A high-fat diet and regulatory T cells influence susceptibility to endotoxin-induced liver injury

Authors

  • Xiong Ma,

    1. Department of Gastroenterology, Renji Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai Institute of Digestive Disease, Shanghai, China
    2. Department of Medicine, Johns Hopkins University, Baltimore, MD
    Search for more papers by this author
  • Jing Hua,

    1. Department of Gastroenterology, Renji Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai Institute of Digestive Disease, Shanghai, China
    2. Department of Medicine, Johns Hopkins University, Baltimore, MD
    Search for more papers by this author
  • Abdiaziz R. Mohamood,

    1. Department of Pathology, Johns Hopkins University, Baltimore, MD
    Search for more papers by this author
  • Abdel Rahmin A. Hamad,

    1. Department of Pathology, Johns Hopkins University, Baltimore, MD
    Search for more papers by this author
  • Rajani Ravi,

    1. Department of Oncology, Johns Hopkins University, Baltimore, MD
    Search for more papers by this author
  • Zhiping Li

    Corresponding author
    1. Department of Gastroenterology, Renji Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai Institute of Digestive Disease, Shanghai, China
    2. Department of Medicine, Johns Hopkins University, Baltimore, MD
    • Johns Hopkins University, 912 Ross Building, 720 Rutland Avenue, Baltimore, MD 21205===

    Search for more papers by this author
    • fax: 410-955-9677.


  • Potential conflict of interest: Nothing to report.

Abstract

In nonalcoholic fatty liver disease, the pathogenesis of progression from simple steatosis to steatohepatitis has not been fully clarified. Many factors, including oxidative stress and hepatic immune regulation, contribute to the inflammation in steatosis. Because regulatory T cells (Tregs) are important components of immune regulation, we have now investigated their role in the pathogenesis of nonalcoholic steatohepatitis. Wild-type C57BL/6 mice were fed a high-fat (HF) diet to induce steatosis, and the hepatic lymphocyte population was analyzed by flow cytometry. HF-induced steatosis was associated with the depletion of hepatic Tregs and led to up-regulation of the inflammatory tumor necrosis factor-α signaling pathway. When challenged by exogenous lipopolysaccharide, the HF-fed mice developed liver inflammation. In contrast, the adoptive transfer of Tregs decreased inflammation in HF-fed mice. In comparison with effector T cells, Tregs had a lower expression of Bcl-2 and, therefore, increased susceptibility to oxidative stress-induced apoptosis. The treatment of mice with the antioxidant Mn(III)tetrakis(4-benzoic acid)porphyrin chloride reduced Treg apoptosis, increased the number of hepatic Tregs, and decreased hepatic inflammation in HF-fed mice. Conclusion: Our results indicate that increased oxidative stress in a fatty liver causes the apoptosis of Tregs, reduces the number of hepatic Tregs, and leads to a lowered suppression of inflammatory responses. This scenario is likely one of the pathogenetic mechanisms that facilitate the transformation of simple steatosis into steatohepatitis when a fatty liver is exposed to second or third hits. (HEPATOLOGY 2007.)

Obesity and its associated conditions, such as nonalcoholic fatty liver disease (NAFLD), have emerged as a major health problem in the United States.1, 2 The spectrum of NAFLD extends from simple hepatic steatosis to nonalcoholic steatohepatitis (NASH) to cirrhosis. Liver-related morbidity and mortality are correlated with the histological severity of NAFLD.3 Patients with simple fatty livers (hepatic steatosis) are relatively asymptomatic and rarely, if ever, die from liver disease.3 Some individuals with steatosis develop NASH, a more aggressive form of liver damage that is characterized by the hepatic accumulation of inflammatory cells and associated liver cell death.4 Many individuals with NASH are also asymptomatic. Over time (that is, 5-10 years), however, chronic subclinical NASH incites a fibrotic response in a sizable subgroup of patients, leading to bridging fibrosis in up to 50% and cirrhosis in as many as 10% of individuals with biopsy-proven NASH.5, 6 The rate of liver-specific mortality in patients with NAFLD-induced cirrhosis appears to be similar to that of well-compensated patients who develop cirrhosis from other chronic liver diseases, that is, about 10% per decade.7 Therefore, understanding the mechanisms that lead to the progression from simple steatosis to NASH is clearly important when rational treatment strategies are designed for those who have developed progressive disease.

The 2-hit hypothesis proposed by Day and James8 remains the prevailing theory for the pathogenesis of NASH. Prolonged overnutrition causes the accumulation of free fatty acids and triglycerides within the liver (steatosis, the first hit). The progression from steatosis to NASH is associated with other factors (the second hit), such as oxidative stress, mitochondrial injury, fatty acid lipotoxicity, immune system activation, and inflammatory cytokine production. Among these factors, the immune regulation of cytokine production is thought to be particularly important.9 A key player in hepatic immune regulation that has recently emerged is the lymphocyte subpopulation known as regulatory T cells (Tregs).10

Tregs are a group of heterogeneous T cells that are actively engaged in the negative control of a variety of physiological and pathological immune responses, including transplant tolerance,11 viral hepatitis,12 autoimmune hepatitis,13 and hepatocellular carcinoma.14 Tregs usually coexpress CD4 and CD25 with other surface markers that are not specific for this subpopulation. The most specific marker of Treg is the transcription factor forkhead box protein 3 (Foxp3), which is a key factor in Treg development and function in vivo.15 In the liver, the over-regulation/suppression of Tregs has been shown to contribute to chronic hepatitis B virus/hepatitis C virus infection12, 16 and hepatocellular carcinoma.14 In contrast, inadequate Treg regulation contributes to autoimmune hepatitis,13 primary biliary cirrhosis,17 and acute rejection of transplant grafts.18 However, whether Tregs play a role in the pathogenesis of NASH has not been studied.

In this study, we have evaluated the role of Tregs in the pathogenesis of NASH. For this purpose, we have used an animal model of a diet-induced fatty liver: Mice fed a high-fat (HF) diet develop obesity, steatosis, and insulin resistance similar to those seen in humans. Under normal, pathogen-free conditions, HF-fed mice do not develop spontaneous steatohepatitis. However, theses mice are unusually susceptible to a small dose of exogenous lipopolysaccharide (LPS) and develop liver inflammation when exposed to LPS.19 We hypothesize that regulation by Tregs is decreased in a fatty liver, and this situation exacerbates the inflammation when the fatty liver is exposed to secondary injury, such as that caused by LPS. We also have elucidated the mechanisms leading to a reduction in the number of Tregs in a fatty liver.

Abbreviations

7-AAD, 7-aminoactinomycin D; ALT, alanine aminotransferase; Bcl-2, B-cell lymphoma 2; BrdU, bromodeoxyuridine; CFSE, carboxyfluorescein succinimidyl ester; Foxp3, forkhead box protein 3; H2O2, hydrogen peroxide; HF, high-fat; HMNC, hepatic mononuclear cell; IKK-β, inhibitor kappa beta kinase; LPS, lipopolysaccharide; MnTBAP, Mn(III)tetrakis(4-benzoic acid)porphyrin chloride; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; ND, normal diet; NF-κB, nuclear factor kappa B; NKT, natural killer T cell; ROS, reactive oxygen species; SD, standard deviation; Teff, effector T cell; TNF-α, tumor necrosis factor-α; Treg; regulatory T cell.

Materials and Methods

Animal Experiments.

Adult male wild-type C57BL/6 mice, 6-8 weeks old, were purchased from the Jackson Laboratory (Bar Harbor, ME). The mice were fed commercial diets containing either a high amount of fat (HF diet; 50% of the total kilocalories; F3282, BioServ, Inc., Frenchtown, NJ) or a normal amount of fat [normal diet (ND); 11% of the total kilocalories] for 6-8 weeks. All mice were maintained in a temperature-controlled and light-controlled facility and allowed to consume water and pellet chow ad libitum. Some HF-fed mice also received Mn(III)tetrakis(4-benzoic acid)porphyrin chloride (MnTBAP; Calbiochem, La Jolla, CA). MnTBAP was dissolved in sterile 0.1 N NaOH as a 20 mg/mL stock and then diluted with sterile phosphate-buffered saline to a final concentration of 2 mg/mL immediately before use. MnTBAP was injected intraperitoneally into mice every 3 days for 8 weeks (10 mg/kg). For LPS experiments, the mice were injected intraperitoneally with a single dose of Escherichia coli LPS (100 μg/mouse; Sigma, St. Louis, MO); after 6 hours, they were killed, and the serum and liver tissue were obtained. The control groups in both the MnTBAP and LPS experiments received vehicle alone. All animal experiments fulfilled the National Institutes of Health and Johns Hopkins University criteria for the humane treatment of laboratory animals.

Isolation and Cell-Surface Labeling of Hepatic Mononuclear Cells (HMNCs).

Mouse livers were perfused briefly with a sterile saline solution to remove blood cells and then were carefully removed and homogenized in a Stomacher80 homogenizer (Seward, London, United Kingdom) for 2 minutes. The liver homogenate was passed through a 100-μm wire mesh to remove connective tissue. After slow centrifugation at 50g to remove hepatic parenchyma cells, the nonparenchymal cells were collected by centrifugation at 450g. The mononuclear cell fraction was then isolated from the interface between 80% and 40% of the Percoll step gradient (Amersham Pharmacia Biotech) after centrifugation at 750g for 20 minutes. Anti-mouse fluorescent antibodies against CD3, CD25, or CD4 were obtained from Pharmingen (San Diego, CA). For apoptosis assays, HMNCs were stained with Annexin V and the vital dye 7-aminoactinomycin D (7-AAD). Apoptotic cells were defined as Annexin V+ and 7-AAD. After surface labeling, HMNCs were evaluated by flow cytometry (Becton Dickinson, Palo Alto, CA), and the data were analyzed with Cell Quest software (Becton Dickinson).

Intracellular Cytokine, Bromodeoxyuridine (BrdU), Foxp3, and B-Cell Lymphoma 2 (Bcl-2) Labeling of Liver Mononuclear Cells.

For intracellular staining of cytokines, we used an intracellular cytokine staining kit (Pharmingen). Briefly, HMNCs were incubated with a leukocyte activation cocktail, which included phorbol 1,2-myristate 1,3-acetate (50 ng/mL), ionomycin (500 ng/mL), and GolgiPlug (1 μL/mL). Cells were labeled with a surface antibody, as described previously, and then permeabilized with Cytoperm/Cytofix (Pharmingen) according to the manufacturer's instructions. After permeabilization, the cells were further labeled with antibodies detecting various intracellular cytokines, such as anti-mouse tumor necrosis factor-α (TNF-α; Pharmingen), and then evaluated by fluorescence-activated cell sorting.

For intracellular Foxp3 or Bcl-2 staining, HMNCs and splenic mononuclear cells were first stained with surface antibodies recognizing CD3, CD4, or CD25 and then washed, permeabilized, and stained with antibodies against Foxp3 (eBioscience, San Diego, CA) or Bcl-2 (Pharmingen). For in vivo BrdU labeling, mice were injected intraperitoneally with 1.0 mg of BrdU (Sigma) every 12 hours for 3 days. HMNCs were isolated, and CD4+CD25+ Foxp3+ cells were identified as described previously. BrdU staining was performed with a BrdU flow kit (Pharmingen) according to the manufacturer's protocol.

Cell Purification and Adoptive Transfer.

Mouse CD4+CD25+ Tregs were isolated from the spleen with a MACS Treg isolation kit (Miltenyi Biotec, Auburn, CA). The purity of the cell separation was ∼95%, as assessed by flow cytometry. After isolation, 1.5 × 106 CD4+CD25+ Tregs or CD4+CD25 effector T cells (Teffs) were adoptively transferred to each recipient mouse via tail vein injection. The effects of Treg adoptive transfer on cytokine production, inflammatory signaling, and susceptibility to LPS were evaluated 8-12 hours after the adoptive transfer. For tracking adoptive transfer Tregs, purified CD4+CD25+ Tregs were stained with 5 μm of carboxyfluorescein succinimidyl ester (CFSE), and then 1.5 × 106 CD4+CD25+ Tregs were injected into the recipient. HMNCs were isolated, as described previously, 8-12 hours after the adoptive transfer. The CFSE positive cells were identified by flow cytometry.

Western Blot Analysis of the Inhibitor Kappa Beta Kinase (IKK-β) Activity.

The whole-protein extracts of whole liver tissue were resolved by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis according to standard procedures. The samples were then transferred to 0.45-μm nitrocellulose membranes (Bio-Rad, Hercules, CA) and incubated overnight with an antibody to total IκB-α (inhibitor of kappa B) (Cell Signaling). After the detection of the total IκB-α, the membranes were then stripped and reprobed with an antibody to phosphorylated IκB-α (Cell Signaling). Chemiluminescent signals were quantitated by density measurements with ImageJ software (National Institutes of Health). The ratio of the expression level of the phosphorylated IκB-α to that for the total IκB-α was taken to represent the IKK-β activity.

Nuclear Factor Kappa B (NF-κB) DNA Binding Activity Analysis.

The NF-κB DNA binding activity was assessed in hepatic nuclear extracts with electrophoretic mobility shift analysis, as previously described,20 with 10 μg of the nuclear protein (obtained from each individual mouse) being used in each assay. The P32-labeled, double-stranded oligonucleotide contained the immunoglobulin gene NF-κB binding site (Santa Cruz Biotechnology, Santa Cruz, CA).21 To demonstrate the identities of the proteins in the DNA binding complexes, antibodies recognizing NF-κB p50 or p65 subunits (Santa Cruz Biotechnology) were added to some assays.

In Vitro Lymphocyte Apoptosis Assays.

Splenic lymphocytes were isolated and cultured in an RPMI 1640 serum-free medium supplemented with murine recombinant interleukin-2 (100 IU/mL; Sigma). Apoptosis assays were performed, as described previously, after the cells had been exposed to 5 μM hydrogen peroxide (H2O2) for 6 hours. In separate experiments, the cells were treated with 150 μM MnTBAP for 2 hours prior to exposure to H2O2.

DNA Microarray Analyses.

Tregs (CD4+CD25+) or Teffs (CD4+CD25) were isolated as described previously. The total RNA was isolated from Tregs and Teffs with the RNeasy kit (Qiagen, Valencia, CA). RNA samples were processed and then hybridized to the Affymetrix murine genome GeneChip array MOE430 as previously described.22 Also, as previously described, stringent statistical methods were used to analyze the microarray data and to identify the genes that were up-regulated or down-regulated in Tregs in comparison with Teffs.22

Serum Alanine Aminotransferase (ALT).

Serum ALT values were measured on a multichannel autoanalyzer in the Clinical Chemistry Laboratory of the Department of Comparative Medicine at Johns Hopkins University.

Statistical Analysis.

All values are expressed as the means ± the standard deviation (SD). The group means were compared by a Student t test with Microsoft Excel (Microsoft, Redmond, WA). P values of less than 0.05 were considered statistically significant.

Results

Depletion of Hepatic Foxp3+CD4+CD25+ Tregs in Steatosis.

Our previous study has shown that mice fed HF diets develop steatosis and gain significantly more weight than pair-fed ND control mice.19 In this study, we evaluated the hepatic levels of Tregs in HF diet–fed mice by flow cytometry. Because the transcription factor Foxp3 determines the differentiation that produces the Treg phenotype and distinguishes Tregs from activated T cells,23–25 the expression of Foxp3 in association with CD25 and CD4 (Foxp3+ CD25+CD4+) cells was used to identify Tregs. The majority of CD4+CD25+ T cells in the livers were found to express Foxp3, and this confirmed that these cells were Tregs (Fig. 1A). Because the total HMNCs recovered from each liver were constant between the HF-fed and ND-fed mice, the percentage of Tregs identified from flow cytometry represented the amount of Tregs in the liver.

Figure 1.

Depletion of CD4+ CD25+ Foxp3+ Tregs in HF diet–induced steatosis. Wild-type C57BL/6 mice were fed an ND or HF diet. HMNCs were isolated and analyzed by flow cytometry at different times. The total liver mononuclear cell numbers were similar for the mice fed an HF diet and for the mice fed an ND. Therefore, the percentage of Tregs reflects the number of Tregs in the liver. (A) Representative dot plots of gated CD4+ cells from each animal of 1 experiment. (B) Mean (±SD) percentage of splenic and hepatic CD4+ CD25+ Foxp3+ Tregs in comparison with those from the ND-fed mice at various times during 8 weeks (n = 10/exp; the experiments were repeated 3 times). *P < 0.05 or **P < 0.01 versus ND-fed control mice. (C) Representative dot plots of gated total HMNCs from each animal of 1 experiment.

The hepatic Tregs gradually decreased during the administration of the HF diet, and at the end of 8 weeks, the levels were less than half of those in the ND-fed mice (Fig. 1B). In contrast, the levels of Tregs in the spleens of HF diet–fed mice remained unchanged (Fig. 1B), and this indicated that this HF diet–induced depletion of Tregs was liver-specific. In addition, the HF diet reduced CD4+ T cells in the liver (Fig. 1C). We previously reported that an HF diet also reduced hepatic CD4+ natural killer T (NKT) cells. Therefore, it is unlikely that the HF diet altered the polarization of CD4+ T cells, which decreased the number of hepatic Treg cells associated with the increased number of other CD4+ T cells. However, because the total HMNCs remained unchanged, the reduction of Tregs in the HF-fed mice should alter the polarization of other types of mononuclear cells. We also asked whether this diet-induced hepatic Treg depletion was reversible. Mice were fed the HF diet for 8 weeks, and then half of the HF-fed mice were switched to the ND for an additional 4 weeks. The other half were maintained on the HF diet. The mice that were switched to the ND lost the extra weight that they had gained, and their hepatic steatosis decreased. The level of hepatic Tregs also returned to the normal level in the mice that were returned to the ND, but they were still depleted in the mice that continued to receive the HF diet (data not shown). Thus, this diet-induced hepatic Treg depletion appears to be reversible and is likely associated with the development of steatosis.

Tregs Modulate Inflammation in Steatosis.

Our previous study showed that there is increased expression of proinflammatory cytokines (for example, TNF-α) by hepatic lymphocytes in HF diet–induced steatosis.19 We, therefore, asked whether the depletion of hepatic Tregs contributes to this increased inflammatory condition. We transiently elevated the hepatic Treg content in fatty livers by adoptive transfer. Tregs were isolated from the ND-fed mice and adoptively transferred to the HF diet–fed mice. The control mice received Teffs (CD4+CD25) by adoptive transfer. The hepatic lymphocyte expression of TNF-α was measured by intracellular cytokine staining and flow cytometry. As observed in our previous study, the expression of TNF-α was increased in CD3 T cells from HF diet–induced fatty livers.19 However, the adoptive transfer of Tregs significantly reduced TNF-α expression (Fig. 2A,B), and this indicated a role for Tregs in regulating the production of this inflammatory cytokine by T lymphocytes. In order to rule out the possibility that the adoptive transfer of Tregs might displace or repopulate the liver with non–TNF-producing cells and, therefore, reduce hepatic TNF, we tracked the adoptive-transferred Tregs with CFSE staining to evaluate the percentage of repopulation by adoptive transfer. Adoptive-transferred Tregs represented only a very small portion of the total HMNCs (0.03%-0.04%) 8-12 hours after adoptive transfer (Fig. 2E). Therefore, it is unlikely that adoptive transfer caused significant hepatic repopulation of non–TNF-producing cells. Although there were fewer adoptive-transferred Tregs in HF-fed mice than in ND-fed mice, the difference was too small to be statistically significant (Fig. 2E). In addition, the adoptive transfer of Tregs to ND-fed mice had little impact on their TNF-α expression (data not shown).

Figure 2.

The depletion of Tregs in mice with HF diet–induced steatosis increases TNF inflammatory signaling. Wild-type C57BL/6 mice were fed an ND or HF diet for 8 weeks. Some HF diet–fed mice also received CD4+ CD25+ Tregs (HF+Treg) or CD4+CD25 Teffs (HF+Teff) by adoptive transfer. The experiments were performed 8-12 hours after adoptive transfer. (A) Representative flow cytometry plots show intracellular TNF-α staining of the total HMNCs from each animal of 1 experiment. (B) Mean (±SD) results of triplicate intracellular TNF-α staining experiments (n = 4/exp). (C) Western blot of IκBα performed on whole liver extracts. Because IKK-β phosphorylates IκBα, the ratio of phospho-IκBα to total IκB-α represents the IKK-β activity. The same membrane was first probed with anti–total IκB-α antibody and then was stripped and reprobed with anti–phospho-IκBα antibody. Representative autoradiographs are shown. Mean (±SD) results of triplicate experiments are graphed (n = 4/exp). (D) NF-κB binding activity from whole liver extracts. Representative autoradiographs are shown. Mean (±SD) results of duplicate experiments are graphed (n = 4/exp). (E) Representative flow cytometry plots show adoptive-transferred Tregs with positive CFSE staining (gated at the total HMNC) from each animal of 1 experiment. *P < 0.05 versus ND-fed mice. §P < 0.01 versus ND-fed mice. **P < 0.05 versus HF diet–fed mice.

We then evaluated downstream TNF-α signaling. TNF-α is known to activate IKK-β, which phosphorylates IκB-α and causes the release of NF-κB. This NF-κB is translocated to the nucleus, where it regulates transcription.26 Although phospho-IκB-α is quickly degraded, the ratio of phospho-IκB-α to total IκB-α represents the activity of IKK-β at a given moment. We measured the expression of total IκB-α through the western blotting of whole liver extracts from mice fed an ND or HF diet and from mice that were fed the HF diet but also received Tregs by adoptive transfer. We then stripped the membranes and reprobed to measure the levels of phospho-IκB-α. The DNA-binding activity of NF-κB as a downstream signal within the IKK-β pathway was also evaluated. Mice fed the HF diet showed significantly increased IKK-β activity, as indicated by an increased ratio of phospho-IκB-α to total IκBα, and increased NF-κB binding activity. The adoptive transfer of Tregs reduced both the IKK-β activity and NF-κB binding activity in HF diet–fed mice (Fig. 2C,D). Therefore, Tregs appear to modulate inflammatory signaling in HF diet–induced steatosis.

TNF-α production can also be induced in hepatic lymphocytes by their exposure to bacterial LPS. Small amounts of LPS and other products of the endogenous intestinal flora that escape into the portal venous blood provoke low levels of cytokine production within downstream organs such as the liver.27 Our previous study has shown that HF diet–induced steatosis increases susceptibility to LPS-induced liver injury.19 Because Tregs reduce inflammatory signaling, we examined the possibility that Tregs could reduce LPS-induced hepatotoxicity in HF diet–induced steatosis. We found that, after the adoptive transfer of Tregs, HF diet–fed mice indeed showed less liver injury after exposure to a small amount of LPS, as reflected by histology and serum ALT (Fig. 3). On the basis of these results, we suggest that the reduced levels of Tregs in steatosis contribute to increased inflammatory signaling and susceptibility to LPS-induced injury in HF-fed mice. These results also have led us to identify the mechanism of Treg depletion in HF diet–induced steatosis.

Figure 3.

The depletion of Tregs in mice with HF diet–induced steatosis increases LPS-induced liver injury. Wild-type C57BL/6 mice were fed an ND or HF diet for 8 weeks. Some HF diet–fed mice also received CD4+ CD25+ Tregs (HF+Treg) or CD4+CD25 Teffs (HF+Teff) by adoptive transfer. The mice, 8-12 hours after adoptive transfer, were injected intraperitoneally with a single dose of 100 μg of LPS. The serum and liver tissue were obtained 6 hours after LPS injection. (A) Representative hematoxylin-eosin stain of the liver histology (scale bar = 10 μm). (B) Mean (±SD) serum ALT levels of triplicate experiments (n = 4/exp). *P < 0.05 versus ND-fed mice. **P < 0.05 versus HF diet–fed mice.

Oxidative Stress Contributes to Increased Treg Apoptosis in Steatosis.

The Treg depletion in HF diet–induced steatosis could be due to either decreased proliferation or increased apoptotic death of Tregs. We first evaluated Treg proliferation by using in vivo 5-bromo-2′-deoxyuridine (BrdU) labeling and flow cytometry. BrdU was injected every 12 hours for 3 days into mice that had been fed either the ND or HF diet for 8 weeks, and the presence of intracellular BrdU staining was determined in hepatic Tregs (CD4+ CD25+ Foxp3+) identified as described previously. These experiments showed no difference in the BrdU labeling between the HF and ND groups (Fig. 4A), indicating that the Treg proliferation was not affected in the HF diet–fed mice. We then used Annexin V staining to evaluate the level of hepatic Treg apoptosis in HF diet–fed mice while concurrently incubating the cells with 7-AAD to assess cell necrosis. Apoptotic Tregs were defined as Annexin V+/7-AAD. There was a significant increase in the apoptosis of hepatic Tregs in HF diet–fed mice, whereas Teff (CD4+CD25) apoptosis was not affected (Fig. 4B,C). This result indicates that hepatic Tregs have an increased susceptibility to apoptosis, which contributes to their reduction in HF diet–induced steatosis. Our next step, therefore, was to identify the agent responsible for the apoptosis of the Tregs.

Figure 4.

Increased apoptosis of hepatic CD4+ CD25+ Foxp3+ Tregs in HF diet–induced steatosis. HMNCs were isolated from mice fed either the ND or HF diet. (A) Representative dot plots (gated on CD4+ CD25+ cells) show Treg (CD4+CD25+Foxp3+) proliferation, as assessed by in vivo labeling with BrdU, from each animal of 1 experiment. The experiments were repeated 3 times. (B) Representative dot plots of apoptosis assays involving Annexin V staining and concurrent incubation with 7-AAD from each animal of 1 experiment. The apoptotic cells are Annexin V+/7-AAD. The upper panels were gated on CD4+ CD25+ Tregs; the lower panels were gated on CD4+ CD25 Teffs. (C) Mean (±SD) levels of apoptosis of CD4+ CD25+ or CD4+ CD25 cells (n = 2/exp; the experiments were repeated 5 times).

In obesity and steatosis, increased fatty acid metabolism leads to increased mitochondrial respiratory activity and excessive production of mitochondrial reactive oxygen species (ROS) in the liver,8, 28 which are known to regulate T-cell apoptosis.29 High levels of ROS probably kill cells directly, whereas low levels probably mediate apoptosis indirectly through their effects on signal transduction and gene expression. To determine whether ROS might contribute to the increased apoptosis of Tregs that we had observed, we used H2O2 to induce oxidative stress and thus allow the study of Treg apoptosis in vitro. When we isolated splenic lymphocytes and analyzed apoptosis among the various lymphocyte subpopulations, we found a higher percentage of apoptosis of Tregs than of Teffs at the baseline (Fig. 5). When subjected to a low level of H2O2-induced oxidative stress, Tregs showed significantly increased apoptosis in comparison with Teffs (Fig. 5). This result suggests that Tregs are more susceptible to ROS-induced apoptosis than Teffs are. Splenic lymphocytes, instead of hepatic lymphocytes, were used in the in vitro study because in vivo experiments showed no reduction in splenic Tregs in HF diet–fed mice (Fig. 1), although the results of the in vitro study indicate that splenic Tregs are still susceptible to ROS-induced apoptosis. Thus, it appears that the selective depletion of Tregs in the liver is likely due to local ROS-induced apoptosis in HF diet–induced steatosis. Therefore, we proceeded to identify the mechanisms that cause the increased susceptibility of Tregs to ROS-induced apoptosis.

Figure 5.

CD4+ CD25+ Tregs are more susceptible to apoptosis induced by ROS. Splenic CD4+ CD25+ Tregs or CD4+ CD25 Teffs were isolated with the MACS Treg isolation kit. H2O2 was added to the culture medium to induce oxidative stress. (A) Representative dot plots of apoptosis assays combining Annexin V staining with concurrent incubation with 7-AAD from 1 experiment using 2 animals. The apoptotic cells are Annexin V+/7-AAD. The upper panels show cells in the control medium; the lower panels show cells in the medium with H2O2. (B) Mean (±SD) results of apoptosis assays (n = 2/exp; the experiments were repeated 5 times).

Tregs Express a Low Level of Bcl-2.

Bcl-2 is known to protect cells from ROS-induced apoptosis,30 and ROS can sensitize T cells to apoptosis by decreasing the expression of Bcl-2.31 Bcl-2 itself does not possess antioxidant activity, but it increases the levels and activities of endogenous antioxidants.32, 33 To evaluate the possible involvement of Bcl-2 in the increased apoptosis of hepatic Tregs, we purified Tregs and Teffs for a microarray analysis of gene expression. As expected, the genes encoding Foxp3 and interleukin 2 receptor alpha chain (CD25) were highly expressed in Tregs (Fig. 6A). However, the expression of Bcl-2 was much lower in Tregs than in Teffs, and this difference was statistically significant using a previously described analysis.22 To confirm that the difference in the Bcl-2 gene expression is associated with a functional difference in the protein expression, we also examined Bcl-2 protein levels by intracellular staining. Our results showed that Tregs also had a much lower level of Bcl-2 protein than Teffs (Fig. 6B,C). This difference likely contributes to the increased susceptibility of Tregs to ROS-induced apoptosis in comparison with Teffs. Interestingly, the low expression of Bcl-2 was a result of the heterogeneous expression of Bcl-2 among different Treg populations rather than an overall low expression of Bcl-2 in all Tregs (Fig. 6C).

Figure 6.

CD4+ CD25+ Tregs express a low level of Bcl-2. CD4+ CD25+ Tregs or CD4+ CD25 Teffs were sorted by flow cytometry. (A) Microarray analysis of the gene expression of CD4+ CD25+ Tregs or CD4+ CD25 Teffs. The relative ratios of the expression of individual genes for CD4+ CD25+ cells and CD4+ CD25 cells are plotted. (B) Mean (±SD) results of Bcl-2 intracellular staining (n = 2/exp; the experiments were repeated 3 times). (C) Representative histogram of Bcl-2 intracellular staining that shows the heterogeneous expression of Bcl-2 among CD4+ CD25+ Foxp3+ cells from 1 experiment using 2 animals.

Antioxidant Treatment Prevents the Apoptosis of Tregs, Increases the Number of Hepatic Tregs, and Reduces LPS-Induced Injury in Steatosis.

Next, MnTBAP, a nonpeptidic mimic of manganese superoxide dismutase, was used to determine whether an antioxidant could prevent ROS-induced apoptosis of Tregs in a fatty liver. This antioxidant has been shown to decrease free-radical generation by inhibiting lipid peroxidation in ob/ob mice.34 To evaluate whether MnTBAP could prevent ROS-induced apoptosis of Tregs in vivo, we treated HF diet–fed mice with MnTBAP for 8 weeks. Unlike the previous study of ob/ob mice,34 MnTBAP had little effect on the body weight of HF diet–fed mice over the course of the treatment period. However, it did significantly reduce Treg apoptosis and reverse Treg depletion in the HF diet–fed mice (Fig. 7A-C,E). We further confirmed the antiapoptotic effect of MnTBAP in vitro by adding MnTBAP to the culture medium of Tregs before exposing them to H2O2. MnTBAP also reduced the H2O2-induced apoptosis of Tregs in this in vitro setting (Fig. 7D). Finally, we evaluated whether the reduced apoptosis and increased number of hepatic Tregs that we observed in response to MnTBAP were also associated with a decrease in LPS-induced liver injury in HF diet–induced steatosis. MnTBAP was able to significantly reduce the LPS-induced liver jury, as reflected by much lower serum ALT levels after the LPS treatment (Fig. 7F).

Figure 7.

An antioxidant treatment prevents the apoptosis of Tregs, increases the number of hepatic Tregs, and reduces LPS-induced injury. Wild-type C57BL/6 mice were fed an ND or HF diet for 8 weeks. Some HF diet–fed mice also received MnTBAP, a known antioxidant (HF+MnTBAP). (A) Representative dot plots (gated on CD4+ CD25+ cells) of Annexin V and 7-AAD staining from each animal of 1 experiment and (C) mean (±SD) results of hepatic Treg apoptosis assays (n = 3/exp; the experiments were repeated 5 times). The apoptotic cells are Annexin V+/7-AAD. (B) Representative flow cytometry dot plots (gated on CD4+ cells) from each animal of 1 experiment and (E) mean (±SD) results of hepatic CD4+ CD25+ Foxp3+ Treg measurements (n = 3/exp; the experiments were repeated 5 times). (D) In vitro apoptosis assay. Splenic CD4+ CD25+ Tregs were isolated as described in the Materials and Methods section. H2O2 and MnTBAP were added to the culture medium. After being washed, the cells were analyzed by flow cytometry. The mean (±SD) CD4+ CD25+ cell apoptosis is graphed (n = 2/exp; the experiments were repeated 5 times). (F) Mean (±SD) of the serum ALT levels 6 hours after LPS stimulation (n = 4/exp; the experiments were duplicated). *P < 0.05 versus ND-fed mice. §P < 0.05 versus HF diet–fed mice.

Discussion

The mechanisms that regulate the transition from simple steatosis to steatohepatitis in NAFLD are not yet fully understood, although the 2-hit theory is currently the prevailing theory to explain the pathogenesis of NASH.8 The accumulation of free fatty acids and triglycerides within the liver and the formation of steatosis constitute the first hit. Subsequent lipid peroxidation, oxidative stress, and other factors lead to transformation from simple steatosis to steatohepatitis (the second hit).

In this study, we have demonstrated that oxidative stress also causes regulatory T-cell apoptosis and depletion from the steatotic liver. Because Tregs play an important role in maintaining the immune regulation of hepatic inflammatory activity,35 their depletion leads to increased inflammation and activation of the TNF-α signaling pathway, resulting in further liver injury when the liver is exposed to LPS, which can be endogenously produced by intestinal flora and delivered to the liver.27 These downstream effects of ROS-induced Treg modulation further contribute to the inflammatory process in steatosis. It is important to emphasize that an HF diet induces only simple steatosis and does not progress to steatohepatitis spontaneously, despite activated inflammatory signaling pathways, such as TNF-α. However, an HF diet makes the liver more susceptible to second or third hits, such as endotoxin and immunoregulatory dysfunction, as reflected by the reduction of Tregs. To the best of our knowledge, this is the first time that Tregs have been linked to the inflammatory process in diet-induced steatosis. In addition, this study reveals that a local factor (hepatic ROS) can selectively affect a subpopulation of T lymphocytes (Tregs) because of their endogenous susceptibility (low Bcl-2 expression). This local modulation of immunoregulation may have implications in other disease mechanisms (that is, autoimmune hepatitis).

T-cell homeostasis is achieved through the balance of the production and proliferation of T cells with their apoptotic cell death. ROS regulate cell death in a variety of cell types, including T cells. It has been known for some time that Bcl-2 expression protects cells from apoptosis mediated by ROS.30 We show here that freshly isolated Tregs are highly sensitive to ROS-mediated apoptosis, whereas other T-cell populations are relatively resistant to ROS-induced apoptosis. Furthermore, we have shown that the relatively low expression of Bcl-2 in Tregs, in comparison with other T-cell populations, may contribute to the susceptibility of Tregs toward ROS-mediated apoptosis. Thus, we have identified a mechanism in which increased oxidative stress and ROS selectively induce Treg apoptosis and reduce Treg numbers in fatty livers. These changes, in turn, lead to increased inflammation and susceptibility to further injury. This chain of multiple insults (a third hit) directly contributes to the pathogenesis of NASH. A particularly interesting finding is our recognition of heterogeneity in the expression of Bcl-2 among Treg subpopulations (Fig. 5C). Currently, we are evaluating the characteristics of these subpopulations. The unique apoptosis-related phenotype of Tregs merits further exploration, particularly as the potential basis for novel therapeutic modulations of Tregs in NASH.

Although Tregs and NKT cells are 2 distinct populations of T lymphocytes that can independently regulate adaptive and innate immune responses, a recent report has provided evidence of crosstalk between Tregs and NKT cells.36 Activated NKT cells promote Treg expansion through interleukin-2–dependent mechanisms in both mice37 and human.38 Our previous study showed decreased hepatic NKT cells in HF diet–induced steatosis.19 Whether the depletion of NKT cells also contributes to the reduction of Tregs in HF diet–induced steatosis is still under investigation.

In conclusion, our study demonstrates that CD4+CD25+ Foxp3+ Tregs may play a critical role in controlling hepatic inflammation. Because one of the important pathological factors of steatosis, oxidative stress, can induce the apoptosis of Tregs, we speculate that the depletion of Tregs may be the key event in the process of transition from simple steatosis to steatohepatitis. The findings presented here expand our current understanding of the pathogenesis of NASH and provide a potential interventional strategy for managing NASH by modulating the survival or apoptosis of Tregs. The results also suggest that strategies aimed at increasing the number and/or function of Tregs should be explored to improve the prognosis of NAFLD.

Acknowledgements

We thank the Hopkins Digestive Disease Basic Research Development Center (R24 DK064388-04) for providing technical support and Dr. James Potter and Dr. Deborah McClellan for their editorial assistance.

Ancillary

Advertisement