatosteatosis following long-term CDD feeding. (A) C57BL/6 mice were fed CDD for 0, 10, or 20 weeks. Hepatic IL-12 expression was determined by quantitative real-time PCR of whole liver tissue. (B) Hematoxylin and eosin (H&E)-stained liver histology after 0, 10, and 20 weeks of CDD in C57BL/6 (WT) and IL-12−/− mice. Representative photomicrographs are presented at 100× magnification with 400× inserts (bottom right of each image). (C) Serum alanine aminotransferase, liver weight to body weight ratio (LW/BW), and hepatic triglycerides were measured following either 0, 10, or 20 weeks of CDD feeding. No significant differences were observed between WT and IL-12−/− mice in any of these parameters. Values are means ± SEM, *P < 0.05, n = 4-8 animals per group.

Deficiency in IL-12 Does Not Affect the Progression of CDD-Induced Hepatosteatosis.

To study the potential implications of IL-12 in hepatosteatosis, IL-12-deficient mice were fed in addition to WT mice for 0, 10, and 20 weeks with CDD. As shown in Fig. 2B, microvesicular steatosis developed in both WT and IL-12−/− mice following 10 weeks of CDD feeding, which progressed in both to macrovesicular lipid accumulation following 20 weeks of feeding. This increased lipid accumulation occurred in the absence of significant hepatocellular injury as measured by serum alanine aminotransferase levels (Fig. 2C). The liver weight to body weight ratio increased following 20 weeks of CDD feeding in both C57Bl/6 and IL-12−/− mice when compared to 0 weeks of feeding (Fig. 2C). Finally, measurement of hepatic triglycerides confirmed the histopathological findings demonstrating similar levels of hepatic triglycerides after 10 and 20 weeks of CDD feeding in WT and IL-12−/− mice (Fig. 2C). Together, it is clear that IL-12 does not significantly alter the progression of hepatosteatosis following long-term feeding of the CDD.

IL-12 Promotes the Induction of Th1-Associated Cytokines in CDD-Induced Hepatosteatosis.

In addition to increased levels of IL-12, other key Th1-associated cytokines like TNF-α and IFN-γ are elevated in different models of hepatosteatosis, whereas Th2-associated cytokines are not affected.12, 15 Chronic CDD resulted in a significant increase in TNF-α and IFN-γ mRNA in WT mice following 10 or 20 weeks of feeding, where only mild changes were observed in IL-4 or IL-10 expression. Surprisingly, the absence of IL-12 abrogated the steatosis-induced increases in Th1-associated cytokines, although mice deficient in IL-12 present with a comparable amount of hepatic fat accumulation (Fig. 3).

Figure 3.

IL-12 promotes Th1 cytokine expression within the steatotic liver. Hepatic Th1 (TNF-α and IFN-γ) and Th2 (IL-4 or IL-10) cytokine gene expression was measured by real-time PCR in WT and IL-12−/− mice fed CDD for 0, 10, or 20 weeks. Values are means ± SEM, *P < 0.05; n = 4-8 animals per group.

IL-12 Is Involved in CDD-Induced Hepatic NKT Cell Depletion.

In the current study of CDD-induced hepatosteatosis, we observed a severity-dependent reduction in the hepatic NKT cell population in WT mice. After 10 weeks of diet nearly 70% and after 20 weeks up to 98% of the resident NKT cell population was depleted (Fig. 4A,B), a finding which was confined to the liver, as splenic NKT cell populations were not affected by CDD feeding (Fig. 4C,D). This significant reduction in liver NKT cells was abrogated in IL-12-deficient mice. After 10 weeks and 20 weeks of CDD, no changes in hepatic NKT cell numbers were observed when compared to 0-week CDD mice (Fig. 4). Together, these results suggest that not IL-12 itself, but a combination of IL-12 in conjunction with changes in the fatty hepatic microenvironment are likely responsible for the depletion of hepatic NKT cells.

Figure 4.

IL-12 is involved in steatosis-associated hepatic NKT cell depletion. C57BL/6 (WT) and IL-12−/− mice were fed CDD for 0, 10, or 20 weeks and total liver mononuclear cells or splenocytes were analyzed by flow cytometry. Hepatic mononuclear cells (A) or splenocytes (C) were stained with antimouse T cell receptor β (TCR) and NK 1.1 antibodies. T cells, NKT cells, and NK cells are outlined in boxes. Values represent percent of total liver mononuclear cells analyzed. Quantitation of hepatic (B) or splenic (D) NKT cells expressed as percent of total hepatic TCRβ+ cells expressing NK1.1. Figure is representative for 3-6 different experiments. Results are representative of 3-6 animals per group.

Hepatic Macrophages Contribute to IL-12 Production.

To evaluate the potential source of IL-12 in hepatosteatosis, we investigated the role of hepatic macrophages in fatty liver. Previous studies reported increased numbers of macrophages in fatty tissue,26 but the reports concerning fatty liver are inconsistent. Evaluation of macrophage numbers by immunohistochemistry with F4/80 did not reveal any increase in their numbers following 10 or 20 weeks of CDD feeding when compared to 0-week CDD mice (Fig. 5A,B). However, steatosis did alter the morphology of the KCs present with the prominent enlargement in cell size and an elongated, spindled shape. Consistent with the potential activation of KCs was a significant increase in hepatic CD14 gene expression in WT mice with hepatosteatosis (Fig. 5C). To evaluate the influence of hepatosteatosis on KC function, mice fed a CDD or control diet for 20 weeks were administered LPS (2.5 mg/kg) by intraperitoneal injection. Six hours following LPS injection there was a large and significant increase in hepatic IL-12 gene expression when compared to LPS-treated CSD-fed mice (Fig. 5D). This increase in IL-12 correlated with significantly increased serum aspartate aminotransferase levels in CDD-fed but not CSD-fed mice, injury which occurred independent of IL-12 (Fig. 5E). Interestingly, in vitro analysis of hepatic macrophage function did not reveal a significant increase in IL-12 expression 3 hours following LPS stimulation in KCs from CDD-fed mice when compared to CSD-fed hepatic macrophages (Fig. 5F), suggesting that secondary factors present within the steatotic liver prime macrophages for cytokine release. In sum, these data demonstrate the enhanced responsiveness of hepatic macrophages to endotoxin within the steatotic liver, a process which likely involves multiple factors for their activation.

Figure 5.

Kupffer cells are activated within the steatotic liver. (A) Liver sections from WT mice fed CDD for 0, 10, or 20 weeks were stained with F4/80. Representative 400× photomicrographs presented. (B) Quantitation of F4/80 positive cells per 400× field. Ten fields per section from each mouse were analyzed. (C) Hepatic CD14 gene expression was measured using real-time PCR. (D) Hepatic IL-12 gene expression 6 hours following LPS injection in wild type mice fed either a choline sufficient or choline deficient diet for 30 weeks. (E) Serum aspartate aminotransferase levels 6 hours following LPS (2.5 mg/kg) injection in WT or IL-12−/− mice fed CSD or CDD for 20 weeks. (F) IL-12 expression 3 hours following treatment with either saline or LPS (10 μg/mL) in isolated hepatic macrophage (Mf) from WT mice fed either a CSD or CDD for 10 weeks. Representative data from two independent experiments shown. Values are mean ± SEM, *P < 0.05 versus respective control, +P < 0.05 versus LPS-treated CSD-fed mice; n = 4-8 animals per group.

KCs Contribute to NKT Cell Loss Within the Steatotic Liver.

In order to study the functional role of macrophages in hepatosteatosis and specifically their impact on resident NKT cells numbers, mice with established fatty liver received clodronate by intravenous injection according to the time-table presented in Fig. 6A. Clodronate is engulfed selectively by macrophages and induces apoptosis. The control group with fatty liver received saline (Fig. 6B), whereas macrophage depletion was achieved effectively by clodronate treatment. The absence of macrophages after clodronate treatment was confirmed using immunohistochemistry (Fig. 6B) and did not result in any changes in liver injury at either the 13-or 25-week time-point when compared to PBS encapsulated liposome-treated CDD-fed controls (data not shown), although loss of macrophages did show a trend toward a reduction in hepatic lipid accumulation (32.40 ± 8.17 versus 17.89 ± 4.62 for CDD-fed mice treated with PBS liposomes versus CDD-fed mice treated with clodronate encapsulated liposomes, P = 0.104).

Figure 6.

Clodronate depletes Kupffer cells and ameliorates increased Th1-associated cytokine expression in steatotic liver tissue. (A) Dosing regimen for macrophage depletion studies. Mice with established hepatosteatosis after 10 weeks of CDD were injected every 4-5 days with clodronate for the following 3 weeks, mice after 20 weeks of CDD received clodronate every 4-5 days for the following 5 weeks while continuously fed CDD. (B) Liver sections stained with F4/80 antibody from WT mice administered vehicle or clodronate as described in (A). Representative photomicrographs at 400× magnification presented. (C) Hepatic Th1 and Th2 cytokine expression as assessed by real-time PCR in vehicle or clodronate-treated animals. Values are mean ± SEM, *P < 0.05. n = 4-6 animals per group.

KCs play a major role in innate immunity and are a likely source of IL-12 in the liver.27 Therefore, we measured hepatic IL-12 and IFN-γ expression as well as IL-4 and IL-10 gene expression using real-time PCR after clodronate (or vehicle) treatment in WT mice (Fig. 6C). KC depletion blunted steatosis-induced hepatic IL-12 and IFN-γ gene expression and had no effect on Th2-associated cytokines. In addition, KC depletion led to a complete repopulation of hepatic NKT cells following 10 weeks of CDD (Fig. 7). Even after 20 weeks of CDD with an almost complete loss of NKT cells in hepatosteatosis, 5 weeks of clodronate-treatment was sufficient to restore NKT cells. To eliminate possible nonspecific effects of clodronate treatment, IL-12-deficient mice fed CDD for 10 weeks were administered clodronate for 3 weeks. Hepatic NKT cell populations were not changed from untreated CDD-fed mice (data not shown). These results suggest that KCs, either directly or indirectly, contribute to the production of IL-12 and continually deplete hepatic NKT cells within the steatotic liver.

Figure 7.

Kupffer cells contribute to steatosis-induced hepatic NKT cell depletion. Hepatic mononuclear cells were isolated from vehicle or clodronate-treated WT mice fed CDD for indicated time periods and subpopulations of lymphocytes assessed by flow cytometry. Results are representative of 4-6 different experiments, P < 0.01.

NKT Cell Populations Are Reduced in Human Nonalcoholic Hepatosteatosis.

Recently, reduced peripheral NKT cell numbers in the blood of patients suffering from nonalcoholic fatty liver disease have been reported,28 but no data exist regarding NKT-cell numbers in human steatotic livers. Human liver samples were obtained from the Liver Diseases Unit at the University of North Carolina and from the Department of Pathology at the University of Heidelberg from patients with different degrees of hepatosteatosis. Samples were selected based on the absence of severe inflammation, fibrosis, or cirrhosis to more clearly evaluate the influence of steatosis alone on hepatic NKT cells. As shown in Fig. 8, liver sections from patients with steatosis present in 6%-32% of hepatocytes had reduced numbers of NKT cells. With severe steatosis, hepatic NKT cells are even further reduced when compared to healthy control livers. Together, these data suggest that, as is seen in the rodent model of hepatosteatosis, nonalcoholic hepatosteatosis in humans results in a decrease in the numbers of hepatic NKT cells.

Figure 8.

Hepatosteatosis is associated with decreased numbers of NKT and NK cells in humans. Human liver biopsies were divided into three groups as described in Materials and Methods. Slides were stained for CD3 and CD57, a marker for human NK cells. CD3-positive cells stained blue, NK-positive cells present red. Double-positive cells are termed NKT cells. (A) Representative microphotographs of stained tissue are presented, NKT cells are pointed out by arrows. (B) Quantitation of NKT cells were counted from at least 10 400× pictures of the tissue. Representative results for n = 3-5 per group. Values are means ± SEM.


NKT cells represent a population of highly specialized lymphocytes involved in regulation of the immune system. As NKT are 20%-25% of hepatic mononuclear cells and share features of both classical T cells and NK cells, they represent an important link between innate and adaptive immunity through production of both Th1-and Th2-associated cytokines.2 Findings from the current study confirm the ability of hepatocellular lipid accumulation to decrease the numbers and function of this important resident lymphocyte population both in mice and humans and experimentally define an IL-12-and KC-associated mechanism for their depletion. Together, these studies provide new and important information regarding the impact of hepatosteatosis on resident NKT cells.

Previous studies in fatty livers of obese, leptin-deficient ob/ob mice reported reduced NKT cell numbers within the liver,16 and this reduction was confirmed in a number of different models of diet-induced hepatosteatosis.12, 14, 15 Interestingly, either adoptive transfer of WT NKT cells or activation of NKT cells with the naturally occurring glycolipid glucocerebroside in leptin-deficient ob/ob mice led to changes in hepatic lipid content, specifically a movement from macrovesicular to microvesicular fat, and improvement in glucose tolerance suggesting that NKT cells may be inhibitors of hepatosteatosis.29, 30 However, our findings suggest that in CDD-induced hepatosteatosis, the restoration of NKT cells, and suppression of Th1 cytokine response in IL-12−/− mice is insufficient to affect hepatic triglyceride accumulation. The reasons for the lack of effect on lipid accumulation is not known with certainty. It is clear that other inflammatory cytokines, TNF-α in particular, are capable of promoting nonalcoholic31 and alcoholic steatosis32 and studies from our own laboratory suggest that TNF-α is active and important for lipid accumulation in this model (unpubl. obs.). In the absence of IL-12, TNF-α levels are reduced to at or near control levels. It could be that their presence, even if at a very minimal level, is capable of promoting and sustaining lipid accumulation. Further study is required to better understand the critical factors for steatosis development. Nevertheless, the current study demonstrates the inability of IL-12 alone to promote lipid accumulation. In sum, as the severity-dependent reduction of hepatic NKT cell numbers is restricted to the liver and peripheral NKT cells in spleen are not affected, the local hepatic environment seems to effect the loss of resident NKT cell numbers, a process that involves IL-12 production.

Over the time-course of 20 weeks, CDD results in severe hepatosteatosis with absence of inflammation or necrosis. It is associated with a significant increase in hepatic IL-12 expression. IL-12 is known to stimulate NKT and other immune-competent T cells and NK cells to release IFN-γ,33 but conflicting data exist concerning their viability after IL-12 exposure. Whereas Ito et al.33 found increased numbers of hepatic NKT cells after IL-12 stimulation, Takahashi et al.34 report the ability of IL-12 to reduce NKT cell viability, and Matsui et al.35 also observed a reduction in NKT cell numbers by IL-12 through activation-induced cell death. In the current study, we demonstrate a significant increase in hepatic IL-12 expression and an IL-12-dependent loss of hepatic NKT cells within the steatotic liver. It is likely that multiple factors present within the steatotic liver sensitize NKT cells specifically to IL-12-associated depletion. Alternatively, IL-12 could operate through other cytokines or factors, as IL-12-deficient mice show significant reductions in the expression of TNF-α and IFN-γ and the function of these mediators on NKT cell function has not been elucidated. Inasmuch as IL-12 promotes loss of NKT cells, it may also provide a supportive signal for hepatic NK cells. Originally described as an NK cell-inducing factor,18 IL-12 likely supports NK cell recruitment and/or survival, although the net effect this has on NKT cell function or hepatosteatosis development remains to be determined. Together, however, findings from the current study strongly implicate IL-12 as a key factor in the depletion of this important immunomodulatory cell population within the murine steatotic liver.

In addition to demonstrating the importance of IL-12 in the depletion of NKT cells from the steatotic liver, the current study further investigated the potential source of IL-12 implicating KCs in this production. Upon activation by endotoxin, resident hepatic macrophages produce proinflammatory mediators including TNF-α and IL-12.13, 36 In alcoholic liver disease, these mediators are associated with the promotion of liver injury and hepatocyte lipid accumulation.32 The impact of nonalcoholic steatosis on KC activation is not well understood. In leptin-deficient mice that present with a fatty liver, hepatic macrophage cytokine production, including IL-12 production, is increased.13 Findings from the current study further suggest that steatosis alone may induce significant KC activation. Using hepatic CD14 gene expression within the steatotic liver as an indirect measure of macrophage activation, we provide data showing a correlation between the severity of hepatosteatosis and the level of CD14 expression independent of alterations in hepatic macrophage numbers. In addition, upon stimulation with LPS in vivo, steatotic livers express significantly greater quantities of IL-12 and show significantly enhanced hepatocellular injury, although this injury occurs independently of IL-12 induction. Interestingly, KCs isolated from CDD-fed WT mice did not express increased levels of IL-12 either basally or when activated by LPS in vitro compared to comparably treated CSD-fed WT mouse KCs, suggesting that secondary factors present within the steatotic liver promote IL-12 production in the presence of LPS. Alternatively, KC-associated IL-12 is not the only source of this important, NKT cell-depleting cytokine. Previous studies in ischemia and reperfusion models have demonstrated the ability of hepatocytes to produce IL-12.37 In the in vivo setting, changes in fat metabolism with unusually high levels of metabolites, increased oxidative stress with consequent lipid peroxidation could promote heightened KC responsiveness38, 39 and the release of proinflammatory cytokines like TNF-α or IL-12 by hepatocytes and lymphocytes.15, 40 Likewise, the potential impact of KCs on the development of hepatosteatosis in this model remains to be determined. It is clear that, unlike loss of IL-12, loss of hepatic macrophages leads to a substantial, although not significant, reduction in hepatic lipid accumulation following 10 and 20 weeks of CDD feeding, indicating that other factors produced by KCs could potentially be important for hepatic lipid accumulation. Further study will be required to concretely identify these factors in the setting of severe hepatosteatosis. The data presented here, however, do implicate hepatic macrophages either directly or indirectly in the production of IL-12, as depletion of these cells reduced hepatic IL-12 expression, balanced the Th cytokine profile of the liver, and led to a reconstitution of NKT cells within the livers of mice with established hepatosteatosis. Together, these data provide additional new insights into the complexities and potential sources of IL-12 production within the steatotic liver.

Hepatosteatosis results in a shifted, Th1-predominated cytokine response.12 The reasons for this shift likely involve increased oxidant production, reduced oxidant defense capacity, and/or modulation of intrahepatic immune cell populations or function. Data from the current study implicate IL-12 and/or loss of NKT cells in the promotion of a Th1-shifted response. IL-12 is known to induce T cells to produce large amounts of IFN-γ and activate Th1 transcription factors, including STAT4.20 NKT cells may also balance the local cytokine response within the liver. Li and Diehl1 and others reported an increased susceptibility of leptin deficient ob/ob mice to endotoxin-induced liver damage with an increase in the hepatic production of Th1 cytokines, including IFN-γ, a known mediator of endotoxin-induced liver damage. Loss of IL-12 reduced hepatic IFN-γ expression in the current study, but also preserved the hepatic NKT cell populations, making it difficult to determine which of these are responsible for the disrupted Th balance in the steatotic liver. Nevertheless, findings from the current study demonstrate the importance of IL-12 in the depletion of hepatic NKT cells in the setting of severe steatosis, a process that is associated with restoration of a balanced hepatic Th cytokine profile.

Utilizing biopsies from patients with mild to severe hepatosteatosis, we present intriguing evidence to suggest a similar correlation between the degree of steatosis and the numbers of hepatic NKT cells present in humans where NKT cell numbers decreased when hepatosteatosis was moderate to severe. Due to small sample size, it was not possible to determine if IL-12 was up-regulated in these samples, as observed experimentally, although previous investigations have demonstrated increased pro-inflammatory cytokines within the human steatotic liver.41 Nevertheless, the findings presented here provide the first line of evidence in humans to suggest that hepatosteatosis may indeed be a key modulator of hepatic lymphocyte populations.

In summary, these results demonstrate a connection between hepatosteatosis-induced IL-12 production and the loss of resident NKT cells and implicate NKT cells in the suppression of Th1-type cytokine production within the steatotic liver. Moreover, we provide evidence for the first time that suggests a similar connection between fatty liver and loss of hepatic NKT cells in humans. Several questions remain, including the nature of the stimulus responsible for increasing KC-associated IL-12 expression and the clear mechanism for how IL-12 mediates its effects on hepatic NKT cell populations. Also, it is of great interest what other effects a loss of NKT cells might have on the steatotic liver. It is well appreciated that both NKT cells and NK cells play a significant role in tumor cell surveillance and clearance and that disruption of NKT and/or NK cell function may predispose the liver to secondary pathologies, including carcinogenesis. In conclusion, the current study provides important new mechanistic information regarding NKT cell depletion within the steatotic liver and identifies potential new therapeutic targets to preserve this critical resident immune cell population of the liver.