Psoralen Suppresses Lipid Deposition by Alleviating Insulin Resistance and Promoting Autophagy in Oleate-Induced L02 Cells

Non-alcoholic fatty liver disease (NAFLD) held a high global prevalence in recent decades. Hepatic lipid deposition is the major characteristic of NAFLD. We aim to explore the mechanisms of psoralen on lipid deposition in NAFLD. The effects of psoralen on insulin resistance, lipid deposition, the expression and membrane translocation of glucose transporter type 4 (GLUT4), autophagy, and lipogenesis enzymes were determined on sodium oleate-induced L02 cells. Chloroquine and 3-MA were employed. The AMP-activated protein kinase alpha (AMPKα) was knocked down by siRNA. Psoralen alleviated insulin resistance in sodium oleate-induced L02 hepatocytes by upregulating the expression and membrane translocation of GLUT4. Psoralen inhibited lipid accumulation by decreasing the expression of key lipogenesis enzymes. Psoralen promotes autophagy and the autophagic flux to enhance lipolysis. Psoralen promoted the fusion of the autophagosome with the lysosome. Both chloroquine and 3-MA blocked the effects of psoralen on autophagy and lipid accumulation. The AMPKα deficiency attenuated the effects of psoralen on autophagy and lipid accumulation. Our study demonstrated that as an antioxidant, psoralen attenuates NAFLD by alleviating insulin resistance and promoting autophagy via AMPK, suggesting psoralen to be a promising candidate for NAFLD.


Introduction
The prevalence of non-alcoholic fatty liver disease (NAFLD) is on a steady upward trend worldwide in the last decades. The latest prevalence was estimated to be around 25.24% of the global population and 29.2% of the Chinese population [1,2]. NAFLD is mainly featured with insulin resistance and hepatic lipid accumulation, which encompasses a series of liver diseases, from simple hepatic steatosis to non-alcoholic steatohepatitis (NASH). It has been reported that about 59.10% of NAFLD patients evolve to NASH in 5 years and approximately 41% of the NASH patients progress to fibrosis and even cirrhosis [3]. Notably, among patients with NASH cirrhosis, the estimated annual incidence of hepatocellular carcinoma ranges from 0.5% to 2.6% [4,5].
Up to now, there are no effective medical interventions for NAFLD except for lifestyle modifications. Although vitamin E and pioglitazone have been demonstrated to be beneficial for NAFLD patients, the United States Food and Drug Administration hasn't approved any specific drugs [6]. However, our previous studies have demonstrated that flavonoids such as naringenin and icaritin have obvious effects on ameliorating NAFLD [7,8]. As a kind of flavonoid, psoralen is the major antioxidant of Medik (syn. Psoralea corylifolia L) which is a traditional Chinese medicine termed "Buguzhi". Clinically, psoralen along with ultraviolet A was used to treat hyperproliferative skin disorders such as psoriasis [9]. Psoralen was reported to ameliorate lipid accumulation in L02 cells [10,11]. However, the mechanism of psoralen ameliorating hepatic lipid accumulation remains unclear.
We aim to determine the effects and explore the mechanisms of psoralen on insulin resistance and lipid deposition in NAFLD. We established the NAFLD cell model by inducing L02 cells with sodium oleate. Then the impacts of psoralen on insulin resistance, lipid accumulation, lipogenesis (especially autophagy), and the activation of Adenosine 5 -monophosphate (AMP)-activated protein kinase (AMPK) pathway were determined. Our study demonstrated that psoralen ameliorates NAFLD by alleviating insulin resistance and promoting autophagy via AMPK. Since insulin resistance and hepatic lipid deposition contribute to several metabolic disorders such as type II diabetes and atherosclerosis, psoralen may serve as a candidate for metabolic diseases such as NAFLD.

Cell Culture and Treatments
The human hepatic L02 cell line was obtained from the American Type Culture Collection (ATCC), cultured with RPMI-1640 medium (Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (Biological Industries, Cromwell, CT, USA) and 1% penicillinstreptomycin (HyClone, Logan, UT, USA) at 37 • C and 5% CO 2 . When the confluence reached 40 to 50 percent, L02 cells were induced with sodium oleate at 100 µM for 24 h to induce lipid deposition. Then, L02 cells were stimulated with insulin (10 µg/mL) for 30 min and then treated with psoralen at 0.37, 1.1, and 3.3 µM for 24 h. Thereafter, glucose content was measured. To detect the effects of psoralen on autophagy, sodium oleate-induced L02 cells were treated with psoralen (3.3 µM) alone, or in combination with chloroquine (CQ, 10 µM), or 3-methyladenine (3-MA, 10 mM) for 24 h.

Glucose Content Assay
Glucose content was measured by the glucose content assay kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The culture medium was collected and mixed with the glucose assay solution. Then the mixture was transferred into a 96-well plate and determined by a SpectraMax 190 microplate reader (Molecular Devices, San Jose, CA, USA).

Reverse Transcription PCR
We performed semi-quantitative reverse transcription PCR. Total RNA was extracted from L02 cell lysate by a TransZol Up kit (TransGen, Beijing, China) according to the manufacturer's instructions. Then, cDNA was generated using super M-MuLV reverse transcriptase (Sangon, Shanghai, China), oligo(dT)12-18 primer (Sangon, Shanghai, China), and RNase inhibitor (BBI, Shanghai, China). The mixture was incubated at 50 • C for 45 min and reactions were terminated by incubation at 80 • C for 15 min. The cDNA was amplified using the following PCR conditions: 94 • C for 5 min, 30 cycles of 94 • C for 30 s, 60 • C for 30 s and 72 • C for 30 s, followed by 72 • C for 5 min, and 4 • C for 30 s. In addition, the PCR products were electrophoresed on 1% agarose gel. Finally, images were taken by ultraviolet exposure. Primers for SREBP1, FASN, CPT1a, β-actin, α-tubulin, and GAPDH are listed in Supplementary Table S1.

Nile Red Staining
L02 cells were stained with Nile red (5 µM) staining solution (Sangon, Shanghai, China) at 37 • C in darkness for 30 min. Then, cells were fixed with 4% PFA for 20 min after washing. Finally, nuclei were stained with Hoechst 33258 (Sigma, St. Louis, MO, USA) for 3 min. Pictures were taken by confocal microscopy (Carl Zeiss, Jena, Germany).

RNA Interference
After L02 cells reached 60-70% confluence, cells were transfected with control scrambled siRNA or AMPKα siRNA (Sangon, Shanghai, China) utilizing the TurboFect transfection reagent (Thermo Fisher, Waltham, MA, USA) for 72 h. The efficiency of knocking down the target protein, AMPKα, was evaluated by Western blotting.

Statistical Analysis
All experiments were conducted in triplicate. The gray value and fluorescence intensity were analyzed by ImageJ Version 1.53c. Data were expressed as the mean ± standard deviation (SD). The colocalization analysis was carried out by using the line-scan function of ImageJ. Representative cells were selected from each group and a 160 pixel-long line across the cell was drawn to scan the fluorescence dot on the cell membrane, cytoplasm, and nucleus. The fluorescence peak overlapping represented the colocalization. All experiment data were analyzed with a Student's t-test by GraphPad Prism version 7.0. A p-value less than 0.05 was regarded as statistically significant.

Psoralen Alleviates Insulin Resistance in Sodium Oleate-Induced L02 Cells by Increasing the Expression and Membrane Translocation of GLUT4
The NAFLD cell model was established by inducing L02 cells with sodium oleate (100 µM) for 24 h. As shown in Figure 1A, control group cells exhibited decreased glucose content upon insulin stimulation. Compared with the control group cells, the 0 µM group cells exhibited much higher glucose content upon insulin stimulation, indicating that insulin resistance was established in sodium oleate-induced cells. Notably, psoralen significantly reduced the glucose content of sodium oleate-induced L02 cells in a dose-dependent manner. Since GLUT4 plays important role in glucose utilization, decreased location of GLUT4 on the cell membrane was reported to induce insulin resistance [12]. While GLUT4 is found primarily in adipose tissues and striated muscle, GLUT2 is a major isoform in the liver and insulin-independent. We measured the expression of GLUT4 and GLUT2 by Western blotting. As shown in Figure 1B, psoralen significantly upregulated the expression of GLUT4 at 0.37, 1.1, and 3.3 µM in a dose-dependent manner, while the expression of GLUT2 was not significantly changed by psoralen. Therefore, the expressions of several proteins including insulin receptor (INSR), p-INSR, insulin receptor substrate 1 (IRS1), and p-IRS1 in the insulin signaling cascade were detected by Western blotting. Psoralen enhanced the ratio of p-INSR/INSR and p-ISR1/ISR1 in a dose-dependent manner ( Figure 1C). Colocalization analysis showed that sodium oleate significantly decreased the colocalization of GLUT4 and F-actin as detected by immunofluorescence ( Figure 1D,E). Psoralen significantly enhanced the colocalization of GLUT4 and F-actin, especially at 3.3 µM, compared with sodium oleate-induced L02 cells. In contrast, the expression of GLUT2 and colocalization with F-actin were not changed by either sodium oleate or psoralen compared with control cells ( Figure 1F,G). These results indicated that psoralen ameliorates sodium oleate-induced insulin resistance by increasing GLUT4 expression and membrane translocation.

Psoralen Alleviates Lipid Accumulation by Reducing Lipogenesis
As shown in Figure 2A, sodium oleate induced obvious lipid deposition in L02 cells as detected by Nile red staining, which may result from increased lipogenesis or reduced lipolysis. Therefore, we determined expressions of lipogenesis regulating proteins such as SREBP1 and FASN, and lipolysis rate-limiting enzyme, CPT1a, by Western blotting. As shown in Figure

Psoralen Attenuates Lipid Accumulation by Enhancing Autophagy and Promoting the Autophagic Flux
After clarifying the effects of psoralen on lipogenesis, we then focused on its impact on lipolysis. It has been reported that autophagy, as a kind of lipolysis, had important roles in human diseases resulting from lipid accumulation [13]. Therefore, we further investigated the relation between autophagy and newly generated lipid droplets by utilizing the GFP plasmids of lipid droplet marker DFCP1 [14]. As shown in Figure 3A,B, sodium oleate induced remarkable expression of DFCP1 in L02 cells, which was decreased by psoralen. Sodium oleate induced obvious endogenous LC3 puncta in L02 cells, which was further enhanced by psoralen ( Figure 3C,D). To further elucidate the impact of psoralen on autophagy, we measured the ratio of LC3II/LC3I which is a reliable marker of the synthesis of the autophagosome, and the expression of the autophagy substrate, p62, which indicates the synthesis of autolysosome [15,16]. As shown in Figure 3E, psoralen significantly increased the ratio of LC3II/LC3I and decreased the expression of p62 at 0.37, 1.1, 3.3 µM in sodium oleate-induced L02 cells. Notably, CQ, the autophagy inhibitor, blocked the effects of psoralen on alleviating lipid deposition in sodium oleate-induced L02 cells ( Figure 3F,G). These results demonstrated that psoralen ameliorates lipid accumulation by enhancing autophagy. Consistently, psoralen increased the number of endogenous LC3 puncta in sodium oleate-induced L02 cells in a dose-dependent manner ( Figure 4A,B). Since autophagy is a continuous multistep process, we further determined the effects of psoralen on the autophagic flux and employed CQ as an autophagic flux inhibitor [17]. Both CQ and psoralen remarkably increased endogenous LC3 puncta in L02 cells. Psoralen increased endogenous LC3 puncta, which was further enhanced by CQ ( Figure 4C,D). Consistently, Western blotting results showed that CQ further enhanced the ratio of LC3II/LC3I and the expression of p62 ( Figure 4E).

Psoralen Enhances Lipolysis by Promoting the Initiation and Autophagosome-Lysosome Fusion of Autophagy
To further elucidate the mechanism of psoralen on the autophagic flux, we transfected L02 cells with the mRFP-EGFP-LC3 plasmid to track exogenous autophagosomes. Psoralen increased the ratio of red fluorescent mRFP-LC3 puncta to total LC3 puncta, which represents the ratio of LC3 in lysosomes to total LC3 in sodium oleosome-induced L02 cells ( Figure 5A,B). Furthermore, we examined the colocalization of LC3 and LAMP1 in sodium oleate-induced L02 cells by immunofluorescence. As shown in Figure 5C, psoralen increased endogenous LC3 puncta at 0.37, 1.1, and 3.3 µM in a dose-dependent manner. As shown in Figure 5D, psoralen dose-dependently promoted the colocalization of LC3 and LAMP1 at 0.37, 1.1, and 3.3 µM. These results demonstrated that psoralen promoted the fusion of the autophagosome and lysosome. Thereafter, we wondered whether psoralen promoted autophagosome synthesis. The PI3K/AKT pathway was reported to promote autophagosome synthesis [18]. So 3-methyladenine (3-MA), the class III PI3K inhibitor, was employed. As shown in Figure 6A,B, 3-MA blocked the effects of psoralen on LC3 puncta. As detected by Western blotting, psoralen significantly decreased the expression of p62 and increased the ratio of LC3II/LC3I, which was reversed by 3-MA ( Figure 6C). As detected by Nile red staining, 3-MA also blocked the effects of psoralen on lipid deposition ( Figure 6D,E). Consistently, psoralen suppressed the expression of FASN and promoted the expression of CPT1a, which was reversed by 3-MA ( Figure 6F). (D) Colocalization efficiency of LC3 and LAMP1 was quantified by line scan analysis (60 pixels with two ends on the membrane) by observing the overlap of fluorescence intensity peaks across the contours of multiple L02 cells (n ≥ 3 cells). All values were expressed as the mean ± SD from three independent experiments. * p < 0.05, *** p < 0.001 vs. sodium oleate-induced group. ## p < 0.01, ### p < 0.001, vs. psoralen treated group. Abbreviations: CQ, chloroquine; LAMP1, lysosomal associated membrane protein 1; Pso, psoralen.

Psoralen Improves Insulin Resistance and Attenuates Lipid Accumulation by Activating AMPK
The LKB1/AMPK/ACC pathway plays a major role in promoting GLUT4 translocation, autophagy, and reducing lipogenesis [19][20][21]. LKB1 is generally regarded as the main protein activating AMPK [22]. Therefore, we wondered whether psoralen plays a role through the AMPK/ACC pathway. As shown in Figure 7A, psoralen promoted the phosphorylation of AMPK and ACC at 0.37, 1.1, and 3.3 µM in a dose-dependent manner. As shown in Figure 7B, psoralen promoted the phosphorylation of LKB1 in a dose-dependent manner. AMPKα is the catalytic subunit that is closely associated with lipogenesis [23]. Activation of AMPK by AMPKα subunit has been reported to improve energy expenditure [24]. So, we employed Compound C (Comp C) which is the AMPK inhibitor. As shown in Figure 7C, psoralen induced the phosphorylation of AMPK, which was blocked by Comp C. To further demonstrate the crucial role of the AMPKα subunit in psoralen's effect, we knocked down AMPKα by siRNA. As detected by Western blotting, AMPKα siRNA significantly reduced the expression of AMPK ( Figure 8A) and ACC ( Figure 7D). Consistently, AMPKα siRNA blocked the effects of psoralen on the expression of GLUT4 ( Figure 8A). Psoralen decreased the expression of FASN and SREBP1 and enhanced the ratio of LC3II/LC3I of L02 cells, which was also blocked by siAMPKα ( Figure 8A). Knocking down AMPKα also significantly decreased endogenous LC3 puncta (Figure 8 B,C). As shown in Figure 8D,E, psoralen significantly alleviated lipid deposition in L02 cells, while knocking down AMPKα impaired the effects of psoralen.

Discussion
Psoralen plus ultraviolet therapy is widely used for hyperproliferative skin disorders such as vitiligo, psoriasis, and eczema because of the high ultraviolet absorbance of psoralen [25]. Although psoralen had been reported to induce skin cancer in treating PUVA at 100 µM because of its UV-mediated activation property, we confirmed the effects of a far lower dose of psoralen (3.3 µM) on NAFLD by employing the sodium oleate-induced L02 cells [26]. Insulin resistance and hepatic lipid accumulation are the two main pathological characteristics of NAFLD. We observed that psoralen significantly ameliorated insulin resistance by upregulating the expression and membrane translocation of GLUT4 in a dose-dependent manner. Moreover, psoralen does inhibit lipogenesis by decreasing the expression of Srebp1 and Fasn. Psoralen also promotes lipolysis through enhancing autophagy, whose effects were blocked by CQ or 3-MA. Our study demonstrated that psoralen does attenuate NAFLD, suggesting psoralen to be a potential candidate for NAFLD or insulin resistance-induced metabolic disorders. Considering the skin cancer-inducing effects of psoralen, NAFLD patients with hyperproliferative skin disorders or photosensitive seizures should not be given psoralen in large doses [27]. Psoralen should also be avoided to be used in combination with other medicines with phototoxicity such as doxycycline.
Various studies have indicated the tight link between NAFLD and mitochondria dysfunction, ER stress, and oxidative stress [28]. These cellular stress can be attenuated by general autophagy. Moreover, lipophagy is also a critical pathway for lipid droplet degradation. Promoting lipophagy has been demonstrated to be beneficial for alleviating lipid accumulation [13]. These reports suggested an important role of autophagy in ameliorating NAFLD. Notably, damaged autophagy has been revealed to promote NAFLD progression [29]. Promoting autophagy shall be an efficient way to ameliorate NAFLD progression. It has been reported that isopsoralen enhanced the autophagic flux to ameliorate apoptosis of rat chondrocytes [30]. However, whether psoralen improves damaged autophagy in NAFLD remains unclear. Our study demonstrated that psoralen enhanced autophagy, the initiation of autophagy, and the autophagosome-lysosome fusion in the autophagic flux in NAFLD. We employed DFCP1, which is the marker of newly generated lipid droplets. We observed the negative correlation between DFCP1-marked lipid droplet and LC3-marked autophagosome in psoralen-treated L02 cells, which indicated that psoralen enhances lipolysis by promoting autophagy. Therefore, we employed the lysosome marker, LAMP1, and the autophagosome marker, LC3, to detect the effects of psoralen on the autophagosome-lysosome fusion. Psoralen promotes the colocalization of LAMP1 and LC3, which demonstrated that psoralen promoted autophagosome-lysosome fusion. However, the increasing autophagosome-lysosome fusion may be the result of the psoraleninduced excessive endogenous autophagosome. To determine whether psoralen directly promoted the autophagosome-lysosome fusion process, we utilized mRFP-EGFP-LC3 plasmid to track exogenous autophagosomes. Psoralen increased the ratio of red mRFP-LC3 puncta/total LC3 compared to the sodium oleate-induced group. These results verified that psoralen promotes autophagosome-lysosome fusion. We also observed that psoralen enhanced total LC3 puncta compared to sodium oleate-induced L02 cells. Therefore, we concluded that psoralen promotes autophagosome synthesis. Following our hypothesis, the class III PI3K inhibitor, 3-MA, evidently blocked psoralen's effect on promoting LC3 puncta. Notably, although psoralen decreased expression of SREBP1 and FASN, both CQ and 3-MA did suppress the effects of psoralen on alleviating lipid deposition. Our results demonstrated that psoralen promotes autophagy to alleviate lipid accumulation in NAFLD. To our knowledge, our study is the first to reveal the exact effects of psoralen on impaired autophagy in NAFLD.
AMPK has been well recognized as the key energy sensor and regulator which holds a central position in energy status [31]. The AMPK pathway plays important role in lipid and glucose metabolism, and autophagy [21]. The activity of AMPK is reduced by obesity, inflammation, and diabetes [32]. Enhancing the activity of AMPK has been considered as a feasible strategy for ameliorating NAFLD [33]. Our study showed that psoralen attenuates lipid deposition in sodium oleate-induced L02 cells by activating the LKB1/AMPK/ACC pathway. Notably, knocking out the AMPKα subunit by siRNA blocked the effects of psoralen on the expression of Srebp1, Fasn, and autophagy. Most importantly, knocking out the AMPKα subunit did reverse the inhibition of psoralen on lipid accumulation. In addition to NAFLD, chronic energy imbalance leads to many metabolic diseases such as obesity and type 2 diabetes mellitus [34,35]. Activating AMPK enhances energy expenditure, which helps to restore energy balance. So, our study suggested that psoralen may be a potential candidate for these metabolic diseases by activating AMPK.
The lipid accumulation in hepatocytes has been reported partly due to defects in lipoprotein synthesis and secretion [36]. Increased lipid peroxidation may disrupt the membrane structure in cells such as endoplasmic membrane and mitochondria and lead to lipid metabolism disorders [37]. As an antioxidant, psoralen may play lipid-lowering effects by decreasing lipid peroxidation. Psoralen was reported to reduce the expression of cyclooxygenase-2 (COX-2) and inducible nitric oxide synthase (iNOS) in lipopolysaccharide (LPS)-induced mouse macrophages (RAW264.7) in vitro [38]. Therefore, we believe that psoralen may attenuate NAFLD by ameliorating lipid deposition, decreasing lipid peroxidation, and promoting the expression of COX-2 and iNOS.
Although no specific drugs have been approved by the FDA for NAFLD, some promising medicines such as Sirtuin 4 (SIRT4) targeting drugs are already in the pipeline [39]. SIRT4 has been identified as a key molecule in NAFLD. SIRT4 coordinates the balance between lipid synthesis and catabolism [40]. More importantly, SIRT4 regulates fatty acid oxidation and mitochondrial gene expression in both hepatocytes and muscle cells. These characteristics make SIRT4 a pharmacological target for NAFLD since NAFLD is a complicated metabolic disease involving the liver, muscle, and adipose [41]. Enhanced circulating levels of SIRT4 have been identified as a potential marker for oxidative stress in NAFLD patients [42]. Natural products such as flavonoids have been extensively demonstrated as modulators of sirtuins including SIRT4 [43]. As a kind of flavonoid, psoralen may also be a potential modulator of SIRT4, which needs to be verified in further studies.
Our study has some limitations. The effects of psoralen for NAFLD need to be further confirmed by animal experiments and clinical trials. Furthermore, the binding of psoralen to the AMPKγ1 subunit was determined by molecular docking and siRNA, which will be strengthened by a co-immunoprecipitation assay.
In conclusion, our study demonstrated that psoralen, as an antioxidant, ameliorates NAFLD by alleviating insulin resistance and restoring lipid metabolism homeostasis. Psoralen exerts effects by promoting autophagosome synthesis and the autophagic flux via activating AMPK. Our study suggested that psoralen is a promising candidate for metabolic diseases by activating AMPK to enhance energy expenditure and alleviate insulin resistance.