Role of Carbon Monoxide in Oxidative Stress-Induced Senescence in Human Bronchial Epithelium

Prolonged or excessive stimulation from inhaled toxins may cause oxidative stress and DNA damage that can lead to stress-induced senescence in epithelial cells, which can contribute to several airway diseases. Mounting evidence has shown carbon monoxide (CO) confers cytoprotective effects. We investigated the effects of CO on oxidative stress-induced senescence in human airway epithelium and elucidated the underlying molecular mechanisms. Here, CO pretreatment reduced H2O2-mediated increases in total reactive oxygen species (ROS) production and mitochondrial superoxide in a human bronchial epithelial cell line (BEAS-2B). H2O2 treatment triggered a premature senescence-like phenotype with enlarged and flattened cell morphology accompanied by increased SA-β-gal activity, cell cycle arrest in G0/G1, reduced cell viability, and increased transcription of senescence-associated secretory phenotype (SASP) genes. Additionally, exposure to H2O2 increased protein levels of cellular senescence markers (p53 and p21), reduced Sirtuin 3 (SIRT3) and manganese superoxide dismutase (MnSOD) levels, and increased p53 K382 acetylation. These H2O2-mediated effects were attenuated by pretreatment with a CO-containing solution. SIRT3 silencing induced mitochondrial superoxide production and triggered a senescence-like phenotype, whereas overexpression decreased mitochondrial superoxide production and alleviated the senescence-like phenotype. Air-liquid interface (ALI) culture of primary human bronchial cells, which becomes a fully differentiated pseudostratified mucociliary epithelium, was used as a model. We found that apical and basolateral exposure to H2O2 induced a vacuolated structure that impaired the integrity of ALI cultures, increased goblet cell numbers, decreased SCGB1A1+ club cell numbers, increased p21 protein levels, and increased SASP gene transcription, consistent with our observations in BEAS-2B cells. These effects were attenuated in the apical presence of a CO-containing solution. In summary, we revealed that CO has a pivotal role in epithelial senescence by regulating ROS production via the SIRT3/MnSOD/p53/p21 pathway. This may have important implications in the prevention and treatment of age-associated respiratory pathologies.


Introduction
Cellular senescence is defined as an irreversible arrest of cell cycle progression, which occurs either after successive rounds of cellular replication (replicative senescence) or may be triggered by oncogenes and stress (oncogene-induced senescence and stress-induced premature senescence) [1,2]. The airway epithelium acts as the first line of defense against inhaled toxins (cigarette smoke, harmful microbes, pollutants, and allergens) by recognizing and responding to "danger signals" to repair defects that arise as a result of injury [3]. Meanwhile, prolonged or excessive stimulus from inhaled toxins may cause oxidative stress and induce DNA damage that can lead to stress-induced senescence in epithelial cells. Senescence in airway epithelial cells is notable because these cells are widely recognized for their role in the onset and progression of age-related lung disease, such as chronic obstructive pulmonary disease (COPD) and lung fibrosis [4,5].
Mitochondria are the main organelle involved in regulating cellular oxidative metabolism. Mitochondrial-derived reactive oxygen species (ROS) are generated in the form of superoxide by complex I and III in the electron transport chain (ETC) as byproducts of inefficient electron transport. In turn, ROS accumulation results in mitochondrial dysfunction and a further decrease in electron transport efficiency, which forms a positive feedback loop of ROS production and oxidative stress that leads to cell dysfunction, including senescence [6]. Sirtuin (SIRT) proteins are a group of epigenetic mediators, which are highly conserved and widely expressed in mammals [7]. There are seven members of SIRT family (SIRT1-7), which participate in various biological processes as nicotinamide adenine dinucleotide-(NAD+-) dependent deacetylases. Among them, SIRT3 is a major mitochondrial deacetylase that plays a pivotal role in response to oxidative stress [8]. The anti-oxidative effects of SIRT3 were mediated via activation of a mitochondrial detoxification enzyme, manganese superoxide dismutase (MnSOD) [9]. In vivo and in vitro studies have demonstrated SIRT3 is an ROS scavenger and regulator of cellular redox balance. SIRT3 knockout mice were susceptible to aberrant levels of cellular ROS and impaired cellular respiration [10]. Likewise, SIRT3 overexpression mice exhibited decreased ROS levels and mitochondrial oxidative stress [11]. Furthermore, Kong et al. reported SIRT3 is an integrator of energy metabolism and ROS generation at the cellular level [12]. In the lung, SIRT3 inhibits airway epithelial mitochondrial oxidative stress in a cigarette smoke-induced COPD model [13]. Also, SIRT3 mediated fibroblasts-derived extracellular vesicle-induced epithelial cell senescence in pulmonary fibrosis [14]. It was recently reported that SIRT3 could ameliorate lung senescence and improve alveolar epithelial cell function [15].
Inhaled carbon monoxide (CO) is lethal at high concentrations due to CO binding to the heme iron centers of hemoglobin, resulting in asphyxia [16]. Due to its role as a silent asphyxiant, CO was previously described as toxic. Endogenous CO is primarily generated from heme catabolism by two isoforms of heme oxygenase (HO): the constitutively expressed HO-2 and the inducible HO-1 [17]. The traditional notion that CO is solely a toxic chemical has been challenged over the last decade. Not only has this notion been challenged, the protective effects and therapeutic potential of CO have been revealed [18,19]. Low concentrations of CO are now considered a cyto-and tissueprotector in respiratory and other systems [18,20]. A series of studies demonstrated that exogenous CO has cytoprotective effects, including anti-oxidation, anti-inflammation, anti-apoptosis, and anti-proliferation [17,20]. Also, our previous studies revealed the role of CO on cytokine and chloride secretion [21], intracellular calcium regulation [22], and anti-inflammatory activity [23] in bronchial epithelial cells. However, it is unclear whether CO exerts any effect on oxidative stress-induced airway epithelial senescence.
Thus, we investigated whether CO could protect human airway epithelium from oxidative stress-induced senescence and the role of SIRT3.

Materials and Methods
2.1. Cell and Air-Liquid Interface (ALI) Cultures. BEAS-2B cells (a human bronchial epithelial cell line) were generously provided by Prof. Wong. These cells were transformed by adenovirus 12-SV40 virus hybrid (Ad12SV40) and maintain the characteristics of primary human bronchial epithelial cells, which were widely used in the epithelium mechanism studies [24][25][26][27]. BEAS-2B cells were cultured in Minimum Essential Medium (MEM) with 5% fetal bovine serum supplemented with 1% L-glutamine and 1% penicillin/streptomycin, in an incubator with humidified atmosphere containing 5% CO 2 (v/v) at 37°C.
For ALI cultures, HBECs were purchased from ScienCell Research Laboratories (#3210, Carlsbad, CA, USA) and grown in submerged culture in PneumaCult-Ex Plus Expansion Medium (#05040, StemCell, Tukwila WA). Subcultured HBECs were seeded on Poly-L-Lysine-coated 24well Transwell permeable inserts. Expansion medium was applied to the apical and basolateral chambers before the cells reached confluence. After monolayer formation, cells were air-exposed by removing the apical medium, and the basolateral medium was replaced with PneumaCult-ALI Maintenance Medium (#05001, StemCell, Tukwila WA), which was set as the beginning (Day 0; D0) of ALI phase. The basolateral ALI medium was changed every two days, while a pseudostratified layer was obtained, and immunofluorescence staining was performed to observe different cell type markers on ALI-D28.

Preparation of CO-Containing Solution and
Detection of Intracellular CO and ROS Levels. Different concentrations of CO-containing solution were prepared as previously described [21]. Briefly, Krebs-Henseleit (KH) solution was saturated with CO gas (99.99%) for 30 min to generate a 100% CO-containing solution. Different concentrations were prepared by diluting the 100% CO-containing stock solution with an unsaturated KH solution. Different cellpermeable fluorescent indicators were used to individually or simultaneously detect intracellular CO and ROS levels in living BEAS-2B cells. A Nile red CO fluorescent probe was generously provided by Prof. Klán and applied to determine the intracellular CO level [28]. BEAS-2B cells were stained with 5 μM CO fluorescent probe and 2 μg/ml Hoechst for 45 min in the presence of different CO concentrations. Intracellular CO level and total ROS production were simultaneously measured by incubating BEAS-2B cells with 20 μM H 2 DCFDA (#C400, Invitrogen, Carlsbad, CA, USA) and fluorescent 5 μM CO probe with 2 μg/ml Hoechst in presence of KH solution or CO-containing solution for 45    Oxidative Medicine and Cellular Longevity methods to detect intracellular CO level, including infrared, electrochemical, colorimetric, and fluorescent molecular chemosensors [29][30][31][32]. In the current study, we applied a fluorescent molecular chemosensor, Nile red CO fluorescent probe [28] to determine the intracellular CO level.
There was a dose-dependent increase in the fluorescence level in the presence of different concentrations (0, 10, 25, 50, 75, 100%) of CO-containing solution compared to the control solution (Figures 1(a) and 1(b)), suggesting this CO probe can semi-quantify the intracellular CO level in living cells. To determine the optimal concentration of hydrogen peroxide (H 2 O 2 ) and CO for airway epithelial cell treatment, we performed a CCK-8 assay to detect cell viability using different concentrations of H 2 O 2 or CO. As shown in Figure S1a, a relative decrease in cell viability emerged at 100 μM H 2 O 2 , though more cell death was observed at the concentrations of 150 and 200 μM. Furthermore, there was no significant difference in relative cell viability with different concentrations of CO-containing solution (range from 25% to 100%), as shown in Figure S1b.  [34] and MitoTracker [35], to measure mitochondrial-derived superoxide (O 2 •-) and to localize mitochondria, respectively. As shown in Figures 1(f) and 1(g), MitoTracker fluorescence intensity increased after H 2 O 2 exposure but decreased with CO pretreatment. This increase in mitochondria was consistent with previous findings and may be a response oxidative stress [36]. Likewise, the H 2 O 2 -mediated increase in mitochondrial-derived O 2

CO Attenuated H 2 O 2 -Induced Cellular Senescence.
Given the established role of oxidative stress in regulating the onset and maintenance of cellular senescence, we hypothesized CO may protect against H 2 O 2 -induced epithelial cell senescence. Compared to the control group, exposure of BEAS-2B cells to H 2 O 2 triggered an enlarged but flattened cell morphology accompanied with an increased number of SA-β-Gal-positive cells (Figures 2(a) and 2(b)). Furthermore, cell cycle analysis revealed H 2 O 2 treatment increased the number of cells in G1 phase (p < 0:05) but decreased the number of cells in S phase (p < 0:05) (Figure 2(c)). In addition, H 2 O 2 exposure reduced cell viability (Figure 2(d)). Critically, 1 h pretreatment with 50% CO-containing solution reversed all of these effects. The senescence-associated secretory phenotype (SASP) includes     (e) mRNA expression of SASP genes was detected by RT-qPCR with pretreated with control, 50% CO-containing solution followed by H 2 O 2 exposure for 12 h. The data were analyzed by one-way ANOVA (b and d) or two-way ANOVA (c and e) followed by Tukey's multiple comparisons test and expressed as the mean ± SEM (n =3). * p < 0:05; * * p < 0:01; * * * p < 0:001; * * * * p < 0:0001.

6
Oxidative Medicine and Cellular Longevity senescence effects of SIRT3 in airway epithelial cells. Previous reports demonstrated SIRT3 mediates the protective effect of hydrogen by inhibiting ROS-induced increased expression of p53 and p21, which are cell cycle regulatory proteins and potent markers of cellular senescence [37]. Therefore, we sought to investigate whether SIRT3 mediates the effects of CO. As shown in Figures 3(a) and 3(b), exposure to 50% CO-containing solution prior to H 2 O 2 treatment in BEAS-2B cells prevented increased expression of p53 and p21, which was consistent with the senescent phenotypes demonstrated in Figure 2. Interestingly, CO pretreatment also reversed the decreased expression of SIRT3 and MnSOD in cells treated with H 2 O 2 . It was recently reported that SIRT3 deacetylates p53, which reduces p53 occupancy of mitochondrial DNA thereby decreasing oxidative stress [38]. We also found CO could reduce p53 acetylation (ace-p53) at K382. Next, we performed gain and loss of function studies to further elucidate the role of SIRT3 in H 2 O 2 -induced oxidative stress in epithelial cells. SIRT3 silencing by small interfering RNA (siRNA) transfection reduced MnSOD expression and increased p21, p53, and ace-p53 expression (Figures 3(c) and 3(d)), which were augmented by H 2 O 2 treatment. Conversely, p21, p53, and ace-p53 (Figures 3(e) and 3(f)) expression levels were decreased in BEAS-2B cells transfected with a SIRT3 overexpression plasmid. Unexpectedly, MnSOD expression decreased following SIRT3 overexpression (Figures 3(e) and 3(f)).   •production and cell viability, BEAS-2B cells with SIRT3 knocked down and overexpressed were co-stained with MitoTracker and MitoSox or incubated with CCK-8 reagent. As expected, SIRT3 knockdown triggered an increase in mitochondrialderived O 2 •- (Figures 4(a)-4(c)) and reduced cell viability (Figure 4(d)). Conversely, mitochondrial-derived O 2 •- (Figures 4(a)-4(c)) was decreased, which was accompanied by increased cell viability (Figure 4(d)) in SIRT3 overexpressing BEAS-2B cells. Differentiated ALI cultures were apically pretreated with a control or 50% CO-containing solution for 30 min, followed by apical and basolateral exposure to 200 μM H 2 O 2 for 12 h. The treatment was repeated every other day for a total of four times. The samples were collected at day 8 after the first treatment. Results from PAS staining showed that H 2 O 2 exposure induced a vacuolated structure and impaired ALI culture integrity. However, the ALI cultures treated with CO prior to H 2 O 2 just showed increase in mucus secreting goblet cell and a relative intact structure was maintained (Figure 6(a)). To detect the effects of H 2 O 2 and CO on different cell types, cell-specific marker was detected after the treatment. In consistent with the results  Figure 4: SIRT3 regulated mitochondrial superoxide production and airway epithelial cell viability. BEAS-2B cells were transfected with a SIRT3 siRNA or overexpression plasmid for 48 h (a-c), then stained with MitoTracker (green) and MitoSOX (red) to measure mitochondrial superoxide production and (d) CCK-8 assay to measure cell viability. The data were analyzed by Student's t-test and expressed as the mean ± SEM (n =4). * * p < 0:01; * * * p < 0:001.

8
Oxidative Medicine and Cellular Longevity from the PAS staining, a vacuolated structure was indicated by the discrete fluorescence signal of Hoechst in the ALI cultures exposed to H 2 O 2 (Figures 6(b) and 6(c)). Additionally, cell type-specific marker of goblet cells, MUC5AC was increased after H 2 O 2 treatment (Figure 6(b)). However, the number of SCGB1A1+ club cells, which are responsible for repairing damage in airway epithelium, was decreased after H 2 O 2 treatment (Figure 6(c)). Similarly, these effects were partially reversed in the presence of the CO-containing solution.
To evaluate the effect of CO on H 2 O 2 -induced senescence, measurement of SASP and p21 immunofluorescence staining were performed. Immunofluorescence showed that H 2 O 2 treatment increased SASP gene transcription (Figure 7(a)) and p21 levels at both apical and basolateral sides (Figure 7(b)). These effects were also attenuated by CO pretreatment, which was consistent with the observations in BEAS-2B cells. These data suggested that CO might also attenuate oxidative stress-induced senescence in ALI-cultured human airway epithelium.

Discussion
Accumulation of senescent cells is presumed to contribute to age-related diseases pathogenesis by at least two distinct mechanisms. First, senescent cells are increasingly considered malignant "zombie cells" because they are resistant to cell death but are unable to proliferate to repair tissue defects. Second, senescent cells act as pro-inflammatory mediators with the capacity of inducing chronic inflammation and 9 Oxidative Medicine and Cellular Longevity subsequent tissue dysfunction [39][40][41]. An accumulation of senescent airway epithelial cells has been found in the lungs of chronic lung disease patients, including COPD and idiopathic pulmonary fibrosis (IPF) [42,43]. Therefore, an understanding of the molecular mechanism can be used to intervene in the process of airway epithelial senescence, which may provide new strategies for the prevention and treatment of age-related respiratory diseases. To our knowledge, the current study is the first to provide evidence suggesting exogenous application of CO protects the airway epithelium from oxidative stress-induced intracellular ROS production and senescence in bronchial epithelial cells and ALI-cultured epithelium models. The data presented here demonstrated CO exerts a protective effect by inhibiting   In the cell, senescence can be triggered when cells are exposed to oxidative stress, and H 2 O 2 is the most commonly used inducer of stress-induced premature senescence (SIPS) [44]. As a result of H 2 O 2 exposure, DNA damage and double-strand breaks can occur, resulting in overactivity of ADP-ribose polymerase, as well as depletion of NAD+ and ATP. Further, excessive ROS production by H 2 O 2 exposure inhibits antioxidant activity, further increasing the production of endogenous free radicals. As a result of these changes, SIPS develops morphological, biomolecular, and genetic alterations [45]. H 2 O 2 exposure affects premature senescence differently depending on the cell type, exposure time, dose, and solvent used [46,47]. We tested different concentrations of H 2 O 2 to determine the most effective concentration to establish a premature senescence model since there is no fixed protocol for H 2 O 2 exposure. When H 2 O 2 -induced premature senescence of fibroblasts was studied, the proportion of SA-β-Gal-positive cells increased significantly after treatment with 25 and 50 μM H 2 O 2 . The percentage decreased, however, in fibroblasts treated with 100 μM or higher H 2 O 2 , presumably due to a sharp reduction in cell numbers [46]. We observed the same results in BEAS-2B cells. In general, H 2 O 2 -induced premature senescence is a chronic process that involves low concentrations and continuous processes, whereas high concentrations and transient stimulation cause apoptosis or necrosis instead of senescence.
Though ROS production can lead to senescence, ROS production is also a consequence of senescence. Excessive oxidative stress can cause DNA damage, leading to cellular senescence. On the other hand, the accumulation of dysfunctional mitochondria and other organelles within senescent cells might in turn promote oxidative stress and enhance intracellular ROS production [41,48]. In the present study, we simultaneously measured intracellular CO and ROS levels and verified that CO negatively regulated H 2 O 2 -induced ROS production, which might contribute to the resulting airway epithelial cell senescence. The mitochondria and the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX) family are the major sources of intracellular ROS [49]. Evidence suggests that the anti-oxidative properties of CO may be due to its ability to bind to heme-containing proteins, NOX family members, and mitochondrial complexes of ETC, which are responsible for ROS production after oxidative stress [50,51]. As reported previously, CORM-2 inhibits NADPH oxidase cytochrome b558 and respiratory chain activity in complex I of airway smooth muscle cells to inhibit proliferation [50]. Nitric oxide is produced by endothelial nitric oxide synthase in vascular endothelium, which directly or indirectly regulates mitochondrial ROS production [52]. Additionally, the metal-binding nature of CO suggests that cyclooxygenase, cytochrome p450, and cytochrome c oxidase could also serve as CO sensors [51]. Therefore, we also demonstrated that CO inhibited H 2 O 2 -induced mitochondrial-derived superoxide production, which is consistent with results from other studies using different cell types [33,53]. SIRT3 plays a critical role in mitochondrial metabolism and is mainly localized in the mitochondrial matrix [54]. Mitochondrial dysfunction is closely associated with several respects of senescence, including reduced metabolic enzyme activity and increased oxidative damage [55]. SIRT3 is a demonstrated major risk factor of senescence and aging as decreased SIRT3 expression is frequently observed in many age-related diseases [56]. It has been reported that SIRT3 deficiency promoted pulmonary fibrosis [57,58]. More recent studies have demonstrated SIRT3 can prevent airway epithelial mitochondrial oxidative stress in cigarette smokeinduced COPD, possibly through the regulation of MnSOD expression levels and activity [13]. In the current study, we first demonstrate that SIRT3/MnSOD mediated the antisenescent effects of CO in H 2 O 2 -induced oxidative stress in airway epithelium cells. A previous study discovered SIRT3 regulated mitochondrial gene expression, mitochondrial oxygen consumption, and ROS production in Alzheimer's disease by reducing p53 acetylation levels. Consistent with this report, we found SIRT3 attenuated H 2 O 2 -induced increased ROS production and p53 acetylation levels, as well as expression of its downstream protein, p21. Though SIRT3 overexpression unexpectedly decreased MnSOD levels, it still exerted anti-senescent and anti-oxidative effects. This result is likely due to the positive feedback inhibition and effects of MnSOD, which depends on its acetylation level and activity in addition to its expression level [59,60].
The ALI culture technique provides a mucociliary model to study lung biology that recapitulates normal airway epithelium biology in vitro [61]. Optimal and stable differentiation of ALI cultures was achieved from commercial human primary bronchial epithelial cells (HBECs) and culture medium in the current study. The expression of cell type-specific markers and senescence-associated genes was evaluated after treatments. MUC5AC upregulation is an important characteristic of goblet cell metaplasia and mucus hypersecretion, which is always observed in airway inflammation diseases, such as asthma and COPD [62][63][64][65]. SCGB1A1, also known as CC10, is a member of the secret globin family, which is mainly produced by club cells [66]. SCGB1A1 is especially abundant in the peripheral airways and patients with asthma, COPD, and obliterative bronchiolitis after lung transplantation have deficient SCGB1A1 expression in their airways [67][68][69][70]. These observations sug-gest club cells and SCGB1A1 play an important role in the control of the airway integrity and repair response upon injury. Indeed, several studies have demonstrated SCGB1A1 may be responsible for exerting club cell effects, such as immunomodulation, inhibiting mucus secretion, and modulating epithelial remodeling. Moreover, there are in vivo studies demonstrating that SCGB1A1+ club cells possessed selfrenewal capacity and could give rise to new ciliated cells after lung injury, where they contribute to tracheal repair [71,72].
The results of our current study showed H 2 O 2 treatment increased the number of MUC5AC+ goblet cells but decreased the number of SCGB1A1+ club cells. Ineffective club cell restoration, which is likely due to oxidative stressinduced senescence in club cells, reduced SCGB1A1 expression. In turn, this contributed to mucus secretion, increased p21 protein levels, and increased SASP gene transcription. These effects were attenuated in the presence of the COcontaining solution.

Conclusions
Overall, the present study verified that CO protects the airway epithelium from H 2 O 2 -induced senescence. From the data presented here, CO exerts a protective mechanism at the cellular levels by inhibiting intracellular ROS formation by activating the SIRT3/MnSOD/p53/p21 pathway ( Figure 8). However, CO might play different roles in ALI cultures depending on cell type. These findings improve our understanding of the underlying molecular mechanisms of CO in oxidative stress-induced senescence, which may contribute to age-related lung disease pathogenesis. However, further in vivo validation and detailed  Figure 8: Schematic summary of the role and underlying mechanisms of CO in oxidative stress-induced senescence in human airway epithelium. The oxidative stress might lead to cellular senescence by causing DNA damage. In turn, the senescent cells could also augment oxidative stress by SASP and form a vicious cycle. The present study verified that H 2 O 2induced airway epithelium senescence, and the effects could be partially reversed by CO. This protective mechanism, at least in part, by inhibiting the mitochondrial ROS production and mediated by the mitochondrial deacetylase SIRT3.