Introduction

Nonalcoholic fatty liver disease (NAFLD), the most common chronic liver disease worldwide, is becoming an escalating public health problem [1]. Nonalcoholic steatohepatitis (NASH) is the severely advanced stage of NAFLD and is characterized by hepatic steatosis, inflammation, hepatocyte apoptosis, and fibrosis [2, 3]. Unfortunately, no specific drugs are available for NASH, and lifestyle intervention through exercise and diet modification is considered the primary treatment strategy for NASH [4, 5]. Exercise is an effective treatment for NAFLD even in the absence of a marked reduction in body weight [6, 7]. However, the mechanisms responsible for the protective effect of exercise against NASH are not fully understood.

Inflammasomes are cytoplasmic multiprotein complexes that sense a range of exogenous and endogenous stimuli and promote the production of the proinflammatory cytokines IL-1β and IL-18 by activating Caspase-1 and Caspase-11. Most inflammasomes consist of a sensor, an adapter and an effector and are named after the sensor protein [8, 9]. The NACHT, LRR, and PYD domains-containing protein 3 (NLRP3) inflammasome, which is highly expressed in the liver, is the best-characterized inflammasome [10]. After activated, the sensor protein NLRP3 undergoes oligomerization and recruits the adapter protein apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC), which promotes cleavage of pro-Caspase-1 to produce active Caspase-1. The effector of the NLRP3 inflammasome, namely active Caspase-1, processes pro-IL-1β, and elicits the release of mature IL-1β [11]. Many activators of the NLRP3 inflammasome promote reactive oxygen species (ROS) generation. ROS act as triggers of NLRP3 inflammasome activation [12]. NLRP3 inflammasome-mediated inflammation has been demonstrated to play a crucial role in NASH. Inhibition of NLRP3 inflammasome activation by the drug MCC-950 reduced hepatic inflammation and fibrosis in atherogenic and methionine and choline-deficient (MCD) diet-induced NASH models [10]. During MCD diet feeding, mice deficient in NLRP3 were found to display less severe NASH than the wild-type littermate control mice [13]. On the other hand, mice with global activation of the NLRP3 inflammasome exhibited more severe hepatic inflammation and fibrosis [14].

Adropin, a secreted protein encoded by the Energy Homeostasis Associated gene (Enho), was initially identified in the liver by Kumar et al. in 2008. Adropin participates in energy homeostasis and regulates glucose and lipid metabolism [15, 16], and studies have suggested that this protein protects against NAFLD [17, 18]. In addition to reducing the hepatic lipid content, adropin treatment decreased the expression of the proinflammatory cytokines IL-1β, IL-6, and TNFα in mice with MCD diet-induced NASH [17], indicating its anti-inflammatory effect in NASH.

NLRP3 inflammasome-induced inflammation is a crucial mechanism in NASH progression. However, whether exercise impacts the hepatic NLRP3 inflammasome is unknown, and the underlying mechanism is undetermined. In the present study, using murine models of high-fat diet (HFD)-induced and MCD diet-induced NASH, we identified that exercise inhibited HFD-induced and MCD-induced NLRP3 inflammasome activation and ameliorated NASH. Moreover, serum adropin levels were significantly enhanced by exercise and were negatively correlated with serum IL-1β levels. Since activation of the NLRP3 inflammasome led to an increase in IL-1β production, we speculated that exercise-induced inhibition of NLRP3 inflammasome activation is mediated by adropin. As expected, adropin inhibited ROS production and NLRP3 inflammasome activation in hepatocytes and Kupffer cells treated with palmitic acid (PA). Collectively, these results indicate that exercise may suppress NLRP3 inflammasome activation via adropin induction, providing a potential mechanism underlying the protective effect of exercise against NASH.

Materials and methods

Animal models

NASH was induced by HFD or MCD diet feeding. Male C57BL/6J mice aged 6 weeks were purchased from the Experimental Animal Center of Guangdong Province (Guangzhou, China). After a 2-week adaptation period, mice were randomly divided into six groups (n = 9 mice/group): HFD control (HFD-C) group, HFD group, HFD plus exercise (HFD + EXE) group, MCD control (MCD-C) group, MCD group, and MCD plus exercise (MCD + EXE) group. Mice in the HFD-C group received a low-fat diet (LFD) containing 10% kcal from fat (D12450J, Research Diets Inc., New Brunswick, NJ, USA) for 24 weeks. Mice in the HFD group and HFD + EXE group mice were fed a HFD containing 60% kcal from fat (D12492, Research Diets Inc.) for 24 weeks, and after 12 weeks of HFD feeding, mice in the HFD + EXE group were subjected to exercise training for 12 weeks. Mice in the MCD group were placed on a MCD diet (A02082002B, Research Diets Inc.) for 8 weeks, and mice in the MCD + EXE group received a MCD diet and exercise intervention for 8 weeks. Mice in the MCD-C group were fed an identical diet supplemented with methionine and choline (A02082003B, Research Diets Inc.) for 8 weeks. All animals were maintained on a 12 h light–dark cycle at 22–24 °C and were provided free access to food and water. After the experimental period, all mice were kept sedentary for 2 days and were fasted overnight before sacrifice for collection of blood and liver tissue. All aspects of this research were conducted in accordance with the Chinese Guidelines for Animal Welfare and Experimental Protocols. Approval was obtained from the Animal Experiment Administration Committee of Guangzhou Sport University.

Exercise protocol

Mice were trained on a six-lane small rodent treadmill (Exer-3/6, Columbus Instruments, Columbus, OH, USA) at a 0% grade 5 days per week. A 5 min of warmup period at 6 m/min, a 20 min main exercise period at 10 m/min, and a 5 min cooldown period at 6 m/min were implemented and electric stimulation (0.34 mA) was used to promote running during the first week for adaptation. From the 2nd week to the final week, a 5 min warmup period at 9 m/min, a 50 min main exercise period at 12 m/min (75% maximum oxygen consumption) [19], and a 5 min cooldown period at 9 m/min were implemented. Electric stimulation was removed, and mice were encouraged to run using a soft brush pressed against their posterior region to reduce the potential stress of continued electric shocking.

Blood analysis

Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels in serum were measured using commercial kits (Jiancheng Bioengineering Institute, Nanjing, China). Serum adropin (Phoenix Pharmaceuticals, Belmont, CA, USA) and IL-1β (R&D Systems, Minneapolis, MN, USA) levels were measured with ELISA kits, according to the manufacturer’s instructions.

Hepatic triglyceride (TG) and total cholesterol (TC) analysis

Hepatic TG and TC levels were measured with commercial kits (Applygen Technologies Inc., Beijing, China).

Histological analysis of liver sections

Fresh liver tissues were fixed with 4% paraformaldehyde solution for 24 h, embedded in paraffin, and sliced into 4-µm sections for hematoxylin-eosin (H&E) staining and Sirius Red staining. Cryosections were fixed and dehydrated and were then used for Oil red O staining. The NAFLD activity score (NAS) was calculated according to the guidance provided by the Pathology Committee of the NASH Clinical Research Network [20]: steatosis (<5% = 0, 5–33% = 1, 33–66% = 2, >66% = 3), lobular inflammation (none = 0, <2 foci = 1, 2–4 foci = 2, >4 foci = 3), and hepatocellular ballooning (none = 0, few = 1, prominent = 2). All features were scored in a blinded manner based on six fields of view per sample. The individual scores for each field of view were summed to calculate the NAS for each animal. Histological assessments were performed by a pathologist who was blinded to the treatment.

Immunohistochemical analysis of liver sections

Liver tissues were fixed with 4% paraformaldehyde solution. After blocking, slides were incubated with specific primary antibodies (anti-F4/80 or anti-Collagen I) overnight at 4 °C and were then incubated with secondary antibodies for 2 h at 37 °C.

Isolation of primary hepatocytes and Kupffer cells

Hepatocytes and Kupffer cells were isolated from the livers of male C57BL/6J mice using a two-step collagenase perfusion technique. In brief, the liver was perfused via the portal vein with Liver Perfusion Medium (Thermo Fisher Scientific, Waltham, MA, USA) and was then perfused with Hank’s balanced salt solution (HBSS) containing 0.05% type IV collagenase (GIBCO, Thermo Fisher Scientific). After the liver was digested, it was removed, cut into small pieces, and passed through a 100 μm strainer. The strained sample was centrifuged for 3 min at 50 × g and 4 °C. The hepatocytes were in the precipitate, and the Kupffer cells were in the supernatant. For hepatocyte isolation, the hepatocyte pellet was resuspended in HBSS and centrifuged two times for 3 min each at 50 × g and 4 °C. Hepatocytes were purified by Percoll gradient centrifugation (50% v/v, Sigma). For Kupffer cell isolation, the supernatant from the first centrifugation step of hepatocyte isolation was loaded on a Percoll gradient (25 and 50%) and centrifuged for 30 min at 2300 rpm and 4 °C. The interphase band containing Kupffer cells was collected and washed two times with PBS. Cell viability was estimated with the trypan blue exclusion test.

Treatment of primary hepatocytes and Kupffer cells

Primary mouse hepatocytes and Kupffer cells were incubated in serum-free DMEM containing 2% fatty acid-free bovine serum albumin in the presence or absence of PA (400 μM) or PA plus adropin (100 ng/ml) for 24 h.

Western blot analysis

Protein was extracted from mouse livers and primary mouse cells. Total protein concentrations were measured with a BCA protein assay kit (Thermo Fisher Scientific). Equal amounts of total protein were separated using SDS-PAGE and transferred to PVDF membranes. Membranes were then blocked and incubated with primary antibodies against adropin (Novus Biologicals, Littleton, CO, USA, NBP1-26387), pro-IL-1β and IL-1β (Cell Signaling Technology, Beverly, MA, USA, #31202), pro-Caspase-1 and active Caspase-1 P10 (Santa Cruz Biotechnology, CA, USA, sc-56036), NLRP3 (Cell Signaling Technology, #15101), ASC (Santa Cruz Biotechnology, sc-514414) and GAPDH (Cell Signaling Technology, #5174). Membranes were incubated for 1 h with the following secondary antibodies: mouse anti-rabbit IgG-HRP (Santa Cruz Biotechnology, sc-2357) and m-IgGκ BP-HRP (Santa Cruz Biotechnology, sc-516102). Immunoreactions were detected by enhanced chemiluminescence (Thermo Fisher Scientific). Band intensities were quantified using Quantity One software.

Quantitative real-time PCR

The primer sequences used are shown in Supplementary Table 1. Expression levels were normalized to those of the housekeeping gene β-Actin.

Caspase-1 activity analysis

Caspase-1 activity was evaluated with a Caspase-1 assay kit (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions.

Measurement of ROS in mouse livers and cells

Dyhidroethidium (DHE) was used to detect ROS production in liver tissues. In brief, flash-frozen liver sections were incubated with DHE for 30 min at 37 °C in a dark room. Images were acquired with a fluorescence microscope (Nikon TE2000; Nikon Corporation, Tokyo, Japan) using an excitation wavelength of 520–540 nm and a rhodamine emission filter, and the fluorescence intensity was analyzed.

The relative levels of ROS in primary hepatocytes and Kupffer cells were measured with a DCFH-DA assay. Cells were incubated with DCFH-DA for 30 min. The fluorescence intensity was measured in a microplate fluorometer at wavelengths of 488 nm (excitation) and 535 nm (emission).

Statistical analysis

All data were expressed as mean ± SEM values. Statistical significance was evaluated by one-way ANOVA with the Bonferroni test for multiple comparisons. The correlation of serum adropin levels with serum IL-1β levels was analyzed by Pearson correlation analysis. All analyses were performed using SPSS 20.0. Differences with a P value of less than 0.05 were considered significant.

Results

Exercise ameliorates HFD-induced and MCD diet-induced hepatic steatosis, inflammation, and fibrosis

We established two murine models of NASH by feeding the mice a HFD or a MCD diet. Exercise suppressed HFD-induced increases in body weight and liver weight but did not affect MCD diet-induced decreases in body and liver weight (Table 1). H&E staining and Oil red O staining were performed (Fig. 1A, B and Supplementary Fig. 1A), and hepatic TG and TC levels were detected (Supplementary Fig. 1B). Our results showed that hepatic steatosis caused by HFD feeding and MCD diet feeding was significantly reduced by exercise.

Table 1 Body weight and liver weight of mice in each study group.
Full size table
Fig. 1: Exercise reduced HFD-induced and MCD diet-induced hepatic steatosis, inflammation, and fibrosis.
figure 1

A Representative images (×20 objective) of liver sections stained with H&E. B Hepatic histological analysis of steatosis, inflammation, ballooning, and NAFLD activity score (NAS). C Hepatic immunohistochemical staining with an anti-mouse F4/80 antibody. Images were observed using a ×20 objective. D Hepatic mRNA expression of inflammation-related genes. E Representative images (×20 objective) of liver sections stained with Sirius Red. F Collagen content determined by counting Sirius Red positive areas in six randomly selected fields using Image Pro Plus 6.0 software. Data are presented as the mean ± SEM; n = 8–9 per group. ap < 0.05 for difference from control diet mice (HFD-C group or MCD-C group); bp < 0.05 for difference from diet-induced NASH mice (HFD group or MCD group).

Full size image

Histological parameters reflecting hepatic inflammation, such as the lobular inflammation score, ballooning score, and total NAS (Fig. 1B), were significantly reduced by exercise. Exercise lowered serum ALT and AST levels (Supplementary Fig. 2A, B). To further investigate the effect of exercise on liver inflammation, we performed staining for the macrophage marker F4/80 and measured the mRNA levels of inflammation-related genes. Exercise alleviated hepatic macrophage infiltration and reduced the levels of proinflammatory cytokines such as IL-1β, TNFα, and MCP-1 in HFD-fed and MCD diet-fed mice (Fig. 1C, D).

Sirius Red staining revealed that exercise suppressed HFD-induced and MCD diet-induced collagen accumulation (Fig. 1E, F). Collagen I levels were lower in HFD + EXE group mice and MCD + EXE group mice than in HFD mice and MCD mice, respectively (Supplementary Fig. 2C, D).

Exercise suppresses HFD-induced and MCD diet-induced NLRP3 inflammasome activation

Both the high-fat and MCD diets activated NLRP3 inflammasome, as indicated by increased levels of NLRP3 inflammasome components, enhanced Caspase-1 enzymatic activity, and elevated IL-1β production in HFD group mice and MCD group mice compared with the mice in the corresponding control diet groups (Fig. 2). Exercise decreased the protein levels of NLRP3, ASC, pro-Caspase-1, active Caspase-1, pro- IL-1β, and IL-1β in HFD-fed mice (Fig. 2A), and reduced the levels of all these proteins except pro-IL-1β in MCD diet-fed mice (Fig. 2B). In addition, the mRNA levels of NLRP3, ASC, and Caspase-1 were reduced by exercise training (Fig. 2C). Furthermore, exercise significantly decreased Caspase-1 enzymatic activity (Fig. 2D) and serum IL-1β levels (Fig. 2E) in both NASH models. These data suggest the inhibitory effect of exercise on NLRP3 inflammasome activation in mice with diet-induced NASH.

Fig. 2: Exercise blocked HFD-induced and MCD diet-induced NLRP3 inflammasome activation.
figure 2

A Western blot analyses of NLRP3, ASC, pro-Caspase-1, active Caspase-1, pro-IL-1β, and IL-1β in the liver of HFD-C, HFD, and HFD + EXE mice. B Western blot analyses of NLRP3, ASC, pro-Caspase-1, active Caspase-1, pro-IL-1β, and IL-1β in the liver of MCD-C, MCD, and MCD + EXE mice. C Hepatic mRNA levels of NLRP3, ASC, and Caspase-1. D Hepatic Caspase-1 activity. E Serum IL-1β levels. Data are presented as the mean ± SEM; n = 8–9 per group. ap < 0.05 for difference from the control diet mice (HFD-C group or MCD-C group); bp < 0.05 for difference from the diet-induced NASH mice (HFD group or MCD group).

Full size image

ROS act as triggers for NLRP3 inflammasome activation [21]. DHE staining showed that ROS generation was notably increased in mice with either HFD-induced or MCD diet-induced NASH. Exercise reduced ROS production in both HFD-fed and MCD diet-fed mice (Fig. 3A, B).

Fig. 3: Exercise suppressed HFD-induced and MCD diet-induced hepatic ROS production.
figure 3

A Representative images (×20 objective) of liver DHE staining. B Quantitative analysis of DHE staining. Data are presented as the mean ± SEM; n = 8–9 per group. ap < 0.05 for difference from the control diet mice (HFD-C group or MCD-C group); bp < 0.05 for difference from the diet-induced NASH mice (HFD group or MCD group).

Full size image

Adropin levels are enhanced by exercise and are negatively correlated with serum levels of IL-1β

To examine the effect of exercise on adropin expression, we measured serum adropin levels and hepatic expression levels of the Enho gene, which encodes adropin. As shown in Fig. 4, both HFD feeding and MCD diet feeding significantly reduced serum levels of adropin, and these reductions were almost completely reversed by exercise (Fig. 4A). Moreover, exercise blocked the reductions in hepatic Enho gene expression induced by HFD feeding and MCD diet feeding (Fig. 4B). Studies have demonstrated that hepatic Enho mRNA expression is regulated by the liver X-receptor alpha (LXRα) [15]. We examined the effect of exercise on the mRNA expression level of LXRα and found that exercise caused a significant reduction in hepatic LXRα mRNA expression in both HFD-fed and MCD diet-fed mice (Fig. 4C).

Fig. 4: Exercise inhibited HFD-induced and MCD diet-induced decrease in serum adropin levels, and serum adropin levels were negatively correlated with serum IL-1β levels.
figure 4

A Serum levels of adropin. B Hepatic mRNA expression of Enho. C Hepatic mRNA levels of LXRα. D Correlation of serum adropin levels with serum IL-1β levels. Data are presented as the mean ± SEM; n = 8–9 per group. ap < 0.05 for difference from the control diet mice (HFD-C group or MCD-C group); bp < 0.05 for difference from the diet-induced NASH mice (HFD group or MCD group).

Full size image

Furthermore, circulating adropin levels were negatively associated with serum IL-1β levels in mice with diet-induced NASH (Fig. 4D). Given that IL-1β levels are increased upon NLRP3 inflammasome activation, we hypothesize that the protective effect of exercise against NLRP3 inflammasome activation in NASH may be mediated by adropin.

Adropin inhibits PA-induced NLRP3 inflammasome activation in hepatocytes and Kupffer cells

Hepatocytes are the major cells in the liver, and Kupffer cells play a key role in inflammation in NASH [22]. We further investigated the effect of adropin on the NLRP3 inflammasome in hepatocytes and Kupffer cells. Protein and mRNA expression levels of NLRP3 inflammasome components, Caspase-1 activity, and IL-1β expression were increased in both hepatocytes and Kupffer cells upon PA treatment (Fig. 5A–C). Adropin treatment significantly reduced the protein and mRNA levels of Caspase-1 and IL-1β in hepatocytes and Kupffer cells and markedly decreased the protein and mRNA levels of NLRP3 in hepatocytes. Adropin lowered the mRNA level of NLRP3 but did not significantly affect its protein level in Kupffer cells. No significant changes were observed in ASC expression in hepatocytes or Kupffer cells upon adropin treatment (Fig. 5A, B). Importantly, adropin inhibited PA-induced enhancement of Caspase-1 activity in both hepatocytes and Kupffer cells (Fig. 5C). These findings suggest that PA-induced NLRP3 inflammasome activation in hepatocytes and Kupffer cells can be suppressed by adropin treatment.

Fig. 5: Adropin treatment suppressed PA-induced NLRP3 inflammasome activation and ROS generation in hepatocytes and Kupffer cells.
figure 5

A Western blot analyses of NLRP3, ASC, pro-Caspase-1, active Caspase-1, pro-IL-1β, and IL-1β in hepatocytes and Kupffer cells. B mRNA levels of NLRP3, ASC, and Caspase-1 in hepatocytes (up) and Kupffer cells (down). C Caspase-1 activity of hepatocytes (up) and Kupffer cells (down). D The relative levels of ROS in hepatocytes (up) and Kupffer cells (down). Data are presented as the mean ± SEM. ap < 0.05 vs. control (CTRL), bp < 0.05 vs. PA, (n = 4 for all the experiments).

Full size image

The saturated fatty acid PA activates NLRP3 inflammasome by increasing ROS generation [23]. Thus, we further detected ROS production in hepatocytes and Kupffer cells. As shown in Fig. 5D, PA-induced ROS production in hepatocytes and Kupffer cells was markedly reduced by adropin treatment. These results indicate that adropin may suppress NLRP3 inflammasome activation through ROS inhibition.

Discussion

The major finding of our study was that exercise can suppress NLRP3 inflammasome activation in mice with diet-induced NASH and that this effect may be mediated by adropin induction.

A sedentary lifestyle and unhealthy eating habits contribute to the high prevalence of NAFLD. NASH, the inflammatory subtype of NAFLD, has attracted extensive attention due to its association with an increased risk of liver cirrhosis and hepatocellular carcinoma [24]. Currently no approved NASH-specific pharmacological treatment is available. Lifestyle modification, including exercise and dietary changes, is still the primary treatment strategy and the cornerstone of successful therapy for NASH [5, 25]. Population studies have demonstrate that lifestyle modification is effective in treating NASH not only in overweight patients but also in people with morbid obesity [26, 27]. Exercise is an important strategy for managing NASH [28]. Regular exercise has been reported to reduce hepatic lipid content, inhibited inflammation, and decrease ROS overproduction in the liver [29, 30]. In the present study, we also confirmed that exercise not only reduced hepatic lipid accumulation but also attenuated hepatic inflammation and fibrosis in mice with either HFD-induced or MCD diet-induced NASH.

NLRP3 inflammasome activation and subsequent IL-1β secretion lead to NASH progression, and the expression levels of NLRP3 inflammasome components are markedly increased in patients with NASH and in animal models of NASH [13, 31, 32]. Moreover, blockade of NLRP3 inflammasome activation by NLRP3 inflammasome-associated genes knockout or NLRP3 inhibitor treatment was found to ameliorate hepatic inflammation and fibrosis in NASH mice [10, 13, 33]. After activation, the NLRP3 inflammasome induces cleavage of the IL-1β precursor to produce active IL-1β. IL-1β functions as an important contributor to inflammation in NASH. Indeed, blocking the IL-1β-activated intracellular signaling pathway via IL-1 receptor knockout decreased plasma ALT levels, reduced hepatic inflammation, and lowered the mRNA expression of fibrosis-related genes in NAFLD mice [34, 35]. Acute exercise activated the NLRP3 inflammasome in the myocardium [36], and long-term exercise inhibited NLRP3 inflammasome activation in the hippocampus [37, 38]. However, currently, no published studies investigating the impact of exercise on hepatic NLRP3 inflammasome activation in NASH are available. In the present study, we observed that exercise reduced the mRNA and protein levels of NLRP3 inflammasome components, decreased Caspase-1 activity and suppressed IL-1β production in mice with either HFD-induced or MCD diet-induced NASH. These results suggest that exercise can suppress both HFD-induced and MCD diet-induced NLRP3 activation, which may be a mechanism underlying the beneficial effects elicited by exercise in NASH.

Our results revealed that exercise significantly reduced pro- IL-1β levels in HFD-fed mice, but did not modulate the expression of pro-IL-1β in MCD diet-fed mice. Activation of NLRP3 inflammasome proceeds through two distinct steps: (1) the priming step, in which a first hit leads to the synthesis of biologically inactive pro-IL-1β and transcriptional induction of NLRP3 inflammasome components, and (2) the activation step, in which a second hit causes homotypic oligomerization of NLRPs and formation of the active NLRP3 inflammasome, which cleaves pro-IL-1β and facilitates the production of active IL-1β [39, 40]. Our data suggest that the mechanisms underlying exercise-induced NLRP3 inflammasome inhibition may differ between mice with HFD-induced NASH and mice with MCD diet-induced NASH. Exercise may inhibit the priming step of pro-IL-1β synthesis in mice with HFD-induced NASH but suppress the activation step of NLRP3 inflammasome-mediated cleavage of pro-IL-1β into active IL-1β in mice with MCD diet-induced NASH.

The peptide hormone adropin controls energy metabolism and is regulated by the metabolic status. Serum levels of adropin decrease during fasting. During the consumption of purified diets for 48 h, HFD-fed mice were found to exhibit higher levels of serum adropin than LFD-fed mice; in addition, circulating adropin levels were significantly decreased after 12 weeks of HFD feeding [41]. Since adropin is closely associated with energy metabolism, and exercise is an effective strategy to regulate energy metabolism, it is natural to speculate that exercise may affect adropin levels. A few recent studies have reported that long-term exercise elevates circulating adropin levels in both obese and healthy adults [42,43,44]. Consistent with these results, we observed that serum adropin levels were significantly increased by exercise in mice with diet-induced NASH. Adropin is encoded by Enho, which is expressed mainly in the brain and liver. LXRα regulates hepatic Enho expression; for example, LXRα activation reduces hepatic Enho mRNA expression [15]. We evaluated the hepatic expression of Enho and LXRα and found that exercise increased the mRNA level of Enho and reduced that of LXRα in the liver. Thus, exercise may increase adropin levels via downregulation of LXRα.

Adropin exerts beneficial effects on NAFLD. Circulating adropin levels are decreased in animals and patients with NAFLD [17, 18]. Upon HFD or MCD diet feeding, mice lacking adropin develop more severe hepatic steatosis and inflammation than the wild-type littermate control mice. Furthermore, administration of adropin to NASH mice was found to protect against liver injury [17]. Accumulating evidence suggests that adropin not only regulates glucose and lipid metabolism but also acts as a potential anti-inflammatory factor. A recent study conducted by Chen et al. demonstrated that adropin treatment alleviated hepatic lipid accumulation along with inflammation in NASH mice [17]. Adropin has been suggested to be a link between energy regulation and inflammation regulation [45]. Here, we demonstrated that exercise elevated adropin levels with concomitant inhibition of NLRP3 inflammasome activation in mice with either HFD-induced or MCD diet-induced NASH. More importantly, circulating adropin levels were negatively correlated with serum levels of IL-1β. These results led us to hypothesize that adropin might be involved in the protective effect of exercise against NLRP3 inflammasome activation in NASH mice.

To test the hypothesis, we investigated whether adropin can regulate NLRP3 inflammasome activation in hepatocytes and Kupffer cells, two of the best-studied cell types in NAFLD in which the NLRP3 inflammasome is expressed and exerts its biological activity. Adropin inhibited the PA-induced increases in the expression level and activity of Caspase-1 and decreased the expression level of mature IL-1β in hepatocytes and Kupffer cells, suggesting that adropin can suppress PA-elicited NLRP3 inflammasome activation in hepatocytes and Kupffer cells. Notably, adropin significantly reduced the protein level of NLRP3 in hepatocytes but not in Kupffer cells. Induction of NLRP3 expression and posttranslational modification, such as deubiquitination, are thought to be important in the process of NLRP3 inflammasome formation and activation [46]. Decreased NLRP3 expression may contribute to adropin-mediated inhibition of NLRP3 inflammasome activation in hepatocytes, while adropin may regulate NLRP3 inflammasome activation via posttranslational modification of NLRP3 or other mechanisms in Kupffer cells.

Notably, ASC, the adapter protein in the NLRP3 inflammasome, was downregulated by exercise in mice with diet-induced NASH. Unexpectedly, we did not observe a significant effect of adropin on ASC expression in hepatocytes or Kupffer cells. Therefore, we deduced that adropin does not mediate the exercise-induced decrease in ASC expression and the protective effect of exercise against NLRP3 inflammasome activation in NASH is not mediated completely by adropin induction.

In the present study, we found that exercise-induced elevation of adropin expression was accompanied with decreased levels of ROS. In vitro data showed that adropin inhibited ROS production in both hepatocytes and Kupffer cells. Since ROS have been proposed to be an important regulator of NLRP3 inflammasome activation [47], we speculated that adropin suppresses NLRP3 inflammasome activation by reducing ROS levels. Further investigations is required to evaluate whether adropin interacts directly with NLRP3 inflammasome.

Taken together, our data suggest that exercise increases adropin levels and may thereby suppress NLRP3 inflammasome activation in mice with diet-induced NASH.