Linking Oxidative Events to Inflammatory and Adaptive Gene Expression Induced by Exposure to an Organic Particulate Matter Component

Background: Toxicological studies have correlated inflammatory effects of diesel exhaust particles (DEP) with its organic constituents, such as the organic electrophile 1,2-naphthoquinone (1,2-NQ). Objective: To elucidate the mechanisms involved in 1,2-NQ–induced inflammatory responses, we examined the role of oxidant stress in 1,2-NQ–induced expression of inflammatory and adaptive genes in a human airway epithelial cell line. Methods: We measured cytosolic redox status and hydrogen peroxide (H2O2) in living cells using the genetically encoded green fluorescent protein (GFP)-based fluorescent indicators roGFP2 and HyPer, respectively. Expression of interleukin-8 (IL-8), cyclooxygenase-2 (COX-2), and heme oxygenase-1 (HO-1) mRNA was measured in BEAS-2B cells exposed to 1,2-NQ for 1–4 hr. Catalase overexpression and metabolic inhibitors were used to determine the role of redox changes and H2O2 in 1,2-NQ–induced gene expression. Results: Cells expressing roGFP2 and HyPer showed a rapid loss of redox potential and an increase in H2O2 of mitochondrial origin following exposure to 1,2-NQ. Overexpression of catalase diminished the H2O2-dependent signal but not the 1,2-NQ–induced loss of reducing potential. Catalase overexpression and inhibitors of mitochondrial respiration diminished elevations in IL-8 and COX-2 induced by exposure to 1,2-NQ, but potentiated HO-1 mRNA levels in BEAS cells. Conclusion: These data show that 1,2-NQ exposure induces mitochondrial production of H2O2 that mediates the expression of inflammatory genes, but not the concurrent loss of reducing redox potential in BEAS cells. 1,2-NQ exposure also causes marked expression of HO-1 that appears to be enhanced by suppression of H2O2. These findings shed light into the oxidant-dependent events that underlie cellular responses to environmental electrophiles.


Research
Oxidant stress is a commonly described mecha nistic feature of the toxicity of environ mental contaminants (Monks et al. 1992). Multiple patho physiological effects of environ mental exposures, including cancer, fibrosis, and inflammation, have been associated with oxidant damage to macro molecules such as lipids, proteins, and DNA (Kelly et al. 1998). Oxidant stress induced by a toxicant is invariably a multi faceted process involving exogenous and endogenous reactions between xeno biotic and cellular macro molecules. Toxic exposures often elicit cellular responses that are intrinsically oxidant in that they involve production of reactive oxygen species (ROS) and/or the loss of intra cellular reducing potential. Oxidative cellu lar responses to exposure to oxidizing agents can also occur, and thus the elucidation of the events involved and the order in which they occur presents significant analytical challenges (Ercal et al. 2001;SantaMaria et al. 2005;Steinberg et al. 1990;Valko et al. 2005).
Oxidant stress is believed to play an important role in air pollutant-mediated tox icity in the respiratory tract. Transition metals and organic chemical components of diesel exhaust particles (DEP) have been shown to induce the generation of various ROS (Monks et al. 1992), including the super oxide radi cal, hydrogen peroxide (H 2 O 2 ), and nitric oxide (Kumagai et al. 1997;Li et al. 2003). The relation ship between oxidative stress and altered expression of inflammatory and adaptive genes has been well established for a variety of air pollutants (Becker et al. 2005;Rahman and MacNee 2000).
Although established methods for the measure ment of oxidant damage to cells and tissues exist, they are relatively insensi tive and often provide only inferential mecha nistic information. In contrast, detection of primary oxidative events resulting from environ mental exposures is inherently chal lenging because of the transient nature of the events involved, as well as the relatively low levels of oxidant reactants that are generated. Imaging approaches offer the distinct advan tages of providing high temporal and spatial resolution, as well as the high sensitivity nec essary to detect early indicators of oxidative stress in cells exposed to environ mental agents.
Recently, we described an integrated imaging approach for the realtime measurement of redox potential changes and H 2 O 2 genera tion resulting from mitochondrial dysfunction in living cells exposed to the non redoxactive transition metal Zn 2+ (Cheng et al. 2010). In the present study, we expanded this approach to include an investigation of the relation ship between specific oxidant events in the cytosol and mitochondria and altered gene expression induced by the redoxactive air contaminant, 1,2naphthoquinone (1,2NQ).
1,2NQ, a reactive electrophile associated with diesel exhaust particles (DEP) (Bai et al. 2001;Rodriguez et al. 2004), has been shown to have cyto toxic, mutagenic, and immuno toxic effects (Monks et al. 1992). Quinone toxicity has been found to involve two primary initiating mechanisms: a) a 1,4Michael addition reaction leading to covalent modification of cellular targets (Endo et al. 2007;Miura et al. 2011) and b) ROS genera tion through redox cycling . Previous studies have shown that 1,2NQ attacks protein-tyrosine phosphatases (Iwamoto et al. 2007;Kikuno et al. 2006;Sun et al. 2006), which has been associated with the activation of signaling pathways that can lead to the expression of pro inflammatory proteins such as interleukin8 (IL8) and cyclo oxygenase2 (COX2) (Kuwahara et al. 2006;Tsatsanis et al. 2006) and the adaptive protein HO1 (Kuroda et al. 2010). Multiple studies have suggested a role for ROS generation and inflammatory processes, but the link between oxidant stress Background: Toxicological studies have correlated inflammatory effects of diesel exhaust particles (DEP) with its organic constituents, such as the organic electrophile 1,2-naphthoquinone (1,2-NQ). oBjective: To elucidate the mechanisms involved in 1,2-NQ-induced inflammatory responses, we examined the role of oxidant stress in 1,2-NQ-induced expression of inflammatory and adaptive genes in a human airway epithelial cell line. Methods: We measured cytosolic redox status and hydrogen peroxide (H 2 O 2 ) in living cells using the genetically encoded green fluorescent protein (GFP)-based fluorescent indicators roGFP2 and HyPer, respectively. Expression of interleukin-8 (IL-8), cyclooxygenase-2 (COX-2), and heme oxygenase-1 (HO-1) mRNA was measured in BEAS-2B cells exposed to 1,2-NQ for 1-4 hr. Catalase over expression and metabolic inhibitors were used to determine the role of redox changes and H 2 O 2 in 1,2-NQ-induced gene expression. results: Cells expressing roGFP2 and HyPer showed a rapid loss of redox potential and an increase in H 2 O 2 of mitochondrial origin following exposure to 1,2-NQ. Over expression of catalase diminished the H 2 O 2 -dependent signal but not the 1,2-NQ-induced loss of reducing potential. Catalase over expression and inhibitors of mitochondrial respiration diminished elevations in IL-8 and COX-2 induced by exposure to 1,2-NQ, but potentiated HO-1 mRNA levels in BEAS cells. conclusion: These data show that 1,2-NQ exposure induces mitochondrial production of H 2 O 2 that mediates the expression of inflammatory genes, but not the concurrent loss of reducing redox potential in BEAS cells. 1,2-NQ exposure also causes marked expression of HO-1 that appears to be enhanced by suppression of H 2 O 2 . These findings shed light into the oxidant-dependent events that underlie cellular responses to environmental electrophiles. key words: confocal microscopy, hydrogen peroxide, mitochondrial dysfunction, oxidative stress, quinones, reactive oxygen species, real-time imaging, ROS. Here we report that exposure to 1,2NQ results in a rapid loss of intra cellular reducing potential and increased production of H 2 O 2 of mitochondrial origin, and that these end points associate differentially with the induction of inflammatory and adaptive gene expression.

Materials and Methods
Reagents. Tissue culture media and supple ments were obtained from Lonza (Walkersville, MD). Adenoviral vectors were procured from the Gene Therapy Center Virus Vector Core Facility (University of North Carolina at Chapel Hill). Common laboratory reagents were obtained from Sigma Chemical Co. (St. Louis, MO). Basic laboratory sup plies were purchased from Fisher Scientific (Raleigh, NC).
Synthesis of fluorescent reporter genes in lenti viral vector. The genetically encoded fluorescent reporter roGFP2 is a redox sensitive ratio metric probe established for detection of oxidative stress in the cytosol and mitochondria (Hanson et al. 2004). The plasmid for this protein was a generous gift from S.J. Remington (University of Oregon, Eugene, OR). HyPer is a genetically encoded probe specific for H 2 O 2 detection and was purchased from Evrogen (Axxora, San Diego, CA). The two genes, roGFP2 and HyPer, were isolated from pEGFPN1 and pQE30 vector by BamHI and HindIII digest and cloned into the lenti viral transfer vector pTLRED [U.S. Environmental Protection Agency (EPA)]. HEK293T cells were cotransfected with puri fied transfer vector plasmids and lenti viral packing mix (Open Biosystems, Huntsville, AL). The resulting supernatants from the individual transfections were concentrated  once by lowspeed centrifugation through an Amicon Ultra 100kD centrifuge filter unit (Millipore, Billerica, MA), and the retentates were aliquoted and stored at -80°C. Viral titers were determined in HEK293T cells stably expressing the rTTA3 trans activator (E10 cells) by transduction with serially diluted vector stocks as previously described (Simmons et al. 2011). Cell culture and viral transduction. Transformed human airway epithelial cells (BEAS2B) (Reddel et al. 1988) were main tained in serumfree keratino cyte growth medium (KGMGold; Lonza). For imaging purposes, BEAS2B cells grown to 50% con fluency were transduced with lenti viral vectors carrying roGFP2 or HyPer genes targeting them to either the cytosol (roGFP2cyto and HyPercyto) or mitochondria under the multi plicity of infection (MOI) of 5, as previously described . For catalase over expression, BEAS2B cells were transduced with an adenoviral vector encoding human catalase (AdCAT), green fluorescent protein (AdGFP), or empty vector for 4 hr using an MOI of 100. The adenoviral constructs were removed after transduction, and the cells were passaged in KGMGold.
Cell exposure. Growth factordeprived BEAS2B cells were exposed to DMSO con trol or 10-150 µM 1,2NQ for 0-4 hr. Cells expressing roGFP2 or HyPer were treated under observation with a Nikon Eclipse C1Si confocal imaging system (Nikon Instruments Inc., Melville, NY). In separate experi ments, cells were analyzed using a PolarStar Optima microplate reader (BMG Labtech, Durham, NC) prior to and during treatment with 1,2NQ. For gene expression analyses, BEAS2B cells were exposed to 1-10 µM 1,2NQ for 4 hr, and changes in the levels of specific transcripts were analyzed using real time polymerase chain reaction (RTPCR).

Measurement of redox potential and H 2 O 2 .
Confocal microscopy analy ses were conducted using a C1Si system equipped with an Eclipse Ti microscope (Nikon). Green fluorescence was derived from excitations at 404 and 488 nm, and emission was detected using a bandpass filter of 525/50 nm (Chroma, Bellows Falls, VT). The results were calculated as ratios of the emissions excited by 488 nm and 404 nm lasers sequentially with a scanning frequency of 60 sec. The optical settings for the plate reader were similar to those used in the microscope, with excitation at 485/12 nm and 400/10 nm and emission at 520/30 nm (Chroma).
RT-PCR. Subconfluent BEAS2B cells were exposed to varying concentrations of 1,2NQ for 0-4 hr. Relative gene expres sion in BEAS2B cells was quantified using the realtime PCR, ABI Prism 7500 Sequence Detection System (Applied Biosystems, Foster City, CA). Total RNA was isolated using an RNeasy kit (Qiagen, Valencia, CA) and reverse transcribed to generate cDNA using a High Capacity cDNA Reverse Transcription kit (Applied Biosystems). Oligonucleotide primer pairs and duallabeled fluorescent probes for IL-8, COX-2, heme oxygenase1 (HO-1), β-actin, and catalase were obtained from Applied Biosystems. The relative abundance of mRNA levels was determined using TaqMan Universal Master Mix (Applied Biosystems) and the 2 −ΔΔCT method (Livak and Schmittgen 2001). β-Actin mRNA was used to normalize levels of the mRNAs of interest.
Statistical analysis. Imaging data were collected with Nikon EZC1 software. An average of 5-10 cells was collected as regions of interests in each experiment, and data were quantified using Nikon Elements software (Nikon). Data are expressed as mean ± SE of three repeated experiments. The linear regres sion of plate reader results was calculated with GraphPad Prism (GraphPad Software, La Jolla, CA), and the slope of the regression line was plotted against 1,2NQ concentra tions. Pairwise comparisons were carried out using Student's ttest, with p < 0.05 taken as statistically significant.

1,2-NQ induces rapid oxidant changes.
We used the genetically encoded fluorescent probes roGFPcyto and HyPercyto to moni tor changes in redox potential and H 2 O 2 production, respectively, in BEAS2B cells exposed to 1,2NQ. The cells were observed for 5 min to establish a baseline signal prior to treatment with either vehicle control or 100 µM 1,2NQ for 15 min. As shown in Figure 1, treatment with 100 µM 1,2NQ induced a rapid increase in the ratiometric fluorescence intensity of cytosolic roGFP2, corresponding to a marked loss of intra cellular reducing potential that peaked and stabilized at 20 min ( Figure 1C-E). The intra cellular redox potential in cells exposed to vehicle alone remained stable during the same time period ( Figure 1E). Similarly, cells expressing HyPercyto responded with an increase in fluorescence ratio intensity, indicating elevated levels of H 2 O 2 after exposure to 1,2NQ compared with control cells (Figure 1H-J).

Overexpression of catalase blunts 1,2-NQinduced H 2 O 2 production.
To explore the inter action between changes in redox potential and H 2 O 2 generation, we studied the effect of 1,2NQ in BEAS2B cells over expressing catalase. Preliminary experi ments established that catalase mRNA levels were 4 times higher in BEAS2B cells transduced with AdCAT compared with controls (data not shown). Treatment of catalase overexpressing BEAS2B cells with 100 µM 1,2NQ induced a loss of reducing potential that was not significantly different from that observed in BEAS2B cells transduced with an empty vector (Figure 2A-E). In contrast, 1,2NQ-induced H 2 O 2 production was effectively ablated in BEAS2B cells over expressing catalase ( Figure 2F-J).
Overexpression of catalase differentially inhibits 1,2-NQ-induced gene expression. We next examined the effect of 1,2NQ exposure on the expression of the pro inflammatory genes IL-8 and COX-2 and the adaptive, oxidant responsive gene HO-1. Exposure of BEAS2B cells to 1-10 µM 1,2NQ or vehicle for 0-4 hr resulted in dose and time dependent inductions in IL-8, COX-2, and HO-1 mRNA ( Figure 3A-F), with maximal respective increases of 5, 4, and 30fold relative to vehicle controls observed at 4 hr of exposure. To test the mechanistic link between gene expression and oxidant responses, we determined the effect of 1,2NQ exposure on the induction of IL-8, COX-2, and HO-1 transcripts in BEAS2B cells over expressing catalase. Relative to control cells transduced with AdGFP, over expression of catalase blunted the increases in IL-8 and COX-2 mRNA induced by treatment with 10 µM 1,2NQ for 4 hr ( Figure 3G,H). However, the induction of HO-1 gene expression by 1,2NQ was significantly augmented in BEAS2B cells that over expressed catalase ( Figure 3I), indicating a differential role for H 2 O 2 in 1,2NQ-induced inflammatory and adaptive gene expression.

1,2-NQ induces intra cellular production of H 2 O 2.
To identify the source of H 2 O 2 production shown in Figure 1, we investigated possible mechanisms through which 1,2NQ exposure of BEAS2B cells could result in the generation of H 2 O 2 . To investigate the possibility that 1,2NQ generates H 2 O 2 extra cellularly, we used a plate reader assay to monitor fluorescence changes in BEAS2B cells expressing roGFPcyto or HyPer cyto with various 1,2NQ concentrations (10-150 µM) in the presence or absence of exogenous catalase. As shown in Figure 4B, the inclusion of extra cellular catalase did not significantly affect the magnitude or time of onset of 1,2NQ-induced H 2 O 2 generation in BEAS2B cells, as detected by HyPercyto. However, in agreement with the microscopy findings shown in Figure 2, adenoviralmediated over expression of catalase in BEAS2B cells ablated H 2 O 2 production induced by 1,2NQ treatment ( Figure 4B). Neither extra cellular catalase nor over expression of catalase had any effect on the loss of cyto plasmicreducing potential observed in roGFPcyto expressing BEAS2B cells treated with 1,2NQ ( Figure 4A). Figure 4 indicated that 1,2NQ exposure elevates the intra cellular concentra tion of H 2 O 2 , suggesting the involvement of a cellular process. We therefore examined potential cellular sources of H 2 O 2 genera tion in 1,2NQtreated cells. We first tested the involvement of H 2 O 2 generation at the cell membrane by pre treating the cells with the specific NADPH oxidoreductase inhibi tor DPI 30 min prior to the addition of 10-150 µM 1,2NQ. We observed no sig nificant differences in the production of H 2 O 2 in cells exposed to 1,2NQ in the presence of DPI relative to cells pretreated with vehicle alone ( Figure 4D). We therefore turned our attention to possible mito chondrial sources of H 2 O 2 , using the mitochondrial inhibitors CCCP, NaN 3 , KCN, CyA, and rotenone. Of these inhibitors, CCCP (a mitochon drial membrane potential uncoupler) and rotenone (a mitochondrial complex I inhibi tor) showed an effect on 1,2NQ-induced H 2 O 2 in BEAS2B cells ( Figure 4F,H). None of the inhibitors showed significant effects on 1,2NQ-induced redox changes ( Figure 4C,E,G). These findings implicated the mitochondrial respiratory chain as the source of 1,2NQ-induced H 2 O 2 production.

Identification of the mitochondrion as the source of 1,2-NQ-induced H 2 O 2 . The data shown in
We then examined BEAS2B cells expressing Hypermito, a version of the H 2 O 2 sensor that is targeted to the mitochon drial inner membrane. Exposure to 100 µM 1,2NQ resulted in an elevation of ratio metric HyPermito fluorescence signal inten sity, indicating elevated concentrations of H 2 O 2 in the mitochondria ( Figure 5A,B,E).
1,2NQ-induced production of mitochon drial H 2 O 2 was effectively suppressed by pre treatment of the cells with 10 µM CCCP ( Figure 5C-E). These data showed 1,2NQinduced generation of H 2 O 2 in the mitochon drion and further established mitochondrial respiration as the source of H 2 O 2 production.

1,2-NQ-induced gene expression is differentially linked to mitochondrial activity and H 2 O 2 availability.
We examined the role of mitochondrial metabolism in 1,2NQinduced inflammatory and adaptive gene Rotenone inhibited the induction of IL-8 and COX-2 expression by 1,2NQ ( Figure 6A,B). In marked contrast, the induction of HO-1 mRNA by 1,2NQ was potentiated by rote none pretreatment ( Figure 6C). This find ing, combined with the earlier observation that catalase over expression also enhanced the induction of HO-1 mRNA by 1,2NQ, led us to hypothe size that H 2 O 2 limits 1,2NQinduced increases in HO-1 mRNA. We there fore tested this hypothesis directly by adding 30 µM H 2 O 2 immediately before 1,2NQ treatment of BEAS2B cells. As shown in Figure 6F, the addition of exogenous H 2 O 2 significantly blunted the induction of HO-1 expression by 1,2NQ ( Figure 6F). H 2 O 2 pre treatment had no effect on 1,2NQ-induced IL-8 and COX-2 expression ( Figure 6D,E).

Discussion
Oxidant effects are a commonly reported mechanistic feature of the toxicity of environ mental agents. In the present study, we expanded our previously described integrated imaging approach to the investigation of mito chondrial dysfunction (Cheng et al. 2010) to include inflammatory and adaptive gene expres sion changes induced by an environ mental electro phile capable of inducing multiple types of oxidant stress. This study presents a mecha nistic link between early oxidant events result ing from exposure to 1,2NQ and downstream toxicological effects, specifically alterations in the expression of genes involved in inflamma tory and adaptive responses in a human bron chial epithelial cell line. In preliminary studies we observed similar oxidant responses and changes in gene expression in primary cultures of human airway epithelium (Cheng WY, unpublished data). As one of the organic components of the ubiquitous air contaminant DEP Inoue et al. 2007b;Jakober et al. 2007), 1,2NQ has been shown to induce air way inflammation and to initiate deleterious effects through covalent modification or ROS generation (Inoue et al. 2007a(Inoue et al. , 2007b. Both activating and inhibitory effects of 1,2NQ have been reported. For instance, Kikuno et al. (2006) reported that 1,2NQ induces vanilloid receptor and epidermal growth factor receptor signaling, leading to guinea pig tracheal con traction. Inhibitory signaling effects associated with 1,2NQ include impairment of cAMP response element-binding protein (CREB) (Endo et al. 2007) and lipopolysaccharide induced nuclear factor kappa B (NFκB) DNA binding activities (Sumi et al. 2010). In addi tion, the cytoplasm, endoplasmic reticulum, nucleus, and mitochondrion are all major tar gets for 1,2NQ-induced toxicity through protein modification in lung epithelial cells (Lame et al. 2003). Thus, the high reactivity of 1,2NQ can result in a diversity of molecular effects that are likely dependent on concentra tion and also show cell type specificity.
In this study, we used the genetically encoded fluorescence reporters roGFP2 and HyPer to detect redox changes and H 2 O 2 production, respectively. The exposure of 1,2NQ induced rapid responses in both roGFP2 and HyPer in the cytosol of BEAS2B cells, indicating an acute oxidative burden stimulated by this compound. The generation of ROS and changes in redox balance can be seen as related events. However, the observation that catalase over expression blunted the 1,2NQ-induced increase in H 2 O 2 production Vehicle CCCP without affecting the changes in redox potential suggests that H 2 O 2 production is not the cause of the redox changes. Furthermore, over expression of catalase also protected against 1,2NQ-induced IL-8 and COX-2 expression, indicating that 1,2NQ-stimulated H 2 O 2 production is involved in the induction of inflammatory responses. This is in agreement with reports of the involvement of H 2 O 2 in the activation of signaling pathways that regulate pro inflammatory genes, such as NFκB, p38, and JNK (Groeger et al. 2009). However, the addition of 30 µM H 2 O 2 did not induce a statistically significant increase in IL-8 or COX-2 expression. This may reflect a requirement for H 2 O 2 to act as a second messenger at specific sub cellular compartments in order to initiate inflammatory gene expression. An unexpected finding is that 1,2NQinduced HO-1 expression in BEAS2B cells was not mediated by H 2 O 2 . On the contrary, the magnitude of HO-1 induction by 1,2NQ was enhanced by removal of H 2 O 2 . Specifically, catalase expression and impairment of mitochondrial electron transport, which effectively decrease H 2 O 2 concentrations and production, respectively, both potentiated 1,2NQ-induced increases in HO-1 mRNA. Furthermore, direct evidence for the suppressive effect of H 2 O 2 on 1,2NQ-induced HO-1 expression was also obtained using exogenous H 2 O 2 . A similar finding was reported by Miura et al. (2011), who showed that pretreatment with catalase did not protect against 1,2NQinduced activation of nuclear factor (erythroid derived 2)like 2 (Nrf2), which is a regulator of HO-1 gene expression. This is a seemingly paradoxical finding, as H 2 O 2 is a known inducer of the Nrf2 pathway that regulates HO-1 expression (Fourquet et al. 2010). One explanation for these observations may be that 1,2NQ-induced HO-1 expression requires electrophilic attack on a susceptible regulatory target, possibly a protein thiol, that is rendered unreactive to 1,2NQ when oxidized by H 2 O 2 .
A parallel for H 2 O 2 mediated inactiva tion of protein thiols is found in redox regu la tion of protein tyrosine phosphatases, in which the cysteine thiolate in the catalytic center of the enzyme is reversibly oxidized by H 2 O 2 (Samet and Tal 2010). Using benzo quinone as the model toxi cant, Mason and Liebler (2000) observed cysteine as a preferred target for quinoneinduced toxicity. Recently, Miura et al. (2011) reported that Nrf2 acti vation by 1,2NQ was mediated by covalent modification and subsequent degradation of Keap1. These studies point to cellular cysteine thiol groups as primary targets of electro philic naphtho quinone attack by covalent modifica tion (Lame et al. 2003). From this perspective, it is intriguing that 1,2NQ has been shown to attack and inactivate the protein tyrosine phosphatase PTP1B, albeit at an allosteric site (Iwamoto et al. 2007). These observations lead us to speculate that bio molecular cova lent modifications by 1,2NQ are involved in HO-1 gene expression induced by electrophilic attack. Detailed studies will be needed to elu cidate the signaling mechanisms that underlie 1,2NQ-induced gene expression.
A variety of metabolic processes are potential targets for xenobioticinduced ROS production. Although quinone species that undergo redox cycling can generate ROS in cellfree aqueous environments (Le et al. 2007), the lack of an effect of extra cellular catalase in suppressing the 1,2NQ-induced HyPer signal excluded an extra cellular redox process as a source of the H 2 O 2 . The presence of exogenous catalase would also be expected to scavenge H 2 O 2 generated by membrane oxido reductases because NADPH oxidases generate H 2 O 2 in the extra cellular space . The failure of the oxido reductase activity inhibitor DPI to suppress HyPer signals is consistent with this notion and thus helped shift the focus to the mitochondria as a source of 1,2NQinduced H 2 O 2 in this study.
The observation of H 2 O 2 dependent fluo rescence in the mitochondria confirmed that the mitochondrion is the site of H 2 O 2 production Figure 6. Differential role of mitochondrial H 2 O 2 in 1,2-NQ-induced gene expression in BEAS-2B cells pretreated with DMSO vehicle, the mitochondrial complex I inhibitor rotenone (Rot; 10 µM, 30 min), or H 2 O 2 (30 µM, 10 sec) prior to the addition of 10 µM 1,2-NQ for 4 hr. mRNA levels of IL-8 (A,D), COX-2 (B,E), and HO-1 (C,F) were measured using TaqMan-based RT-PCR, normalized to levels of β-actin mRNA, and expressed as fold increases over vehicle control (mean ± SE; n = 3). **p < 0.01. ? X X X X X volume 120 | number 2 | February 2012 • Environmental Health Perspectives in BEAS2B cells exposed to 1,2NQ. Of the variety of mitochondrial inhibitors used in this study that target membrane potential (CCCP), complex I (rotenone), complex IV (KCN and NaN 3 ), and the permeability transition pore (CyA), only CCCP and rotenone blunted 1,2NQ-induced HyPer signals, indicating that the molecular target for 1,2NQ-stimulated H 2 O 2 is associated with components of the upstream mitochondrial respiratory chain. A similar mitochondrial dysfunction was observed by Xia et al. (2004) who exposed a mouse mac rophage cell line to a quinoneenriched polar fraction of DEP. Furthermore, in the present study, pre treatment with rotenone diminished 1,2NQ-induced IL-8 and COX-2 gene expres sion, further establishing the functional link between the formation of mitochondrial H 2 O 2 and inflammatory gene expression.
The ambient concentration of 1,2NQ has been reported to range from 13 to 53 µg/g DEP Valavanidis et al. 2006). Given the ubiquitous nature of DEP as a constituent of ambient particulate mat ter (PM), plausible realworld scenarios may result in exposure of airway epithelial cells to deposited doses of 1,2NQ during a 3hr inha lational exposure that are about 10fold lower than those used in this study [see Supplemental Material (http://dx.doi.org/10.1289/ ehp.1104055) for supporting calculations and assumptions]. Moreover, 1,2NQ is repre sentative of a class of organic constituents of ambient PM that includes other quinones as well as poly aromatic hydro carbons that may be metabolized to redox active quinones Valavanidis et al. 2006).
Most studies on environmental electro philes such as 1,2NQ have focused on the highly reactive electrophilic properties of these com pounds. Here we were able to measure early oxidative events in real time and correlate them mechanistically to gene expression changes asso ciated with adverse responses to electrophilic exposure. In this study, we demon strate that 1,2NQ induces mitochondrial H 2 O 2 produc tion that leads to inflammatory gene expres sion but not the accompanying loss of reducing potential observed in the cytosol. 1,2NQ also induces HO-1 expression; however, our data show that it does so through a mechanism that is actually opposed by the availability of H 2 O 2 . Thus, these findings reveal dissociation between H 2 O 2 production and the loss of reduc ing potential induced by a frank electro phile (Figure 7). Ascertaining whether the loss of reducing potential is a consequence or a cause in the induction of HO1 by 1,2NQ requires further investigation. Taken as a whole, our experimental strategy in this study represents an integrated approach for the systematic study of oxidative events that underlie adverse cel lular responses to xeno biotic exposure. From a public health perspective, the inflammatory and adaptive responses induced by 1,2NQ are consistent with the inflammatory and immuno toxic effects that are associated with human exposure to DEP and ambient PM.