Ferulic Acid Alleviates Lipotoxicity-induced Hepatocellular Death Through SIRT1 Activating-regulated Autophagy Pathway, and Independently of AMPK and Akt in AML-12 Hepatocytes

Background: Lipotoxicity-induced cell death plays a detrimental role in the pathogenesis of metabolic diseases. Ferulic acid, widespread in vegetative food, is conferred as a radical scavenger with multiple bioactivities. However, the benets of Ferulic acid against hepatic lipotoxicity are largely unclear. Here, we investigated the protective role of ferulic acid on palmitate-induced lipotoxicity and clarify its potential mechanisms in AML-12 hepatocytes. Methods: AML-12 mouse hepatocyte was employed and exposed to palmitate to mimic lipotoxicity. Different doses (25, 50, and 100 μM) of ferulic acid were added 2 h before palmitate induction. Cell viability was detected by the measurements of lactate dehydrogenase release, nuclear staining, and the expression of cleaved-caspase3. Intracellular reactive oxygen species and mitochondrial membrane potential were analyzed by uorescent probe. The potential mechanisms were explored by molecular biological methods, including Western-blot and quantitative real-time PCR, and further veried by siRNA interference. Results: Our data showed that ferulic acid signicantly reversed palmitate-induced cell death, rescued mitochondrial membrane potential, reduced reactive oxidative species accumulation, and improved inammatory factors activation, including IL-6 and IL-1beta. Ferulic acid signicantly stimulated autophagy in hepatocytes, while, autophagy suppression blocked ferulic acid-protected lipotoxicity. Ferulic acid-activated autophagy, which was triggered by Sirt1 upregulation, was mechanistically involved in its anti-lipotoxicity benets. Sirt1 silencing blocked ferulic acid-induced protable alterations. Conclusions: We demonstrated that ferulic acid was a protective phytochemical against hepatic lipotoxicity in plant-based food, through Sirt1/autophagy pathway. Increasing ferulic acid-enriched food intake is a potential strategy to prevent and/or improve metabolic diseases with lipotoxicity as a typical pathological feature.


Introduction
Metabolic diseases, including non-alcoholic fatty liver disease (NAFLD), diabetes mellitus, and obesity, are worldwide epidemic and commonly featured with hyperlipidemia as a hallmark. Under normal physiological condition, adipose tissues have a high ability to store excessive fat, and maintain lipolysis homeostasis; however, in the pathological state of insulin resistance, enhanced lipolysis of adipose tissue leads to elevated free fatty acids (FFAs) heterotopic deposition in non-adipose tissues, such as liver, skeletal muscle, and pancreas, resulting in cell dysfunction and even cell death, which is called lipotoxicity or lipoapoptosis [1]. As the central organ of metabolism, hepatic cell death induced by lipotoxicity accelerates the occurrence and development of metabolic diseases. We previously reported that alleviating hepatic lipotoxicity effectively improved high-fat diet-induced NAFLD in mice [2]. FFAs are chemically divided into saturated fatty acids (SFAs) and unsaturated fatty acids (USFAs). Among which, lipotoxicity-caused cell death is commonly induced by SFAs via stimulating oxidative stress and endoplasmic reticulum stress [3,4], whereas USFAs tend to improve SFA-induced lipotoxicity [5].
Palmitate acid (PA, C16:0) is the most abundant natural SFA in food and human body and is widely used as lipotoxicity inducer for scienti c investigations [6,7].
Autophagy is a highly conserved physiological process by which intracellular components can be degraded and removed for quality and nutritional purposes. People with lipids metabolic disorder of liver are usually suffered from impaired autophagy, such as NAFLD patients [2]. Recent studies including ours have con rmed that activating autophagy is an effective way to improve lipotoxicity-induced hepatocellular injury in both cultured cells and animal models [2,8]. Although the penetrating mechanisms behind autophagy-regulated lipotoxicity are not fully illustrated, autophagy activation helped to eliminate damaged organelles, such as mitochondria and endoplasmic reticulum, which were termed as mitophagy and reticulophagy, respectively, was involved in its lipotoxicity protective role via improving oxidative stress and endoplasmic reticulum stress [9,10]. Additionally, the activation of autophagy degraded intracellular accumulation of triglycerides, which was termed as lipophagy [11], and in turn alleviating obesity-associated metabolic disorder in liver [12].
Several mechanisms have been identi ed implicating in the regulation of autophagy. Among which, adenosine monophosphate-activated protein kinase (AMPK) activation acts as a positive regulator of autophagy via inhibiting mammalian target of rapamycin (mTOR) in the state of intracellular energy de ciency [13]. Sirtuin 1 (SIRT1), a homologue of mammalian silencing information regulator 2, is a NAD + dependent deacetylase that regulates protein activity via modifying the acetylation of molecules, transcriptional factors and enzymes. It is known that SIRT1 participated in the regulation of autophagy via affecting autophagy-related genes 3 (Atg3), Atg7 and microtubule-associated protein 1 (MAP1) light chain 3 (LC3) deacetylation. LC3, an initiator of autophagosomes, can be deacetylated by SIRT1 in nucleus, by which LC3 will be selectively activated and allowed to engage in autophagy [14,15]. We previously reported that SIRT1 activation protected lipotoxicity-induced hepatic cell death via stimulating autophagy [16]. Akt (known as protein kinase B, PKB), a key regulator for cellular survival, has also been con rmed participating in the regulation of autophagy and lipotoxicity-induced hepatic cell death [17,18].
There are no safe and effective clinical drugs to treat metabolic diseases so far. Accumulating evidence supports that patients with abnormal liver metabolism enhanced their physical health indicators by adjusting dietary structure, like increasing the intakes of fruits, vegetables, and whole grains [19]. Ferulic acid (FA),, chemically known as 4-hydroxy-3-methoxy cinnamic acid, is mainly cross-linked with cytoderm polysaccharides and lignin to form part of cytoderm in plants. FA is commonly spread in the seeds of whole grains (bran, rice, wheat, etc.), vegetables (tomato, celery, spinach, etc.), and fruits (pineapple, grape, blackberry, etc.), with the highest content in whole grains (up to 1000 mg/kg in rye) [20][21][22]. An accurate nutrition survey about FA intake is lacking. Consumption of food source FA can be estimated in daily intake of 150-200 mg [23]. Several biological functions of FA have been reported, such as antidiabetes, anti-oxidation, anti-in ammation, anti-cancer, and lowering blood lipid. FA is more easily absorbed by the body than other phenolic acids and stays in the blood for longer [24]. FA acts as antioxidant due to its scavenging on radicals, instead of the formation of phenoxyl radicals, as well as inhibition of reactive oxygen species (ROS) generation through donation of a hydrogen atom. Previous studies showed that FA improved thioacetamide-and CCl4-induced hepatic brosis [25,26] and diosbulbin B-, cadmium-, and streptozotocin-induced liver damage [27][28][29]. FA supplementation signi cantly improved hepatic lipids metabolic disorder and decreased liver injury in high-fat diet-induced obese mice [30,31]. However, limited study has been conducted to analyze the effect of FA on lipotoxicity-induced hepatic cell death, and the mechanisms are largely unclear.
The present study was designed to emphasize the in uence of FA on lipotoxicity-induced hepatocytes impairments, including apoptosis, mitochondrial function, oxidative stress, and its potential molecular mechanisms. PA was chosen to establish hepatic lipotoxicity model in vitro. We observed that FA incubation markedly ameliorated PA-induced apoptosis, LDH leakage, mitochondrial membrane potential (MMP) reduction, ROS generation, and in ammatory activation. FA triggered SIRT1 upregulation, which in turn activated autophagy, was mechanistically involved in its bene cial role against lipotoxicity. Hence, our study reveals that FA is a potential effective phytochemical compound resistant to hepatic lipotoxicity.

Chemicals
All of chemicals, including PA, bovine serum albumin (BSA), dimethylsulfoxide (DMSO), and choloroquine (CQ) were purchased from Sigma-Aldrich (St. Louis, MO). FA was provided by Chengdu Herbpurify Co., Ltd (Sichuan, China). PA-BSA conjugates were prepared as described previously [16]. All experiments contained a control group/vehicle, which was exposed to a same amount of solvent (e.g. BSA, DMSO).

Cell death assays
Cells were seeded into plate and allowed to grow to approximate 80% con uency, and then were incubated with indicated treatments. Cell viability was detected by MTT test, lactate dehydrogenase (LDH) release, and Hoechst staining. For MTT assay, MTT (Solarbio, Beijing, China) was added to each well to nal 5 mg/mL and maintained at room temperature for 4 hours. Then DMSO was added and incubated on a plate shaker for 10 min. The absorption was detected at 470 nm using microplate reader (Dynatech, El Paso, Texas, MR-4100). For LDH assay, the medium was collected for LDH analyzing using LDH kit (Pierce, Rockford, lL) according to the manufacturer's instructions. The absorption values were measured at 340 nm using microplate reader (Dynatech, El Paso, Texas, MR-4100). Cell viability was also assessed by Hoechst 33342 (Sigma-Aldrich, St. Louis, MO) staining. The plate was washed twice with chilled phosphate buffered saline (PBS) after staining for 30 min. The nuclear morphological changes were examined by uorescence microscope (Nikon, Tokyo, Japan, TE2000-U).
DCFH-DA (10 μM as nal concentration) was added to each well and stained at room temperature for 30 min. Chilled PBS were used to wash cells for three times. The uorescence intensity was measured with inverted uorescent microscope (Nikon, Tokyo, Japan, TE2000-U). Image J 1.51 software was used to quantify the mean uorescence intensity (MFI) from ve random elds.

MMP assay
MMP was assessed via uorescent dye Rhodamine 123 (Rh123, Solarbio, Beijing, China) staining. Cells were stained with Rh123 (100 μg/mL as nal concentration) for 45 min at room temperature. Then the cells were washed by chilled PBS to remove excess dye. The uorescence intensity was measured with inverted uorescent microscope (Nikon, Tokyo, Japan, TE2000-U). Image J 1.51 software was used to analyze MFI from ve random elds.

RNA interference
Small interfering RNA (siRNA) for mouse SIRT1 was purchased from GenePharma Co., Ltd (Shanghai, China). SiRNA-Mate (GenePharma, Shanghai, China) was utilized to deliver siRNA to the targeted cells according to the manufacturer's protocol. Scrambled siRNA (GenePharma, Shanghai, China) was applied in negative control group. Silencing e ciency was veri ed by quantitative real-time PCR and western-blot analysis.

RNA extraction and quantitative real-time PCR
Intracellular total RNA was harvested by Trizol (Invitrogen, Carlsbad, CA). qRT-PCR was performed with StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA). The data were analyzed by 2 -(△△Ct) method. Rn18s was used as house-keeping gene for calibration. Primers' sequences were listed in Table 1.

Analysis of autophagic ux
The autophagic ux was measured as previously described [32]. In brief, chloroquine (CQ) was added prior to the FA treatment to inhibit lysosome acidi cation. The autophagic ux was determined by detecting GFP-LC3 puncta with laser scanning confocal microscope (Zeiss, Jena, German, LSM880) as well as LC3-II expression by western blot. For GFP-LC3 uorescence detection, cells were transiently infected with recombinant GFP-LC3 lentivirus (GeneChem, Shanghai, China).

Statistical analysis
All experimental data were from at least three independent experiments and shown as mean ± SD.
Statistical signi cance was compared by Student's t-test or one-way ANOVA followed by Tukey's multiple comparisons test. All tests were performed with SPSS 19.0. A value of p < 0.05 was considered signi cant.

Results
3.1 FA protects against hepatocyte cell death induced by PA.
The cytotoxicity of FA was initially determined on AML12 hepatocytes. There was no signi cant cytotoxicity observed at dose up to 800 µM (Fig. 1A). Combined with previous studies [8,33], an incremental dose of FA (0, 25, 50, and 100 µM) were selected to evaluate the protective effect of FA on palmitate-induced lipotoxicity. The result indicated that FA signi cantly reversed PA-induced hepatocyte cell death in a dose-dependent manner (Fig. 1B). FA exposure also prevented PA-induced caspase-3 cleavage, a typical indicator of apoptosis, with an optimal dose at 100 µM (Fig. 1C). In addition, PAinduced DNA condensation and nuclei fragmentation were also reversed by FA pretreatment (Fig. 1D). To test the generality of our nding, we conducted the similar experiments in human hepatocytes cell line (HepG2). Our results showed that FA incubation markedly reversed PA-induced cell death in HepG2 cells (Fig. S1). All these results suggested that FA exhibited a strong protective role against hepatic lipotoxicityinduced by PA.

FA improves lipotoxicity-induced mitochondrial dysfunction in hepatocytes.
In view of the strong antioxidant effect of FA, we subsequently tested protective ability of FA on mitochondrial, the main source of intracellular ROS. Our results showed that PA exposure signi cantly potentiated intracellular ROS levels, while FA pretreatment effectively reversed PA-triggered excessive ROS generation ( Fig. 2A). Meanwhile, FA pretreatment also dramatically recovered MMP suppressed by PA (Fig. 2B). Moreover, the alteration of apoptosis-related mitochondrial proteins, including Bcl-2 and Bax, was signi cantly reversed by FA pretreatment (Fig. 2C).

Autophagy activation contributes to FA-inhibited lipotoxicity.
We previously reported that activation of autophagy protects hepatocytes from PA-induced lipotoxicity [8]. In this study, we found that FA exposure signi cantly stimulated hepatic autophagy in AML12 cells, evidenced by the observation of increased Beclin1 and LC3 II conversion, the well-established markers of autophagy induction, whereas reduced p62, a protein speci cally degraded by autophagy, in a dosedependent and time-course manner (Fig. 3A & B). Moreover, FA incubation markedly stimulated autophagic ux, which was con rmed by increased LC3-II conversion and LC-3 puncta formation in the presence of chloroquine (Fig. 3C & D). Importantly, FA-protected lipotoxicity was robustly blocked by autophagy inhibition (Fig. 3E-G), which indicating autophagy activation participated in the bene cial role of FA against lipotoxicity.

SIRT1 upregulation participates in FA-induced autophagy.
Subsequently, we analyzed upstream regulators by which FA-stimulated autophagy. Our data showed that FA treatment did not increased expressions of phosphorylated-AMPK and -Akt (Fig. 4A). However, SIRT1 was upregulated by FA intervention in a dose-dependent and time-course manner ( Fig. 4A & B). PA blunted SIRT1 protein abundance was obviously reversed by FA incubation (Fig. 4C). Importantly, FA treatment failed to stimulated autophagy in SIRT1-silenced hepatocytes (Fig. 4D-F), indicating that SIRT1 was participated in FA-stimulated autophagy. We also observed that more LC3 II was detected in SIRT1knockingdown cells (Fig. 4F), which probably due to the fact that SIRT1 de ciency led to the impediment of autophagic ux and in hence blocked LC3 II degradation. To further certi ed that, we tested the expression of p62, which could be specially degraded by autophagy. In support of our hypothesis, p62 was obviously accumulated in Sirt1-knockingdown cells (Fig. S2). This phenomenon was in agree with a previous study [14].

Sirt1-regulated autophagy is mechanistically involved in FA-alleviated lipotoxicity.
We then analyzed the participation of Sirt1-mediated autophagy in FA-protected lipotoxicity in AML12 hepatocytes. The results showed that genetically knocking-down Sirt1 signi cantly blocked FA-protected lipotixicity-induced by PA treatment, evidenced by detections of LDH release, cleaved-caspase3 expression, and nuclear staining (Fig. 5A-C).

FA improves PA-induced proin ammatory cytokines activation in hepatocytes.
The anti-in ammatory role of FA was investigated in this study. Our results showed that PA exposure signi cantly transcriptionally stimulated pro-in ammatory factors, including IL-1beta and IL-6, while FA pretreatment markedly reversed PA-induced in ammatory reactions (Fig. 6A & B).

Discussion
In this study, we identi ed that FA, a safe phenolic acid, exerted strong anti-apoptosis role in PA-induced hepatic lipotoxicity. FA intervention signi cantly alleviated lipotoxicity-caused mitochondrial dysfunction and in ammation in AML-12 hepatocytes. Our data suggested that Sirt1-mediated autophagy signaling pathway contributed to the bene cial roles of FA mentioned above.
Lipotoxicity plays an essential pathological role in the development of several metabolic diseases [34]. In liver, excessive FFAs may initiate hepatocyte injury, in ammation, and even apoptosis, which in turn leads to liver dysfunction and promotes the occurrence of various metabolic diseases [35,36]. It is commonly recognized that the major detrimental role of lipotoxicity was not sourced from neutral triglyceride deposition, but originated from excessive free SFAs, among which, PA was the most selected lipotoxicityinducer [5,7,37]. Accumulated evidence has proved that strategy on alleviating hepatic lipotoxicity effectively improved metabolic diseases, such as NAFLD [38]. To date, no safe and e cient clinical drug targeting on lipotixicity prevention have been o cially approved. Plenty of epidemiological studies have reported that improving dietary habits, especially increasing the intake of plant foods, such as wholegrain food, was bene t for the prevention of hepatic metabolic disorder in NAFLD patients [39]. Therefore, phytochemicals extracted from plant foods or medical herbs provide a feasible alternative for the treatment of lipotixicity-related metabolic diseases. FA, which is widely existing in whole grains with a dose of 1000 mg/kg in rye [22], has recently been reported to improve high-fat diet-induced hepatic metabolic disorder in experimental animal [30,31,40]. Several studies have demonstrated that FA was a strong scavenger of excessive ROS, which induction was mechanistically involved in lipotoxicity-induced apoptosis [41,42]. However, limited study has conducted to investigate the protective role of FA on lipotoxicity-induced cell death in hepatocytes. In this study, we reported for the rst time that FA intervention signi cantly alleviated PA-induced apoptosis in hepatocytes, which was con rmed by the detections of LDH release, caspase-3 cleavage, and nuclear morphology.
We subsequently analyzed the potential mechanisms behind the protective effects of FA. Autophagy is a conservative and complex quality control pathway that plays a crucial part in eliminating damaged proteins and organelles [43]. Upon autophagy induction, LC3 (microprotein 1 light chain 3), a homologue of Atg8 in mammalian cells, controls the formation of autophagosomes and lysosome, as well as the degradation of substrates. LC3-I is regulated with phosphatidylethanolamine at the glycine residue, and becomes LC3-II, which is bound to both the outer and the inner membrane of the autophagosome [44]. Beclin1 (also termed as BECN1), a homologue of yeast Atg6, not only participates in the formation of autophagosomes, but also regulates autophagy activity [45]. SQSTM1 (sequestrome1, also known as p62) is a selective autophagy receptor, which transports ubiquitinated targets to autophagosomes [46]. It has been well-considered that autophagic ux was damaged in the liver of metabolic diseases, such as NAFLD [47][48][49]. The activation of autophagy could improve hepatic metabolic disorders by removing excessive lipids droplets from hepatocytes, and alleviate liver injury [50]. We previous reported that autophagy activation protected hepatocyte from SFAs-induced hepatic apoptosis [8]. Therefore, we hypothesized that autophagy activating contributed to the bene cial role of FA. In support of our vision, FA incubation markedly stimulated autophagy in hepatocytes, based on the observations of increased autophagic ux, along with enhanced Beclin1 expression and LC3-II conversion, and reduced p62 expression. In line with our observations, FA has also been reported to stimulate autophagy in other types of cells, including renal cells, brain microvascular endothelial cells, and myocardial cells [51][52][53].
Importantly, the inhibition of autophagy signi cantly blocked the protective role of FA against PA-induced apoptosis, which indicating autophagy induction was mechanistically involved in FA-alleviated lipotoxicity in hepatocytes.
Lipotoxicity-induced mitochondrial dysfunction, excessive ROS production, and even programmed apoptosis played a critical role in the pathological process in hepatic metabolic diseases [4]. Selective degradation of damaged mitochondrial by autophagy was also termed as mitophagy, which helped to maintain the integrity of the cell function. Recent evidence has provided that the induction of mitophagy prevented high-fat diet-induced liver injury [54,55]. Besides, PA-decreased mitophagy, leading to mitochondrial dysfunction as characterized by extensive mito-ROS production and loss of MMP [56]. Data from our study clearly revealed that FA treatment signi cantly abrogated PA-induced mitochondrial dysfunction, by the facts of improved MMP and intracellular ROS level. However, more direct evidence on mitophagy of FA-treated hepatocytes was still needed in the future studies.
Several mechanisms have been identi ed in the regulation of autophagy. Among which, Sirt1, an NAD +dependent deacetylase, regulated autophagy initiation via mediating LC3 deacetylation [57]. Sirt1 activation exerted positive effect in the regulation of liver lipids metabolism, oxidation and in ammation [58], whereas, Sirt1 depletion accelerated hepatic injury in the pathogenesis of NAFLD [58,59]. We recently reported the protective effect of Sirt1 induction on PA-induced hepatocellular death [16]. These evidences promoted us to hypothesize that Sirt1-regulated autophagy contributed to FA-protected lipotoxicity. This notion was indeed supported by the following evidences. Firstly, FA treatment obviously stimulated Sirt1 expression in a dose-dependent and time-course manner. Such performance was also observed in FA-treated skeletal muscle cells, bone, and testis [60][61][62]. Secondly, genetically knockingdown Sirt1 robustly abolished FA-stimulated autophagy induction. Last but importantly, FA-protected lipotoxicity was strongly blocked in Sirt1-silenced hepatocytes. Additionally, AMPK, a central sensor of intracellular energy, is a key regulator of autophagy via inhibiting the downstream targets mTOR complex 1, which is a negative regulator of autophagy. Several studies have reported the reciprocal regulatory relationship between Sirt1 and AMPK [63]. The activation of AMPK can signi cantly eliminate lipotoxicityinduced hepatocyte death [64]. FA was showed to active AMPK in skeletal muscle cells and cardiac myocytes [60,65]. We therefore analyzed the involvement of phosphorylated-AMPK in FA-treated hepatocytes. Beyond our expectation, FA incubation did not activate AMPK phosphorylation in AML12 hepatocytes, which excluded the participation of AMPK in FA-protected lipotoxicity. Phosphatidylinositol 3-kinase (PI3K)/Akt pathway, which activation promoted cell survival under lipotoxicity, was mechanistically involved in autophagy induction [17]. The reports on the regulation of FA on PI3K/Akt were inconsistent. It seemed like that PI3K/Akt was inhibited by FA in tumor cells [66] but stimulated in detrimental stimuli-induced cellular dysfunction [67]. In this study, we observed that Akt phosphorylation was not signi cantly regulated by FA treatment in AML12 hepatocytes, implying that PI3K/Akt pathway was not participated in FA-induced autophagy and further bene cial roles against lipotoxicity.
Pro-in ammatory factors-triggered cytotoxicity played a detrimental role in the development of hepatic metabolic disorders. Accumulated studies including ours have indicated that PA exposure transcriptionally stimulated expressions of pro-in ammatory factors, including IL-1beta and IL-6 [68][69][70]. In the present study, we observed that FA intervention markedly reversed PA-caused activation of proin ammatory factors in hepatocytes, which was in line with the facts that FA-inhibited pro-in ammatory reactions in other types of cells [71,72]. Therefore, we presumed that FA-blocked in ammation might be mechanistically participated in its lipotoxicity-alleviative role. In view of the crosstalk between oxidative stress and in ammation, we are still not sure whether FA inhibited lipotoxicity by directly acting on the in ammatory signaling pathway or by improving oxidative stress, which needs further investigations.

Conclusions
In summary, our study reported that via activating Sirt1/autophagy pathway, FA exposure in hepatocytes protects lipotoxicity-induced apoptosis. Our ndings provided a new mechanism that may help to understand the biological values of FA in hepatic metabolic diseases. We highlighted the potential value of FA as a dietary supplement in preventing and/or treating liver diseases with lipotoxicity as a typical pathological feature.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
Not applicable.

Competing interests
The authors declare no con ict of interest. Authors' contributions XD, SL and HC designed of the work; TX made contributions to the acquisition and analysis; QS drafted the manuscript; LZ, HN and QQ made contributions to the interpretation of data; QH and HP revised the manuscript. All authors read and approved the nal manuscript. Table   Table 1 List of primers.

Gene
Forward primer (