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.

Keywords: ESI-MS2; actin; bromine; discovery; halogen; proteomics; systems biology; vascular permeability.

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.

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.

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.

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.

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.

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).

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).

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.

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.

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.

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.