Nitrated Fatty Acids Reverse Cigarette Smoke-Induced Alveolar Macrophage Activation and Inhibit Protease Activity via Electrophilic S-Alkylation

Nitrated fatty acids (NFAs), endogenous products of nonenzymatic reactions of NO-derived reactive nitrogen species with unsaturated fatty acids, exhibit substantial anti-inflammatory activities. They are both reversible electrophiles and peroxisome proliferator-activated receptor γ (PPARγ) agonists, but the physiological implications of their electrophilic activity are poorly understood. We tested their effects on inflammatory and emphysema-related biomarkers in alveolar macrophages (AMs) of smoke-exposed mice. NFA (10-nitro-oleic acid or 12-nitrolinoleic acid) treatment downregulated expression and activity of the inflammatory transcription factor NF-κB while upregulating those of PPARγ. It also downregulated production of inflammatory cytokines and chemokines and of the protease cathepsin S (Cat S), a key mediator of emphysematous septal destruction. Cat S downregulation was accompanied by decreased AM elastolytic activity, a major mechanism of septal destruction. NFAs downregulated both Cat S expression and activity in AMs of wild-type mice, but only inhibited its activity in AMs of PPARγ knockout mice, pointing to a PPARγ-independent mechanism of enzyme inhibition. We hypothesized that this mechanism was electrophilic S-alkylation of target Cat S cysteines, and found that NFAs bind directly to Cat S following treatment of intact AMs and, as suggested by in silico modeling and calculation of relevant parameters, elicit S-alkylation of Cys25 when incubated with purified Cat S. These results demonstrate that NFAs’ electrophilic activity, in addition to their role as PPARγ agonists, underlies their protective effects in chronic obstructive pulmonary disease (COPD) and support their therapeutic potential in this disease.


Introduction
Chronic obstructive pulmonary disease (COPD) is the third leading cause of death in the US [1], characterized by chronic inflammation of small airways and destruction of alveolar septa with consequent emphysema [2]. Such septal destruction reduces the surface area available for pulmonary O 2 and CO 2 exchange, and combines with airway inflammation and mucus production to impair respiratory mechanics. Because COPD is progressive and often fatal, and no effective therapies are available to impede its course, identification of new potential therapeutic targets is needed.
Among the many cell types that contribute to COPD pathophysiology, alveolar macrophages (AMs) are especially critical [3]. They are directly exposed to cigarette smoke, the major risk factor for COPD, and other noxious airborne agents. Such agents activate AMs to produce pro-inflammatory cytokines that activate other cells, and chemokines that attract neutrophils and T lymphocytes, prominent in COPD-associated inflammation. Such activated and recruited cells likewise secrete mediators that further activate AMs in a vicious positive-feedback cycle. Yet AMs also exert an inflammation-dampening influence by phagocytosing apoptotic neutrophils and epithelial cells that, if allowed to become necrotic, would further feed the inflammatory cycle. Such phagocytotic activity of AMs is reduced in COPD, thereby exacerbating pathogenesis [4].
AMs are also the primary sources of elastin-degrading proteases that largely drive septal destruction [5]. Most recent attention has focused on the Zn-containing matrix metalloproteinases (MMPs), but cathepsins, including cathepsins K, L, and S, are also important mediators of septal degradation [6]. Production of inflammation-associated oxidants by AMs can also contribute to septal destruction.
Inflammation-associated oxidants include NO and derived reactive nitrogen species that react nonenzymatically with unsaturated fatty acids to produce nitrated fatty acids (NFAs), including 10-nitro-oleic acid (OA-NO 2 ) and 12-nitrolinoleic acid (LNO 2 ) [7], the most prevalent NO reaction products in the human bloodstream [8]. NFAs are potentially important endogenous modulators of inflammatory processes [9][10][11], acting at least in part as agonists of the nuclear hormone receptor peroxisome proliferator activated receptor-γ (PPARγ) [12]. They are also reversible electrophiles [13]. As such, they can react with thiol groups including those of cysteine residues within proteins to produce an S-alkyl bond that is cleavable (thus reversible) by reaction with free intracellular thiols such as glutathione (GSH). It has long been observed that specific cysteines are preferentially alkylated, but the factors that render such specific cysteines more readily or stably alkylated are unknown. Known cysteine targets of NFA Salkylation include one within the ligand-binding site of PPARγ that contributes significantly to agonist potency [14].
Reversible Cys S-alkylation provides a unique regulatory mechanism for biologically active proteins. For example, reversible electrophiles regulate oxidative stress by reversibly S-alkylating the inhibitory protein Keap-1, allowing activation of the antioxidant transcription factor Nrf2 [15]. NFAs may also alleviate inflammation by S-alkylating the pro-inflammatory transcription factor NF-κB [10]. No such influence of reversible electrophiles in general or NFAs in particular has been recognized in the pathophysiology of COPD. We hypothesized that antiinflammatory and related actions of NFAs may mitigate COPD pathogenesis, including secretion of the proteases that mediate septal destruction, and tested this idea and the mechanisms involved in AMs from smoke-exposed mice.

Materials and Methods
Animals C57BL/6 and Tie2 Cre-PPARγ flox/flox mice [16] were expanded from breeding pairs. Mice were housed in microisolator cages under specific pathogen-free conditions and fed autoclaved food. Male mice aged 6-8 weeks (20-25 g) were used in all experiments. All studies were performed according to protocols reviewed and approved by the VA Pittsburgh Healthcare System Institutional Animal Care and Use Committee.

Smoke Exposure
Mice were exposed to cage-air cigarette smoke (CS) or to filtered air for 14 days. CS was generated by burning five 3RF4 research cigarettes (Tobacco Research Institute, University of Kentucky, Lexington, KY) according to the Federal Trade Commission (FTC) protocol, each puff being of 2-second duration and 35-ml volume, in an automated TE-10 smoking machine (Teague Enterprises, Davis, CA). The machine was adjusted to produce 89% sidestream and 11% mainstream smoke. The chamber atmosphere was monitored to maintain TPM at 250 mg/m 3 . Twenty-four hours following the last exposure, mice were euthanized and BAL fluid and lungs were collected for further analysis.
BAL Fluid Collection, Cell Count and AM Isolation from Mice BAL fluid was collected by flushing 3 × 1 ml of PBS containing 0.1 mM EDTA into the lung via a tracheal cannula. The pooled BAL fluid was centrifuged at 500 × g at 4°C for 5 min. Pelleted cells were then resuspended in 1 ml of PBS. Total cell number was counted by hemocytometer and differential cell count was performed using cytospin preparations stained with Diff-Quik (Siemens, Newark, DE). AMs were isolated as described previously [17].

Measurement of Cytokine, Chemokine and Cathepsin S Levels
Cell culture medium was collected and stored at -80°C. Levels of TNF-α and KC were measured using ELISA kits (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. Cat S levels in media samples were measured using an ELISA kit (MyBioSource, Inc, San Diego, CA) according to the manufacturer's instructions.

Immunoprecipitation
Total protein extracts were immunoprecipitated using the Dynabeads Protein G Immunoprecipitation kit (Invitrogen, Carlsbad, CA). Antibodies were bound to Dynabeads Protein G, and the Dynabeads-Ab complex was used to precipitate target proteins from the total protein extracts. Precipitates were washed to remove unbound proteins and complexes were eluted. All samples were separated by electrophoresis on SDS-polyacrylamide gels, transferred to membranes, and Western blotting was performed. EDTA), 25 mM DTT and 2.5% Tween 20. Samples were then separated on 5% non-denaturing polyacrylamide gels in 1× Tris-Borate EDTA buffer (130 mM Tris, pH 8.3, 45 mM boric acid, 2.5 mM EDTA). The infrared signal was detected using an Odyssey Infrared Imager (LI-COR).

Monocyte Chemotaxis Assay
Mouse blood monocytes were isolated using an EasySep mouse monocyte enrichment kit (Stem-Cell Technologies Inc, Vancouver, BC, Canada). Briefly, leukocyte pools from blood were incubated with an antibody mixture against non-monocytes followed by a biotin selection mixture and magnetic particles. Labeled cells were removed using DynaMag-2 magnet (Invitrogen), resulting in highly significant enrichment of monocytes in the remaining cell preparation.
Blind-well chemotaxis chambers separated with 8-mm pore polycarbonate membrane (Neuro Probe Inc., Gaithersburg, MD) were used to assess migration of monocytes toward cell culture media. The upper wells contained monocytes isolated as described above; the lower wells were loaded with culture media supernatant from macrophages isolated from mice exposed to cigarette smoke for 14 days and then treated either with vehicle or with OA-NO 2 for 6 h. The chambers were incubated at 37°C for 1 h, after which the membranes were fixed and stained with Diff-Quik stains. The membranes were retrieved and mounted on glass slides, covered with a coverslip, and examined under a microscope. The number of cells that migrated through the membrane into the lower chamber and cells that were attached on the lower side of the membrane were counted and transmigration was expressed as % of monocytes added in upper chamber.

Elastolytic Activity Assay
Intracellular and extracellular elastin degradation studies were performed using soluble and insoluble fluorescein (FITC) labeled elastin respectively (Elastin Products Company, Inc, Owensville, MO). To determine extracellular elastolytic activity insoluble FITC-elastin was added to culture media or BAL fluid and treated as indicated. After incubation, insoluble elastin was removed by centrifugation and fluorescence in supernatants was determined using excitation at 490 nm and emission at 520 nm. Intracellular elastolytic activity was determined by incubating AMs treated as indicated above with soluble FITC-elastin. After incubation cells were washed and mounted with Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA). The slides were viewed by an Olympus Fluoview FV1000 confocal microscope (Olympus, Center Valley, PA) using a 60× objective lens along with Fluoview confocal software (FV10-ASW v1.7, Olympus).

ChIP Assay
The ChIP assay was performed using SimpleChIP Enzymatic Chromatin Immunoprecipitation kit with magnetic beads (Cell Signaling Technology, Beverly, MA, USA). Briefly, cellular  chromatin was crosslinked with 1% formaldehyde for 10 min at room temperature, the crosslinking was stopped with 0.125 M glycine, and cells were washed twice with ice cold PBS. Nuclei were pelleted and digested by micrococcal nuclease. Following sonication and centrifugation, equal amounts of sheared chromatin were incubated overnight at 4°C with antibodies, IgG as negative control, or RNA polymerase II as positive control. Protein G magnetic beads were then added and the chromatin was incubated with rotation for 2 h at 4°C. An aliquot of chromatin that was not incubated with any antibody was used as the input control sample. Antibody-bound protein/DNA complexes were eluted and subjected to real-time PCR as described previously [18] with specific primers for Ctss (Cat S), CD36, and α-satellite ( Table 2).

Molecular Modeling and Computer Simulations of Binding of OA-NO 2 with Cat S
Molecular modeling and covalent docking analysis was carried out as described previously [19]. The X-ray structure of human Cat S {PDB ID 1NPZ; [20]} and its C25S mutant {PDB ID 1GLO; [21]} were obtained from the RCSB Protein Data Bank. Covalent docking was performed with OA-NO 2 (ligand) using Discovery Studio 2.5 (Accelrys Inc., San Diego, CA).

Cat S Enzyme Inhibition Assay
Cat S enzyme inhibition assay was performed using the Cathepsin S Drug Discovery Kit (BML-AK431-0001; Enzo Life Sciences, Farmingdale, NY) according to the manufacturer's instructions.

Cat S Enzyme Activity Assay
Cat S enzyme activity in cell culture media samples was measured by incubating these samples with Cat S-specific fluorescent substrate (Ac-KQKLR-AMC; 600742, Cayman Chemical) according to the manufacturer's instructions.

Cat S Cys Mutants
A plasmid containing the human Cat S gene was obtained from Addgene (#11251) and subcloned into a pGEX 6P1 plasmid. Cysteine residues at positions 12, 25 and 110 or all three (12,25 and 110) were mutated to serine using site directed mutagenesis (custom cloning core facility at Emory University, Atlanta, GA). Recombinant proteins were expressed in BL21-DE3 cells (Sigma-Aldrich, St. Louis, MO), purified using GST SpinTrap columns (GE Healthcare Life Sciences), and the GST tag was removed with PreScission Protease (GE Healthcare Life Sciences) according to the manufacturer's instructions.

Statistical Analysis
Data are presented as mean ± SD. Differences between groups were analyzed using an unpaired t-test or analysis of variance, followed by Bonferroni multiple comparison correction. Analyses were performed using GraphPad Prism 5.03 (GraphPad Software, La Jolla, CA). Differences with P < 0.05 were considered significant.

OA-NO 2 reduces cigarette smoke-induced AM activation in mice
Considering the pivotal role of AMs in the response to smoke inhalation and COPD pathogenesis, we tested the effects of NFAs on AM activities relevant to inflammation and septal destruction in a mouse model of COPD. We exposed mice to CS delivered via cage air for 14 days. Smoke exposure greatly increased the numbers of macrophages in BAL fluid, with lesser increases in the number of neutrophils (Fig 1A and 1B). Because AMs represented the majority of inflammatory cells found in BAL fluid of smoke-exposed mice, we assessed them further.
In vitro OA-NO 2 treatment (no toxicity was observed, S1 Fig) of AMs isolated from BAL fluid of smoke-exposed mice reduced both nuclear expression and DNA-binding activity of the proinflammatory transcription factor NF-κB (Fig 1C and 1D), but increased those of the antiinflammatory nuclear hormone receptor PPARγ (Fig 1C and 1E). Based on these changes in activity of pro-and anti-inflammatory transcriptional regulators, we predicted that expression of pro-inflammatory genes would be reduced, and indeed found that secretion of TNF-α ( Fig 1F) and KC (Fig 1G) by AMs was suppressed. Secretion of chemokines was also reduced, as seen using a monocyte transmigration assay (Fig 1H and 1I). Assessing emphysema-causing protease secretion and elastin degradation, we found that OA-NO 2 treatment markedly reduced secreted Cat S levels and activity in culture medium (Fig 1J and 1K). Degradation of elastin by culture medium of OA-NO 2 -treated AMs (Fig 1L) and intracellular elastin degradation by whole AMs (Fig 1M) were similarly reduced. Thus, OA-NO 2 suppressed both the inflammatory and proteolytic activities induced by smoke exposure in AMs. As these cells play a prominent role in COPD pathophysiology [3], these effects carry significant implications for modulation of this disease.
OA-NO 2 Effects on Cathepsin S Expression, but not Activity, are PPARγ-dependent As NFAs are known to act both as reversible electrophiles and via PPARγ activation, we tested which of these mechanisms was responsible for the observed reductions in Cat S expression and activity using AMs from wild-type (WT) and PPARγ knockout (KO) mice following 14-day in vivo smoke exposure and 6-h in vitro exposure to OA-NO 2 or vehicle. Absence of PPARγ in AMs from KO mice was confirmed by Western blotting, which also revealed modest increases in Cat S (Fig 2A). OA-NO 2 treatment reduced Cat S expression in AMs from WT but not PPARγ KO mice (Fig 2B), indicating that NFA-induced suppression of Cat S expression requires PPARγ activation. In contrast, NFAs reduced Cat S activity in culture media of AMs to approximately the same extent irrespective of the presence or absence of PPARγ ( Fig 2C). Therefore, inhibition of Cat S activity does not depend on PPARγ, and we thus hypothesized it reflects S-alkylation by NFAs. Supporting this hypothesis, we found that NFA bound directly to Cat S following incubation of Cat S with biotin-labeled OA-NO 2 , immunoprecipitation, and Western blotting using anti-biotin antibody (Fig 2E). Conversely, supporting a PPARγ-dependent mechanism of NFA effects on Cat S expression, we found by chromatin immunoprecipitation (ChIP) that OA-NO 2 treatment increased PPARγ binding to both the Cd36 promoter, which is known to be upregulated by PPARγ activation, and also the Ctss (Cat S) promoter ( Fig  2D). This was consistent with prior findings that under certain circumstances binding of PPARγ to DNA inhibits rather than stimulates transcription [22]. Thus, the observed NFA effects on Cat S expression and activity were mediated by distinct, complementary mechanisms in which electrophilic addition plays a prominent role.

OA-NO 2 Inhibits Cathepsin S by S-alkylating Cys25
We next sought to specifically identify the site of NFA-Cat S binding and the type of bond involved. Molecular docking studies revealed covalent bond formation between the electrophilic carbon of OA-NO 2 and sulfur atom Cys25 of Cat S (Fig 3A). No covalent bond Following AM isolation and OA-NO 2 (1 μM) treatment as above, chromatin was crosslinked and immunoprecipitated with PPARγ antibody; the antibody-bound DNA-protein complexes were then subjected to real-time PCR with primers specific for PPRE sites in Cat S and CD36 promoter regions, α-satellite was used as control. (E) As above, AMs were collected from CS-exposed mice and treated with 5 μM biotin-labeled OA-NO 2 or Veh for 6 h. Total protein extracts were prepared, Cat S was immunoprecipitated and Western blotting was performed with anti-biotin and -Cat S antibodies. Data are representative of two to three independent experiments with cells obtained from n = 14-18 mice/group (cells from 3 mice pooled together). *P < 0.05, **P < 0.01, ***P < 0.001, n.s. = nonsignificant.  formation was seen when Cat S with Cys25 to Ser mutated was used, as the oxygen in serine is not sufficiently nucleophilic to bond with the NFA's electrophilic carbon (Fig 3B).
We confirmed that NFA binds to Cat S protein by incubating recombinant human Cat S in vitro with biotin-labeled OA-NO 2 followed by Western blotting (Fig 3C). Incubation with either OA-NO 2 or LNO 2 dose-dependently inhibited human Cat S activity, and at the highest dose tested exhibited inhibition similar to that seen with the classic Cat S inhibitor E-64 ( Fig  3D). This inhibition was blocked by the thiol exchange inhibitor bacitracin, thus confirming Salkylation as the mechanism of inhibition.
We next sought to specifically identify the site of NFA-Cat S binding and the type of bond involved. We modeled Cat S in silico (Fig 4) to 1) determine the presence or absence of disulfide bonding for each Cys residue and the S atom's partial positive charge ( Fig 4A); and 2) calculate solvent accessibility (Fig 4B), H-bonding potential (Fig 4C), and pKa ( Fig 4D) for each of its cysteine residues. The results suggested that Cys25 is the most nucleophilic and thus the most likely site of S-alkylation. This conclusion was supported by in silico modeling of OA-NO 2 bonded to Cys25, indicating appropriate solvent accessibility and the presence of stabilizing H-bonds (Fig 4B-4D).
To test whether Cys25 was the specific Cat S residue to which NFAs bond covalently, we incubated biotin-labeled OA-NO 2 with WT Cat S or Cat S with Cys residues mutated to Ser [either individually at Cys12 (C12S), Cys25 (C25S), or Cys110 (C110S), or at all three positions, (C12;25;110S)]. OA-NO 2 binding was then determined by Western blotting (Fig 5). OA-NO 2 was bound to Cat S by S-alkylation only in Cat S WT, C12S, and C110S mutants; the interaction was absent when Cys25 was mutated to Ser (C12;25;110S and C25S). These results indicate that the S-alkylation of Cat S by OA-NO 2 occurs specifically at Cys25.
To further confirm the covalent interaction we incubated OA-NO 2 (or vehicle) with a synthetic Cat S peptide (Cat S23-29) containing either the native Cys25 (sequence GACWAFS) or a Cys25Ser substitution (GASWAFS), and analyzed the products by mass spectrometry, which indicated that NFA was S-alkylated to the Cys25-containing fragment (Fig 6B and table). As predicted, Cys25Ser mutation abolished NFA S-alkylation (Fig 6D and table) and no addition was seen in the absence of NFA (Fig 6A and 6C and table). Our initial prediction, based on the theoretical calculations shown in Fig 4, that Cys25 would be the alkylated residue was thus confirmed.

Discussion
We find that NFA treatment of activated AMs from CS-exposed mice downregulates expression and activity of Cat S, a protease heavily involved in the septal destruction that leads to emphysema. Because NFAs are known PPARγ agonists [23], we tested whether PPARγ mediates their suppressive effects on Cat S, finding that NFAs downregulate both Cat S gene expression and its enzymatic activity. These actions are mechanistically distinct, however, as PPARγ deletion eliminated NFA-induced downregulation of Cat S expression but not NFA-induced inhibition of Cat S enzymatic activity. Further studies demonstrated that NFAs inhibit Cat S activity via electrophilic S-alkylation. As our data show, this reversible addition of an NFA to the Cat S protein is not random, but is directed toward a specific catalytic cysteine (Cys25) that exhibits appropriate nucleophilicity, solvent accessibility, and potential of surrounding amino acids for formation of stabilizing H-bonds. Thus, we found that NFAs act through distinct PPARγ-dependent and -independent mechanisms that nonetheless act in a dual, functionally concerted manner to reduce Cat S levels, leading to the observed decrease in AMs' extracellular and intracellular elastolytic activity. Such a reduction in vivo would be expected to alleviate COPD severity and progression toward emphysema.
Bonnaci and colleagues similarly observed PPARγ-dependent decreases in expression of MMPs, another class of COPD-related proteases, in conjunction with PPARγ-independent effects on enzyme activity [24]. Mechanisms of the apparent MMP inhibition were not investigated, however. Reversible electrophiles, including NFAs, were found earlier to S-alkylate other regulatory proteins, including the p65 subunit of NF-κB [10], protein kinase CB [25] and Keap-1 [15,26], but these actions have not previously been linked to mechanisms of disease pathogenesis. Indeed, despite NFAs' recognized ability to electrophilically alkylate targeted protein cysteines via Michael addition, the physiological and pathogenetic relevance of these actions has remained largely unknown prior to our current investigations.
The biological roles of endogenous NFAs are likewise uncertain, but available evidence suggests that they may be important inflammatory modulators. They are produced by nonenzymatic reaction of unsaturated fatty acids with NO-derived reactive nitrogen species [7], production of which is upregulated during inflammation. Inflammation also upregulates cholesteryl hydrolase activity [27], which drives release of free, active NFAs from their esters. Even in the absence of inflammation, the combined concentrations of free and esterified fractions of NFAs in blood are within the range required to activate PPARγ [8]. Endogenous NFAs may also downregulate inflammation by S-alkylating NF-κB, and may alleviate inflammation-produced oxidative stress by S-alkylating Keap-1 [26]. Definitive proof of NFAs' endogenous role has remained elusive, however, because traditional techniques are not applicable: Endogenous NFA production cannot be effectively blocked and their multiple mechanisms of action, as illustrated by our present results, render it difficult to block their effects. Indeed, it will be challenging to block S-alkylation experimentally in intact cells.
Beyond Cat S regulation, we found that NFAs downregulate activation of AMs obtained from smoke-exposed mice, as assessed by a variety of inflammatory markers including NF-κB activity, cytokine/chemokine production, and transmigration. AMs are pivotal contributors to inflammation, particularly in COPD [3]. Through their production of inflammatory cytokines and chemokines [28], activated AMs serve to activate and recruit other cells, complementing their own direct pro-inflammatory actions. Our finding that treating AMs with OA-NO 2 reduces ability of their conditioned medium to recruit additional macrophages reveals a potentially powerful pathway via which NFAs can act to alleviate inflammation. As key sources of the principal elastolytic proteases, MMPs and cathepsins [5], AMs also contribute crucially to emphysema-causing septal destruction. Thus, our finding that NFAs downregulate total Cat S activity via both PPARγ-and electrophile-dependent mechanisms, while others found NFAs downregulate MMP expression [24], points to a potentially important protective role of endogenous or therapeutically-aimed NFAs in COPD (Fig 7).
In summary, our findings demonstrate that NFAs act as both electrophiles and PPARγ agonists to suppress the pathophysiology of CS-induced COPD. In particular, NFAs downregulate expression of the septum-destroying protease Cat S while specifically S-alkylating the enzyme's  NFAs Inhibit Emphysema-related Cathepsin via S-Alkylation catalytic Cys25, thus inhibiting its elastolytic activity and ability to promote emphysema. These results provide mechanistic insights into the part NFAs may play in modulating COPD severity, highlighting the novel role of their electrophilic activity, and support their potential therapeutic role in this disease.
Supporting Information S1 Fig. OA-NO 2 treatment has no toxic effects on mouse AMs. AMs were isolated and cultured as described and were treated with OA-NO 2 (0.1, 0.5, 1, 5 and 10 μM) for 6 h. After treatment AM viability, cytotoxicity and apoptosis were assayed as indicated in Materials and Methods. Data are representative of two independent experiments with n = 3/group. (TIFF) S1 File. Materials and Methods. Viability, Cytotoxicity and Apoptosis. Mouse AMs were isolated and treated as indicated with OA-NO 2 . Cell viability, cytotoxicity and apoptosis were measured according to manufacturer's instructions using ApoTox-Glo Triplex Assay (Promega, Madison, WI). Briefly, at the end of the OA-NO 2 treatment period 20 μl of viability/ cytotoxicity reagent containing both GF-AFC substrate and bis-AAF-R110 substrate was added to all wells, briefly mixed by shaking, and incubated for 30 min at room temperature. Viability and cytotoxicity were assessed by fluorescence measured using a plate reader (VIC-TOR X, PerkinElmer; Waltham, MA). After measurement, 100 μl of Caspase-Glo 3/7 reagent was added to each well, briefly mixed by shaking, and incubated for 30 min at room temperature. Apoptosis was assessed by luminescence measured using a plate reader. (DOCX)