A Gedunin-Type Limonoid, 7-Deacetoxy-7-Oxogedunin, from Andiroba (Carapa guianensis Aublet) Reduced Intracellular Triglyceride Content and Enhanced Autophagy in HepG2 Cells

The seed oil of Carapa guianensis Aublet (Andiroba) has been used in folk medicine for its insect-repelling, anti-inflammatory, and anti-malarial activities. This study aimed to examine the triglyceride (TG) reducing effects of C. guianensis-derived limonoids or other commercially available limonoids in human hepatoblastoma HepG2 cells and evaluate the expression of lipid metabolism or autophagy-related proteins by treatment with 7-deacetoxy-7-oxogedunin (DAOG; 1), a principal limonoid of C. guianensis. The gedunin-type limonoids, such as DAOG (% of control at 20 μM: 70.9 ± 0.9%), gedunin (2, 74.0 ± 1.1%), epoxyazadiradione (4, 73.4 ± 2.0%), 17β-hydroxyazadiradione (5, 79.9 ± 0.6%), 7-deacetoxy-7α-hydroxygedunin (6, 61.0 ± 1.2%), andirolide H (7, 87.4 ± 2.2%), and 6α-hydroxygedunin (8, 84.5 ± 1.1%), were observed to reduce the TG content at lower concentrations than berberine chloride (BBR, a positive control, 84.1 ± 0.3% at 30 μM) in HepG2 cells pretreated with high glucose and oleic acid. Andirobin-, obacunol-, nimbin-, and salannin-type limonoids showed no effect on the intracellular TG content in HepG2 cells. The TG-reducing effect of DAOG was attenuated by the concomitant use of compound C (dorsomorphin), an AMPK inhibitor. Further investigation on the detailed mechanism of action of DAOG at non-cytotoxic concentrations revealed that the expressions of autophagy-related proteins, LC3 and p62, were upregulated by treatment with DAOG. These findings suggested that gedunin-type limonoids from Andiroba could ameliorate fatty liver, and that the action of DAOG in particular is mediated by autophagy.


Introduction
Highly oxygenated tetranortriterpenoids, such as limonoids, are structurally formed by the loss of four terminal carbons from the side chain in apotirucallane or apoeuphane skeleton, followed by cyclization to form the 17-furan ring [1][2][3]. Compounds belonging to this group exhibit a diverse range of biological functions, such as insecticidal, insect antifeedant, and growth-regulating activities in insects, as well as anti-bacterial, antifungal, anti-malarial, anti-cancer, anti-viral, and several other pharmacological activities in humans [1,2,4]. Limonoids are mainly found in the Meliaceae and Rutaceae families and less frequently in the Cneoraceae and Harrisonia sp. of Simaroubaceae family [1][2][3][4]. A Meliaceae plant, Carapa guianensis Aublet (local name Andiroba), is distributed in the tropical rainforests of countries, such as Brazil and Colombia. The seeds, flowers, and bark of this plant have been used as traditional medicines due to its insect-repelling, antiinflammatory, and anti-malarial properties. The Andiroba seed oil has been reported to exhibit highly efficient analgesic, anti-bacterial, anti-inflammatory, anti-cancerous, antitumor, anti-fungal, and anti-allergic properties. In addition, it was found to be effective against wounds, bruises, herpes ulcers, rheumatism, ear infections, and insect bites as a repellent [5]. During characterization studies on the biological activities of constituents from the seed and flower oils of C. guianensis, we reported their cytotoxic [6][7][8][9][10], antimalarial [11], anti-inflammatory [12][13][14][15], hepatoprotective [16], and collagen synthesispromoting activities [17].
Non-alcoholic fatty liver disease (NAFLD) is induced by changes in diet and other lifestyle habits and is characterized by excessive triglyceride (TG) accumulation in hepatocytes, which is often associated with obesity and diabetes. As NAFLD progresses, it develops into non-alcoholic steatohepatitis (NASH) accompanied by inflammation, necrosis, or fibrosis of hepatocytes, often leading to hepatocirrhosis or hepatocellular carcinoma [18,19]. Additionally, NAFLD is closely associated with several other diseases, including chronic kidney disease, cardiovascular disorder, thyroid disease, polycystic ovary syndrome (PCOS), and colon cancer [20][21][22][23][24][25]. The first line of treatment for NAFLD/NASH coexisting with obesity as defined by the ESPEN guideline is intensive lifestyle intervention leading to weight loss and exercise leading to hepatic fat reduction [26]. Reports documenting effective NAFLD/NASH treatment include pioglitazone, a sodium-glucose cotransporter 2 inhibitor, or glucagon-like peptide-1 receptor agonist for type 2 diabetes [27][28][29], statins for hypercholesterolemia [30], angiotensin II receptor blockers and angiotensin IIconverting enzyme inhibitors for hypertension [31,32], high-dose ursodeoxycholic acid [33], the anti-oxidant vitamin E, [34] and obeticholic acid [35], a nuclear receptor FXR agonist acting as bile acid. However, no specific medicine has been approved; therefore, the development of a NAFLD/NASH-specific medicine is essentially required [36].
We have been searching for constituents from natural products that can contribute to the prevention and improvement of NAFLD or any other lifestyle-related disease in vivo models using mice or in cell-based assay using HepG2 cells. Studies on hepatic triglyceride metabolism often use HepG2 cells and other cell lines derived from liver, as well as liver slices or primary hepatocytes [37]. We identified several flavonoids [38][39][40], megastigmanes [41], and diterpenes [42][43][44] with anti-fatty liver principles, which showed inhibitory effects on TG accumulation or accelerated TG metabolism in hepatocytes. In addition, sesquiterpene glycosides were obtained from artichoke (the leaves of Cynara scolymus L.) [45], saponins from tea flower buds of Camellia sinensis (L.) Kuntze [46,47], daisy flowers of Bellis perennis L. [48,49], pericarps of Sapindus rarak DC. [50], mate leaves of Ilex paraguariensis A. St.-Hil. [51], and oligostilbenes from the bark of Shorea roxburghii G. Don [52] showed anti-hyperlipidemic effects in olive oil-treated mice. In the previous report, we elucidated that limonoids isolated from Andiroba strongly reduced the intracellular TG content in high glucose-pretreated HepG2 cells [53]. This study examined the relationship between the different types of limonoids and their TG-reducing activity against accumulated TG in HepG2 cells concomitantly pretreated with high glucose and oleic acid. Furthermore, the mechanism of action of 7-deacetoxy-7-oxogedunin (DAOG; 1), a principal limonoid constituent of Andiroba, was investigated to study the involvement of autophagy in TG-reducing activity.

Effects of AMPK Inhibitor on Intracellular TG Content as Well as AMPK and ACC Phosphorylation
To investigate the AMPK pathway involvement in the intracellular TG-reducing effect of DAOG (1), the combined effect of a potent AMPK inhibitor, compound C [56], was confirmed. Compound 1 (% of control: 71.3 ± 1.6% at 20 μM) and BBR (82.6 ± 2.4% at 30 μM) showed a concentration-dependent reduction of TG accumulated intracellularly using high glucose and oleic acid concentrations, as indicated by the blank bars in Figure  4A. In contrast, the activity of DAOG (1) and BBR was significantly attenuated by the concomitant use of compound C ( Figure 4A, filled bars, 97.9 ± 0.9% at 20 μM of 1, 149.0 ± 2.6% at 30 μM of BBR), as fully described in Table S2. The attenuation of the TG-reducing effect by inhibiting AMPK with compound C was more effective in combination with BBR. Our results are consistent with previous studies, reporting a beneficial effect of BBR on NAFLD through the AMPK pathway [57,58].

Effects of AMPK Inhibitor on Intracellular TG Content as Well as AMPK and ACC Phosphorylation
To investigate the AMPK pathway involvement in the intracellular TG-reducing effect of DAOG (1), the combined effect of a potent AMPK inhibitor, compound C [56], was confirmed. Compound 1 (% of control: 71.3 ± 1.6% at 20 µM) and BBR (82.6 ± 2.4% at 30 µM) showed a concentration-dependent reduction of TG accumulated intracellularly using high glucose and oleic acid concentrations, as indicated by the blank bars in Figure 4A. In contrast, the activity of DAOG (1) and BBR was significantly attenuated by the concomitant use of compound C ( Figure 4A, filled bars, 97.9 ± 0.9% at 20 µM of 1, 149.0 ± 2.6% at 30 µM of BBR), as fully described in Table S2. The attenuation of the TG-reducing effect by inhibiting AMPK with compound C was more effective in combination with BBR. Our results are consistent with previous studies, reporting a beneficial effect of BBR on NAFLD through the AMPK pathway [57,58].
Next, the expression of phosphorylated AMPK (the active form of AMPK) was analyzed by immunoblotting. Compound 1 or BBR was added to HepG2 cells in which TG accumulation was induced using high glucose and oleic acid concentrations; post-24 h, the cells were collected, and the proteins were extracted for western blotting. Figure 4B,C indicate that BBR markedly increased the p-AMPK/AMPK ratio (3.21 ± 0.66-fold increase, Table S3). In addition, compound 1 also tended to increase the p-AMPK/AMPK ratio (1.43 ± 0.07-fold increase, Table S3), however, the change was insignificant, and relative to the control (only vehicle-treated). Acetyl-CoA carboxylase (ACC) is a key molecule in fatty acid metabolism and is negatively regulated by AMPK. ACC catalyzes the conversion of acetyl-CoA to malonyl-CoA, which inhibits carnitine palmitoyltransferase (CPT)-1. Because CPT-1 catalyzes the transfer of the fatty acid groups from acyl-CoA to carnitine and increases their transport from cytosol to mitochondria, inactivation of ACC by AMPK enhances fatty acid degradation via β-oxidation [59]. In our study, ACC phosphorylation was not clearly observed despite AMPK activation (0.86 ± 0.09-fold increase for 1, 0.96 ± 0.14-fold increase for BBR, Table S3). However, the p-ACC/ACC ratio significantly decreased due to treatment with compound C (Table S3).

Involvement of Autophagy in the Mechanism of Action of DAOG (1)
In Section 2.2, the lipid-reducing effect of DAOG (1) was implicated to be mediated by AMPK; however, the western blot analysis results suggested that the enhancement of β-oxidation is not very strongly induced through the AMPK/ACC pathway. Therefore, we investigated the effects of compound 1 and its involvement in lipophagy, the selective breakdown of droplet-stored lipids by autophagy. Time-dependent expression of several autophagy-related proteins, such as microtubule-associated protein 1 light chain 3 (MAP1LC3, also known as LC3), sequestosome 1 (SQSTM1/p62), and Run domain Beclin-1 interacting and cysteine-rich containing protein (Rubicon), was analyzed by western blotting (Figure 5A,B). Compound 1 stimulated the conversion of LC3-I into LC3-II, and the LC3-II expression at 24 h after the addition of compound 1 (1.55 ± 0.17-fold increase, Table S4) was significantly increased compared to the initial value. SQSTM1/p62 is an autophagy adaptor that directly binds to LC3 to induce autophagosome formation [60]. Typically, SQSTM1/p62 is degraded during autophagy, but our results showed that compound 1 caused a rapid increase in SQSTM1/p62 protein expression (2.74 ± 0.15-fold increase at 6 h, Table S4), followed by a decrease over time. Meanwhile, the expression of Rubicon, a negative regulator of autophagy [61], was observed to decrease in response to treatment with compound 1 (0.62 ± 0.08-fold increase at 24 h, Table S4). Since the accumulation of LC3-II and SQSTM1/p62 is also caused by inhibition of autophagy, an autophagy flux assay was performed to determine whether the present results obtained with compound 1 were due to enhancement or inhibition of autophagy. Chloroquine (CQ) inhibits autophagy at its final stage by suppressing the fusion of autophagosomes with lysosomes [62]. The results in Figure 5C showed that LC3-II expression in the control group was significantly increased in the presence of CQ (1.84 ± 0.05-fold increase, Table S5); however, an even more significant increase was observed with the combination of compound 1 and CQ (2.58 ± 0.21-fold increase, Table S5). Furthermore, cells were collected 4 h after sample addition, and mRNA expression was confirmed by real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR); the mRNA expression levels of SQSTM1 and RUBCN were observed to increase and decrease, respectively, compared to the control (relative mRNA level was 2.93 ± 0.07 for SQSTM1, 0.63 ± 0.02 for RUBCN, Table S6). This indicated that compound 1 enhanced autophagy in HepG2 cells pretreated with high glucose and oleic acid concentrations.

Discussion
Disruption of hepatic lipid homeostasis resulting in hepatic TG accumulation is commonly described based on five concepts: (1) increased fatty acid influx, (2) increased de novo lipogenesis, (3) reduced TG secretion, (4) reduced fatty acid oxidation, and (5) dysfunctional autophagy [60,63,64]. As a research strategy to find naturally occurring ingredients that contribute to the prevention and improvement of fatty liver (NAFLD) associated with lifestyle-related diseases, we constructed a screening system to induce intracellular TG accumulation using specific culture medium and screened various herbal medicines and food-derived ingredients, with an increase or decrease in intracellular TG content as indicators. In our study, HepG2 cells were cultured in DMEM containing high glucose with oleic acid. After 24 h, the amount of intracellular TG content greatly increased compared to the control. The cells were treated with several limonoids, and most of the gedunin-type limonoids isolated from Andiroba reduced intracellular TG content by 10-40% compared to the control at 24 h (Table 1). Gedunin-type limonoids (1)(2)(3)(4)(5)(6)(7)(8) are intact ABCD limonoid ring skeleton or the D-ring-seco limonoids, while compound 9 is the B,D-ring-seco limonoid, compounds 10-12 are the A,D-ring-seco-limonoids, and compounds 13 and 14 are members of the C-ring-seco class ( Figure 1). Therefore, the intact A, B, and C-rings of limonoid may be essential for the activity expression, however, further evaluation of wider range of limonoids is required. In the gedunin-type limonoids, 7-deacetoxy-7α-hydroxygedunin (6) with a hydroxy group at C-7 position showed the strongest activity, while the activity vanished in the case of 6α-acetoxygedunin (3), with two acetoxy groups at C-6 and C-7 positions. The activity was enhanced by the presence of an oxo-substituent at C-7 position. This indicates that the substitutions at these positions in the gedunin skeleton affected the activity.
There have been several reports on the anti-obesity effects of limonoids; nomilin (10) and obacunone (11) are agonists of the bile acid-specific G protein-coupled receptor, TGR5, and are known to improve insulin resistance and increase energy expenditure by activating TGR5 in L cells of the small intestine and skeletal muscle [65][66][67]. In addition, nimbolide (14) has exhibited anti-obesity effects by reducing oxidative stress in previous studies using primary cultured hepatocytes and high-fat diet-loaded rats [68,69]. Ceramicine B derived from Chisocheton ceramicus has been reported to suppress adipocyte differentiation and lipid-droplet accumulation due to interruption of the phosphorylation of Forkhead box O1 (Foxo1) in the mouse preadipocyte cell line MC3T3-G2/PA6 [70]. Among the Andirobaderived limonoids, DAOG (1) has been reported to enhance the brown adipose tissue to improve hyperlipidemia and insulin resistance in diet-induced obese mice and suppress lipid accumulation by reducing glucose transporter (GLUT) 4 expression in adipocytes via insulin receptor substrate (IRS)-1/Akt axis [71,72]. However, the involvement of AMPK in the anti-obesity mechanism of limonoids has not been described clearly. Therefore, in the present study, we investigated the involvement of AMPK in the mechanism of action of gedunin-type limonoids, particularly compound 1, in reducing the effects of intracellular TG.
AMPK is the key metabolic sensor for cellular energy status and is a serine/threonine protein kinase that plays a central role in regulating cellular metabolism and energy balance in mammalian cells. AMPK is activated by changes in intracellular levels of AMP:ATP and ADP:ATP ratios, or by upstream kinases, such as liver-kinase-B1 or calcium/calmodulindependent kinase 2, also known as CAMKKβ [73]. Activated AMPK (phosphorylation at Thr172 in the AMPKα subunit) induces ATP-generating pathways, such as glycolysis, fatty acid oxidation, and lipolysis, while the anabolic pathways, such as gluconeogenesis, cholesterol, and protein synthesis, are inhibited. Therefore, AMPK acts comprehensively on lipid metabolism by promoting fatty acid oxidation and inhibiting cholesterol and TG synthesis [73]. AMPK has attracted attention as a therapeutic target for various diseases, and many naturally occurring compounds have shown the ability to activate AMPK, including BBR or other constituents belonging to the polyphenol family [74]. BBR has several beneficial functions, such as antioxidant [75][76][77], cardiovascular protective [78,79], anti-hyperlipidemic [80], anti-hyperglycemic [81], hepatoprotective [82,83], nephroprotective [84,85], and immunomodulatory effects [86,87]. BBR has been reported to attenuate NAFLD in a clinical study [88] and inhibit TG levels in steatotic hepatocytes [89]. Our results indicated that both compound 1 and BBR, in combination with the AMPK inhibitor compound C, exhibit AMPK-mediated intracellular TG-reducing effects ( Figure 4A). However, western blot analysis showed that the AMPK activation effect of compound 1 was less prominent than that of BBR ( Figure 4B,C). In addition, no obvious change in ACC phosphorylation (inactivation) associated with AMPK activation was observed, suggesting that further optimization of experimental conditions, such as the cell collection timing, is necessary. Furthermore, compound 1 have been reported to repress the intracellular TG accumulation in adipocytes [72]. Since AMPK activation also suppresses de novo lipogenesis, it is consistent with our result of compound 1 activating AMPK. Thus, we presume that compound 1 regulates lipid metabolism via AMPK.
Autophagy is an evolutionarily conserved cytoplasmic degradation system that controls autolysosomal degradation and recycling of cellular components. Lipophagy is a specific form of autophagy that mediates the selective breakdown of lipid droplets in the liver and enhances β-oxidation [63,64,90,91]. Disturbances in lipophagy have been reported to be relevant to NAFLD and hepatic steatosis [92][93][94]. Autophagy is mainly regulated by mechanistic targeting of rapamycin complex 1 (mTORC1), which is responsible for anabolism and is activated by starvation. In addition, the activated AMPK promotes autophagy directly by phosphorylating autophagy-related proteins, such as mTORC1 and unc-51 like autophagy activating kinase 1 (ULK1) [95,96]. Hence, AMPK activation by compound 1 could affect intracellular TG content through the enhancement of autophagy.
Our results showed that compound 1 increased the expression of LC3-II over time and transiently increased p62 ( Figure 5A,B). LC3-I is cleaved from LC3 by autophagy-related gene 4, which binds to phosphatidylethanolamine (PE) to become LC3-II and participates in the elongation of isolation membranes [91]. Therefore, the expression level of LC3-II is frequently used as a marker of autophagy activation. However, since increased levels of LC3-II analyzed by immunoblotting may result from dysfunction of the fusion process with lysosomes, an increase in LC3-II alone cannot be considered to determine autophagy enhancement [96]. Therefore, we confirmed alterations in LC3-II expression through an autophagy flux assay in which an autophagy inhibitor artificially induces a dysfunctional state of autophagy. The results showed that LC3-II accumulation by CQ in the presence of compound 1 was significantly higher than that by CQ alone ( Figure 5C). This indicates that compound 1 could enhance autophagy. SQSTM1/p62 is a receptor protein that directs ubiquitinated autophagy substrate proteins to the isolation membrane, where they interact directly with LC3 and are degraded in an autophagy-dependent manner [97][98][99]. SQSTM1/p62 also accumulates in cells due to inhibition of autophagy; however, compound 1 increased SQSTM1/p62 mRNA expression transiently, which declined gradually. Evaluation of SQSTM1/p62 mRNA expression 4 h after the addition of compound 1 revealed that it was significantly higher than that in controls ( Figure 5D). This suggests that compound 1 stimulates SQSTM1/p62 expression and enhances autophagy. Although increased SQSTM1/p62 expression has been reported to trigger the enhancement of autophagy [100][101][102], further investigation is needed to determine whether a similar mechanism applies to compound 1. Furthermore, compound 1 decreased the protein and mRNA expression levels of Rubicon ( Figure 5B,D), an inhibitor of autophagy that has been reported to be associated with NAFLD because of its increased concentration in the liver of high-fat diet-loaded mice, resulting in inhibition of lipid droplet degradation [61]. In addition, Rubicon is reported as one of the signatures of aging because autophagic dysfunction with age in several organisms is induced by an age-dependent increase in Rubicon [103]. The reduction of Rubicon by compound 1 supports autophagy activation; however, a decrease in mRNA expression suggests that compound 1 may also be useful in aging caused by autophagy deficiency (Figure 5B,D). Few reports have shown that nimbolide (14) and azadirachtin induce autophagy [104,105], while gedunin (2) is reported to downregulate it [106], which is indicative of the relationship between limonoids and autophagy. However, our study is the first report demonstrating that limonoids improve intracellular lipid metabolism through autophagy. Presently, only gedunin-type limonoids have been studied, and more research data should be accumulated in the future.
In the current study, only in vitro evaluation using cells was conducted. The pharmacokinetics of limonoids have been well studied, especially for limonin (12). Generally, the bioavailability of limonin is very low due to its low solubility and poor permeability [107,108]. Although there are no reports on the pharmacokinetics of compound 1, it is also sparingly soluble in water, so the bioavailability is thought to be low. However, Matsumoto et al. elucidated the anti-obesity and the anti-diabetic effects of DAOG (1) by an in vivo model using high-fat-diet-fed mice [71]. Thus, compound 1 is expected to exhibit oral absorbability to some extent.

Plant Material
The flower and seed oils of C. guianensis Aublet (Meliaceae) were collected in Amazon, Brazil, in March of 2006, 2011, and 2013. Voucher specimens (CG-01-1, CGS-01-1, and CGS-01-2) were deposited at the Herbarium of the Laboratory of Medicinal Chemistry, Osaka University of Pharmaceutical Sciences, as described previously [16].

Effect on TG Content in HepG2 Cells Pretreated with High Glucose and Oleic Acid
HepG2 cells were inoculated into a 48-well tissue culture plate (10 5 cells/well in 150 µL/well MEM). After 24 h of incubation, the medium was replaced with 150 µL/well of DMEM (FUJIFILM Wako) containing low glucose (1000 mg/L), and the cells were cultured until 90% confluency. The medium was then replaced with DMEM containing high glucose (4500 mg/L) and 5% (v/v) oleic acid albumin (Sigma) for induction of intracellular TG accumulation. After 24 h, the medium was replaced with low glucose-DMEM containing a test sample or vehicle, and the cells were cultured for another 24 h. Subsequently, the medium was removed, and the cells were homogenized in distilled water (105 µL/well) by sonication. The TG and protein content in the homogenate were determined using commercial kits (LabAssay™ Triglyceride and Protein Assay BCA Kit, FUJIFILM Wako); data were expressed as the % of control of TG/protein (µg/mg). Each test compound was dissolved in dimethyl sulfoxide (DMSO) and added to the medium (final DMSO concentration was up to 1%). Compound C was applied in combination with the sample.

Oil Red O Staining
After 24 h of the sample application in TG-accumulated HepG2 cells, the cells were washed twice with phosphate-buffered saline (PBS) (−), followed by fixation using 4% neutral-buffered formalin for 10 min. The cells were again washed by PBS (−) two times and stained with 1.8 mg/mL Oil red O in 60% isopropanol for 30 min. Following staining and washing with distilled water twice, the cells were observed by microscopy.

Cell Viability
MTT assay was performed for cell viability determination. Briefly, MTT solution (5 mg/mL in PBS (−)) was diluted using DMEM (final MTT concentration: 0.5 mg/mL). After 24 h of sample treatment in HepG2 cells, the DMEM was replaced with MTT-DMEM and incubated for an additional 2 h. The medium was then removed, and 0.04 N HCl in isopropanol (100 µL/well) was added, followed by 30 min of incubation at RT to dissolve the formazan. The absorbance was measured on a microplate reader (Multiskan Sky, Thermo Fisher Scientific) using a test wavelength of 560 nm and a reference wavelength of 670 nm. Cell viability was determined using the following equation: where A is the absorbance with the test sample and B is the absorbance with the control.

Autophagy Flux Assay
The induction of TG accumulation in HepG2 cells was performed as mentioned in Section 4.6. After 18 h of sample application, 50 µL of DMEM containing CQ (final concentration; 100 µM) was added, and the cells were further incubated for 6 h. Total protein was then extracted, followed by the determination of LC3-II and β-actin expression levels by western blot analysis. β-Actin was used as a loading control. Autophagic flux was determined by comparing differences in LC3-II expression between control and sample groups with and without CQ addition.

qRT-PCR
The samples were added to HepG2 cells after induction of intracellular TG accumulation using high glucose and oleic acid containing DMEM. After 4 h of sample application, the cells were collected, and the total RNA was extracted using RNeasy Mini Kit (QIAGEN, Venlo, The Netherlands) according to the manufacturer's instructions. The concentration and purity of the RNA were determined by the absorbance ratio at 260 and 280 nm using NanoDrop Lite (Thermo Fisher Scientific). Complementary DNAs (cDNAs) were synthesized from 1 µg total RNA using iScript™ Reverse Transcription Supermix (Bio-Rad). The template cDNAs obtained were incubated with gene-specific primers and SsoAdvanced™ Universal SYBR ® Green Supermix (Bio-Rad) in a CFX Connect RT-PCR Detection System (Bio-Rad). The abundance of each gene product was calculated by relative quantification using the 2 −∆∆Cq method, with values for the target genes normalized to RPLP0. The thermal cycling program had an initial denaturation step (98 • C for 30 s), followed by 40 cycles of denaturation (98 • C for 15 s) and annealing/extension (60 • C for 30 s). The primer pairs used were: SQSTM1 primers, 5 -TGCCCAGACTACGACTTGTG-3 and 5 -AGTGTCCGTGTTTCACCTTCC-3 [101]; RUBCN primers, 5 -GATTACTGGCAGTTCGTGAAAGA-3 and 5 -CTGCTCTGGTCGTT-CTCGTG-3 [112]; and RPLP0 primers, 5 -TCTACAACCCTGAAGTGCTTGAT-3 and 5 -CAATCTGCAGACAGACACTGG-3 [113].

Statistics
The experiments were repeated at least four times. Values are expressed as mean ± S.E.M. Student's t-test and one-way analysis of variance, followed by Dunnett's or Tukey-Kramer's HSD test, were used for statistical analysis with JMP ® 9.0.2 (SAS Institute Inc., NC, USA). Probability (p) values less than 0.05 were considered significant.

Conclusions
Several types of limonoids, including Andiroba (C. guianensis)-derived limonoids, were examined for their intracellular TG-reducing effects, and we found that the gedunintype limonoids characteristically reduced intracellular TG content. Consideration of the structure-activity relationship suggested that the substituents at the C-6 and C-7 positions in the gedunin skeleton affected the activity. DAOG (1), a principal limonoid of Andiroba, was observed to be involved in AMPK activation as part of its mechanism of action. In addition, compound 1 increased autophagy markers and decreased the expression of autophagy suppressors, suggesting that its action is mediated by autophagy.

Data Availability Statement:
The data supporting the findings of this study are available from the corresponding author upon reasonable request.