Vascular permeability disruption explored in the proteomes of mouse lungs and human microvascular cells following acute bromine exposure

Abstract

Bromine (Br₂) is an organohalide found in nature and is integral to many manufacturing processes. Br₂ is toxic to living organisms, and high concentrations can prove fatal. To meet industrial demand, large amounts of purified Br₂ are produced, transported, and stored worldwide, providing a multitude of interfaces for potential human exposure through either accidents or terrorism. To identify the key mechanisms associated with acute Br₂ exposure, we have surveyed the lung proteomes of C57BL/6 male mice and human lung-derived microvascular endothelial cells (HMECs) at 24 h following exposure to Br₂ in concentrations likely to be encountered in the vicinity of industrial accidents. Global discovery proteomics applications combined with systems biology analysis identified robust and highly significant changes in proteins associated with three biological processes: 1) exosome secretion, 2) inflammation, and 3) vascular permeability. We focused on the latter, conducting physiological studies on isolated perfused lungs harvested from mice 24 h after Br₂ exposure. These experiments revealed significant increases in the filtration coefficient (K_f) indicating increased permeability of the pulmonary vasculature. Similarly, confluent monolayers of Br₂ and Br-lipid-treated HMECs exhibited differential levels of zona occludens-1 that were found to be dissociated from cell wall localization, an increase in phosphorylation and internalization of E-cadherin, as well as increased actin stress fiber formation, all of which are consistent with increased permeability. Taken as a whole, our discovery proteomics and systems analysis workflow, combined with physiological measurements of permeability, revealed both profound and novel biological changes that contribute to our current understanding of Br₂ toxicity.

INTRODUCTION

Bromine (Br₂) is the most common organohalide found in nature (25). Addition of Br₂ to water forms hydrobromic (HBr) and hypobromous (HOBr) acids (Fig. 1; Refs. 25, 48). In biological systems, peroxidases catalyze the oxidation of bromide and hydrogen peroxide, generated by eosinophils to produce HOBr, a bactericidal agent (Fig. 1; Refs. 5, 42, 53). Similar to other halogens, bromine is extracted from subterranean deposits, sea brines, and salt lakes to produce its pure form, Br₂ (brown acrid liquid and gas; Ref. 52). The Dead Sea alone is estimated to contain nearly a billion tons of bromine and is a major site of harvesting for purification into Br₂ (52). This form of Br₂ is commonly used throughout the chemical industry with world production currently believed to reach upward of nearly half a billion pounds annually (56a). Primary countries involved in the production and use of Br₂ include the United States, Israel, Jordan, and China (52).

Fig. 1.

Lipid-bromine reactions. Although Br₂ is slightly soluble in aqueous (aq.) solutions (~300 µg/mL) at neutral pH, it will also react with H₂O to form bromide ions (Br⁻) and hypobromous acid (HOBr). Hypobromous acid is also formed through the peroxidase-driven catalytic reaction between the bromide ions and H₂O₂. The most specific peroxidase driving this reaction is eosinophil peroxidase, but other peroxidases (such as myeloperoxidase) will form HOBr when Br⁻ exist at high concentrations. Plasmalogens are a subclass of ethanolamine glycerophospholipids (PE) and choline glycerophospholipids (PC) that are reactive toward HOBr and have been shown to form 2-bromopalmitaldehyde (2-BrPALD), which may be either reduced (Red) to 2-bromopalmitoyl alcohol or oxidized (OX) to form 2-bromopalmitic acid. 2-BrPALD may also react with glutathione (GSH), which exists in abundance in lung tissues and bronchoalveolar lavage to form 2-GS-PALD. Lysophospholipids (LPLs) such as lysophosphatidic acid (LPA) are also formed, which activate lysophospholipid receptors (LPLRs) within the G protein-coupled receptor (GPCR) family. Eth, ethyl; SG, conjugated glutathione; SNx, position on the glycerophospholipid (GPL) backbone.

Br₂ is used in the production of many products, including brominated fire retardants, drilling fluids, agriculture materials, and biocides for the treatment of water (44). To meet broad industrial demands, large amounts of purified Br₂ are produced, transported, and stored worldwide, providing a multitude of interfaces for potential human exposure through either accident or terrorism (22). Exposure of humans and mouse models to liquid Br₂ can cause severe burns, whereas inhalation of Br₂ damages the respiratory track, potentially leading to pneumonia and death from respiratory failure (26, 44). Similar to humans, surviving mice often develop chronic lung diseases such as pulmonary fibrosis (3). Thus there is a need to develop animal models of Br₂ toxicity to identify which proteins and biological processes are significantly affected and, with this knowledge, to develop countermeasures.

Previously, we (4) have shown that C57BL/6 mice exposed to Br₂ in environmental chambers and returned to room air develop severe acute lung injury resembling human acute respiratory distress syndrome (ARDS). The observed phenotype is characterized by hypoxemia, hypercapnia, inflammation, lung cell apoptosis and necrosis, and the appearance of protein-rich edema in the alveolar spaces due to increased permeability of the pulmonary endothelial and alveolar epithelial cells to plasma and plasma proteins (4). Approximately 50% of the mice exposed to 600 ppm Br₂ for 30 min died within 5 days of exposure. Surviving mice developed increased lung compliance and enlarged alveolar spaces, a phenotype resembling that of human emphysema (3, 4).

Common biological substrates of Br₂ include unsaturated fatty acids and various conjugated metabolites, including free aromatic amino acids and amino acids found within the backbone of peptides and proteins (i.e., tryptophan, phenylalanine, histidine, and tyrosine; Refs. 18, 39, 42, 53, 57). Br₂ and its hydrolysis product, HOBr, are likely to react with targets on airway epithelia and, therefore, unlikely to reach distal lung spaces. Consequently, injury to alveolar epithelial and systemic organs such as the heart are thought to be mediated by reaction products. One such class of the products includes long-chain brominated fatty acids and aldehydes. These products form primarily by the reaction of Br₂ and HOBr with plasmalogens to create what we will refer to hereon as brominated lipids (Br-lips; Fig. 1; Refs. 3, 5, 23). However, there is limited information on which proteins and biological processes are affected by Br₂, HOBr, and Br-lips. Identification of underlying pathological mechanisms may lead to the development of novel targeted countermeasures, and investigation of global proteomic changes offers significant insight.

In our first series of experiments, we exposed C57BL/6 mice to bromine regimens (600 ppm for 30 min) in environmental chambers and then returned the animals to ambient air for 24 h. Through the application of a common proteomics workflow combined with systems biology analysis (Fig. 2), we identified robust and highly significant changes in three key processes: 1) exosome secretion, 2) inflammation, and 3) vascular permeability. To gain additional insight into the latter process, we exposed human lung microvascular endothelial cells (HMECs) to Br-lips (the most likely mediator of these changes; Refs. 3, 5, 23) and performed similar global discovery proteomics studies. We then validated our findings by performing a number of ex vivo and in vitro measurements to assess functional changes in vascular permeability of isolated perfused mouse lungs [Br₂-treated (Tx)] while also identifying and assessing key protein changes responsible for vascular permeability disruption across HMEC monolayers (Br₂ Tx and Br-lips). To the best of our knowledge, this is the first report in which global proteomics and systems analysis approaches have been combined with physiological measurements to investigate halogen-induced acute lung injury.

Fig. 2.

Schematic of workflow applied to the global discovery proteomics experiments. Proteins were extracted from the lungs of C57BL/6 mice 24 h after exposure to Br₂ (600 ppm for 30 min) or human lung endothelial microvascular (HLME) cells (100 ppm for 10 min) after a 4-h incubation with 2-bromopalmitic acid and 2-bromopalmitaldehyde and analyzed separately through the given workflow for the purpose of identifying novel biological processes and potential drug targets. The more detailed steps for this workflow are described in materials and methods. 1D, 1-dimensional; LCMS2, liquid chromatography-tandem mass spectrometry; MW, molecular weight.

MATERIALS AND METHODS

Materials

The chemical reagents dithiothreitol (cat. no. D9779) and iodoacetamide (cat. no. I1149) were purchased from Sigma-Aldrich (St. Louis, MO). Acetonitrile (ACN) was purchased from Thermo Fisher Scientific (cat. no. A996SK; St. Louis, MO). The generation of 2-bromopalmitaldehyde (2-BrPALD) and 2-bromopalmitic acid (2-BrPA) along with their stable heavy-labeled standards for absolute quantification by mass spectrometry were carried out using strategies previously employed for the synthesis of chlorinated aldehydes (54, 60). All other disposables are referenced within the manuscript.

Animals

C57BL/6 8- to 12-wk-old male (20–25 g body wt) mice were purchased from Charles River (Wilmington, MA). All experimental procedures involving animals were approved by the University of Alabama at Birmingham (UAB) Institutional Animal Care and Use Committee (IACUC; APN20950).

Mice were exposed to Br₂ gas (600 ppm/30 min) in a cylindrical glass chamber, two at a time, located in an environmental hood inside a negative pressure room, as previously described (4). Br₂ tanks (mixed with air at the desired concentrations) and compressed air (used for control experiments) were purchased from Airgas (Birmingham, AL), which certified the Br₂ concentration to within 2% of its nominal value. The concentration of Br₂ into the chamber was measured randomly, at least once per week, using an Interscan detector (model RM34-1000m; Simi Valley, CA). The gas flow rate (5 L/min) was chosen to allow the chamber Br₂ concentration to reach its nominal value within a few minutes and maintain the chamber CO₂ concentration <0.05%, as measured with a CO₂ meter (https://www.co2meter.com/). Chamber humidity was at 40–50% throughout the exposure.

Following exposure, Br₂ flow was discontinued, and mice were returned to their cages and observed during the day at least every 2 h and once during the night. They were allowed access to food and water ad libitum. For the survival studies, 24 males and 15 females were observed closely for 3 wk (21 days) to determine the level of mortality between each group.

Discovery Proteomics (Animals and HMECs)

Sample preparation of mouse lung tissues (n = 4) and HMECs (n = 3).

Culture conditions are described below (section Immunotargeted and Global Proteomic Measurements in HMECs Tx Br₂ and Br-Lips). For the animal experiments described above (section Animals), at 24 h after exposure, mice were euthanized according to UAB IACUC protocols, and mouse lungs were harvested, flash-frozen in liquid nitrogen, and stored at −80°C until used for proteomics analysis. The global proteomics analysis was carried out as previously referenced with minor changes (Fig. 2; Ref. 41). The protein fractions were quantified using the Pierce BCA Protein Assay Kit (cat. no. PI23225; Thermo Fisher Scientific), and ~20 µg of protein per sample was then diluted to 35 µL using NuPAGE LDS Sample Buffer (1× final concentration; cat. no. NP0007; Invitrogen). Proteins were then reduced with dichlorodiphenyltrichloroethane and denatured at 70°C for 10 min before loading onto Novex NuPAGE 10% Bis-Tris Protein Gels (cat. no. NP0315BOX; Invitrogen) and separated approximately halfway (15 min at 200-V constant). The gels were stained overnight with Novex Colloidal Blue Staining Kit (cat. no. LC6025; Invitrogen). Following destaining, each lane was partitioned into three separate molecular weight (MW) fractions and equilibrated in 100 mM ammonium bicarbonate. Each gel plug was then digested overnight with Trypsin Gold, Mass Spectrometry Grade (cat. no. V5280; Promega) following manufacturer’s instructions. Peptide extracts were reconstituted in 0.1% formic acid/double-distilled water (ddH₂O) at 0.1 µg/µL.

Mass spectrometry: nano-HPLC electrospray ionization multistage tandem mass spectrometry analysis and database searches.

Peptide digests (8 µL each) were injected onto a 1260 Infinity nano-HPLC stack (Agilent Technologies, Santa Clara, CA) and separated using a 100-µm inside diameter-by-13.5-cm pulled tip C18 column (Jupiter 5 µm C18 300 Å; Phenomenex, Torrance, CA). This system runs in line with a Thermo Scientific Orbitrap Velos Pro hybrid mass spectrometer equipped with a nanoelectrospray source (Thermo Fisher Scientific, Waltham, MA), and all data were collected in collision-induced dissociation mode. The nano-HPLC was configured with binary mobile phases that included solvent A (0.1% formic acid in ddH₂O) and solvent B (0.1% formic acid in 15% ddH₂O-85% ACN), programmed as follows: 10 min at 5% solvent B (2 µL/min, load), 90 min at 5–40% solvent B (linear: 0.5 nL/min, analyze), 5 min at 70% solvent B (2 µL/min, wash), 10 min at 0% solvent B (2 µL/min, equilibrate). Following each parent ion scan (300–1,200 mass-to-charge ratio at 60,000 resolution), fragmentation data (tandem mass spectrometry, MS2) were collected on the top-most-intense 15 ions. For data-dependent scans, charge state screening and dynamic exclusion were enabled with a repeat count of 2, a repeat duration of 30 s, and an exclusion duration of 90 s.

The Xcalibur *.RAW files were collected in profile mode, centroided, and converted to mzXML using ReAdW version 3.5.1 software (Seattle Proteome Center, Institute for Systems Biology; http://tools.proteomecenter.org/wiki/index.php?title=Formats:mzXML). The data were searched using SEQUEST (20), which was set for two maximum missed cleavages, a precursor mass window of 20 ppm, trypsin digestion, and variable modifications C at 57.0293 and M at 15.9949. Searches were carried out using species-specific subsets [Mus musculus (mouse) and Homo sapiens (human)] of the UniProtKB database (https://www.uniprot.org/).

Peptide filtering, grouping, and quantification.

The list of peptide identifications generated based on SEQUEST (Thermo Fisher Scientific) search results were filtered using Scaffold (Protein Sciences, Portland, OR). Scaffold filters and groups all peptides to generate and retain only high-confidence identifiers while also generating normalized spectral counts (NSCs) across all samples for the purpose of relative quantification. The filter cutoff values were set with minimum peptide length of more than five amino acids, with no single charged peptides (MH⁺1 charge states), with peptide probabilities of >80% confidence interval, and with the number of peptides per protein at least two. The protein probabilities were then set to a >99.0% confidence interval and a false discovery rate (FDR) <1.0. Scaffold incorporates the two most common methods for statistical validation of large proteome data sets, the FDR and protein probability (33, 46, 59). Relative quantification across experiments was then performed via spectral counting (40, 50), and when relevant, spectral count abundances were then normalized between samples (9).

Statistical and systems biology analysis.

Two separate nonparametric statistical analyses were performed for each pairwise comparison of all proteomic data, including 1) the calculation of weight values by significance analysis of microarray (SAM; cutoff >|0.6|) combined with 2) t test (single tail, unequal variance, cutoff of P < 0.05). Data were then sorted according to the highest statistical relevance in each comparison. SAM (24, 61) is a statistically derived function that approaches significance as the distance between the means (μ₁ − μ₂) for each group increases and the standard deviation (δ₁ − δ₂) decreases using the formula W = (μ₁ − μ₂)/(δ₁ − δ₂). For protein abundance ratios determined with NSCs, we set a 1.5- to 2.0-fold change as the threshold for significance. This significance threshold was determined empirically by analyzing the interquartile data from the control experiment indicated above using ln − ln plots, where the Pearson correlation coefficient (R) was 0.98, and >99% of the normalized intensities fell between ±1.5-fold. In each case, any two of the three tests (SAM, t test, or fold change) had to pass. The most significant proteomic changes associated with permeability were graphed with GraphPad (Prism 8.0; San Diego, CA), whereas the volcano plot, heat map, and principal component analysis were all generated using Qlucore Omics Explorer (Lund, Sweden). Gene ontology assignments and pathway analysis were carried out using MetaCore (GeneGo, St. Joseph, MI), UniProtKB database, and Database for Annotation, Visualization and Integrated Discovery (DAVID; https://david.ncifcrf.gov/). Interactions identified within MetaCore are manually correlated using full text articles. Detailed algorithms have been described previously (11, 19).

Biological Measurements of Brominated Lipids

LC-MS and GC-MS quantitative methods for the detection of 2-BrPA and 2-BrPALD in mouse lungs have been developed by D. A. Ford (n = 6–12 per group; Refs. 5, 58). Tandem mass spectrometric analysis of 2-BrPA revealed that these molecules fragment, losing HBr, which is similar to the fragmentation of 2-chloropalmitic acid, which loses HCl (6, 14, 60). This fragmentation was used for selected reaction monitoring (SRM) detection of 2-BrPA resolved by LC; 2-BrPA was quantified by comparison with the SRM detection of its stable labeled isotope, used as internal standard. The ratio of the ion intensity response of the natural 2-BrPA to that of the internal standard was linear over three orders of magnitude. Similarly, using GC-MS with negative-ion chemical ionization (NICI) detection, the ratio of the ion intensity response for 2-Br-[d₄]-PALD to 2-BrPALD following derivatization to their pentafluorobenzyl (PFB) oximes is linear over three orders of magnitude. This method has previously been used to quantify 2-BrPALD (5). C57BL/6 mice (n ≥ 6) were exposed to Br₂ and euthanized, and their lungs were removed and flash-frozen as indicated for the proteomics experiments. For these experiments, the specimens were packed in dry ice and sent overnight to D. A. Ford’s laboratory for analysis. Lipids were extracted by the method of Bligh and Dyer (12). Esterified and nonesterified 2-BrPA were prepared in samples that were either subjected or not subjected to base hydrolysis. 2-BrPA was then quantified by LC-MS using SRM detection under column and mobile phase conditions similar to that employed for assessing 2-chlorofatty acid levels (23). Additionally, lung lipid extracts were used to derive the PFB oxime of 2-BrPALD and other molecular species of 2-BrFALD, which were then quantified by NICI-GC-MS using 2-Br-[d₄]-PALD internal standard (5).

Functional Studies in Br₂ Tx Mouse Lungs and Br-Lips Tx HMECs

Measurements of lung filtration coefficient in isolated perfused lungs (n = 6).

Measurements of the K_f were performed as described previously in detail (8). In brief, mice were anesthetized with an intraperitoneal injection of pentobarbital (50 mg/kg) to achieve an adequate plane of anesthesia. Their tracheas were cannulated and connected to a mouse mechanical ventilator (Harvard Apparatus, South Natick, MA) and ventilated with 95% O₂-5% CO₂. Ventilation was maintained with a tidal volume of 0.2 mL (10 mL/kg), a respiratory rate of 40–60 breaths/min (ventilation of 8–12 mL/min), and 2 cmH₂O of positive end-expiratory pressure throughout the experiment. Heparinization was achieved with intraperitoneal injection of 50 units of heparin immediately before tracheal cannulation to allow adequate time for systemic absorption. Next, the peritoneum was entered, the diaphragm was carefully dissected, and the thoracic cavity was exposed via a high median sternotomy. The pulmonary artery and the left atrium were cannulated sequentially via incisions in the right and left ventricles, respectively. The heart and lungs were then removed en bloc and perfused with Earle’s balanced salt solution with NaHCO₃ (CAS no. 90604-29-8; Sigma Life Science) and 4% fat-free (<0.2% fat) bovine serum albumin (Albumin Fraction V; Sigma Life Science) with a measured pH between 7.35 and 7.40 ensured before lung perfusion. The perfusate was maintained at 37°C. At time of isolation, the lungs were flushed with a sufficient volume of perfusate until clear of blood. The total recirculating perfusate volume was 30 mL.

Lungs were perfused initially with a pulmonary venous pressure of 15 cmH₂O maintained for 15 min and then decreased to 5 cmH₂O for 15 min with flow adjusted to maintain no net change in weight for ≥10 min before measurements. Pulmonary artery and left atrial pressures were transduced with continuous monitoring using ADInstruments PowerLab 8/30 and analyzed using LabChart 8 (ADInstruments, Colorado Springs, CO). Capillary pressures were measured every 15 min using the double occlusion technique (31, 56). The K_f was calculated as the linear rate of weight gain over 15 min after a 15-cmH₂O increase in pulmonary venous pressure. Values were normalized per gram of lung dry weight, measured at the end of the experiment by placing the lungs in an oven at 80°C for 48 h.

Measurements of transendothelial resistance across HMECs.

Human lung microvascular endothelial cell (HMEC) transendothelial electrical resistance (TER), a measure of endothelial permeability to lipid-insoluble molecules, was measured using an Electric Cell-substrate Impedance Sensing (ECIS) system (Applied Biophysics, Troy, NY) as described previously in detail (31). Briefly, HMECs (1 × 10⁵ cells/mL; passages 5–15; n = 6) were plated onto 8W10E arrays in normal culture medium [500 mL of DMEM, 50 mL of FBS, 5 mL of Antibiotic Antimycotic Solution (100×), Stabilized, Sigma cat. no. A5955, containing penicillin: 50,000 U/500 mL DMEM, streptomycin: 50 mg/500 mL DMEM, amphotericin B: 12 µg/500 mL DMEM] until they reached resistances of ~900 Ω, usually 2–3 days after seeding, at which time they were considered confluent. Resistance values were recorded every 15 min for the duration of the experiments. At that time, stock solutions of 16C brominated lipids (2-BrPA and 2-BrPALD; 3 mM each) were diluted to final concentrations of 100 µM in normal culture medium, and 50 µL was added into the well arrays for a final concentration of 10 µM each. Similar amounts of palmitic acid (1-hexadecanoic acid; CAS no. 57-10-3; MilliporeSigma cat. no. P0500) and palmitaldehyde (1-hexadecanalaldehyde; CAS no. 629-80-1; Sigma-Aldrich product identifier SY3H6E41694C, Synthonix cat. no. P67608) were added to control wells. Resistance values were measured continuously for the next 3 days.

Immunofluorescence, Western Blots, and Global Proteomic Measurements in HMECs

Immunofluorescence staining.

HMECs (n = 5) were cultured with DMEM, 10% FBS, and 1% antibiotics on 12-mm round coverslips in a 24-well plate until they became confluent as determined by light microscopy examination. For the Br₂-treated cells, cells were exposed to Br₂ at 100 ppm for 10 min and returned in an incubator vented with 95% air-5% CO₂ for 6 h, as previously described for Cl₂ (36). For the Br-lips treatments, cells were incubated with vehicle, 2-BrPA (10 µM), and 2-BrPALD (10 µM) or their corresponding nonbrominated compounds for 4 h. Stock solutions were diluted in normal culture medium as indicated above (under the TER measurements), which was added directly to the cells for 24 h, and cells were then immediately washed with PBS, fixed in 4% paraformaldehyde (10 min), and permeabilized in 0.3% Triton X-100 in PBS for 10 min. Cells were incubated for 30 min with antibodies against F-actin labeled with Alexa Fluor 488 (1:1,000; cat. no. A12379; Molecular Probes, Thermo Fisher Scientific, Rockford, IL) and phosphorylated VE-cadherin against the Tyr658 residue labeled with Alexa Fluor 555 (1:1,000; cat. no. ABT1760-AF555; Thermo Fisher Scientific). After washing thrice with PBS, nuclei were counterstained with DAPI (Thermo Fisher Scientific). Cells were then mounted with ProLong Gold and imaged by using a Zeiss fluorescent microscope (Carl Zeiss USA, San Diego, CA).

Measurement of RhoA and ROCK2 activities.

The Rho activity was measured and quantified by using the RhoA Activation Assay Biochem Kit (cat. no. BK036; Cytoskeleton) based on the Rhotekin pull-down assay as per the manufacturer’s instructions. In brief, HMECs (n = 6) were cultured and treated with 10 µM each palmitic acid and palmitaldehyde or 10 µM each 2-BrPA and 2-BrPALD for 15–30 min. The cell lysate was prepared in radioimmunoprecipitation assay (RIPA) buffer containing protease inhibitors (cat. no. 78425; Thermo Fisher Scientific) immediately after treatment and protein estimation was done, and equal amounts of protein were incubated with Rhotekin-RBD beads (cat. no. RT02) for 1 h at 4°C. After the beads were washed with wash buffer, proteins were removed from the beads with Laemmli buffer and then subjected to Western blotting.

Measurement of ROCK2 phosphorylation.

In brief, HMECs (n = 6) were cultured in DMEM and treated with 10 µM each palmitic acid and palmitaldehyde or 10 µM each 2-BrPA and 2-BrPALD for 15–30 min. The cell lysates were prepared in RIPA buffer containing protease inhibitors (Thermo Fisher Scientific, Rockford, IL) immediately after treatment. Protein estimation was carried out using the BCA assay. Equal amounts of protein in Laemmli buffer were loaded in 10% Tris·HCl Criterion precast gels, and proteins were transferred to PVDF membranes. Membranes were probed with ROCK2 (1:1,000 dilution; cat. no. 8236; Cell Signaling Technology, Beverly, MA) or pTyr⁷²²-ROCK2 antibody (cat. no. SAB4301564; MilliporeSigma). Bands were detected by the chemiluminescent horseradish peroxidase substrate. Protein loading was normalized by reprobing the membranes with an antibody specific to β-actin.

Proteomics analysis of HMEC.

In the last series of experiments, we treated HMECs (n = 3) with 2-BrPA and 2-BrPALD (10 µM each) for 4 h and performed the same discovery global proteomic analysis from sample preparation to systems analysis as described above [section Discovery Proteomics (Animals and HMECs)].

Statistical analysis.

The details for the statistical validation of proteomics data are shown in the pertinent sections above [Discovery Proteomics (Animals and HMECs)]. For all other studies, data are presented as means ± standard error of the mean. Statistical analysis among means was performed with analysis of variance (ANOVA; 1- or 2-way) followed by Bonferroni post hoc comparisons. Data were considered significant if P < 0.05. Data were graphed using GraphPad Prism 7 for Windows.

RESULTS

Survival (Br₂-Exposed Mice)

None of the animals exposed to Br₂ (600 ppm in 30 min) died during the exposure period. However, mice exhibited signs of respiratory distress, including labored breathing and flaring of the nostrils during inspiration. Respiratory rates also decreased acutely during exposure (data not shown). Approximately 50% of the mice died by 8 days after exposure, in agreement with our previous reports (Fig. 3; Refs. 32, 35). No statistically significant differences in survival were observed among male and female mice. Thus we opted to use male mice for all experiments reported in this paper to avoid variabilities associated with the estrous cycle.

Fig. 3.

Kaplan–Meier survival curves after Br₂ exposure. Adult male and female C57BL/6 mice, 8–10 wk old, were exposed to Br₂ gas (600 ppm for 30 min) in environmental chambers and returned to room air as detailed in materials and methods. Mice were considered dead when they lost >30% of their initial body weight (University of Alabama at Birmingham Institutional Animal Care and Use Committee guidelines) or stopped breathing for >5 min. In the case that they were found dead, survival was calculated as the mean time between the last observation and the time of discovery. The cross-correlate P value (P = 0.554) was calculated at ~0.6 across all time points, indicating that there was no significant difference in terms of pathology between the 2 groups.

Global Proteomics and Systems Biology Analysis (After Br₂-Exposed Mice)

For the global proteomics workflow carried out on lung tissues, we applied a standard 1D-PAGE-LC-MS2 approach, whereby the proteins extracted from an entire lung specimen for each animal were loaded onto a 1-dimensional (1D) denaturing gel and separated by MW halfway. The entire lane for each sample was cut into 3 equally sized MW fractions and digested with trypsin, and the resultant peptides for each fraction were then separated using nano-HPLC and analyzed in line with a high-resolution MS2 mass spectrometer (Fig. 2). This approach yielded 1,623 proteins with >99% confidence and <1% FDR that were present in the lungs of both the experimental (24 h after Br₂ Tx) and control (24 h after air) groups (Supplemental Table S1; all supplemental material is available at https://doi.org/10.6084/m9.figshare.12229463). Quantification was carried out using normalized spectral counting (NSCs), and the applied pairwise statistical analysis (as described in materials and methods) yielded 95 differentially abundant proteins (Supplemental Table S2), with 34 increased (Table 1) and 61 (Table 2) decreased in Br₂-exposed versus air controls. These data were plotted to better visualize the entire data set while highlighting the most significantly changed proteins (Fig. 4, A and B; Supplemental Tables S1 and S2). The volcano plot (Fig. 4B) illustrates the distribution of all data points derived from the entire data set, with upper limits (above the line) indicating statistically significant changes and outer limits (to the right and left of each vertical line) indicating significant fold changes as outlined in materials and methods under statistics. Please note that although fold change is visualized as log fold (log₂), with a cutoff value of ±1.5 applied to fold changes before logging the data, this value equates to the indicated ±0.6 log₂ limits illustrated in the figure. Various proteins that play a role in vascular permeability are highlighted within the plot by arrows and network names.

Table 1.

Proteomic changes in mouse lung: Br₂-treated vs. air controls (34 proteins increased)

				Statistics			UniProtKB, DAVID, and GeneGo MetaCore Definitions
UniProtKB Name	UniProt ID	Entrez ID	Network ID	SAM	t-Test P Value	Fold∆ (Br2/Air)	GO Biological Processes
CD177 antigen	Q8R2S8	68891	NB1*‡§	2.96	0.017	8.0	Cell-to-cell junction/adhesion; innate immune response; regulated endocytosis
Cathelin-related antimicrobial peptide	P51437	12796	CAMP‡§	2.11	0.001	4.5	Cytokine response and regulation; regulated exocytosis
Serum paraoxonase/arylesterase 1	P52430	18979	PON1	1.40	0.047	3.3	Aromatic compound catabolic process; positive regulation of binding
Heterogeneous nuclear ribonucleoprotein L	Q8R081	15388	hnRNP L†	2.99	0.002	3.0	Regulation of adhesion molecules; RNA binding and processing
Myeloperoxidase	P11247	17523	PERM‡§	1.06	0.019	2.9	Peroxidase activity; defense response; regulated exocytosis
Tubulin-specific chaperone D	Q8BYA0	108903	TBCD*	4.69	0.006	2.8	Adherens and tight junction assembly; β-tubulin binding
Neutrophil gelatinase-associated lipocalin	P11672	16819	NGAL‡§	1.07	0.044	2.8	Innate immune response; iron homeostasis; regulated exocytosis
KN motif and ankyrin repeat domain-containing protein 2	Q8BX02	235041	ANKRD25*	0.96	0.030	2.6	Regulation of Rho signal transduction; apoptotic process; stress fiber assembly
Squamous cell carcinoma antigen recognized by T cells 3	Q9JLI8	53890	Tip110	0.94	0.048	2.6	Histone and RNA binding; cellular morphogenesis
Lmo7 protein	B7ZN52	380928	LMO7†	0.90	0.022	2.6	Regulation of cell adhesion; adherens junction; actinin binding
60S ribosomal protein L34	Q9D1R9	68436	RPL34	0.89	0.035	2.5	Translation
40S ribosomal protein SA	P14206	16785	LAMR1*	0.89	0.041	2.4	Cell-to-cell adhesion; differentiation
1,4-alpha-glucan-branching enzyme	Q9D6Y9	74185	GLGB	2.15	0.011	2.4	Carbohydrate metabolic process
60S ribosomal protein L26	B1ARA5	19941	RPL26	1.34	0.005	2.3	Translation
Protein S100A9	P31725	20202	Calgranulin B†/S100‡§	1.46	0.014	2.3	Actin cytoskeleton reorganization; innate immune response; regulated exocytosis
Endoplasmic reticulum resident protein 29	P57759	67397	ERp29	0.95	0.034	2.2	Regulation of gene and protein expression; protein secretion and phosphorylation
Estradiol 17-β-dehydrogenase 8	P50171	14979	HSD17B8	1.72	0.007	2.2	Estrogen and androgen metabolic process
Myo1b	Q7TQD7	17912	MYO1B*‡	1.63	0.012	2.2	Actin binding and filament bundle assembly; vesicle-mediated transport
Guanine nucleotide-binding protein subunit alpha-13	P27601	14674	G-protein alpha-12 family†	1.33	0.028	2.2	Regulation of cell shape and migration; Rho signal transduction; G protein receptor signaling
60S ribosomal protein L36	P47964	54217	RPL36	1.02	0.026	2.1	Ribosome; synapse
Actin-related protein 2/3 complex subunit 2	Q9CVB6	76709	ARPC2*§	0.89	0.034	2.1	Actin cytoskeleton reorganization; immune response-associated
60S ribosomal protein L22	P67984	19934	RPL22*	0.90	0.037	2.1	Focal adhesion component; heparin binding; glutamatergic synapse
Sorbin and SH3 domain-containing protein 1	Q62417	20411	SORBS1*	0.78	0.036	2.0	Focal adhesion assembly; insulin receptor signaling
Coronin-1C	Q9WUM4	23790	CORO1C*	1.07	0.038	1.9	Actin cytoskeleton organization; endosomal transport
Alpha-2-HS-glycoprotein	P29699	11625	Fetuin-A‡§	0.79	0.048	1.9	Acute phase and inflammatory response; regulated exocytosis
Succinate dehydrogenase [ubiquinone] flavoprotein subunit, mitochondrial	Q8K2B3	66945	SDHA	0.92	0.033	1.8	Nervous system development; succinate metabolic process
Collagen alpha-1(XIV) chain	Q80X19	12818	Collagen XIV*	0.71	0.048	1.8	Cell adhesion; collagen fibril organization
A-kinase anchor protein 2	O54931	11641	AKAP2	1.00	0.050	1.8	Mediate multiprotein complex protein kinase A
Electron transfer flavoprotein subunit β	Q9DCW4	110826	ETFB	0.78	0.039	1.7	Fatty acid β-oxidation
Adiponectin	Q60994	11450	Adiponectin	0.98	0.040	1.6	Response to ox-stress, cAMP, epinephrine, and insulin; glucose homeostasis, etc.
Hsp90 co-chaperone Cdc37	Q61081	12539	CDC37§	0.98	0.040	1.6	Regulation interferon-γ signaling; protein folding and stabilization
Alpha-crystallin B chain	P23927	12955	Alpha crystallin B	0.91	0.045	1.6	Response and regulation of apoptosis and ox-stress
Haloacid dehalogenase-like hydrolase domain-containing protein 2	Q3UGR5	76987	HDHD2	5.20	0.000	1.5	Phosphatase activity; metal ion binding
Isoaspartyl peptidase/L-asparaginase	Q8C0M9	66514	Asrgl1	5.20	0.000	1.5	Asparagine catabolic process

All proteins listed were identified as significantly changed in Br₂-treated vs. air control-treated animal lungs. This is from the top-95 protein list (Supplemental Table S2) with fold changes (∆) of at least +1.5. All proteins have been gene ontology (GO)-annotated using a combination of UniProtKB, Database for Annotation, Visualization and Integrated Discovery (DAVID), and GeneGo MetaCore definitions. ID, identifier; ox, oxidative; SAM, significance analysis of microarray.

^*Permeability-associated (Fig. 7),

^†permeability-associated and highlighted in volcano plot (Fig. 4B),

^‡regulated exocytosis-associated,

^§immune response-associated.

Table 2.

Proteomic changes in mouse lung: Br₂-treated vs. air controls (61 proteins decreased)

				Statistics			UniProtKB, DAVID, and GeneGo MetaCore Definitions
UniProtKB Name	UniProt ID	Entrez ID	Network ID	SAM	t-Test P Value	Fold∆ (Br2/Air)	GO Biological Processes
RAB14 protein variant	Q50HX3	68365	Rab-14‡§	1.35	0.006	−1.5	Regulated exocytosis; immune response-associated
CDGSH iron-sulfur domain-containing protein 1	Q91WS0	52637	CISD1	1.07	0.019	−1.5	Regulation of cellular respiration; iron ion binding
E3 ubiquitin-protein ligase HUWE1	Q7TMY8	59026	HUWE1‡§	1.05	0.020	−1.5	Base excision repair; exocytosis; immune response-associated
Cullin-associated NEDD8-dissociated protein 1	Q6ZQ38	71902	TIP120A‡§	0.85	0.042	−1.6	Protein ubiquitination; exocytosis; innate immune response
Peptidyl-prolyl cis-trans isomerase FKBP2	P45878	14227	FKBP2	0.70	0.048	−1.6	Peptidyl-prolyl cis-trans isomerase activity
Heterogeneous nuclear ribonucleoprotein A1	P49312	15382	hnRNP A1	0.82	0.031	−1.6	RNA binding and splicing
Ras-related protein Rab-1A	Q5SW87	19324	Rab-1A	1.04	0.038	−1.6	GTPase binding and activity
Bleomycin hydrolase	Q8R016	104184	Bleomycin hydrolase	1.93	0.001	−1.7	Response to drug and toxic substance; peptidase activity
Alcohol dehydrogenase [NADP(+)]	Q9JII6	58810	ALDX	1.01	0.019	−1.7	Ascorbic acid biosynthetic process; detoxification of aldehyde
Niban-like protein 1	Q8R1F1	227737	FAM129B*	0.74	0.046	−1.7	Negative regulation of apoptosis; regulation of cell-to-cell adhesion; adherens junction
Ras suppressor protein 1	Q01730	20163	RSU-1*	1.66	0.002	−1.7	Positive regulation of cell-substrate adhesion; Ras signal transduction
JNK-interacting leucine zipper protein long form	B8X349	70834	SPAG9	1.15	0.016	−1.7	Mitogen-activated protein kinase scaffold activity
Stathmin	P54227	16765	Stathmin*	0.94	0.036	−1.7	Tubulin binding; regulation of cytoskeleton organization; Rho signal transduction
Coatomer subunit delta	Q5XJY5	213827	COPD‡	0.89	0.024	−1.7	Golgi localization and vesicle transport; Golgi apparatus
Heat shock protein HSP 90-beta	P11499	15516	HSP90‡§	0.79	0.047	−1.8	Regulation of apoptosis, protein kinase B signaling, exocytosis and immune response, etc.
Tetratricopeptide repeat protein 38	A3KMP2	239570	TTC38	0.83	0.042	−1.8	Structural motif; protein-protein interaction
Selenium-binding protein 1	P17563	20341	SELENBP1	0.95	0.038	−1.8	Selenium binding; methane thiol oxidase; protein transport
ATP-citrate synthase	Q91V92	104112	ACLY‡§	0.83	0.036	−1.8	Citrate and oxaloacetate metabolic process; immune response-associated; exocytosis
ATP-dependent RNA helicase A	O70133	13211	DDX9§	0.79	0.035	−1.9	Response to stress; innate immune response; regulation of cytokines
Dimethylaniline monooxygenase [N-oxide-forming] 2	Q8K2I3	55990	FMO2	1.15	0.016	−1.9	Xenobiotic, toxin, drug, and oxygen metabolic process
Cap-specific mRNA (nucleoside-2'-O-)-methyltransferase 1	Q9DBC3	74157	CMTR1	1.29	0.023	−1.9	mRNA methylation
STE20-like serine/threonine-protein kinase	O54988	20874	SLK*	1.07	0.021	−1.9	Apoptotic process; regulation of focal adhesion assembly; protein kinase signaling
Protein Sec24c	G3X972	218811	Sec24‡	1.00	0.022	−1.9	Golgi vesicle-mediated transport; enables zinc ion and SNARE binding
Hypoxia upregulated protein 1	Q9JKR6	12282	HYOU1*‡	1.09	0.014	−1.9	Regulate vascular cell adhesion molecule-1; response to hypoxia and stress; exocytosis
Proteasome-associated protein ECM29 homolog	Q6PDI5	230249	ECM29	0.96	0.022	−1.9	Proteasome assembly; ubiquitin ERAD pathway
Junction plakoglobin	Q02257	16480	Plakoglobin*‡§	0.79	0.034	−2.0	Cell-to-cell adhesion and migration; exocytosis; immune response; angiogenesis
Protein Sec24d	Q6NXL1	69608	Sec24‡	1.37	0.031	−2.0	SNARE and zinc ion binding; Golgi vesicle-mediated transport
Cytoplasmic dynein 1 heavy chain 1	Q9JHU4	13424	Dynein 1, heavy chain‡§	1.24	0.020	−2.0	Golgi vesicle-mediated transport; immune response-associated; stress granule assembly
AP-2 complex subunit beta	Q9DBG3	71770	Beta-adaptin 2	1.34	0.005	−2.0	Clathrin binding and coat assembly; regulation of neuronal death; regulation of endocytosis
Coatomer subunit beta'	O55029	50797	COPB2	0.89	0.023	−2.0	Intracellular Golgi vesicle-mediated transport; intracellular protein and toxin transport
Glutathione S-transferase Mu 1	P10649	14862	GSTM1	0.88	0.028	−2.0	Response to drug; glutathione metabolic process; glutathione transferase activity
Transforming growth factor beta-1-induced transcript 1	Q62219	21804	HIC5†	1.57	0.006	−2.1	Endothelial cell migration; Wnt and Tgfb signaling; response to stress
Eukaryotic translation initiation factor 5	P59325	217869	eIF5	1.15	0.034	−2.1	Activation of GTPase activity; translation activity
Phosphofurin acidic cluster sorting protein 1	Q8K212	107975	PACS-1	1.04	0.043	−2.1	Ion channel binding; COPI-coated vesicle
Deoxynucleoside triphosphate triphosphohydrolase	Q60710	56045	SAMHD1§	0.73	0.044	−2.1	Innate immune response; response to DNA damage
Long-chain specific acyl-CoA dehydrogenase, mitochondrial	P51174	11363	ACADL	0.84	0.033	−2.1	Regulation of fatty acid biosynthesis; temperature homeostasis; lipid catabolic process
Regulator of nonsense transcripts 1	Q9EPU0	19704	RENT1	1.07	0.023	−2.1	DNA replication and repair; regulation of telomere maintenance
Nuclear pore complex protein Nup214	Q80U93	227720	NUP214*	0.92	0.050	−2.1	Regulation of nuclear envelope permeability; regulation of cell cycle
Alpha-glucosidase 2 alpha neutral subunit	A1A4T2	14376	GANAB	1.26	0.014	−2.2	Carbohydrate metabolic process; N-glycan processing
Eukaryotic translation initiation factor 4B	Q8BGD9	75705	eIF4B	1.50	0.017	−2.2	Neuronal postsynapse; translation initiation
d-dopachrome decarboxylase	O35215	13202	DDT§	1.00	0.036	−2.2	Regulation of ERK1 and ERK2 cascade; regulation of inflammatory process
SEC14-like protein 4	Q8R0F9	103655	SEC14L4 (TAP3)	1.11	0.020	−2.3	Lipid binding
182 kDa tankyrase-1-binding protein	P58871	228140	TAB182*	1.05	0.012	−2.4	Double-strand break repair; adherens junction; regulation of protein phosphorylation
Programmed cell death 6-interacting protein	Q9WU78	18571	Alix*‡	1.13	0.010	−2.6	Regulation of membrane permeability; apoptotic process; exosome assembly and secretion
1-Phosphatidylinositol 4,5-bisphosphate phosphodiesterase gamma-1	Q62077	18803	PLC-gamma†§	1.55	0.007	−2.6	Regulation of endothelial cell migration and apoptosis; T cell receptor signaling
Proteasome activator complex subunit 1	P97371	19186	PSME1§	1.40	0.034	−2.6	Regulation of proteasomal protein catabolic process; antigen processing and presenting
E3 ubiquitin-protein ligase UBR4	A2AN08	69116	UBR4‡	1.41	0.006	−2.7	Ubiquitin-dependent protein catabolic process; calmodulin binding; regulated exocytosis
GDH/6PGL endoplasmic bifunctional protein	Q8CFX1	100198	G6PE	1.14	0.011	−2.7	Carbohydrate metabolic process; pentose phosphate shunt
Epidermal growth factor receptor substrate 15-like 1	Q60902	13859	EPS15R	0.99	0.042	−2.7	Endocytosis; calcium ion binding
AP-3 complex subunit beta-1	Q9Z1T1	11774	AP-3 beta subunits‡§	2.26	0.001	−2.7	Antigen processing and presentation; vesicle-mediated transport
Signal transducer and activator of transcription 1	P42225	20846	STAT1§	1.01	0.034	−2.7	Cellular response to cytokines; receptor signaling pathway via JAK-STAT, etc.
Dynein light chain 2, cytoplasmic	D6RIN4	68097	DLC2/DYNLL‡§	1.96	0.001	−2.9	Microtubule-based process; regulated exocytosis; immune response-associated
40S ribosomal protein S10	P63325	67097	RPS10†	1.66	0.005	−3.0	Ribosomal small subunit assembly; focal adhesion component
Ribosomal protein S23	Q497E1	66475	RPS23	1.43	0.018	−3.0	Translation
Coagulation factor XIII A chain	Q8BH61	74145	Coagulation factor XIII‡	4.34	0.004	−3.1	Blood coagulation; protein cross-linking; regulated exocytosis
Uteroglobin	Q06318	22287	Uteroglobin§	0.91	0.047	−3.1	Regulation and response to cytokine; regulation of inflammation; phospholipase A2 inhibition
60S acidic ribosomal protein P2	P99027	67186	RPLP2*	0.93	0.035	−3.2	Translational elongation; focal adhesion component
Importin-7	Q9EPL8	233726	RANBP7§	3.37	0.001	−3.2	Innate immune response; histone, SMAD, and Ran GTPase binding
Ras-related protein Rab-11B	P46638	19326	Rab-11B*‡	0.86	0.047	−3.4	Cell-to-cell adhesion and migration; regulated exocytosis; myosin V binding
Serine/threonine-protein kinase WNK1	P83741	232341	WNK1*§	2.92	0.005	−4.1	Cellular response to Ca ion; negative regulation of cell-to-cell adhesion; regulation of T cell
Four and a half LIM domains 1, isoform CRA_b	A2AEX8	14199	FHL1 (SLIM1)†	1.00	0.033	−4.1	Colocalize with integrins at cell adhesion sites; muscle organ development

All proteins listed were identified as significantly changed in Br₂-treated vs. air control-treated animal lungs. This is from the top-95 protein list (Supplemental Table S2) with fold changes (∆) of at least −1.5. All proteins have been gene ontology (GO)-annotated using a combination of UniProtKB, Database for Annotation, Visualization and Integrated Discovery (DAVID), and GeneGo MetaCore definitions. COPI, coat protein complex I; ERAD, endoplasmic reticulum-associated degradation; ID, identifier; SAM, significance analysis of microarray; SNARE, soluble N-ethylmaleimide-sensitive factor attachment protein receptor; Tgfb, transforming growth factor-β.

^*Permeability-associated (Fig. 7),

^†permeability-associated and highlighted in volcano plot (Fig. 4B),

^‡regulated exocytosis-associated,

^§immune response-associated.

Fig. 4.

Global protein changes in air vs. Br₂. Adult male C57BL/6 mice, 8–10 wk old, were exposed to Br₂ gas (600 ppm for 30 min) or air in environmental chambers and returned to room air as detailed in materials and methods. Twenty-four hours later, their lungs were removed and proteins were processed for global proteomics analysis as discussed in materials and methods. A: Venn diagram demonstrating the total number of proteins identified across both groups in addition to those proteins found to be significantly changed following exposure to Br₂ (increased vs. the other group). B: volcano plot of the log₁₀ P value vs. log₂ fold change (Br₂/air) demonstrating the distribution of the entire data set of proteins with upper limits (above the line) indicating statistically significant changes and outer limits (to the right and left of each line) indicating significant fold changes as outlined in materials and methods under statistics. Note that although fold change is visualized as log₂, the cutoff value of ±1.5 was applied to the fold change before logging, thereby yielding the indicated ±0.6 limits. Various proteins that play a role in vascular permeability are identified by the arrows.

For the purposes of qualitatively visualizing the most significant proteome changes, we pulled out those proteins from Tables 1 and 2 presenting with at least a twofold change while also passing both statistical tests (SAM > 0.8, P < 0.05). Two visually confirming qualitative analyses were carried out on the resultant and highly significant 53 proteins as indicated by network name within Fig. 5A using a 2-dimensional hierarchical clustering analysis heat map and a principal component analysis (Fig. 5B). These figures visually verify the statistical significance of the marked differences identified in proteins not only between the 2 groups, but also for each protein across all animals. In each case, a close clustering of proteins of the 2 different groups (24 h after exposure to Br₂ vs. air) is appreciable.

Fig. 5.

Heat map and principal component analysis (PCA) plots for proteins exposed to Br₂ vs. air. Adult male C57BL/6 mice, 8–10 wk old, were exposed to Br₂ gas (600 ppm for 30 min) in environmental chambers and returned to room air as detailed in materials and methods. Twenty-four hours later, their lungs were removed, and proteins were processed for global proteomics analysis as discussed in materials and methods. The top 53 proteins (pulled from Tables 1 and 2) that passed a 2-tiered statistics test with −2.0≤|≥2.0 fold change were used for these analyses. A: the 2-dimensional hierarchical analysis heat map demonstrates which proteins are increased (red) or decreased (blue) in Br₂- vs. air-treated mice. Notice similar behavior for each animal in each group and each protein across groups. No outliers were indicated. B: PCA complements the heat map by using a similar cluster approach that determines which animal (based on protein quantification for all proteins in the top list) is similar across all animals analyzed. Notice tight clustering for air (blue)- and Br₂-exposed mice (yellow) with a clear separation between the 2 groups. The number in parenthesis denotes the animal number for each group analyzed.

Systems biology analysis.

The top 96 significantly differential protein list (Tables 1 and 2) was then utilized for systems biology analysis. This allowed us to identify key gene ontology (GO) annotations listed in the tables in addition to focusing on GO localizations and biological functions associated with Br₂ exposure. Following this analysis, pie charts were generated to better visualize the more common cellular locations that the differential proteins tended to exist for overlapping animal and HMEC studies (Fig. 6A; Supplemental Table S3) in addition to the animal and HMEC overlapping biological processes following Br₂ exposure (Fig. 6B; Supplemental Table S4). The primary overlapping cellular and biological relationships are highlighted within discussion. The cellular localization analysis revealed that the majority of significantly changed proteins are known to associate within seven different compartments [1) focal adhesion/adherens/anchoring/cell junction, 2) extracellular vesicle and extracellular space, 3) ribosomal, 4) endoplasmic reticulum, 5) cytosol, 6) synapse, and 7) ribosome]. Similarly, the biological processes associated with these proteins indicate that nine primary cellular changes are taking place; whereby we are especially interested in 1) vesicle-mediated exocytosis, 2) immune response and neutrophil activation processes, and 3) general transport and secretion-based mechanisms. Because of our specific interest in both “overall” and vascular changes in tissue permeability, we have identified those proteins associated with similar corresponding cellular localized and biological processes. For that purpose, we have identified 30 proteins associated with cell junction, cell-to-cell interactions, and cell adhesion, which are all visualized in a bar graph (Fig. 7).

Fig. 6.

Gene ontology (GO)-annotated cellular localization and biological processes were identified for proteins affected by exposure to Br₂. Adult male C57BL/6 mice, 8–10 wk old, were exposed to Br₂ gas (600 ppm for 30 min) in environmental chambers and returned to room air as detailed in materials and methods. System analysis using the top 95 statistically significant proteins with a Br₂/air fold change of at least ±1.5 (Tables 1 and 2; Supplemental Table S2) allowed us to categorize them according to cellular locations (A) and biological processes (B). The GO annotations can be found in Supplemental Tables S3 and S4. The number of proteins associated with each location or process were summed and normalized to 100 (of note, each protein can be associated with >1 location or process). The resultant pie charts are indicative of the normalized percentage of proteins associated with each category within cellular localizations and biological functions. We have added an asterisk next to all segments that we are particularly interested in for each pie chart; these cellular locations and biological functions are of interest.

Fig. 7.

Relative changes in abundance of proteins associated with vascular permeability. Relative abundance of the 30 proteins [by gene ontology (GO) annotation; Supplemental Tables S3 and S4] within the top-95 list that are known to associate with cell junction and adhesion as described for each protein in Tables 1 and 2 and Supplemental Table S2 is depicted. The protein names were converted to GO identifiers, which are listed along with their UniProtKB names in Tables 1 and 2. Values as generated in GraphPad Prism are presented as means ± standard deviation (n = 4 for each value). Each value in the 24 h after Br₂ (600 ppm/30 min) is statistically different from its corresponding air value (P < 0.05, t test).

Biological Measurements of Brominated Lipids in Mice

Based on previous studies (23), the levels of Br-lips, 2-BrPA, and 2-BrPALD are believed to be major biologically reactive intermediates in Br₂-exposed mice. These intermediates we, therefore, quantified in the lungs of Br₂-treated mice. The 16C adducts were found to exist at significantly high concentrations in lung tissues following Br₂ treatments from 30 min up to 24 h (Fig. 8, A and B). 2-BrPALD also reacts readily with glutathione, which exists in abundant quantities in lung tissue (18); as a result, significant levels of the glutathionylated 2-GS-PALD adducts (Fig. 1) were also observed in lung tissue 4 h after exposure (data not shown).

Fig. 8.

Exposure of mice to Br₂ leads to formation of brominated lipids (Br-lips). Br₂ and Br-lips increase microvascular permeability both in vivo and ex vivo. A and B: mice were exposed to Br₂ (600 ppm for 30 min) and returned to room air for the indicated periods of time. At those times mice were euthanized, their lungs were removed, and both 2-bromopalmitic acid (2-BrPA; A) and 2-bromopalmitaldehyde (2-BrPALD; B) were quantified by mass spectrometry. C and D: each point represents 1 mouse. C: C57BL/6 mice were exposed to Br₂ (600 ppm for 30 min) and returned to room air. At 24 h after exposure, all mice were euthanized, and their lungs were removed. The filtration coefficient (K_f) was measured as described in materials and methods. Data shown are all values (each point corresponds to a different mouse) as well as means ± 1 SE; n = 6 for each group. D: human lung microvascular cells were cultured on Electric Cell-substrate Impedance Sensing (ECIS) plates in DMEM with 10% FBS and 1% antibiotics until they reached confluence (resistance > 800 Ω); they were then incubated with either vehicle alone or a mixture of either 10 µM each PA and PALD or 10 µM 2-BrPA and 2-BrPALD. Transendothelial resistance (TER) was measured every hour for ≤25 h and expressed as the ratio at each time point divided by the control value (values are means ± 1 SE; n = 6 for each group).

Functional Studies in Br₂ Tx Mouse Lung and Br-Lips Tx HMECs

Data illustrated in the mouse global proteomics Br₂ study (above) indicate that 30 of the 95 significantly affected proteins (~32%) are involved in the cell junction, cell-to-cell interactions, and cell adhesion category, many of which are known to contribute to the control of vascular permeability (Fig. 7; Tables 1 and 2; Supplemental Table S2). Thus our next series of experiments aimed to identify whether exposure of mice to Br₂ or HMEC to brominated lipids, respectively, caused increased permeability. We exposed mice to Br₂ (600 ppm for 30 min) and then returned them to ambient air. Twenty-four hours later, we removed their lungs and measured the filtration coefficient (K_f), as described in materials and methods. Our data (Fig. 8C) indicate that there was a threefold increase in K_f at 24 h after Br₂ exposure compared with control (air). K_f values revealed two clusters at 24 h after Br₂ exposure with 50% of the mice demonstrating a twofold increase and the remaining 50% increasing fourfold. K_f values for control mice exposed to air showed little variance.

Vascular permeability changes in human lung microvascular endothelial cells.

To gain additional insight as to Br₂-induced changes in vascular permeability while also identifying the proteins involved, we incubated human microvascular cells with Br-lips, 2-BrPA, and 2-BrPALD at 10 µM each compared with the same concentrations of nonbrominated palmitate and palmitaldehyde. We then measured changes in TER, as illustrated (Fig. 8D). Br-lips, but not the nonbrominated lips, caused a significant decrease of TER, consistently from just a few hours until the final 24-h time point at approximately twofold.

Immunotargeted and Global Proteomic Measurements in HMECs Tx Br₂ and Br-Lips

Proteins associated with vascular permeability are altered following exposure of HMECs to Br-lips.

Pathological changes in F-actin stress fiber formation, internalization of VE-cadherin phosphorylation (Tyr658), and dysregulated/disrupted zona occludens-1 (ZO-1) in paracellular junctions were all observed in Br-lips Tx HMECs, and all are indicators of increased permeability. We illustrated the appearance of F-actin stress fibers (Fig. 9A) and internalization of phosphorylated VE-cadherin (Fig. 9B) along with marked disruption of ZO-1 staining in paracellular junctions of HMECs (Fig. 9C) following incubation with 2-BrPA and 2-BrPALD (10 µM each) compared with vehicle alone or nonbrominated PA and PALD lips for 24 h. Similarly, exposure of these cells to Br₂ (100 ppm) for 10 min also caused significant internalization of phosphorylated VE-cadherin and appearance of F-actin fibers (Fig. 10).

Fig. 9.

Exposure of human lung-derived microvascular endothelial cells with brominated lipids disrupts/activates markers of cell permeability. Human lung microvascular cells were cultured with DMEM, 10% FBS, and 1% antibiotics until confluent as determined by light microscopy examination. They were then incubated with vehicle, 2-bromopalmitic acid (2-BrPA; 10 µM), 2-bromopalmitaldehyde (2-BrPALD; 10 µM), or their corresponding nonbrominated compounds for 24 h. At that time, cells were immunostained with antibodies against F-actin (A; green), phosphorylated (phospho-) VE-cadherin against the Tyr658 residue (B; red) and F-actin (B; green), and zona occludens-1 (ZO-1; C; green). Nuclei were counterstained with DAPI (blue color). Notice the appearance of F-actin stress fibers (top right), the appearance of red color (middle right) indicating the internalization of VE-cadherin, in addition to the disruption of ZO-1 in cells treated with brominated lipids. Characteristic figures are illustrated, which were reproduced ≥5 times with different cells and on 2 different days.

Fig. 10.

Exposure of human lung-derived microvascular endothelial cells with Br₂ disrupts/activates markers of cell permeability. Human lung microvascular cells were cultured with DMEM, 10% FBS, and 1% antibiotics until confluent as determined by light microscopy examination. They were then exposed to Br₂ (media infused up to 100 ppm for 10 min) and returned in an incubator vented with 95% air-5% CO₂. Six hours later, cells were immunostained with antibodies against phosphorylated (phospho-) VE-cadherin (P-Tyr658; red) and F-actin (green). Nuclei were counterstained with DAPI (blue color). Notice the appearance of F-actin stress fibers, in addition to the appearance of red color, indicating internalization of phospho-VE-cadherin in cells treated with Br₂. These are characteristic figures, which were reproduced ≥5 times with different cells over 2 different days.

Incubation of HMECs to Br-lips upregulates and activates RhoA/phosphorylated RhoA and ROCK2/phosphorylated ROCK2.

To gain additional understanding on the mechanisms responsible for the increase in vascular permeability, we incubated HMECs with Br-lipids as discussed above and then measured activation of RhoA and ROCK2. As illustrated in Fig. 11, Br-lips induced phosphorylation (activation) of both RhoA (Fig. 11, A and B) and ROCK2 (Fig. 11, C and D), but not ROCK1 (data not shown), within 30 min after incubation.

Fig. 11.

Incubation of human lung-derived microvascular endothelial cells with brominated lipids activates RhoA and ROCK2. Human lung microvascular cells were cultured with DMEM, 10% FBS, and 1% antibiotics until confluent as determined by light microscopy examination. They were then incubated with vehicle, 2-bromopalmitic acid (2-BrPA; 10 µM), and 2-bromopalmitaldehyde (2-BrPALD; 10 µM) or their corresponding nonbrominated compounds for 30 min. At that time, cell lysates were prepared as indicated in materials and methods; for RhoA, equal amounts of protein were immunoenriched, washed, and subjected to Western blotting following 1-dimensional PAGE immediately on return to room air. RhoA activity and protein levels were measured as mentioned in materials and methods. Then, the ratio of active RhoA to total RhoA for after treatment was divided by the corresponding air control ratio for the same experiment (fold increase). A: characteristic gel showing phosphorylated (p-) RhoA and total RhoA. B: graph of ratios, active p-RhoA to total RhoA. Individual values as well as means ± 1 SE are shown. C: characteristic Western blot illustrating p-ROCK2 and total ROCK2. D: graph of ratios, active p-ROCK2 to total ROCK2; values were expressed as fold increase compared with untreated controls. Individual points and means ± 1 SE are shown; analysis of variance was followed by Tukey test.

HMEC (Br-lips) proteomics, Western blot, and immunofluorescence combined results for pathway analysis.

Human lung microvascular cells were cultured with DMEM, 10% FBS, and 1% antibiotics until confluent as determined by light microscopy examination. They were then incubated with vehicle, 2-BrPA (10 µM), and 2-BrPALD (10 µM) or their corresponding nonbrominated compounds for 30 min. At that time, cell lysates were prepared as indicated in materials and methods for discovery proteomics analysis; 1,127 proteins were identified with high confidence (Supplemental Table S5). The most significant signaling pathways were then identified using GeneGo MetaCore where 106 proteins were found to be significantly changed in abundance (Table 4; Supplemental Table S6). As indicated previously, the discovery proteomics data were complemented with Western blot analysis carried out on key permeability-associated proteins along with phosphorylation status in addition to cytoimmunofluorescence studies to visualize cellular disruption, which were combined for the following pathway analysis. The top pathway (Fig. 12; Supplemental Fig. S1) included cytoskeleton remodeling, regulation of actin cytoskeleton organization by the kinase effectors of Rho GTPases, in agreement with the physiological changes we observed in our experiments. The primary proteins identified in the pathway shown include α-actinin, ezrin, radixin, and moesin proteins, talin, myosin heavy chain, RhoA, ROCK, actin cytoskeletal, vinculin, RhoA-related, F-actin cytoskeleton, and moesin. There were a number of additional cytoskeleton remodeling/permeability-associated proteins and pathways identified similarly that are indicated in Tables 3 and 4, respectively. There were 30 out of 106 significantly changed proteins that were found to be permeability-associated by way of GeneGo MetaCore, UniProtKB, and DAVID GO (Table 3; Supplemental Tables S3 and S4). The localization and processes mapped using MetaCore were cross-correlated to similar mouse lung-derived proteins. The majority of proteins, across species, mapped by cellular location and function but not by protein/gene (Supplemental Tables S3 and S4). The top pathways generally mapped to associations with cytoskeletal remodeling/permeability-associated (Table 4).

Table 4.

HMECs enrichment by pathway maps (permeability-associated)

Network Maps	Total	P Value	FDR	In Data	Network Objects from Active Data
Cytoskeleton remodeling_Regulation of actin cytoskeleton organization by the kinase effectors of Rho GTPases	58	4.0e−13	2.6e−10	11	α-Actinin, ERM proteins, talin, MyHC, RhoA, ROCK, actin cytoskeletal, vinculin, RhoA-related, F-actin cytoskeleton, moesin (MSN)
Cell adhesion_Histamine H1 receptor signaling in the interruption of cell barrier integrity	45	1.3e−09	4.3e−07	8	α-Actinin, talin, RhoA, ROCK, actin cytoskeletal, vinculin, VE-cadherin, β-catenin
Cell adhesion_Role of tetraspanins in the integrin-mediated cell adhesion	37	9.3e−09	1.5e−06	7	Talin, RhoA, ROCK, actin cytoskeletal, vinculin, ezrin (VIL2), actin
Cell adhesion_Endothelial cell contacts by junctional mechanisms	26	3.1e−08	4.1e−06	6	α-Actinin, AF-6, actin cytoskeletal, ZO-2, VE-cadherin, β-catenin
Cell adhesion_Integrin-mediated cell adhesion and migration	48	6.2e−08	6.8e−06	7	α-Actinin, talin, MyHC, RhoA, ROCK, actin cytoskeletal, vinculin
Cytoskeleton remodeling_Fibronectin-binding integrins in cell motility	32	1.2e−07	9.8e−06	6	α-Actinin, talin, RhoA, ROCK, actin cytoskeletal, vinculin
Cytoskeleton remodeling_ESR1 action on cytoskeleton remodeling and cell migration	23	6.5e−07	3.8e−05	5	ROCK2, ERM proteins, RhoA, F-actin cytoskeleton, moesin (MSN)
Cytoskeleton remodeling_Hyaluronic acid/CD44 signaling pathways	43	7.5e−07	3.8e−05	6	RhoA, ROCK, actin cytoskeletal, ezrin (VIL2), actin, moesin (MSN)
Cell adhesion_Cadherin-mediated cell adhesion	26	1.3e−06	5.9e−05	5	AF-6, actin cytoskeletal, VE-cadherin, actin, β-catenin
Chemotaxis_Lysophosphatidic acid signaling via GPCRs	129	5.5e−06	1.9e−04	8	RhoA, ROCK, Rho GTPase, G protein-γ12, actin cytoskeletal, vinculin, F-actin cytoskeleton, β-catenin
Cytoskeleton remodeling_Integrin outside-in signaling	49	3.2e−05	8.4e−04	5	α-Actinin, talin, actin cytoskeletal, vinculin, β-catenin
E-cadherin signaling and its regulation in gastric cancer	36	1.5e−04	3.1e−03	4	α-Actinin, RhoA, actin, β-catenin
Cytoskeleton remodeling_Role of PKA in cytoskeleton reorganization	41	2.5e−04	4.4e−03	4	RhoA, ROCK, actin cytoskeletal, F-actin cytoskeleton
Cytoskeleton remodeling_Regulation of actin cytoskeleton nucleation and polymerization by Rho GTPases	46	3.8e−04	5.9e−03	4	RhoA, actin cytoskeletal, RhoA-related, F-actin cytoskeleton

Pathway analysis output stemming from the 106 significantly changed proteins in human lung-derived microvascular endothelial cells (HMECs) treated with a combination of equal amounts of 2-bromopalmitic acid and 2-bromopalmitaldehyde vs. vehicle (Supplemental Table S6), including the antibody-targeted verification experiments. Most of the pathways mapped through GeneGo MetaCore were identified as permeability-associated; the top pathway is also illustrated (Fig. 12). ERM proteins, ezrin, radixin, and moesin; ESR1, estrogen receptor 1; FDR, false discovery rate; GPCRs, G protein-coupled receptors; MyHC, myosin heavy chain; ZO-2, zona occludens-2.

Fig. 12.

Human lung-derived microvascular endothelial cells. Top pathway map following brominated lipids treatment is shown. Human lung microvascular endothelial cells were incubated with vehicle, 2-bromopalmitic acid (10 µM), and 2-bromopalmitaldehyde (10 µM) or their corresponding nonbrominated compounds for 30 min, and cell lysates were prepared as indicated in materials and methods for discovery proteomics analysis. The most significant signaling pathways were then identified using GeneGo MetaCore; these were from the 106 proteins found to be significantly changed in abundance by liquid chromatography-tandem mass spectrometry (Table 3; Supplemental Table S6) in addition to immunotargeted proteins for which phosphorylation status or location/disruption were confirmed (immunocytoflourescence and Western blot analysis). The top pathway included cytoskeleton remodeling, regulation of actin cytoskeleton organization by the kinase effectors of Rho GTPases. The primary proteins identified in the pathway shown include α-actinin, ezrin, radixin, and moesin (ERM) proteins, talin, myosin heavy chain (MyHC), RhoA, ROCK, actin cytoskeletal, vinculin, RhoA-related, F-actin cytoskeleton, and moesin (MSN). The figure key can be found in Supplemental Fig. S1.

Table 3.

HMECs: cell adhesion/junction-associated proteins

				Statistics			UniProtKB, DAVID, and GeneGo MetaCore Definitions
UniProtKB Name	UniProt ID	Entrez ID	Network ID	SAM	t-Test P Value	Fold∆ (16CBr/Veh)	GO Localizations/Biological Processes
40S ribosomal protein S9	P46781	6203	RPS9	1.35	0.016	3.0	Anchoring and adherens junction/focal adhesion
Tight junction protein ZO-2	Q9UDY2	9414	ZO-2	1.11	0.029	2.6	Adherens and tight junction/cadherin binding
Myosin-9	P35579	4627	MYH9	1.93	0.009	2.6	Actin, cadherin, and calmodulin binding
60S ribosomal protein L7a	P62424	6130	RPL7A	1.11	0.010	2.0	Focal adhesion/cadherin binding
Tight junction protein 1 (zona occludens-1)	G5E9E7	7082	ZO-1	1.00	0.015	1.9	Cell and tight junction-associated
60S acidic ribosomal protein P0	P05388	6175	RPLP0	0.90	0.023	1.9	Anchoring and adherens junction/focal adhesion
Laminin subunit gamma-1	P11047	3915	LAMG1	1.47	0.006	1.8	Cell adhesion and migration
Desmin	P17661	1674	Desmin	1.19	0.009	1.7	Cytoskeletal and intermediate filament organization
Polyadenylate-binding protein 1	P11940	26986	PABPC1	1.53	0.006	1.7	Cell substrate junction
Catenin beta-1	P35222	1499	Beta-catenin	0.86	0.027	1.6	α-Catenin and cadherin binding/adherens junction assembly
Plectin	Q15149	5339	Plectin 1	1.76	0.013	1.6	Cell junction/actin and cadherin binding
Catalase	P04040	847	Catalase	1.04	0.025	1.5	Cell and adherens junction
Actin, alpha cardiac muscle 1	P68032	70	ACTC	0.93	0.021	−1.5	Myosin binding/actin and myosin filament movement
Filamin-C	Q14315	2318	Filamin C	1.68	0.005	−1.6	Actin and ankyrin binding/cell junction assembly
IQ motif containing GTPase activating protein 1	A4QPB0	8826	IQGAP1	1.28	0.025	−1.6	Anchoring and adherens junction
Moesin	P26038	4478	MSN (moesin)	1.45	0.005	−1.6	Actin and cell adhesion molecule binding/cell-to-cell adhesion
Talin-1	Q9Y490	7094	Talin-1	1.42	0.012	−1.6	Actin filament and cadherin binding/cell-to-cell junction
Ubiquitin-like modifier-activating enzyme 1	P22314	7317	UBE1	1.08	0.018	−1.8	Anchoring and cell junction
Laminin subunit alpha-5	O15230	3911	LAMA5	1.19	0.020	−1.8	Integrin binding/regulation of cell-to-cell adhesion
Filamin-B	O75369	2317	Filamin B (TABP)	1.57	0.002	−1.9	Actin and cadherin binding/actin cytoskeleton organization
Spectrin alpha chain, nonerythrocytic 1	Q13813	6709	Alpha-fodrin	1.48	0.015	−1.9	Actin, cadherin, and calmodulin binding/actin filament capping
14-3-3 protein epsilon	P62258	7531	14-3-3	1.00	0.025	−1.9	Cadherin binding/membrane reorganization
Ezrin	P15311	7430	VIL2 (ezrin)	1.63	0.005	−2.0	Actin filament, cadherin and cell adhesion binding
60S ribosomal protein L5	P46777	6125	RPL5	0.98	0.025	−2.0	Anchoring and adherens junction/focal adhesion
Nucleoside diphosphate kinase B	P22392	4831	NDPK B	0.87	0.046	−2.0	Intermediate filament binding/cell adhesion
Afadin	P55196	4301	AF-6	1.40	0.009	−2.1	Actin filament, cadherin and cell adhesion binding/cell adhesion
Alpha-actinin-4	O43707	81	Alpha-actinin 4	2.71	0.000	−2.6	Actin filament and integrin binding/tight junction assembly
Vinculin	P18206	7414	Vinculin	1.57	0.005	−2.9	Actin, catenin, and cadherin binding/adherens and apical junctions
Alpha-actinin-1	P12814	87	Alpha-actinin 1	2.90	0.000	−2.9	Actin filament, vinculin and integrin binding/focal adhesion assembly
Serine/threonine-protein phosphatase 2A	P30153	5518	PP2A structural	1.06	0.032	−3.0	Phosphorylation and degradation β-catenin/regulation cell adhesion

Systems analysis annotation output stemming from the 106 significantly changed proteins in human lung-derived microvascular endothelial cells (HMECs) treated with a combination of equal amounts of 2-bromopalmitic acid and 2-bromopalmitaldehyde vs. vehicle (16CBr/Veh; Supplemental Table S6). The proteins illustrated were all associated with cellular permeability-related functions. These 30 resultant proteins have been gene ontology (GO)-annotated using a combination of the UniProtKB, Database for Annotation, Visualization and Integrated Discovery (DAVID), and GeneGo MetaCore databases. ∆, Change; ID, identifier; SAM, significance analysis of microarray.

DISCUSSION

Br₂ is toxic to living organisms. This is reflected in the safety data sheet for reagent-grade Br₂ (CAS no. 7726-95-6, Revision Date 18-Jan-2018; Thermo Fisher Scientific) where Br₂ is listed as a category 1 toxin for skin, eye, and inhalation exposures. Br₂ is considered potentially fatal if inhaled with a permissible exposure limit of <0.1 ppm (45a). The National Institute for Occupational Safety and Health Immediately Dangerous to Life or Health limit for Br₂ gas exposure was recently reduced to 3 ppm based on additional review of previous human exposures (45a).

The broad application of Br₂ in diverse industrial settings increases the risk of accidental or malicious Br₂ release and consequent human exposure. Despite widespread use of Br₂, the specific mechanisms of toxicity remain unclear, hindering development of effective countermeasures and therapeutics. To our knowledge, there is no prior literature exploring the impact of Br₂ on organ-specific proteomes derived from translational animal models in vivo or in human cell lines ex vivo. To further characterize the mechanisms underlying Br₂ toxicity, we have employed the latest proteomics and bioinformatics tools combined with established functional and immunotargeted assays to cross-correlate both a translational mouse model of Br₂ exposure with human lung microvascular endothelial cells exposed to both Br₂ and Br-lips (Fig. 13).

Fig. 13.

Experimental summary. Our preliminary experiment involved the use of Br₂ exposure in mice to survey the global lung proteome at 24 h after exposure. This led to the identification of 95 proteins that changed significantly, with 30 that are known to be associated with permeability-related mechanisms. This was followed by the measurement of a functional experiment K_f to confirm lung permeability as a confounding endpoint pathology linked to Br₂ exposure. Brominated lipids (Br-lips) were then quantified in lung tissues with focus on 2-bromopalmitic acid (2-BrPA) and 2-bromopalmitaldehyde (2-BrPALD) derived from plasmalogens. This was a focus as a result of previously published data derived from similar experimental designs using halogens such as chlorine. High levels of Br-lips in lung tissues were identified, which led to the 2nd tier of experiments focused on the treatment of human lung-derived microvascular endothelial cells (HMECs) with both Br₂ and Br-lips separately. For this, we carried out a classic functional permeability experiment using transendothelial electrical resistance (TER), followed by immunofluorescence-established vascular permeability-associated proteins to confirm that disruptions were occurring within a single human vascular cell line, compared with the complex makeup of the lung. In addition, we carried out similar discovery proteomics analysis as we had in animals. This led to the identification of 106 proteins, whereby ~30 additional key permeability-associated proteins were confirmed, leading to a more complete set of systems analysis results with vascular permeability as a lead mechanism following Br₂ and Br-lips toxicity. FA’s, F-actins; IHC, immunohistochemical; p, phosphorylated; PE, ethanolamine glycerophospholipids; Veh., vehicle; WB, Western blot; ZO-1, zona occludens-1.

Advantages and Limitations

There are key advantages and limitations to global discovery-based proteomic studies. Proteomic applications can help investigators to gain a bird’s-eye viewpoint while at the same time potentially identifying previously unknown small-scale details. This is especially true when combined with systems analysis applications. In this way, proteomics can provide both a broader contextual overview while also identifying specific small-scale changes that may warrant additional investigation.

We chose a common discovery proteomic workflow with the expectation that this would provide us with clarity, context, and a better sense of direction to pursue for future investigations into novel mechanisms and potential therapeutic interventions following Br₂-induced lung injury. This approach, however, presents with key limitations that warrant emphasis: 1) unlike transcriptomic experiments, there is the inability to amplify proteins, thereby imposing a limit of detection for low-abundance proteins, and 2) as applied in this study, there is an inability to distinguish between causation and association. Causal inferences are more challenging to determine in discovery omics applications due primarily to the large number of quantified transcripts or proteins identified by modern techniques. As the number of proteins identified increases, the number of associations arising by random chance will increase correspondingly. To address this risk, we employed statistical techniques in concert with complementing proteomics data, Western blots, and cytoimmunofluorescence. It remains important for the reader when reviewing our tables to recall that the presence of a single protein or group of proteins should not be interpreted in isolation as causative for a phenotype or pathway.

In addition to these points, the individual proteins identified and their relative levels present in tissue fluctuate over time. Discovery proteomics applications are limited to capturing a snapshot of proteins frozen at a predetermined time point. Although this provides a still-frame of protein abundance at that moment, our analysis is incapable of describing temporal changes in abundance. Similarly, although phosphorylation status was explored for key permeability-associated proteins identified in our analysis, the measurement of specific posttranslational modifications of proteins that may otherwise be required for gain/loss of function present with an additional set of challenges and, therefore, was not explored at the global level here.

We utilized a proteomics workflow to serve as a guide for devising new molecular and translational experiments to confirm and further probe particular pathways of interest identified by this analysis. Our findings of the observed biological changes are consistent with permeability disruption following Br₂ exposure. This outcome directly prompted novel investigations into the capability of Br₂ to damage the pulmonary microvasculature using a measurement of filtration coefficient while also demonstrating previously unknown changes in endothelial cell actin stress filaments and VE-cadherin internalization specifically in halogen injury.

In all, our findings highlight three significant biological processes: 1) exosome secretion, 2) inflammation, and 3) vascular permeability. Although exosome secretion is of great interest in terms of understanding cellular cross talk-associated mechanisms, in addition to serum and urinary biomarkers for future interests, we have chosen to discuss recent work in relation to inflammatory pathways while primarily focusing our current experiments on changes in vascular permeability.

Inflammation

An acute inflammatory response has been well-described following toxic gas exposures, and our previous work has demonstrated that Br₂-induced lung injury is no exception. Br₂ exposure can induce increases in plasma levels of inflammatory cytokines (including IL-6, TNF-α, and keratinocyte chemoattractant/growth-related oncogene; Ref. 35) and increased inflammatory cells (macrophages and neutrophils; Refs. 2, 4) in bronchoalveolar lavage fluid. Our proteomic analysis identified a number of changes in proteins involved in regulation of the immune response, including neutrophil/granulocyte activation and myeloid cell activation.

We noted a 4.5-fold increase in cathelin-related antimicrobial peptide (CRAMP) in mice after Br₂ exposure compared with control animals. CRAMP is an ortholog of human cathelicidin/LL-37, functions as a leukocyte chemoattractant, and enhances the adaptive immune response in mice (34).

Serum paraoxonase 1 (PON1) was increased 3.3-fold in mice exposed to Br₂. PON1 is protective against organophosphate toxicity, but recent work has demonstrated its role in the organisms’ response to oxidative stress and inflammation along with the preservation of vascular endothelial function. Notably, the antioxidant function of PON1 centers around the hydrolyzation of oxidized lipids (15). The role that PON1 may play in response to halogens or halogenated lipids remains to be determined.

The heme-containing enzyme myeloperoxidase (MPO) was increased 2.9-fold in mice exposed to Br₂. MPO serves a variety of functions in host immune response regulation and influences both innate and adaptive immunity (7). MPO plays an important role in termination of an inflammatory response as well as promotion of inflammation via granulocyte recruitment after release from necrotic cells (49). Increases in MPO may reflect a quantitative increase in granulocytes following Br₂-induced lung injury. Of note, a significant increase of MPO staining was found in the lungs of mice 24 h after exposure; S100 calcium-binding protein A9 (S100A9) was increased 2.3-fold in Br₂-exposed mice (4). Also known as myeloid-related protein-14, this protein works through Toll-like receptor 4 activation to regulate vascular inflammation and leukocyte recruitment (17). Elevations in plasma S100A9 are predictive of adverse cardiovascular events, and S100A9 is a ligand of receptor for advanced glycation endproducts, which is overexpressed in pulmonary artery smooth muscle cells in patients with pulmonary arterial hypertension. Future work is warranted to further explore this pathway. It is possible that chronic alterations in S100A9 could lead to adverse pulmonary vascular remodeling (45).

Additional proteins that are involved in inflammation and the immune response and found in the lungs of mice following Br₂ exposure in addition to HMECs following Br-lips treatment are highlighted in Tables 1 and 2 (mouse lung), Supplemental Tables S3 and S4 (mouse lung cross-correlated with HMEC), and Supplemental Table S6 (HMEC), respectively. This vast array of new evidence supports the important role of the host inflammatory response after acute Br₂-induced pulmonary injury. The degree to which the host immune response may be contributing to or exacerbating acute lung damage in this setting remains to be determined. It is possible that an overzealous host inflammatory response may lead to additional pulmonary injury in susceptible patients in a manner analogous to ARDS.

Vascular Permeability

Br₂ and other halogen gases cause lung injury, in part, through disrupting pulmonary barrier function. We (29) have previously demonstrated systemic endothelial dysfunction in rats following Cl₂ exposure with decreased endothelial nitric oxide synthase expression and activity. Prior work looking at Br₂ exposure in pregnant mice has found evidence of an increased risk for endothelial dysfunction in these animals that appears to manifest in the lungs as severe pulmonary edema and systemically by triggering a pre-eclamptic-like syndrome (1, 2, 35). Interestingly, as indicted in the referenced materials, this phenotype is partially rescued with administration of the type 5 phosphodiesterase inhibitor tadalafil or by administration of exogenous VEGF-121, suggesting that cGMP and VEGF pathways are, respectively, involved in the pathogenesis of injury in pregnant animals.

Our current proteomic analysis has identified several targets for investigation related to vascular permeability and endothelial function. In mice exposed to Br₂, these include an increase in tubulin-specific chaperone D [2.8-fold increase (↑)], involved in adherens and tight junction disassembly, and an increase of guanine nucleotide-binding protein subunit α-13 (2.2-fold↑), which functions to regulate the actin cytoskeleton along with actin-related protein 2/3 complex subunit 2 (2.1-fold↑; Table 1; Refs. 10, 36, 55). In HMECs, Br-lips exposure led to increased abundance of a host of proteins associated with cellular adhesion and junctional integrity (Table 3). These include tight junction protein zona occludens-2 (ZO-2; 2.6-fold↑) and ZO-1 (1.9-fold↑) along with catenin β-1 (1.6-fold↑), which are all involved in cadherin binding (Table 3). Vinculin, α-actinin-1, and serine/threonine-protein phosphatase 2A were all found to be decreased by 2.9-, 2.9-, and 3.0-fold, respectively (Table 3). All three are important in junctional integrity and cadherin binding (13, 30, 43).

In concert with our proteomic and systems biology approach, we designed and pursued experiments that further characterized pulmonary vascular permeability changes in vivo. We then reported changes in HMEC transendothelial resistance (a measure of vascular permeability) and measured changes in specific permeability-associated proteins and pathways within HMECs following Br₂, and Br-lips exposure, including F-actin stress fiber formation, VE-cadherin phosphorylation and internalization, and disruption of ZO-1, along with noting that Br-lips induced phosphorylation (activation) of both RhoA and ROCK2. These combined approaches illustrated in Fig. 13 allowed us to directly explore an important permeability-associated cytoskeletal pathway implicated by our discovery proteomic analysis (Fig. 12; Supplemental Fig. S1).

Our combined workflow presents with several strengths. This is the first time an isolated lung perfusion assay was employed to assess the global changes in pulmonary vascular permeability following halogen gas exposure. The marked increase in K_f at 24 h is consistent with increased pulmonary vascular permeability and provides the strongest evidence to date accounting for the role of permeability in the pulmonary edema observed in these animals. HMEC transendothelial resistance provides cellular evidence of these changes and helps further confirm that the proteomic signal we observed may be biologically meaningful.

Future Studies

As a whole, this work has provided a base for future studies. Although we have focused herein on pulmonary endothelial injury, it will be important in the future to explore halogen-associated injury to the alveolar epithelium, particularly as our finding of an increased filtration coefficient in mice exposed to Br₂ may also occur in the context of injury to lung epithelium. Similarly, disruptions in ion transport are likely involved; more specifically, we believe that alterations in calcium signaling and transport across the pulmonary epithelium following halogen exposure warrants additional research, similar to previous reports following exposure of human airway smooth cells to chlorine and bromine (37, 38). Calcium may also play an important role in the regulation of the systemic vascular endothelium, particularly in the pathogenesis of the systemic hypertension and endothelial dysfunction observed in pregnant mice following exposure to Br₂ (1, 2, 35). Additionally, although we have focused our efforts on the acute effects of Br₂ injury, it will be important in the future to explore the subacute and chronic changes related to pulmonary endothelial dysfunction and inflammation as well as expanding our investigations to other halogenated agents, including chlorine and phosgene. Several therapeutic options for treatment of acute halogen injury deserve attention based on our findings, including modulators of the inflammatory response and agents that affect endothelial function, including phosphodiesterase inhibitors and exogenous nitric oxide. Indeed, previous studies have shown that nitrite administered intramuscularly to mice after Cl₂ prolongs survival and decreases mortality (28, 62). Similarly, administration of the phosphodiesterase inhibitor rolipram decreased chlorine-induced lung injury in mice (16). Furthermore, our data are congruent with prior work demonstrating an increase in filtration coefficient after hydrochloric acid tracheal instillation (63). Acute lung injury was mitigated by administration of high-molecular-weight hyaluronan, which also served to decrease ROCK2 phosphorylation (63). This observation points to a common pathway (RhoA-ROCK2) between the two separate mechanisms of injury, and consequently treatment with high-molecular-weight hyaluronan may also be beneficial in halogen gas-induced injury.

Therefore, further studies into the reported biological functions and pathways are in order; this type of data helps to guide future directions, as these cross-correlates lend a higher level of confidence to the global acute effects of Br₂ exposure in vivo (51).

In this report, we have identified potentially significant biological functions to help guide us toward specific molecular pathologies following acute Br₂ exposure. We have generated proof of principal that this type of robust proteomics workflow coupled to a series of systems biology analysis can be applied to a number of understudied models of toxic exposures to better understand what biological systems may be influenced downstream and over time. We expect that our data contained within this manuscript and those available through the ProteomeXchange Consortium will aid others in their efforts to better characterize and understand acute toxic lung injury.

Data Sharing

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the data set identifier PXD015462 (see endnote).

GRANTS

Funding was provided by the CounterACT Program, National Institutes of Health Office of the Director, the National Institute of Neurological Disorders and Stroke, and the National Institute of Environmental Health Sciences, Grants 5UO1 ES026458 03, 3UO1 ES026458 03S1, and 5UO1 ES027697 02 to S. Matalon; National Heart, Lung, and Blood Institute, National Institutes of Health, Grant T32HL129948 to D. R. Addis; and UAB Comprehensive Cancer Center, agency: National Institutes of Health, institute: National Cancer Institute, Project no. P30CA013148 to J. A. Mobley. This study was supported by research funding from the National Institute of General Medical Sciences R01 GM-115553 to D. A. Ford.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

S.A., S.M., and J.A.M. conceived and designed research; D.R.A., S.F.D., M.-Y.J., I.A., K.K., D.A.F., and J.A.M. performed experiments; D.R.A., S.A., S.F.D., M.-Y.J., I.A., K.K., D.A.F., and J.A.M. analyzed data; D.R.A., S.A., S.M., and J.A.M. interpreted results of experiments; D.R.A., S.A., S.F.D., M.-Y.J., K.K., D.A.F., S.M., and J.A.M. prepared figures; D.R.A., S.M., and J.A.M. drafted manuscript; D.R.A., S.A., K.K., D.A.F., S.M., and J.A.M. edited and revised manuscript; D.R.A., S.A., S.F.D., M.-Y.J., I.A., K.K., D.A.F., S.M., and J.A.M. approved final version of manuscript.

ENDNOTE

At the request of the authors, readers are herein alerted to the fact that additional materials related to this manuscript may be found at a website that at the time of publication the authors indicate is: https://doi.org/10.6019/PXD015462. These materials are not a part of this manuscript and have not undergone peer review by the American Physiological Society (APS). APS and the journal editors take no responsibility for these materials, for the website address, or for any links to or from it.