Ananas comosus L. Leaf Phenols and p-Coumaric Acid Regulate Liver Fat Metabolism by Upregulating CPT-1 Expression

In this study, we aimed to investigate the effect and action mechanisms of pineapple leaf phenols (PLPs) on liver fat metabolism in high-fat diet-fed mice. Results show that PLP significantly reduced abdominal fat and liver lipid accumulation in high-fat diet-fed mice. The effects of PLP were comparable with those of FB. Furthermore, at the protein level, PLP upregulated the expression of carnitine palmitoyltransferase 1 (CPT-1), whereas FB had no effects on CPT-1 compared with the HFD controls. Regarding mRNA expression, PLP mainly promoted the expression of CPT-1, PGC1a, UCP-1, and AMPK in the mitochondria, whereas FB mostly enhanced the expression of Ech1, Acox1, Acaa1, and Ehhadh in peroxisomes. PLP seemed to enhance fat metabolism in the mitochondria, whereas FB mainly exerted the effect in peroxisomes. In addition, p-coumaric acid (CA), one of the main components from PLP, significantly inhibited fat accumulation in oleic acid-induced HepG2 cells. CA also significantly upregulated CPT-1 mRNA and protein expressions in HepG2 cells. We, firstly, found that PLP enhanced liver fat metabolism by upregulating CPT-1 expression in the mitochondria and might be promising in treatment of fatty liver diseases as alternative natural products. CA may be one of the active components of PLP.


Introduction
The prevalence of nonalcoholic fatty liver diseases (NAFLDs) is about 20% in mainland China and even higher in developed countries or areas [1]. NAFLDs exhibit adverse levels of liver fibrosis and cardiometabolic risk factors [2]. Besides genetic factors, environmental factors, such as high-fat diets, have an important function in the development of NAFLDs [3]. NAFLDs are considered to be diseases of affluence [4].
The management of NAFLDs still faces a great challenge [5]. Fibrates exert lipid-lowering effects on the blood and liver by targeting peroxisomal proliferator-activated receptor alpha (PPAR ) and promoting fat metabolism in the liver [6]. However, the effects of fibrates on liver histology are minimal [7]. Statins lower cholesterol synthesis in the liver by inhibiting HMGCoA reductase, and their uses in patients with hyperlipidemia and NAFLD are justified; however, neither of the trials reported possible histological changes in NAFLD when subjected to statin therapy [8]. Current drugs used for the treatment of NAFLDs require further improvement.
Pineapple (Ananas comosus L.) is grown largely in Hawaii, the Philippines, Caribbean area, Malaysia, Taiwan, Thailand, Australia, Mexico, Kenya, South Africa, and southern China. Pineapple has various agricultural utilities, such as fruits for nutrition, and some folk medicinal uses have been found. Pineapple leaves can help in digestion [9]. Pineapple leaf phenols (PLPs) demonstrated hypolipidemic effects in previous studies [10,11]. PLPs have different hypolipidemic mechanisms from fenofibrate (FB). However, whether PLPs have an effect on NAFLDs remains unclear. Moreover, the molecular mechanisms of PLP are unknown.
Carnitine palmitoyltransferase 1 (CPT-1) is a mitochondrial membrane protein that converts long-chain fatty 2 Evidence-Based Complementary and Alternative Medicine acyl-CoA molecules to their corresponding acylcarnitine molecules [12]. CPT-1 deficiency causes a disorder of longchain fatty acid oxidation [13]. The upregulation of CPT-1 is associated with the inhibition of fatty liver formation [14]. Some plant phenols have been reported to increase CPT-1 expression and attenuate hepatic steatosis [15].
In this study, we investigated the effect of PLP on the formation of fatty livers in high-fat diet-fed mice and found that PLP inhibited liver fat accumulation by regulating CPT-1 expression.

Animals and Diets.
Three-week-old male NIH mice were purchased from Guangdong Medical Animal Center (Guangzhou, China). The animals were housed in an environmentally controlled animal room (temperature: 20 ∘ C ± 2 ∘ C; humidity: 60% ± 5%; and a 12 h dark/light cycle). Animals were fed chow diets and water ad libitum. The study was strictly carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of Institutional Animal Care and Use Committee of Tsinghua University. The protocol was approved by the Animal Welfare and Ethics Committee of Tsinghua University, China (2013-XWD-BC). Both normal chow and high-fat diets were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. Normal chow diets contained 12% calories in fat while highfat diets contained 38% calories in fat.

Drugs.
Pineapple leaves were collected from Boao, Hainan, China. PLPs (batch number 051201) were prepared as previously prescribed [11]. Phytochemical assays showed that PLP mainly contained total phenols (more than 60%, w/w, in terms of the extract). p-coumaric acid (CA) was a compound separated from PLPs. PLP contained 1.5% CA (w/w). HPLC figures of PLP can be tracked from the previous study [11]. FB was purchased from Sigma-Aldrich. Drugs were freshly prepared before administration.

Experimental Procedure.
Male NIH mice were divided into four groups, namely, normal control mice (normal), high-fat diet-fed control mice (HFD), and PLP-and FBtreated high-fat diet-fed groups (PLP and FB, resp.). First, mice were fed high-fat diets (HFD, PLP, and FB groups) or normal chow diets (normal group) for four weeks. PLP and FB were orally administered once daily at 300 and 200 mg/kg, respectively. The dosages of PLP and FB were selected according to a previous report (200-800 mg/kg) [11] with a slight modification. The drugs were suspended with distilled water. Normal and HFD control mice were treated with identical volumes of distilled water. Diet intake was periodically measured in metabolic cages within 24 h. After four weeks of administration, the mice were weighed. Livers and abdominal adipose tissues were removed and weighed. Parts of the liver tissues were stored at −80 ∘ C or soaked in 10% formalin solution for further biochemical or histopathology assays, respectively.

Biochemical
Assays. Liver fat assays were performed according to a previous protocol with slight modification [16]. In brief, five mice from each group were randomly selected. About 35 ± 5 mg of liver from each mouse was weighed precisely and homogenized with 0.2 mL of PBS plus 1 mL of chloroform-methanol mixture solution (CHCl 3 -CH 3 OH; 2 : 1, v/v) in an Eppendorf tube. The extraction solution was centrifuged at 10,000 ×g for 5 min. A 20 (for triglyceride assay) and 40 L (for cholesterol assay) aliquot of the lower phase were added to another Eppendorf tube and air-dried at 37 ∘ C, respectively. The dried lipids at the bottom of the tubes were used to assay triglycerides and cholesterol levels in livers. Triglycerides and cholesterols in livers were assayed by using triglyceride and cholesterol reagent kits (BioSino Biotechnology and Science Inc., Beijing, China) according to the GPO-PAP and CHOD-PAP calorimetric methods, respectively. Briefly, the dried lipids in the bottom of the tubes were incubated with 200 L of triglyceride or cholesterol reagents at 37 ∘ C for 30-60 min. Then, the reaction liquids were transferred into a 96-well plate and measured at 500 nm through a Microplate Reader (Thermo Scientific Varioskan Flash). The lipid contents were normalized with liver weight, and data were expressed as lipids/liver weight (mg/g).

Histopathology.
Five mice from each group were randomly selected. A liver tissue mass (2 mm diameter) from each mouse was fixed with 10% formalin and processed for routine paraffin-wax histology. Sections were stained with hematoxylin and eosin (H&E). Five continuous slices (thickness of 10 m; gap of 20 m) were obtained to measure the formation of fat droplets in the liver of each mouse. We transferred the H&E pictures into black and white ones. Fat droplets formed in liver sections could be transferred into white, bright areas while other liver tissue areas' background became black by adjusting the contrast and brightness appropriately. These grey density values of white, bright areas from the liver fat droplets can be calculated by Gel-Pro software (Media Cybernetics, USA) and subjected to further statistical analysis.

mRNA Analysis by Quantitative Real-Time PCR (qPCR).
Five mice from each group were randomly selected. Total RNA of liver tissues was extracted using RNAiso plus reagent (Takara, Dalian, China) according to the manufacturer's instructions. Reverse transcriptase (RT) was performed using a kit containing reverse transcriptase M-MLV (RNase H − , code number D2639A, Takara, Dalian, China) according to the manufacturer's instructions. RT was conducted using an Alpha Unit Block Assembly for DNA Engine systems (Bio-Rad, USA) with a thermocycler program consisting of cDNA synthesis at 42 ∘ C for 60 min, RNase inactivation at 70 ∘ C for 15 min, and sample cooling at 4 ∘ C for 10 min. Primers were synthesized from Invitrogen (Table 1). Actin was used as an internal control for normalization. qPCR analysis was conducted using SYBR Green analysis was performed in two steps. First, cDNA samples were predenatured at 95 ∘ C for 60 s. Second, denatured cDNA samples were amplified with 40 cycles at 95 ∘ C for 15 s, 60 ∘ C for 15 s, and 72 ∘ C for 45 s. Data were analyzed by the raw relative quantitation method (2 −ddCt ).

Western Blot.
Five mice from each group were randomly selected. Liver tissues were homogenized and lysed with NETN buffer, and lysates were centrifuged at 10,000 ×g at 4 ∘ C for 10 min. Supernatants were collected, and the protein concentration was determined using a BCA assay kit (Nanjing Jiancheng Biotech, China). Western blot analysis was carried out according to the manufacturer's protocol. Goat polyclonal antibody against CPT-1A of human origin (also recommended for the detection of mouse origin, Santa Cruz; 1 : 500-1000 dilution), rabbit polyclonal antibody against enoyl coenzyme A hydratase 1, peroxisomal (Ech1) of human origin (also recommended for the detection of mouse origin, Beijing Aviva Systems Biology, China; 1 : 500-1000 dilution), and mouse monoclonal antibody against betaactin of chicken origin (also recommended for the detection of mouse origin, Santa Cruz; 1 : 500-1000 dilution) were used. Protein expression was visualized with secondary antibodies (donkey versus goat, goat versus rabbit, and rabbit versus mouse IgG-HRP; Amersham Biosciences, USA; 1 : 2000) and enhanced chemiluminescence (KPL, USA). Grey density values of Ech1 and CPT-1A protein expression were normalized with beta-actin (reference protein).  For RT-PCR, human CPT-1A (forward: ATCAATCGG-ACTCTGGAAACGG, reverse: TCAGGGAGTAGCGCA-TGGT), CPT-1B (forward: GCGCCCCTTGTTGGATGAT, reverse: CCACCATGACTTGAGCACCAG), and actin (forward: CATGTACGTTGCTATCCAGGC, reverse: CTCCT-TAATGTCACGCACGAT) primers were synthesized from Invitrogen. For Western blot, we assayed the expressions of CPT-1A and actin. Actin was served as internal control for normalization. RT-PCR and Western blot were conducted as described above, respectively.

Statistical
Analysis. Data are expressed as mean ± SD, and mean comparisons in groups were performed using one-way ANOVA. Tukey-Kramer comparisons were used to determine the source of significant differences. Differences with < 0.05 were considered to be statistically significant.

Body Weight Index and Diet
Intake. The high-fat diet-fed mice showed no significant increase in body gain compared with the normal controls (Figure 1(a)). Both PLP and FB did not show significant effects on the high-fat diet-fed mice compared with the HFD controls. However, HFD mice showed higher abdominal fat accumulation than normal controls; increased fat accumulation in HFD mice was attenuated by both PLP and FB (Figure 1(b)). In addition, HFD mice showed a slight increase in liver weight than normal controls; PLP did not affect liver weight, but FB increased liver weight significantly (Figure 1(c)). HFD mice demonstrated a slight increase in dietary calorie intake compared with normal controls. PLP seemed to have no significant effect on calorie intake in HFD mice, but FB showed a slight increase in calorie intake in HFD mice compared with that in HFD controls (Figure 1(d)).

Fat Contents in the Livers.
Biochemical assays showed that the fat contents in livers of mice from the HFD group significantly increased, but this increase was attenuated by both PLP and FB. FB seemed to have a stronger effect than PLP ( Figure 2). However, FB significantly increased the liver weights, as reported previously.

Liver Tissue Sections.
According to the liver histochemical sections stained with H&E, high-fat diet-fed mice showed a significant accumulation of fat droplets in the liver ( Figure 3). However, PLP and FB significantly attenuated the formation of fat droplets in the liver. FB was more effective than PLP in inhibiting fat accumulation.

Protein Expression Determined by Western Blot. Ech1
was downregulated in high-fat diet-fed mice. However, PLP slightly upregulated Ech1 expression, whereas FB largely upregulated Ech1 expression ( Figure 5). CPT-1A is one main subtype of CPT1 expressed in livers. Here, CPT-1A was also significantly downregulated in high-fat diet-fed mice. PLP significantly upregulated the expression of CPT-1A, whereas FB had no significant effect on CPT-1A expression. Thus, PLP may attenuate the formation of fatty livers by increasing CPT-1A expression. However, the compounds contributing to this effect remain unclear.

Effects of PLP and CA in Oleic Acid-Induced HepG2
Cells. AC is one of the main components of PLPs [17] and demonstrates good pharmacokinetic behavior after oral administration [18]. In this study, we selected PLP and CA for further validation in vitro. Incubation of 0.6 M oleic acid for 24 h significantly increased fat accumulation in HepG2 cells ( Figure 6). However, PLP (final concentrations of 10 and 50 g/mL, resp.) and CA (final concentrations of 1 and 10 g/mL, resp.) significantly attenuated fat accumulation in HepG2 cells. FB (final concentration of more than 10 g/mL) showed significant cell toxicity responses in HepG2 cells. Thus, a low concentration (1.4 g/mL) was used for further investigation. However, FB at 1.4 g/mL did not show a significant effect (data not shown). Nevertheless, these results show that CA could be an active component of PLP.

Effects of PLP and CA on CPT-1 Expression in HepG2
Cells. The immunofluorescence assay showed that PLP (final concentrations of 10 and 50 g/mL, resp.), CA (final concentrations of 1 and 10 g/mL, resp.), and FB (final concentration of 1.4 g/mL) significantly upregulated CPT-1A expression in HepG2 cells (Figure 7). Furthermore, RT-PCR and Western blot assays were used to validate the effects of PLP and CA on CPT-1A expression as shown in the immunofluorescence assay. The results showed that both PLP and CA significantly upregulated CPT-1A mRNA and protein expressions in either oleic acid-treated or oleic acid-untreated HepG2 cells (Figure 8). For mRNA expressions, although the absolute amount of CPT-1B mRNA expression was less than that of CPT-1A (data were not shown), the effects of PLP and CA on CPT-1B and CPT-1A mRNAs were similar. The results of PLP were consistent with the results in vivo. However, the effect of FB on CPT-1 expression in vitro differed from that in vivo. This difference might be due to the fact that the different action times of FB had variable effects on CPT-1 expression (the trials in vitro were within 24 h, whereas the trials in

Discussion
In this study, high-fat diet-fed mice showed a significant increase in abdominal fat and liver lipid accumulation, which was consistent with our previous study [19]. Both PLP and FB significantly attenuated the increase in abdominal fat and liver lipid accumulation in high-fat diet-fed mice. High-fat diet-fed mice had higher calorie intake than normal controls. The increase in calorie intake mainly contributed to the increase in fat accumulation or storage in tissues instead of body weight gain. Despite this effect, the hypolipidemic effects of PLP seemed to have no relationship with dietary calorie intake. FB slightly increased dietary calorie intake but did not contribute to a significant increase in body gain or increase in fat accumulation. Thus, enhanced fat oxidation metabolism may contribute to the effects of PLP.
The liver is one of the main organs for fat oxidation metabolism. The mitochondria and peroxisomes are the main cell organelles responsible for fat oxidation metabolism [20]. Ech1 is a key enzyme that dominates fat oxidation metabolism in peroxisomes [21]. The downregulation of Ech1 contributes to high-fat diet-induced hepatic steatosis [22]. In this study, PLP slightly upregulated the protein expression of Ech1 in HFD mice, but this effect was much weaker than that of FB. By contrast, FB significantly upregulated the mRNA and protein expressions of Ech1. Acox1, Acaa1, and Ehhadh are important genes responsible for fat oxidation in peroxisomes [23]. Higher expression levels of these genes may indicate more fatty acid oxidation metabolism in peroxisomes. In this study, FB promoted the expression of these peroxisome genes, but PLP had no significant effects. These results suggest that peroxisomes could be targeted cell organelles for FB but not for PLP.
CPT-1 is a key rate-limiting enzyme in charge of transporting long-chain fatty acids into mitochondria. The overexpression of CPT-1 in skeletal muscles in vivo increases fatty acid oxidation and reduces triacylglycerol esterification [24]. Moreover, the stimulation of systemic CPT-1 activity may be an attractive means to accelerate peripheral fatty acid oxidation [25]. CPT-1A is mainly expressed in livers. CPT-1B is mainly expressed in muscles and also expressed in livers. However, CPT-1B expressed in livers is far less than CPT-1A. Here, PLP significantly promoted the mRNA and protein expression of CPT-1. However, four weeks of FB administration seemed to have no significant effect on the expression of CPT-1 in vivo. PLP attenuated fat accumulation in the liver, which could be related to the upregulation of CPT-1 expression. Furthermore, PLP induced a greater increase in AMPK and UCP-1 expression than FB. AMPK and UCP-1 are key factors responsible for mitochondrial energy metabolism [26,27]. These results indicate that PLP acted mainly through targeting mitochondrial functions.
PPAR is a transcription factor and major regulator of lipid metabolism in the liver. The activation of PPAR promotes uptake, utilization, and catabolism of fatty acids by upregulating genes involved in peroxisomal and mitochondrial fatty acid -oxidation [28]. PGC-1 is a regulator of mitochondrial biogenesis and function [29]. The activation of PPAR promotes the expression of Ech1 [30], which was consistent with our results. Both PPAR and PGC1 stimulated the transcription of the CPT-1A gene [31]. PLP significantly promoted the expression of CPT-1 by upregulating the expression of PPAR and PGC1 . However, FB seemed to have no effect on the protein expression of CPT-1 in the livers of mice even though FB activated PPAR . In animal trials, FB activated PPAR for a long period (about four weeks), which possibly triggered the negative regulation of PPAR expression because of the homeostatic responses in vivo. These responses may explain why FB decreased the expression of PPAR at a late stage; this decrease could directly affect the regulation of CPT-1 expression. In livers of mice after acute administration of tetradecylglycidic acid, a PPAR ligand, CPT1A, and CPT1B were significantly upregulated [32]. It seemed that the expressions of CPT1A and CPT1B might vary with the time of PPAR activation or might not be completely mediated by PPAR pathway.
CA is one of the main components of PLP [17]. In this study, both PLP and CA showed a significant effect on the attenuation of fat accumulation in oleic acid-induced HepG2 cells. Furthermore, both PLP and CA upregulated the expression of CPT-1 in HepG2 cells. The results of PLP in vitro were consistent with those of the animal study in vivo. In our previous study, CA demonstrated preferable pharmacokinetic parameters after oral administration in mice [18]. Thus, CA could be considered as a marker component in PLP. However, further investigations in vivo should be conducted.
Beside impaired fatty acid beta oxidation, liver fat accumulation can be traced by the increased incidence of de novo lipogenesis [33]. In this study, we mainly investigated the effect of PLP and CA on hepatic fat oxidation metabolism. However, the effect of PLP and CA on hepatic lipogenesis might require further investigation in the future. In addition, PLP and CA have strong antioxidant properties against the production of oxidative stress in vitro in our preliminary trials (data were not shown). It remains interesting to investigate whether the effects of PLP and CA are associated with the inhibition of prooxidative state induced by HFD in vivo and likely by oleic acid treatment in vitro in the future since NAFLDs were associated with increased oxidative stress [34].

Conclusion
Taken together, we firstly found that PLP inhibited the formation of fatty livers and enhanced fat oxidation metabolism in the mitochondria of livers possibly by targeting CPT-1. By contrast, FB mainly targeted the related genes in peroxisomes. CA could be one of the main active components of PLP. PLP likely served as a natural component with novel hypolipidemic mechanisms and may be a promising natural alternative to improve NAFLDs.