Cytochrome P450 expression-associated multiple-hit pathogenesis of non-alcoholic fatty liver disease (NAFLD) in HepG2 cells

Purpose: To establish a free fatty acid (FFA)-induced non-alcoholic fatty liver disease (NAFLD) model in HepG2 cells. Methods: HepG2 cells were incubated with 0.1, 1, or 5 mM oleic acid (OA) or palmitic acid (PA) for 24 h. Histological features were examined by oil-red-O staining. Expression levels of metabolic genes (peroxisome proliferator activated receptors α/γ, sterol regulatory element binding proteins 1a/1c, acetylCoA carboxylase, acyl-CoA oxidase, and fatty acid synthase), antioxidative genes (catalase and superoxide dismutases 1/2), and cytochrome P450 genes (CYP1A2, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP4A11) were determined by reverse transcription-real time polymerase chain reaction (RT-qPCR). Results: Intracellular lipid storage was observed in cells treated with 1 mM OA or PA while cell shrinkage was present at 5 mM concentrations of both. Expression of all metabolic genes were elevated by 1 mM PA and 5 mM OA and PA. Expression of all antioxidative genes were diminished by 5 mM OA whereas 5 mM PA only reduced superoxide dismutase-2 expression. Expression of CYP1A2, CYP2D6, and CYP3A4 genes were down-regulated by both FFAs, CYP2C19 was induced by PA, while CYP2E1 and CYP4A11 were up-regulated in a concentration-dependent manner. Conclusion: PA was the more potent steatogenic agent in an OAor PAinduced NAFLD model in HepG2 cells. Increase in intracellular hepatic lipid and expression of metabolic genes, suppression of antioxidative genes, suppression of CYP1A2, CYP2D6, and CYP3A4, and induction of CYP2E1 and CYP4A11 correlated with the multiple-hit pathogenesis model of NAFLD. These findings suggest that PA-induced NAFLD model in HepG2 cells is a suitable in vitro model for studying novel therapeutic approaches to NAFLD treatment.


INTRODUCTION
Non-alcoholic fatty liver disease (NAFLD) is a major chronic liver disease affecting around 30 % of Western and Asian populations [1]. To date, fatty liver disease patients have been given lifestyle modification advice comprised of recommendations on diet and encouragement to exercise as no effective treatment has been reported. The development of NAFLD is associated with excess fat consumption and metabolic syndromes [1]. Fatty acids are involved in lipogenesis and development of hepatic fat accumulation. Oleic acid (OA) and palmitic acid (PA) are the most common dietary fatty acids. OA is the most abundant fatty acid in triglyceride (TG) stored in human adipose tissue followed by PA [2]. Monounsaturated OA (C18:1 n-9) is found in vegetable oils including olive, sunflower, safflower, and canola and in animal fats from beef, pork, camel, and krill [3]. Saturated PA (C16:0) is found in beef, palm oil, lard, and unsalted butter [4]. Previous studies in primary hepatocytes, immortalized hepatic cell lines, mice, geese, and humans have suggested induction of NAFLD by OA, PA, and their combinations occurs through cytotoxic effects, disruption of the lipid metabolic and oxidantantioxidant systems, and alteration of cytochrome P450 (CYP) profiles [5].
CYP is a superfamily of mono-oxygenase enzymes that are highly abundant in the liver and play a key role in metabolism of drugs, xenobiotics, and toxic chemicals [6]. CYPs have been implicated in the pathogenesis of fatty liver disease by promoting oxidative stress and inflammation, however the metabolic pathways have not been sufficiently described [6].
This study aimed to establish an in vitro NAFLD model in HepG2 cells to investigate the modulation of metabolic systems during steatohepatitis. NAFLD pathogenesis is believed to occur through multiple-hit including insulin resistance, mitochondrial dysfunction, endoplasmic reticulum (ER) stress, oxidantantioxidant imbalance, and inflammation, all of which are targets for NAFLD therapies. Therefore, hepatosteatosis was induced in HepG2 cells with OA and PA to examine the association of CYP450 regulation with multiplehit pathogenesis of NAFLD.

Experimental design and treatment
HepG2 cells (ATCC ® HB-8065, Manassas, USA) were cultured in DMEM supplemented with 10% FBS and 10,000 units/mL penicillin/streptomycin at 37ºC under 95% humidity and 5% CO2. The cells were seeded in a 6-well plate at a density of 5×10 5 cells/well until reaching to 80% confluence. Stock solutions (1 M) of OA and PA were prepared by dissolving in isopropanol and diluted with medium to 0.1, 1, and 5 mM before incubating with the cells for 24 h. Then the medium and the cells were collected for further analysis.

Examination of intracellular fat by oil red O staining
Oil red O solution (ORO; 0.18 %) was freshly prepared in 60 % isopropanol. At 24 h after the last treatment, the medium was removed, and the monolayer cells were washed with phosphate buffered saline (PBS). The cells were fixed by immersing in 10% neutral-buffered formalin and washed twice with distilled water, followed by immersing in 60% isopropanol. Then the fixed cells were stained with ORO. The background was cleaned via immersing in 60% isopropanol, followed by washing with distilled water. The histological features were evaluated using a Motic AE2000 inverted microscope at 10× magnification (Motic, Kowloon, Hong Kong). The image was recorded and analyzed on screen using a Motic image plus 3.0 software [7].

Histomorphology of OA and PA-induced NAFLD in HepG2 cells
The histological features of ORO-stained HepG2 cells are shown in Figure

Effect of OA-or PA-induced NAFLD on mRNA expression of CYPs in HepG2 cells
Expression of CYP1A2 mRNA was significantly suppressed by all OA or PA treatments, except OA at the concentration of 0.1 mM ( Table 5). Expression of CYP2C19 mRNA was suppressed by OA (1 and 5 mM) but induced by PA (0.1 and 1 mM) ( Table 5). OA did not change expression of CYP2D6, while PA induced CYP2D6 at the lowest concentration but suppressed it at the two higher concentrations (1 and 5 mM, Figure 5 A). The expression of CYP3A4 was suppressed by OA and PA in a concentration-dependent manner ( Table 5). The expression of CYP2E1 and CYP4A11 mRNA were increased by OA (5 mM) and all PA treatments (Table 6).

DISCUSSION
This study optimized an in vitro model of NAFLD in HepG2 cells and explored the effect of OAand PA-induced NAFLD on intracellular lipid levels and mRNA expression of metabolic, antioxidative, and CYP genes. This model was developed using HepG2 cells at the density of 5 × 10 5 cells/well with OA and PA at various concentrations that correlated to human diets  [9]. The advantages of HepG2 cells are that this cell type carries an adiponutrin variant responsive to hepatic fat accumulation, they are easy to handle and culture with high reproducibility, low cost, and relatively stable in gene expression profiles. The HepG2 adiponutrin variant Ile 148 Met is emphatically associated with liver fat content, especially TG, the main fat accumulated in the liver and the major cause of fatty liver disease. This HepG2 adiponutrin variant has been shown to increase intracellular TG content and TG concentrations in media [10]. In the current study, accumulation of intracellular lipid was demonstrated by changes to histomorphological features. HepG2 cells treated with PA at 1 mM and OA and PA at 5 mM demonstrated fat overloading with karyopyknotic cell shrinkage and the presence of apoptotic bodies indicating up-regulation of metabolic genes. The findings that PA provided more steatogenic effects and apoptotic induction was in agreement with previous studies [11].
One of the markers of multiple-hit pathogenesis of NAFLD comprises induction of FFA synthetic pathways through PPAR-, PPAR-γ, SREBP-1a, SREBP-1c, ACC, ACOX, and FAS, which not only directly affect lipogenesis but also influence insulin resistance [12,13]. In our hands, OA or PA dramatically increased expression of those metabolic genes, and PA at 1 mM was the optimal concentration for NAFLD induction in the HepG2 model.
Regarding NAFLD, fat accumulates in the liver as TG by esterification of FFAs and glycerol. Once FFAs enter hepatocytes, they form fatty acyl-CoAs through acyl-CoA synthase. Fatty acyl-CoAs could then pass into either the βoxidation pathway or esterification, which results in hepatic fat accumulation [14]. The highest dose of OA and all doses of PA induced SREBP-1C, resulting in up-regulation of hepatic de novo lipogenesis and hepatic fat accumulation. Moreover, upregulation of ACC, ACOX, and FAS prompt insulin resistance.
Insulin resistance inhibits β-oxidation of FFAs, promoting fat accumulation in the liver [13]. This supports PA activating insulin resistance more than OA. At the same concentration, OA stimulated only ACOX expression while PA triggered increased ACC, ACOX, and FAS expression. High dose of FFAs in hepatocytes induces FAS to activate serine-kinase to induce a defect in insulin signaling pathways, resulting in insulin resistance. Insulin suppresses adipose tissue lipolysis; hence, increased insulin resistance diminishes this repression, leading to increased supply of FFAs to the liver [15]. Although OA at 5 mM induced FAS, PA demonstrated more specificity increasing expression at the lowest concentration.
The observed reductions of CAT and SOD in our model indicated oxidative stress. SOD consists of 3 subtypes [16]. SOD1 is located in the cytoplasm while SOD2 is in the mitochondria and SOD2 polymorphisms are significantly correlated with liver injury. SOD3 is found extracellularly.
Alteration of mitochondrial or peroxisomal function impairs a cells ability to handle an increase in lipid flux, resulting in destruction of fat homeostasis and generation of toxic metabolites via overproduction of lipid and reactive oxygen species (ROS), finally causing hepatocyte necroinflammation and exacerbation of mitochondrial damage [14]. The highest concentration of OA led to down-regulation of CAT, SOD1, and SOD2, while PA suppressed only SOD2, but it did so from the lowest concentration. Suppression of SOD2 enhances oxidative stress via mitochondrial dysfunction. Since this occurred even at the lowest concentration of PA, this suggests PA was a greater pro-oxidant than OA.
The modulation of CYP expression observed in our model correlated to the degree of steatosis. While the highest concentration of OA upregulated CYP2E1 and CYP4A11, the lowest concentration of PA was able to significantly increase expression of CYP2E1 and CYP4A11. These results correlate with previous studies that showed OA and PA increase the level of CYP2E1 and CYP4A11 in rodents, human hepatocytes, and differentiated human cells [17,18]. Induction of CYP2E1 and CYP4A11 impairs both the ER and mitochondria. Mitochondrial impairment leads to electron leakage during the mitochondrial respiratory chain process and metabolism phase, resulting in lipid peroxidation, ROS overproduction, and oxidative stress. ER impairment activates the unfolded protein response (UPR) and activates c-Jun N-terminal kinase (JNK), which trigger inflammation, apoptosis, and insulin resistance [13,19].
Furthermore, increased CYP2E1 and CYP4A11 expression and concomitant exposure to their substrate drugs can lead to severe cellular injury due to the over-production of toxic metabolites, such as acetone from CYP2E1 and ketone bodies from CYP4A11 [20,21]. These findings support that CYP2E1 induces fatty liver disease whether it is induced by alcohol, or not [22,23]. Previous investigations have demonstrated a relationship between NAFLD progression and decreased activity of CYP1A2, CYP2D6, and CYP3A4 [18,24]. In our model, PA reduced expression of CYP1A2, CYP2D6, and CYP3A4, while OA down-regulated CYP1A2 and CYP3A4. The expression of CYP2C19 has been reported either as induction or inhibition in humans [18,25]. Concentrations of 1 and 5 mM OA decreased CYP2C19 expression while 0.1 and 1 mM PA induced CYP2C19. These findings suggest that modulation of CYP2C19 expression was dependent on the structure of the FFA.
The present model revealed the relevance of the FFA-concentration to the degree of steatosis. The activation of PPAR-, PPAR-γ, SREBP-1a, SREBP-1c by fat accumulation in the liver causes lipotoxicity, mitochondrial dysfunction, and ER stress. Furthermore, the induction of ACC, ACOX, and FAS, which are down-stream of the carbohydrate-responsive element-binding protein (ChREBP) and SREBP, indicates the development of insulin resistance while the down-regulation of CAT and SOD2 mRNA expression causes oxidative stress. Taken together, the increases in CYP2E1 and CYP4A11 expression, decreases in CYP1A2, CYP2D6, and CYP3A4 expression and changes in CYP2C19 expression affect the progression of NAFLD via ER and mitochondrial injury and potentially through an increase in toxic products from clinical drug metabolism and the associated imbalance in the expression of metabolic and anti-oxidative genes potentially authorizes gene mutations. This correlation of multiple changes to CYP450 regulatory profiles and metabolic and anti-oxidative gene expression support the multiple-hit pathogenesis model of NAFLD progression.

CONCLUSION
This study has established a novel in vitro NAFLD model in HepG2 cell derived multiple-hit pathogenesis associated with regulation of CYP450s. PPAR-, PPAR-γ, SREBP-1a, SREBP-1c, ACC, ACOX, and FAS were induced while CAT and SOD2 were suppressed. CYP2E1 and CYP4A11 were elevated while CYP1A2, CYP2C19, CYP2D6 and CYP3A4 were attenuated. PA was a more potent steatogenic agent with less apoptotic effects than OA. PA at a concentration of 1 mM was optimal for induction of NAFLD in HepG2 cell.