Cells

The BEAS-2B cell line (SV40-immortalized human airway bronchial epithelial cell line) were purchased from the American Culture of Tissue Collection, and were grown in complete growth medium (RPMI 1640 supplemented with heat-inactivated 10% FBS and 1% L-glutamine) at 37 °C / 5% CO₂. Before use, cells were starved for 24 hours in minimum medium (RPMI 1640 supplemented with 1% FBS and 1% L-glutamine). In the BSO pretreatment model, cells were pre-incubated with BSO for 16–18 hours. Various inhibitors or activators were added 30 min prior to the cigarette smoke extract (CSE) exposure. Primary cells were obtained as follows: human airway epithelial cells of bronchial origin (hAEC) and air-liquid interface (ALI) cultured human bronchial epithelial cells (HBEC) were from Epithelix-Sarl (Geneva, Swiss); human pulmonary artery endothelial cells (hPAEC) were from Lonza (#CC-2530, Slough, UK). For the analysis of clinical samples, human primary bronchial epithelial cells were extracted from lung tissue from patients with chronic obstructive pulmonary disease (COPD) undergoing lung resection surgery at the Royal Brompton Hospital. The characteristics of patients with chronic obstructive pulmonary disease (COPD) were as follows (mean ± SD); 3 male and 1 female, 68.75 ± 9.8 years old, pack year (number of cigarettes smoked per day/20 x duration of smoking): 82.0 ± 58.7, FEV₁ (forced expiratory volume in one second): 1.56 ± 3.1L (62.0 ± 18.2% predicted normal), FEV₁/FVC (forced vital capacity): 49.0 ± 14.0%. All subjects gave informed written consent and the study was approved by the NRES London-Chelsea Research Ethics committee (study number 09/H0801/85). The cells were cultured as monolayers in LHC-9 media (Invitrogen, Paisley, UK) on collagen (1% w/v) coated plates.

Preparation of CSE

One full-strength Marlboro cigarette with filter removed (Phillip Morris, London, UK) was bubbled into 10 ml of minimum medium, at a rate of one cigarette per 1.5 minutes. CSE was then passed through a 0.2μm filter to sterilize and remove particulate matter and was used immediately. The optical density was measured at 320λ wavelength, and the solution was diluted to be OD = 0.85 (this is original stock as 100%). The stock CSE was thereafter diluted with culture media to appropriate percentages of CSE solution.

Protein extraction and Western blotting (WB)

Cytoplasmic fraction was prepared using modified radio-immuno-precipitation assay (RIPA) buffer (50 mM Tris HCl pH 7.4, 0.5% NP-40, 0.5% w / v Na-deoxycholate, and 150 mM NaCl) with freshly added complete protease inhibitor cocktail. This lysis buffer was confirmed to extract cytoplasmic rich fraction based on β-actin and Lamin A/C staining in Western blotting. This method has been used for the experiments shown in Figs ​Figs11 and ​and2,2, S1 and S2 Figs. In following experiments, cytoplasmic fraction and nuclear fraction were prepared using Nuclear Extract Kit (#40010, Active Motif, La Hulpe, Belgium) according to the manufacture’s instruction.

Open in a separate window

Fig 1

Localization of SIRT1 in human bronchial epithelial cells.

(A) SIRT1, SIRT6 and HDAC2 protein expressions are shown in cytoplasmic and nuclei fraction. Lamin A/C and α-tubulin was used for loading control of nuclear and cytoplasmic fractions, respectively. (B) SIRT1 localization in BEAS-2B cells was detected by immunocytochemistry. Upper panel showed SIRT1 immunoreactivity, and lower panel indicated negative control (× 40 in BD Pathway 435 bioimager). (C) SIRT1 subcellular localization was examined for the different human primary cells, such as human airway epithelial cells of bronchial origin (hAEC, passage 2 and 4), air-liquid interface cultured human bronchial epithelial cells (ALI) and human pulmonary artery endothelial cells (hPAEC).

Open in a separate window

Fig 2

PI3Kα is involved in the CSE-induced SIRT1 reduction in the presence of L-buthionine-sulfoximine (BSO) in cytoplasmic fraction.

(A) After pretreatment with 100 μM BSO for 16 hours, BEAS-2B cells were exposed to the 3% CSE for up to 24 hours. SIRT1 protein and phosphorylated Akt/Akt ratio(p-Akt / Total-Akt) were assayed at different time point by SDS-PAGE / WB in cytoplasmic fraction. (B) SIRT1 protein levels were examined at different concentration of CSE with or without 100 μM BSO pretreatment. All values were mean value ± SEM of at least three experiments. ** p < 0.01, compared with the values of non-treatment group; ^†† p < 0.01 between the absence or presence of BSO. (C) Various inhibitors against PI3K signaling pathway molecules were added 30 min prior to the CSE exposure (0.1 μM of PIK75 for PI3Kα inhibitor, 10 μM of GSK2636771 for PI3Kβ inhibitor, 10 μM of AS605240 for PI3Kγ inhibitor, 5 μM of IC87114 for PI3Kδ inhibitor, and 0.02 μM of rapamycin for mTOR inhibitor). The Akt activation status (p-Akt / Total-Akt) and SIRT1 protein levels were evaluated by SDS-PAGE / WB. (D, E) The concentration dependent effects of PIK75 on the SIRT1 and Akt phosphorylation levels were examined using CSE exposure with BSO pretreatment model. The ratio of p-AkT/Total-Akt or SIRT1 levels were fitted with the sigmoid-curve, and calculated for the IC₅₀ of PIK75 (E, respectively). All values were mean value ± SEM of at least three experiments. ** p < 0.01, compared with the values of non-treatment group; ^†† p < 0.01 between the two groups.

The cell extracts prepared above were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose membrane, and then incubated overnight with anti-SIRT1 antibody (1:1,000 dilutions, #sc-15404), anti-SIRT6 antibody (1:3,000 dilutions), anti-HDAC2 antibody (1:200,000 dilutions), anti-p-Akt antibody (1:500 dilution), anti-Total-Akt antibody (1:1,000 dilutions), anti-Nrf2 antibody (1:1000, dilutions) or anti-CRM1 antibody (1:200 dilutions). To standardize the expression of each protein, the membranes were re-probed with anti-β-actin antibody (1:200,000 dilutions) or anti-α-tubulin antibody (1:2,000 dilutions), anti-Lamin A/C antibody (1:500 dilutions). The membranes were then incubated with the appropriate peroxidase-conjugated secondary antibodies (1:3000 dilutions each). The bound antibodies were visualized by chemiluminescence (ECL plus; GE healthcare, Buckinghamshire, UK).

MTT assay

After the removal of supernatant, cells were washed with phosphate buffered saline (PBS) once, and incubated with 1 mg / ml thiazolyl blue tetrazolium bromide solution for 1 hour at 37 °C / 5% CO₂. Formazan crystals were dissolved in DMSO, and optical density was measured at 550 λ wavelength. Cell viabilities were represented as the relative ratio to that of non-treatment group.

Protein half-life assay

After 16 hours pretreatment with or without 100 μM BSO, BEAS-2B cells were exposed with or without 3% CSE from 1 to 24 hours in the presence of 30 min pre-incubation with cycloheximide (CHX) (100 ng / ml). At different time points, cells were harvested and SIRT1 protein levels were assayed by SDS-PAGE / WB.

Reverse transcriptase quantitative PCR (RT-qPCR)

Total cellular RNA was extracted using RNeasy mini kit (Qiagen, Valencia, CA, USA) and cDNA was prepared by using Multiscribe reverse transcriptase (Applied Biosystems, Warrington, UK). The RT-qPCR analysis of SIRT1 and GNB2L as a house keeping gene, were performed using Taqman primers and probe set from Applied Biosystems in a Corbett Rotor-Gene 3000 (Corbett Research Sortlake, NSW, Australia).

Analysis of nascent protein synthesis

Nascent protein synthesis was measured by modifying the protocol for Click-IT^® L-homopropargylglycine (HPG) Alexa Fluor Protein Synthesis Assay kits (#C10429, ThermoFisher scientific, Rockford, IL, USA). In brief, cells were at first cultured with methionine free RPMI medium (#14517–01, ThermoFisher scientific) for 30 min. Then medium was replaced with CSE- or vehicle-containing methionine free RPMI supplemented with 50 μM Click-IT^® HPG, which was incorporate into the newly synthesized proteins instead of methionine. After 6 hours, cells were harvested and lysed for WCEs preparation. The WCEs products were immuno-precipitated by rabbit-derived anti-SIRT1 antibody (#sc-15404, Santa Cruz Biotechnology) with the Protein A magnetic beads (#10001D, Life Technologies), and additionally labelled with Biotin-Azide (#B10184, ThermoFisher scientific) in the presence of cupper, according to the manufacture’s instruction. After separation with SDS-PAGE / WB, biotin labelled SIRT1 (biotin-SIRT1) was then detected by the HRP-conjugated streptavidin (1:1,000 dilutions, ThermoFisher scientific). To standardize the biotin-SIRT1, the membranes were re-probed with mouse-derived anti-SIRT1 antibody (1:8,000 dilutions, #SAB1404972, Sigma-Aldrich), and then incubated with the goat-derived anti-mouse peroxidase-conjugated secondary antibodies (1:3000 dilutions). Newly synthesized SIRT1 was expressed as relative intensity calculated by the ratio of the mean band intensity for biotinylated SIRT1 to the mean band intensity for the total SIRT1 protein.

SIRT1 motif change by oxidative stress

The various antibodies against different part of SIRT1 were used for examining the protein motif changes caused by oxidative stress. We used anti-SIRT1 antibodies from; Santa-cruz technology (#sc-15404 for ① indicated in Panel F in S2 Fig. (F)), Sigma-Aldlich (#5322 for ②, #5447 for ③, #SAB1404972 for ④); Cell signaling technology (#9475 for ⑤); and OriGene (Rockville, MD, USA) (#TA303784 for ⑥).

Immunofluorostaining for SIRT1 in BEAS-2B cells

Immunostaining of SIRT1 was done using Histostain-Plus^® IHC kit (#85–9043, Invitrogen, Paisely, UK). Briefly, BEAS-2B cells were cultured on the chamber slide, and then fixed in 3% formaldehyde fixative solution for 15 min, and peameabilized by 0.1% triton X-100 in 10 mM PBS for 30 min. After blocking with serum blocking solution (#85–9043, Invitrogen), the cells were incubated with FITC-conjugated anti-SIRT1 polyclonal antibody (1:100 dilution, #orb103489, biorbyte, Cambridge, UK) overnight at 4 °C. The immunoreactions were then embedded in the DAPI (4’,6-diamidino-2-phenylindole) containing mounting solution (ProLong^® Diamond Antifade Mountant with DAPI, ThermoFisher scientific, Rockford, IL, USA), as manufacture’s instruction. The slides were analyzed by BD Pathway 435 bioimager system (with 40X Olympus Plan Fluorite objective, 0.75 numerical aperture, Lumenera-Infinity 3–1 12 bit CCD camera) at room temperature using BD Atto Vision softeware (BD, New Jersey, US).

Malondialdehyde (MDA) assay

MDA was quantified using OxiSelect^™ TBARTS assay kit (#STA-330, Cambridge Bioscience, Cambridge, UK). Briefly, the cytoplasm samples were thawed on ice and centrifuged for 1 or 2 min at 4°C, 13000rpm, before use. Twenty μl of diluted samples (0.5μg/mL protein) or standards were mixed and incubated with equal volume of SDS lysis solution for 5min at room temperature (RT) with shaking in PCR plates. TBA reagent was then added and mixed with the samples and after 5–15 min at RT was set up for 1 hour at 100°C in a PCR cycler. The plates were then incubated for 15 min at 4°C in a refrigerator. Fifty μl of this solution was then transferred to a half- area 96-well black plate and MDA was detected by fluorescence with 540λ excitation and 590λ emission. The amount of MDA in the samples was calculated using the standard curve.

Cigarette smoke extracts decreased cytoplasmic SIRT1

As shown in Panel A and B in S1 Fig, CSE exposure caused a reduction in SIRT1 protein expression in cytoplasmic fractions in a concentration and time-dependent manner. SIRT1 protein expression was significantly decreased at 20% CSE although the cell viability was significantly reduced at higher than 10% CSE (61 ± 12% reduction).

Therefore, cells were treated with L-buthionine sulfoximine (BSO: an irreversible inhibitor of γ-glutamylcystein synthetase [25]), which depletes intracellular glutathione storage, making cells more susceptible to the exogenous oxidative stress [25] without increasing cell toxicity. In fact, it has been reported that available glutathione storage is decreased in the chronic smokers [26,27]. As shown in Fig 2A and 2B, in the presence of BSO, cytoplasmic SIRT1 was decreased by lower concentrations of CSE (3%) after 24 hours exposure. No significant reduction of cell viability was seen.

At same time, the 3% CSE exposure with the pretreatment of 100 μM BSO (3% CSE/BSO) significantly increased the phosphorylation of Akt (Fig 2A), suggesting the phosphatidylinositol 3-kinase (PI3K) signaling pathway was activated in this condition. To identify the isoform of PI3K involved in SIRT1 reduction, we tested various inhibitors of PI3K pathway (0.1 μM of PIK75 for PI3Kα inhibitor, 10 μM of GSK2636771 for PI3Kβ inhibitor, 10 μM of AS605240 for PI3Kγ inhibitor, 5 μM of IC87114 for PI3Kδ inhibitor, and 0.02 μM of rapamycin for mTOR inhibitor). As shown in Fig 2C, only PIK75 among all inhibitors tested suppressed the CSE-induced Akt phosphorylation effectively (Fig 2C, upper panel), and significantly reversed the reduction of SIRT1 protein under oxidative stress (Fig 2C, lower panel). PIK75 concentration-dependently reversed SIRT1 suppression (Fig 2D and 2E) with an IC₅₀ value of 14.3 nM and suppressed p-Akt/Total-Akt ratio (Fig 2E and 2F) with an IC₅₀ value of 9.1 nM, respectively; these values were very close to the previously reported IC₅₀ of PIK75 for PI3Kα enzyme activity (IC₅₀: 5.8 nM) [28]. Therefore, these results suggested that PI3K, especially α subunit, plays an important role in regulating SIRT1 suppression under oxidative stress. Furthermore, we also examined the effect of NU7026 (DNA dependent protein kinase (DNA-PK) inhibitor, 3.3 μM) as PIK75 is also known to inhibit DNA-PK [28], and also tested other signaling inhibitors such as 1,1-dimethyl-biguanide hydrochloride (5’ AMP-activated protein kinase (AMPK) activator, 5 mM), BIRB796 (p38 mitogen activated protein kinase (MAPK) inhibitor, 1 μM) and U1026 (MAPK / extracellular-signal-regulated kinases (ERK) kinase (MEK) 1 / 2 inhibitor, 1 μM), however none of which had any effects on SIRT1 suppression.

To clarify the mechanism of reduction of cytoplasmic SIRT1, we investigated role of proteasome degradation using proteasome inhibitors, Z-Leu-Leu-Leu-al (MG-132: 8 μM) or N-Acethyl-Leu-Leu-Norleu-al (ALLN: 10 μM) (Panel A in S2 Fig), SIRT1 protein half-life in the presence of cycloheximide (Panel B in S2 Fig) and the translational regulation of SIRT1 protein using Click-IT^® assay (Panel C in S2 Fig), but none of systems were responsible for the reduction of cytoplasmic SIRT1. We also investigated the level of SIRT1 gene transcription determined by SIRT1 mRNA levels by reverse transcriptase quantitative polymerase chain reaction (RT-qPCR). Although the 24 hours exposure with CSE decreased the SIRT1 mRNA only in the presence of BSO (Panel D in S2 Fig), the reduction was not reversed by PIK75 (Panel E in S2 Fig). Additionally, we had concerns on the modification of the antibody recognition motif of SIRT1 by oxidative stress, resulting in failure of SIRT1 detection by the antibody used rather than actual reduction of SIRT1 protein. Therefore, we tested various antibodies that recognize epitopes within different regions of the SIRT1 protein, however, all antibodies confirmed the reduction of SIRT1 protein levels by CSE exposure with BSO pretreatment (Panel F in S2 Fig), therefore antibody detection motif on SIRT1 was not modified by oxidative stress. Finally, to investigate the possibility of secretion of SIRT1 as the mechanism of cellular SIRT1 reduction, we also detected SIRT1 in the cell culture supernatant, but no significant increase in SIRT1 protein after CSE exposure was observed (Panel G in S2 Fig).

Cigarette smoke extracts induced shuttling of SIRT1 from cytoplasmic to nuclei

As shown above, none of the mechanisms such as transcriptional, translational, or post-translational regulations, seemed to be associated with SIRT1 reduction in cytoplasmic fraction under oxidative stress. Therefore, we next investigated the localization of SIRT1 under oxidative stress. Interestingly, cytoplasmic SIRT1 dramatically translocated to nucleus after 24 hour exposure of CSE in the presence of BSO in BEAS-2B cells (Fig 3A and 3B, S3 Fig, Panel A in S4 Fig). This dynamic shuttling between cytoplasm and nucleus seemed to be SIRT1 specific phenomenon, because both SIRT6 and HDAC2 stayed within the nucleus irrespective of the CSE stimulation with BSO pretreatment; in case of SIRT6 or HDAC2, only nuclear protein fractions had a tendency of reduction under oxidative stress without changes in its location. Immunocytochemistry also showed nuclear localization after CSE treatment (Fig 3C) compared to the baseline in Fig 1B. Most importantly, the shuttling of SIRT1 to nuclei by CSE was also confirmed in primary cells, such as hAEC or ALI-cultured HBEC (Fig 3D and 3E or 3F and 3G, respectively). The shuttling of SIRT1 seemed to be caused by the active transport because its dynamics was in parallel to that of chromosome maintenance region 1 (CRM1) protein (Fig 3A and 3B), which is known to recognize nuclear export sequence and function as a cargo for SIRT1 nuclear export [22]. Furthermore, the nuclear translocation of SIRT1 under oxidative stress in BEAS-2B cells was almost completely reversed by PIK75, the PI3Kα specific inhibitor (Fig 3H and 3I). Thus, the SIRT1 located mainly in cytoplasm in bronchial epithelial cells at normal condition, and translocated from cytoplasm to nucleus under oxidative stress, which was mediated mainly through PI3Kα signaling. The SIRT1 shuttling was not induced by nicotine, an Akt activator [29] as shown in Fig 3J; in addition, Akt inhibitor did not inhibit 3% CSE-induced SIRT1 nuclear translocation, either (Fig 3K). This suggests that the translocation of SIRT1 is regulated downstream of PI3Kα before Akt comes into play. The SIRT1 shuttling was not induced by anisomycin, a JNK activator at 6 hours or 24 hours after treatment (Fig 3J). SP-600125, a JNK inhibitor, did not inhibit CSE-induced SIRT1 shuttling, either (right panel in Fig 3L). Although Nasrin and colleagues reported that JNK is involved in SIRT1 shuttling to the nuclei in hydrogen peroxide stimulated human embryonic kidney (293T) cells [30], it was not the case in our current study with cigarette smoke conditioned media. At least, N-acetyl cysteine clearly inhibited CSE-induced SIRT1 shuttling, suggesting oxidative stress to drive the shuttling process (left panel in Fig 3L).

Open in a separate window

Fig 3

CSE-induced SIRT1 reduction is actually due to shuttling of cytoplasmic SIRT1 to nuclei.

BEAS-2B cells were pretreated with or without 100 μM BSO in the presence or absence of 3% CSE. After 24 hours, localization of SIRT1, SIRT6, HDAC2 and CRM1 in cytoplasm (A) and nuclei (B) were examined by SDS-PAGE / WB. (C) SIRT1 localization after CSE exposure for 24 hours in the presence of BSO in BEAS-2B cells were detected by immunocytochemistry (× 40 in BD Pathway 435 bioimager). (D-G) In the 3% CSE exposure with 100 μM BSO pretreatment model, the SIRT1 shuttling was examined in the human primary bronchial epithelial cells (D, E) or air-liquid interface cultured human bronchial epithelial cells (F, G). (H, I) After CSE exposure with BSO pretreatment in BEAS2B, the effect of PIK75 (0.1 μM) on the SIRT1 shuttling was examined. (J) SIRT1 in nuclear fraction after nicotine (10 μM) or anisomycin (10 μg / ml) treatment for 6 hours in BEAS-2B. (K) Effects of Akt inhibitor (10 μM) on SIRT1 protein in nuclei from cells stimulated with 3% CSE for 24 hours. (L) Effects of N-acetyl cysteine (NAC, 10 mM) and SP-600125 (SP, 10 μM) on SIRT1 protein in nuclei from cells stimulated with 3% CSE for 24 hours. All values were mean value ± SEM of at least three experiments. * p < 0.05, ** p < 0.01, compared with the values of non-treatment group; ^†† p < 0.01 between the two groups.

The shuttling of cytoplasmic SIRT1 to the nucleus seemed to be cell-protective, because nuclear SIRT1 is strongly associated with the mRNA induction of some anti-oxidative proteins; not only with the well-known SIRT1 downstream target superoxide dismutase 2 (SOD2) [31], but also with the hemoxygenase-1 (HO-1) (Fig 4A), SOD3 (Fig 4B) and NADPH quinone oxidoreductase-1 (NQO-1). Such beneficial effect (mRNA induction of anti-oxidative proteins) was also confirmed to be inhibited by the PI3Kα inhibitor, PIK75 (Fig 4A). As well as in mRNA levels, HO-1 and SOD2 proteins were also induced by CSE and inhibited by PIK-75 (Fig 4B). Furthermore, although FOXO3a is known to be an important transcription factor to induce these anti-oxidant gene expression and also the function is regulated by SIRT1 [32], we found 3% CSE also increased FOXO3a in nuclei together with SIRT1 (Fig 4C). Taken together, SIRT1 translocation from cytoplasm to nucleus is a normal protective mechanism against “single” oxidative stress, by induction of potent anti-oxidant enzymes.

Open in a separate window

Fig 4

CSE-induced SIRT1 nuclear translocation was associated with increases in anti-oxidant gene and protein expression.

(A) The messenger RNA levels of superoxide dismutase 2 (SOD2), SOD3, NADPH quinone oxidoreductase-1 (NQO-1) and hemoxygenase-1 (HO-1), all of which were normalized to mRNA expression of a GNB2L1 housekeeping gene, were assessed by the reverse transcription-quantitative polymerase chain reaction (RT-qPCR) 24 hours after 3% CSE exposure with BSO. All values were mean values ± SEM of at least three experiments. * p < 0.05, ** p < 0.01, compared with the values of non-treatment group; ^†† p < 0.01, compared between the two groups. (B) The protein levels of HO-1, NQO-1 and SOD2 were assessed by SDS-PAGE/WB. (C) SIRT1 and FOXO3a protein in cytoplasmic fraction (top) and nuclear fraction (bottom) 24 hours after 3% CSE exposure with BSO. PIK75 (0.1 μM) was also treated before CSE exposure. The band density of SIRT1 and FOXO3a were also calculated and corrected to that of Lamin A/C. * p < 0.05, ** p < 0.01, compared with the values of BSO only group; ^† p < 0.05, compared between with or without PIK75.

This study was supported by the Wellcome Trust Programme Grant (093080/Z/10/Z). S.Y is a recipient of a Banyu Life Science Foundation International fellowship. CV and JB are recipients of a Wellcome Trust grant shown above. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Pulmocide Ltd provided support in the form of salaries for authors [KI, TC], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.