Disruption of Adenosine 2A Receptor Exacerbates NAFLD through Increasing Inflammatory Responses and SREBP1c Activity

Abstract

Adenosine 2A receptor (A_2AR) exerts protective roles in endotoxin- and/or ischemia-induced tissue damages. However, the role for A_2AR in non-alcoholic fatty liver disease (NAFLD) remains largely unknown. We sought to examine the effects of global and/or myeloid cell-specific A_2AR disruption on the aspects of obesity-associated NAFLD and to elucidate the underlying mechanisms. Global and/or myeloid cell-specific A_2AR-disrupted mice, as well as control mice were fed a high-fat diet (HFD) to induce NAFLD. Also, bone marrow-derived macrophages and primary mouse hepatocytes were examined for inflammatory and metabolic responses. Upon feeding an HFD, both global A_2AR-disrupted mice and myeloid cell-specific A_2AR-defcient mice revealed increased severity of HFD-induced hepatic steatosis and inflammation compared with their respective control mice. In in vitro experiments, A_2AR-deficient macrophages exhibited increased proinflammatory responses, and enhanced fat deposition of wild-type primary hepatocytes in macrophage-hepatocyte co-cultures. In primary hepatocytes, A_2AR deficiency increased the proinflammatory responses and enhanced the effect of palmitate on stimulating fat deposition. Moreover, A_2AR deficiency significantly increased sterol regulatory element-binding protein 1c (SREBP1c) abundance in livers of fasted mice and in hepatocytes upon nutrient deprivation. In the absence of A_2AR, SREBP1c transcription activity was significantly increased in mouse hepatocytes. Taken together, these results demonstrate that disruption of A_2AR in both macrophage and hepatocytes accounts for increased severity of NAFLD, likely through increasing inflammation and through elevating lipogenic events due to stimulation of SREBP1c expression and transcription activity.

INTRODUCTION

Hepatic steatosis is a hallmark of non-alcoholic fatty liver disease (NAFLD) (1, 2). When the liver develops overt inflammatory damage, simple steatosis progresses to non-alcoholic steatohepatitis (NASH) (2, 3). The latter is the advanced form of NAFLD and is considered as a leading causal factor of cirrhosis and hepatocellular carcinoma (4, 5). In addition, NAFLD critically contributes to the development of dyslipidemia and significantly increases the incidence of atherogenic cardiovascular diseases (6), thereby serving as a key component of metabolic syndrome (7).

Numerous studies from both human subjects and rodent models demonstrate that obesity significantly increases the incidence of NAFLD (8–10). Accordingly, obesity-associated inflammation is accepted as a critical factor that initiates or exacerbates NAFLD. For instance, inflammation can cause hepatic insulin resistance, which, in turn elevates hepatic steatosis, at least in part, through increasing the expression of genes for lipogenic enzymes such as acetyl-CoA carboxylase 1 (ACC1) and fatty acid synthase (FAS), and decreasing the expression of genes for fatty acid oxidation including carnitine palmitoyltransferase 1a (CPT1a) (11–13). In addition, proinflammatory mediators, e.g., cytokines, can exert direct effects on hepatocytes to increase lipogenic events (14), and act on both hepatocytes and liver macrophages/Kupffer cells to accelerate liver inflammation. This exemplifies how inflammation serves as “a second hit” to drive the progression of simple steatosis to NASH. However, exactly how inflammation is regulated in the context of NAFLD pathophysiology remains largely unclear.

Adenosine receptors, including A₁, A_2A, A_2B and A₃, belong to the superfamily of G-protein-coupled receptors, and mediate various physiological functions of adenosine (15). Among the four adenosine receptors, A_2AR displays powerful anti-inflammatory effects in immune cells such as macrophages and neutrophils (15, 16). Several studies in animal models have indicated that A_2AR deficiency exacerbates concanavalin A- or endotoxin-induced liver damage, which is attributable to prolonged and enhanced expression of proinflammatory cytokines including tumor necrosis factor alpha (TNFα) and interleukin-6 (IL-6) (17). Also, there is a study showing that A_2AR deficiency exacerbates the severity of aspects of alcoholic fatty liver disease (ALD) in mice (18). Additionally, A_2AR activation displays beneficial effects on aspects of NASH in rodents (19, 20), which is largely mediated through the anti-inflammatory effect of A_2AR. In contrast, treatment of mice with an A_2AR antagonist increased the severity of CCl₄-induced liver fibrosis; although the same treatment reversed the effect of ethanol on exacerbating CCl₄-induced liver fibrosis (21). Collectively, these findings suggest that A_2AR has a protective role in liver damage. However, whether and how A_2AR coordinates hepatocyte and macrophage metabolic and/or inflammatory responses to alter NAFLD development and progression remains poorly understood. It is also not clear whether and how A_2AR regulates hepatocyte lipogenic events, whose increase is of particular importance in the pathogenesis of hepatic steatosis. The present study aimed to examine how global and/or myeloid cell-specific A_2AR disruption influences NAFLD aspects in mice and provides the primary evidence to support a protective role for the A_2AR in both macrophages and hepatocytes in the pathogenesis of obesity-associated NAFLD. In addition, A_2AR has a previously unidentified role in repressing sterol regulatory element-binding protein 1c (SREBP1c), which contributes to the anti-steatotic effect of A_2AR.

MATERIALS AND METHODS

Animal experiments

Wild-type (WT) C57BL/6J were obtained from Jackson Laboratory (Bar Harbor, ME). A_2AR^−/−, A_2AR^+/−, and A_2AR^+/+ mice were generated as described (22). Myeloid cell-specific A_2AR-disrupted mice were generated using the Cre-LoxP strategy. Briefly, homozygous myeloid cell-specific A_2AR-disrupted (LysMCre⁺-A_2AR^F/F) mice, heterozygous A_2AR-disrupted (LysMCre⁺-A_2AR^F/+) mice, and LysM control (LysMCre⁺-A_2AR^+/+) mice were used. All mice were maintained on a 12:12-h light-dark cycle (lights on at 06:00) and subjected to studies involving chow-diet, high-fat diet (HFD, 60% fat calories) and low-fat diet (LFD, 10% fat calories) as detailed in Supplementary Information (SI). All study protocols were reviewed and approved by the Institutional Animal Care and Use Committee of Texas A&M University.

Cell culture and treatment

Bone marrow cells were prepared from A_2AR-disrupted mice and WT mice, as well as LysMCre⁺-A_2AR^F/F mice and LysMCre⁺-A_2AR^+/+ mice, and differentiated into macrophages (BMDM) and examined for the proinflammatory activation as described (23). Primary hepatocytes were isolated from free-fed mice (24, 25) and examined for metabolic and inflammatory responses. For macrophage-hepatocyte co-culture study, bone marrow cells were prepared from A_2AR-disrupted mice and WT mice at 6 days prior to hepatocyte isolation. After differentiation, BMDM were trypsinized and added to WT primary mouse hepatocytes at a ratio of 1:10 based on the published method (12). Some hepatocytes were incubated in the absence of macrophages and served as the control. After incubation for 48 hr, the co-cultures and control hepatocytes were subjected to the selected assays detailed in SI. Additional experiments involving hepatocytes were also detailed in SI.

Histological, biochemical and molecular assays

Liver sections were subjected to histological and immunohistochemical assays. Plasma parameters were measured using metabolic assay kits and ELISA kits. Also, tissue and/or cell samples were subjected to selected assays including Western blots analysis, real-time PCR, the reporter assays, and chromatin immunoprecipitation (ChIP) assay. Details are provided in SI.

Statistical Methods

Numeric data are presented as means ± SE (standard error). Statistical significance was assessed by unpaired, two-tailed ANOVA or Student’s t tests. Differences were considered significant at the two-tailed P < 0.05.

RESULTS

HFD feeding increases liver A_2AR abundance and the proinflammatory status

Tissue distribution results revealed that the liver is one of the organs in which A_2AR abundance was at the highest levels (Figure 1A). Next, we examined the pathophysiological relevance of hepatic expression of A_2AR in C57BL/6J mice, which display obesity-associated NAFLD upon feeding HFD (9, 10). HFD-fed C57BL/6J mice displayed a significant increase in liver A_2AR mRNAs compared to LFD-fed mice (Figure 1B). HFD-fed mice also displayed significant increases in liver protein levels of A_2AR, as well as CD39 and CD73 (ectoenzymes responsible for extracellular adenosine production) (Figure 1C). By immunohistochemistry in liver sections, HFD-fed mice revealed increased A_2AR expression in liver cells including hepatocytes compared with LFD-fed mice (Supplemental Figure S1). Additionally, HFD-fed C57BL/6J mice displayed significant increases in liver phosphorylation states of Jun N-terminal kinase (JNK) p46 and nuclear factor kappa B (NFκB) p65 as well as mRNA and protein levels of TNFα, interleukin 1 beta (IL-1β), and/or IL-6 compared with LFD-fed mice (Figure 1D,E). These results suggest an association between A_2AR amount and diet-induced liver inflammation.

Figure 1. HFD feeding increases hepatic A_2AR abundance and enhances liver proinflammatory responses.

Wild type C57BL/6J mice were fed as described in Methods. (A) Tissue distribution of A_2AR. SI, small intestine; WAT, white adipose tissue. (B) Liver A_2AR mRNA levels. AU, arbitrary unit. (C) Liver amount of A_2AR, CD39 (ecto-nucleoside triphosphate diphosphohydrolase 1), and CD73 (ecto-5′-nucleotidase). (D) Liver proinflammatory signaling. (E) Liver cytokine mRNA (left panel) and protein (right two panels) levels. For A, C, and D, liver lysates were used for Western blot analysis. Blots were quantified using densitometry (bar graphs in C and D). For B and E, liver mRNAs were quantified using real-time RT-PCR. For E, liver protein levels of cytokines were quantified using ELISA kits. For all bar graphs, data are means ± SE. n = 6 – 8. Statistical difference between HFD and LFD: *, P < 0.05 and **, P < 0.01 in B and E (right two panels) or in bar graphs of C, D, and E (left two panels) for the same protein or gene.

A_2AR disruption exacerbates HFD-induced hepatic steatosis and inflammation

The role of A_2AR in obesity-associated NAFLD remains largely unknown, and was examined using both male and female A_2AR-disrupted (A_2AR^−/− and/or A_2AR^+/−) mice, and their WT (A_2AR^+/+) littermates (Supplemental Figure S2A) upon HFD feeding for 12 weeks. After the feeding period, hepatic A_2AR expression was examined and verified for A_2AR disruption (Supplemental Figure S2B). Regardless of the gender, A_2AR^−/− mice displayed much more severe phenotype of insulin resistance and glucose intolerance compared with A_2AR^+/− mice and/or A_2AR^+/+ mice; although A_2AR disruption exacerbated HFD-induced obesity in male, but not female mice (Supplemental Figures S3–S6). Next, we examined the effect of A_2AR disruption on NAFLD pathology. Among male mice, HFD-fed A_2AR-disrupted mice revealed a greater increase in liver weight compared with control mice and this increase was gene-dose-dependent (Figure 2A). Consistently, the severity of HFD-induced hepatic steatosis in A_2AR^−/− mice or A_2AR^+/− mice was greater than that in A_2AR^+/+ mice. Also, HFD-induced hepatic steatosis in A_2AR^−/− mice was much pronounced compared to A_2AR^+/− mice as demonstrated by H&E or Oil Red O staining in liver sections (Figure 2B). Additionally, both plasma and hepatic levels of triglycerides in HFD-fed A_2AR^−/− mice were greater than their respective levels in HFD-fed A_2AR^+/− mice and/or A_2AR^+/+ mice (Supplemental Figure S7A,B). Among female mice, A_2AR disruption caused a slight increase in the severity of HFD-induced hepatic steatosis without significantly altering liver weight (Supplemental Figure S7C).

Figure 2. A_2AR disruption exacerbates HFD-induced hepatic steatosis and inflammation.

Male A_2AR^−/− mice, A_2AR^+/− mice, and A_2AR^+/+ mice, at 5 – 6 weeks of age, were fed an HFD for 12 weeks. Age- and gender-matched A_2AR^+/+ mice were fed an LFD for 12 weeks and served as normal control. (A) Liver weight. (B) Representative images of H&E, Oil-Red-O, and/or F4/80 staining for liver sections. (C) Liver proinflammatory signaling. (D) Liver mRNA levels. (E) Liver SREBP1c abundance. pSREBP1c, precursor SREBP1c; nSREBP1c, nuclear SREBP1c. (F) Liver insulin signaling. Prior to tissue harvest, HFD-fed mice were injected with or without insulin (1 U/kg) into the portal vein for 5 min. IR, insulin receptor. For C, E, and F, liver lysates (C and F), as well as liver cytosolic and nuclear fractions (E) were subjected to Western blot analysis. Bar graphs, quantification of blots. For A – F, numeric data are means ± SE. n = 10 – 12 (A) or n = 8 – 10 (C – F). Statistical difference between A_2AR^−/−-HFD and A_2AR^+/+-HFD: *, P < 0.05 and **, P < 0.01 in bar graphs of A, C, and E, in D within the same gene, or in bar graphs of F under insulin-stimulated condition; statistical difference between A_2AR^−/−-HFD and A_2AR^+/−-HFD: ^†, P < 0.05 and ^††, P < 0.01 in bar graphs of A, C, and E, in D within the same gene, or in bar graphs of F under insulin-stimulated condition; statistical difference between A_2AR^+/−-HFD and A_2AR^+/+-HFD: ^‡, P < 0.05 and ^‡‡, P < 0.01 in bar graphs of A and C, in D within the same gene, or in bar graphs of F under insulin-stimulated condition.

Because HFD-induced phenotype in male mice was more pronounced, we examined other liver phenotypes, i.e., inflammation and insulin sensitivity, mainly in male mice. Upon HFD feeding, A_2AR-disrupted and control mice displayed comparable numbers of macrophages/Kupffer cells in the liver (Figure 2B). However, either HFD-fed A_2AR^−/− mice or HFD-fed A_2AR^+/− mice revealed a significant increase in liver proinflammatory signaling through JNK p64 and/or NFκB p65 (Figure 2C). The same trends were also observed in TNFα and IL-1β mRNAs (Figure 2D). Since increased lipogenesis is key to the development of hepatic steatosis, we examined the mRNA levels of ACC1 and FAS, which are key enzymes that stimulate liver lipogenesis by converting acetyl-CoA to malonyl-CoA and synthesizing long-chain fatty acids (26, 27). Compared with those in livers of HFD-fed A_2AR^+/+ mice, the mRNA levels of ACC1 and FAS in either HFD-fed A_2AR^−/− mice or HFD-fed A_2AR^+/− mice were significantly elevated, and these elevations were A_2AR gene-dependent (Figure 2D). Additionally, hepatic mRNA levels of SREBP1c in HFD-fed A_2AR-disrupted mice were significantly elevated in an A_2AR gene-dependent manner (Figure 2D). We also measured hepatic mRNA levels of CPT1a, a master regulator that transfers acyl CoA into mitochondria for oxidation (28, 29). Compared with the control, CPT1a mRNAs in HFD-fed A_2AR^−/− mice were slightly higher than in HFD-fed A_2AR^+/+ mice or A_2AR^+/− mice (Figure 2D), suggesting a compensatory response to increased hepatic steatosis. Since SREBP1c stimulates lipogenic gene expression, we analyzed hepatic SREBP1c expression. Both cytosolic and nuclear SREBP1c levels were significantly increased in HFD-fed A_2AR^−/− mice compared to the values of HFD-fed A_2AR^+/+ mice (Figure 2E). When insulin signaling was analyzed, insulin-stimulated insulin receptor phosphorylation and Akt phosphorylation (S473) were significantly decreased in HFD-fed A_2AR-disrupted mice compared to A_2AR^+/+ mice; the decreases were A_2AR gene-dose-dependent (Figure 2F). These results suggest that A_2AR disruption exacerbates diet-induced hepatic steatosis and inflammation in male mice. Moreover, A_2AR disruption enhances liver lipogenic events.

Myeloid cell-specific A_2AR disruption exacerbates aspects of NAFLD

Macrophages critically regulate liver inflammation and hepatic steatosis. We sought to address a role for the A_2AR in macrophages (myeloid cells) in the pathophysiology of NAFLD. We fed homozygous myeloid cell-specific A_2AR disrupted (LysMCre⁺-A_2AR^F/F) mice, heterozygous myeloid cell-specific A_2AR disrupted (LysMCre⁺-A_2AR^F/+) mice, and their WT (LysMCre⁺-A_2AR^+/+) littermates (Supplemental Figure S8) an HFD and examined systemic insulin sensitivity and NAFLD aspects. Compared with HFD-fed LysMCre⁺-A_2AR^+/+ mice, HFD-fed LysMCre⁺-A_2AR^F/F mice and HFD-fed LysMCre⁺-A_2AR^F/+ mice revealed similar increases in the severity of systemic insulin resistance although all mice displayed similar body weight and consumed comparable amount of foods (Supplemental Figures S9 and S10A,B). However, liver weight as well as the severity of HFD-induced hepatic steatosis in LysMCre⁺-A_2AR^F/F or LysMCre⁺-A_2AR^F/+ mice was greater compared to HFD-fed LysMCre⁺-A_2AR^+/+ mice (Figure 3A,B). Additionally, plasma and hepatic levels of triglycerides in HFD-fed LysMCre⁺-A_2AR^F/F mice were significantly greater than their respective levels in HFD-fed LysMCre⁺-A_2AR^+/+ mice or LysMCre⁺-A_2AR^F/+ mice (Supplemental Figure S10C,D). The severity of HFD-induced hepatic steatosis in LysMCre⁺-A_2AR^F/F mice was lower compared to that observed in global A_2AR-deficient (A_2AR^−/−) mice. When liver proinflammatory signaling was examined, the phosphorylation states of JNK p46 and/or NFκB p65 and the mRNA levels of cytokines in HFD-fed LysMCre⁺-A_2AR^F/F mice were greater than in HFD-fed HFD-fed LysMCre⁺-A_2AR^+/+ mice or LysMCre⁺-A_2AR^F/+ mice (Figure 3C, D). Consistent with increased hepatic steatosis, liver mRNA expression of ACC1 and FAS in HFD-fed LysMCre⁺-A_2AR^F/F mice was also greater than their respective levels observed in HFD-fed LysMCre⁺-A_2AR^+/+ or LysMCre⁺-A_2AR^F/+ mice (Figure 3D). Taken together, these results suggest that A_2AR disruption in macrophages (myeloid cells) is sufficient to exacerbate diet-induced NAFLD.

Figure 3. Myeloid cell-specific A_2AR disruption exacerbates HFD-induced hepatic steatosis and proinflammatory responses.

Male LysMCre⁺-A_2AR^F/F mice, LysMCre⁺-A_2AR^F/+ mice, and LysMCre⁺-A_2AR^+/+ mice, at 5 – 6 weeks of age, were fed an HFD for 12 weeks. (A) Liver weight. (B) Representative images of H&E, Oil-Red-O, and/or F4/80 staining for liver sections. (C) Liver lysates were examined for the proinflammatory signaling using Western blot analysis. Bar graphs, quantification of blots. AU, arbitrary unit. TLR4, Toll-like receptor 4. (D) Liver mRNA levels were quantified using real-time RT-PCR. For A, C, and D, numeric data are means ± SE. n = 8 – 10. Statistical difference between LysMCre⁺-A_2AR^F/F and LysMCre⁺-A_2AR^+/+: *, P < 0.05 and **, P < 0.01 in A or in bar graphs of C and D within the same protein or gene; statistical difference between LysMCre⁺-A_2AR^F/F and LysMCre⁺-A_2AR^{F/+: †}, P < 0.05 and ^††, P < 0.01 in bar graphs of C and D within the same protein or gene; statistical difference between LysMCre⁺-A_2AR^F/+ and LysMCre⁺-A_2AR^{+/+: ‡}, P < 0.05 and ^‡‡, P < 0.01 in A or in bar graphs of C and D within the same protein or gene.

A_2AR-disrupted macrophages display increased proinflammatory responses and exacerbate hepatocyte fat deposition and cytokine expression

Since LysMCre⁺-A_2AR^F/F mice revealed increased severity of diet-induced NAFLD, we sought to verify a direct role played by the A_2AR in macrophages in regulating hepatocyte responses in vitro. Initially, we isolated bone marrow cells from A_2AR^−/− and/or A_2AR^+/+ mice, differentiated the cells into macrophage (BMDM), and analyzed BMDM proinflammatory responses. Under LPS-stimulated conditions, the phosphorylation states of JNK p46 and NFκB p65 as well as the secretion of TNFα and IL-6 were significantly increased in A_2AR^−/− BMDM compared to the values observed in A_2AR^+/+ BMDM (Figure 4A,B). Similarly, LPS-induced phosphorylation states of JNK p46 and NFκB p65 in BMDM from LysMCre⁺-A_2AR^F/F mice were significantly increased compared with those in BMDM from LysMCre⁺-A_2AR^+/+ mice (Figure 4C).

Figure 4. A_2AR disruption aggravates macrophage proinflammatory activation.

Bone marrow-derived macrophages (BMDM) were prepared as described in Methods. (A, B) Proinflammatory signaling (A) and cytokine production (B) of BMDM from A_2AR^−/− mice and A_2AR^+/+ mice. For A, BMDM were treated with or without lipopolysaccharide (LPS, 100 ng/ml) for 30 min prior to harvest. Cell lysates were subjected to Western blot analysis. Bar graphs, quantification of blots. For B, cytokine concentrations in BMDM conditioned-media. (C) Proinflammatory signaling of BMDM from LysMCre⁺-A_2AR^F/F mice and LysMCre⁺-A_2AR^+/+ mice. Cells were treated and analyzed as described in A. Bar graphs, quantification of blots. For A – C, numeric data are means ± SE. n = 4 – 6. Statistical difference between A_2AR^−/− and A_2AR^+/+: *, P < 0.05 and **, P < 0.01 in bar graphs of A (under LPS-stimulated condition) or in B; statistical difference between LysMCre⁺-A_2AR^F/F and LysMCre⁺-A_2AR^{+/+: †}, P < 0.05 and ^††, P < 0.01 in bar graphs of C (under LPS-stimulated condition).

Next, we co-cultured WT primary mouse hepatocytes with A_2AR^−/− or A_2AR^+/+ BMDM. Under palmitate-stimulated conditions, hepatocytes co-cultured with A_2AR^−/− BMDM accumulated more fat and revealed significantly higher levels of triglycerides than hepatocytes co-cultured in the absence of BMDM or in the presence of A_2AR^+/+ BMDM (Figure 5A). When the expression of genes/enzymes related to fat metabolism was analyzed, ACC1, FAS and SREBP1c mRNAs in hepatocytes co-cultured with A_2AR^−/− BMDM were significantly higher than their respective levels in hepatocytes co-cultured in the absence of BMDM or in the presence of A_2AR^+/+ BMDM (Figure 5B). In contrast, CPT1a mRNAs in hepatocytes co-cultured with A_2AR^−/− BMDM were significantly lower than those in hepatocytes co-cultured in the absence of BMDM or in the presence of A_2AR^+/+ BMDM. When the expression of cytokines was analyzed, TNFα and IL-1β mRNAs in hepatocytes co-cultured with A_2AR^−/− BMDM were significant higher than their respective levels in hepatocytes co-cultured in the absence of BMDM or in the presence of A_2AR^+/+ BMDM under basal conditions (Figure 5C, top panel). Strikingly, under LPS-stimulated conditions, TNFα, IL-1β, and IL-6 mRNAs in hepatocytes co-cultured with A_2AR^+/+ or A_2AR^−/− BMDM were markedly higher than their respective levels in hepatocytes cultured in the absence of BMDM (Figure 5C, bottom panel). Among co-cultures, TNFα, IL-1β, and IL-6 mRNAs in hepatocytes/A_2AR^−/− BMDM co-cultures were also significantly higher than their respective levels in hepatocytes/A_2AR^+/+ BMDM co-cultures. Additionally, the phosphorylation states of Akt in hepatocytes co-cultured with A_2AR^−/− BMDM were significantly lower than those in control hepatocytes or control co-cultures (Figure 5D). Taken together, these results suggest that the A_2AR in macrophages protects against macrophage proinflammatory activation. The latter has detrimental effects on increasing hepatocyte fat deposition and proinflammatory responses and on impairing hepatocyte insulin sensitivity.

Figure 5. A_2AR disruption exacerbates the effects of macrophages on increasing hepatocyte fat deposition, cytokine expression, and insulin resistance.

Macrophage-hepatocyte co-cultures were performed as described in Methods. A set of primary hepatocytes were incubated without macrophages and followed by the same treatments as co-cultures. (A) Hepatocyte fat deposition. The cells were treated with palmitate (Pal, 250 μM, conjugated in bovine serum albumin (BSA)) or BSA for the last 24 hr of the 48 hr incubation period, and stained with Oil Red O for 1 hr. Bar graphs, quantification of fat content and triglyceride levels. (B, C) The mRNA levels of genes related to fat metabolism (B) and proinflammatory cytokines (C) were examined using real-time RT-PCR. Prior to harvest, cells were treated without or with LPS (20 ng/ml) for 6 hr. (D) Insulin signaling. Prior to harvest, cells were treated with insulin (100 nM) or PBS for 30 min. Cell lysates were subjected to Western blot analysis. Bar graph, quantification of blots. For A – D, numeric data are means ± SE, n = 4 – 6. Statistical difference between co-cultures with A_2AR^−/− BMDM and co-cultures with A_2AR^+/+ BMDM: *, P < 0.05 and **, P < 0.01 in A, in B and C within the same gene, or in bar graph of D under insulin-stimulated condition; statistical difference between co-cultures with A_2AR^−/− BMDM and hepatocytes cultured without BMDM (None): ^†, P < 0.05 and ^††, P < 0.01 in A, in B and C within the same gene, or in bar graph of D under insulin-stimulated condition; statistical difference between co-cultures with A_2AR^+/+ BMDM and hepatocytes/None: ^‡, P < 0.05 in bar graph of D under insulin-stimulated condition.

A_2AR disruption exacerbates hepatocyte fat deposition and proinflammatory responses

A_2AR appeared to play a protective role in NAFLD. Next, we examined the direct effects of A_2AR disruption on hepatocyte responses. When fat deposition was analyzed, hepatocytes from A_2AR^−/− mice accumulated much more fat and revealed significantly higher levels of triglycerides than hepatocytes from A_2AR^+/− mice or A_2AR^+/+ mice under palmitate-stimulated conditions (Figure 6A). When proinflammatory signaling was analyzed, the phosphorylation states of JNK p46 and NFκB p65 in hepatocytes from A_2AR^−/− mice were significantly stronger than those in hepatocytes from A_2AR^+/− mice or A_2AR^+/+ mice under LPS-stimulated conditions (Figure 6B). When insulin sensitivity was analyzed, the states of insulin-stimulated Akt phosphorylation in hepatocytes from A_2AR^−/− or A_2AR^+/− mice were significantly lower than those in hepatocytes from A_2AR^+/+ mice (Figure 6C). Consistently, A_2AR inhibition increased hepatocyte fat deposition and proinflammatory responses, and decreased hepatocyte insulin signaling (Supplemental Figure S11). Taken together, these results indicate a direct role of A_2AR in protecting hepatocytes from palmitate-induced fat deposition and from LPS-induced proinflammatory responses.

Figure 6. A_2AR disruption exacerbates hepatocyte fat deposition and proinflammatory responses and impairs hepatocyte insulin sensitivity.

Primary hepatocytes were isolated from chow-diet-fed male A_2AR^−/− mice, A_2AR^+/− mice, and A_2AR^+/+ mice, at 10 – 12 weeks of age. (A) Hepatocyte fat deposition. After attachment, hepatocytes were treated with palmitate (Pal, 250 μM) or BSA for 24 hr, and stained with Oil Red O for 1 hr. Bar graphs, quantification of fat content and triglyceride levels. (B, C) Hepatocyte proinflammatory (B) and insulin (C) signaling. Prior to harvest, cells were treated with LPS (100 ng/ml), insulin (100 nM), or PBS for 30 min. Cell lysates were subjected to Western blot analysis. For A – C, numeric data are means ± SE, n = 4 – 6. Statistical difference between A_2AR^−/− vs. A_2AR^+/+: *, P < 0.05 and **, P < 0.01 in A, in B under LPS-stimulated condition, or in C under insulin-stimulated condition; statistical difference between A_2AR^−/− and A_2AR^{+/−: †}, P < 0.05 in A or in B under LPS-stimulated condition; statistical difference between A_2AR^+/− and A_2AR^{+/+: ‡‡}, P < 0.01 in C under insulin-stimulated condition.

A_2AR disruption increases the expression and transcription activity of SREBP1c

A_2AR disruption exacerbated hepatic steatosis and increased lipogenic events. To gain mechanistic insights for A_2AR regulation of lipogenesis, we examined the effect of A_2AR disruption on SREBP1c in mice and in isolated primary mouse hepatocytes. Since feeding increases the amount and/or transcription activity of SREBP1c, we examined whether A_2AR disruption alters the effect of feeding on SREBP1c. Compared with fasting, feeding caused a significant increase in the abundance of both cytosolic and nuclear SREBP1c in livers of A_2AR^+/+ mice, but not A_2AR^−/− mice. Strikingly, the abundance of either cytosolic or nuclear SREBP1c in livers of A_2AR^−/− mice was greater than that in livers of A_2AR^+/+ mice under fasted conditions, but not fed conditions (Figure 7A). Consistently, the abundance of cytosolic or nuclear SREBP1c was increased in hepatocytes from A_2AR^−/− mice compared with that in hepatocytes from A_2AR^+/+ mice under conditions mimicking fasting (Figure 7B). We then performed the reporter assay and observed that SREBP1c transcription activity in hepatocytes from A_2AR^−/− mice was significantly higher than that in hepatocytes from A_2AR^+/+ mice under basal conditions (in the absence of insulin) (Figure 7C). Similar changes were also observed in hepatocytes under insulin-stimulated conditions. However, insulin increased SREBP1c transcription activity in hepatocytes from A_2AR^+/+ but not A_2AR^−/− mice. Next, we performed the ChIP assay. The binding of SREBP1c to SRE/E (a binding element in FAS promoter region) was significantly increased in nuclear extracts of hepatocytes from A_2AR^−/− mice compared with that observed in hepatocytes from A_2AR^+/+ mice (Figure 7C). Also, the mRNAs of key target genes of SREBP1c, e.g., glucokinase and FAS, were greater in hepatocytes from A_2AR^−/− mice compared to those of hepatocytes from A_2AR^+/+ mice (Figure 7D). These results suggest that A_2AR is capable of repressing SREBP1c expression under fasted conditions and suppressing SREBP1c transcription activity.

Figure 7. A_2AR deficiency elevates hepatic abundance of SREBP1c under fasting/nutrient deprivation and enhances SREBP1c transcription activity.

(A) Liver SREBP1c abundance. Male A_2AR^−/− mice and A_2AR^+/+ mice, at 10 – 12 weeks of age, were free-fed or fasted for 18 hr. (B) Hepatocyte SREBP1c abundance. Primary hepatocytes were incubated in M199 in the absence of fetal bovine serum for 24 hr. For A and B, liver (A) or hepatocyte (B) cytosolic and nuclear proteins were subjected to Western blot analysis. (C) Hepatocyte SREBP1c transcription activity. Left panel, primary hepatocytes were incubated in M199 in the absence of fetal bovine serum and transfected with a reporter construct in which luciferase expression is under the control of SRE sequences on fatty acid synthase (pFAS-SRE-luc) or a control (pGL3-luc) for 24 hr; right panel, primary hepatocytes were treated as described in B. Hepatocyte chromatins were immunoprecipitated with antibodies against SREBP1c. The resultant DNA were analyzed for SRE/E sequences of FAS promoter. (D) Hepatocyte mRNAs. Cells were treated as described in B. (E) Liver AMPK phosphorylation states. Male A_2AR^−/− mice, A_2AR^+/− mice, and A_2AR^+/+ mice were fed as described in Figure 2. (F) Hepatocyte ADP/ATP ratio. Cells were treated as described in B. For B – D and F, primary hepatocytes were isolated from chow-diet-fed male A_2AR^−/− mice and A_2AR^+/+ mice, at 10 – 12 weeks of age. For A, B, and E, blots were quantified using densitometry. For all bar graphs, data are means ± SE, n = 4 – 6 (in A – C) or n = 6 – 8 (in D – F). Statistical difference between A_2AR^−/− and A_2AR^+/+: *, P < 0.05 and **, P < 0.01 in A and B for the same fraction, in C for the same construct under the same condition, in D for the same gene, or in E and F; statistical difference between A_2AR^−/− and A_2AR^{+/−: †}, P < 0.05 in E and F; statistical difference between insulin-treated A_2AR^+/+ cells and control-treated A_2AR^+/+ cells (transfected with pFAS-SRE-luc): ^‡, P < 0.05 in C; statistical difference between A_2AR^+/− and A_2AR^{+/+: ‡‡}, P < 0.01 in F.

We also examined the phosphorylation status of AMP-activated protein kinase (AMPK), a regulator that exerts a powerful anti-steatotic effect. Compared with those in HFD-fed A_2AR^+/+ mice or A_2AR^+/− mice, hepatic phosphorylation states of AMPK in HFD-fed A_2AR^−/− mice were significantly decreased (Figure 7E). Since cellular ADP: ATP ratio critically influences AMPK phosphorylation, we performed cellular assays and observed that the ADP: ATP ratios were significantly lower in hepatocytes from A_2AR^−/− mice than those of hepatocytes from A_2AR^+/+ mice or A_2AR^+/− mice (Figure 7F). These results were consistent with changes in the status of hepatic steatosis.

DISCUSSION

We established an association between A_2AR and obesity-associated NAFLD using HFD-fed WT mice, in which increased liver proinflammatory responses were accompanied by increased liver abundance of A_2AR. Considering that A_2AR activation exerts powerful anti-inflammatory effects (19, 20, 30, 31) whereas A_2AR deficiency exacerbates proinflammatory responses (32), we speculated that the increase in liver A_2AR abundance in HFD-fed WT mice was a defensive response. As substantial evidence, global A_2AR deficient mice displayed a significant increase in the severity of HFD-induced liver inflammation and other NAFLD aspects. Of importance, this evidence from A_2AR deficient mice establishes a cause-and-effect relationship between A_2AR disruption and NAFLD. In global A_2AR-deficient mice, increased NAFLD aspects may be attributable to the lack of A_2AR in macrophages, hepatocytes, or both. Considering the importance of macrophages in regulating inflammation, we speculated that the A_2AR in macrophages was needed for protecting mice from obesity-associated NAFLD and we found this was the case. Therefore, our three lines of in vivo evidence makes it conceivable that A_2AR acts through, at least partially, suppressing inflammation to protect against obesity-associated NAFLD.

A_2AR activation was previously shown to ameliorate methionine- and choline-deficient diet (MCD)-induced NASH in rats (19). A follow-up study confirmed the beneficial effect of A_2AR activation in MCD-fed mice (20), which was associated with decreased proinflammatory responses in both hepatocytes and lymphocytes. These results suggest anti-inflammation as an essential mechanism by which A_2AR activation improves NASH. However, the studies by Carini et al. did not address how macrophages respond to A_2AR activation in the context of protecting against NASH. Indeed, the observed decreases in cytokine production by hepatic mononuclear cells of MCD-fed mice upon A_2AR activation could be secondary to the anti-inflammatory effects of A_2AR activation on hepatocytes and/or macrophages. Considering that macrophages have been implicated to critically determine the development and progression of obesity-associated NAFLD (11, 12), it is necessary to address how the A_2AR in macrophages alters NAFLD aspects. In the present study, we confirmed that A_2AR disruption exacerbates macrophage proinflammatory activation. Moreover, we revealed, for the first time, that myeloid cell-specific A_2AR deficiency exacerbated HFD-induced hepatic steatosis and inflammation. Because A_2AR was disrupted only in myeloid cells, increased hepatic steatosis and inflammation in HFD-fed LysMCre⁺-A_2AR^F/F mice appeared to be secondary to the consequences of A_2AR disruption in myeloid cells. As supporting evidence, hepatocytes co-cultured with A_2AR-disrupted macrophages revealed significant increases in the expression of proinflammatory cytokines and in palmitate-induced fat deposition and in the expression of ACC1, FAS, and SREBP1c compared with hepatocytes co-cultured with control macrophages. Because the initial differences among co-cultures existed solely in macrophages (e.g., the presence or absence of A_2AR), the outcomes of hepatocyte lipogenic events were attributable to the effects secondary to A_2AR disruption-associated increase in macrophage proinflammatory activation. The latter has been previously shown to generate macrophage-derived factors, i.e., proinflammatory cytokines, to promote hepatocyte lipogenic events (14), enhance hepatocyte proinflammatory responses, and decrease hepatocyte insulin signaling (33). Therefore, the results from myeloid cell-specific A_2AR deficient mice and macrophage-hepatocyte co-cultures demonstrated that A_2AR has a role in coordinating macrophage actions on hepatocytes to protect against NAFLD aspects.

A_2AR also critically regulates hepatocyte metabolic and inflammatory responses. In support of this, A_2AR disruption brought about deleterious effects on palmitate-induced fat deposition and on LPS-induced proinflammatory responses in primary mouse hepatocytes. In addition, treatment of WT mouse hepatocytes with an A_2AR antagonist generated effects mimicking A_2AR disruption. These two lines of evidence not only recapitulated the findings in HFD-fed A_2AR-disrupted mice, but also served as complementary evidence to support the anti-inflammatory and anti-steatotic effects of A_2AR activation (19, 20). However, it remains to be explored, within hepatocytes, whether and how the anti-inflammatory effect of A_2AR interplays with the anti-lipogenic effect of A_2AR. Given the co-existence of these effects of A_2AR, it is possible that A_2AR acts through suppressing hepatocyte inflammatory signaling to inhibit lipogenic events. Alternatively, A_2AR may act through two parallel pathways to suppress hepatocyte proinflammatory responses and to inhibit hepatocyte fat deposition.

It is a significant finding that A_2AR disruption elevated hepatic SREBP1c abundancy. Additionally, increased liver abundancy of SREBP1c in HFD-fed A_2AR-deficient mice was accompanied with increases in hepatic expression of lipogenic enzymes and in the severity of hepatic steatosis. These results are consistent with an established paradigm concerning the critical contribution of lipogenesis, at increased levels, to the development of hepatic steatosis (34–36). Physiologically, hepatic SREBP1c is stimulated by feeding and repressed by fasting (36), which accounts for increased hepatic fat content under fed states and decreased hepatic fat content under fasted states, respectively. Strikingly and interestingly, in the present study, A_2AR deficiency increased liver SREBP1c abundance under fasted conditions, where SREBP1c expression and activity should be suppressed physiologically. Consistently, A_2AR disruption increased the abundance of SREBP1c in hepatocytes upon nutrient deprivation. These findings support a repressive effect A_2AR on SREBP1c. Additional to altering SREBP1c abundance, A_2AR disruption also decreased hepatic AMPK phosphorylation states, likely through a mechanism involving decreases in hepatocyte ADP: ATP ratio because AMPK is activated by increased ratios of AMP: ATP and ADP: ATP. Considering the role of AMPK in phosphorylating and inactivating SREBP1c (37), A_2AR deficiency-associated decreases in AMPK phosphorylation appeared to also contribute to the effects of A_2AR deficiency on increasing hepatic lipogenesis and on exacerbating hepatic steatosis. Indeed, at the cellular level, A_2AR disruption increased SREBP1c transcription activity and enhanced the binding of SREBP1c to the promoter region of FAS. Collectively, our findings suggest that A_2AR signaling pathways are involved in repression of SREBP1c, which in turn contributes to the effects of A_2AR on suppressing lipogenic events, and on protecting against the development of hepatic steatosis.

In summary, we validated a protective role for A_2AR in obesity-associated NAFLD. Mechanistically, this role of A_2AR is attributable to the direct effects of A_2AR on altering inflammatory and metabolic responses of macrophages and hepatocytes. Specifically, A_2AR suppresses macrophage and hepatocyte proinflammatory responses. A_2AR suppression of macrophage activation also generates secondary effects on inhibiting the inflammatory responses and lipogenic events of hepatocytes. Lastly, A_2AR can directly suppress hepatocyte fat deposition, which is attributable to the effects of A_2AR on repressing SREBP1c expression under fasted states and on suppressing SREBP1c transcription activity. Therefore, targeting A_2AR to suppress inflammation and lipogenesis is a viable therapeutic strategy for the treatment of NAFLD and inflammation-associated human liver diseases.

Abbreviations

ACC1

acetyl-CoA carboxylase 1

A_2AR

adenosine 2A receptor

ALD

alcoholic fatty liver disease

AMPK

AMP-activated protein kinase

BMDM

bone marrow-derived macrophages

BSA

bovine serum albumin

CD

chow diet

ChIP

chromatin immunoprecipitation

ChREBP

carbohydrate-responsive element-binding protein

CPT1a

carnitine palmitoyltransferase 1a

DMEM

Dulbecco’s modified Eagle’s medium

FAS

fatty acid synthase

FBS

fetal bovine serum

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

GTT

glucose tolerance test

H&E

hematoxylin and eosin

HFD

high-fat diet

LFD

low-fat diet

IL-1β

interleukin 1β

IL-6

interleukin 6

ITT

insulin tolerance test

LPS

lipopolysaccharide

JNK

c-Jun N-terminal kinases

MCD

methionine- and choline-deficient diet

NAFLD

non-alcoholic fatty liver disease

NASH

non-alcoholic steatohepatitis

NFκB

nuclear factor kappa B

PBS

phosphate-buffered saline

P-Akt

phosphorylated Akt

Pp65

phosphorylated p65 subunit of NFκB

Pp46

phosphorylated JNK1 (p46)

RQ

respiratory quotient

SREBP1c

sterol regulatory element-binding protein 1c

TG

triglycerides

TNFα

tumor necrosis factor α

WAT

white adipose tissue