1 Introduction

Diet-induced obesity is caused by chronic overnutrition and characterized by excessive body fat storage, which is associated with the development of metabolic diseases, including nonalcoholic fatty liver disease (NAFLD).^{[1, 2]} In the early stage of obesity, excess energy was preserved in the form of triglyceride (TG) in lipid droplets and leads to adipocyte hyperplasia.^[3] Further, visceral adipocyte expansion reaches the maximum and promotes adipocyte dysfunction and lipolysis. Adipocyte lipolysis releases high levels of non-esterified fatty acids (NEFA), which are unable to be sufficiently absorbed and transformed in adipocytes due to adipocyte dysfunction,^[4] and eventually results in the ectopic fat deposited in hepatocytes and the development of NAFLD.^{[5, 6]}

Lipolysis, i.e., neutral hydrolysis of TGs to NEFA and glycerol, requires the involvement of three neutral lipases, including adipose triglyceride lipase (ATGL), hormone sensitive lipase (HSL), and monoacylglycerol lipase (MGL). In adipose tissue, ATGL and HSL are responsible for more than 90% of TG hydrolysis.^[7] Sirtuin 1 (SIRT1) regulates the acetylation status and functional activity of Forkhead box O1 (FoxO1), which directly binds to the ATGL promoter and drivers the activation of adipocyte lipolysis.^[8] The activation of lipolysis triggers the triglyceride hydrolysis and elevates the phosphorylation of Adenosine 5’-monophosphate-activated protein kinase (AMPK).^[9] In addition, intracellular Ca²⁺ accumulation promotes lipolysis by increasing cGMP level.^[10] In mouse studies, epididymal adipose tissue (EAT) is the most commonly used visceral fat depot. In EAT from the mice fed with a high-fat diet (HFD), the elevated serotonin (5-HT) receptor 2B (HTR2B) signaling promotes Ca²⁺-cGMP-PKG cascade and adipocyte lipolysis, which leads to the development of obesity-related complications.^[11] Therefore, 5-HT signaling may drive adipocyte lipolysis through promoting the activation of Ca²⁺-cGMP-PKG cascade and SIRT1/FoxO1/AMPK pathway, which ultimately leads to ectopic fat deposit in hepatocytes.

Luteolin (3´,4´,5,7-tetrahydroxy flavone) is a natural flavonoid, widely distributed in a variety of fruits and vegetables, such as chrysanthemum flowers, celery, broccoli, and parsley.^[12] Some studies have demonstrated that dietary luteolin reduces body weight gain and associated hepatic steatosis in HFD-fed animals.^{[13, 14]} Thus, the improvement of fatty liver may be a secondary and concomitant outcome of body weight loss. Independent of the body weight loss, whether dietary luteolin can substantially reduce hepatic steatosis need to be further investigation.

In this study, the effect of luteolin on 5-HT-induced lipolysis was firstly assessed in mature 3T3-L1 adipocytes. To investigate the mechanism by which luteolin inhibits 5-HT-induced adipocyte lipolysis, we detected Ca²⁺-PKG cascade and SIRT1/FoxO1/AMPK signaling and simulated the binding of HTR2B and luteolin utilizing a molecular docking program. Next, the mice were fed with a low-fat diet (LFD) or an HFD for 16 weeks. Meanwhile, some HFD-fed mice were switched to a luteolin-containing HFD at the 6th, 8th, or 10th weeks of HFD feeding. Further, the substantial action of dietary luteolin on hepatic steatosis, visceral adiposity, and lipolytic molecules were identified in HFD-fed mice.

2 Results

2.1 Luteolin Addition Suppressed 5-HT-Induced Adipocyte Lipolysis

The elevated HTR2B signaling promotes visceral adipocyte lipolysis and leads to the development of obesity-related complications.^[11] Here, to examine whether 5-HT directly induce adipocyte lipolysis, 3T3-L1 pre-adipocytes were fully differentiated to mature adipocytes, and subsequently treated with 5-HT at physiological concentration (Figure 1A). In differentiated 3T3-L1 adipocytes, 5-HT treatment significantly reduced intracellular fat accumulation (Figure 1B,C). Along with the alternation of intracellular lipid level, 5-HT treatment also significantly elevated the release of free glycerol and NEFA from 3T3-L1 cells (Figure 1D,E). Meanwhile, the expression of lipolytic enzyme gene Atgl, Hsl, and Mgl (Figure 1F–H), the protein level of ATGL (Figure 1I,J), and the phosphorylation of HSL (Figure 1I,K) were upregulated by 5-HT treatment. Additionally, the expression of lipogenesis-related gene Dgat1, Dgat2, Gpat3, and Mogat2 was unaffected by 5-HT treatment (Figure S1, Supporting Information). The results affirmed that 5-HT treatment induced the lipolysis in 3T3-L1 adipocytes.

Further, to assess the effect of luteolin on 5-HT-induced adipocyte lipolysis, as shown in Figure 1A, 10, 20, or 50 µM luteolin were added. In the context of 5-HT treatment, luteolin addition increased the intracellular lipid levels, although the increase did not reached the statistical significance after 10 or 20 µM luteolin addition (Figure 1B,C). Consistent with the enhancement of intracellular lipid levels, luteolin addition successfully abolished the effects of 5-HT on the release of glycerol and NEFA (Figure 1D,E), the expression of Atgl and Mgl (Figure 1F–H), the protein level of ATGL (Figure 1I,J), and the phosphorylation of HSL (Figure 1I,K). Additionally, in 3T3-L1 adipocytes without 5-HT treatments, the release of glycerol and NEFA and the expression of lipolytic genes were unaffected by luteolin addition (Figure S2, Supporting Information). Therefore, luteolin addition inhibited 5-HT-induced adipocyte lipolysis.

2.2 Luteolin Inhibited 5-HT-Induced Adipocyte Lipolysis via SIRT1/FoxO1/AMPKα Pathway

SIRT1/FoxO1/AMPKα pathway has been implicated in adipocyte lipolysis and FoxO proteins promote triacylglycerol catabolism through regulating ATGL and its inhibitor G0/S1 switch 2 (G0S2).^{[8, 15]} Along with the induction of lipolysis, 5-HT treatment promoted the expression of G0s2, Sirt1, FoxO1, and Ampkα in 3T3-L1 adipocytes (Figure 2A–D). In addition, 5-HT treatment inhibited FoxO1 acetylation, increased the level of SIRT1 and phosphorylated-AMPKα (Figure 2E–H). Except for the expression of Ampkα, luteolin addition totally abated the effects of 5-HT on the aforementioned molecules involved in adipocyte lipolysis (Figure 2), suggesting that luteolin inhibited 5-HT-induced adipocyte lipolysis via SIRT1/FoxO1/AMPKα pathway.

2.3 Luteolin Competitively Bound to HTR2B and Blocked Downstream Ca²⁺-PKG Cascade

5-HT exerts specific biological activities through the different 5-HT receptors (HTRs). Of the HTRs, HTR1A and HTR1B play the primary roles at the neural central, whereas HTR2A and HTR2B are mainly expressed in peripheral tissue and have been implicated in the regulation of metabolic homeostasis.^{[11, 16, 17]} Here, 5-HT treatment upregulated the expression of Htr1a and Htr2b, but not Htr1b and Htr2a, in 3T3-L1 adipocytes (Figure 3A–D). Luteolin addition reversed the upregulation of Htr1a and Htr2b (Figure 3A,D). To identify the possible interaction between luteolin and HTRs, we performed a molecular docking simulation using publicly available molecular structures of proteins and found that the binding energy of HTR2B to luteolin was the highest in the tested HTRs (Figure 3E and Figure S3, Supporting Information). Like 5-HT, luteolin also bound to the active pocket 1 of HTR2B (Figure 3E). And even, relative to 5-HT, luteolin possessed the smaller molecular distance and the higher binding energy to HTR2B (Figure 3F,G). These results suggest that luteolin can competitively space-occupy the active pocket 1 of HTR2B to block 5-HT binding to HTR2B.

As a G protein-coupled receptor, HTR2B activation leads to intracellular Ca²⁺ accumulation and cGMP production, which in turn induces PKG activation and adipocyte lipolysis.^{[11, 18]} As expected, 5-HT treatment significantly increased intracellular Ca²⁺ levels and PKG1 expression (Figure 3H–J). Luteolin addition blunted 5-HT-induced intracellular Ca²⁺ influx and PKG1 expression (Figure 3H–J). Thus, luteolin competitively binds to HTR2B to suppress Htr2b expression and blunt intracellular Ca²⁺-PKG1 cascade, and thereby inhibits the 5-HT-induced adipocyte lipolysis.

2.4 Diet Luteolin Supplementation before the Formation of Fatty Liver Alleviates Hepatic Steatosis in HFD-fed Mice

To assess the effect of dietary luteolin on obesity-related hepatic steatosis, we firstly detected the lipid droplet and triglyceride content in liver in HFD-fed mice and identified that fat liver was formed since the 10th week of HFD feeding (Figure S4A–C, Supporting Information). Next, the other 5-week-old mice were randomly divided into five groups: LFD, HFD, HFD+Lu6, HFD+Lu8, and HFD+Lu10 groups (Figure 4A). In HFD-fed mice, body weight gain and body component were unaffected by dietary luteolin (Figure 4B). However, relative to those in the HFD group, HFD+Lu6 and HFD+Lu8 groups had the lower liver and SAT weights (Figure 4C,D). Intriguingly, EAT weights (Figure 4E) were obviously elevated in all luteolin-treated groups and BAT weights (Figure 4F) were unaffected by dietary luteolin. Along with the reduction of liver weights, the livers from HFD+Lu6 and HFD+Lu8 groups were redder and more elastic (Figure 4G) and contained smaller fat vacuoles (Figure 5A,B). Additionally, diet luteolin supplementation at 6th week of HFD feeding significantly reduced the liver levels of TG, TCHO, NEFA, and glycerol (Figure 5C–F), and down-regulated the expression of gene (Fabp4, CD36, and Fas) involved in lipid transport and synthesis (Figure 5G–I). Consistently, serum TG, LDL-c, NEFA, glycerol, and ALT levels were also improved in HFD+Lu6 group (Table 1). Thus, independent of the body weight loss, diet supplementation of luteolin before fatty liver formation improves hepatic steatosis in HFD-fed mice.

Table 1. Luteolin supplementation before the formation of fatty liver reduced blood lipid and liver injury

Parameter	LFD	HFD	HFD+Lu6	HFD+Lu8	HFD+Lu10
	(N = 10)	(N = 10)	(N = 10)	(N = 10)	(N = 10)
ALT [U L−1]	8.100 ± 0.314****	21.800 ± 2.225	14.000 ± 1.795*	17.700 ± 2.864	19.400 ± 2.045
AST [U L−1]	29.100 ± 2.718	34.100 ± 4.212	28.400 ± 3.205	30.000 ± 3.204	32.600 ± 3.618
ALP [U L−1]	18.600 ± 0.702**	30.400 ± 3.413	27.400 ± 2.141	24.800 ± 2.453	24.300 ± 1.826
TG [mmol L−1]	0.398 ± 0.025	0.403 ± 0.031	0.322 ± 0.020*	0.347 ± 0.018	0.344 ± 0.010
TCHO [mmol L−1]	0.530 ± 0.028****	1.572 ± 0.071	1.617 ± 0.060	1.537 ± 0.054	1.562 ± 0.074
LDL-c [mmol L−1]	0.078 ± 0.002****	0.237 ± 0.010	0.1763 ± 0.013***	0.181 ± 0.007***	0.197 ± 0.011*
HDL-c [mmol L−1]	0.421 ± 0.013****	1.148 ± 0.046	1.196 ± 0.040	1.142 ± 0.033	1.133 ± 0.050
Glycerol [µg mL−1]	277.900 ± 20.471****	446.844 ± 12.020	384.511 ± 12.354*	428.822 ± 7.910	430.156 ± 21.582
NEFA [mmol L−1]	0.388 ± 0.016****	0.691 ± 0.029	0.590 ± 0.029*	0.620 ± 0.020	0.631 ± 0.033

• Note: ATL, Alanine transaminase; AST, Aspartate transaminase; ALP, Alkaline phosphatase; TG, Triglycerides; TCHO, Total cholesterol; LDL-c, Low density lipoprotein cholesterol; HDL-c, High density lipoprotein cholesterol; NEFA, Non-esterified fatty acid.
• Each group was compared with the HFD group, using one-way ANOVA test.
• * p < 0.05;
• ** p < 0.01;
• *** p < 0.001;
• **** p < 0.0001.

2.5 Dietary Luteolin Inhibits EAT Lipolysis and Associated Lipolytic Signaling

In HFD-fed mice, the liver weights and lipid levels are negatively correlated with EAT weights during the development of obesity-induced fatty liver. The activation HTR2B signaling and associated adipocyte lipolysis in EAT leads to ectopic lipid deposition in liver and the development of fatty liver.^[11] Given diet supplementation of luteolin before fatty liver formation elevated EAT weights (Figure 4E) and improved hepatic steatosis (Figure 5), the levels of lipolytic enzymes and the activation of lipolytic signaling molecules should be depressed in EAT in the mice fed with a luteolin-containing HFD. Consistent with this notion, relative to those from HFD group, the mice from HFD+Lu6 group had the lower expression of Atgl, Mgl, and G0s2, protein level of ATGL, and phosphorylation of HSL (Figure 6A–G). Although EAT 5-HT levels were unaffected by dietary luteolin (Figure S5, Supporting Information), diet luteolin supplementation at the 6th week of HFD feeding reversed the action of HFD on the expression of Sirt1, FoxO1, and Htr2b, the acetylation of FoxO1, the level of SIRT1 and PKG, and the phosphorylation of AMPKα (Figure 6H–P). Altogether, diet supplementation of luteolin before fatty liver formation improves hepatic steatosis in obese mice by inhibiting visceral adipose tissue lipolysis and associated Ca²⁺-PKG cascade, and SIRT1/FoxO1/AMPKα signaling.

3 Discussion

It has been reported that diet supplementation of luteolin at start time of HFD feeding could protect the mice against diet-induced obesity and related complications, such as hepatic steatosis and insulin resistance.^{[14, 19, 20]} However, diet supplementation of luteolin at the 10th week of HFD feeding does not affect mouse body weight gain but ameliorates obesity-related insulin resistance.^[20] Thus, the improved action of luteolin on diet-induced obesity and associated metabolic diseases may greatly depend on the start time and duration of dietary luteolin intervention. The study reveals that diet supplementation of luteolin before fatty liver formation improves hepatic steatosis in HFD-fed mice by inhibiting visceral adipose tissue lipolysis, suggesting a reasonable therapeutic and interventional course of dietary luteolin for HFD-induced hepatic steatosis.

Luteolin mainly exists in the glycoside form in various edible plants and is primarily metabolized and absorbed in the intestinal tract.^[21] When rats are orally administrated with luteolin-rich Chrysanthemum extract (approximately equivalent to 200 mg luteolin per kg body weight), plasma luteolin level reaches to 4 µg mL⁻¹ (13.9 µM).^[22] Serum luteolin level is higher than 14.1 µM after the gavage of 5 µmol kg⁻¹ (approximately equivalent 14.3 mg kg⁻¹ body weight) luteolin.^[22] At the similar gavage dosage (14.7 mg kg⁻¹ body weight), the maximum plasma concentration of luteolin is 100 µM.^{[22, 23]} Therefore, in this study, 0.01% luteolin (approximately equivalent to 10 mg kg⁻¹ body weight) used as the dosage of dietary intervention not only refers to human recommended daily intake of luteolin (2–125 mg day⁻¹),^[24] but also is comparable with the dose of the previous animal studies.^{[22, 23, 25]}

Physiological utilization of luteolin is dependent on the degradation and transformation of intestinal microorganisms.^[26] However, HFD feeding for more than 4 weeks significantly changes mouse intestinal microbial population,^[27] and the post-dieting dysbiosis leads to the degradation of dietary flavonoids, which in turn interferes with host energy expenditure.^[28] Supporting the notions, our previous study^[24] and this study show that diet luteolin supplementation after the 6–10 weeks of HFD feeding does not affect body weight gain of HFD-fed mice (Figure 4B). Intriguingly, the improvement of dietary luteolin on obesity-associated insulin sensitivity^[20] and hepatic steatosis (Figures 4 and 5) does not depend on the change of mouse body weight in HFD-fed mice. The observations suggest that some microbiome-degraded products of luteolin may possess the substantial activities to combat insulin resistance and hepatic steatosis, a hypothesis that needs further investigation.

Adipose tissue gradually loses its function in the progress of obesity. In the early stage of HFD-induced obesity, basal adipocyte lipolysis is reduced. However, in the late stage of HFD-induced obesity, basal lipolysis is significantly elevated in white adipose tissue and finally leads to the ectopic deposition of lipids.^{[11, 29]} As the lipolysis regulator and energy sensor, the activation of SIRT1/FoxO1/AMPKα signaling is also consistent with the lipolysis level in EAT. In our prior study, 12 weeks of HFD feeding prominently depresses EAT lipolysis and SIRT1/FoxO1/AMPKα signaling, and dietary luteolin reverses the action of HFD feeding.^[30] In this study, EAT lipolysis and SIRT1/FoxO1/AMPKα signaling are remarkably induced by 16 weeks of HFD feeding and are inhibited by dietary luteolin. These results support the regulated synchronism between the level of visceral adipocyte lipolysis and the activation of SIRT1/FoxO1/AMPKα signaling.

Although we demonstrate the regulation of luteolin on 5-HT-induced 3T3-L1 lipolysis and HFD-induced hepatic steatosis, this study possesses some potential limitations. Firstly, besides HTR2B signaling, several classical stimulations, such as insulin^{[31, 32]} and cold^{[33, 34]} also induce adipose tissue lipolysis. As a multi-functional compound, luteolin may also affect adipocyte lipolysis and the development of fatty liver through regulating other signaling pathways in response to the stimulations, which needs future research. Additionally, although the improvements of dietary luteolin on hepatic steatosis in HFD+Lu6 and HFD+Lu8 groups are more significant than those in HFD+Lu10 group, the durations of luteolin treatments in HFD+Lu6 and HFD+Lu8 groups are also longer than those in HFD+Lu10 group. Since the levels of some EAT lipolytic molecules, including the phosphorylated levels of HSL and AMPK and the protein expression of PKG1, are altered in the mice from HFD+Lu10 group (Figure 6), we cannot exclude the presumption that the mice may have the improvement of fatty liver when the duration of diet luteolin supplementation is prolonged. However, our study shows that the earlier supplementation of diet luteolin is more beneficial to protect HFD-fed mice against hepatic steatosis.

In conclusion, diet luteolin supplementation before the formation of fatty liver alleviates HFD-induced hepatic steatosis by inhibiting EAT lipolysis, suggesting a reasonable therapeutic and interventional course for obesity-related fatty liver.

4 Experimental Section

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

J.W. designed and performed experiments, analyzed data, generated the figure panels, and wrote the draft manuscript. Z.Z. helped to perform the experiments. Y.L., J.C., F.W., and L.W. helped to design experiments. J.L. and X.Z. led the project, designed the experiments, and revised the manuscript.