Endothelial knockdown of the tumor suppressor, WWOX, increases inflammation in ventilator-induced lung injury

Abstract

Chronic cigarette smoke exposure decreases lung expression of WWOX which is known to protect the endothelial barrier during infectious models of acute respiratory distress syndrome (ARDS). Proteomic analysis of WWOX-silenced endothelial cells (ECs) was done using tandem mass tag mass spectrometry (TMT-MS). WWOX-silenced ECs as well as those isolated from endothelial cell Wwox knockout (EC Wwox KO) mice were subjected to cyclic stretch (18% elongation, 0.5 Hz, 4 h). Cellular lysates and media supernatant were harvested for assays of cellular signaling, protein expression, and cytokine release. These were repeated with dual silencing of WWOX and zyxin. Control and EC Wwox KO mice were subjected to high tidal volume ventilation. Bronchoalveolar lavage fluid and mouse lung tissue were harvested for cellular signaling, cytokine secretion, and histological assays. TMT-MS revealed upregulation of zyxin expression during WWOX knockdown which predicted a heightened inflammatory response to mechanical stretch. WWOX-silenced ECs and ECs isolated from EC Wwox mice displayed significantly increased cyclic stretch-mediated secretion of various cytokines (IL-6, KC/IL-8, IL-1β, and MCP-1) relative to controls. This was associated with increased ERK and JNK phosphorylation but decreased p38 mitogen-activated kinases (MAPK) phosphorylation. EC Wwox KO mice subjected to VILI sustained a greater degree of injury than corresponding controls. Silencing of zyxin during WWOX knockdown abrogated stretch-induced increases in IL-8 secretion but not in IL-6. Loss of WWOX function in ECs is associated with a heightened inflammatory response during mechanical stretch that is associated with increased MAPK phosphorylation and appears, in part, to be dependent on the upregulation of zyxin.

NEW & NOTEWORTHY Prior tobacco smoke exposure is associated with an increased risk of acute respiratory distress syndrome (ARDS) during critical illness. Our laboratory is investigating one of the gene expression changes that occurs in the lung following smoke exposure: WWOX downregulation. Here we describe changes in protein expression associated with WWOX knockdown and its influence on ventilator-induced ARDS in a mouse model.

INTRODUCTION

The WW domain-containing oxidoreductase gene, WWOX, consists of a 1.2 kb exonal product that spans 1.1 Mb of the human genome on chromosome 16q23, making it one of the most active common chromosomal fragile sites in humans (1). As such, genotoxic environmental exposures can cause homozygous deletions, loss of heterozygosity, and translocation breakpoints at the WWOX locus (2). This includes exposure to cigarette smoke, and thus WWOX function in the lung is of interest from the standpoint of tobacco-related disease pathogenesis. Previous observations in humans indeed suggest significant downregulation of gene and protein expression of WWOX in the lung following chronic cigarette smoke exposure (3).

The implications of WWOX knockdown for lung disease pathogenesis have been studied in two previous papers. Intratracheal instillation of Wwox-targeting siRNA in mice was observed to result in further potentiation of neutrophilic inflammation in a lipopolysaccharide (LPS)-induced murine model of lung injury (4). Endothelial cell-specific Wwox knockout (EC-Wwox KO) mice stimulated by intratracheal LPS or methicillin-resistant Staphylococcus aureus (MRSA) exhibited significantly greater levels of vascular leak and histological lung injury (3).

These observations are of particular interest in the context of observed associations between prior cigarette smoke exposure and decreased resilience against acute respiratory distress syndrome (ARDS) during conducive conditions. ARDS is a critical illness that occurs following events such as acute aspiration, pneumonia (bacterial or viral), sepsis, blunt force trauma, and nonlung protective mechanical ventilation often required to support patients with severe respiratory failure (including ARDS) (5–7). An increased progression to clinically detectable ARDS has been observed in cigarette smokers versus nonsmokers following blunt force trauma, cardiac surgery, transfusion of blood products, or nonpulmonary sepsis (8–12). Cigarette smoke worsens human (8, 12, 13) and murine forms of ARDS induced by pulmonary sepsis (14) where it has been observed to exacerbate vascular leak.

The mechanistic underpinning for the observed Wwox-deficient lung phenotype is unknown. In the current study, tandem mass tag-mass spectrometry (TMT-MS) was leveraged to determine the most differentially expressed proteins in human endothelial cells during WWOX knockdown relative to wild-type controls. Furthermore, 2,051 distinct proteins were identified and quantitated at a 1% false discovery rate. Among the most significant differentially expressed proteins, two were identified that have previously been biologically connected to the inflammatory and endothelial barrier-susceptible phenotypes observed in prior work. These were the mechanotransducer protein, zyxin, which was upregulated during WWOX knockdown, and the mitochondrial superoxide dismutase, SOD2, which was downregulated. Zyxin is a known modulator of inflammatory responses in endothelial cells following mechanical stretch (15). Along with known interactions between WWOX and the MAP kinases (16), which are also involved in transducing gene transcriptional responses to stretch in ECs (17), these observations led to an examination of the potential involvement of WWOX in VILI-associated inflammation.

MATERIALS AND METHODS

Reagents

RBC lysis buffer, DMEM, bovine serum albumin, DNase, Dispase II, and RIPA buffer were purchased from Sigma Aldrich. Bicinchoninic acid (BCA) Protein Assay Kit, Shandon Kwik-Diff kit, siPORT Amine transfection reagent, Lipofectamine 2000, High-Capacity cDNA Reverse Transcription kit, TaqMan Gene Expression Master Mix, and TaqMan Gene Expression Assays for GAPDH, zyxin, and WWOX were from Thermo Fisher. Mini-PROTEAN TGX Precast Gels were from Bio-Rad. QIAShredder and RNeasy Mini kits were from Qiagen. Anti-WWOX antibody (ab137126;GR134023-45) was from Abcam (Cambridge, MA). Human lung microvascular endothelial cells (ECs), and endothelial cell growth media were purchased from Lonza (Basel, Switzerland). Heparin was purchased from Hospira (Lake Forest, IL). Collagenase was from Worthington (Lakewood, NJ). ELISA kits for mouse cytokines (KC, MCP-1, IL-6, IL-1β), anti-mouse CD16/32 (101302;B26487), anti-mouse CD31-PE/Cy7 (102524;B221707), and anti-mouse CD45-Alexa700 (304023;B304957) antibodies were from BioLegend (San Diego, CA) and R&D systems (Minneapolis, MN). siGENOME WWOX-targeting and scrambled control siRNA was from Dharmacon/GE, and the zyxin siRNA was from Santa Cruz Biotechnology (Santa Cruz, CA). pCMV entry vector containing the ORF corresponding to (Myc-DDK-tagged)-human WW domain-containing oxidoreductase (WWOX), transcript variant 1, as well as a (Myc-DDK-tagged)-empty vector were purchased from OriGene. Anti-zyxin antibody (PA125162; XJ3727705) was from Thermo Fisher Scientific, Inc., (Waltham, MA), and the concentration used was 1:1,000 for Western blot and 1:100 for IF. Anti-phospho-p44/42 MAPK(Erk1/2 (9154), anti-phospho-p38 MAPK (Thr180/Tyr182) (9211), anti-phospho-SAPK/JNK (Thr183/Tyr185) (4668), and anti-β-actin antibodies (4967) were from Cell Signaling (Danvers, MA). Unless specified, concentrations of antibodies used in each application were as per the manufacturer’s specifications.

Cell Culture

Human pulmonary artery endothelial cells (ECs) obtained from Lonza were grown in cell culture using endothelial growth media containing supplemental growth factors and 2% fetal bovine serum according to the manufacturer’s instructions. Cells underwent a maximum of three passages for use in experiments.

For separate studies, ECs were isolated from the lungs of 8-wk-old C57BL/6-Wwox^flox/flox and tamoxifen-treated Cdh5-CreERT2/Wwox^flox/flox mice as detailed in the study by Kawasaki et al. (18). Following ketamine/xylazine administration, the thoracic cavity was surgically opened. Sterile PBS (20 mL) containing 10 U/mL heparin was instilled into the pulmonary vasculature via injection through the right ventricle until the lungs were blood free. Mouse lungs were then harvested, minced, and digested in an enzyme complex as previously described. Washed cell pellets were resuspended and serially incubated with anti-mouse CD16/32 (to block Fc receptors), anti-mouse CD31-PE/Cy7, and anti-mouse CD45-Alexa700 antibodies. After washing with PBS/BSA, cells were sorted by flow cytometry (MoFlo Astrios Cell Sorter, Beckman Coulter, Indianapolis, IN) to isolate CD31 + CD45- ECs (99% purity) which were then placed in EC growth media on gelatin-coated plates.

Mechanical Cyclic Stretch of Endothelial Cells

ECs were plated onto six-well Flexcell plates which were coated with type I collagen (FlexCell international, Hillsborough, NC) and grown to 75%–80% before transfection of siRNAs as described in In Vitro Silencing of Wwox and Zyxin in Human Pulmonary Endothelial Cells below. Three days after the transfection of siRNA, mechanical stretch was performed on the FlexCell Strain Unit (FX-3000, FlexCell International). The device cyclic stretch cells by a vacuum system. The stretch condition was 18% elongation at a frequency of 30 cycles per minute (0.5 Hz) for 4 h which are considered physiologically relevant and have been used for the study of VILI mechanisms in cell culture by leading experts in the field (19). After finishing cyclic stretch, the medium was harvested for ELISA assay and the cell lysates were harvested with cell lysate buffer for Western blot (19).

Animals

All experiments and animal care procedures were approved by the University of Illinois at Chicago Animal Care and Use Committee. C57BL/6-Wwox^flox/flox mice were donated generously by Dr. C. Marcelo Aldaz, of MD Anderson Center at the University of Texas. C57BL/6-Cdh5-CreERT2 mice are shared within our larger investigative group, originally provided by Dr. Ralf Adams (20). They are currently commercially available at Taconic Biosciences. The two strains were crossed to produce Cdh5-CreERT2/Wwox^flox/flox mice as in the study by Ludes-Meyers et al. (21). The result is a tamoxifen-inducible, endothelial cell-specific Wwox knockout mouse strain that has been previously validated by our group (3). To further address any concerns about a recombination event in an EC progenitor cell leading to tamoxifen-induced knockdown in an immune cell, we measured Wwox expression in bronchoalveolar lavage (BAL) immune cells from these mice (Supplemental Fig. S1). They were housed in cages in a temperature-controlled room with a 12-h dark/light cycle, and with free access to food and water.

Murine high tidal volume ventilation.

Six-week-old male Cdh5-CreERT2/Wwox^flox/flox mice and control animals received intraperitoneal injections of tamoxifen (75 mg/kg) for 5 consecutive days. Seven days later they were subjected to high tidal volume mechanical ventilation (VT = 40 mL/kg, 65 breaths/min, PEEP 15–20, 21% $\text{F}{\text{I}}_{{\text{O}}_2}$, 25°C) for 4 h to induce VILI as we have previously described (19). Bronchoalveolar lavage (BAL) fluid was collected. BAL fluid was assessed for total cell counts, protein concentration, and cytokine levels. Meanwhile, the mouse lungs were flushed and harvested for histological analysis (19).

Bronchoalveolar Lavage Fluid Collection and Analysis

Mice were anesthetized with ketamine/xylazine. The trachea was exposed surgically, and a small incision was made on the anterior surface. An 18-gauge blunt-end cannula was inserted into this opening and secured with a suture tied around the trachea. Sterile saline (1 mL) was infused through this cannula and slowly withdrawn. This BALF was centrifuged to pellet cells and separate supernatant. RBCs were removed from the cell pellet using RBC lysis buffer according to the manufacturer’s instructions. The remaining cells (leukocytes) were pelleted and resuspended in 200 μL of Hank’s buffer for measurement of cell concentration with an automated cell counter. The suspension was cytospun onto a glass slide for staining with the Shandon Kwik-Diff kit. A manual differential cell count was performed on 10 high-powered fields to determine relative percentages of neutrophils in the leukocyte population. Protein concentration was measured using the BCA Protein Assay kit, and the corresponding ELISA kits were used to measure concentrations of KC, MIP-2, MCP-1, IL-6, IL-1β, and TNF-α.

Mouse Lung Endothelial Cells for RT-PCR, Western Blotting, and Histology

Following BALF collection, the right ventricle was infused with sterile saline until the lungs were free of intravascular blood. The right middle lobe was excised and snap-frozen in liquid nitrogen for RT-PCR and Western blotting. The remaining lobes were excised and placed in 10% formalin for 24 h followed by transfer to 70% ethanol for storage until paraffin embedding and sectioning for H&E staining.

For RT-PCR, lungs were thawed, lysed, and homogenized, and RNA was isolated using the QIAShredder and RNeasy Mini kits per the manufacturer’s instructions. cDNA was produced for each sample using the high-capacity cDNA Reverse Transcription kit from ThermoFisher Scientific. Real-time PCR was performed using the TaqMan Gene Expression Master Mix and Assays as per the manufacturer’s instructions. Relative quantifications between controls and experimental samples of Wwox and zyxin gene expression normalized for GAPDH and 18S, respectively, were determined using the ΔΔC_t method.

For Western blotting, snap-frozen lungs were thawed and homogenized in RIPA buffer. Homogenized samples were centrifuged to pellet debris and isolate protein lysates. Protein concentration of lysates was measured using the BCA Protein Assay kit, and samples were diluted in RIPA buffer until all had the same protein concentration. After dilution with SDS buffer, samples were loaded into 4%–20% Mini-PROTEAN TGX Precast Gels and subjected to SDS-PAGE. Gels were transferred to nitrocellulose membranes which were then subjected to Western blotting using antibodies targeting proteins of interest in accordance with the manufacturer’s instructions. Densitometric analysis was performed using ImageJ software.

In Vitro Silencing of Wwox and Zyxin in Human Pulmonary Endothelial Cells

In vitro silencing of Wwox in human ECs (Lonza) was achieved using transfection of siRNA as described in the study by Wolfson et al (22). WWOX-targeting (CCAAGGACGGCUGGGUUUA) and/or Zyxin-targeting (GUGUUACAAGUGUGAGGACTT) versus scrambled control siRNA (UAGCGACUAAACACAUCAA) was complexed with siPORT Amine transfection reagent. Serum-free media was used to bring the concentration of siRNA to 100 nM. In experiments in which dual-silencing of WWOX and Zyxin was achieved, 45 nM concentrations of each siRNA were used together. Transfection of ECs was then performed as described previously (22). Seventy-two hours after transfection, ECs were analyzed by Western blotting for WWOX and zyxin expression.

Human or mouse endothelial immunochemistry assay.

Human endothelial cells were grown to 75% confluence on glass slides in 12-well plates, the cells were silenced with 100 nM of human scrambled control, WWOX-targeting siRNA, zyxin siRNA, or WWOX/zyxin siRNA for 3 days. The cells were fixed with paraformaldehyde followed by 0.1% Triton incubation. Cells were then incubated with 1% BSA for 30 min followed by overnight incubation with the antibodies indicated. On the next day, the cells were incubated with secondary antibodies for 1 h. After washing with PBS, the slides were mounted with DAPI mount solution. After the slides were dried, wide-field images were acquired using an Olympus BX51 microscope with ×20/0.4 numerical aperture (NA) PLN and ×40/0.75 NA UPLFLN objectives.

Nuclear Translocation of Zyxin

Images obtained for the four different siRNA conditions were analyzed by a reviewer blinded to the conditions. The red (DAPI) and green (zyxin) channels were merged and the resultant multifluorescent images were analyzed for the presence of yellow nuclear staining by visual inspection. At least one hundred cells were counted per condition by a blinded reviewer. The presence of yellow color in the nucleus scored a cell as positive for zyxin-nuclear translocation. The number of zyxin-nuclear positive cells divided by 100 equals the percent of nuclear zyxin for each condition.

Proteomic Analysis

TMT-MS was used to obtain relative quantitation of protein expression between control and WWOX-silenced human ECs as described previously (23–25). Briefly, samples were digested, and 50 µg of peptides from each condition were labeled with TMT tags, combined, fractionated, and desalted with C18 spin columns. The labeled peptides were subjected to MS analysis and run in triplicate. Data were analyzed against a human database using Scaffold 5. In total, 2,051 proteins, including 20 decoys, were identified. Differential protein expression between groups was assessed by comparing mean normalized intensity values utilizing Scaffold’s unpaired t test and considered significant upon correction for multiple testing at a Benjamini–Hochberg false discovery rate <1%. Log₂ fold change was defined as the difference of log₂ normalized intensity values between WWOX-silenced human ECs and controls.

Statistical Analysis

GraphPad Prism software was used to carry out the statistical analyses. In studies in which one independent variable was manipulated to contribute to the final outcome, a Student’s t test was performed to evaluate the differences between groups. In studies in which two independent variables were manipulated and contributed to the final outcome being measured (e.g., BAL cytokine level as influenced by spontaneous breathing vs. VILI and control vs. WWOX KO), a two-way ANOVA was used to determine which of the two independent variables influenced the outcome as well as whether there was any interaction between the two independent variables. Following examination of the F-statistics and P values for each variable which verified no interaction between independent variables and the significance of each, a Tukey’s HSD post hoc test was used to determine which groups were significantly different from each other. The P values reported in the figures for these studies are the result of Tukey’s HSD test. In all cases, a P value < 0.05 was considered significant.

RESULTS

TMT-MS Proteomics and Validation Studies Identified Zyxin as a Significant Differentially Expressed Protein in WWOX-Silenced ECs Compared with Wild-Type Controls

TMT-MS was used to profile the proteome of control versus WWOX-silenced ECs. Silencing was confirmed by Western blotting and immunostaining (Fig. 1, A and B). Results of the proteomic analysis are shown in Fig. 1C and Supplemental Table S1. Of the most significant differentially expressed proteins were the mechanotransducer, zyxin, and the mitochondrial superoxide dismutase, SOD2. Validation studies for zyxin were performed using RT-PCR and Western blotting which reproduced the TMT-MS result at both the mRNA and protein levels (Fig. 1, D and E).

Figure 1.

WWOX knockdown is associated with upregulation of the mechanosensitive transducer, zyxin. A: ECs were transfected with either scrambled, control, or WWOX-silencing siRNA. Western blotting illustrates the degree of WWOX knockdown achieved. B: immunostaining for WWOX confirms the knockdown of the predominantly cytosolic protein in siRNA-transfected ECs. C: TMT-MS differential protein expression between WWOX-silenced and control ECs is depicted for uniquely identified proteins on the volcano plot and is detailed in Supplemental Table S1. The horizontal dotted line indicates the t test’s significance threshold (Benjamini–Hochberg adjusted P value < 1%). Vertical dotted lines indicate a log₂ fold change of −0.5 or 0.5. SOD2 (superoxide dismutase 2) and ZYX (zyxin) were among the most significant differentially expressed proteins and were labeled for reference. D: zyxin upregulation during WWOX knockdown was confirmed by Western blotting. E: qPCR for zyxin revealed that the upregulation occurs at the mRNA level (n = 6 separate transfections, means ± SD, *P < 0.05). EC, endothelial cell; TMT-MS, tandem mass tag mass spectrometry.

Loss of WWOX Promotes EC Cytokine Production in Response to Cyclic Stretch

An 18% elongation cyclic stretch (CS) of WWOX-silenced endothelial cells was used to mimic high tidal volume ventilation in vitro. After 4 h of cyclic stretch (18% elongation, 0.5 Hz), cytokine assays done on EC culture media showed that silencing of WWOX significantly increased CS-mediated production of IL-6 (239.5 ± 21.8 vs. 113.77 ± 5.0 pg/mL) and IL-8 (128.42 ± 3.11 vs. 76.10 ± 3.5 pg/mL) (Fig. 2).

Figure 2.

Silencing of WWOX enhanced endothelial inflammatory response to cyclic stretch. Human ECs were seeded on a Flexcell plate and silenced with siRNA control or siRNA of WWOX for 3 days. After cyclic stretch, the media were harvested for ELISA. IL-6 and IL-8 production increased significantly in control ECs subjected to stretch compared with static controls. Compared with the stretched ECs transfected with siRNA control, IL-6 and IL-8 production from stretched WWOX-silenced ECs was significantly increased (n = 6 separate transfections, means ± SD, ****P < 0.05). EC, endothelial cell.

MAPK signaling is a well-known effector pathway of mechanosensitive gene expression changes in ECs (26). Therefore, ERK and JNK phosphorylation status were examined in WWOX-silenced endothelial cells. As shown in Fig. 3, CS induced time-dependent ERK phosphorylation in control siRNA-transfected endothelial cells with peak phosphorylation occurring at 5 min. However, silencing of WWOX induced an increase in basal and CS-mediated ERK phosphorylation compared with control siRNA-transfected endothelial cells (Fig. 3A). Similarly, silencing of WWOX also induced an increase of basal and CS-mediated JNK phosphorylation compared with control siRNA-transfected endothelial cells (Fig. 3A). Interestingly, silencing of WWOX significantly decreased basal and CS-mediated p38 MAPK phosphorylation compared with controls (Fig. 3A).

Figure 3.

Silencing of WWOX enhanced MAPK activity in endothelial inflammatory response to cyclic stretch. A: human ECs were seeded on a Flexcell plate and silenced with siRNA control or siRNA of WWOX for 3 days. After cyclic stretch, cellular lysates were harvested for Western blotting. The phospho-kinase lanes and the total kinase lanes are derived from two separate gels. Actin bands corresponding to each gel are shown. 18% elongation, 0.5 Hz cyclic stretch induced Erk 1/2, SAPK/JNK (Thr183/Tyr185), and p38 phosphorylation in both control and WWOX-silenced ECs. Silencing of WWOX induced significantly higher Erk 1/2 and SAPK/JNK phosphorylation and lower p38 phosphorylation during both basal and cyclic stretch conditions. Note that for SAPK and JNK isoforms, there are two bands each corresponding to singly and doubly phosphorylated proteins. B: the bar graph illustrates a densitometric analysis of the two gels. The density of each band was normalized to the corresponding actin band, and the ratio of actin-normalized phospho- to total kinase densities was calculated for each condition. EC, endothelial cell; MAPK, mitogen-activated kinases.

EC Wwox KO Mice Exhibit More Injury during VILI

To further investigate the role of WWOX in VILI, Wwox^flox/flox (control) mice and EC Wwox KO mice underwent high tidal volume ventilation for 4 h (40 mL/kg, 65 breath/min). BAL protein and cell counts assay indicated that EC Wwox KO mice had significantly increased BAL protein concentration (0.82 ± 0.16 vs. 0.53 ± 0.08 mg/mL, ECWX-KO vs. ECWX-Flox) and cell infiltration (1.59 ± 0.34 vs. 0.91 ± 0.17 × 10⁵/ml, ECWX-KO vs. ECWX-Flox) following VILI (Fig. 4A). H&E staining of lung sections showed VILI-induced cell infiltration in both ECWX-Flox and ECWX-KO lung interstitium. Consistent with BAL cell counts, VILI induced significantly higher cell infiltration in ECWX-KO lung compared with ECWX-Flox (Fig. 4B).

Figure 4.

Endothelial-specific Wwox KO mice exhibit greater vascular leak during high tidal volume ventilation. EC Wwox KO and control mice were subjected to high tidal volume mechanical ventilation as described in materials and methods. A: VILI induced significantly higher BAL protein concentrations and cell counts in control mice. Significantly higher protein concentration and cell counts were present in BAL fluid from EC Wwox KO mice subjected to VILI compared with corresponding controls (n = 10 separate mice per group, means ± SD, */**/****P < 0.05). B: H&E stain of sections of lung tissue showed increased VILI-induced interstitial and alveolar cellular infiltration in EC Wwox KO mice lung compared with controls. BAL, bronchoalveolar lavage; EC, endothelial cell; KO, knockout.

Increased BAL protein and alveolar cellular infiltration in ECWX-KO versus ECWX-Flox mice after VILI correlated with increased BAL IL-6, KC, IL-1β, and MCP-1 production in WXEC-KO mice groups (Fig. 5).

Figure 5.

Endothelial-specific Wwox KO mice exhibit a greater inflammatory response to high tidal volume ventilation. BAL cytokines in VILI-exposed EC Wwox KO and control mice were measured by ELISA. These showed that VILI induced significantly higher IL6, KC, IL-1β, and MCP-1 generation in EC Wwox KO mice (n = 10 separate mice per group, means ± SD, */**/***/****P < 0.05). BAL, bronchoalveolar lavage; EC, endothelial cell; KO, knockout; MCP, monocyte chemoattractant protein.

ECs Isolated from EC Wwox KO Mice Produce More Cytokines following Cyclic Stretch

Endothelial cells were isolated from ECWX Flox (control) and ECWX-KO mice as described previously (3). Flow cytometry for CD31+ CD45− cells was performed, and these endothelial cells were then cultured on gel-coated plates.

Cyclic stretch-mediated cytokine generation from isolated WXEC-flox and WXEC-KO endothelial cells was measured. As shown in Fig. 6, 4 h of 18% elongation cyclic stretch stimulated the production of IL-6, KC, IL-1β, and MCP-1 in WXEC-Flox endothelial cells medium compared with static control. Elevations of all these cytokines were greater in WXEC-KO ECs undergoing stretch when compared with CS-mediated elevations in WXEC-Flox cells: IL-6 (16.43 ± 11.59 pg/mL vs. 44.17 ± 13.47 pg/mL, P < 0.05); KC (16.08 ± 3.76 pg/mL vs. 65.67 ± 17.84 pg/mL, P < 0.05); IL-1β (13.43 ± 17.14 pg/mL vs. 38.28 ± 8.97 pg/mL, P < 0.05) and MCP-1 (41.12 ± 29.82 pg/mL vs. 90.02 ± 10.46 pg/mL, P < 0.05) (Fig. 6).

Figure 6.

Endothelial cells derived from EC Wwox KO mice exhibit greater cyclic stretch-induced cytokine generation. ECs were isolated from EC Wwox KO mice and controls. These were grown on FlexCell plates subjected to cyclic stretch. After 4 h, the medium was harvested for ELISA-based measurement of cytokines. ECs from KO mice exhibited significantly higher IL6, KC, IL-1β, and MCP-1 generation following cyclic stretch compared with controls (n = 5 separate cultures, means ± SD, *P < 0.05). EC, endothelial cell; KO, knockout; MCP, monocyte chemoattractant protein.

Silencing of Zyxin during WWOX Knockdown in ECs Mitigates Stretch-Induced Increases in IL-8 Production

The mechanotransducer, zyxin, has previously been shown to be an enhancer for IL-8 transcription during cyclic stretch of ECs (15). To determine the relative contribution of zyxin upregulation to the increases in stretch-induced increases in IL-8 production during WWOX knockdown, cosilencing of WWOX and zyxin in human ECs was utilized (Fig. 7). Zyxin protein levels were noted to be higher during dual WWOX/zyxin knockdown compared with single gene knockdown of zyxin. To investigate this further, zyxin mRNA levels were examined in all of the conditions. They were found to be increased during WWOX knockdown and back to baseline control levels during dual WWOX/zyxin knockdown. This implies that WWOX knockdown increases zyxin mRNA levels either by increasing zyxin promoter activity or decreasing zyxin mRNA degradation. Silencing of zyxin during WWOX knockdown resulted in a 35% reduction in CS-induced IL-8 secretion compared with ECs in which WWOX was solely silenced (Fig. 7). Interestingly, IL-6 secretion which is increased during WWOX knockdown was not diminished by simultaneous loss of zyxin expression (Fig. 7).

Figure 7.

Silencing of Zyxin during WWOX knockdown decreases IL-8 production following cyclic stretch. A: human ECs were treated with control, WWOX-targeting, and zyxin-targeting siRNA. Dual-silencing of WWOX and zyxin was achieved with treatment involving both siRNAs. A representative Western blot of lysates is shown. B: densitometric analysis of several blots from these experiments is reported in the bar graph (n =4 separate transfections, means ± SD, *P < 0.05). C: qPCR for zyxin mRNA during silencing of WWOX alone reveals an increase in mRNA expression. During dual silencing of WWOX/zyxin, zyxin mRNA reverts to control levels. (n = 6 separate transfections, means ± SD, *P < 0.05) D: following cyclic stretch, culture medium was harvested for measurement of IL-8 generation. The greater amount of IL-8 generation exhibited by WWOX-silenced ECs compared with controls following cyclic stretch appears to be abrogated by zyxin knockdown (n = 3 separate transfections, means ± SD, *P < 0.05). E: the same culture medium was analyzed by ELISA for IL-6 generation. Although there was an increase in IL-6 generation associated with WWOX silencing, this was not abrogated by zyxin knockdown. (n = 3 separate transfections, means ± SD, *P < 0.05). EC, endothelial cell.

Silencing of WWOX Increases Nuclear Localization of Zyxin

An assessment of zyxin nuclear localization during WWOX knockdown, necessary for zyxin’s IL-8 enhancing function, was performed. As shown in Fig. 8, silencing of WWOX was associated with significantly increased nuclear localization of zyxin.

Figure 8.

Increased zyxin nuclear localization upon WWOX knockdown. A and B: ECs were treated with control siRNA, WWOX siRNA, WWOX and zyxin siRNA, and zyxin siRNA and labeled with anti-zyxin antibody as well as with the nuclear dye DAPI. The percentages of cells with zyxin nuclear staining in each condition were counted. Scale bar = 20 µm. (n = 3 separate transfections, means ± SD, *P < 0.05). Arrow indicates nuclear localization of zyxin. EC, endothelial cell.

DISCUSSION

In this study, observations of the WWOX-deficient lung phenotype, which is associated with tobacco smoke exposure, have been extended to high tidal volume ventilation-mediated acute lung injury. Silencing of the WWOX gene in ECs results in a proinflammatory, barrier-sensitive phenotype characterized, in part, by the upregulation of the mechanosensitive IL-8 enhancer, zyxin. Furthermore, the transduction of signals in ECs following mechanical stress occurs through focal adhesion complexes via mitogen-activated kinases (MAPK). These signaling pathways appear to be potentiated in WWOX-deficient ECs with increases in both CS-mediated ERK and JNK phosphorylation. EC Wwox KO mice subjected to a VILI model exhibited, relative to VILI-exposed wild-type mice, increased BAL protein, accumulation of alveolar cellular infiltrates, and proinflammatory cytokines (IL6, KC, IL-1β, and MCP-1). Isolated endothelial cells from EC-Wwox KO mice subjected to CS resulted in significantly higher production of cytokines when compared with stretched mouse EC-Wwox flox endothelial cells.

In pursuing the mechanisms underlying the WWOX-deficient endothelial cell phenotype, TMT-MS for analysis of differential protein expression between WWOX-silenced ECs and control cells was performed. Two significant differences in protein expression were identified. The mitochondrial superoxide dismutase, SOD2 was markedly downregulated in WWOX-silenced ECs. Interestingly, a proteomic study in Drosophila has previously demonstrated genetic interactions between WWOX and various superoxide dismutases (27). SOD2 converts mitochondrial superoxide to hydrogen peroxide (28). In human ECs, loss of SOD2 such as that which occurs in sickle cell disease, is associated with decreased mitochondrial potential without impairment in respiration/ATP production and increased EC layer permeability (28). Lipopolysaccharide (LPS) or bacterial endotoxin, an ARDS-relevant agonist, is known to increase EC barrier permeability, at least in part, by an increase in mitochondrial ROS and subsequent VE-Cadherin phosphorylation (29). Therefore, the loss of SOD2 during WWOX knockdown is currently being studied in the context of LPS- and other bacterial infection-related stimuli in ECs.

The other significant proteomic profile finding in WWOX-silenced ECs is the upregulation of the mechanotransducer, zyxin. Zyxin is primarily located in focal adhesion plaques but can shuttle to actin stress fibers and the nucleus in response to mechanical stimulation (30). Nuclear translocation of zyxin has been shown to be associated with gene expression changes, including upregulation of IL-8 during mechanical stretch of ECs (15, 30). In the current study, the analysis of zyxin nuclear translocation is limited by the absence of nuclear fractionation studies or confocal imaging with z-stacking. However, the data do strongly suggest that zyxin upregulation during WWOX-silencing plays a role in the exaggerated IL-8 response of WWOX-deficient ECs during mechanical stretch and VILI. The connection between zyxin and WWOX appears to be at the level of mRNA regulation. One possible mechanism underlying this phenomenon is via the previously reported direct inhibitory interaction between SMAD3 and WWOX (31), and SMAD-3-dependent upregulation of zyxin (32). The relationship between increased MAPK signaling and zyxin upregulation during mechanical stretch of WWOX-deficient ECs is also the subject of future study.

The pathological change of VILI includes overdistention of alveoli from high tidal volume and repetitive opening and closing of alveoli, with resultant inflammatory response leading to disruption of the endothelial barrier and interstitial and alveolar infiltration of monocytes and neutrophils (33). Why some patients sustain VILI and recover quickly and others do not may be, in part, related to chronic gene expression changes associated with past exposures such as cigarette smoke. WWOX downregulation appears to be one of these changes with an impact on the molecular mechanisms of stretch-induced inflammatory responses in ECs.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Table S1 and Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.25137089.

GRANTS

This work was supported by the National Natural Science Foundation of China (82060361; 82360373) and Natural Science Foundation of Jiangxi Province Grants 20224ACB206020 (to Z.Z.); R01-HL-127342, R01-HL-133951, and 2R01 HL111656-06 (to R.F.M.); and 1K08 HL140222-01A1 (to S.S.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

E.A., W.C., R.F.M., and S.S. conceived and designed research; Z.Z., E.A., and W.C. performed experiments; Z.Z., E.A., W.C., C.A., and S.S. analyzed data; Z.Z., E.A., W.C., C.A., T.P., J.R.J., S.M.D., V.N., C.M.A., R.F.M., and S.S. interpreted results of experiments; Z.Z., E.A., W.C., C.A., and T.P. prepared figures; Z.Z. and T.P. drafted manuscript; E.A., W.C., C.A., T.P., J.R.J., S.M.D., V.N., C.M.A., R.F.M., and S.S. edited and revised manuscript; Z.Z., E.A., W.C., C.A., T.P., J.R.J., S.M.D., V.N., C.M.A., R.F.M., and S.S. approved final version of manuscript.