Death-associated Protein Kinase 1 Mediates Ventilator-induced Lung Injury in Mice by Promoting Alveolar Epithelial Cell Apoptosis

Mouse alveolar epithelial cells (MLE-12; American Type Culture Collection, USA) were cultured and passaged in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (Gibco, USA) and antibiotics (100 units/ml penicillin, 100 μg/ml streptomycin). The culture was maintained in a humidified atmosphere containing 5% carbon dioxide at 37°C. The cells were harvested for experiments until 70 to 80% confluent.

Mouse alveolar epithelial cells (5 × 10⁵/well) were inoculated on collagen 1–precoated Bioflex plates. After a 24-h culture, the medium was replaced by fresh medium containing 0.5% fetal bovine serum for another 12 h.

The cells were randomly split for cyclic stretch, static stretch, or no stretch. The cells in the cyclic stretch group were exposed to cyclic stretch by using an FX-5000T Flexercell Tension Plus system (Flexcell International, USA), at 30 cycles/min and 30% amplitude. The stretching times were 6, 12, or 24 h. For static stretching, a pattern with 30% elongation for up to 24 h was used. The cells in the control group were sown on Bioflex plates treated in the same manner as the stretched cells, but they were not subjected to any stretch. After 6, 12, or 24-h culture, the cells and the culturing media were collected for further detection.

The silencing of DAPK1 expression in mouse alveolar epithelial cells was conducted via small interfering RNA. Three different DAPK1 small interfering RNAs were used (Supplemental Digital Content, table S1, https://links.lww.com/ALN/C452). The small interfering RNA and transfection reagent were purchased from RiboBio (China).

The cells (5 × 10⁵/well) were sown on Bioflex plates. The cells were divided into three groups: control, control small interfering RNA + cyclic stretch, and DAPK1 small interfering RNA + cyclic stretch. After 24 h, the transfection mixture (small interfering RNA and transfection reagent) was prepared in a final volume of 1,000 μl to achieve a final small interfering RNA concentration of 100 nmol/l. This mixture was incubated for 7 min at room temperature to allow complex formation and then added drop-by-drop onto the cells. A control small interfering RNA was used as a negative control.

After a 12-h incubation, the medium was replaced by fresh medium containing 0.5% fetal bovine serum, and the cells were stretched at 30 cycles/min for 24 h to 30% elongation. The cells in the control group were not stretched. Then the cells and culture media were collected for further procedures.

DAPK1 Δcalmodulin is an active form of the death-associated protein kinase, which lacks the calmodulin regulatory and binding domain (i.e. amino acids 266 to 312; named DAPK1Δcalmodulin). The corresponding plasmid for DAPK1Δcalmodulin was constructed as described in a previous report.¹⁷ 

Mouse alveolar epithelial cells (3 × 10⁵) were seeded in a 6-well culture plate for 24 h. A transfection mixture (i.e., plasmid, Lipofectamine 2000, and culture medium) was prepared to a final volume of 500 μl with 0.01 μg/μl DAPK1Δcalmodulin vector or control vector (empty plasmid). Lipofectamine 2000 transfection reagent (Invitrogen, USA) was used by following the manufacturer’s instructions. The transfection mixture was incubated with cells for 6 h. Then the medium was changed to Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum. After another 48-h culture, the cells were collected for further detection.

The media from the cultures of stretch cells and control cells were analyzed for lactic acid dehydrogenase activity by using a lactic acid dehydrogenase assay kit in accordance with the manufacturer’s instructions (Beyotime, China). The kit was calibrated using lactic acid dehydrogenase standards at 2 μg/ml.

Cell apoptosis was analyzed with an annexin V–fluorescein isothiocyanate apoptosis detection kit (BD Pharmingen, USA). The cells were digested by trypsin (0.05%) EDTA (0.2 g/l) for 90 s. Both attached and nonattached cells were harvested and stained with fluorescein isothiocyanate–conjugated annexin V and propidium iodide and then analyzed by flow cytometry (FACSort). Alveolar epithelial cell apoptosis was expressed as a percentage of annexin V–positive cells to total cells.

Caspase 3 activity was analyzed using a caspase 3 activity assay kit (Beyotime) in accordance with the manufacturer’s instructions. Briefly, the cell lysate supernatant was mixed with buffer containing the substrate peptides for caspase attached to p-nitroanilide. The release of p-nitroanilide was quantified by measuring the absorbance at 405 nm by an enzyme-linked immunosorbent assay reader. The caspase activities were expressed as a percentage of the control.

The Ethics Committee of Huazhong University of Science and Technology approved and supervised this study. The animal care and use committee of Tongji Medical College of Huazhong University of Science and Technology permitted all the animal experiments.

Healthy male C57BL/6 mice (specific pathogen-free grade, aged 6 to 8 week) weighing 20 to 22 g were purchased from Wuhan University (Wuhan, China). All mice were housed in specific pathogen-free conditions for 5 days before experimental use. The mice were tested in sequential order.

For adeno-associated virus infection, DAPK1 inhibitor pretreatment, and mechanical ventilation, experiments were performed in specific pathogen-free environment with fixed temperature and humidity. The other experiments were performed in our laboratory. All the animal experiments were performed in the morning (8 to 10 am).

All the mice were anesthetized with intraperitoneal injections of ketamine (80 mg/kg) and xylazine (10 mg/kg) and intubated through the trachea. The mice were ventilated (ventilator model: Harvard, 55-7062) either at 8 ml/kg with 120 breaths/min (low tidal volume, low V_T), or at 35 ml/kg with 80 breaths/min (high tidal volume, high V_T) for 4 h; the positive end-expiratory pressure was set at nil for animals subjected to ventilation, whereas the control mice underwent tracheotomy but breathed spontaneously. All animals were ventilated with room air.

For some experiments, mice received DAPK1 inhibitor ((4Z)-2-[(E)-2-phenylethenyl)-4-(3-pyridinylmethylene)-5(4H)-oxazolone, 500 μg · kg⁻¹ · day⁻¹, intraperitoneally; Merck, Germany) or an identical dose of placebo-control solution (dimethyl sulfoxide) for 6 weeks. The mice were then ventilated at 35 ml/kg at 80 breaths/min for 4 h. All animals were euthanized by exsanguination at the end of the mechanical ventilation. The primary outcomes of the animal experiments were lung injury score and alveolar epithelial cell apoptosis.

A small hairpin RNA (Supplemental Digital Content, table S1, https://links.lww.com/ALN/C452) used for gene knockdown of DAPK1 was expressed from plasmid with a U6 promoter by using an adeno-associated virus expression system (DAPK1 virus; Shanghai Genechem, China). The same adeno-associated virus containing a scramble small hairpin RNA sequence for nonmammalian gene was used as the control (referred to here as the “control-virus”).

C57BL/6 mice were divided into three groups: the control group, the control-virus group + high V_T, and the DAPK1-virus group + high V_T. The mice were anesthetized with intraperitoneal injections of ketamine (80 mg/kg) and xylazine (10 mg/kg) and intubated through the trachea with a 20-gauge needle. The mice were inoculated through the trachea with 10⁶ genomic copies of DAPK1-virus in a final volume of 30 μl. The same dose of control-virus (10⁶ genomic copies, 30 μl) was administered through the trachea to the control-virus + high V_T group. Control mice received phosphate-buffered saline only. The mice in the control-virus + high V_T and DAPK1-virus + high V_T groups received high tidal volume ventilation (35 ml/kg, 4 h) 30 days after the adeno-associated virus instillation. Control mice received no treatment. All animals were euthanized by exsanguination at the end of mechanical ventilation.

Right middle lung specimens were fixed in 4% paraformaldehyde and embedded in paraffin. The embedded lung tissues were divided into 4-μm sections for staining with hematoxylin and eosin for light microscopy. Lung injury pathologic changes were blindly examined by two histologists without knowing the experimental groups. Images were captured randomly from five nonoverlapping fields per slide. Three slides were captured for each animal, and six animals were included for each group. The lung injury score calculated by following a previously published criteria (Supplemental Digital Content, table S2, https://links.lww.com/ALN/C453),¹⁸  in which each animal was examined as the unit of analysis.

The lung wet-to-dry ratio was used as an index of pulmonary edema formation. The wet right middle lungs were weighed and recorded as the wet weight. Then they were weighed after the incubation in an oven at 60°C for 72 h and recorded as the dry weight.

Bronchoalveolar lavage was performed with phosphate-buffered saline (3×0.5 ml) and analyzed as described.¹⁹  The total number of bronchoalveolar lavage cells was determined by a hemocytometer. The protein concentration in bronchoalveolar lavage was measured by a bicinchoninic acid assay kit (Beyotime).

Bronchoalveolar lavage and lung tissues were measured for tumor necrosis factor-α, interleukin-6, and interleukin-1β by using commercially available enzyme-linked immunosorbent assay kits (Elabscience Biotechnology Company, China).

In animal models, alveolar epithelial cell apoptosis was visualized with terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate in situ nick-end labeling staining. Lungs were pressure-fixed and processed for paraffin embedding. Paraffin-embedded lung sections were deparaffinized and rehydrated in a graded ethanol series. After antigen retrieval and quenching of endogenous peroxidase, the sections were incubated with anti-cytokeratin 18 antibody (1:200 dilution; Abcam, China) at 4 °C overnight.

On the second day, after phosphate-buffered saline wash, The sections were incubated with cell apoptosis reaction mixture (Roche, Canada) for 1 h. The sections were washed and incubated with secondary antibodies at room temperature for 1 h. 4,6-Diamino-2-phenylindole was added to stain the sections for 5 min. Images of the stained lung sections were analyzed under a confocal microscope. Images were captured as described for histologic analysis. Alveolar epithelial cell apoptosis was examined with each animal as the unit of analysis. The data were presented by a percentage of double-positive (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate in situ nick-end labeling staining and anti-cytokeratin 18 staining) cells to total cells observed in the field.

At the end of the experiments, total protein extracts were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and then transferred to polyvinylidene difluoride membranes as previously reported.²⁰  The transformed membranes were probed with the following primary antibodies: DAPK1 (1:1,000; Sigma–Aldrich, USA), phosphorylated DAPK1 (1:100; Cell Signaling Technology, USA), phosphorylated myosin light chain (1:500; Cell Signaling Technology), myosin light chain (1:1,000; Abcam, USA), B-cell lymphoma 2 (1:1,000; Santa Cruz, USA), and B-cell lymphoma 2–associated X (1:1,000; Santa Cruz).

No statistical power calculation was performed before the study; sample sizes were the estimates based on experience and preliminary data. Except for the evaluation of the lung injury score, investigators were not blinded during data collection or analysis. Outliers were evaluated with no necessary action. No data was excluded or missing. The data are presented in scatter plots (n ≤ 4) or bar graphs (n ≥ 5) with means ± SD. The distribution of all data was tested for normality with Shapiro–Wilk tests. ANOVA was two-way repeated, and then the Bonferroni post hoc test was used for analyzing changes in lactic acid dehydrogenase concentration, DAPK1 levels and activation at different time points. Two independent sample t tests were performed for analysis of the changes in two-group comparisons. One-way ANOVA and the Bonferroni post hoc test were used for comparing more than two groups. All the statistical analyses were performed using Prism 5 (GraphPad Software, USA). A two-tailed P value less than 0.05 was considered statistically significant.

In this study, cyclic stretch was utilized as a stress model to simulate ventilator-induced lung injury. To characterize alveolar epithelial cell under this stress model, several cellular and molecular responses toward cyclic stretch of different durations were compared (fig. 1). Although 6-h stretch had little impact to the cell morphology throughout the culture, most of the cells are round and clustered together after a 12-h stretch, and nearly all cells were detached and suspended in the culture medium after a 24-h stretch (fig. 1A).

Consistent with the morphology changes, the differential apoptotic percentage started to be significant in the 12-h stretch group by flow cytometry analysis compared with the nonstretched control, which was even more severe after a 24-h stretch (45 ± 4% fold vs. 6 ± 1% fold; P < 0.0001; n = 4/group; fig. 1B). Cell necrosis was not significantly increased in any of these three groups (data not shown). The lactic acid dehydrogenase concentration in the stretched cells was elevated along with the increase in the stretch duration (fig. 1C).

At the molecular level, only the 24-h group showed a significant increase in DAPK1 expression in the stretched cells over the nonstretched control (1.8 ± 0.2 fold vs. 0.5 ± 0.1 fold; P < 0.0001; n = 4/group; fig. 1D), whereas the phosphorylated DAPK1 (the inactive form of DAPK1) was significantly reduced (0.5 ± 0.2 fold vs. 1.3 ± 0.2 fold; P < 0.0001; n = 4/group; fig. 1E). The ratio of phosphorylated myosin light chain (a catalytic substrate of DAPK1)–to–myosin light chain was higher after 24 h of stretching (2.0 ± 0.3 fold vs. 0.6 ± 0.1 fold; P < 0.0001; n = 4/group; fig. 1F).

To confirm whether the above observed changes were specific to the cyclic stretch model, the same evaluation was done for static stretch model (n = 4/group). A morphologic change was also observed after static stretch of cells (Supplemental Digital Content, fig. S1A, https://links.lww.com/ALN/C444). However, compared with control group, there was no significant change observed in lactic acid dehydrogenase concentration and the apoptotic percentage of cells subjected to static stretch (Supplemental Digital Content, fig. S1, B–D, https://links.lww.com/ALN/C444). The levels of DAPK1 and phosphorylated DAPK1 and the phosphorylated myosin light chain–to–myosin light chain ratio were also similar between the stretched and control cells (Supplemental Digital Content, fig. S1, E–G, https://links.lww.com/ALN/C444).

These results indicated that it was not likely that static stretching could induce alveolar epithelial cell apoptosis or DAPK1 activation. Therefore, static stretching was not chosen in the experiments following.

To determine whether the expression of DAPK1 affects cyclic stretch–induced alveolar epithelial cell apoptosis, we knocked down DAPK1 by using small interfering RNA in vitro. DAPK1 small interfering RNA transfection alone had no effect on cell morphology, apoptosis, or myosin light chain phosphorylation but verified the transfection efficiency (Supplemental Digital Content, fig. S2, https://links.lww.com/ALN/C445). Meanwhile, the cyclic stretch–induced DAPK1 activation was also inhibited after DAPK1 small interfering RNA transfection (Supplemental Digital Content, fig. S3, https://links.lww.com/ALN/C446). The changes in morphology after cyclic stretch in control small interfering RNA transfected cells were not observed in the DAPK1 small interfering RNA transfected cells. The lactic acid dehydrogenase release in DAPK1 small interfering RNA transfected cells was less than that in the control small interfering RNA group (829 ± 142 U/l vs. 1,056 ± 152 U/l, P < 0.0001; n = 4/group; fig. 2, A and B). More importantly, both cyclic stretch–induced alveolar epithelial cell apoptotic percentage and caspase 3 activity were significantly reduced in the transfected cells compared with the control small interfering RNA transfected cells (27 ± 5% fold vs. 53 ± 8% fold; 2.3 ± 0.2 fold of control vs. 4.6 ± 0.7 fold of control; P = 0.005, P < 0.0001; n = 4/group; fig. 2, C–E).

To explore the downstream pathways of DAPK1, we also examined the intrinsic mitochondrial pathway molecules. The results suggested associations between DAPK1 knockdown and enhanced B-cell lymphoma 2 expression (1.6 ± 0.2 fold vs. 1.1 ± 0.1 fold; P = 0.004; n = 4/group; fig. 2F) and between DAPK1 knockdown and repressed B-cell lymphoma 2–associated X expression (5.4 ± 0.3 fold vs. 6.1 ± 0.4 fold; P = 0.003; n = 4/group; fig. 2G). Thus, the underlying mechanism of cyclic stretch–induced alveolar epithelial cell apoptosis may involve intrinsic mitochondrial pathways.

To confirm that DAPK1 participated in cyclic stretch–induced alveolar epithelial cell apoptosis, the influence of DAPK1 overexpression on alveolar epithelial cell apoptosis was examined. It was noted that DAPK1Δcalmodulin plasmid transfection was associated with significantly higher DAPK1 expression and activation compared with the control plasmid group (Supplemental Digital Content, fig. S4, https://links.lww.com/ALN/C447), as well as cell floating and shrinkage (fig. 3A) and higher lactic acid dehydrogenase concentration (852 ± 99 U/l vs. 510 ± 71 U/l; P = 0.003; n = 4/group; fig. 3B). DAPK1 overexpression was also associated with a higher percentage of apoptotic cells (49 ± 6% fold vs. 14 ± 3% fold; P < 0.0001; n = 4/group; fig. 3, C and D), increased caspase 3 activity (4.8 ± 0.3 fold of control vs. 1.3 ± 0.2 fold of control; P < 0.0001; n = 4/group; fig. 3E), lower B-cell lymphoma 2 expression (1.1 ± 0.2 fold vs. 2.4 ± 0.5 fold; P = 0.004; n = 4/group; fig. 3F), and higher B-cell lymphoma 2–associated X expression (4.4 ± 0.1 fold vs. 2.8 ± 0.3 fold; P < 0.0001; n = 4/group; fig. 3G), relative to the control plasmid group. These data suggest that DAPK1 overexpression is capable of inducing alveolar epithelial cell apoptosis and activating intrinsic mitochondrial pathways.

We have confirmed the proapoptotic effect of DAPK1 in alveolar epithelial cells that received cyclic stretch in vitro. To determine the related effect of DAPK1 in ventilator-induced lung injury, the effect of different tidal ventilation on ventilator-induced lung injury, alveolar epithelial cell apoptosis, and DAPK1 activation were investigated in vivo (fig. 4). Compared with the lung tissues of the low V_T group, those in the high V_T group showed greater histologic changes, including damage of the alveolar structure with septal thickening and hemorrhage and significantly higher lung injury score (0.50 ± 0.05 score/field vs. 0.27 ± 0.13 score/field; P = 0.005; n = 6/group; fig. 4, A and B). The high V_T also had significant increases in the total number of cells ([8.3 ± 0.7] × 10⁴ cells vs. [4.6 ± 0.7] × 10⁴ cells; P = 0.002; n = 6/group; fig. 4C), protein concentration in bronchoalveolar lavage (657 ± 58 µg/ml vs. 516 ± 38 µg/ml; P = 0.006; n = 6/group; fig. 4D), wet-to-dry lung ratio (5.7 ± 0.9 fold vs. 4.1 ± 0.2 fold; P = 0.038; n = 6/group; fig. 4E), and levels of the proinflammatory cytokines tumor necrosis factor-α, interleukin-6, and interleukin-1β in bronchoalveolar lavage and lung tissues (fig. 4, F–K). Concomitantly, the terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate in situ nick-end labeling staining indicated a significantly higher percentage of apoptotic events in the lungs of the high V_T group compared with the low V_T group (43 ± 6% fold vs. 4 ± 2% fold; P < 0.0001; n = 6/group; fig. 4, L and M). The above evidence suggested that, in vivo, high V_T ventilation induced alveolar epithelial cell apoptosis.

Moreover, the ratios of DAPK1/phosphorylated DAPK1 (1.8 ± 0.3 fold vs. 0.5 ± 0.2 fold; P = 0.005; n = 6/group; fig. 4N) and phosphorylated myosin light chain/myosin light chain (1.7 ± 0.4 fold vs. 0.6 ± 0.1 fold; P = 0.002; n = 6/group; fig. 4O) were significantly higher in the lung tissues of the high V_T group compared with the low V_T group. All of the above indicated that ventilator-induced lung injury was induced by high V_T ventilation and associated with alveolar epithelial cell apoptosis and DAPK1 activation.

To investigate whether DAPK1 knockdown in vivo had the same effect as that in vitro, we knocked down DAPK1 in lungs and evaluated alveolar epithelial cell apoptosis and ventilator-induced lung injury. The lung injury score and the alveolar epithelial cell apoptotic percentage of the DAPK1–adeno-associated virus–infected mice were comparable with that of the control-virus group (Supplemental Digital Content, fig. S5, https://links.lww.com/ALN/C448), but the high V_T–induced DAPK1 level and activation were lower (Supplemental Digital Content, fig. S6, https://links.lww.com/ALN/C449). More importantly, compared with the control-virus group, DAPK1–adeno-associated virus infection was associated with attenuation in high V_T–induced destruction of lung tissue (lung injury score, 0.40 ± 0.06 score/field vs. 0.54 ± 0.05 score/field; P < 0.0001; n = 6/group; fig. 5, A and B), cell infiltration ([5.3 ± 0.7] × 10⁴ cells vs. [7.5 ± 0.8] × 10⁴ cells; P < 0.0001; n = 6/group; fig. 5C), protein exudation in bronchoalveolar lavage (582 ± 50 µg/ml vs. 713 ± 35 µg/ml; P < 0.001; n = 6/group; fig. 5D), and lung edema (wet to dry ratio, 4.4 ± 0.4 fold vs. 5.1 ± 0.2 fold; P = 0.001; n = 6/group; fig. 5E). At the same time, DAPK1–adeno-associated virus–pretreated mice showed lower levels of the proinflammatory cytokines tumor necrosis factor-α, interleukin-6, and interleukin-1β in bronchoalveolar lavage and lung tissues, compared with the mice in the control-virus group (fig. 5, F–K), whereas terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate in situ nick-end labeling staining indicated smaller alveolar epithelial cell apoptotic percentage induced by high V_T (35 ± 5% fold vs. 47 ± 4% fold; P < 0.0001; n = 6/group; fig. 5, L and M).

Because the activation of DAPK1 included an increase of DAPK1 expression and an enhanced activity, we used DAPK1 inhibitor to suppress DAPK1 activity in vivo to verify the detrimental effects of DAPK1 in ventilator-induced lung injury. The influence of DAPK1 inhibition on high V_T–induced ventilator-induced lung injury was investigated (fig. 6). In the DAPK1 inhibitor + high V_T group, the DAPK1 level and activation were less than that of the control group given dimethyl sulfoxide (Supplemental Digital Content, fig. S7, https://links.lww.com/ALN/C440). In DAPK1 inhibitor–pretreated mice with no or low V_T ventilation, the lung injury score, DAPK1, and alveolar epithelial cell apoptotic percentage were comparable (Supplemental Digital Content, fig. S8, https://links.lww.com/ALN/C451).

After high V_T ventilation, DAPK1 inhibitor–pretreated mice showed less lung damage (lung injury score, 0.46 ± 0.03 score/field vs. 0.60 ± 0.04 score/field; P < 0.0001; n = 6/group; fig. 6, A and B), less cell infiltration ((5.3 ± 0.7) × 10⁴ cells vs. (7.1 ± 1.5) × 10⁴ cells; P = 0.003; n = 6/group; fig. 6C), and less protein exudation in bronchoalveolar lavage (579 ± 48 μg/ml vs. 706 ± 35 μg/ml; P = 0.002; n = 6/group; fig. 6D) and alleviated lung edema (wet to dry ratio, 5.0 ± 0.3 fold vs. 5.9 ± 0.4 fold; P = 0.004; n = 6/group; fig. 6E), compared with that of the control group given dimethyl sulfoxide. Meanwhile, DAPK1 inhibitor pretreatment was also associated with lower levels of the proinflammatory cytokines tumor necrosis factor-α, interleukin-6, and interleukin-1β in bronchoalveolar lavage and lung tissues (fig. 6, F–K), and terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate in situ nick-end labeling staining showed a lower high V_T–induced apoptotic percentage (48 ± 6% fold vs. 34 ± 4% fold; P < 0.0001; n = 6/group; fig. 6, L and M).

The present study investigated the following three hypotheses: (1) DAPK1 activation is elevated in response to cyclic stretch–induced alveolar epithelial cell apoptosis in vitro, (2) cyclic stretch–induced alveolar epithelial cell apoptosis is enhanced by DAPK1, and (3) DAPK1-induced alveolar epithelial cell apoptosis correlates with the development of ventilator-induced lung injury in vivo.

Alveolar epithelial cell apoptosis is associated with epithelial dysfunction during acute respiratory distress syndrome and ventilator-induced lung injury. A previous study reported that a stretching amplitude increasing surface area by 30% would deform the alveolar basement membrane as much as total lung inflation could, which was previously reported to be associated with ventilator-induced lung injury.²¹  Therefore, this amplitude was chosen for the current study. The actual stretch experienced during mechanical ventilation of the injured and inhomogeneous lungs is very likely identical to or even more severe than the chosen stretching amplitude, even when ventilation is conducted under a protective protocol.²²  Considering apoptosis secondary to overdistension cannot be completely avoided even with minimal standard ventilation; reducing stretch-induced apoptosis can be a potential strategy to prevent ventilator-induced lung injury.

Among the numerous cell death molecules, DAPK1 is crucial for the induction of apoptosis via several cell death signals, including those generated by death receptors, cytokines, matrix detachment, etc. In the current study, gene knockdown and overexpression technology were used to evaluate the relationship between alveolar epithelial cell apoptosis and DAPK1 activation. The findings suggested DAPK1 as a positive controller of cyclic stretch–induced alveolar epithelial cell apoptosis and a transducer delivering mechanical stress.

It has been reported that mechanical stress leads to lower cellular levels of two anti-apoptotic proteins, B-cell lymphoma 2 and B-cell lymphoma–associated X,⁴  which stabilize mitochondria. Cyclic stretch–induced alveolar epithelial cell apoptosis models have revealed alterations in the intrinsic apoptosis pathway, specifically in the following components: caspase 9, cytochrome c, mitochondrial membrane potential, and reactive oxygen species.²³  In stretch-induced apoptosis models, it was reported that the extrinsic death pathway is activated via the factor-associated suicide ligand/ factor-associated suicide system.²⁴  In the present study, intrinsic mitochondrial pathway components responded to both DAPK1 suppression and activation, especially the B-cell lymphoma 2 family proteins: B-cell lymphoma 2 and B-cell lymphoma 2–associated X. The results suggested that DAPK1 promoted cyclic stretch–induced cell apoptosis via the intrinsic apoptosis pathway and eventually activated caspase 3 leading to cell death. Moreover, a prominent feature of DAPK-induced cell death is the effect on the cytoskeleton, including a loss of matrix attachment.²⁵  Activation of DAPK1 promotes apoptosis by uncoupling stress fibers and focal adhesions and by breaking the balance between contractile and adhesion forces.²⁶  This may also explain the correlation between cyclic stretch–mediated alveolar epithelial cell apoptosis and DAPK1 observed in the present study. Additional research is needed to further clarify this.

In vivo, adverse pulmonary permeability usually results from the damage to alveolar epithelium, in which alveoli encounter abnormal edema and high concentrations of extravasated macromolecules.²⁷  Mechanical stretch is one of factors damaging the alveolar–epithelial barrier. It has been well accepted that dysfunction of the alveolar epithelial barrier caused by mechanical stretch can induce alveolar flooding and the infiltration of inflammatory cells in ventilator-induced lung injury.²⁸  Although it is widely accepted that inflammation is a direct cause of lung injury during mechanical ventilation, alveolar epithelial cell apoptosis also adversely affects the epithelial barrier. In addition, excessive apoptosis will exacerbate the inflammatory response and worsen lung injury.²⁹ 

In the present study, mice were subjected to severely injurious ventilation by applying a high V_T of 35 ml/kg for 4 h, whereas the in vitro stretching time did not correspond with that. This is because mechanical ventilation in vivo may adversely affect many physiologic processes,³⁰  such as pulmonary immune responses, or induce inflammation.³¹  In addition, lung injury to an animal must be investigated within a feasible time, and therefore the duration of mechanical ventilation in vivo differs from the stretching time needed in vitro.

In the current study, the experiments with different tidal volume ventilation showed that injurious mechanical ventilation could damage alveolar epithelial barrier, and trigger inflammation and apoptosis in the lungs. A significant lung stress or strain in vivo is required for DAPK1 activation.

This study suggests that DAPK1 promoted ventilator-induced lung injury in vivo by inducing alveolar epithelial cell apoptosis. First, after the DAPK1 knockdown in the lungs, the integrity of histologic structures and the reduction in lung injury score indicated a protective effect during the DAPK1 deficiency. In addition, there was lower permeability, less high V_T–induced pulmonary inflammation, and reduced high V_T–induced alveolar epithelial cell apoptotic percentage in lungs deprived of DAPK1. Because DAPK1 activation was measured by two indicators: DAPK1 level and phosphorylated myosin light chain,¹²  DAPK1 inhibitor was applied to inhibit DAPK1 activation to strengthen the results obtained by the adeno-associated virus infection. To summarize, this study suggests that DAPK1 contributes to the in vivo development of ventilator-induced lung injury through its proapoptotic effect, whereas inhibition of DAPK1 maintains epithelial permeability and attenuates inflammation through the reduction of alveolar epithelial cell apoptotic percentage. However, we cannot attribute all the harmful effects of DAPK1 in ventilator-induced lung injury to the promotion of alveolar epithelial cell apoptosis in vivo. DAPK influences many pathways, including inflammation, cell skeleton remodeling, autophagy, etc.^25,32,33  More studies are still needed to further characterize DAPK1 in ventilator-induced lung injury.

This study has some limitations. First, the study focus was apoptosis, whereas the role of inflammation was not investigated, especially in vivo. Inflammation is a vital feature in the development of ventilator-induced lung injury, and DAPK1 also has an important function in inflammation. Second, although the in vivo studies indicated a significant increase in alveolar epithelial cell apoptosis percentage after high V_T, alveolar epithelial cell necrosis was not studied. Necrosis constitutes irreversible changes that occur after cell death and contribute to tissue damage.³⁴  More studies remain to be done to confirm whether apoptosis or necrosis is predominant in ventilator-induced lung injury. Third, the mechanism of ventilator-induced lung injury is difficult to investigate in a clinical study. More approaches are needed to testify our idea in a clinical setting. Finally, only male mice were used in the in vivo study, because the female sex hormones may affect the immune response.³⁵  Therefore, there may be some bias due to missing data for female mice.

This study suggests that DAPK1 participates in cyclic stretch–induced alveolar epithelial cell apoptosis through intrinsic apoptotic pathways and promotes ventilator-induced lung injury by inducing alveolar epithelial cell apoptosis in the lungs. Inhibition of DAPK1 expression mitigated cyclic stretch–induced alveolar epithelial cell apoptosis and ventilator-induced lung injury.

The authors thank Lu Youming, Ph.D., Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei province, China, for the gift of death-associated protein kinase 1 Δcalmoduli plasmid, and Yuwen Li, M.D., Ph.D., Hayward Genetics Center, Tulane University School of Medicine, New Orleans, Louisiana, for the modification of the manuscript.

Supported by grant Nos. 81372036 and 81601669 from the National Natural Science Foundation of China and grant No. WJ2017Q020 from the Young Scientists Funds of the Health Department in Hubei Province.

The authors declare no competing interests.