PM2.5 Induced the Expression of Fibrogenic Mediators via HMGB1-RAGE Signaling in Human Airway Epithelial Cells

Background The aim of the present study was to test whether fine particulate matter (PM2.5) induces the expression of platelet-derived growth factor-AB (PDGF-AB), PDGF-BB, and transforming growth factor-β1 (TGF-β1) in human bronchial epithelial cells (HBECs) in vitro via high-mobility group box 1 (HMGB1) receptor for advanced glycation end products (RAGE) signaling. Methods Sprague-Dawley rats were exposed to motor vehicle exhaust (MVE) or clean air. HBECs were either transfected with a small interfering RNA (siRNA) targeting HMGB1 or incubated with anti-RAGE antibodies and subsequently stimulated with PM2.5. Results The expression of HMGB1 and RAGE was elevated in MVE-treated rats compared with untreated rats, and PM2.5 increased the secretion of HMGB1 and upregulated RAGE expression and the translocation of nuclear factor κB (NF-κB) into the nucleus of HBECs. This activation was accompanied by an increase in the expression of PDGF-AB, PDGF-BB, and TGF-β1. The HMGB1 siRNA prevented these effects. Anti-RAGE antibodies attenuated the activation of NF-κB and decreased the secretion of TGF-β1, PDGF-AB, and PDGF-BB from HBECs. Conclusion PM2.5 induces the expression of TGF-β1, PDGF-AB, and PDGF-BB in vitro via HMGB1-RAGE signaling, suggesting that this pathway may contribute to the airway remodeling observed in patients with COPD.


Introduction
Chronic obstructive pulmonary disease (COPD) is characterized by partially reversible air-ow obstruction that is frequently ascribed to airway remodeling. PM2.5 (particles with an aerodynamic diameter less than 2.5 μm) have been associated with an increased risk of COPD, respiratory symptoms, and impaired lung function in epidemiological studies [1][2][3].
High-mobility group box 1 (HMGB1) is a ubiquitous nuclear protein that acts as a crucial factor in acute lung injury progression and pulmonary brosis [4]. HMGB1 is actively or passively released from various cells in response to stimulation with endogenous proin ammatory cytokines [5]. e receptors for HMGB1 include receptor for advanced glycation products (RAGE), Toll-like receptor 2 (TLR2), and TLR4 [6]. Elevated HMGB1 expression in COPD airways might sustain in ammation and remodeling through interactions with RAGE [7]. e interaction of RAGE with its ligand HMGB1 induces the production of a number of pro brotic cytokines, such as PDGF and TGF-β1, in the lungs [8]. TGF-β1 and PDGF are among the most important brogenic mediators that participate in the pathophysiology of pulmonary brosis.
To our knowledge, no study has systematically examined the e ects of PM2.5 on human bronchial epithelial cells (HBECs) in vitro and the possible molecular mechanisms that regulate the expression of TGF-β1 and PDGF. According to epidemiological studies, PM2.5 might regulate the expression of HMGB1-RAGE signaling intermediates and associated mechanisms in elderly men [9]. us, in the present study, we investigated whether PM2.5 induces the expression of TGF-β1 and PDGF in HBECs in vitro via HMGB1-RAGE signaling. e anti-NF-kB p65 antibody, anti-HMGB1 antibody, HMGB1 siRNA (h), and the siRNA transfection reagent were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). e anti-TLR4 antibody, neutralizing anti-RAGE antibody, isotype-matched control antibody (IgG), and TGF-β1, PDGF-AB, and PDGF-BB ELISA kits were obtained from R&D Systems (Minneapolis, MN, USA). e anti-TLR2 antibody was purchased from Novus Biologicals (Novus, CO, USA). e HMGB1 ELISA kit was purchased from Shino-Test (Tokyo, Japan).

Animals.
Eighteen female Sprague-Dawley rats (body weight 180-200 g, 6-8 weeks old) were socially housed (up to four rats per cage) in the Laboratory Animal Center of the First A liated Hospital, Guangzhou Medical University, which approved the use of experimental animals in this study. e rats were randomly divided into a MVE group and a clean air control group. e rats in the MVE group were exposed to MVE (1.5 mg/m 3 ) for 2 h periods, 5 days per week for 1 month. In the MVE exposure room, the concentrations of PM2.5, PM10, and PM1 were 1.46 ± 0.034 mg/m 3 , 1.47 ± 0.034 mg/m 3 , and 1.45 ± 0.035 mg/m 3 , respectively. e O 2 , CO, NO 1 , NO X , and SO 2 levels in the exposure rooms were 20.95 ± 0.006%, 67.524 ± 3.565 ppm, 0.50 ± 0.211 ppm, 0.50 ± 0.211 ppm, and 0.35 ± 0.181 ppm, respectively [10].

Sampling Lung Tissues.
Rats were anesthetized via an intraperitoneal injection of 3% pentobarbital (1 ml/kg). e anesthesia was maintained at a light surgical plane for the duration of testing. Rats were sacri ced by intraperitoneal injection of sodium pentobarbital (100 mg/kg). e left or right lung was in ated and xed using 4% paraformaldehyde (pH 7.40) at 25 cmH 2 O pressure for 24 h. e lungs were then embedded in para n and cut into 4 μm thick sections [10].

Collection and Extraction of PM2.5.
Brie y, tra c ambient PM2.5 was collected by aerodynamic impactors equipped with a glass ber lter, quartz lter, or te on membrane, depending on the speci c purpose. Four batches of PM2.5 lter samples were extracted with DMSO. Mean concentrations of tra c ambient PM2.5 in spring and summer were 79.95 ± 23.89 μg/m 3 and 46.29 ± 12.87 μg/m 3 , respectively. Mean concentrations of these extracts were 11.96 ± 0.96 mg/ml, whereas the recovery from a single membrane was 90.7% ± 5.9%. Subsequently, the quantication and characterization of PM2.5, including polynuclear aromatic hydrocarbons (PAHs), n-alkanes, metals, and water-soluble inorganic ions, were performed using a gravimetric analysis, thermal desorption-gas chromatographymass spectrometry (TD-GC-MS), energy-dispersive X-ray uorescence (ED-XRF) spectrometry, and ion chromatography (IC). e mean concentration of PAHs in PM2.5 was 108.453 μg/g, and the concentration in DMSO extracts was 48.392 μg/g, with a nal PAH recovery of 44.62%. e mean concentration of n-alkanes in PM2.5 was 18,670.883 μg/g. e DMSO extracts had a concentration of 164.675 μg/g [11,12].

Cell
Culture. Human bronchial epithelial cells (HBECs; ATCC) were cultured in 2 ml of DMEM containing 10% FBS and penicillin (100 U/ml) at a density of 5-6 × 10 5 cells per 6-well plate. Prior to the experiments, the cells were serum starved for 24 h, transfected with the HMGB1 siRNA or control siRNA for 48 h or pretreated with anti-RAGE antibodies (10 g/ml) and IgG for 1 h, and subsequently stimulated with 5-20 μg/ml of PM2.5 in medium for 12-48 h. e cells were then harvested for further analyses.

Extraction of Cytoplasmic and Nuclear Proteins and Western Blotting.
e cells were lysed and incubated in a cytoplasmic extract bu er. e remaining nuclei were washed and resuspended in nuclear extract bu er. Protein concentrations were quanti ed using the bicinchoninic acid method. Subsequently, the proteins were separated by SDS-PAGE, transferred to polyvinylidene di uoride membranes, and probed with a rabbit polyclonal anti-RAGE antibody (1 : 1000), mouse polyclonal anti-TLR2 antibody (1 : 1000), mouse polyclonal anti-TLR4 antibody (1 : 1000), or mouse polyclonal anti-NF-κB p65 antibody (1 : 1000). Antibody binding was detected using chemiluminescence according to the manufacturer's instructions. GAPDH (1 : 5000), β-tubulin (1 : 2000), and lamin B (1 : 500) antibodies were used as loading controls. e bands obtained for HMGB1 receptors and cytosolic and nuclear levels of NF-κB were normalized to GAPDH, β-tubulin, and lamin B, respectively, and expressed as a fold change relative to the control.
2.9. Immuno uorescence. Cells were seeded on sterile round coverslips placed in 12-well plates. PM2.5 was added to a subset of wells at a nal concentration of 20 μg/ml. Fortyeight hours later, the cells were stained with the anti-RAGE antibody (1 : 50) for 1 h at room temperature. Antibody binding was detected using a peroxidase-conjugated anti-mouse or anti-rabbit antibody, according to the manufacturer's instructions. Cell nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI). Immuno uorescence was examined using a Leica confocal microscope at ×400 magni cation.
2.10. Immunohistochemistry. Para n sections of lung samples were stained with antibodies against HMGB1 and RAGE for 1 h at room temperature. HMGB1 and RAGE (1 : 50) antibody binding was detected using a peroxidaseconjugated anti-mouse or anti-rabbit antibody and DAB, according to the manufacturer's instructions. Expression was visualized using a confocal microscope at ×400 magni cation. We determined the optical density (OD) of HMGB1-and RAGE-positive cells using a semiquantitative scoring system.

ELISA.
Cell supernatants were collected, and HMGB1, TGF-β1, and PDGF levels were measured using HMGB1, TGF-β1, and PDGF ELISA kits, respectively, according to the manufacturer's instructions. Samples with concentrations exceeding the standard curve limits were diluted until an accurate reading was obtained. Four replicate wells were used to obtain all data points, and all samples were processed in duplicate and averaged.

Statistical Analysis.
Data from at least 3 independent sets of experiments were analyzed using the SPSS 17.0 statistical software and expressed as means + SD. Statistical evaluations of continuous data were performed using ANOVA or the independent samples t-test for between-group comparisons. e level of signi cance was set to P < 0.05.

HMGB1 and RAGE Expression in the Representative Lung
Tissue.
e distribution of HMGB1 and RAGE in lung tissue sections from experimental Sprague-Dawley rats was determined by immunostaining. A large number of HMGB1and RAGE-positive bronchial epithelial cells and alveolar epithelial cells were detected in MVE-treated rats compared with untreated rats (Figures 1(a) and 1(b)). Because the translocation of HMGB1 from the nucleus to the cytoplasm is considered a hallmark of the active secretion of this protein in the extracellular milieu [13], we also determined the subcellular localization of HMGB1 and RAGE. e HMGB1 protein was primarily located in the nucleus and cytoplasm of epithelial cells. RAGE was expressed in the cytoplasm of bronchial epithelial cells and alveolar epithelial cells (Figure 1 (a)).
us, MVE induced the expression of HMGB1 and RAGE in rats, and these cells represent a potential source of the secreted soluble form of HMGB1 in the airways.

PM2.5 Induces the Secretion of HMGB1 and Upregulates RAGE Expression in HBECs.
Based on our observations, PM2.5 was one of the main components detected in the PM exposure room; therefore, we endeavored to further assess whether PM2.5 induced HMGB1 secretion and upregulated the expression of its receptors in HBECs. According to the ELISA data, the cells did not show signi cant changes in HMGB1 secretion after 12 h of PM2.5 (20 g/ml) stimulation, but the levels were signi cantly increased at 24 and 48 h compared with the baseline levels ( Figure 2(a)). A 48 h exposure to PM2.5 (5-20 g/ml) increased HMGB1 secretion from HBECs in a concentration-dependent manner (Figure 2(b)). Moreover, the relative levels of the RAGE mRNA were increased after 48 h of PM2.5 treatment, but signi cant di erences in the levels of TLR2 and TLR4 were not detected, although a slight increase in expression was observed (Figure 2(c)). Western blot data were consistent with the qRT-PCR data (Figure 2(d)), and as shown in Figure 2(e), PM2.5 induced RAGE expression.

e Role of HMGB1 in the PM2.5-Induced Expression of HMGB1 Receptors in HBECs.
We used siRNAs to deplete HMGB1 expression and further assess the extracellular e ect of HMGB1. Cells were transfected with an HMGB1 siRNA or control siRNA for 48 h and subsequently stimulated with PM2.5 (20 μg/ml) for 48 h. e transfection of the HMGB1 siRNA into the cells signi cantly decreased HMGB1 production in both PM2.5-treated and untreated cells. Compared to the negative control siRNA, which did not inhibit the enhanced HMGB1 production observed after PM2.5 treatment, the HMGB1 siRNA transfection suppressed the increase in HMGB1 production; however, HMGB1 production was higher than the untreated control (Figure 3(a)). Cells transfected with the HMGB1 siRNA showed decreased RAGE production compared with cells stimulated with PM2.5 (Figure 3(b)), indicating that the upregulation of RAGE in HBECs exposed to PM2.5 was mediated by HMGB1.

HMGB1-RAGE Signaling Contributes to PM2.5-Induced
NF-κB Activity in HBECs. We characterized the potential downstream signaling events in cells with increased RAGE expression. Accordingly, western blot results did not reveal signi cant changes in NF-κB levels after 12 h of PM2.5 stimulation compared with baseline levels, but signi cantly decreased cytosolic levels of NF-κB and increased nuclear Canadian Respiratory Journal 3 levels were observed at 24 and 48 h of PM2.5 stimulation (Figure 4(a)). Cells transfected with the HMGB1 siRNA showed increased cytosolic levels of NF-κB and attenuated nuclear NF-κB levels compared with cells stimulated with PM2.5 (Figure 4(b)). Furthermore, cells were pretreated with anti-RAGE antibodies and IgG for 1 h and subsequently exposed to PM2.5 (20 μg/ml) for 48 h. Anti-RAGE antibodies signi cantly increased the cytosolic levels of NF-κB and decreased the nuclear levels of NF-κB (Figure 4(c)) compared to the levels observed when cells were pretreated with a negative control antibody (IgG), indicating that HMGB1-RAGE signaling was involved in PM2.5-induced NF-κB activation.

e Role of HMGB1-RAGE Signaling in PM2.5-Induced TGF-β1 and PDGF Production in HBECs.
We examined the secretion of TGF-β1 and PDGF in HBECs cultured with PM2.5 using ELISAs to determine the molecular mechanism by which PM2.5 activated HBECs. Based on the ELISA data, the level of the TGF-β1 protein was increased in these cells   Canadian Respiratory Journal after a 48 h exposure to PM2.5 (Figures 5(a) and 5(b)). Dimeric isoforms of PDGF-A and PDGF-B chains, such as PDGF-AB and PDGF-BB, play important roles in the pathogenesis of brosis. A 48 h exposure to PM2.5 also increased PDGF-AB and PDGF-BB production in HBECs (Figures 5(a) and 5(b)). Cells transfected with the HMGB1 siRNA or incubated with the anti-RAGE antibody showed decreased secretion of the pro brotic cytokines TGF-β1, PDGF-AB, and PDGF-BB compared with cells stimulated with PM2.5 (Figures 5(a) and 5(b)). Based on these results, PM2.5 induces TGF-β1 and PDGF production in HBECs through an HMGB1-RAGE-dependent mechanism.

Discussion
Many studies have reported associations between air pollution and the exacerbation of preexisting COPD [14], although tobacco smoking is the primary cause of COPD. Many other environmental and occupational exposures contribute to its pathology. As shown in our previous study, MVE exposure causes airway cells to release multiple cytokines that are capable of inducing pronounced COPD in rats [10]. In the present study, a large number of HMGB1and RAGE-positive bronchial epithelial cells and alveolar epithelial cells were detected in MVE-treated rats compared with untreated rats. ese cells represent a potential source of the secreted soluble form of HMGB1 in the airways. ese results are consistent with a previous study showing that RAGE and HMGB1 mRNA levels were increased in rats exposed to ozone plus diesel exhaust particulate [15]. PM2.5 was one of the main components observed in the PM exposure room. Combined with the results from our previous study, the results from the present study provided the rst evidence that 48 h of PM2.5 stimulation induces HMGB1 secretion, upregulates RAGE expression, and promotes the translocation of NF-κB to the nucleus of HBECs. Moreover, PM2.5 induced the production of the probrotic cytokines TGF-β1, PDGF-AB, and PDGF-BB in HBECs. HMGB1-RAGE signaling increases the levels of pro brotic cytokines, such as TGF-β1, and PDGF, in the lungs [8]. HBECs express TLR2, TLR4, and RAGE, which bind HMGB1 [16], indicating that HBECs respond to HMGB1. Furthermore, HMGB1-RAGE signaling might also be involved in mechanisms related to PM2.5 in vivo [9]. Based on experimental data, HMGB1 activates RAGE signaling and induces NF-κB activation to promote cellular proliferation in hepatocellular carcinoma (HCC) cell lines  Canadian Respiratory Journal [17].
us, HBECs stimulated with PM2.5 exhibited increased expression of extracellular HMGB1 and RAGE, accompanied by the nuclear translocation of NF-κB and production of TGF-β1, PDGF-AB, and PDGF-BB, which are critical factors involved in the development of brosis and airway remodeling.
HMGB1 alone increases the expression of proinammatory cytokines in HBECs, primarily through RAGE [18]. We transfected cells with an HMGB1 siRNA to block the PM2.5-induced expression of cytokines and HMGB1 receptors. Transfection with the HMGB1 siRNA alone signi cantly inhibited the PM2.5-induced secretion of TGF-β1, PDGF-AB, and PDGF-BB, decreased RAGE production, and attenuated the activation of NF-κB, strongly suggesting that HMGB1 is primarily involved in the PM2.5induced production of the pro brotic cytokines TGF-β1, PDGF-AB, and PDGF-BB. Mechanistically, NF-κB signaling has been shown to be important in HMGB1-induced synovial broblast migration, is associated with abnormal proliferation of cells in the lungs [19], and is signi cantly involved in mediating PDGF-BB and TGF-β1 production [20,21].
In addition, RAGE is the main receptor mediating the e ects of HMGB1 on pro brotic cytokines in HBECs. HMGB1 itself enhances the production of proin ammatory cytokines through RAGE in the absence of TLR activation [18]. Consistent with the results from these reports, upregulation of RAGE and the nuclear translocation of NF-κB in cells exposed to PM2.5 were mediated by HMGB1 in our study. Cellular signaling through RAGE has been suggested to play a role in DPM-(diesel particulate matter-) induced NF-κB activation and chemokine responses in a type-I-like epithelial cell line [22]. However, HMGB1 has been reported to increase cytokine production in macrophages in response to proin ammatory cytokines through TLR2 and TLR4 [23]. Based on this evidence, we used anti-RAGE antibodies to block the expression of PM2.5-induced cytokines and activate NF-κB. RAGE in uenced NF-κB activation and the production of the pro brotic cytokines TGF-β1, PDGF-AB, and PDGF-BB following PM2.5 stimulation, suggesting that PM2.5 induces the expression of extracellular HMGB1 that in turn upregulates RAGE expression, which participates in the activation of NF-κB to induce the production of TGF-β1, PDGF-AB, and PDGF-BB.
In summary, we reported previously unknown ndings showing that HMGB1 and RAGE expression was elevated in MVE-treated rats compared with untreated rats. PM2.5 induced the secretion of HMGB1 and promoted TGF-β1, PDGF-AB, and PDGF-BB production in HBECs in a RAGEdependent manner, and this mechanism is partially dependent on NF-κB activation. ese results will contribute to a better understanding of the airway remodeling occurring in patients with COPD who are exposed to PM2.5 and the development of new therapeutic approaches.

Conflicts of Interest
e authors declare that they have no con icts of interest.