- Email: [email protected]
Comprehensive metabolic responses of HepG2 cells to fine particulate matter exposure: Insights from an untargeted metabolomics
Science of the Total Environment 691 (2019) 874–884
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Comprehensive metabolic responses of HepG2 cells to fine particulate matter exposure: Insights from an untargeted metabolomics Guozhu Ye a,b,1, Dongxiao Ding b,c,1, Han Gao b,c, Yulang Chi a,b, Jinsheng Chen a,b, Zeming Wu d, Yi Lin e,⁎, Sijun Dong a,b,⁎⁎ a
Center for Excellence in Regional Atmospheric Environment, Institute of Urban Environment, Chinese Academy of Sciences, 1799 Jimei Road, Xiamen 361021, China Key Laboratory of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, 1799 Jimei Road, Xiamen 361021, China University of Chinese Academy of Sciences, 19 Yuquan Road, Beijing 100049, China d iPhenome Biotechnology (Dalian) Inc., 300-8 Jinlongsi Road, Dalian 116063, China e State Key Laboratory of Molecular Vaccinology and Molecular Diagnostics, School of Public Health, Xiamen University, Xiamen 361102, China b c
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• PM2.5 induced substantial metabolic changes in HepG2 cells. • PM2.5 induced accumulation of FFAs and medium-chain acylcarnitines (6–12 carbons). • PM2.5 induced accumulation of oncometabolites, purine and pyrimidine nucleosides. • PM2.5 inhibited most amino acid pathways. • PM2.5 activated cysteine and methionine metabolism, and glutathione metabolism.
a r t i c l e
i n f o
Article history: Received 29 March 2019 Received in revised form 12 July 2019 Accepted 12 July 2019 Available online 13 July 2019 Editor: Elena Paoletti Keywords: Fine particulate matter Metabolomics HepG2 cell Fatty acid Acylcarnitine Nucleotide
a b s t r a c t Exposure to fine particulate matter (PM2.5) increases the risk of metabolic diseases, such as cancer and cardiovascular disease. Disturbed hepatocyte metabolism accelerates the incidence and progression of metabolic diseases. However, toxic effects of PM2.5 on hepatocyte metabolism remain unclear. Accordingly, an untargeted metabolomics approach based on liquid chromatography–mass spectrometry was used to characterize comprehensive metabolic responses of HepG2 cells to PM2.5 exposure and to discover potential therapeutic targets for PM2.5induced metabolic dysregulation in metabolic diseases. Metabolomics revealed that exposure to liposoluble extracts of PM2.5 samples (LE) triggered substantial changes in 46 metabolic pathways, mainly involved in lipid, amino acid, nucleotide and carbohydrate metabolism, in HepG2 cells. Notably, LE exposure induced accumulation of FFAs and medium-chained acylcarnitines (6–12 carbons), but decreased levels of short-chained acylcarnitines (b5 carbons) in HepG2 cells. Meanwhile, levels of citrate/isocitrate and aconitate were decreased, while 2-hydroxyglutate and succinate accumulated in HepG2 cells treated with LE. Additionally, levels of adenosine triphosphate, guanosine triphosphate, uridine triphosphate and cytidine triphosphate were decreased; however, contents of adenosine monophosphate, guanosine monophosphate, purines and pyrimidines were increased in HepG2 cells treated with LE. Moreover, levels of glutathione, Glu-Cys, Cys-Gly, lipoic acid, methionine
Abbreviations: PM2.5, fine particulate matter; PPAR, peroxisome proliferator-activated receptor; LC-MS, liquid chromatography–mass spectrometry; LE, liposoluble extracts of PM2.5; DMSO, dimethyl sulfoxide; QC, quality control; LPC, lysophosphatidylcholine; LPI, lysophosphatidylinositol; LPE, lysophosphatidylethanolamine; LPG, lysophosphatidylglycerol; TCA, tricarboxylic acid; AhR, aryl hydrocarbon receptor. ⁎ Correspondence to: Y. Lin, School of Public Health, Xiamen University, Xiamen 361102, China. ⁎⁎ Correspondence to: S. Dong, Institute of Urban Environment, Chinese Academy of Sciences, 1799 Jimei Road, Xiamen 361021, China. E-mail addresses: [email protected] (Y. Lin), [email protected] (S. Dong). 1 These authors contributed equally to this work.
https://doi.org/10.1016/j.scitotenv.2019.07.192 0048-9697/© 2019 Elsevier B.V. All rights reserved.
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
875
sulfoxide, methionine and S-adenosyl-L-methionine were increased, while those of most amino acids were decreased in HepG2 cells treated with LE. These data demonstrated that LE exposure triggered accumulation of FAAs and oncometabolites (2-hydroxyglutate and succinate), mitochondrial dysfunctions characterized by incomplete FFA oxidation and reduced energy supply from TCA cycle and oxidative phosphorylation, disturbances in methylation and redox homeostasis, and the inhibition of most amino acid metabolism in HepG2 cells. Above metabolic disorders indicates potential therapeutic targets for treating PM2.5-induced injury and diseases. To the best of our knowledge, this study provides the first evidence that LE exposure triggered accumulation of medium-chain acylcarnitines, oncometabolites, purines and pyrimidines in HepG2 cells. © 2019 Elsevier B.V. All rights reserved.
1. Introduction Substantial epidemiological, animal and in vitro evidence demonstrated that exposure to fine particulate matter (PM2.5) could cause various adverse health effects, such as increased incidences of hepatic fibrosis, nonalcoholic steatohepatitis, diabetes, lung cancer, cardiovascular disease, respiratory disease, and low birth weight (Pimpin et al., 2018; Rajagopalan et al., 2018; Turner et al., 2011; Xu et al., 2018). Short- and long-term exposure to PM2.5 raised the relative risk of acute cardiovascular disease incidence by 1 to 3% and ~ 10%, respectively (Rajagopalan et al., 2018). Additionally, a 1% rise in diabetes incidence was associated with each 10 μg/m3 rise in PM2.5 concentration (Pearson et al., 2010). Moreover, a 10 μg/m3 increase in PM2.5 exposure correlated with a 15–27% rise in lung cancer mortality (Turner et al., 2011). Accordingly, PM2.5 exposure is highly correlated with the incidence and development of metabolic diseases. As one important metabolic organs, the liver regulates digestion, detoxification, synthesis and storage of various biochemical components, and is essential for systemic energy homeostasis in biological systems (Purushotham et al., 2009). Since the liver mediates key aspects of lipid metabolism, amino acid metabolism, steroid metabolism, and many other metabolic pathways, chronic metabolic disorders in the liver would lead to the incidence and development of liver damages, such as non-alcoholic fatty liver diseases ranging from steatosis to non-alcoholic steatohepatitis, fibrosis and cirrhosis (Byrne et al., 2009; Lee et al., 2003; Purushotham et al., 2009; Thomas et al., 2008). Notably, non-alcoholic fatty liver diseases increase the risk of other more serious metabolic diseases, such as diabetes, cardiovascular disease and cancer (Gaggini et al., 2013; Kim et al., 2018; Smith and Adams, 2011). Therefore, insights into metabolic disturbances in the liver/hepatocytes benefit the understanding of molecular pathology of metabolic diseases and relevant potential therapeutic target discovery. Hepatic damages due to PM2.5 exposure have been attracting more and more attentions (Liu et al., 2017; Xu et al., 2018; Zheng et al., 2013; Zheng et al., 2015). Exposure to PM2.5 for 10 weeks triggered impaired glucose intolerance, glycogen storage and insulin resistance, activated inflammatory response pathways mediated by Toll-like receptor 4, nuclear factor kappa B and c-Jun N-terminal kinase, while repressed the expression of peroxisome proliferator-activated receptor (PPAR) α/γ and the signaling mediated by insulin receptor substrate 1 in the liver, leading to hepatic inflammation, steatosis and fibrosis in male C57BL/6 mice (Zheng et al., 2013). Moreover, oxidative stress and inflammation in the liver induced by 24-wek PM2.5 inhalation promoted abnormal hepatic function and lipid accumulation in mice (Xu et al., 2018). However, effects of PM2.5 exposure on hepatocyte/liver metabolism remain unclear. Metabolomics, aiming to comprehensive qualitative and quantitative analysis of metabolites in biological systems, have been used to discover metabolic disturbances induced by PM2.5 exposure (Cui et al., 2019; Fiehn, 2002; Liang et al., 2018; Zhang et al., 2017; Zhang et al., 2019). Nonetheless, comprehensive metabolic responses of hepatocytes/livers to PM2.5 exposure are still undiscovered. Therefore, an untargeted metabolomics approach based on liquid chromatography–mass spectrometry (LC-MS) was employed to present the comprehensive metabolic responses of HepG2 cells to liposoluble
extracts of PM2.5 samples (LE) and to discover potential therapeutic targets for PM2.5-induced metabolic disturbances in metabolic diseases. 2. Materials and methods 2.1. Materials Ammonium bicarbonate (≥99.0%), HPLC-grade methanol, dichloromethane, acetonitrile and formic acid were purchased from SigmaAldrich (Shanghai, China). Ultrapure water was prepared employing a Milli-Q system (Millipore, USA). HepG2 cells were ordered from the Cell Bank of Type Culture Collection, Chinese Academy of Sciences (Shanghai, China). 2.2. PM2.5 sample collection and liposoluble component extraction PM2.5 samples were kindly provided from Prof. Chen (Institute of Urban Environment, Chinese Academy of Sciences). The samples were collected in Gulou Campus in Nanjing University (China) during winter. The liposoluble components were extracted with dichloromethane/ methanol (v/v, 2:1) via accelerated solvent extraction (Dionex 350). Detailed sample information and the extraction could be referred to the relevant study (Hong et al., 2017). LE was dried and redissolved in dimethyl sulfoxide (DMSO) as the stock solution for subsequent cell experiments. The dried liposoluble extracts of blank sampling paper were also reconstituted in DMSO as the control. 2.3. Cell culture and treatment HepG2 cells were cultured in Minimum Essential Medium (Gibco, NY, USA) added with 10% Fetal Bovine Serum (HyClone, Victoria, Australia), 1.5 g/L NaHCO3 and 0.11 g/L sodium pyruvate in a humidified incubator with 5% CO2 at 37 °C. Since that a high-risk individual exposure to 79 μg/m3 of PM2.5 over a 24-h period was estimated to be comparable to in vitro 0.2–20 μg/cm2, and that seasonal average PM2.5 concentrations in Nanjing during the winter sampling period was 78 ± 42 μg/m3 (Li et al., 2003; Nie et al., 2018). Additionally, during haze events in winter and spring, the daily average PM2.5 concentration in major cities in China was above 75 μg/m3 (Chinese pollution standard), and it exceeded 100 μg/m3 in some cities, such as Beijing and Xi'an (Huang et al., 2014). We previously found that HepG2 cell viability was significantly reduced to 69.28% compared to the control after 25 μg/cm2 of LE exposure (Ding et al., 2019). Accordingly, HepG2 cells were exposed to LE at 25 μg/cm2 for 72 h to investigate consequent metabolic responses. 2.4. Cell sample preparation for metabolomics analysis Following cell culture and treatment, the medium was discarded. Cells were immediately quenched in liquid nitrogen and stored at −80 °C prior to PBS washing. The cell samples was added with 1000 μL of ice-cold methanol/water (v/v, 4:1) and harvested with scrapers. The sample was drawn into a centrifuge tube and vortexed for 1.0 min. After cell-sample centrifugation at 13, 000 rpm for 15 min at 4 °C, 600 μL of the supernatant was drawn and vacuum dried in a
876
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
SpeedVac concentrator (Thermo Scientific, USA). To monitor and assess the stability and reliability of the metabolomics approach, the residual supernatant from all samples was withdrawn and mixed evenly, then distributed into 600-μL aliquots as quality control (QC) samples. One QC sample was inserted every 3 analytical samples and treated in the same way as analytical samples during vacuum drying, resuspension, instrumental analysis and data processing. 2.5. Instrumental analysis for the metabolomics approach The dried cell extracts were resuspended with 200 μL MeOH/water (v:v, 1:1) solution, and placed into glass vials for LC-MS analysis. Metabolic profiles were obtained by employing a Ultimate 3000 UHPLC system coupled to Q Exactive quadruple-Orbitrap high resolution mass spectrometer (Thermo Scientific, San Jose, USA). Metabolite separation was achieved on an Acquity BEH C18 column (1.7 μm, 2.1 × 100 mm, Waters Co., USA). Two and five μL of the resuspended metabolite extracts were injected in positive and negative electrospray ionization mode, respectively. Mobile phases A (water containing 0.1% formic acid) and B (acetonitrile containing 0.1% formic acid) were employed in positive mode with the gradient elution: initial 2% B increased to 98% in 12 min prior to 3 min column wash and equilibration. On the other hand, Mobile phases C (5 mM ammonium bicarbonate in water) and D (5 mM ammonium bicarbonate in methanol/acetonitrile mixture, v/v = 20/80) were used in negative mode, with initial 98% C linearly decreased to 0% in 12 min. Metabolites were eluted at a flow rate of 400 μL/min and directly transferred into the mass spectrometer. Parameters of the ion source were set as follows: ion transfer capillary temperature, 320 °C; heater temperature, 375 °C; sheath gas, 45 arbitrary units; auxiliary gas, 10 arbitrary units; ionization voltage, 4.0 kV (positive) and 3.5 kV (negative); S-lens level, 50%. The mass scan range was acquired between 100 and 1000 m/z. Mass resolution of the Orbitrap was set to 70,000 (FWHM) with 1e6 automatic gain control target and 100-ms maximum ion injection time. Full scan-ddMS2 data of QC samples were repeatedly acquired to obtain high-resolution MS2 spectra for structural verification of metabolites. Key parameters of ddMS2 mode were set as follows: mass resolution (FWHM),17,500; automatic gain control target, 1e5; loop count, 7; precursor isolation window, 1.0 Da; stepped normalized collision energy, 15, 30 and 45%; dynamic exclusion, 5 s. 2.6. Data preprocessing for the metabolomics approach Raw spectral data were preprocessed by Compound Discoverer 2.1 (Thermo Scientific, San Jose) for peak alignment, ion detection and deconvolution, gap re-filling, artifacts exclusion, element prediction
and structural elucidation. Metabolites were firstly identified by searching a local endogenous lipid and metabolite database within 3 ppm mass tolerance, and further verified through searching against an in-house built MS2 spectral library and mzCloud library (www. mzcloud.org). The spectral similarity between the retrieved and reference spectra was calculated by a proprietary Highchem HighRes algorithm, and the similarity score threshold was 50. All identified metabolites were exported into TraceFinder software (Thermo Scientific, San Jose) for peak integration and manual check. Ion peaks of metabolites in positive mode were divided by the total ion current and multiplied by 1 × 109. On the other hand, ion peaks of metabolites in negative mode were divided by the total ion current and multiplied by 1 × 1010. Subsequently, the data in positive and negative mode were merged and applied for statistical analysis. 2.7. Statistical analysis Principal component analysis and pathway analysis were conducted via MetaboAnalyst 3.0 (Xia et al., 2015). Significant differences in metabolite levels among different groups were assessed by a nonparametric test (two-tailed Mann-Whitney U test) using MeV 4.9.0 (Saeed et al., 2006). The significant level was further adjusted by the false discovery rate based on Benjamini-Hochberg Correction using MeV 4.9.0, and the false discovery rate was set to 0.12. The level of statistical significance was 0.05. After compound name standardization and data normalization (mean-centered and divided by the standard deviation of each variable), “Global test” and “Relative-betweeness Centrality” algorithms were employed for pathway enrichment analysis and pathway topology analysis, respectively. Subsequently, KEGG metabolic pathways related to Homo sapiens were used for pathways analysis. The heat map was plotted by MeV 4.9.0. 3. Results 3.1. Global metabolic responses of HepG2 cells in response to LE exposure Distribution of QC samples and relative standard deviation of metabolites in QC samples were utilized to evaluate the analytical performance. It was clear in the principal component analysis score plot that 3 QC samples were clustered together (Fig. 1A). Moreover, 279, 285 and 289 metabolites, separately accounting for 94.3, 96.3 and 97.6% of total identified metabolites, had relative standard deviation below 15, 20 and 30%, respectively (Fig. 1B). The data demonstrated high stability and repeatability of the metabolomics approach (Begley et al., 2009; Ye et al., 2014).
Fig. 1. Analytical performance of the metabolomics approach and changes in the metabolic profiling of HepG2 cells induced by LE exposure. (A) Sample distribution in the score plot of principal component analysis. (B) Relative standard deviation distribution of metabolites in the QC samples.
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
We found that metabolic profiles of HepG2 cells treated with LE were significantly different from those of the control group (Fig. 1A). Subsequently, a two-tailed Mann-Whitney U test was utilized to discover differential metabolites (p b 0.05) between the control and exposure group. Among 296 annotated metabolites, 123 metabolites were discovered to be significantly disturbed in HepG2 cells in response to LE exposure, including 10 metabolites involved in carbohydrate metabolism, 39 metabolites involved in amino acid and oligopeptide
877
metabolism, 54 metabolites involved in lipid metabolism, 16 metabolites involved in purine and pyrimidine nucleotide metabolism, and 4 metabolites involved in other metabolic pathways (Table S1 and Fig. 2). Pathways analysis revealed that 46 metabolic pathways were significantly altered in HepG2 cells in response to LE exposure (Table S2). Notably, levels of free fatty acids (FFAs) and most metabolites involved in nucleoside metabolism were all significantly increased, while levels of most amino acids were decreased in HepG2 cells treated
Fig. 2. Heat map of metabolic responses of HepG2 cells to LE exposure. n = 5 per group. After unit variance scaling, levels of metabolites were used to plot the heat map.
878
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
with LE (Fig. 2). Detailed metabolic changes in HepG2 cells treated with LE are provided below.
FFA transport into mitochondria and subsequent oxidation degradation in HepG2 cells in response to LE exposure.
3.2. LE exposure induced FFA accumulation and significant changes in acylcarnitines in HepG2 cells
3.3. LE exposure induced significant changes in phospholipid, sphingolipid and steroid metabolism in HepG2 cells
Significant changes in FFAs and acylcarnitines in HepG2 cells were induced by LE exposure (Figs. 2 and 3). Levels of short-chain, medium-chain and long-chain FFAs, such as FFA(6:0), FFA(12:1), FFA (14:1), FFA(20:5), FFA(22:6) and FFA(24:6), were all significantly increased in HepG2 cells in response to LE exposure (Fig. 3A). In addition, the level of hydroxylated FFA(16:0) was also significantly increased in HepG2 cells treated with LE. FFAs participate in the synthesis of other lipids (such as triglycerides, phospholipids and bioactive signaling lipids) and energy homeostasis via mitochondrial oxidative metabolism. We found that levels of short-chained acylcarnitines (b5 carbons, such as carnitine, acetylcarnitine, propionylcarnitine, succinylcarnitine and hydroxybutyrylcarnitine) were significantly decreased, however, levels of medium-chained acylcarnitines (6 to 12 carbons), including hexanolycarnitine, acylcarnitine(8:0), acylcarnitine(10:1) and acylcarnitine(12:1), were significantly increased in HepG2 cells treated with LE (Fig. 3B). Changes in acylcarnitine levels suggested disturbed
We found significant changes in phospholipid, sphingolipid and steroid metabolism in HepG2 cells in response to LE exposure (Figs. 2 and 4, Table S1). Levels of acetylcholine and glycerophosphocholine were significantly decreased, while the level of phosphorylcholine was significantly increased in HepG2 cells treated with LE (Fig. 4). Besides, levels of lysophosphatidylcholines (LPCs), such as LPC(16:0) sn-1, LPC(16:0) sn-2, LPC(18:2) sn-1 and LPC(18:2) sn-2, were all significantly decreased in HepG2 cells treated with LE (Fig. 4). Moreover, levels of most lysophosphatidylinositols (LPIs), such as LPI(16:0) sn-1, LPI (16:0) sn-2, LPI(18:1) sn-1 and LPI(18:1) sn-2, were significantly increased in HepG2 cells treated with LE (Fig. 4). Significant changes in levels of lysophosphatidylethanolamines (LPEs) and lysophosphatidylglycerols (LPGs) were also observed in HepG2 cells treated with LE (Fig. 4). These data demonstrated that LE exposure induced significant changes in phospholipids, especially phosphatidylcholines and phosphatidylinositols. On the other hand, levels of sphingosine
Fig. 3. LE exposure induced FFA accumulation (A) and significant changes in acylcarnitines (B) in HepG2 cells. Data were represented as the mean + SD. *, p b 0.05, **, p b 0.01, two-tailed Mann-Whitney U test. n = 5 per group.
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
879
Fig. 4. LE exposure induced significant changes in phospholipid metabolism in HepG2 cells. LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; LPI, lysophosphatidylinositol; LPG, lysophosphatidylglycerol; PC, phosphatidylcholine; PE, phosphatidylethanolamine; DG, diglyceride; CDP-DG, 1,2-diacyl-sn-glycero-3-cytidine-5′diphosphate. Data were represented as the mean + SD. *, p b 0.05, **, p b 0.01, two-tailed Mann-Whitney U test. n = 5 per group.
(d16:0), sphingosine(d20:1) and phytosphingosine were significantly decreased, while the level of 7-dehydrodesmosterol was significantly increased in HepG2 cells, which illustrated disturbed sphingolipid and steroid metabolism induced by LE exposure (Fig. 2 and Table S1).
3.4. LE exposure induced significant changes in carbohydrate metabolism in HepG2 cells Levels of fructose 1,6-bisphosphate, 6-phosphogluconic acid, 3phosphoglyceric acid and phosphoenolpyruvic acid were significantly disturbed in HepG2 cells treated with LE, which indicated that LE exposure induced metabolic disturbances in glycolysis/gluconeogenesis, pentose phosphate pathway and pyruvate metabolism in HepG2 cells (Fig. 5). Besides, levels of cis-aconitic acid and citric acid (or isocitric acid) were significantly decreased, while levels of succinic acid and 2hydroxyglutarate were significantly increased in HepG2 cells treated with LE (Fig. 5). The data demonstrated that LE exposure triggered disorders of tricarboxylic acid (TCA) cycle in HepG2 cells. Additionally, metabolites related to amino sugar and nucleotide sugar metabolism, such as uridine diphosphate glucuronic acid and UDP-N-acetylglucosamine, were also disturbed in HepG2 cells treated with LE (Fig. 5).
3.5. LE exposure induced significant changes in amino acid metabolism in HepG2 cells We also observed significant changes in amino acid metabolism in HepG2 cells in response to LE exposure (Fig. 5). Levels of metabolites involved in alanine, aspartate and glutamate metabolism (such as glutamine, glutamate, aspartate and N-Acetylaspartylglutamic acid) and arginine and proline metabolism (i.e., proline, 4-hydroxyproline, spermidine and N1-acetylspermidine) were all significantly decreased in HepG2 cells treated with LE (Fig. 5). On the contrary, levels of metabolites involved in cysteine and methionine metabolism (such as Methionine, methionine sulfoxide and S-adenosyl-L-methionine), glutathione metabolism (i.e., glu-cys, cys-gly, glutathione and Smethyl glutathione) and leucine metabolism (such as leucine and 3hydroxy-3-methylglutarate) were significantly increased in HepG2 cells treated with LE (Fig. 5). Changes in cysteine and methionine metabolism and glutathione metabolism indicated the increase in methylation and oxidative stress in HepG2 cells induced by LE exposure. Moreover, levels of most metabolites in aromatic amino acid metabolism (such as tyrosine, hydroxyphenyllactic acid, tryptophan, kynurenine and 5-hydroxy-L-tryptophan), glycine, serine and threonine metabolism (such as serine, threonine and betaine), histidine
880
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
Fig. 5. LE exposure induced significant changes in carbohydrate and amino acid metabolism in HepG2 cells. Data were represented as the mean + SD. *, p b 0.05, **, p b 0.01, two-tailed Mann-Whitney U test. n = 5 per group.
metabolism, and lysine metabolism were significantly decreased in HepG2 cells treated with LE (Fig. 5 and Table S1). 3.6. LE exposure induced significant changes in purine and pyrimidine nucleotide metabolism in HepG2 cells Significant changes in purine and pyrimidine nucleotide metabolism occurred in HepG2 cells in response to LE exposure (Fig. 6). Levels of
most metabolites involved in adenine and guanine nucleotide metabolism, such as xanthine, adenosine monophosphate, guanine, guanosine and guanosine monophosphate, were significantly increased in HepG2 cells treated with LE, except adenosine triphosphate and guanosine triphosphate (Fig. 6A). We observed that levels of uracil, uridine and uridine 5′-monophosphate involved in uridine nucleotide metabolism were significantly increased, while levels of uridine 5′-diphosphate and uridine triphosphate were significantly decreased in HepG2 cells
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
881
Fig. 6. LE exposure induced significant changes in purine (A) and pyrimidine (B) nucleotide metabolism in HepG2 cells. Data were represented as the mean + SD. *, p b 0.05, **, p b 0.01, two-tailed Mann-Whitney U test. n = 5 per group.
treated with LE (Fig. 6B). Additionally, levels of cytidine 5′-diphosphate and cytidine triphosphate involved in cytidine nucleotide metabolism were significantly decreased in HepG2 cells treated with LE (Fig. 6B). These data suggested disturbed energy production and/or nucleic acid synthesis in HepG2 cells treated with LE. 4. Discussion Owing to high separation efficiency, resolution, sensitivity, specificity and throughput as well as the development of various derivatization technologies, chromatography-mass spectrometry has become the mainstream technology for the development and application of metabolomics, especially LC-MS. However, there is currently no single technology that could cover all the metabolites in biological systems, and each technology has its own technical advantages and disadvantages. Compared to gas chromatography–mass spectrometry commonly employed in metabolomics analysis, time-consuming chemical derivatization produces are usually not needed for LC-MS analysis. Additionally, metabolites difficult to volatilize or to be derivatized can be detected by LC-MS, but not by gas chromatography–mass spectrometry, such as carnitines, cholines, and lysophospholipids. However, some types of polar metabolites, such as saccharides and short-chain organic acids, cannot be effectively separated by LC-MS, but can be effectively separated by gas chromatography–mass spectrometry after proper derivatization. In this study, 50% methanol solution was used to dissolve the dried sample prior to LC-MS analysis to reduce the solvent effect, which improved the peak shape of metabolites and subsequent analytical reproducibility and metabolite coverage. On the other hand, there are N1000 metabolites with accurate parent ions and their fragments as well as retention times in our self-built spectral library. Totally, 296 metabolites were accurately identified and quantified, whereas
metabolites insoluble in 50% methanol solution, such as triglycerides, cannot be detected in this study. Non-polar metabolites insoluble in 50% methanol solution can be effectively determined using either untargeted metabolomics or lipidomics based on LC-MS after dissolution in methanol or other non-polar solvents. Accumulated data revealed that PM2.5 exposure increased the incidence of metabolic diseases, and that metabolic disturbances in the liver/hepatocytes promoted the risk of metabolic diseases. However, metabolic disturbances in the liver/hepatocytes in response to PM2.5 exposure remain unclear. Accordingly, an untargeted metabolomics approach based on high-resolution LC-MS was employed to uncover metabolic responses of HepG2 cells to LE exposure. We found that LE exposure induced significant disturbances in many metabolic pathways, especially lipid, amino acid and nucleotide metabolism. Notably, FFA accumulated in HepG2 cells in response to LE exposure. Furthermore, disturbed acylcarnitine shuttle system, responsible for FFA transport for mitochondrial oxidation and working as an acyl-CoA pool, was also observed in HepG2 cells treated with LE. The decreased levels of short-chain acylcarnitines, including acetylcarnitine, propionylcarnitine, succinylcarnitine and hydroxybutyrylcarnitine, and the increased levels of medium-chain acylcarnitines, including hexanolycarnitine, acylcarnitine(8:0), acylcarnitine(10:1) and acylcarnitine(12:1) indicated incomplete FFA oxidation, accelerated FFA oxidation and lipid overload in mitochondrial in HepG2 cells treated with LE (Wolf et al., 2013; Yu et al., 2012). Medium-chain acylcarnitine accumulation might be due to an imbalance between FFA influx and oxidation (Overmyer et al., 2015). AhR (aryl hydrocarbon receptor) as the receptor of polycyclic aromatic hydrocarbons (major components of LE) could be activated after LE exposure, and then stimulated CD36 expression and resultant increases in FFA uptake and accumulation (Matsumoto et al., 2007; Yao et al., 2016). Excessive FFAs could be
882
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
stored in neutral lipids (such as triglycerides) in HepG2 cells in response to LE exposure (Ding et al., 2019). Moreover, AhR activation induced by LE exposure could increase the activity of PPARα via up-regulation of genes involved in circadian rhythmicity, thus promoting FFA transport for mitochondrial oxidation via carnitine O-palmitoyltransferase 1 (Wang et al., 2011). Therefore, accumulation of FFAs and mediumchain acylcarnitines was probably due to accelerated FFA influx and/or incomplete oxidation induced by disturbed AhR and PPAR signaling pathways in HepG2 cells after LE exposure. Short-chained acylcarnitines, especially carnitine and propionylcarnitine could improve metabolic functions (Ramsay and Zammit, 2004). For example, oral administration of propionyl-carnitine attenuated abnormalities of cardiac function and reduced plasma lipids in streptozotocin-diabetic rats (Terada et al., 1998). Additionally, both the carnitine and propionyl-carnitines are able to reduce mitochondrial acetyl-CoA, leading to the activation of pyruvate dehydrogenase and resultant increase in pyruvate oxidation (Ramsay and Zammit, 2004). However, levels of short-chained acylcarnitines were decreased in our study, which was not conducive to the improvement of metabolic function in HepG2 cells treated with LE (Ramsay and Zammit, 2004). Moreover, we also found a significant increase in the level of 3phosphoglyceric acid and phosphoenolpyruvic acid, precursors for pyruvate synthesis, which indicated disturbed pyruvate metabolism in HepG2 cells treated with LE. The intermediate products of FFA oxidation could enter the TCA cycle. As observed, levels of citrate (or isocitrate) and aconitate were decreased in HepG2 cells treated with LE, which suggested less materials for the oxidative decarboxylation to produce energy. Furthermore, we found significant decreases in levels of adenosine triphosphate, guanosine triphosphate, uridine triphosphate and cytidine triphosphate in HepG2 cells treated with LE. PM exposure induced a decrease in CCAA T/enhancer binding protein alpha, which led to a decline in the mRNA expression of genes involved in NADH ubiquinone dehydrogenase (NDUFA2 and NDUFS4), NADH dehydrogenase in mitochondrial complex I (NDUFA1 and NDUFC2) and adenosine triphosphate synthase in mitochondrial complex V (ATP5H), resulting in the suppression of mitochondrial oxidative phosphorylation and adenosine triphosphate synthesis (Li et al., 2017). These data indicated that LE exposure triggered a decrease in energy generation from TCA cycle and oxidative phosphorylation in HepG2 cells. Moreover, we found succinate accumulation in HepG2 cells treated with LE. Excessive succinate production activated hypoxia-inducible factor-1α expression, which accelerated glycolysis and the synthesis of lactate and interleukin 1β, aggravating the inflammatory status in HepG2 cells induced by PM2.5 (Appari et al., 2018; Xu et al., 2018). As revealed in A549 cells, exposure to water-insoluble and total fractions of Ottawa urban dust (EHC-93) triggered an increase in the release of lactate dehydrogenase and a decrease in cell proliferation and adenosine triphosphate production (Vuong et al., 2017). Of note, both 2-hydroxyglutarate and succinate accumulation could inhibit the activity of 2-ketoglutarate-dependent dioxygenases, including the ten-eleven translocation family of cytosine hydroxymethylases, prolylhydroxylases and the Jumonji family of histone demethylases, thus leading to hypermethylation in HepG2 cells treated with LE (Raffel et al., 2017; Sullivan et al., 2016). Moreover, we found significant increases in levels of methionine and S-adenosyl-L-methionine, which indicated more methyl donors available for the hypermethylation in HepG2 cells treated with LE. Genome-wide DNA methylation analysis of fasting venous blood samples from students revealed that methylation levels were changed with the concentration of PM2.5 exposure (24-h averages: 53.1 vs. 24.3 μg/m3) at 49 CpG loci, where 31 CpG sites were annotated to specific genes, and that DNA methylation of the annotated genes was increased with elevated PM2.5 exposure, which was involved in lipid and glucose metabolism, insulin resistant, oxidative stress, platelet activation, inflammation and cell survival and apoptosis (Li et al., 2018). These data indicated an import role of methylation in the molecular toxicology of PM2.5 exposure.
The glycolytic intermediates could be utilized for the synthesis of nucleotides, amino acids and substrates in pentose phosphate pathway to generate NADPH. Since that PM exposure induced decreases cell proliferation and adenosine triphosphate production, and that most metabolites involved in amino acid metabolism were decreased in HepG2 cells treated with LE in this study, the glycolytic intermediates were probably utilized to provide substrates in pentose phosphate pathway for NADPH generation to defense against reactive oxygen species (ROS) in HepG2 cells after LE exposure，in addition to producing lactate. The decreased activity of mitochondrial electron transport chain in our study could lead to mitochondrial ROS generation in HepG2 cells treated with LE (Appari et al., 2018). We found significant increased levels of glutathione and its precursors (Glu-Cys and Cys-Gly), lipoic acid and methionine sulfoxide (a biomarker of oxidative stress), which suggested that the synthesis of glutathione and ROS was increased, and that antioxidant system was activated in HepG2 cells treated with LE. PM2.5 exposure for 24 weeks at the concentration of 230 ± 2.5 μg/m3 demonstrated significant increases in oxidative stress, inflammatory response, dyslipidemia (such as accumulation of triglycerides, total cholesterol and FFAs) and hepatic dysfunctions in mice (Xu et al., 2018). On the other hand, blocking oxidative stress and NF-κB signaling pathway with pyrrolidine dithiocarbamate and N-acetyl-L-cysteine, respectively, reduced expression levels of genes related to oxidative stress, inflammation and lipid accumulation in vitro, suggesting PM2.5induced oxidative stress and inflammation as major promoters of dyslipidemia in mice liver (Xu et al., 2018). We also shown that levels of most amino acids were reduced in HepG2 cells treated with LE, which indicated less amino acids available for the generation of energy, proteins, nucleotides and lipids to sustain cell proliferation and growth. PPARs governs the expression of many genes involved in amino acid metabolism, including those related to transamination, deamination, amino acid inter-conversions, generation of amino acid derivatives and alpha-keto acid oxidation (Kersten et al., 2001). Accordingly, decreases in levels of most amino acids were probably related to disturbances in AhR and PPAR signaling induced by LE exposure in HepG2 cells (Matsumoto et al., 2007; Wang et al., 2011; Yao et al., 2016). 5. Conclusions An untargeted metabolomics approach based on LC-MS was applied to characterize metabolic responses of HepG2 cells to LE exposure. We discovered that LE exposure induced significant changes in many metabolic pathways (including lipid, amino acid and nucleotide metabolism,
Fig. 7. Overview of metabolic responses of HepG2 cells to LE exposure. Red/blue fonts: most relevant metabolites significantly increased/decreased in HepG2 cells treated with LE. Black fonts: metabolites not detected or not significantly altered in HepG2 cells treated with LE. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
and TCA cycle) in HepG2 cells (Fig. 7). Notably, LE exposure induced accumulation of FFAs and medium-chained acylcarnitines (6–12 carbons), while decreased levels of short-chained acylcarnitines (b5 carbons) in HepG2 cells. Besides, levels of citrate/isocitrate and aconitate were decreased, but 2-hydroxyglutate and succinate accumulated in HepG2 cells treated with LE. We also found that levels of adenosine triphosphate, guanosine triphosphate, uridine triphosphate and cytidine triphosphate were decreased, however, those of adenosine monophosphate, guanosine monophosphate, purines and pyrimidines were increased in HepG2 cells treated with LE. Moreover, levels of glutathione, Glu-Cys, Cys-Gly, lipoic acid, methionine sulfoxide, methionine and S-adenosyl-L-methionine were increased, while those of most amino acids were decreased in HepG2 cells treated with LE. These data demonstrated that LE exposure induced accumulation of FAAs and oncometabolites (2-hydroxyglutate and succinate), mitochondrial dysfunctions characterized by incomplete FFA oxidation and reduced energy supply from TCA cycle and oxidative phosphorylation, disturbances in methylation and redox homeostasis, and the inhibition of most amino acid metabolism in HepG2 cells. Above metabolic disturbances provide potential therapeutic targets for treating PM2.5-induced injury and diseases. To the best of our knowledge, this study provides the first demonstration that LE exposure triggered accumulation of medium-chain acylcarnitines, oncometabolites, purines and pyrimidines in HepG2 cells. Author contributions G.Y., D.X., Y.L. and S.D. conceived and designed this study. G.Y. performed the metabolomics analysis and statistical analysis, interpreted the data, wrote and revised the manuscript. H.G. and Y.C. participated in the cell sample preparation for metabolomics analysis. Z.W. conducted the instrumental analysis and subsequent data processing for metabolomics analysis. J.C. provided the PM2.5 sample. Declaration of Competing Interest The authors declare that they have no conflict of interest. Acknowledgements This work was supported by the National Natural Science Foundation of China [grant numbers 21507128, 41390240, 21777158, 21477124 and 21677140]; the Natural Science Foundation of Fujian Province, China [grant numbers 2018J01020 and 2017J01028]; the Key Laboratory of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences [grant number J008]; and the Department of Science and Technology of Fujian Province, China [grant number 2016T3005]. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2019.07.192. References Appari, M., Channon, K.M., McNeill, E., 2018. Metabolic regulation of adipose tissue macrophage function in obesity and diabetes. Antioxid. Redox Signal. 29, 297–312. Begley, P., Francis-McIntyre, S., Dunn, W.B., Broadhurst, D.I., Halsall, A., Tseng, A., et al., 2009. Development and performance of a gas chromatography-time-of-flight mass spectrometry analysis for large-scale nontargeted metabolomic studies of human serum. Anal. Chem. 81, 7038–7046. Byrne, C.D., Olufadi, R., Bruce, K.D., Cagampang, F.R., Ahmed, M.H., 2009. Metabolic disturbances in non-alcoholic fatty liver disease. Clin. Sci. 116, 539–564. Cui, J., Fu, Y., Lu, R., Bi, Y., Zhang, L., Zhang, C., et al., 2019. Metabolomics analysis explores the rescue to neurobehavioral disorder induced by maternal PM2.5 exposure in mice. Ecotox. Environ. Safe. 169, 687–695.
883
Ding, D., Ye, G., Lin, Y., Lu, Y., Zhang, H., Zhang, X., et al., 2019. MicroRNA-26a-CD36 signaling pathway: pivotal role in lipid accumulation in hepatocytes induced by PM2.5 liposoluble extracts. Environ. Pollut. 248, 269–278. Fiehn, O., 2002. Metabolomics - the link between genotypes and phenotypes. Plant Mol. Biol. 48, 155–171. Gaggini, M., Morelli, M., Buzzigoli, E., DeFronzo, R.A., Bugianesi, E., Gastaldelli, A., 2013. Non-alcoholic fatty liver disease (NAFLD) and its connection with insulin resistance, dyslipidemia, atherosclerosis and coronary heart disease. Nutrients 5, 1544–1560. Hong, Z., Hong, Y., Zhang, H., Chen, J., Xu, L., Deng, J., et al., 2017. Pollution characteristics and source apportionment of PM2.5-bound n-alkanes in the Yangtze River Delta, China. Aerosol Air Qual. Res. 17, 1985–1998. Huang, R.-J., Zhang, Y., Bozzetti, C., Ho, K.-F., Cao, J.-J., Han, Y., et al., 2014. High secondary aerosol contribution to particulate pollution during haze events in China. Nature 514, 218–222. Kersten, S., Mandard, S., Escher, P., Gonzalez, F.J., Tafuri, S., Desvergne, B., et al., 2001. The peroxisome proliferator-activated receptor alpha regulates amino acid metabolism. FASEB J. 15, 1971–1978. Kim, G.A., Lee, H.C., Choe, J., Kim, M.J., Lee, M.J., Chang, H.S., et al., 2018. Association between non-alcoholic fatty liver disease and cancer incidence rate. J. Hepatol. 68, 140–146. Lee, C.-H., Evans, R.M., Olson, P., 2003. Minireview: lipid metabolism, metabolic diseases, and peroxisome proliferator-activated receptors. Endocrinology 144, 2201–2207. Li, N., Hao, M., Phalen, R.F., Hinds, W.C., Nel, A.E., 2003. Particulate air pollutants and asthma. A paradigm for the role of oxidative stress in PM-induced adverse health effects. Clin. Immunol. 109, 250–265. Li, X., Yang, H., Sun, H., Lu, R., Zhang, C., Gao, N., et al., 2017. Taurine ameliorates particulate matter-induced emphysema by switching on mitochondrial NADH dehydrogenase genes. Proc. Natl. Acad. Sci. U. S. A. 114, E9655–E9664. Li, H., Chen, R., Cai, J., Cui, X., Huang, N., Kan, H., 2018. Short-term exposure to fine particulate air pollution and genome-wide DNA methylation: a randomized, double-blind, crossover trial. Environ. Int. 120, 130–136. Liang, D., Moutinho, J.L., Golan, R., Yu, T., Ladva, C.N., Niedzwiecki, M., et al., 2018. Use of high-resolution metabolomics for the identification of metabolic signals associated with traffic-related air pollution. Environ. Int. 120, 145–154. Liu, C., Xu, X., Bai, Y., Zhong, J., Wang, A., Sun, L., et al., 2017. Particulate air pollution mediated effects on insulin resistance in mice are independent of CCR2. Part. Fibre Toxicol. 14, 017–0187. Matsumoto, Y., Ide, F., Kishi, R., Akutagawa, T., Sakai, S., Nakamura, M., et al., 2007. Aryl hydrocarbon receptor plays a significant role in mediating airborne particulateinduced carcinogenesis in mice. Environ. Sci. Technol. 41, 3775–3780. Nie, D., Chen, M., Wu, Y., Ge, X., Hu, J., Zhang, K., et al., 2018. Characterization of fine particulate matter and associated health burden in Nanjing. Int. J. Environ. Res. Public Health 15. Overmyer, Katherine A., Evans Charles, R., Qi Nathan, R., Minogue Catherine, E., Carson Joshua, J., Chermside-Scabbo Christopher, J., et al., 2015. Maximal oxidative capacity during exercise is associated with skeletal muscle fuel selection and dynamic changes in mitochondrial protein acetylation. Cell Metab. 21, 468–478. Pearson, J.F., Bachireddy, C., Shyamprasad, S., Goldfine, A.B., Brownstein, J.S., 2010. Association between fine particulate matter and diabetes prevalence in the U.S. Diabetes Care 33, 2196–2201. Pimpin, L., Retat, L., Fecht, D., de Preux, L., Sassi, F., Gulliver, J., et al., 2018. Estimating the costs of air pollution to the National Health Service and social care: an assessment and forecast up to 2035. PLoS Med. 15, e1002602. Purushotham, A., Schug, T.T., Xu, Q., Surapureddi, S., Guo, X., Li, X., 2009. Hepatocytespecific deletion of SIRT1 alters fatty acid metabolism and results in hepatic steatosis and inflammation. Cell Metab. 9, 327–338. Raffel, S., Falcone, M., Kneisel, N., Hansson, J., Wang, W., Lutz, C., et al., 2017. BCAT1 restricts alphaKG levels in AML stem cells leading to IDHmut-like DNA hypermethylation. Nature 551, 384–388. Rajagopalan, S., Al-Kindi, S.G., Brook, R.D., 2018. Air pollution and cardiovascular disease: JACC state-of-the-art review. J. Am. Coll. Cardiol. 72, 2054–2070. Ramsay, R.R., Zammit, V.A., 2004. Carnitine acyltransferases and their influence on CoA pools in health and disease. Mol. Asp. Med. 25, 475–493. Saeed, A.I., Bhagabati, N.K., Braisted, J.C., Liang, W., Sharov, V., Howe, E.A., et al., 2006. TM4 microarray software suite. Method Enzymol 411, 134–193. Smith, B.W., Adams, L.A., 2011. Nonalcoholic fatty liver disease and diabetes mellitus: pathogenesis and treatment. Nat. Rev. Endocrinol. 7, 456–465. Sullivan, L.B., Gui, D.Y., Vander Heiden, M.G., 2016. Altered metabolite levels in cancer: implications for tumour biology and cancer therapy. Nat. Rev. Cancer 16, 680–693. Terada, R., Matsubara, T., Koh, N., Nakamura, J., Hotta, N., 1998. Effects of propionyl-Lcarnitine on cardiac dysfunction in streptozotocin-diabetic rats. Eur. J. Pharmacol. 357, 185–191. Thomas, C., Pellicciari, R., Pruzanski, M., Auwerx, J., Schoonjans, K., 2008. Targeting bileacid signalling for metabolic diseases. Nat. Rev. Drug Discov. 7, 678. Turner, M.C., Krewski, D., Pope 3rd, C.A., Chen, Y., Gapstur, S.M., Thun, M.J., 2011. Longterm ambient fine particulate matter air pollution and lung cancer in a large cohort of never-smokers. Am. J. Resp. Crit. Care. 184, 1374–1381. Vuong, N.Q., Breznan, D., Goegan, P., O'Brien, J.S., Williams, A., Karthikeyan, S., et al., 2017. In vitro toxicoproteomic analysis of A549 human lung epithelial cells exposed to urban air particulate matter and its water-soluble and insoluble fractions. Part. Fibre Toxicol. 14, 017–0220. Wang, C., Xu, C.X., Krager, S.L., Bottum, K.M., Liao, D.F., Tischkau, S.A., 2011. Aryl hydrocarbon receptor deficiency enhances insulin sensitivity and reduces PPAR-alpha pathway activity in mice. Environ. Health Perspect. 119, 1739–1744. Wolf, M., Chen, S., Zhao, X., Scheler, M., Irmler, M., Staiger, H., et al., 2013. Production and release of acylcarnitines by primary myotubes reflect the differences in fasting fat oxidation of the donors. J. Clin. Endocrinol. Metab. 98, E1137–E1142.
884
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
Xia, J., Sinelnikov, I.V., Han, B., Wishart, D.S., 2015. MetaboAnalyst 3.0-making metabolomics more meaningful. Nucleic Acids Res. 43, W251–W257. Xu, M.X., Ge, C.X., Qin, Y.T., Gu, T.T., Lou, D.S., Li, Q., et al., 2018. Prolonged PM2.5 exposure elevates risk of oxidative stress-driven nonalcoholic fatty liver disease by triggering increase of dyslipidemia. Free Radical Bio. Med. 130, 542–556. Yao, L., Wang, C., Zhang, X., Peng, L., Liu, W., Liu, Y., et al., 2016. Hyperhomocysteinemia activates the aryl hydrocarbon receptor/CD36 pathway to promote hepatic steatosis in mice. Hepatology 64, 92–105. Ye, G., Liu, Y., Yin, P., Zeng, Z., Huang, Q., Kong, H., et al., 2014. Study of induction chemotherapy efficacy in oral squamous cell carcinoma using pseudotargeted metabolomics. J. Proteome Res. 13, 1994–2004. Yu, Z., Zhai, G., Singmann, P., He, Y., Xu, T., Prehn, C., et al., 2012. Human serum metabolic profiles are age dependent. Aging Cell 11, 960–967. Zhang, S.Y., Shao, D., Liu, H., Feng, J., Feng, B., Song, X., et al., 2017. Metabolomics analysis reveals that benzo[a]pyrene, a component of PM2.5, promotes pulmonary injury by
modifying lipid metabolism in a phospholipase A2-dependent manner in vivo and in vitro. Redox Biol. 13, 459–469. Zhang, Y., Ji, X., Ku, T., Li, B., Li, G., Sang, N., 2019. Ambient fine particulate matter exposure induces cardiac functional injury and metabolite alterations in middle-aged female mice. Environ. Pollut. 248, 121–132. Zheng, Z., Xu, X., Zhang, X., Wang, A., Zhang, C., Huttemann, M., et al., 2013. Exposure to ambient particulate matter induces a NASH-like phenotype and impairs hepatic glucose metabolism in an animal model. J. Hepatol. 58, 148–154. Zheng, Z., Zhang, X., Wang, J., Dandekar, A., Kim, H., Qiu, Y., et al., 2015. Exposure to fine airborne particulate matters induces hepatic fibrosis in murine models. J. Hepatol. 63, 1397–1404.
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Comprehensive metabolic responses of HepG2 cells to fine particulate matter exposure: Insights from an untargeted metabolomics Guozhu Ye a,b,1, Dongxiao Ding b,c,1, Han Gao b,c, Yulang Chi a,b, Jinsheng Chen a,b, Zeming Wu d, Yi Lin e,⁎, Sijun Dong a,b,⁎⁎ a
Center for Excellence in Regional Atmospheric Environment, Institute of Urban Environment, Chinese Academy of Sciences, 1799 Jimei Road, Xiamen 361021, China Key Laboratory of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, 1799 Jimei Road, Xiamen 361021, China University of Chinese Academy of Sciences, 19 Yuquan Road, Beijing 100049, China d iPhenome Biotechnology (Dalian) Inc., 300-8 Jinlongsi Road, Dalian 116063, China e State Key Laboratory of Molecular Vaccinology and Molecular Diagnostics, School of Public Health, Xiamen University, Xiamen 361102, China b c
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• PM2.5 induced substantial metabolic changes in HepG2 cells. • PM2.5 induced accumulation of FFAs and medium-chain acylcarnitines (6–12 carbons). • PM2.5 induced accumulation of oncometabolites, purine and pyrimidine nucleosides. • PM2.5 inhibited most amino acid pathways. • PM2.5 activated cysteine and methionine metabolism, and glutathione metabolism.
a r t i c l e
i n f o
Article history: Received 29 March 2019 Received in revised form 12 July 2019 Accepted 12 July 2019 Available online 13 July 2019 Editor: Elena Paoletti Keywords: Fine particulate matter Metabolomics HepG2 cell Fatty acid Acylcarnitine Nucleotide
a b s t r a c t Exposure to fine particulate matter (PM2.5) increases the risk of metabolic diseases, such as cancer and cardiovascular disease. Disturbed hepatocyte metabolism accelerates the incidence and progression of metabolic diseases. However, toxic effects of PM2.5 on hepatocyte metabolism remain unclear. Accordingly, an untargeted metabolomics approach based on liquid chromatography–mass spectrometry was used to characterize comprehensive metabolic responses of HepG2 cells to PM2.5 exposure and to discover potential therapeutic targets for PM2.5induced metabolic dysregulation in metabolic diseases. Metabolomics revealed that exposure to liposoluble extracts of PM2.5 samples (LE) triggered substantial changes in 46 metabolic pathways, mainly involved in lipid, amino acid, nucleotide and carbohydrate metabolism, in HepG2 cells. Notably, LE exposure induced accumulation of FFAs and medium-chained acylcarnitines (6–12 carbons), but decreased levels of short-chained acylcarnitines (b5 carbons) in HepG2 cells. Meanwhile, levels of citrate/isocitrate and aconitate were decreased, while 2-hydroxyglutate and succinate accumulated in HepG2 cells treated with LE. Additionally, levels of adenosine triphosphate, guanosine triphosphate, uridine triphosphate and cytidine triphosphate were decreased; however, contents of adenosine monophosphate, guanosine monophosphate, purines and pyrimidines were increased in HepG2 cells treated with LE. Moreover, levels of glutathione, Glu-Cys, Cys-Gly, lipoic acid, methionine
Abbreviations: PM2.5, fine particulate matter; PPAR, peroxisome proliferator-activated receptor; LC-MS, liquid chromatography–mass spectrometry; LE, liposoluble extracts of PM2.5; DMSO, dimethyl sulfoxide; QC, quality control; LPC, lysophosphatidylcholine; LPI, lysophosphatidylinositol; LPE, lysophosphatidylethanolamine; LPG, lysophosphatidylglycerol; TCA, tricarboxylic acid; AhR, aryl hydrocarbon receptor. ⁎ Correspondence to: Y. Lin, School of Public Health, Xiamen University, Xiamen 361102, China. ⁎⁎ Correspondence to: S. Dong, Institute of Urban Environment, Chinese Academy of Sciences, 1799 Jimei Road, Xiamen 361021, China. E-mail addresses: [email protected] (Y. Lin), [email protected] (S. Dong). 1 These authors contributed equally to this work.
https://doi.org/10.1016/j.scitotenv.2019.07.192 0048-9697/© 2019 Elsevier B.V. All rights reserved.
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
875
sulfoxide, methionine and S-adenosyl-L-methionine were increased, while those of most amino acids were decreased in HepG2 cells treated with LE. These data demonstrated that LE exposure triggered accumulation of FAAs and oncometabolites (2-hydroxyglutate and succinate), mitochondrial dysfunctions characterized by incomplete FFA oxidation and reduced energy supply from TCA cycle and oxidative phosphorylation, disturbances in methylation and redox homeostasis, and the inhibition of most amino acid metabolism in HepG2 cells. Above metabolic disorders indicates potential therapeutic targets for treating PM2.5-induced injury and diseases. To the best of our knowledge, this study provides the first evidence that LE exposure triggered accumulation of medium-chain acylcarnitines, oncometabolites, purines and pyrimidines in HepG2 cells. © 2019 Elsevier B.V. All rights reserved.
1. Introduction Substantial epidemiological, animal and in vitro evidence demonstrated that exposure to fine particulate matter (PM2.5) could cause various adverse health effects, such as increased incidences of hepatic fibrosis, nonalcoholic steatohepatitis, diabetes, lung cancer, cardiovascular disease, respiratory disease, and low birth weight (Pimpin et al., 2018; Rajagopalan et al., 2018; Turner et al., 2011; Xu et al., 2018). Short- and long-term exposure to PM2.5 raised the relative risk of acute cardiovascular disease incidence by 1 to 3% and ~ 10%, respectively (Rajagopalan et al., 2018). Additionally, a 1% rise in diabetes incidence was associated with each 10 μg/m3 rise in PM2.5 concentration (Pearson et al., 2010). Moreover, a 10 μg/m3 increase in PM2.5 exposure correlated with a 15–27% rise in lung cancer mortality (Turner et al., 2011). Accordingly, PM2.5 exposure is highly correlated with the incidence and development of metabolic diseases. As one important metabolic organs, the liver regulates digestion, detoxification, synthesis and storage of various biochemical components, and is essential for systemic energy homeostasis in biological systems (Purushotham et al., 2009). Since the liver mediates key aspects of lipid metabolism, amino acid metabolism, steroid metabolism, and many other metabolic pathways, chronic metabolic disorders in the liver would lead to the incidence and development of liver damages, such as non-alcoholic fatty liver diseases ranging from steatosis to non-alcoholic steatohepatitis, fibrosis and cirrhosis (Byrne et al., 2009; Lee et al., 2003; Purushotham et al., 2009; Thomas et al., 2008). Notably, non-alcoholic fatty liver diseases increase the risk of other more serious metabolic diseases, such as diabetes, cardiovascular disease and cancer (Gaggini et al., 2013; Kim et al., 2018; Smith and Adams, 2011). Therefore, insights into metabolic disturbances in the liver/hepatocytes benefit the understanding of molecular pathology of metabolic diseases and relevant potential therapeutic target discovery. Hepatic damages due to PM2.5 exposure have been attracting more and more attentions (Liu et al., 2017; Xu et al., 2018; Zheng et al., 2013; Zheng et al., 2015). Exposure to PM2.5 for 10 weeks triggered impaired glucose intolerance, glycogen storage and insulin resistance, activated inflammatory response pathways mediated by Toll-like receptor 4, nuclear factor kappa B and c-Jun N-terminal kinase, while repressed the expression of peroxisome proliferator-activated receptor (PPAR) α/γ and the signaling mediated by insulin receptor substrate 1 in the liver, leading to hepatic inflammation, steatosis and fibrosis in male C57BL/6 mice (Zheng et al., 2013). Moreover, oxidative stress and inflammation in the liver induced by 24-wek PM2.5 inhalation promoted abnormal hepatic function and lipid accumulation in mice (Xu et al., 2018). However, effects of PM2.5 exposure on hepatocyte/liver metabolism remain unclear. Metabolomics, aiming to comprehensive qualitative and quantitative analysis of metabolites in biological systems, have been used to discover metabolic disturbances induced by PM2.5 exposure (Cui et al., 2019; Fiehn, 2002; Liang et al., 2018; Zhang et al., 2017; Zhang et al., 2019). Nonetheless, comprehensive metabolic responses of hepatocytes/livers to PM2.5 exposure are still undiscovered. Therefore, an untargeted metabolomics approach based on liquid chromatography–mass spectrometry (LC-MS) was employed to present the comprehensive metabolic responses of HepG2 cells to liposoluble
extracts of PM2.5 samples (LE) and to discover potential therapeutic targets for PM2.5-induced metabolic disturbances in metabolic diseases. 2. Materials and methods 2.1. Materials Ammonium bicarbonate (≥99.0%), HPLC-grade methanol, dichloromethane, acetonitrile and formic acid were purchased from SigmaAldrich (Shanghai, China). Ultrapure water was prepared employing a Milli-Q system (Millipore, USA). HepG2 cells were ordered from the Cell Bank of Type Culture Collection, Chinese Academy of Sciences (Shanghai, China). 2.2. PM2.5 sample collection and liposoluble component extraction PM2.5 samples were kindly provided from Prof. Chen (Institute of Urban Environment, Chinese Academy of Sciences). The samples were collected in Gulou Campus in Nanjing University (China) during winter. The liposoluble components were extracted with dichloromethane/ methanol (v/v, 2:1) via accelerated solvent extraction (Dionex 350). Detailed sample information and the extraction could be referred to the relevant study (Hong et al., 2017). LE was dried and redissolved in dimethyl sulfoxide (DMSO) as the stock solution for subsequent cell experiments. The dried liposoluble extracts of blank sampling paper were also reconstituted in DMSO as the control. 2.3. Cell culture and treatment HepG2 cells were cultured in Minimum Essential Medium (Gibco, NY, USA) added with 10% Fetal Bovine Serum (HyClone, Victoria, Australia), 1.5 g/L NaHCO3 and 0.11 g/L sodium pyruvate in a humidified incubator with 5% CO2 at 37 °C. Since that a high-risk individual exposure to 79 μg/m3 of PM2.5 over a 24-h period was estimated to be comparable to in vitro 0.2–20 μg/cm2, and that seasonal average PM2.5 concentrations in Nanjing during the winter sampling period was 78 ± 42 μg/m3 (Li et al., 2003; Nie et al., 2018). Additionally, during haze events in winter and spring, the daily average PM2.5 concentration in major cities in China was above 75 μg/m3 (Chinese pollution standard), and it exceeded 100 μg/m3 in some cities, such as Beijing and Xi'an (Huang et al., 2014). We previously found that HepG2 cell viability was significantly reduced to 69.28% compared to the control after 25 μg/cm2 of LE exposure (Ding et al., 2019). Accordingly, HepG2 cells were exposed to LE at 25 μg/cm2 for 72 h to investigate consequent metabolic responses. 2.4. Cell sample preparation for metabolomics analysis Following cell culture and treatment, the medium was discarded. Cells were immediately quenched in liquid nitrogen and stored at −80 °C prior to PBS washing. The cell samples was added with 1000 μL of ice-cold methanol/water (v/v, 4:1) and harvested with scrapers. The sample was drawn into a centrifuge tube and vortexed for 1.0 min. After cell-sample centrifugation at 13, 000 rpm for 15 min at 4 °C, 600 μL of the supernatant was drawn and vacuum dried in a
876
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
SpeedVac concentrator (Thermo Scientific, USA). To monitor and assess the stability and reliability of the metabolomics approach, the residual supernatant from all samples was withdrawn and mixed evenly, then distributed into 600-μL aliquots as quality control (QC) samples. One QC sample was inserted every 3 analytical samples and treated in the same way as analytical samples during vacuum drying, resuspension, instrumental analysis and data processing. 2.5. Instrumental analysis for the metabolomics approach The dried cell extracts were resuspended with 200 μL MeOH/water (v:v, 1:1) solution, and placed into glass vials for LC-MS analysis. Metabolic profiles were obtained by employing a Ultimate 3000 UHPLC system coupled to Q Exactive quadruple-Orbitrap high resolution mass spectrometer (Thermo Scientific, San Jose, USA). Metabolite separation was achieved on an Acquity BEH C18 column (1.7 μm, 2.1 × 100 mm, Waters Co., USA). Two and five μL of the resuspended metabolite extracts were injected in positive and negative electrospray ionization mode, respectively. Mobile phases A (water containing 0.1% formic acid) and B (acetonitrile containing 0.1% formic acid) were employed in positive mode with the gradient elution: initial 2% B increased to 98% in 12 min prior to 3 min column wash and equilibration. On the other hand, Mobile phases C (5 mM ammonium bicarbonate in water) and D (5 mM ammonium bicarbonate in methanol/acetonitrile mixture, v/v = 20/80) were used in negative mode, with initial 98% C linearly decreased to 0% in 12 min. Metabolites were eluted at a flow rate of 400 μL/min and directly transferred into the mass spectrometer. Parameters of the ion source were set as follows: ion transfer capillary temperature, 320 °C; heater temperature, 375 °C; sheath gas, 45 arbitrary units; auxiliary gas, 10 arbitrary units; ionization voltage, 4.0 kV (positive) and 3.5 kV (negative); S-lens level, 50%. The mass scan range was acquired between 100 and 1000 m/z. Mass resolution of the Orbitrap was set to 70,000 (FWHM) with 1e6 automatic gain control target and 100-ms maximum ion injection time. Full scan-ddMS2 data of QC samples were repeatedly acquired to obtain high-resolution MS2 spectra for structural verification of metabolites. Key parameters of ddMS2 mode were set as follows: mass resolution (FWHM),17,500; automatic gain control target, 1e5; loop count, 7; precursor isolation window, 1.0 Da; stepped normalized collision energy, 15, 30 and 45%; dynamic exclusion, 5 s. 2.6. Data preprocessing for the metabolomics approach Raw spectral data were preprocessed by Compound Discoverer 2.1 (Thermo Scientific, San Jose) for peak alignment, ion detection and deconvolution, gap re-filling, artifacts exclusion, element prediction
and structural elucidation. Metabolites were firstly identified by searching a local endogenous lipid and metabolite database within 3 ppm mass tolerance, and further verified through searching against an in-house built MS2 spectral library and mzCloud library (www. mzcloud.org). The spectral similarity between the retrieved and reference spectra was calculated by a proprietary Highchem HighRes algorithm, and the similarity score threshold was 50. All identified metabolites were exported into TraceFinder software (Thermo Scientific, San Jose) for peak integration and manual check. Ion peaks of metabolites in positive mode were divided by the total ion current and multiplied by 1 × 109. On the other hand, ion peaks of metabolites in negative mode were divided by the total ion current and multiplied by 1 × 1010. Subsequently, the data in positive and negative mode were merged and applied for statistical analysis. 2.7. Statistical analysis Principal component analysis and pathway analysis were conducted via MetaboAnalyst 3.0 (Xia et al., 2015). Significant differences in metabolite levels among different groups were assessed by a nonparametric test (two-tailed Mann-Whitney U test) using MeV 4.9.0 (Saeed et al., 2006). The significant level was further adjusted by the false discovery rate based on Benjamini-Hochberg Correction using MeV 4.9.0, and the false discovery rate was set to 0.12. The level of statistical significance was 0.05. After compound name standardization and data normalization (mean-centered and divided by the standard deviation of each variable), “Global test” and “Relative-betweeness Centrality” algorithms were employed for pathway enrichment analysis and pathway topology analysis, respectively. Subsequently, KEGG metabolic pathways related to Homo sapiens were used for pathways analysis. The heat map was plotted by MeV 4.9.0. 3. Results 3.1. Global metabolic responses of HepG2 cells in response to LE exposure Distribution of QC samples and relative standard deviation of metabolites in QC samples were utilized to evaluate the analytical performance. It was clear in the principal component analysis score plot that 3 QC samples were clustered together (Fig. 1A). Moreover, 279, 285 and 289 metabolites, separately accounting for 94.3, 96.3 and 97.6% of total identified metabolites, had relative standard deviation below 15, 20 and 30%, respectively (Fig. 1B). The data demonstrated high stability and repeatability of the metabolomics approach (Begley et al., 2009; Ye et al., 2014).
Fig. 1. Analytical performance of the metabolomics approach and changes in the metabolic profiling of HepG2 cells induced by LE exposure. (A) Sample distribution in the score plot of principal component analysis. (B) Relative standard deviation distribution of metabolites in the QC samples.
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
We found that metabolic profiles of HepG2 cells treated with LE were significantly different from those of the control group (Fig. 1A). Subsequently, a two-tailed Mann-Whitney U test was utilized to discover differential metabolites (p b 0.05) between the control and exposure group. Among 296 annotated metabolites, 123 metabolites were discovered to be significantly disturbed in HepG2 cells in response to LE exposure, including 10 metabolites involved in carbohydrate metabolism, 39 metabolites involved in amino acid and oligopeptide
877
metabolism, 54 metabolites involved in lipid metabolism, 16 metabolites involved in purine and pyrimidine nucleotide metabolism, and 4 metabolites involved in other metabolic pathways (Table S1 and Fig. 2). Pathways analysis revealed that 46 metabolic pathways were significantly altered in HepG2 cells in response to LE exposure (Table S2). Notably, levels of free fatty acids (FFAs) and most metabolites involved in nucleoside metabolism were all significantly increased, while levels of most amino acids were decreased in HepG2 cells treated
Fig. 2. Heat map of metabolic responses of HepG2 cells to LE exposure. n = 5 per group. After unit variance scaling, levels of metabolites were used to plot the heat map.
878
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
with LE (Fig. 2). Detailed metabolic changes in HepG2 cells treated with LE are provided below.
FFA transport into mitochondria and subsequent oxidation degradation in HepG2 cells in response to LE exposure.
3.2. LE exposure induced FFA accumulation and significant changes in acylcarnitines in HepG2 cells
3.3. LE exposure induced significant changes in phospholipid, sphingolipid and steroid metabolism in HepG2 cells
Significant changes in FFAs and acylcarnitines in HepG2 cells were induced by LE exposure (Figs. 2 and 3). Levels of short-chain, medium-chain and long-chain FFAs, such as FFA(6:0), FFA(12:1), FFA (14:1), FFA(20:5), FFA(22:6) and FFA(24:6), were all significantly increased in HepG2 cells in response to LE exposure (Fig. 3A). In addition, the level of hydroxylated FFA(16:0) was also significantly increased in HepG2 cells treated with LE. FFAs participate in the synthesis of other lipids (such as triglycerides, phospholipids and bioactive signaling lipids) and energy homeostasis via mitochondrial oxidative metabolism. We found that levels of short-chained acylcarnitines (b5 carbons, such as carnitine, acetylcarnitine, propionylcarnitine, succinylcarnitine and hydroxybutyrylcarnitine) were significantly decreased, however, levels of medium-chained acylcarnitines (6 to 12 carbons), including hexanolycarnitine, acylcarnitine(8:0), acylcarnitine(10:1) and acylcarnitine(12:1), were significantly increased in HepG2 cells treated with LE (Fig. 3B). Changes in acylcarnitine levels suggested disturbed
We found significant changes in phospholipid, sphingolipid and steroid metabolism in HepG2 cells in response to LE exposure (Figs. 2 and 4, Table S1). Levels of acetylcholine and glycerophosphocholine were significantly decreased, while the level of phosphorylcholine was significantly increased in HepG2 cells treated with LE (Fig. 4). Besides, levels of lysophosphatidylcholines (LPCs), such as LPC(16:0) sn-1, LPC(16:0) sn-2, LPC(18:2) sn-1 and LPC(18:2) sn-2, were all significantly decreased in HepG2 cells treated with LE (Fig. 4). Moreover, levels of most lysophosphatidylinositols (LPIs), such as LPI(16:0) sn-1, LPI (16:0) sn-2, LPI(18:1) sn-1 and LPI(18:1) sn-2, were significantly increased in HepG2 cells treated with LE (Fig. 4). Significant changes in levels of lysophosphatidylethanolamines (LPEs) and lysophosphatidylglycerols (LPGs) were also observed in HepG2 cells treated with LE (Fig. 4). These data demonstrated that LE exposure induced significant changes in phospholipids, especially phosphatidylcholines and phosphatidylinositols. On the other hand, levels of sphingosine
Fig. 3. LE exposure induced FFA accumulation (A) and significant changes in acylcarnitines (B) in HepG2 cells. Data were represented as the mean + SD. *, p b 0.05, **, p b 0.01, two-tailed Mann-Whitney U test. n = 5 per group.
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
879
Fig. 4. LE exposure induced significant changes in phospholipid metabolism in HepG2 cells. LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; LPI, lysophosphatidylinositol; LPG, lysophosphatidylglycerol; PC, phosphatidylcholine; PE, phosphatidylethanolamine; DG, diglyceride; CDP-DG, 1,2-diacyl-sn-glycero-3-cytidine-5′diphosphate. Data were represented as the mean + SD. *, p b 0.05, **, p b 0.01, two-tailed Mann-Whitney U test. n = 5 per group.
(d16:0), sphingosine(d20:1) and phytosphingosine were significantly decreased, while the level of 7-dehydrodesmosterol was significantly increased in HepG2 cells, which illustrated disturbed sphingolipid and steroid metabolism induced by LE exposure (Fig. 2 and Table S1).
3.4. LE exposure induced significant changes in carbohydrate metabolism in HepG2 cells Levels of fructose 1,6-bisphosphate, 6-phosphogluconic acid, 3phosphoglyceric acid and phosphoenolpyruvic acid were significantly disturbed in HepG2 cells treated with LE, which indicated that LE exposure induced metabolic disturbances in glycolysis/gluconeogenesis, pentose phosphate pathway and pyruvate metabolism in HepG2 cells (Fig. 5). Besides, levels of cis-aconitic acid and citric acid (or isocitric acid) were significantly decreased, while levels of succinic acid and 2hydroxyglutarate were significantly increased in HepG2 cells treated with LE (Fig. 5). The data demonstrated that LE exposure triggered disorders of tricarboxylic acid (TCA) cycle in HepG2 cells. Additionally, metabolites related to amino sugar and nucleotide sugar metabolism, such as uridine diphosphate glucuronic acid and UDP-N-acetylglucosamine, were also disturbed in HepG2 cells treated with LE (Fig. 5).
3.5. LE exposure induced significant changes in amino acid metabolism in HepG2 cells We also observed significant changes in amino acid metabolism in HepG2 cells in response to LE exposure (Fig. 5). Levels of metabolites involved in alanine, aspartate and glutamate metabolism (such as glutamine, glutamate, aspartate and N-Acetylaspartylglutamic acid) and arginine and proline metabolism (i.e., proline, 4-hydroxyproline, spermidine and N1-acetylspermidine) were all significantly decreased in HepG2 cells treated with LE (Fig. 5). On the contrary, levels of metabolites involved in cysteine and methionine metabolism (such as Methionine, methionine sulfoxide and S-adenosyl-L-methionine), glutathione metabolism (i.e., glu-cys, cys-gly, glutathione and Smethyl glutathione) and leucine metabolism (such as leucine and 3hydroxy-3-methylglutarate) were significantly increased in HepG2 cells treated with LE (Fig. 5). Changes in cysteine and methionine metabolism and glutathione metabolism indicated the increase in methylation and oxidative stress in HepG2 cells induced by LE exposure. Moreover, levels of most metabolites in aromatic amino acid metabolism (such as tyrosine, hydroxyphenyllactic acid, tryptophan, kynurenine and 5-hydroxy-L-tryptophan), glycine, serine and threonine metabolism (such as serine, threonine and betaine), histidine
880
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
Fig. 5. LE exposure induced significant changes in carbohydrate and amino acid metabolism in HepG2 cells. Data were represented as the mean + SD. *, p b 0.05, **, p b 0.01, two-tailed Mann-Whitney U test. n = 5 per group.
metabolism, and lysine metabolism were significantly decreased in HepG2 cells treated with LE (Fig. 5 and Table S1). 3.6. LE exposure induced significant changes in purine and pyrimidine nucleotide metabolism in HepG2 cells Significant changes in purine and pyrimidine nucleotide metabolism occurred in HepG2 cells in response to LE exposure (Fig. 6). Levels of
most metabolites involved in adenine and guanine nucleotide metabolism, such as xanthine, adenosine monophosphate, guanine, guanosine and guanosine monophosphate, were significantly increased in HepG2 cells treated with LE, except adenosine triphosphate and guanosine triphosphate (Fig. 6A). We observed that levels of uracil, uridine and uridine 5′-monophosphate involved in uridine nucleotide metabolism were significantly increased, while levels of uridine 5′-diphosphate and uridine triphosphate were significantly decreased in HepG2 cells
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
881
Fig. 6. LE exposure induced significant changes in purine (A) and pyrimidine (B) nucleotide metabolism in HepG2 cells. Data were represented as the mean + SD. *, p b 0.05, **, p b 0.01, two-tailed Mann-Whitney U test. n = 5 per group.
treated with LE (Fig. 6B). Additionally, levels of cytidine 5′-diphosphate and cytidine triphosphate involved in cytidine nucleotide metabolism were significantly decreased in HepG2 cells treated with LE (Fig. 6B). These data suggested disturbed energy production and/or nucleic acid synthesis in HepG2 cells treated with LE. 4. Discussion Owing to high separation efficiency, resolution, sensitivity, specificity and throughput as well as the development of various derivatization technologies, chromatography-mass spectrometry has become the mainstream technology for the development and application of metabolomics, especially LC-MS. However, there is currently no single technology that could cover all the metabolites in biological systems, and each technology has its own technical advantages and disadvantages. Compared to gas chromatography–mass spectrometry commonly employed in metabolomics analysis, time-consuming chemical derivatization produces are usually not needed for LC-MS analysis. Additionally, metabolites difficult to volatilize or to be derivatized can be detected by LC-MS, but not by gas chromatography–mass spectrometry, such as carnitines, cholines, and lysophospholipids. However, some types of polar metabolites, such as saccharides and short-chain organic acids, cannot be effectively separated by LC-MS, but can be effectively separated by gas chromatography–mass spectrometry after proper derivatization. In this study, 50% methanol solution was used to dissolve the dried sample prior to LC-MS analysis to reduce the solvent effect, which improved the peak shape of metabolites and subsequent analytical reproducibility and metabolite coverage. On the other hand, there are N1000 metabolites with accurate parent ions and their fragments as well as retention times in our self-built spectral library. Totally, 296 metabolites were accurately identified and quantified, whereas
metabolites insoluble in 50% methanol solution, such as triglycerides, cannot be detected in this study. Non-polar metabolites insoluble in 50% methanol solution can be effectively determined using either untargeted metabolomics or lipidomics based on LC-MS after dissolution in methanol or other non-polar solvents. Accumulated data revealed that PM2.5 exposure increased the incidence of metabolic diseases, and that metabolic disturbances in the liver/hepatocytes promoted the risk of metabolic diseases. However, metabolic disturbances in the liver/hepatocytes in response to PM2.5 exposure remain unclear. Accordingly, an untargeted metabolomics approach based on high-resolution LC-MS was employed to uncover metabolic responses of HepG2 cells to LE exposure. We found that LE exposure induced significant disturbances in many metabolic pathways, especially lipid, amino acid and nucleotide metabolism. Notably, FFA accumulated in HepG2 cells in response to LE exposure. Furthermore, disturbed acylcarnitine shuttle system, responsible for FFA transport for mitochondrial oxidation and working as an acyl-CoA pool, was also observed in HepG2 cells treated with LE. The decreased levels of short-chain acylcarnitines, including acetylcarnitine, propionylcarnitine, succinylcarnitine and hydroxybutyrylcarnitine, and the increased levels of medium-chain acylcarnitines, including hexanolycarnitine, acylcarnitine(8:0), acylcarnitine(10:1) and acylcarnitine(12:1) indicated incomplete FFA oxidation, accelerated FFA oxidation and lipid overload in mitochondrial in HepG2 cells treated with LE (Wolf et al., 2013; Yu et al., 2012). Medium-chain acylcarnitine accumulation might be due to an imbalance between FFA influx and oxidation (Overmyer et al., 2015). AhR (aryl hydrocarbon receptor) as the receptor of polycyclic aromatic hydrocarbons (major components of LE) could be activated after LE exposure, and then stimulated CD36 expression and resultant increases in FFA uptake and accumulation (Matsumoto et al., 2007; Yao et al., 2016). Excessive FFAs could be
882
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
stored in neutral lipids (such as triglycerides) in HepG2 cells in response to LE exposure (Ding et al., 2019). Moreover, AhR activation induced by LE exposure could increase the activity of PPARα via up-regulation of genes involved in circadian rhythmicity, thus promoting FFA transport for mitochondrial oxidation via carnitine O-palmitoyltransferase 1 (Wang et al., 2011). Therefore, accumulation of FFAs and mediumchain acylcarnitines was probably due to accelerated FFA influx and/or incomplete oxidation induced by disturbed AhR and PPAR signaling pathways in HepG2 cells after LE exposure. Short-chained acylcarnitines, especially carnitine and propionylcarnitine could improve metabolic functions (Ramsay and Zammit, 2004). For example, oral administration of propionyl-carnitine attenuated abnormalities of cardiac function and reduced plasma lipids in streptozotocin-diabetic rats (Terada et al., 1998). Additionally, both the carnitine and propionyl-carnitines are able to reduce mitochondrial acetyl-CoA, leading to the activation of pyruvate dehydrogenase and resultant increase in pyruvate oxidation (Ramsay and Zammit, 2004). However, levels of short-chained acylcarnitines were decreased in our study, which was not conducive to the improvement of metabolic function in HepG2 cells treated with LE (Ramsay and Zammit, 2004). Moreover, we also found a significant increase in the level of 3phosphoglyceric acid and phosphoenolpyruvic acid, precursors for pyruvate synthesis, which indicated disturbed pyruvate metabolism in HepG2 cells treated with LE. The intermediate products of FFA oxidation could enter the TCA cycle. As observed, levels of citrate (or isocitrate) and aconitate were decreased in HepG2 cells treated with LE, which suggested less materials for the oxidative decarboxylation to produce energy. Furthermore, we found significant decreases in levels of adenosine triphosphate, guanosine triphosphate, uridine triphosphate and cytidine triphosphate in HepG2 cells treated with LE. PM exposure induced a decrease in CCAA T/enhancer binding protein alpha, which led to a decline in the mRNA expression of genes involved in NADH ubiquinone dehydrogenase (NDUFA2 and NDUFS4), NADH dehydrogenase in mitochondrial complex I (NDUFA1 and NDUFC2) and adenosine triphosphate synthase in mitochondrial complex V (ATP5H), resulting in the suppression of mitochondrial oxidative phosphorylation and adenosine triphosphate synthesis (Li et al., 2017). These data indicated that LE exposure triggered a decrease in energy generation from TCA cycle and oxidative phosphorylation in HepG2 cells. Moreover, we found succinate accumulation in HepG2 cells treated with LE. Excessive succinate production activated hypoxia-inducible factor-1α expression, which accelerated glycolysis and the synthesis of lactate and interleukin 1β, aggravating the inflammatory status in HepG2 cells induced by PM2.5 (Appari et al., 2018; Xu et al., 2018). As revealed in A549 cells, exposure to water-insoluble and total fractions of Ottawa urban dust (EHC-93) triggered an increase in the release of lactate dehydrogenase and a decrease in cell proliferation and adenosine triphosphate production (Vuong et al., 2017). Of note, both 2-hydroxyglutarate and succinate accumulation could inhibit the activity of 2-ketoglutarate-dependent dioxygenases, including the ten-eleven translocation family of cytosine hydroxymethylases, prolylhydroxylases and the Jumonji family of histone demethylases, thus leading to hypermethylation in HepG2 cells treated with LE (Raffel et al., 2017; Sullivan et al., 2016). Moreover, we found significant increases in levels of methionine and S-adenosyl-L-methionine, which indicated more methyl donors available for the hypermethylation in HepG2 cells treated with LE. Genome-wide DNA methylation analysis of fasting venous blood samples from students revealed that methylation levels were changed with the concentration of PM2.5 exposure (24-h averages: 53.1 vs. 24.3 μg/m3) at 49 CpG loci, where 31 CpG sites were annotated to specific genes, and that DNA methylation of the annotated genes was increased with elevated PM2.5 exposure, which was involved in lipid and glucose metabolism, insulin resistant, oxidative stress, platelet activation, inflammation and cell survival and apoptosis (Li et al., 2018). These data indicated an import role of methylation in the molecular toxicology of PM2.5 exposure.
The glycolytic intermediates could be utilized for the synthesis of nucleotides, amino acids and substrates in pentose phosphate pathway to generate NADPH. Since that PM exposure induced decreases cell proliferation and adenosine triphosphate production, and that most metabolites involved in amino acid metabolism were decreased in HepG2 cells treated with LE in this study, the glycolytic intermediates were probably utilized to provide substrates in pentose phosphate pathway for NADPH generation to defense against reactive oxygen species (ROS) in HepG2 cells after LE exposure，in addition to producing lactate. The decreased activity of mitochondrial electron transport chain in our study could lead to mitochondrial ROS generation in HepG2 cells treated with LE (Appari et al., 2018). We found significant increased levels of glutathione and its precursors (Glu-Cys and Cys-Gly), lipoic acid and methionine sulfoxide (a biomarker of oxidative stress), which suggested that the synthesis of glutathione and ROS was increased, and that antioxidant system was activated in HepG2 cells treated with LE. PM2.5 exposure for 24 weeks at the concentration of 230 ± 2.5 μg/m3 demonstrated significant increases in oxidative stress, inflammatory response, dyslipidemia (such as accumulation of triglycerides, total cholesterol and FFAs) and hepatic dysfunctions in mice (Xu et al., 2018). On the other hand, blocking oxidative stress and NF-κB signaling pathway with pyrrolidine dithiocarbamate and N-acetyl-L-cysteine, respectively, reduced expression levels of genes related to oxidative stress, inflammation and lipid accumulation in vitro, suggesting PM2.5induced oxidative stress and inflammation as major promoters of dyslipidemia in mice liver (Xu et al., 2018). We also shown that levels of most amino acids were reduced in HepG2 cells treated with LE, which indicated less amino acids available for the generation of energy, proteins, nucleotides and lipids to sustain cell proliferation and growth. PPARs governs the expression of many genes involved in amino acid metabolism, including those related to transamination, deamination, amino acid inter-conversions, generation of amino acid derivatives and alpha-keto acid oxidation (Kersten et al., 2001). Accordingly, decreases in levels of most amino acids were probably related to disturbances in AhR and PPAR signaling induced by LE exposure in HepG2 cells (Matsumoto et al., 2007; Wang et al., 2011; Yao et al., 2016). 5. Conclusions An untargeted metabolomics approach based on LC-MS was applied to characterize metabolic responses of HepG2 cells to LE exposure. We discovered that LE exposure induced significant changes in many metabolic pathways (including lipid, amino acid and nucleotide metabolism,
Fig. 7. Overview of metabolic responses of HepG2 cells to LE exposure. Red/blue fonts: most relevant metabolites significantly increased/decreased in HepG2 cells treated with LE. Black fonts: metabolites not detected or not significantly altered in HepG2 cells treated with LE. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
and TCA cycle) in HepG2 cells (Fig. 7). Notably, LE exposure induced accumulation of FFAs and medium-chained acylcarnitines (6–12 carbons), while decreased levels of short-chained acylcarnitines (b5 carbons) in HepG2 cells. Besides, levels of citrate/isocitrate and aconitate were decreased, but 2-hydroxyglutate and succinate accumulated in HepG2 cells treated with LE. We also found that levels of adenosine triphosphate, guanosine triphosphate, uridine triphosphate and cytidine triphosphate were decreased, however, those of adenosine monophosphate, guanosine monophosphate, purines and pyrimidines were increased in HepG2 cells treated with LE. Moreover, levels of glutathione, Glu-Cys, Cys-Gly, lipoic acid, methionine sulfoxide, methionine and S-adenosyl-L-methionine were increased, while those of most amino acids were decreased in HepG2 cells treated with LE. These data demonstrated that LE exposure induced accumulation of FAAs and oncometabolites (2-hydroxyglutate and succinate), mitochondrial dysfunctions characterized by incomplete FFA oxidation and reduced energy supply from TCA cycle and oxidative phosphorylation, disturbances in methylation and redox homeostasis, and the inhibition of most amino acid metabolism in HepG2 cells. Above metabolic disturbances provide potential therapeutic targets for treating PM2.5-induced injury and diseases. To the best of our knowledge, this study provides the first demonstration that LE exposure triggered accumulation of medium-chain acylcarnitines, oncometabolites, purines and pyrimidines in HepG2 cells. Author contributions G.Y., D.X., Y.L. and S.D. conceived and designed this study. G.Y. performed the metabolomics analysis and statistical analysis, interpreted the data, wrote and revised the manuscript. H.G. and Y.C. participated in the cell sample preparation for metabolomics analysis. Z.W. conducted the instrumental analysis and subsequent data processing for metabolomics analysis. J.C. provided the PM2.5 sample. Declaration of Competing Interest The authors declare that they have no conflict of interest. Acknowledgements This work was supported by the National Natural Science Foundation of China [grant numbers 21507128, 41390240, 21777158, 21477124 and 21677140]; the Natural Science Foundation of Fujian Province, China [grant numbers 2018J01020 and 2017J01028]; the Key Laboratory of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences [grant number J008]; and the Department of Science and Technology of Fujian Province, China [grant number 2016T3005]. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2019.07.192. References Appari, M., Channon, K.M., McNeill, E., 2018. Metabolic regulation of adipose tissue macrophage function in obesity and diabetes. Antioxid. Redox Signal. 29, 297–312. Begley, P., Francis-McIntyre, S., Dunn, W.B., Broadhurst, D.I., Halsall, A., Tseng, A., et al., 2009. Development and performance of a gas chromatography-time-of-flight mass spectrometry analysis for large-scale nontargeted metabolomic studies of human serum. Anal. Chem. 81, 7038–7046. Byrne, C.D., Olufadi, R., Bruce, K.D., Cagampang, F.R., Ahmed, M.H., 2009. Metabolic disturbances in non-alcoholic fatty liver disease. Clin. Sci. 116, 539–564. Cui, J., Fu, Y., Lu, R., Bi, Y., Zhang, L., Zhang, C., et al., 2019. Metabolomics analysis explores the rescue to neurobehavioral disorder induced by maternal PM2.5 exposure in mice. Ecotox. Environ. Safe. 169, 687–695.
883
Ding, D., Ye, G., Lin, Y., Lu, Y., Zhang, H., Zhang, X., et al., 2019. MicroRNA-26a-CD36 signaling pathway: pivotal role in lipid accumulation in hepatocytes induced by PM2.5 liposoluble extracts. Environ. Pollut. 248, 269–278. Fiehn, O., 2002. Metabolomics - the link between genotypes and phenotypes. Plant Mol. Biol. 48, 155–171. Gaggini, M., Morelli, M., Buzzigoli, E., DeFronzo, R.A., Bugianesi, E., Gastaldelli, A., 2013. Non-alcoholic fatty liver disease (NAFLD) and its connection with insulin resistance, dyslipidemia, atherosclerosis and coronary heart disease. Nutrients 5, 1544–1560. Hong, Z., Hong, Y., Zhang, H., Chen, J., Xu, L., Deng, J., et al., 2017. Pollution characteristics and source apportionment of PM2.5-bound n-alkanes in the Yangtze River Delta, China. Aerosol Air Qual. Res. 17, 1985–1998. Huang, R.-J., Zhang, Y., Bozzetti, C., Ho, K.-F., Cao, J.-J., Han, Y., et al., 2014. High secondary aerosol contribution to particulate pollution during haze events in China. Nature 514, 218–222. Kersten, S., Mandard, S., Escher, P., Gonzalez, F.J., Tafuri, S., Desvergne, B., et al., 2001. The peroxisome proliferator-activated receptor alpha regulates amino acid metabolism. FASEB J. 15, 1971–1978. Kim, G.A., Lee, H.C., Choe, J., Kim, M.J., Lee, M.J., Chang, H.S., et al., 2018. Association between non-alcoholic fatty liver disease and cancer incidence rate. J. Hepatol. 68, 140–146. Lee, C.-H., Evans, R.M., Olson, P., 2003. Minireview: lipid metabolism, metabolic diseases, and peroxisome proliferator-activated receptors. Endocrinology 144, 2201–2207. Li, N., Hao, M., Phalen, R.F., Hinds, W.C., Nel, A.E., 2003. Particulate air pollutants and asthma. A paradigm for the role of oxidative stress in PM-induced adverse health effects. Clin. Immunol. 109, 250–265. Li, X., Yang, H., Sun, H., Lu, R., Zhang, C., Gao, N., et al., 2017. Taurine ameliorates particulate matter-induced emphysema by switching on mitochondrial NADH dehydrogenase genes. Proc. Natl. Acad. Sci. U. S. A. 114, E9655–E9664. Li, H., Chen, R., Cai, J., Cui, X., Huang, N., Kan, H., 2018. Short-term exposure to fine particulate air pollution and genome-wide DNA methylation: a randomized, double-blind, crossover trial. Environ. Int. 120, 130–136. Liang, D., Moutinho, J.L., Golan, R., Yu, T., Ladva, C.N., Niedzwiecki, M., et al., 2018. Use of high-resolution metabolomics for the identification of metabolic signals associated with traffic-related air pollution. Environ. Int. 120, 145–154. Liu, C., Xu, X., Bai, Y., Zhong, J., Wang, A., Sun, L., et al., 2017. Particulate air pollution mediated effects on insulin resistance in mice are independent of CCR2. Part. Fibre Toxicol. 14, 017–0187. Matsumoto, Y., Ide, F., Kishi, R., Akutagawa, T., Sakai, S., Nakamura, M., et al., 2007. Aryl hydrocarbon receptor plays a significant role in mediating airborne particulateinduced carcinogenesis in mice. Environ. Sci. Technol. 41, 3775–3780. Nie, D., Chen, M., Wu, Y., Ge, X., Hu, J., Zhang, K., et al., 2018. Characterization of fine particulate matter and associated health burden in Nanjing. Int. J. Environ. Res. Public Health 15. Overmyer, Katherine A., Evans Charles, R., Qi Nathan, R., Minogue Catherine, E., Carson Joshua, J., Chermside-Scabbo Christopher, J., et al., 2015. Maximal oxidative capacity during exercise is associated with skeletal muscle fuel selection and dynamic changes in mitochondrial protein acetylation. Cell Metab. 21, 468–478. Pearson, J.F., Bachireddy, C., Shyamprasad, S., Goldfine, A.B., Brownstein, J.S., 2010. Association between fine particulate matter and diabetes prevalence in the U.S. Diabetes Care 33, 2196–2201. Pimpin, L., Retat, L., Fecht, D., de Preux, L., Sassi, F., Gulliver, J., et al., 2018. Estimating the costs of air pollution to the National Health Service and social care: an assessment and forecast up to 2035. PLoS Med. 15, e1002602. Purushotham, A., Schug, T.T., Xu, Q., Surapureddi, S., Guo, X., Li, X., 2009. Hepatocytespecific deletion of SIRT1 alters fatty acid metabolism and results in hepatic steatosis and inflammation. Cell Metab. 9, 327–338. Raffel, S., Falcone, M., Kneisel, N., Hansson, J., Wang, W., Lutz, C., et al., 2017. BCAT1 restricts alphaKG levels in AML stem cells leading to IDHmut-like DNA hypermethylation. Nature 551, 384–388. Rajagopalan, S., Al-Kindi, S.G., Brook, R.D., 2018. Air pollution and cardiovascular disease: JACC state-of-the-art review. J. Am. Coll. Cardiol. 72, 2054–2070. Ramsay, R.R., Zammit, V.A., 2004. Carnitine acyltransferases and their influence on CoA pools in health and disease. Mol. Asp. Med. 25, 475–493. Saeed, A.I., Bhagabati, N.K., Braisted, J.C., Liang, W., Sharov, V., Howe, E.A., et al., 2006. TM4 microarray software suite. Method Enzymol 411, 134–193. Smith, B.W., Adams, L.A., 2011. Nonalcoholic fatty liver disease and diabetes mellitus: pathogenesis and treatment. Nat. Rev. Endocrinol. 7, 456–465. Sullivan, L.B., Gui, D.Y., Vander Heiden, M.G., 2016. Altered metabolite levels in cancer: implications for tumour biology and cancer therapy. Nat. Rev. Cancer 16, 680–693. Terada, R., Matsubara, T., Koh, N., Nakamura, J., Hotta, N., 1998. Effects of propionyl-Lcarnitine on cardiac dysfunction in streptozotocin-diabetic rats. Eur. J. Pharmacol. 357, 185–191. Thomas, C., Pellicciari, R., Pruzanski, M., Auwerx, J., Schoonjans, K., 2008. Targeting bileacid signalling for metabolic diseases. Nat. Rev. Drug Discov. 7, 678. Turner, M.C., Krewski, D., Pope 3rd, C.A., Chen, Y., Gapstur, S.M., Thun, M.J., 2011. Longterm ambient fine particulate matter air pollution and lung cancer in a large cohort of never-smokers. Am. J. Resp. Crit. Care. 184, 1374–1381. Vuong, N.Q., Breznan, D., Goegan, P., O'Brien, J.S., Williams, A., Karthikeyan, S., et al., 2017. In vitro toxicoproteomic analysis of A549 human lung epithelial cells exposed to urban air particulate matter and its water-soluble and insoluble fractions. Part. Fibre Toxicol. 14, 017–0220. Wang, C., Xu, C.X., Krager, S.L., Bottum, K.M., Liao, D.F., Tischkau, S.A., 2011. Aryl hydrocarbon receptor deficiency enhances insulin sensitivity and reduces PPAR-alpha pathway activity in mice. Environ. Health Perspect. 119, 1739–1744. Wolf, M., Chen, S., Zhao, X., Scheler, M., Irmler, M., Staiger, H., et al., 2013. Production and release of acylcarnitines by primary myotubes reflect the differences in fasting fat oxidation of the donors. J. Clin. Endocrinol. Metab. 98, E1137–E1142.
884
G. Ye et al. / Science of the Total Environment 691 (2019) 874–884
Xia, J., Sinelnikov, I.V., Han, B., Wishart, D.S., 2015. MetaboAnalyst 3.0-making metabolomics more meaningful. Nucleic Acids Res. 43, W251–W257. Xu, M.X., Ge, C.X., Qin, Y.T., Gu, T.T., Lou, D.S., Li, Q., et al., 2018. Prolonged PM2.5 exposure elevates risk of oxidative stress-driven nonalcoholic fatty liver disease by triggering increase of dyslipidemia. Free Radical Bio. Med. 130, 542–556. Yao, L., Wang, C., Zhang, X., Peng, L., Liu, W., Liu, Y., et al., 2016. Hyperhomocysteinemia activates the aryl hydrocarbon receptor/CD36 pathway to promote hepatic steatosis in mice. Hepatology 64, 92–105. Ye, G., Liu, Y., Yin, P., Zeng, Z., Huang, Q., Kong, H., et al., 2014. Study of induction chemotherapy efficacy in oral squamous cell carcinoma using pseudotargeted metabolomics. J. Proteome Res. 13, 1994–2004. Yu, Z., Zhai, G., Singmann, P., He, Y., Xu, T., Prehn, C., et al., 2012. Human serum metabolic profiles are age dependent. Aging Cell 11, 960–967. Zhang, S.Y., Shao, D., Liu, H., Feng, J., Feng, B., Song, X., et al., 2017. Metabolomics analysis reveals that benzo[a]pyrene, a component of PM2.5, promotes pulmonary injury by
modifying lipid metabolism in a phospholipase A2-dependent manner in vivo and in vitro. Redox Biol. 13, 459–469. Zhang, Y., Ji, X., Ku, T., Li, B., Li, G., Sang, N., 2019. Ambient fine particulate matter exposure induces cardiac functional injury and metabolite alterations in middle-aged female mice. Environ. Pollut. 248, 121–132. Zheng, Z., Xu, X., Zhang, X., Wang, A., Zhang, C., Huttemann, M., et al., 2013. Exposure to ambient particulate matter induces a NASH-like phenotype and impairs hepatic glucose metabolism in an animal model. J. Hepatol. 58, 148–154. Zheng, Z., Zhang, X., Wang, J., Dandekar, A., Kim, H., Qiu, Y., et al., 2015. Exposure to fine airborne particulate matters induces hepatic fibrosis in murine models. J. Hepatol. 63, 1397–1404.