Diesel Exhaust Particles Activate the Matrix-Metalloproteinase-1 Gene in Human Bronchial Epithelia in a β-Arrestin–Dependent Manner via Activation of RAS

Background Diesel exhaust particles (DEPs) are globally relevant air pollutants that exert a detrimental human health impact. However, mechanisms of damage by DEP exposure to human respiratory health and human susceptibility factors are only partially known. Matrix metalloproteinase-1 (MMP-1) has been implied as an (etio)pathogenic factor in human lung and airway diseases such as emphysema, chronic obstructive pulmonary disease, chronic asthma, tuberculosis, and bronchial carcinoma and has been reported to be regulated by DEPs. Objective We elucidated the molecular mechanisms of DEPs’ up-regulation of MMP-1. Methods/Results Using permanent and primary human bronchial epithelial (HBE) cells at air–liquid interface, we show that DEPs activate the human MMP-1 gene via RAS and subsequent activation of RAF-MEK-ERK1/2 mitogen-activated protein kinase signaling, which can be scaffolded by β-arrestins. Short interfering RNA mediated β-arrestin1/2 knockout eliminated formation, subsequent nuclear trafficking of phosphorylated ERK1/2, and resulting MMP-1 transcriptional activation. Transcriptional regulation of the human MMP-1 promoter was strongly influenced by the presence of the –1607GG polymorphism, present in 60–80% of humans, which led to striking up-regulation of MMP-1 transcriptional activation. Conclusion Our results confirm up-regulation of MMP-1 in response to DEPs in HBE and provide new mechanistic insight into how these epithelia, the first line of protection against environmental insults, up-regulate MMP-1 in response to DEP inhalation. These mechanisms include a role for the human –1607GG polymorphism as a susceptibility factor for an accentuated response, which critically depends on the ability of β-arrestin1/2 to generate scaffolding and nuclear trafficking of phosphorylated ERK1/2.


Research
The production of diesel exhaust particles (DEPs) by vehicular traffic is a major con tributor to urban particulate matter air pollu tion (McClellan 1987;McClellan et al. 1985;Sydbom et al. 2001;TorresDuque et al. 2008). Inhalation of diesel exhaust is associated with cardiovascular diseases (e.g., atherosclero sis, arrhythmias, thrombosis) and respiratory diseases [e.g., chronic asthma, chronic obstruc tive pulmonary disease (COPD), bronchial cancer], leading to an increase in mortality (Bayram et al. 2006). DEPs form aggregates approximately 0.1-0.5 µm in diameter that can penetrate into more distal branches of the bronchial tree. Because of the large num ber of hazardous chemicals that are present on DEPs, their pathologic effects on airways and lungs are pleiotropic, as documented in numerous studies that have focused on various pathologic mechanisms. Specifically, DEPs have been shown to increase the secretion of proinflammatory cytokines, release phosphati dylcholine, produce reactive oxygen species that lead to oxidative injury, and induce DNA damage, any or all of which may compro mise lung function (Bayram et al. 2006;Cao et al. 2007a; Danielsen et al. 2008; Ghio et al. 2000;Madden et al. 2000;Nikula et al. 1995;Singh et al. 2004;Zhang et al. 2004).
Matrix metalloproteinase1 [MMP1; Ensembl Gene ID ENSG00000196611 (Ensembl 2008)] is a zincdependent endo peptidase that has been shown to exert detri mental effects on respiratory health. MMP1 is secreted from cells as an inactive precursor of the active proteinase, zymogen (Pardo and Selman 2005). MMP1 plays a role in tissue remodeling and repair during development, in inflammation, and in the invasion, migration, and metastasis of malignantly transformed cells (Boire et al. 2005;Ishii et al. 2003). A polymorphism in the MMP-1 5´regulatory region -1607G(G) exerts a powerful effect on transcriptional activation, and the 1607GG sequence forms an Ets transcriptionfactor binding site, which acts as a transcriptional activator (Brinckerhoff and Matrisian 2002;Rutter et al. 1998;Tower et al. 2002).
Activation of MMP-1 has been shown to be of great relevance for airway and lung health and disease. MMP1 is involved in airway extra cellular matrix degradation and alveolar wall stability and is pathogenetically linked to both malignant and nonmalignant chronic respiratory diseases (Elkington et al. 2005;Mercer et al. 2004National Heart, Lung, and Blood Institute 2007;SeguraValdez et al. 2000), including COPD, chronic asthma, emphysema, lung tuberculosis, and bronchial carcinoma.
Two studies have examined the putative role of DEPinduced MMP-1 activation in lung cells. Doornaert et al. (2003) reported a decrease in MMP1 expression when HBE cells (16HBE14o) were exposed to DEPs. In contrast, Amara et al. (2007) investigated the effects of DEPs on MMP1 expression in A549 and NCIH292 lung epithelial tumor cell lines and found it increased and dependent on the NADP(H) oxidase/NOX4 redoxdependent mechanism. Given these seemingly conflicting results and the relevance of increased MMP1 expression for human respiratory health, we addressed this issue in permanent and primary human bronchial epithelial (HBE) cells, the latter assayed at air-liquid interface, using a DEP preparation high in organic content real istically generated by diesel engines in cars, trucks, buses, locomotives, and boats (Bechtold et al. 1985;Hirano et al. 2003).
We found that DEPs led to increased acti vation of MMP-1 in BEAS2B bronchial epi thelia and primary HBE cells that was linked to specific activation of RAS, which leads to activation of RAFMEKERK1/2 signaling. Signaling was fully dependent on scaffolding by both βarrestin isoforms, enabling mito genactivated protein (MAP) kinase signaling, which activates MMP-1 in the nucleus via phosphorylated extracellular signalregulated kinase (phosphoERK1/2). We also found that the regulatory effect of DEPs on the MMP-1 gene critically involved the -1607GG MMP-1 promoter polymorphism that is present in 25% of Caucasians homozygously, and 50% hetero zygously, and with similar frequencies in Asian and AfricanAmerican populations (Fujimoto et al. 2002). Thus, in most humans, breathing DEPpolluted air may trigger increased MMP-1 activation in airway epithelia, making them vul nerable to chronic airway and lung injury.
Primary HBE cells. All investigations with primary human cells were approved by the institutional review board of the EPA and Duke University; all studies complied with all applicable requirements of the United States and customary international regulations. All participants provided written informed consent before the study. We obtained HBE cells from healthy, nonsmoking adult volunteers by cyto logic brushing of the airways during bronchos copy (Ghio and Cohen 2005). We grew these cells to passage 3 in bronchial epithelial growth medium (BEGM) (Clonetics, San Diego, CA), plated them on 12well culture plates with collagencoated filters with a 0.4µm pore size (TransCLR; Costar, Cambridge, MA) at a density of 1 × 10 5 cells/filter. The cells were maintained in a 1:1 mixture of BEGM and HEPESsupplemented Dulbecco's modified Eagle's medium with singlequot supplements, bovine pituitary extracts (13 mg/mL), bovine serum albumin (1.5 mg/mL), and nystatin (20 U/mL) in 0.5 mL in the apical chamber and 1.5 mL in the basal chamber. Fresh medium was provided every 48 hr. Media was removed from the apical chamber at least 24 hr before use to create an air-liquid interface (Ghio and Cohen 2005). For measurement of secreted MMP1, we sampled media from the lower compartment. Cells were used 3-5 days after passage 3; that is, they did not yet display a terminally differentiated phenotype (Turi et al. 2006). Compared with BEAS2B permanent cells, primary HBE cells maintained in air-liq uid interface culture devices had a 5 to 10fold increased cell density, resulting in 5 to 10fold increased absolute number of cells per assay.
DEPs and P90 control carbon nano particles. DEPs were the generous gift of D. Costa and I. Gilmour at the U.S. Environmental Protection Agency (EPA) in Research Triangle Park, NC (Singh et al. 2004). The DEPs were generated at the U.S. EPA main campus in Research Triangle Park (NC, USA) using a 30kW four cylinder Deutz BF4M1008 diesel engine con nected to a 22.3 kW Saylor Bell air compressor to provide load. The emissions from the engine were diluted with filtered air (3:1), the tempera ture adjusted to approximately 35°C, and the emission directed to a small baghouse (Dustex model T63.59 150 ACFM with nine polyes ter felt bags). The emissions were collected by a conventional pulsing system (reverse air puls ing). While the baghouse was pulsecleaned, DEPs filled the bottom of the baghouse in a conical section. After 45 min, the contents were emptied and immediately refrigerated. In order to simulate a more "realistic" environmental condition, "EPA DEP MIX" was generated by mixing DEPs collected at three different engine operations. The engine and compressor were operated at steady state using approxi mately 25% of the engine's rated full load. The organic:elemental carbon ratio of DEP MIX was estimated to be 0.47, calculated from the weight percentage of each DEP in the DEP MIX and the measured organic:elemental car bon ratio of each filter sample collected from the different engine operations using a thermal optical carbon analyzer.
Degussa Printex 90 (P90) carbon nano spheres were the generous gift of W. Moller (GSF National Research Center for Environment and Health, Neuherberg/Munich, Germany) (Moller et al. 2005). Investigations at GSF have demonstrated that P90 nanoparticles comprise relatively "clean" carbon particles, that is, very low in organic contamination and low in metal/watersoluble contamination. In view of this, and because their size is similar to the DEP carbonaceous cores, we used P90 nano particles as control particles.
We applied DEPs and P90 nanoparticles in concentrations between 10 and 100 µg/mL, subjected to rigorous vortexing (30 sec) before application.
Chemicals/pharmacologics. Pharmacologic inhibitors and targeted pathways included PD98059 and UO126 to specifically inhibit MEK MAP kinase, SB203580 for inhibition of p38 MAP kinase, SP600125 for inhibition of JNK MAP kinase, and LBT613 and AAL881 to inhibit RAF (Hjelmeland et al. 2007;Sathornsumetee et al. 2006). All chemicals were purchased from Tocris (Ellisville, MO), except the RAF inhibitors, which were a generous gift of D. Batt, (Novartis, Cambridge, MA). We chose the concentrations based on recom mendations of the supplier (Tocris) or those commonly used in previous studies (Ahn et al. 2003;Walker et al. 2003), usually 5-10 µM, except for LBT613, which we used at 1 µM (Hjelmeland et al. 2007).
Probing gene expression of MMP-1 by realtime quantitative reverse transcriptase polymerase chain reaction. We quantified relative gene expression of MMP-1 in human airway epithelial cells using realtime quantitative reverse transcriptase polymerase chain reaction (qRTPCR). Total RNA was isolated using the RNAeasy kit (Qiagen, Valencia, CA) and reverse transcribed to generate oligodTprimed cDNA. MMP-1 primer/probe sets were obtained as Taqman predeveloped assay reagents [concentrated and preoptimized mix of primers and Taqman probe labeled with 6carboxyfluorescein (FAM)] from Applied Biosystems (Foster City, CA).Quantitative fluoro genic amplification of cDNA was per formed using the ABI Prism 7500 Sequence Detection System (Applied Biosystems), MMP-1 primer/probe sets of interest, and TaqMan Universal PCR Master Mix (Applied Biosystems). We determined the relative abun dance of mRNA levels from standard curves generated from a serially diluted standard pool of cDNA prepared from cultured human air way epithelial cells. The relative abundance of βactin mRNA was used to normalize levels of MMP-1 mRNAs.
Transfection of MMP-1-promoter reporter constructs. The human MMP-1 reporter plas mids -4400, -3292, -2942, -2002, -1546, and -517 used in this study harbored the firefly luciferase (fLUC) reporter gene under the transcriptional control of the MMP-1 pro moter Rutter et al. 1998;Tower et al. 2003). 1 µg DNA from promoter constructs was transiently transfected using ExGen 500 (Fermentas, Glen Burnie, MD) volume 117 | number 3 | March 2009 • Environmental Health Perspectives into BEAS2B cells and plated in 24well plates. As evidenced by fluorescent reporters, transfection efficiency was > 70% (data not shown). After transfection, cells were washed and incubated for 24 hr with or without DEPs and/or chemical modulators. Cell lysates were generated using 25 mM Tris (pH 7.8), 2 mM EDTA, 10% glycerol, 2 mM dithiothreitol, and 1% Triton X100 and were subsequently assayed for luciferase activity using the Dual Luciferase Reporter Assay System (Promega, Thousand Oaks, CA), in a 96wellplate lumi nometer (Veritas Microplate Luminometer, Turner Biosystems, Mountain View, CA). All transfections were carried out in triplicate or in quadruplicate, and cells were cotrans fected with a promoterless Renilla luciferase construct to control for transfection efficiency and toxicity/viability. Data were normalized by Renilla activity and presented as mean ± SE.
Immunocytochemistry. For immuno fluorescence studies, we preincubated sections for 30 min at room temperature in phosphate buffered saline (PBS) containing 0.3% Triton X100 (PBSTriton) and 10% normal don key serum. The sections were then incubated for 24 hr at 4°C with mouse monoclonal MMP1-specific antibody (MAB901; R&D Systems, Minneapolis, MN) that was diluted in PBSTriton and incubated with fluores cein iso thiocyanate-conjugated donkey anti mouse IgG (H+L; Jackson Immunoresearch, Philadelphia, PA). Fluorescent micrographs were recorded using a BX61 Olympus micro scope equipped with the respective filter sets or a Leica SP2 confocal laser scanning platform. Control experiments were conducted using identical amounts of nonspecific isotype mouse antibody (R&D Systems).
Enzyme-linked immunosorbent assays and MMP-1 cleavage activity assay. We initially conducted a cytokine multiplex assay (BioPlex Hu27Plex; BioRad Laboratories, Hercules, CA) and a simplified matrixmetalloproteinase proteomics array (SearchLight Human MMP Arrays 1 and 2; EndogenPierce, Rockford, IL). In agreement with previous studies (Bayram et al. 1998), we found interleukin6 (IL6) to be significantly upregulated after DEP exposure ( Figure 1A), thus assuring the validity of our cytokine multiplex assay.
We performed MMP1 enzymelinked immunosorbent assay (ELISA) following stan dard protocols (Kim et al. 2005;Peake et al. 2005). Briefly, to measure MMP1 in cell cul ture supernatant of BEAS2B cells and in the lower compartment of primary airway epithelia kept at air-liquid interface, Nunc MaxiSorp ELISA plates (eBioscience, Termecula, CA) were coated overnight at 4°C with MAB901 (R&D Systems). Nonspecific binding to plates was blocked with 1% bovine serum albumin in PBS for 1 hr at room temperature. After wash ing the cells with PBS/0.05% Tween 20, culture supernatants/fluids and the MMP1 standard dilutions were added to wells for overnight incu bation at 4°C. After washing, detection antibody (goat antihuman MMP1, R&D Systems) was added for 90 min at room temperature, fol lowed by washing and donkey antigoat IgG horse radish peroxidase-conjugated detection antibody for 45 min. A chromogenic ELISA substrate (Kblue, Sigma) was then added for color development, which was arrested with 1 M H 2 SO 4 . Plates were read at 450 nm in an ELISA plate reader using SoftmaxPro software (Molecular Devices, Sunnyvale, CA), which led directly to determination of MMP1 concentra tions in samples.
MMP1 cleavage activity was meas ured using a commercially available assay (SensolytePlus 520; Anaspec, San Jose, CA). In this assay, MMP1 is captured by immobilized MMP1 antibodies, and its proteolytic activity is measured by a 5FAM/QXLTM520 pep tide that evokes fluorescent resonance energy transfer (FRET), bearing a specific MMP1 cleavage site. The fluorescence of 5FAM, which acts as fluoro phore, is quenched by QXLTM520, which acts as a quencher in the intact FRET peptide. Upon specific cleavage by active MMP1, the fluorescence of 5FAM is recovered, and the dual emission of 490/520 nm is monitored.
None of the colorimetric readouts were impaired by the use of DEPs as stimulus (data not shown).
Trafficking assay of phospho-ERK. After stimulation with DEPs, we fixed cells in 4% paraformaldehyde at 0, 10, 20, 30, and 60min time points. Cells were then immuno labeled for phosphoERK1/2, using a mouse monoclonal phosphoERK1/2-specific (B) After exposure to increasing concentrations of DEPs, secretion of MMP-1 by BEAS-2B cells followed a time course and a dose-response pattern by ELISA. P90 nanoparticles served as control; experiment conducted six times. (C) In primary HBE, DEP (100 µg/mL) exposure evoked robust MMP-1 secretion by ELISA. Note also the high amount of unstimulated MMP-1 secretion. P90 nanoparticles served as control; experiment conducted in triplicate. (D) MMP-1 immunofluorescence (BEAS-2B, primary HBE cells) identifies cell-bound MMP-1 immunoreactivity (green). In BEAS-2B cells, stimulation with P90 nanoparticles does not surpass background; DEP stimulation (100 µg/mL) leads to robust up-regulation, uniformly staining all cells. Specificity of the DEP MMP-1 response could also be observed in primary HBE cells, yet here a more heterogeneous MMP-1 immunoreactivity is apparent. 4'-6'-diamidino-2-phenylindole (DAPI) nuclear stain illustrates increased cell density (blue insert in the lower right-hand micrograph). Scale bars = 20 µm for MMP-1 immunolabeling, 80 µm for the DAPI. *p < 0.05, **p < 0.01, and # p < 0.001 compared with controls.  antibody (Cell Signaling Technology), which was immunodetected by standard second ary reagents (Molecular Probes, Invitrogen, Carlsbad, CA), using an upright fluores cent microscope (Olympus BX61) or a laser confocal scanning microscope (Leica SP2). Nuclear abundance of phosphoERK1/2 was determined densitometrically on micrographs recorded at respective time points with fixed image acquisition settings. ImageJ freeware was used for analysis (National Institutes of Health 2008) and averaged data were plotted against time.
β-Arrestin short interfering RNA experiments. These experiments used established protocols (Ahn et al. 2003;DeWire et al. 2007). Briefly, chemically synthesized, double stranded short interfering RNAs (siRNAs) w e r e p u r c h a s e d f r o m D h a r m a c o n Research (Lafayette, CO). We derived the sequences from Ahn et al. (2003). The siRNA sequences targeting βarrestin1 (GenBank accession no. NM_020251; National Center for Biotechnology Information 2008) and βarrestin2 (NM_004313) are 5´AAAGCCUUCUG CGCGGAGAAU3´ and 5´AAGGACCGC AAAGUGUUUGUG3´ and correspond to positions 439-459 and 201-221, respec tively, relative to the start codon. Another RNA duplex was synthesized, used as a control (5´AAGUGGACCCUGUAGAUGGCG 3´; position 101-120 from the start codon of βarrestin1, sequence common to both arres tins), and found to have no silencing effects on βarrestin1 or 2 expression (Ahn et al. 2003;DeWire et al. 2007). BEAS2B cells were then transfected with siRNA using Lipofectamine transfection reagent (Invitrogen) following established guidelines (Rippmann et al. 2005). Fortyeight hours after transfection, cells were divided into 24well plates for MMP-1-fLUC reporter gene assays and MMP1 ELISA of the supernatant. All assays were performed in triplicates, with two independent experiments.
For cellular immunodetection of βarrestin, we used a mouse monoclonal antibody that recog nizes both arrestin isoforms (BD Bioscience, San Diego CA), and for βarrestin Western blot detection, an antibody generously pro vided by R. Lefkowitz (Duke University) raised against a common region in the Cterminus.
Transfection of dominant-negative RAS. We obtained the dominantnegative RAS mutant N116Y as a fusion to enhanced green fluorescent protein (eGFP) in pcDNA3.1 plas mid as a generous gift from R. Yasuda (Duke University) (Watanabe et al. 2000;Yasuda et al. 2006). This plasmid (500 ng/well in a 24well plate) was transfected into BEAS2B cells using ExGen500 DNA transfection reagent (Fermentas) according to manufac turer instructions.
Statistical analysis. Mean and SEs of quan tified outcome parameters after DEP stimu lation were compared with their respective controls. For experiments involving modula tion of the response, we likewise compared the increases from stimulation with DEPs alone versus DEPs plus inhibitor. Group compari sons were accomplished using fixedeffect one way analysis of variance (ANOVA) with post hoc Scheffe test for multigroup comparisons. Minimum significance was set at the 0.05 level.

DEPs increase MMP-1 secretion in human permanent and primary bronchial epithelia.
As a starting point of our investigation, we used commercially available multiplex ELISAs to screen the supernatant of DEPexposed BEAS2B airway epithelial cells for upregula tion of proinflammatory mediators. MMP1 emerged as the mediator with the strongest upregulation among MMPs. We also found IL6 to be upregulated, which was previously reported for airway cells in response to DEPs (Steerenberg et al. 1998), thus validating our findings. In addition, there was a slight down regulation of several other MMPs as well as their endogenous inhibitors, tissue inhibitor of metalloproteinase1 (TIMP1) and TIMP2 ( Figure 1A). MMP1 secretion was increased by DEPs as a function of time (2-24 hr) and dose (10-100 µg/mL), but not by a compa rable concentration of P90 nanoparticles that served as inert carbon control particles ( Figure  1B). These findings agree with a recent report of MMP1 upregulation (Amara et al. 2007) but not with the other study reporting conflict ing results (Doornaert et al. 2003). MMP1 is also strongly upregulated in primary HBE cells assayed at an air-liquid interface ( Figure 1C, D). Western blotting of BEAS2B cells confirms this result [see Supplemental Material, Figure 1 (http://www.ehponline.org/ members/2008/0800311/suppl.pdf)]. Using immunocytochemistry, we found that HBE cells exhibit a more heterogeneous expression pattern of MMP1 than do BEAS2B cells ( Figure 1D). Density of MMP1-expressing cells was higher in HBE than in BEAS2B cells. Nevertheless, we noticed that in unstimu lated conditions, HBE cells secrete significantly more MMP1 than do BEAS2B cells, exceed ing a factor 10 to account for increased cell density of HBE cells. Finally, we demonstrated that secreted protein correlates with specific MMP1 cleavage activity [see Supplemental Material, Figure 2 (http://www.ehponline. org/ members/2008/0800311/suppl.pdf)].
DEPs increase MMP-1 transcription in an allele-specific manner. The dose dependency and time course of secreted MMP1 protein in response to DEPs suggest transcriptional regulation of the MMP-1 gene. Therefore, we investigated transcriptional regulation of MMP-1 in response to DEPs, and the possible role of the human 1607G(G) polymorphism. Increased transcription of MMP-1 was con firmed by qRTPCR in BEAS2B cells stimu lated with DEP versus P90 nanoparticles (100 µg/mL, 24hr time point; n = 3 independently stimulated dishes per group, Taqman realtime qRTPCR methodology; Figure 2A; results . The large arrow denotes the -1607G(G) human polymorphism, which generates an ETS transcription factor binding site for GG; two smaller arrows denote additional upstream ETS binding sites. Diagram at right illustrates MMP-1-fLUC reporter gene activity (normalized for Renilla), in relative units (RU). Note that in response to DEPs (100 µg/mL; 24-hr incubation; triplicate assays; data based on three or more experiments), fLUC activity was strikingly increased for -1607GG (2G). (C) DNA sequencing of the MMP-1 promoter, encompassing -1607, from BEAS-2B cells. Like approximately 50% of all humans, BEAS-2B cells have a heterozygous genotype (-1607GG; -1607G).
In (A), *p < 0.05 up-regulation of MMP-1. In (B), *p < 0.05, and **p < 0.01, significant difference compared with the unstimulated condition.  Amara et al. 2007). To examine whether increased MMP-1 mRNA abundance in response to DEPs is due to increased mRNA stability or to increased transcription, we con ducted luciferase reporter gene assays using the 4.4kb MMP-1 promoter. We employed both isoforms of the promoter, -1607G and -1607GG, to elucidate the impact of the promoter polymorphism on gene regulation. Figure 2B shows that upregulation of MMP-1 transcription causes increased MMP-1 mRNA abundance. Also, the -1607GG polymorphism potentiated MMP-1 transcription. Figure 2C shows the genotype of BEAS2B cells to be heterozygous -1607GG; -1607G, indicating that transcription of the endogenous MMP-1 gene in these cells is carried by both alleles. A recent study on transcriptional regu lation of MMP-1 in response to cigarette smoke extract, a complex mix of noxious, pro oxidative chemicals/irritants (in this respect similar to DEPs), reported a distal "tobacco response element" from -2.9 to -4.4kb of the MMP-1 promoter. In view of this, we transfected a set of luciferase reporters harbor ing MMP-1 promoters of varying length [see Supplemental Material, Figure 3 (http://www. ehponline.org/members/2008/0800311/suppl. pdf)] into BEAS2B cells and stimulated them with DEPs. The 2.9kb promoter (-1607GG) exhibited the highest activity, followed by 3.3kb and 4.4kb promoters. This suggests the presence of sequences that function to bind transcriptional repressors, in response to DEPs, within the "tobaccoresponse element" (-2.9 to -4.4 kb). This regulation was appreciable for the -1607GG polymorphism. For shorter constructs, iterative reduction of the 2.9kb promoter decreased transcriptional activation. promoter, -1607G and -1607GG). Cells transfected with the -1607G polymorphism showed no effect; further reduction of transcription for the -1607GG polymorphism was statistically significant, yet incomplete. (H) Transcriptional activation of MMP-1 and secretion of MMP-1 after 24 hr measured using MMP-1 ELISA. Transfection of dominant-negative RAS without DEP stimulation had no effect on MMP-1 transcription. Dominant-negative RAS virtually eliminated MMP-1. Experiment conducted in triplicate with two sets of independent experiments. *p < 0.05, **p < 0.01, and # p < 0.001, indicate significant inhibition of the increase caused by DEP. ## p < 0.05, † p < 0.01, and † † p < 0.001 indicate statistically significant differences from control.

RU LUC
RU LUC † † † † † # been shown to be critically linked to mitogen activated protein kinase activation (Cao et al. 2007;Hecht et al. 2007;Pillinger et al. 2007), and involvement of MEKERK was recently reported for DEPstimulated lung tumor cells, depending on NADPH oxidase (Amara et al. 2007). We used BEAS2B bronchial epithelial cells to explore whether MAP kinases function as intracellular signal transducers leading to transcriptional activation of MMP-1. This was accomplished by inhibiting MEK, JNK, and p38 MAP kinases with specific antagonists; RAF with novel specific inhibitors, and RAS using a dominantnegative genetic construct. We used inhibitory compounds on BEAS2B and primary HBE cells, whereas we used the RAS dominantnegative gene construct only in BEAS2B cells.
Specific inhibitors of MEK (UO126, PD98059) downregulated transcription ( Figure 3A), an effect that was more accen tuated for the -1607GG polymorphism. In keeping with the -1607G;-1607GG genotype of BEAS2B cells, secretion of MMP1 was also profoundly downregulated in these cells ( Figure 3B). To examine whether primary HBE cells show an identical activation pattern of MAP kinases, we inhibited MEK in pri mary HBE cells, which also led to a decrease in MMP1 secretion. MEK inhibition led to even lower levels of secreted MMP1 com pared with nonstimulated, nontreated cells, pointing toward a "tonic drive" along the MEKERK MAP kinase pathway in these cells ( Figure 3C). We next examined whether RAF functions upstream of MEK. Inhibition of RAF using the novel specific inhibitors LBT613 and AAL881 was as effective as MEK inhibition ( Figure 3D-F). Based on this find ing, we examined whether RAF is activated by the membrane bound GTPase RAS. In order to inhibit RAS, the dominantnegative RAS isoform was transfected heterologously into BEAS2B cells, which led to a marked downregulation of the MMP1 response ( Figure 3G, H). However, taking into account an estimated 60-70% efficiency for transient transfection of BEAS2B cells, as indicated by transfection of fluorescent reporter genes (data not shown), a partial downregulation of the MMP1 response indicates a powerful impact of dominantnegative RAS on down stream RAF signaling. Regarding the specific ity of these findings, inhibition of JNK with SP600125 and p38 with SB203580 did not significantly (p > 0.3) downregulate tran scription of MMP-1 or MMP1 secretion [see Supplemental Material, Figure 4 (http:// www.ehponline.org/members/2008/0800311/ suppl.pdf)]. The fourth known MAP kinase pathway, involving ERK5, was also tested using phosphoERK5 Western blot, and it was not activated by DEPs (data not shown). Thus, we conclude that activation of RAS, which leads to RAFMEKERK1/2 MAP kinase signaling, but not p38, JNK, or ERK5, is selectively implicated in MMP-1 activation.

MMP-1 up-regulation depends on both isoforms of β-arrestin.
We next examined the role of βarrestins, given their known associa tion with RAFMEKERK1/2 signaling (Ahn et al. 2003;Dasgupta et al. 2006;DeWire et al. 2007;Lohse et al. 1990;McDonald et al. 2000). In recognition of the emerging role of βarrestins as scaffolding proteins that bind directly to RAF and MEK and indirectly to ERK1/2, orchestrating their function in a "signalosome" (Naumann et al. 1999), we investigated the function of βarrestins in the DEP MMP1 response by employing a pre viously reported βarrestin-specific siRNA (Ahn et al. 2003;DeWire et al. 2007;Lohse et al. 1990;McDonald et al. 2000). We used BEAS2B cells because they have been suc cessfully subjected to siRNA gene knockdown (Cao et al. 2007a;Rippmann et al. 2005). Knocking down βarrestins led to a down regu lation of the targeted proteins and resulted in a downregulation of MMP-1 transcription and MMP1 secretion ( Figure 4A,B). MMP1 downregulatory effects of a specific knock down for βarrestin1 were more noticeable than for βarrestin2 ( Figure 4A,B), yet both were statistically significantly different from control (p < 0.01 for protein secretion and transcriptional activation). However, the most complete effect was obtained using combined βarrestin1 and 2 targeting, which amounted to a protein knockout for both βarrestins, as revealed by Western blot ( Figure 4B) and led to complete elimination of transcription and secretion of MMP-1 and MMP1. These striking effects clearly demonstrate that, in response to DEPs, both βarrestins are neces sary for mediating MMP-1 activation.
DEPevoked formation of phospho ERK1/2 and its subsequent transfer to the nucleus depends on RASRAFMEKERK1/2 MAP kinase signaling and βarrestins. βArrestins' function as scaffolds for RAF MEKERK1/2 has been shown to encompass retention of phosphoERK in the cytoplasm (Ahn et al. 2003;DeWire et al. 2007), yet in another study, they were associated with nuclear translocation of phosphoERK (Kobayashi et al. 2005;GestyPalmer et al. 2005). As a concept, activation of MMP-1 via phosphoERK is understood to involve nuclear translocation of the latter (Roberts and Der 2007;Tower et al. 2002). In order to resolve this issue, we established a time course for nuclear translocation of phosphoERK1/2, using immunocytochemistry for phospho ERK1/2 after stimulation with DEPs.
A time course of phosphoERK1/2, using immunofluorescent labeling, indicated that phosphoERK enters the nuclear compart ment as early as 10 min, with a peak at the (A) Up-regulation of MMP-1 transcription using luciferase reporter genes (4.4-kb promoter, -1607G and -1607GG) in response to DEPs (100 µg/mL, 24 hr) and its striking down-regulation after knocking down β-arrestin-1 and -2 either separately or in combination. Fold increase for the -1607G promoter (vs. control stimulation) was considerably weaker and not statistically significantly different. On the other hand, the large increase for the -1607GG promoter was robustly reduced for β-arrestin knockdown, with β-arrestin-2 exhibiting a slightly weaker effect than β-arrestin-1 and a complete elimination for combined knockout. siRNA was transfected 72 hr before DEP stimulation. Experiment conducted in triplicate, with two independent experiments. (B) Fold increase of MMP-1 secretion (BEAS-2B cells) after stimulation with DEPs (100 µg/mL), and the respective effect of knockdown of β-arrestins. Knockdown for β-arrestin-1 lowers the MMP-1 response significantly and is stronger than knockdown of β-arrestin-2, with the highest impact with both arrestins combined. Western blots show the effect of knockdown of both β-arrestins on their respective protein abundance (normalized for β-actin). Pan-arrestin siRNA leads to complete β-arrestin knockout. siRNA was transfected 72 hr before DEP stimulation. Experiment conducted in triplicate, with two independent experiments. *p < 0.05, **p < 0.01, and # p < 0.001 compared with fold increase obtained for DEP stimulation using control siRNA.  . These findings are not in keeping with the concept of cytoplasmic retention of phosphoERK1/2 by βarrestins, but rather with previously reported experiments that demonstrate βarrestin-dependent nuclear translocation (Kobayashi et al. 2005). Our data clearly suggest a logical temporal sequence that involves early entry of phosphoERK1/2 into the nuclear compartment, leading to early transcriptional activation of MMP-1. These data and their interpretation are in good agreement with our 2hr time point findings for activation of the MMP-1 fLUC reporter gene ( Figure 3A). Using the pres ence of nuclear phosphoERK1/2 at 30 min after stimulation as a readout of the effect of DEPs, we found no increase of nuclear phosphoERK1/2 when chemically inhibiting RAF and MEK and when knocking out both isoforms of βarrestin ( Figure 5C-E). The confocal image of βarrestin knockout cells shows a lack of phosphoERK1/2 staining both in nucleus and cytoplasm, and a virtual elimination of wholecell phosphoERK1/2 when using Western blotting, with total ERK constant. Thus, both βarrestins are necessary for formation of phosphoERK1/2. Without phosphoERK1/2, there is no nuclear trans location and subsequent regulation of MMP-1 in response to DEPs.

Discussion
Using human airway epithelial cells, we have shown that DEPs lead to increased transcrip tional activation of the MMP-1 gene and subsequent secretion of MMP1. This mecha nism is powerfully boosted by the -1607GG polymorphism within the MMP-1 promoter, which is present in at least one allele in approximately 75% of humans and forms a known ETS transcription factor binding site (Rutter et al. 1998). Intracellular con stituents that carry this signal transduction are the MEKERK1/2 MAP kinase cascade, with necessary upstream activation of RAF and RAS (see schematic overview in Figure 6). RAFMEKERK1/2 MAP kinase signaling is known to be scaffolded by βarrestins1 and 2. Accordingly, we found that both βarrestins were necessary for formation of phosphoERK1/2, its subsequent trafficking to the nucleus, and transcription of MMP-1. Thus, our findings are suggestive of a mecha nism for this activation. Because MMP1 has been linked to both nonmalignant and malig nant respiratory disorders, results presented here increase our understanding of how air borne DEPs can injure bronchial epithelia, sensitize airway sensory nerve afferents, and thus damage human airways and lungs in a context of several highly relevant respiratory disorders e.g., emphysema, COPD, chronic asthma, lung tuberculosis, and bronchial car cinoma (Kim et al. 2004;Mercer et al. 2004;Rutter et al. 1998;TorresDuque et al. 2008). Two previous investigations reported on the phenomenon of DEPs regulating MMP-1 expression: one reported a decrease (Doornaert et al. 2003), and the other an increase (Amara et al. 2007). We differ from the study report ing a MMP1 decrease, a discrepancy that could possibly be related to particle, cell, or MMP1 ELISA technology. Our results are consistent with the other study showing DEPs increased MMP1 activation based on an NADP(H) oxidase-dependent pathway (for the relevance of DEPs evoking oxidant mediated injury, see Amara et al. 2007;Bayram et al. 2006;Cao et al. 2007b;Ghio et al. 2000;Madden et al. 2000;Zhang et al. 2004; for using tumorderived alveolar cells, see Amara et al. 2007). Here, we have used bronchial epithelial cells, both permanent and primary, with the latter assayed at air-liquid interface. Of note, studies on the regulation of the human MMP-1 gene cannot readily be complemented by studies in mice (or rats), because rodents do not have a valid ortholog of this gene (Brinckerhoff and Matrisian  . In addition, our DEP preparation con tained an increased concentration of organic components compared with the DEPs used in the study with tumorderived alveolar cells (Amara et al. 2007), thus providing a stimulus more realistically reflecting the exhausts from road, rail, and ship traffic. Three particular aspects of the intra cellular signaling mechanisms leading to activation of MMP-1 deserve comment. First, we have shown that the -1607GG MMP-1 poly morphism in the 5´ regulatory region of the human MMP-1 gene exerts a dominant influ ence on MMP-1 transcription and is more effective than the -1607G promoter. From a public health perspective, this is highly rele vant because, regardless of ethnicity, at least one copy of the -1607GG allele is present in 60-80% of humans (Fujimoto et al. 2002). This finding suggests an enhanced susceptibil ity to the detrimental effects of DEP inhala tion, possibly together with tobacco smoking (Mercer et al. 2004(Mercer et al. , 2008TorresDuque et al. 2008), which leads to activation of MMP-1, in most people that populate habitats with significant DEP exposure (Fujimoto et al. 2002).
Second, we have extended our mechanis tic understanding of the role of the MEK ERK1/2 pathway in upregulating the MMP-1 gene in response to DEPs, first reported by Boczkowski's group (Amara et al. 2007). For the first time, we demonstrate an essential role for RAF and RAS in this signaling. We have identified the necessity of MEKERK1/2 in primary HBE and shown its critical influence on transcriptional activation of the MMP-1 gene that is potentiated by the -1607GG allele. In addition, we have shown that after DEP exposure, within 30 min, newly gener ated phosphoERK relocates to the nucleus. Intracellular signaling cascades do not simply operate in a diffusely solubilized environment within the cytoplasm, but are functionally tightly organized by scaffolding proteins, which constitute specific "signalo somes" for any signaling cascade that func tions in response to a specific stimulus (DeWire et al. 2007;GestyPalmer et al. 2005). In this respect, our third contribution is that the cellular response to DEPs is depen dent on cellgrowth-related MAP kinases that are dependent on both βarrestins. In contrast to the welldocumented concept of βarrestin-mediated scaffolding of MAP kinase signaling with cytoplasmic retention of phosphoERK (Ahn et al. 2003;DeWire et al. 2007), we have observed nuclear trans location of phosphoERK in the presence of βarrestin1 and 2, in agreement with a recent study that reported nuclear translocation of phosphoERK dependent on βarrestin2 in immortalized cultured cells transfected with the β 2 adrenoceptor (Kobayashi et al. 2005).
Our finding that both βarrestins are closely linked to this MAP kinase pathway has not been previously reported. Mechanistically, in the absence of both βarrestins, phospho ERK1/2 was not generated ( Figure 5D,E). The study by Dasgupta et al. (2006) serves as an interesting comparison. They examined the role of βarrestin1-mediated scaffolding in lung tumor cells in response to nicotine, with the outcome of cell growth. siRNAmediated knockdown of βarrestin1 led to nonforma tion of phosphoERK in response to nicotine, with subsequent elimination of the growth response. βArrestin1 was recruited to the nicotinergic acetylcholine receptor, which was instrumental for downstream signaling leading to activation of RAF. βArrestins might func tion similarly in bronchial epithelia exposed to DEPs, but a) both βarrestins were nec essary for the DEP MMP1 response, not specifically one of them; b) cell lines in the study of Dasgupta et al. (2006) were derived from bronchial tumors because the focus of their study was cell growth of such tumors in response to nicotine, a single compound that has a specific cognate cell surface receptor; and c) Dasgupta et al. (2006) provided evidence of a functional and dynamic protein-protein interaction of the nicotinergic acetylcholine receptor with βarrestin1, but based on their data, an additional role for βarrestin1 in scaf folding RAFMEK cannot be excluded.
Regarding the potential human health issues associated with this study, RASmediated MAP kinase signaling is a growthrelated pathway also known to be dysfunctionally activated in malignant transformation of tumors (Roberts and Der 2007;Sridhar et al. 2005). In this context, we note that patho genesis of bronchial cancer is also linked to extracellularmatrix-degrading properties of MMP1, which includes tumor cell growth, invasion, and metastatic capability (Pritchard et al. 2001;Rutter et al. 1998). Our find ings in cultured airway epithelia can thus be viewed as a reflection of the wellknown epidemiologic association between breathing polluted, DEPcontaining urban smog and increased incidence of bronchial cancer in humans (Hemminki and Pershagen 1994;McClellan et al. 1985;Parent et al. 2007). The findings from this study position us favor ably to ask whether known cancer risk co factors, namely, cigarette smoking and genetic predisposition, enhance MMP-1 activation in combination with DEP exposure. With respect to cigarette smoking, smoke compo nents can enhance MMP-1 transcription via the -1607GG polymorphism. In the study by Mercer et al. (2008), the distal 1kb of the same 4.4kb promoter that we have used here proved to be essential for activation of the MMP-1 gene by cigarette smoke extract, a "chemicotoxicologic library" of several thou sand compounds, in this respect similar to DEPbound chemicals. As in our study, there was increased GGallele stimulusresponsive activity. Remarkably in contrast, however, the "tobaccoresponsive elements" from -2.9 to -4.4kb were critical for an increase in activ ity in response to cigarette smoke extract, whereas our results point to a strikingly repressive mode of action [see Supplemental Material, Figure 3 (http://www.ehponline. org/members/2008/0800311/suppl.pdf)]. When Mercer et al. (2008) conducted dele tion, bioinformatics, and transcriptionfactor binding studies on the distal segment of the 4.4kb MMP-1 promoter, they found PEA3 transcription factors acting as robust repres sors, with binding sites at a "tandem" position (-3838 and -3824 relative to transcriptional start site). PEA3 and its respective tandem binding sites are attractive candidates to func tion as repressors of MMP-1 transcription in the response to DEPs. However, when comparing the response profile of the differ ent reporter gene constructs of Mercer et al.'s study with ours, it is apparent that detailed transcriptional regulation of MMP-1 reveals different mechanisms, dependent on the airwayinjury-inducing stimulus. Regarding genetic factors as lung cancer risk factors, 30% of all bronchial carcinomas have muta tions in RAS that render its activity more or even constitutively active (Roberts and Der 2007;Sridhar et al. 2005).
In summary, we have elucidated signaling mechanisms operative in HBE and how the diseaseenhancing MMP-1 gene is activated in response to DEPs. We provide evidence that in primary HBE cells the growthrelated MAP kinase signaling pathway is critical for DEPevoked upregulation of MMP-1 and subsequent secretion of MMP1. We also present, for the first time, data that this regu lation depends on activation of RAS and RAF as well as both isoforms of βarrestin, which we found necessary for formation of phospho ERK1/2. In the presence of both βarrestin isoforms, phosphoERK1/2 translocates to the nucleus, peaking at 30 min. This cascade of events leads to transcriptional activation of MMP-1 that is significantly more robust for the -1607GG MMP-1 promoter poly morphism, present in most humans.
In terms of translational medical implica tions, our findings suggest the potential for topi cal delivery of compounds to human air ways that can downregulate RASinduced MAP kinase signaling, culminating in the acti vation of MMP-1, which might be of benefit for patients at risk of developing ultimately dev astating respiratory diseases linked to MMP-1 dysregulation . Such topical applications, in order to inhibit MAP kinase signaling, have also been proposed recently for cystic fibrosis patients Roque et al. 2008).