Small-interfering RNA for c-Jun attenuates cell death by preventing JNK-dependent PARP1 cleavage and DNA fragmentation in nitrogen mustard-injured immortalized human bronchial epithelial cells

Abstract

Sulfur mustard (a type of vesicant) can directly damage lung bronchial epithelium via aerosol inhalation, and prevalent cell death is an early event that obstructs the respiratory tract. JNK/c-Jun is a stress response pathway, but its role in cell death of the injured cells is not clear. Here, we report that JNK/c-Jun was activated in immortalized human bronchial epithelial (HBE) cells exposed to a lethal dose (20 μM) of nitrogen mustard (NM, a sulfur mustard analog). c-Jun silencing using small-interfering RNA (siRNA) rendered the cells resistant to NM-mediated cell death by blocking poly(ADP-ribose) polymerase 1 (PARP1) cleavage and DNA fragmentation. In addition, the transduction of upstream extrinsic (Fasl-Fas-caspase-8) and intrinsic (loss of Bcl-2 and mitochondrial membrane potential, ΔΨm) apoptosis pathways, as well as phosphorylated (p)-H2AX (Ser139), an epigenetic marker contributing to DNA fragmentation and PARP1 activity, was partially suppressed. To mimic the detachment of cells by NM, HBE cells were trypsinized and seeded on culture plates that were pre-coated with poly-HEMA to prevent cell adhesion. The JNK/c-Jun pathway was found to be activated in the detached cells. In conclusion, our results indicate that JNK/c-Jun pathway activation is necessary for NM-caused HBE cell death and further suggest that c-Jun silencing may be a potential approach to protect HBE cells from vesicant damage.

Introduction

Sulfur mustard (SM), one of the classical vesicating mustard agents, has been deployed in WWI, the Iran–Iraq war, and even in the recent Syria civil war [1]. Nitrogen mustard (NM) is an analog of SM with similar toxicities, which is highly toxic to exposed organs or tissues [2]. The lung is one of the primary target organs for vesicants via aerosol inhalation. Vesicant-caused lung injures, including airway ulcer or inflammation, mucosal epithelial degeneration or necrosis, and tracheobronchial obstruction, are great threats to individual and public health [3–6].

Once inhaled, mustard gas is easily adsorbed on the wet upper respiratory tract surface of the exposed victims. Thus, the mucosal layer of the bronchus is more seriously injured than lung parenchyma. Vesicant-caused epithelial cell death is prevalent in the injured lung, and poly(ADP-ribose) polymerase 1 (PARP1) cleavage and nuclear breakdown have been considered the major causes. Cleaved (c) PARP1 is a DNA damage sensor that facilitates DNA repair by recruiting DNA repair machinery to damage sites. On the other hand, PARP1 (c) can disturb cellular metabolism by consuming NAD+ and ATP, which results in energy deficiency in the injured cells [7, 8]. Nuclear breakdown may block the expression of many genes necessary for cell growth or survival, and the amount of DNA fragments (180–2000-bp long) is commonly proportional to the degree of cell death [9–12].

Phosphorylated (p)-H2AX (S139), a DNA damage sensor, is increased in SM- or SM analog-injured tissues or cells [13–15]. p-H2AX (S139) is well known to be necessary for DNA fragmentation [16–18]. Recent studies have reported that PARP1 (c) also promoted chromatin DNA fragmentation by modulating the speed of fork elongation and recognizing the unligated Okazaki fragments [7]. Interestingly, p-H2AX (S139) can stabilize PARP1 association and promote catalysis to facilitate DNA repair [18]. Thus, in vesicant-injured lung cells, the interaction between p-H2AX (S139) and PARP1 (c) during DNA damage repair may be necessary for DNA fragmentation.

JNK/c-Jun, a stress response pathway, can be activated by SM or its analogs and induces inflammatory and proteolytic mediators [14, 19, 20]. JNK/c-Jun had been reported to either protect or damage cells [21–23]. Its specific role may be related to expression duration and level, as long-lasting-JNK/c-Jun and or its over-expression can be harmful. We hypothesized that JNK/c-Jun was important for vesicant-caused cell death in lung cells. The mechanism may involve, first, the JNK/c-Jun-activated cytochrome c-mediated apoptosis pathway [24]; second, JNK/c-Jun-activated Fas-Fasl-caspase-8 [25]; third, increased JNK/c-Jun-induced p-H2AX (S139) levels [17]; and fourth, JNK/c-Jun activation in detached cells, which diminishes survival signals to induce anoikis, a special form of programmed cell death [26–29].

Here, by employing an immortalized human bronchial epithelial (HBE) cell line and NM, an in vitro model similar to other reports was established, and the JNK/c-Jun pathway was found to be activated after a fatal dose of NM exposure [5, 12, 30, 31]. A specific siRNA for c-Jun alleviated cell death and suppressed molecular signal activation, including p-H2AX (S139) induction, PARP1 cleavage, and DNA fragmentation. Therefore, the c-Jun siRNA reported here may offer a potential therapeutic approach for preventing vesicant-induced lung mucosal epithelium damage.

Materials and Methods

Chemicals, reagents, and antibodies

The following reagents were commercially purchased: Mechlorethamine hydrochloride (used as the NM in this study; Shanghai Dipper Chemical, purity >98%); JC-1, 4′,6-diamidino-2-phenylindole (DAPI), YO-PRO-1, propidium iodide (PI), and anisomycin (ANI; activator of JNK) (Beyotime, Shanghai, China); and inhibitors of ERK1/2 (U0126; MCE, Monmouth Junction, NJ, USA), PARP1 (veliparib; MCE), and JNK1/2 (SP600125; Santa Cruz, Dallas, TX, USA). The primary antibodies were raised against c-Jun, PARP1 (c), p-ERK1/2 (Thr202/Tyr204), Fasl, caspase-8 (total and cleaved), caspase-7 (c), caspase-3 (c) (Cell Signaling Technology, Danvers, MA, USA); Bad, Bax, Bcl-2, p-JNK1/2 (Thr183/Thr221) (Beyotime); Fas, p-H2AX (Ser139), p-c-Jun (Ser73, S73) (Sangon, Shanghai, China); and GAPDH and β-actin (Santa Cruz). The corresponding secondary antibodies were purchased from Beyotime. All other reagents were of analytical grade.

Cell culture and treatments

The HBE cell line HBE135-E6E7 (HBE, ATCC® CRL-2741™) was obtained from ATCC (Gaithersburg, MD, USA). Cells were cultured in RPMI 1640 medium (Hyclone, Logan, UT, USA) containing 10% fetal bovine serum (Kang Yuan Biology, Tianjin, China) and were cultured in a cell incubator at 37°C under 5% CO₂.

For NM exposure, cells were trypsinized, seeded on a culture plate, and incubated for 24–72 h to reach 80–90% confluence. In addition to single NM exposure, signal pathway inhibitors SP600125, veliparib, and U0126 were pre-incubated with the cells for 1 h before NM exposure. After removing the culture medium, cells were exposed to 3-, 7-, or 20-μM NM solution (mechlorethamine hydrochloride) freshly prepared in complete culture medium. After 2 h, the toxic solution was replaced with fresh culture medium containing each inhibitor without NM for further culture. At the indicated time points, both detached and adherent cells in each group were harvested for further analysis.

To study the signaling changes of non-adherent cells, the homogeneous anion polymer poly-HEMA (Sigma, St. Louis, MO, USA) was employed to pre-coat the culture plates to prevent cell adhesion. Briefly, 10 mg/ml poly-HEMA in water-free ethanol was shaken for 2 h at 42°C to dissolve and was centrifuged at 4000 g for 2–4 min to remove undissolved particles. Then, 600 μl of the poly-HEMA suspension were applied per well in a 12-well plate and were incubated overnight at 42°C until they were dry. The pre-coated plates were transferred to the refrigerator (+4°C) to cool for 1 h. After ultraviolet disinfection, the plates were ready for cell seeding.

Cell viability assay

For cell viability analysis, the CellTiter 96® Aqueous One Solution Cell Proliferation Assay (Promega, Madison, WI, USA) was employed. HBE cells were cultured in a 96-well culture plate. At the indicated time points after treatment, the culture medium was replaced with RPMI medium containing 10% CellTiter 96® Aqueous One Solution Reagent. After 15–20 min of incubation, the absorbance of the suspension was measured in a microplate reader (SpectraMax i3X, Molecular Devices, Sunnyvale, CA, USA) at 490 nM. Meanwhile, a standard curve was established in each test by measuring the absorbance in wells seeded with known amounts of cells (5 × 10⁴, 3 × 10⁴, 1.5 × 10⁴, 0.75 × 10⁴), which was used to calculate the relative cell viability.

ATP level detection

A commercial kit for detecting ATP levels was purchased (Beyotime). Cells were lysed using lysis buffer and were centrifuged at 12 000 g for 10 min. A volume of 100 μl supernatant mixed with 100 μl of ATP working solution and was added to an opaque white 96-well plate (Corning, NY, USA) and analyzed in the SpectraMax microplate reader by measuring luminescence values. Meanwhile, protein was quantified by BCA (Boster, Wuhan, China). The concentration of ATP in each sample was calculated as nmol/mg protein.

Apoptotic/necrotic cell staining

Both detached and adherent cells in each group were harvested. After centrifugation, 2 × 10⁵ cells were re-suspended in 200 μl RPMI medium containing YO-PRO-1 (2 μg/ml) and PI (2 μg/ml) for 5 min of incubation. Then, cells were seeded on a culture plate for further examination. The excitation/emission wavelength of YO-PRO-1 was 491/509 nm, while wavelength of PI was 535/617 nm. The individual cell staining of PI or YO-PRO-1 indicated necrotic cells (red) or early apoptotic cells (green), respectively. Dual staining of YO-PRO-1 and PI indicated late apoptotic cells (yellow).

DNA fragmentation detection

A DNA Ladder Extraction Kit (Beyotime) was used to detect DNA fragmentation. Briefly, cells were harvested and successively mixed with RNase, protease K, and lysis buffer. Importantly, to avoid unexpected damage, the cells were quickly frozen in liquid nitrogen for 10–15 s after trypsinization. After 10 min incubation at 70°C, ethanol was added to the mixtures to precipitate DNA. The solution was then transferred to a DNA binding column provided in the kit and centrifuged under 8000 g for 1 min. DNA on the binding column was purified with washing buffer and was finally collected with DNA elution buffer after centrifugation at 12 000 g for 1 min.

After quantifying the DNA amounts with a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), ~500 ng DNA from each group was loaded onto a 1% agarose gel containing 2 μg/ml GelRed (Biotium, Fremont, CA, USA). The DNA was separated in TAE or TBE buffer at 90 V for 30 min and was visualized under an ultraviolet transilluminator (Vilber Lourmat, Collegien, France). The DNA fragments varied with an increment of ~185 nucleotides.

Immunofluorescence staining for c-Jun protein

Cells were fixed with 4% formaldehyde for 10 min and then incubated in 1% BSA/10% normal goat serum/0.3 M glycine in 0.1% PBS-Tween for 1 h to permeabilize the cells and block non-specific protein–protein interactions. The cells were then incubated with the primary antibody for c-Jun overnight at +4°C. The secondary antibody (green) was Alexa Fluor® 488 goat anti-Rabbit IgG (H + L) used at a 1:2000 dilution for 1 h. DAPI was used to stain the nuclei of the cells (blue) at a concentration of 2 μg/ml. Under fluorescence microscopy (Olympus, Tokyo, Japan), the fluorescent signals were recorded at the emission wavelength of 529 nm.

Western blotting

Total protein from cells was prepared using RIPA lysates (Boster) containing a phosphatase and protease inhibitor cocktail (Roche, Basel, Switzerland). About 30 μg of denatured proteins from each sample were loaded into each well and were separated by 8 or 10% SDS-PAGE and transferred to a PVDF membrane (Millipore, Bedford, MA, USA). After 1 h blocking with 5% BSA containing 0.05% Tween 20, the membrane was incubated overnight with primary antibody diluted to 1:1000 or 1:2000. The next day, a further 1 h incubation with secondary antibody diluted to 1:1000 was carried out. After a complete wash with Tris-buffered saline, the membrane was mixed with chemiluminescence solution (BeyoECL Star, Beyotime) to display fluorescent signals, which were recorded by the Vilber Lourmat transilluminator. Western blot results were quantified by densitometric analysis of immunoblots and were normalized to β-actin or GAPDH as the internal control.

siRNA transfection

For gene silencing, ~1.25 × 10⁵ HBE cells were seeded in each well of a 12-well culture plate to reach a density of 25–30% for transfection on the next day. siRNA for c-Jun and the negative control were transfected into cells using the siRNA-Mate transfection kit (GenePharma, Shanghai, China) according to the manufacturer’s protocol. After 1 or 2 days of further culture, cells at 80–90% confluency were exposed to NM. The sequences of siRNA were as follows: c-Jun, 5′-GAACGTGACAGATGAGCAG-3′; negative control, 5′-AAGAACCAGCAGAGGUCACAA-3′.

Flow cytometry

JC-1 was used to detect the mitochondrial membrane potential (ΔΨm). JC-1 aggregates in the matrix of mitochondria to form J-aggregates in normal cells but dissociate into monomers in cells with low ΔΨm. Under 488-nm excitation, the emission of JC-1 aggregates, and monomers are 585/42 nm (red) and 530/30 nm (green), respectively. Cells with low ΔΨm exhibit a reduced ratio of red to green fluorescence.

Briefly, ~2.5 × 10⁵ cells were collected. After 5 min centrifuge at 1000 g, cells were re-suspended with 200 μl JC-1 solution (1 μmol/L in RPMI medium) for 10 min of incubation at room temperature. After a wash with PBS, a total of 1 × 10⁴ cells were subsequently analyzed on an FC500 cytometer (Beckman Coulter, Indianapolis, IN, USA) using 488-nm excitation with 530/30 (FL1) and 585/42 nm (FL3) bandpass emission filters. Specific gates were set on the scatter diagram to distinguish normal cells. Results were finally calculated using CXP analysis software (Beckman Coulter).

Statistical analysis

The statistical analysis was conducted with t-tests and one-way analysis of variance using SPSS 13.0 statistical software (IBM Inc, Armonk, NY, USA). A P-value <0.05 was considered to be statistically significant. The Tukey–Kramer post hoc test was applied to significant between-group comparisons. All data represent at least triplicate independent experiments.

Results

PARP1 inhibitor suppressed NM-induced HBE cell death

First, different concentrations of NM (3-, 7-, and 20-μM mechlorethamine hydrochloride) were employed to test its toxicity to HBE cells. After 20-μM NM exposure, the average cell viability was significantly reduced to 91% at 10 h (P < 0.01) and to 66% at 24 h (P < 0.01; Fig. 1A). Western blotting revealed that caspase-3 (c), caspase-7 (c), p-H2AX (S139), PARP1, c-Jun, and p-c-Jun reached maximum levels 24 h after 20-μM NM exposure (Fig. 1B). In addition, p-H2AX (S139), c-Jun, and p-c-Jun appeared to be more sensitive to NM as their levels were increased earlier (4 h) and at a lower dose than other proteins.

Figure 1.

the effect of PARP1 inhibitor on NM-caused cell death; (A) cell viability relative to the non-treated group at 4 h is shown; bars indicate the mean ± SD (n = 4); ^aP < 0.01; (B) analysis of the protein level of caspase-3 (c), caspase-7 (c), PARP1 (c), p-H2AX (S139), c-Jun, and p-c-Jun by western blot; (C and D) analysis of the effect of PARP1 inhibitor veliparib on cell viability and ATP level; ATP levels were measured in cell lysates and normalized to total protein; bars indicate the mean ± SD (n = 3); ^aP < 0.01 and ^bP < 0.05 versus the non-treated group; ^cP < 0.01 and ^dP < 0.05 versus the combined veliparib- and NM-treated group; (E) typical fluorescent images of cells stained with 2 μg/ml YO-PRO-1 and PI; cells emitting red or green (yellow) signals were counted as necrotic or apoptotic cells, respectively; bar: 50 μm; the results were calculated from six random fields (n = 6); ^aP < 0.01 versus single NM-treated cells (comparing apoptotic cells); ^cP < 0.01 versus single NM-treated cells (comparing necrotic cells).

Application of the PARP1 inhibitor veliparib (50 μM) blocked the NM-caused cell viability reduction. Meanwhile, the ATP level was not changed in NM- or veliparib-treated cells (Fig. 1C and D). To further identify the apoptotic and necrotic cells, YO-PRO-1 and PI were employed to stain the cells. In normal cells, fluorescence signals were rarely detected. In NM-injured cells, the fluorescence signals increased, and the number of cells emitting yellow or green (apoptosis) was greater than the cells emitting red (necrosis). Both apoptotic and necrotic cells were significantly suppressed after application of PARP1 inhibitor veliparib (Fig. 1E).

PARP1 inhibitor suppressed p-H2AX (S139) induction and DNA fragmentation in NM-injured cells

It was previously found that ATP level did not decline in NM-injured cells even when PARP1 was cleaved, probably due to sodium pyruvate in the RPMI medium, which could support ATP production according to a previous study [32]. DNA fragmentation is another important cell death pathway, and p-H2AX (S139) can sense DNA damage and promote DNA fragmentation [17, 18]; thus, the effects of PARP1 inhibitor veliparib on p-H2AX (S139) expression and DNA fragmentation were measured. At 24 h, the induction of p-H2AX (S139) by NM was suppressed by veliparib (Fig. 2A). Consistently, veliparib also significantly suppressed the formation of 360-bp DNA fragments (Fig. 2B and C). These results indicated that the PARP1 inhibitor protected HBE cells from NM damage at least partially by directly or indirectly (p-H2AX-dependent) suppressing DNA fragmentation.

Figure 2.

the effect of PARP1 inhibitor on p-H2AX (S139) expression and DNA fragmentation; (A) western blot analysis of p-H2AX (S139) protein change; (B) detection of DNA fragmentation in control and NM-treated cells with and without veliparib treatment at 24 h; (C) quantitation of 360-bp DNA fragments; results are expressed as means ± SD (n = 3); ^aP < 0.01 versus single NM-treated cells.

p-JNK1/2, not p-ERK, contributed to c-Jun activation in NM-injured HBE cells

Signaling cascades responsible for c-Jun activation are involved in the regulation of transcriptional (promoter-dependent), translational (eEF2-dependent), and post-transcriptional (p-JNK1/2-dependent) expression of c-Jun [33]. To explore how c-Jun was activated, JNK and ERK, the candidate up-regulatory signals of c-Jun, were simultaneously screened. Both p-JNK1/2 and p-ERK levels were stimulated to a maximum level 10 h after 20-μM NM exposure (Fig. 3A), while 20-μM U0126 and 30-μM SP600125 could efficiently suppress the induction of p-ERK and p-JNK1/2, respectively. However, only SP600125, not U0126, potently inhibited c-Jun activation (total/p-c-Jun) 10 h after NM exposure (Fig. 3B). Immunofluorescence staining revealed that c-Jun levels were low in normal cells but increased after 20-μM NM exposure and were mainly localized to the nuclei (Fig. 3C).

Figure 3.

regulation and location of the c-Jun protein in NM-injured cells; (A) analysis of the amount of p-JNK1/2 and p-ERK1/2 in control and NM-treated cells for various durations; (B) analysis of the influence of U0126 (ERK inhibitor) and SP600125 (JNK inhibitor) on the NM-dependent increase in p-ERK, p-JNK1/2, p-c-Jun, and c-Jun; (C) immunofluorescent staining of c-Jun protein to identify subcellular location; HBE cells were treated with NM (20 μM) for 10 h; following fixation, the cells were immunostained with a c-Jun antibody (green) and stained with DAPI (blue) to visualize nuclei; the stained cells were captured under fluorescence microscopy; bar: 50 μm; white staining indicates high expression of c-Jun proteins in cellular nuclei.

c-Jun silencing attenuated cell death by suppressing JNK-dependent PARP1 cleavage, p-H2AX (S139) induction, and DNA fragmentation

To explore the function of c-Jun, a specific siRNA for c-Jun was synthesized and transfected into HBE cells. c-Jun expression in either NM- or ANI-treated cells was inhibited (Fig. 4A and F). Molecular signal studies revealed that c-Jun siRNA suppressed caspase-3/7-PARP1 cleavage, p-H2AX (S139) induction, and DNA fragmentation (Fig. 4A, B, F, and G). Furthermore, c-Jun siRNA prevented cell morphological change induced by NM or ANI. Consistently, c-Jun siRNA significantly suppressed the reduction of cell viability and the increase in apoptotic and necrotic cells in ANI (12 h)- or NM (24 h)-injured cells (Fig. 4C–E and F–H). Under higher doses of ANI or NM exposure, the dead cells were mostly necrotic but not apoptotic, and necrotic cell death was also suppressed by c-Jun siRNA (Supplemental Fig. S1).

Figure 4.

the effect of c-Jun silencing on ANI- or NM-caused cell death; (A, F) analysis of the effect of c-Jun silencing (50-nM siRNA) on ANI- or NM-evoked JNK/c-Jun, p-H2AX (S139), PARP1 (c) and other apoptotic proteins; (B, G) DNA fragmentation was detected and the 360-bp DNA fragment was quantified; results are expressed as means ± SD (n = 3); ^aP < 0.01; (C–E and H–J) the change in cell morphology, cell viability, and the percentage of dead cells; for cell viability, bars indicate the mean ± SD (n = 3); ^aP < 0.01, ^bP < 0.05 versus the non-treated group; ^cP < 0.01 versus the combined si-c-Jun-, NM-, and ANI-treated group; for dead cell calculation, fields were randomly chosen (n = 6); ^aP < 0.01 (comparing apoptotic cells) and ^cP < 0.01 (comparing necrotic cells) versus single NM-treated cells.

Noticeably, SP600125 (30 μM) exerted no protective effect for NM-injured cells (data not shown), probably because over 10 μM of SP600125 can promote cell apoptosis independent of JNK inhibition [34]. Thus, only c-Jun siRNA, but not SP600125, was employed in further toxic mechanism experiments.

c-Jun silencing suppressed the activation of intrinsic and extrinsic apoptosis pathways

The activation of Fasl-Fas-caspase-8 is known to induce the extrinsic apoptosis pathway [9]. In HBE cells exposed to NM, the extrinsic apoptosis pathway was indeed activated, as Fasl and Fas were increased at 10 and 24 h, respectively. Caspase-8 was cleaved at 10 h and to a greater extent at 24 h. Meanwhile, caspase-8 (total) was reduced correspondingly. In the c-Jun silenced group, the increases in Fas, Fasl, and caspase-8 (c) were all suppressed (Fig. 5A).

Figure 5.

the effect of c-Jun silencing on the intrinsic and extrinsic apoptosis-signaling pathways; (A) analysis of the effect of c-Jun silencing (50-nM siRNA) on intrinsic and extrinsic apoptotic protein expressions; (B) the scatter diagram of cellular ΔΨm measured by flow cytometry is shown on the left; based on the fact that cells with loss of ΔΨm have a reduced ratio of red (FL3) to green (FL1) fluorescence signals, gates were set to classify normal and abnormal (loss of ΔΨm) cells (gate H: normal, gate J: loss of ΔΨm); the percentage of cells in gate J was calculated and is shown; bars indicate the mean ± SD (n = 3) ^aP < 0.01.

Bcl-2 can form heterodimers with Bax and Bad to reduce their insults to mitochondria [9, 35]. The reduction of Bcl-2 and loss of ΔΨm have been considered to be the inducers of the intrinsic apoptosis pathway [9]. In NM-injured HBE cells, the level of Bcl-2, not Bax, was reduced 10 h after NM exposure and to a minimum level at 24 h. c-Jun silencing did not affect Bad or Bax expression but partially restored the Bcl-2 level at 24 h (Fig. 5A).

Using flow cytometry, we confirmed that the percent of cells with ΔΨm loss in gate J increased to 2.77% at 10 h and to 8.06% at 24 h after NM exposure. Compared to the control group, cells in gate J were significantly reduced in the c-Jun silenced group after NM exposure (1.76% at 10 h and 5.32% at 24 h, P < .01 vs. control group; Fig. 5B).

The JNK/c-Jun pathway was activated in non-adherent cells

Since cell adhesion can be damaged by vesicants and the detached cells may undergo anoikis, the next experiment was designed to identify whether JNK/c-Jun could be activated and induce the expression of PARP1 (c) and p-H2AX in the detached cells. To prevent cell adhesion, the trypsinized cells were seeded on culture plates that were pre-coated with poly-HEMA. The JNK/c-Jun pathway was activated at 0 h, while c-Jun and p-c-Jun reached a higher level at 2 h after activation. Thereafter the JNK/c-Jun pathway returned to a lower level at 5 and 8 h. Except for p-JNK1/2, the induction of c-Jun and p-c-Jun was potently suppressed by c-Jun silencing. Although PARP1 (c) and p-H2AX (S139) were confirmed to be regulated by JNK/c-Jun, their expressions were weakly increased with time in non-adherent cells, probably due to the JNK/c-Jun pathway was only transiently activated (Fig. 6). All in all, the results here proved that JNK/c-Jun could be activated in detached cells, which suggested that c-Jun siRNA might be also useful in preventing anoikis of vesicants-injured lung cells.

Figure 6.

the level of p-JNK1/2, c-Jun, p-c-Jun, p-H2AX (S139), and PARP1 (c) in non-adherent cells transfected with siRNA for c-Jun or control siRNA; HBE cells were pre-transfected with 50 nM c-Jun or control siRNA for 2–3 days; once near confluence, cells were then trypsinized and seeded on another plate pre-coated with poly-HEMA for further culture; at the indicated time points, cells were collected and the total proteins were prepared; the protein levels were further analyzed by western blotting.

Discussion

In this study, we employed NM to explore the role of JNK/c-Jun in HBE cell death in vitro. Because of the strict administrative control on chemicals listed by chemical weapon conventions, SM is not commonly employed in laboratory studies. As an SM analog, NM has been applied by others and us previously due to its similar chemical structure, toxic effects, and safe handling [36–38]. PARP1 can be cleaved in SM- or NM-injured tissues or cells, and inhibitors against PARP1 cleavage can protect HBE cells in vitro from NM damage [8, 12]. Consistently, the current study confirmed the participation of PARP1 (c) in NM-caused HBE cell death. Moreover, a specific siRNA for c-Jun was found to efficiently suppress NM-caused PARP1 cleavage and cell death.

H2AX is the primary histone marker essential for checkpoint-mediated cell cycle arrest and DNA repair after DNA double-strand breaks. We noticed that PARP1 could localize to H2AX-enriched chromatin damage sites to label DNA break sites with poly(ADP-ribose) and that p-H2AX (S139) facilitates DNA repair by stabilizing PARP1 association and promoting catalysis [18]. Interestingly, H2AX can be phosphorylated at the Ser139 site by p-JNK1/2 directly or indirectly, and the latter is dependent on caspase-3-induced DNA-dependent protein kinase expression [17]. The increase in p-H2AX (Ser139) did not contribute to the activation of caspase-3 but was indispensable to caspase-induced DNA fragmentation, which contributes to cell apoptosis [16, 18]. In NM-injured cells, JNK/c-Jun activation was shown to be one of the causes for p-H2AX (S139) induction, and c-Jun siRNA suppressed p-H2AX (S139) expression and DNA fragmentation.

The extrinsic apoptosis pathway Fas/Fasl/caspase-8 can be stimulated in lung cells exposed to SM [9, 39, 40]. Once Fas increases, it binds to the death receptor Fasl to cleave caspase-8, which activates and amplifies downstream apoptotic signals, that is, activating caspase-3 and caspase-7 directly or indirectly through the Bid-cytochrome-c-caspase-9 pathway [9]. Application of small-interfering RNA for the Fas receptor suppressed SM-induced apoptosis in human airway epithelial cells [41]. A recent report showed that c-Jun could regulate Fas and Fasl to mediate apoptosis in mouse intestinal epithelial cancer cells [25]. Similarly, our results confirm the conclusion that c-Jun regulates Fas and Fasl in HBE cells exposed to NM.

The intrinsic apoptosis pathway (loss of Bcl-2 level and ΔΨm) was also partially suppressed by c-Jun silencing in the injured cells. The blockage of cell detachment may be a novel mechanism for c-Jun silencing to maintain the Bcl-2 level since cell adhesion could keep the Bcl-2 level stable [29]. The vesicants of SM and NM are known to reduce cell adhesion and interaction by damaging adhesion-related proteins, such as laminin 5, keratins, and integrin α6β4a [2, 42, 43]. Since JNK belongs to the MAPK pathway which regulates proteolytic mediators [14, 44], c-Jun silencing might block JNK-dependent adhesion-related protein degradation to keep the injured cells adhered to the matrix. The maintenance of cellular ΔΨm by c-Jun silencing can be Bcl-2-dependent or independent [31, 45]; the latter might be explained as the loss of ΔΨm resulting from apoptosome formation in the apoptotic-signaling pathway [45].

We note that integrins at focal adhesions and talin suppress JNK, while the multidirectional static stretch of mechanical stress activates JNK [46]. Besides, vesicants can decrease cell adhesion to the matrix, and JNK is necessary for anoikis [28, 31]. These reports indicate cell cytoskeleton and adhesion change after vesicants exposure modulates JNK activation, and c-Jun siRNA may be useful in preventing JNK-dependent cell anoikis. In non-adherent cells (on Poly-HEMA), the JNK/c-Jun pathway was confirmed to be rapidly activated, and c-Jun siRNA was effective to suppress c-Jun and p-c-Jun expression. However, the increase in PARP1 (c) and p-H2AX levels was modest probably due to the JNK/c-Jun pathway being only transiently activated and the immortalized HBE cell being resistant to anoikis. We assume the expression of JNK/c-Jun, PARP1 (c), and p-H2AX may be different in the non-adherent primary bronchial epithelial cells, and this is more important in promoting cell anoikis.

In this study, c-Jun siRNA was proved to attenuate ANI- and NM-caused cell death. Under a higher dose of ANI or NM exposure, necrotic cells and not apoptotic cells were detected to be significantly reduced (Supplemental Fig. S1), probably due to PARP activity inhibition by c-Jun siRNA suppressed energy consumption, thus changing the cell death pathway according to previous reports [23, 32, 47]. In summary, the current findings provide experimental evidence that JNK/c-Jun pathway activation is necessary for NM-induced HBE cell death and that the specific siRNA for c-Jun reported here may be useful for preventing lung epithelial cell death after NM or SM injury, although additional experimental studies are still needed in primary lung epithelial cells or in an in vivo model.

Conflict of interest statement

The authors declared that they have no conflicts of interest to this work and the publication.

Funding

National Natural Science Foundation of China (grant numbers 81502711 and 81803279); Young Talent Cultivation Program of Military Medical Science and Technology, Army Medical University (grant number 20QNPY001).

Abbreviations

ANI, anisomycin; DAPI, 4′,6-diamidino-2-phenylindole; HBE, human bronchial epithelial cells; NM, nitrogen mustard; PARP1, poly(ADP-ribose) polymerase 1; PI, propidium iodide; SM, sulfur mustard; t/p, total/phosphorylated; ΔΨm; mitochondria membrane potential