Macrophage-Mediated Effects of Airborne Fine Particulate Matter (PM2.5) on Hepatocyte Insulin Resistance in Vitro.

Fine particulate matter (PM2.5) pollution poses significant health risks worldwide, including metabolic syndrome-related diseases with the characteristic feature of insulin resistance. However, the mechanism and influencing factors of this effect are poorly understood. In this serial in vitro study, we aimed at testing the hypothesis that macrophage-mediated effects of PM2.5 on hepatic insulin resistance depend on its chemical composition. Mouse macrophages were exposed to PM2.5 that had been collected during summer or winter in Beijing, which represented different compositions of PM2.5. Thereafter, hepatocytes were treated with macrophage-conditioned medium (CM). PM2.5 induced interleukin-6, tumor necrosis factor-α, and monocyte chemoattractant protein-1 expression and secretion in macrophages, particularly after winter PM2.5 exposure. Correspondingly, winter CM weakened hepatocellular insulin-stimulated glucose consumption. Further investigation revealed that the normal insulin pathway was suppressed in winter CM-treated hepatocytes, with increased phosphorylation of insulin receptor substrate 1 at serine residue 307 (Ser307) and decreased phosphorylation of protein kinase B (PKB/AKT) and forkhead box transcription factor O1 (FoxO1). Moreover, c-Jun N-terminal kinase, a key moderator of the sensitivity response to insulin stimulation, was activated in hepatocytes treated with winter CM. Although further studies are warranted, this preliminary study suggested an association between PM composition and insulin resistance, thus contributing to our understanding of the systemic toxicity of PM2.5.

Introduction

It has been suggested that particulate matter (PM) air pollution is a significant risk factor for human health worldwide. It contributed to about 2.9 million deaths in 2013 according to the Global Burden of Disease study.[1] In China, ambient PM was the fourth major risk factor, resulting in 1.2 million premature deaths in 2010, with it being the major cause of chronic noncommunicable diseases including respiratory and cardiovascular diseases.[2] In addition, a series of epidemiological studies have proposed that PM exposure is associated with several other metabolic syndrome-related chronic diseases such as type 2 diabetes mellitus and atherosclerosis.[3,4]

Among the particles, fine PM refers to those with an aerodynamic diameter of ≤2.5 μm (PM2.5). They have different shapes and chemical compositions[5−7] and thus have different surface properties.[8] Because of their small sizes and large surface areas, PM2.5 can enter deep into the pulmonary alveoli along with adsorbed toxic components.[9] Once inhaled and deposited into the lungs, PM2.5 can induce oxidative stress and local inflammation by activating stress kinases, such as mitogen-activated protein kinase (MAPK), especially after being phagocytosed by pulmonary macrophages, which can generate and release various proinflammatory mediators, such as interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), and monocyte chemoattractant protein-1 (MCP-1).[10] These cytokines are important signaling molecules that respond to environmental stress and can enter the blood, resulting in systemic inflammation.[11] The proinflammatory effects are probably attributed to polycyclic aromatic hydrocarbons (PAHs), transition metals, and endotoxins in PM2.5.[10]

Epidemiological studies have found a significant association between PM2.5 exposure and increased prevalence of metabolic syndrome-originated diseases, such as type 2 diabetes mellitus.[3,12] Moreover, exposure to PM2.5 in mice could lead to abnormal glucose metabolism and insulin resistance,[13,14] which are fundamental for metabolic syndrome. The underlying molecular mechanism is not fully understood; however, several proinflammatory cytokines have been proposed as causative factors of insulin resistance in insulin target cells, such as hepatocytes and skeletal muscle cells.[15,16] In combination with the evidence that exposure to PM2.5 can induce proinflammatory cytokine expression, particularly in alveolar macrophages,[17−19] we hypothesized that macrophages could mediate the effects of ambient PM2.5 on insulin resistance in hepatocytes via the generation and release of proinflammatory cytokines.

PM2.5 pollution is a severe environmental issue throughout China,[20] particularly in metropolitan regions. For instance, in Beijing, the capital of China, with ∼21 million inhabitants, the annual PM2.5 level reached 80.6 μg m–3 in 2015,[21] which was ∼2.3-fold greater than the recommended Interim Target-1 level set by the WHO. Furthermore, distinctive seasonal variations in the PM2.5 concentration and composition based on source have been reported.[22,23] As we have demonstrated previously, PM2.5 samples collected in winter contained much higher levels of PAHs and its derivatives than did those collected in summer in Beijing.[24] However, there is limited information on the proinflammatory and possible downstream effects (e.g., on insulin target cells) of PM2.5 from Beijing, which is composed of various toxic components.

To test the hypothesis that the macrophage-mediated effect of PM2.5 on hepatic insulin resistance is dependent on its chemical composition, we designed a serial in vitro study and exposed mouse macrophages to PM2.5 collected during summer and winter from Beijing, which represented different sources and chemical compositions. Thereafter, we explored the effects of a macrophage-conditioned medium (CM) on insulin-stimulated glucose consumption and the underlying insulin signaling pathway in mouse hepatocytes. This will help clarify the mechanisms that link PM2.5 exposure to insulin resistance and further explore the influencing factors on this effect, especially the sources and chemical composition of PM2.5.

Materials and Methods

Cell Lines and Reagents

The mouse macrophage cell line (Ana-1) was purchased from the cell culture center at the Shanghai Institutes for Life Science of the Chinese Academy of Sciences. The mouse hepatocyte cell line (NCTC clone 1469) was acquired from the cell culture center at the Institute of Basic Medical Sciences of the Chinese Academy of Medical Sciences.

Dulbecco’s minimum essential medium (DMEM), fetal bovine serum (FBS), donor equine serum, penicillin/streptomycin, phosphate-buffered saline (PBS), and trypsin ethylenediaminetetraacetic acid (EDTA) solution (0.05% trypsin, 0.02% EDTA) were purchased from Thermo Fisher Scientific (Rockford, IL). Power SYBR Green PCR Master Mix was obtained from Applied Biosystems (Woolston, Warrington, U.K.). Dihydrorhodamine 123 (DHR 123), TRIzol reagent, bovine insulin, water, phosphatase, and proteinase inhibitors were obtained from Sigma-Aldrich (St. Louis, MO). The PrimeScript RT Reagent Kit was purchased from TaKaRa (Dalian, China). Cytometric bead array (CBA) reagents for mouse IL-6, TNF-α, and MCP-1 were purchased from BD Biosciences (San Diego, CA). The glucose oxidase method kit, protein quantification kit, RIPA buffer, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer, and Super ECL Plus solution were purchased from Applygene (Beijing, China). Antibodies against insulin receptor substrate 1 (IRS1), phosphorylated IRS1 (pIRS1), protein kinase B (PKB/AKT), phosphorylated protein kinase B (pAKT), forkhead box transcription factor O1 (FoxO1), phosphorylated FoxO1 (pFoxO1), glycogen synthase kinase 3β (GSK3β), phosphorylated GSK3β (pGSK3β), c-Jun N-terminal kinase (JNK), phosphorylated JNK (pJNK), and β-actin were purchased from Cell Signaling Technology (Beverly, MA). Finally, the secondary antibody horseradish peroxidase (HRP)-conjugated rabbit anti-goat IgG, from ZSGB-BIO (Beijing, China), was used in the study.

PM2.5 Collection and Preparation

PM2.5 was collected at the Peking University Atmosphere Environmental Monitoring Station in Beijing, China (39°59′21″N, 116°18′25″E, ∼30 m above ground level). PM2.5 was collected daily on Whatman Teflon filters (diameter 47 mm; Clifton, NJ), using a TH-16A Ambient Particulate Sampler (Tianhong Instruments, Wuhan, China), at a flow rate of 16.7 L min–1, from July 10 to August 4, 2013 (summer), and from December 2 to December 24, 2013 (winter). After sampling, the filters were stored in the dark at −20 °C until extraction. The PM2.5 on the filters was prepared following a previous method, with modifications.[25] Briefly, each filter was sonicated three times for 30 min each using sterilized water in an ice bath. The extracts were pooled by season and lyophilized in a vacuum freeze dryer. Then, the particles were weighed, resuspended at 10 mg mL–1 in sterilized water, and stored in the dark at −80 °C. The stock PM2.5 solution was sonicated briefly before dilution for the cell exposure experiments.

Cell Culture and Exposure

Ana-1 macrophages were grown in DMEM supplemented with 10% FBS and penicillin/streptomycin at 37 °C and 5% CO2. At 80–90% confluency, the cells were passaged using trypsin–EDTA, cultured in 12-well plates at a concentration of 500 000 cells mL–1, and incubated for 12 h to adhere. After overnight serum starvation, the cells were exposed to 20 μg cm–2 (∼80 μg mL–1) summer or winter PM2.5 in DMEM supplemented with penicillin/streptomycin for 3–12 h. After exposure, the culture medium was separated via centrifugation at 5000g for 5 min to eliminate residual cells and particles, and the supernatant was collected as the summer or winter CM.

NCTC 1469 hepatic cells were cultured in DMEM supplemented with 10% heat-inactivated donor equine serum and penicillin/streptomycin at 37 °C and 5% CO2. At 80–90% confluency, the cells were seeded in plates at a concentration of 500 000 cells mL–1 and incubated for 12 h to adhere. After serum starvation, the cells were treated with summer or winter CM for 3 h. Next, the CM was replaced with DMEM supplemented with 100 nM bovine insulin and penicillin/streptomycin and incubated for 1 h for the assessment of glucose consumption or for 10 min for western blot analysis.

Gene Expression

Total RNA was extracted from Ana-1 cells using the TRIzol reagent, and the RNA was reverse-transcribed to generate cDNA using the PrimeScript RT Reagent Kit. The relative expression levels of IL-6, TNF-α, and MCP-1 in Ana-1 cells were quantified using Power SYBR Green PCR Master Mix. The two-step polymerase chain reaction (PCR) cycle was as follows: initial denaturation at 95 °C for 10 min, 40 cycles of denaturation at 95 °C for 15 s, and annealing and extension at 60 °C for 1 min. The relative gene expression levels obtained by quantitative real-time reverse transcription PCR (qRT-PCR) in each group were calculated using the 2–ΔΔCT method following normalization to β-actin. The primers used for the measured genes are listed in Table .

Table 1

Primer Sequences for qRT-PCR

primers	sequences
IL-6 forward primer	TCCAGTTGCCTTCTTGGGAC
IL-6 reverse primer	GTGTAATTAAGCCTCCGACTTG
TNF-α forward primer	CATCTTCTCAAAATTCGAGTGACAA
TNF-α reverse primer	TGGGAGTAGACAAGGTACAACCC
MCP-1 forward primer	CTTCTGGGCCTGCTGTTCA
MCP-1 reverse primer	CCAGCCTACTCATTGGGATCA
β-actin forward primer	TCATCACTATTGGCAACGAGC
β-actin reverse primer	AACAGTCCGCCTAGAAGCAC

Cytokine Release

The levels of proinflammatory cytokines IL-6, TNF-α, and MCP-1 in the CM were measured using CBA as per the manufacturer’s instructions. Briefly, the culture medium supernatant was incubated with capture beads for 1 h, followed by incubation with PE-detecting reagents for 1 h at room temperature in the dark. The beads were washed and resuspended in 300 μL washing buffer for flow cytometric analysis (FACSVerse; BD Biosciences). The acquired data were analyzed using FCAP Array software (BD Biosciences).

Reactive Oxygen Species (ROS) Measurement

The ROS levels in NCTC 1469 cells were measured using the fluorescent probe DHR 123. After treatment with CM or DMEM, the probe solution at a final concentration of 10 μM was spiked into the culture medium, and the cells were incubated at 37 °C for 30 min in the dark, followed by two washes with PBS. Finally, the cells were dissociated using trypsin–EDTA solution and resuspended in 0.5 mL PBS. The samples were analyzed using a BD FACSVerse flow cytometer at excitation/emission wavelengths of 488/530 nm. For each sample, 10 000 events were acquired.

Glucose Consumption

After treatment with CM or DMEM, the NCTC 1469 cells were incubated with DMEM supplemented with 100 nM bovine insulin and penicillin/streptomycin for 1 h. The glucose concentration was measured using the glucose oxidase method kit. Hepatocyte glucose consumption was obtained by subtracting the glucose concentration obtained from the wells with cells from that in blank wells.[26]

Protein Isolation and Western Blot Analysis

NCTC 1469 cells were washed with PBS and lysed with RIPA buffer containing phosphatase and proteinase inhibitors at 4 °C. After boiling for 5 min, the harvested proteins were resolved by SDS-PAGE and transferred onto a polyvinylidene difluoride membrane. The membranes were blocked in 5% BSA buffer for 1 h and then blotted with the primary antibody (anti-IRS1, anti-pIRS1, anti-AKT, anti-pAKT, anti-FoxO1, anti-pFoxO1, anti-GSK3β, anti-pGSK3β, anti-JNK1/2, anti-pJNK1/2, and anti-β-actin) at 4 °C for 12 h and the HRP-conjugated secondary antibody for 1 h at room temperature. Finally, the membranes were developed using the Super ECL Plus Western blotting detection system, and images of the immunoreactive bands were acquired using the ChemiDoc XRS System (Bio-Rad, Hercules, CA). The integrated band densities were measured using Image J software (National Institutes of Health, USA).

Statistical Analysis

Data are presented as means ± standard error of the mean (SEM) of at least three parallel experiments. Differences between groups were tested by using Student’s t-test, and p < 0.05 was considered to be statistically significant. All statistical analyses were conducted using SPSS (ver. 16.0; SPSS Inc., Chicago, IL).

Results

Inflammatory Cytokine Expression and Release in Macrophages

As pulmonary cells such as macrophages and alveolar epithelial cells are the first cells to respond to inhaled PM2.5, we determined the capacity of PM2.5 to activate macrophages by measuring the expression of representative proinflammatory cytokines, including IL-6, TNF-α, and MCP-1. PM2.5 induced IL-6, TNF-α, and MCP-1 expression in Ana-1 cells in a time-dependent manner (Figure a–c). Significant increases in the IL-6 and MCP-1 mRNA levels were observed in PM2.5-treated Ana-1 cells at 3 h, especially those exposed to winter PM2.5. The IL-6 mRNA levels remained unchanged after 6 h of incubation but decreased by 12 h. Similar alterations were observed in the MCP-1 mRNA levels. The TNF-α mRNA levels were higher in summer PM2.5-treated cells at 3 and 6 h, whereas no difference was observed at 12 h. This indicated an acute and short response to PM2.5 for Ana-1. Therefore, we measured the secretion of the corresponding proteins after a 3 h incubation. Both summer and winter PM2.5 induced an increase in the release of IL-6, TNF-α, and MCP-1 into the Ana-1 culture medium (Figure d). Macrophages exposed to winter PM2.5 showed a greater increase in TNF-α release, followed by IL-6 and MCP-1, compared to those treated with summer PM2.5.

Figure 1

Effects of 20 μg cm–2 summer and winter PM2.5 exposure on (a) IL-6, (b) TNF-α, and (c) MCP-1 gene expression in Ana-1 cells at different exposure times, and (d) cytokine release in Ana-1 cells after 3 h of treatment. Values are expressed as means ± SEM of three to four independent experiments. *p < 0.05 and **p < 0.01.

ROS Generation and Glucose Consumption in Hepatocytes

Oxidative stress has been reported to induce insulin resistance in hepatocytes and muscle cells.[27,28] In this study, we measured intracellular ROS levels in NCTC 1469 cells. Both summer and, especially, winter CMs enhanced ROS production significantly compared to that in the cells exposed to DMEM or the control CM (Figure a).

Figure 2

Effect of culture media on (a) intracellular ROS levels and (b) insulin-stimulated glucose consumption in NCTC 1469 cells. C-CM, S-CM, and W-CM represent control, summer, and winter CMs, respectively. Values are expressed as means ± SEM of five to six independent experiments. **p < 0.01.

Glucose consumption was examined to investigate the insulin sensitivity of NCTC 1469 hepatocytes. Hepatocytes pretreated with winter CM had a significantly lower glucose consumption than that of the other three groups (Figure b), whereas no significant differences were found among DMEM-, control CM-, or summer CM-exposed hepatocytes. This indicated that exposure to winter CM decreased the insulin sensitivity in NCTC 1469 cells.

Alteration of the Insulin Signaling Pathway

Glycometabolism is regulated mainly by the insulin signaling pathway.[29] To explore the underlying mechanism of decreased glucose consumption in winter CM-treated hepatocytes, we assessed key molecules of the insulin signaling pathway. The results are shown in Figure . In winter CM-exposed hepatocytes, the phosphorylation ratio of IRS1 at Ser307 increased, which could lead to suppressed normal insulin signaling.[30] A decreased ratio of phosphorylated AKT (Ser473)/total AKT was observed in winter CM-treated cells, likely as a result of IRS1 phosphorylation, as AKT is a vital kinase in the insulin pathway downstream of IRS1.[31] Similarly, the phosphorylation ratio of FoxO1 (Ser256), a kinase downstream of the insulin signaling cascade, was decreased significantly. Given that FoxO1 inhibits gluconeogenesis, inhibiting the phosphorylation of FoxO1 could enhance gluconeogenesis,[32] which is a characteristic feature of insulin resistance. For GSK3β, a key regulator of glycogen synthesis,[33] no significant differences in the ratio of phosphorylated (Ser9) versus total GSK3β were observed.

Figure 3

(a) Western blot analysis of phosphorylated and total IRS1, AKT, FoxO1, GSK3β, and JNK1/2 and (b) their levels in hepatocytes treated with DMEM, control CM (C-CM), summer CM (S-CM), and winter CM (W-CM). Values are expressed as means ± SEM of three independent experiments. *p < 0.05.

Suppression of normal insulin pathway could be triggered by the activation of certain stress kinases, like MAPKs. In fact, JNK, a major subset of MAPKs, comprising JNK1 and JNK2,[34] showed higher phosphorylation levels in winter CM-treated hepatocytes, indicating activation of this stress kinase.

On the basis of the results, we proposed a mechanism of hepatic insulin resistance induced by exposure to medium conditioned by macrophages treated with winter PM2.5 (Figure ).

Figure 4

Proposed mechanism of hepatic insulin resistance induced by exposure to medium conditioned by macrophages treated with winter PM2.5.

Discussion

PM pollution is a significant risk factor for human health worldwide, especially PM2.5, which can be inhaled along with absorbed toxic components and deposited in the lungs.[9] Recent epidemiological studies have identified an association between PM2.5 exposure and increased incidence of metabolic syndrome-related diseases, including type 2 diabetes mellitus.[3,4] In addition, an in vivo experiment revealed elevated fasting blood glucose levels and insulin resistance in mice exposed to urban PM2.5.[14] Such adverse health effects have been proposed to be initiated via oxidative stress and systemic inflammation after PM2.5 exposure and could be modified by the chemical composition of PM.[11] To test this hypothesis, we explored the effects of PM2.5 collected during summer and winter in Beijing, which differed in their main sources and chemical compositions, on the insulin sensitivity of hepatocytes in serial in vitro experiments.

Pulmonary cells, such as macrophages and alveolar epithelial cells, are among the first to respond to inhaled particles; therefore, we exposed macrophages to PM2.5. After incubation for 3 h, the intracellular mRNA levels of IL-6 and MCP-1 were increased, although the TNF-α mRNA levels in winter PM2.5-treated macrophages remained unchanged (Figure a–c). The levels of IL-6, TNF-α, and MCP-1 secreted into the culture medium were higher in PM2.5-treated macrophages than those in controls. In particular, macrophages exposed to winter PM2.5 showed a greater increase in IL-6, TNF-α, and MCP-1 release compared to those treated with summer PM2.5, indicating that winter PM2.5 was more cytotoxic to macrophages (Figure d). This was in agreement with the half maximal inhibitory concentration results for winter PM2.5 (45.6 μg cm–2) and summer PM2.5 (209 μg cm–2; Figure S2 in the Supporting Information). This differed from the results obtained for several European cities, which suggested that summer PM2.5 induced higher levels of IL-6 release in rat macrophages than did winter PM2.5.[17] However, Pozzi et al. found that RAW 264.7 mouse macrophages released higher levels of TNF-α after incubation with winter PM2.5 than those on incubation with summer PM2.5 from Rome.[18]

The difference in cytotoxicity between summer and winter PM2.5 could be attributed, at least partly, to their chemical compositions. In fact, a higher concentration of secondary components, such as NH4+, SO42+, and NO3–, was found in the summer PM2.5 solution (Figure S1), whereas the winter PM2.5 solution contained more primary components like K+, a marker for the primary source of biomass burning. Other components, such as organics, were not measured in the present study; however, in a previous study, we found that winter PM2.5 contained more toxic organic components due to primary emissions from coal and biomass burning in the urban and neighboring regions of Beijing.[24] The levels of PM2.5-bound PAHs and nitrated, hydroxylated, and oxygenated derivatives were much higher during winters than during summers at the same site. The benzo(a)pyrene (BaP)-based toxic equivalency (BaPeq) of PM2.5-bound PAHs during winters (median, 0.44 ng BaPeq/μg PM2.5) was 12.5-fold higher than that during summers (median, 0.035 ng BaPeq/μg PM2.5), suggesting that winter PM2.5 could be more toxic.[24]

Studies have suggested that inhaled PM2.5 has effects on tissues and organs beyond the respiratory system.[11] In humans, the liver is the primary organ responsible for glucose and lipid metabolism.[29] After stimulation by insulin, hepatocytes accelerate the glucose consumption rate by triggering the insulin signaling pathway. To simulate this condition, CM was used to model the association between PM2.5-activated macrophages and hepatocyte responses. Although neither summer CM nor winter CM exposure significantly affected hepatocyte viability compared to that on exposure to the DMEM control (Figure S3), increased ROS levels were observed, particularly in cells exposed to winter CM (Figure a). Previous studies have suggested that proinflammatory cytokines, such as TNF-α, can induce hepatocellular ROS generation by activating nicotinamide adenine dinucleotide phosphate oxidase.[35] Therefore, the increased ROS production in the hepatocytes was probably due to high proinflammatory cytokine levels, especially after winter CM exposure.

Meanwhile, compared with DMEM, control, or summer CM, winter CM reduced the insulin-stimulated glucose consumption in hepatocytes (Figure b), suggesting impairment of the insulin signaling pathway. In glucoregulation, normal insulin signaling begins with autophosphorylation of the insulin receptor (IR), followed by phosphorylation of IRS1 at several tyrosine residues and phosphorylation of the downstream protein kinase, AKT.[27] IRS1 phosphorylation of serine residues (e.g., Ser307 and Ser1101) can interfere with tyrosine residue phosphorylation and thus inhibit IRS1-mediated insulin signaling.[30,36,37] For example, Zheng et al. reported that ambient PM2.5 induced IRS1 phosphorylation at Ser636 and Ser1101, impairing hepatic glucose metabolism in mice.[14] In this study, we found increased IRS1 phosphorylation at Ser307 in hepatocytes exposed to winter CM, followed by decreased phosphorylation of AKT at Ser473, which confirmed suppression of the IRS1–AKT signaling pathway.

Downstream of the IRS1–AKT signaling cascade, insulin-stimulated FoxO1 phosphorylation is an essential step in gluconeogenesis inhibition. FoxO1 is part of the forkhead transcription factor family and is a substrate of AKT in hepatocytes. Furthermore, it mediates glucose-6-phosphatase (G6Pase) transcription by binding to its CAAAACAA sequence. Phosphorylation of FoxO1 via AKT stimulation can promote nuclear-to-cytoplasmic translocation and degradation of FoxO1, which can result in lower G6Pase expression and suppression of gluconeogenesis.[36] In this study, hepatocytes treated with winter CM showed attenuated FoxO1 phosphorylation, which confirmed the suppression effect of winter CM on IRS1-mediated insulin signaling and explained the reduced glucose consumption observed. By contrast, no significant changes were observed in hepatocytes treated with summer CM, indicating that differences in chemical composition affected macrophage responses directly and impacted hepatocytes distinctly.

Inflammatory cytokines, such as IL-6 and TNF-α, have been demonstrated to induce insulin resistance in insulin target cells by binding to the corresponding receptors in the cell membrane and activating downstream regulatory proteins, such as JNK.[35] JNK proteins belong to the MAPK family and have three isoforms: JNK1, JNK2, and JNK3. JNK1 and JNK2 are broadly expressed in cells, including hepatocytes.[34] Activating JNK increases IRS1 phosphorylation at serine residues, thereby impairing insulin signaling.[30] In addition to inflammatory cytokines, ROS have been reported to induce JNK1/2 activation in hepatocytes.[35] As proposed in Figure , impaired insulin signaling and glucose consumption in the winter CM-treated hepatocytes might be a result of the enhanced JNK1/2 activation in this study (Figure ).

In conclusion, using an in vitro CM system, we found that winter PM2.5 from Beijing induced insulin resistance in hepatocytes, probably due to the high levels of toxic components in PM2.5 during the winter. These preliminary results provided a potential mechanism linking PM2.5 exposure to the insulin resistance of hepatocytes and thus contributed to our understanding of the systemic toxicity of PM, including its role in metabolic syndrome-related diseases, such as type 2 diabetes mellitus. Nonetheless, it should pointed out that the extraction process could change the surface properties and shape of PM2.5; therefore, further in vitro studies with air–liquid interface exposure devices and in vivo mammal exposure experiments are warranted to identify the main components in PM2.5 that contribute to its inflammatory effects and insulin resistance, which is essential for controlling this important air pollutant.