Acute cytotoxicity and increased vascular endothelial growth factor after in vitro nitrogen mustard vapor exposure

Abstract

Nitrogen mustard (NM) is a highly toxic alkylating agent. Inhalation exposure can cause acute and chronic lung injury. This study’s aims were to develop an in vitro coculture model of mustard-induced airway injury and to identify growth factors contributing to airway pathology. Primary human bronchial epithelial cells cultured with pulmonary endothelial cells were exposed to NM (25, 50, 100, 250, or 500 μM) or PBS (control) for 1 hour. Lactate dehydrogenase (LDH) and transepithelial electrical resistance (TEER) were measured prior to and 24 h after NM exposure. Fixed cultures were stained for H&E or live/dead staining. Culture media was analyzed for eleven growth factors. A 1-h vapor exposure to greater than or equal to 50 μM NM increased supernatant LDH, decreased TEER, and caused airway epithelial cell detachment. Endothelial cell death occurred at 500 μM NM. Vascular endothelial growth factor A (VEGF-A) and placental growth factor (PlGF) expression increased in 500 μM NM–exposed cultures compared PBS control cultures. NM vapor exposure causes differential cytotoxicity to airway epithelial and endothelial injury in culture. Increased VEGF-A and PlGF expression occurred in airway cocultures. Future studies are required to validate the role of VEGF signaling in mustard-induced airway pathology.

Graphical abstract:

The in vitro model of primary airway epithelial and endothelial cells at the air–liquid interface closely models the injury seen after an acute mustard inhalation injury with acute cytotoxicity at increasing NM concentrations. Two important growth factors, VEGF-A and PlGF, related to pulmonary vascular biology and angiogenesis were increased at 24 h after a single-hour NM vapor exposure. Future in vivo and in vitro studies are required to validate the use of VEGF-A or PlGF as a potential biomarker or drug therapy for either acute or chronic vascular pathogenesis after high-concentration mustard inhalation exposure.

Introduction

Sulfur mustard (2,2-dichlorodiethyl; SM) is a highly toxic blistering agent. It is classified as a chemical threat owing to its known use as a warfare agent. First synthesized in the 19th century, SM remains the most utilized chemical weapon on the planet with greater than 100 years of chemical weapon use^1-3. Nitrogen mustard (bis(2-chloroethyl)methylamine; HN₂; NM) is a second vesicant. Similar to SM, one of its mechanisms of cytotoxicity is DNA alkylation. NM was stockpiled by many European countries during World War II. Its primary use is as a chemotherapeutic though NM remains classified as a chemical threat due to its potential use as a warfare agent.

Both SM and NM inhalation exposure can cause both acute and chronic injury to the respiratory tract ^4,5. Acute pulmonary injury manifests as pulmonary edema, pulmonary hemorrhage, airway obstruction, and airway mucosal injury ^6-8. Airway obstruction is the primary cause of acute mortality after high-dose SM exposure ^9-11. In those individuals who survive an acute exposure, late pulmonary sequelae often develop with symptoms of dyspnea and chronic cough as well as signs of persistent hypoxemia ^12-14.

Although acute airway obstruction is the main cause of acute mortality after mustard exposure, limited understanding remains in regards to both acute and chronic mustard-induced lung pathogenesis. Following high-dose mustard inhalation exposure, injury to both the airway epithelial and adjacent endothelial cell layers occurs ^15,16. This acute airway injury results in bronchial vascular leak and subsequent bronchial cast formation ^11,17,18. Fibrin-rich bronchial casts cause significant airway obstruction, and ultimately, death ^9,18. The primary purpose of this manuscript was to develop an in vitro coculture of primary human airway epithelial cells (HAECs) with primary pulmonary artery endothelial cells (PAECs) for modeling mustard-induced airway injury. The secondary purpose of this manuscript was to screen a panel of growth factors for potential biomarker identification of epithelial–endothelial interaction following acute mustard inhalation injury.

Materials and methods

Chemical

NM (mechlorethamine hydrochloride, 98.0% purity) was purchased from Sigma-Aldrich (St. Louis, MO). NM was used over SM owing to international, federal and local restrictions on SM. NM use was approved by the University of Rochester’s Environmental Health and Safety

Primary human airway epithelial cultures

Human EpiAirway™ (AIR-100) generated from primary human bronchial epithelial cells of a healthy, nonsmoking donor (TBE-20), were purchased from MatTek Corporation (Ashland, MA). All tissues were well-differentiated at the air–liquid interface (ALI) on a microporous Transwell^® (9 mm internal diameter) membrane in plastic inserts before receipt and exposure. Upon receipt from MatTek, cultures were placed into 1 mL of culture medium (MatTek, Asthland, MA) in 6-well culture plates for at least 24 h prior to exposure for proper recovery and equilibration.

Primary human PAECs

Primary human PAEC were obtained from Lonza (Walkersville, MD). PAECs were cultured in T-75 flask to 70–80% confluence in endothelial growth medium-2 (EGM; Lonza) with EGM SingleQuots^® Kits (Lonza). Before exposure, PAECs were seeded onto 6-well culture plates and allowed to proliferate until 80% confluence. Once the PAECs reached the required density, a Transwell insert was inserted above the endothelial cultures (Corning^®, Corning, NY), and the AIR-100 cultures were placed on the transwell insert to complete the coculture culture for exposure. All exposures occurred in PAECs that underwent less than five passages.

In vitro NM vapor cup exposure

Immediately prior to exposure, NM was diluted in phosphate-buffered solution (PBS) to final concentrations of 25, 50, 100, 250, and 500 μM. PBS vehicle was used as a negative control. Concentrations were chosen from previously published estimations of vapor concentrations within the vapor cup exposure at 37 °C, and relevant to high-concentration NM exposure in chemical exposures ^19,20. ALI tissue cultures were exposed to NM-derived vapors for 1-h using vapor cups as described previously ²¹. A schematic figure of the coculture model is illustrated in Figure 1A with a representative image of vapor cup and epithelial culture in Figure 1B. Briefly, 50 microliters (μL) NM or PBS was pipetted onto a 6-mm antibiotic sensitivity disk (BD BBL™; Franklin Lakes, NJ) located within a 1.5-mL Eppendorf tube top. The vapor cup was inverted over the tissue culture; sealed tightly onto the plastic well insert; and placed into a 5% CO₂ incubator for 1 hour. After 1 h, the Eppendorf tube top was removed, and the tissue cultures were returned to the incubator prior to analysis at 24 h postexposure.

Figure 1.

(A) Schematic figure of the coculture model with human bronchial airway epithelial cells cultured at the air–liquid interface (ALI) in a porous membrane culture well. Primary pulmonary artery endothelial cells (PAECs) were cultured to confluence in the lower culture plate. Culture media (supplied by MatTek Corporation) was supplemented above the endothelial cells and below the epithelial cells. Supernatant LDH (LDH) was collected from the epithelial cells before and after exposure. Transepithelial electrical resistance (TEER) measurements were taken across the epithelial layer. Growth factors (ProcartaPlex 11 Human Growth Factor) was performed on growth media obtained at 24 h after exposure. Live/dead staining (Invitrogen) was performed on PAECs at 24 h after exposure. (B) Representative image of vapor cup (left) and epithelial culture dish (right) with porous membrane removed.

Cell viability via lactate dehydrogenase (LDH) activity

LDH release was measured in apical washes of the epithelial supernatant as a surrogate marker of cell viability using a commercially available calorimetric kit (490/680 nm; Thermo Scientific Pierce, Rockford, IL). The apical surface of the ALI tissues was gently rinsed with 0.4 mL PBS prior to exposure and at 24 h after exposure (day 1). Apical rinses were centrifuged to remove mucus/debris prior to LDH testing. The absorbance data for each tissue sample was normalized to 100% tissue death (complete LDH release when lysed with 0.2% Triton X-100 in PBS) and baseline LDH release (untreated incubator control tissue) using the following formula: corrected cell viability = 100 - [Abs(X) - Abs(Inc)]/[Abs(Triton) - Abs(Inc)]*100 where Abs(X) is the absorbance of the sample, Abs(Inc) is the absorbance of the incubator control, and Abs(Triton) is the absorbance of the sample incubated in 0.2% Triton X.

Transepithelial electrical resistance (TEER)

Transepithelial electrical resistance (TEER) was measured in all tissue cultures prior to and at 24 h after exposure using silver chloride electrodes (EVOM², World Precision Instruments, Sarasota, FL). Electrodes were connected to the volt-ohmmeter and were equilibrated in a balanced PBS solution (MatTek) for 15 min before use. Zero-point four millimeters of warm PBS solution (MatTek) was added to the apical surface of ALI cultures. TEER was measured by placing the longer electrode into the basal media, and the shorter electrode into the apical transwell insert. Two measurements were taken from each insert. ALI cultures were not used for exposure if the TEER measurement was ≤300 Ohms*cm^{2 22,23}.

Histologic evaluation of primary airway epithelial cultures

At 24 h, tissue cultures exposed to PBS and 25 μM NM were fixed in 10% neutral-buffered formalin (NBF) overnight at 4 °C followed by cold PBS wash. Tissue cultures were excised from transwell insert, placed in a Kim wipe and enclosed within a tissue embedding cassette. Tissue cultures were dehydrated in 80% ethanol, embedded in paraffin, sectioned (5-μm), and mounted on silane-coated glass slides. Sections were stained with hematoxylin and eosin (H&E). For 25 μM NM–exposed cultures, two images were taken at 40x magnification for each stained section roughly 50% and 75% from the left end of the section. The length of the image at 40x magnification was 200 micrometers. All images from 25 μM NM–exposed culture sections were quantified for the number of lesions per mm of length.

Live/dead staining of PAECs

At 24 h after exposure, basolateral cultures of PAECs were labeled with a Live/Dead^® Assay Kit (Invitrogen; Carlsbad, CA). The live/dead assay differentially labels live and dead cells with fluorescent dyes. Live cells stain green to indicate intracellular esterase activity with calcein-AM (Live Green) while dead cells stain red to indicate loss of plasma membrane integrity with ethidium homodimer-1 (Dead Red). The live/dead assay was used in place of the LDH Assay to assess cytotoxicity in the basolateral cultures due to the presence of phenol red in media, which interferes with the LDH Assay. After thawing, 1 mL of Live Green solution was transferred into a 1 μL Dead Red solution. The solution was diluted in 1 mL PBS and was applied to the PAEC cultures. Cells were incubated for 15 min and were then fluorescently imaged at 488 nm and 570 nm, respectively, at 4x magnification. Three representative images were taken of each 6-well insert for further quantification.

Quantification of percent surface area for live/dead imaging

Each image obtained at 488 nm (live cells) and 570 nm (dead cells) was quantified for percent surface area using ImageJ software (The National Institutes of Health, Bethesda, MD). The average pixel intensity was calculated for each image by selecting the entire image and analyzing using the “Measure” tool of ImageJ. For each image, the total calculated area was also recorded in square centimeters. To calculate the average pixel intensity per image, each average pixel intensity was divided by the total calculated surface area. Three images were quantitated for each condition. The percent (%) surface area of dead-to-live cells was then calculated for each image by dividing the surface area of dead cells by the surface area of live cells.

Biomarker analyses—lung growth factors

Twenty-four hours after treatment, the medium was removed and stored at −80 °C. A panel of growth factors in the conditioned medium were analyzed using Luminex multiplex technology (growth factor 11-plex human ProcartaPlex™ Panel, Invitrogen, Carlsbad, CA). This panel measures 11 human-specific protein targets including nerve growth factor-beta (NGF-β), brain-derived neurotrophic factor (BDNF), epidermal growth factor (EGF), fibroblast growth factor 2 (FGF2), hepatocyte growth factor (HGF), leukemic inhibitory factor (LIF), platelet-derived growth factor-BB (PDGF-BB), placental growth factor (PlGF), stem cell factor (SCF), and vascular endothelial growth factors A and D (VEGF-A and VEGF-D). Both assays were performed as recommended by the supplier using a Bio-Plex 100 instrument (BioRad, Hercules, CA).

Statistical analysis

All exposures were performed twice at each concentration for sufficient repeatability. For normally distributed data, results of quantitative measures were expressed as means ± standard deviation (SD) and analyzed using a one-way analysis of variance (ANOVA) followed by Dunnett’s post hoc analysis. All data were analyzed using with GraphPad Prism^® 7.0 software (GraphPad, La Jolla, CA).

Results

Increased supernatant LDH after NM vapor exposure

LDH in airway epithelial supernatants was used as a surrogate marker of acute cytotoxicity after NM exposure. At 24 h post-NM vapor exposure, airway supernatant LDH increased significantly in cultures exposed to NM compared with PBS controls (Fig. 2A; ANOVA; n = 8/group; **** P < 0.0001). Supernatant LDH in airway cultures exposed at the lowest concentration of NM (25 μM) did not statistically differ from PBS controls (P = 0.3698; an ANOVA with Dunnett’s). These results support acute cytotoxicity in the airway epithelial layer when exposed to greater than or equal to 50 μM NM.

Figure 2.

(A) Supernatant LDH sampled apically from airway epithelial cells at 24 h postexposure. LDH absorbance data for each tissue sample was normalized to 100% tissue death (complete LDH release when lysed with 0.2% Triton X-100 in PBS) and baseline LDH release (untreated incubator control tissue). Nitrogen mustard (NM) concentrations included 25, 50, 100, 250, and 500 μM or PBS control. Supernatant LDH in NM-exposed cultures differed significantly from PBS control (ANOVA; n = 8/group; **** P < 0.0001). (B) Transepithelial electrical resistance (TEER) at 24 h post-NM exposure. TEER differed significantly at 24 h in all NM concentrations compared with PBS controls (ANOVA; n = 8/group; **** P < 0.0001).

Decline in transepithelial electrical resistance (TEER) at 24 h after NM exposure

Transepithelial electrical resistance (TEER) was used as a marker of airway epithelial cellular permeability. TEER measurements decreased significantly at 24 h postexposure in all NM-exposed cultures compared with PBS exposed controls (Fig. 2B; ANOVA; n = 8/group; **** P < 0.0001). TEER also progressively decreased with higher NM concentrations. Thus, all airway epithelial cultures demonstrated increased cellular permeability after NM exposures with increasing permeability noted at higher NM concentrations.

Histologic evidence of airway epithelial cell death after nm exposure

All airway epithelial cultures were visually inspected at 24 h after NM vapor exposure prior to histologic fixation. Epithelial cultures exposed to greater than or equal to 50 μM NM demonstrated multiple areas of denudation with a loss of confluence. Conversely, PBS-exposed and 25 μM NM cultures remained confluent by visual inspection. Thus, PBS and 25 μM NM–exposed cultures were fixed for further histologic evaluation and H&E cross-sectional staining. PBS-exposed cultures (Fig. 3A) demonstrated evidence of differentiation with pseudo-stratification and cilia present along the apical surface (Fig. 3A; black arrows). In cultures exposed to 25 μM NM for 1 h, multiple intercellular, debris-filled lesions appeared in H&E cross-sections (Fig. 3B; black arrows). The number of lesions per millimeter (mm) of epithelial culture length was 15.9 ± 11.3 (n = 16 total images analyzed). Additionally, the basement membrane detached from the underlying permeable culture membrane (Fig. 3B; white arrow) in multiple locations in cultures exposed to 25 μM NM, but not in the PBS controls cultures. Thus, airway epithelial cells (AECs) exposed to the lowest concentration of NM (25 μM) histologically demonstrated injury but prior to significant cell death.

Figure 3.

(A) Representative hematoxylin & eosin (H&E), cross-sectional image of pseudostratified bronchial airway epithelium exposed to phosphate-buffered solution (PBS) and fixed at 24 h postexposure (scale bar: 50 μm). Black arrows denotes multiple cells with cilia along the apical surface. (B) Representative H&E image of pseudostratified bronchial airway epithelium exposed to 25 μM nitrogen mustard (NM) for 1 h fixed at 24 h postexposure (scale bar: 50 μm). Black arrow denotes multiple intercellular lesions filled with debris. White arrow identifies detachment of airway epithelium from the adjacent permeable membrane (scale bar: 50 μm).

PAEC viability after NM exposure

PAECs cultured in the basolateral compartment of the coculture system were stained for cellular viability using live/dead staining at 24 h postexposure. Figure 4A provides representative images of PAECs stained for live (green) and dead (red) cells at 24 h postexposure under each exposure condition (PBS, 25, 50, 100, 250, and 500 μM NM). Limited dead cell staining (red) was seen in PAECs exposed to 25, 50, and 100 μM NM or PBS cultures. Conversely, increased dead cell staining in PAECs became more apparent at 250 and 500 μM NM exposure concentrations.

Figure 4.

(A) Representative images of live/dead stained primary artery endothelial cell cultures (PAECs) at 24 h after exposure to PBS (i), 25 μM NM (ii), 50 μM NM (iii), 100 μM NM (iv), 250 μM NM (v), and 500 μM NM (vi). Live cells stain green indicating intracellular esterase activity with calcein-AM (Live Green) while dead cells stain red indicating loss of plasma membrane integrity with ethidium homodimer-1 (Dead Red). (B) Total percent (%) surface area of dead-to-live PAECs at 24 h after NM vapor exposure. Percent surface area of dead-to-live PAECs significantly increased in cultures exposed to 500 μM NM (19.8 ± 5.0 % versus 3.6 ± 3.4 %; n = 3 images per condition; n = 4/group; an ANOVA with Dunnett’s; **** P < 0.0001). No statistically significant difference was present at all other NM concentrations compared with PBS controls.

The percent surface area (SA) of both live and dead cells was quantified under each exposure condition (three images per condition; n = 4/group; Figs. SE1A and SE1B, online only, respectively). The percent SA of live PAECs did not differ between exposure conditions (ANOVA; P = 0.12; Fig. SE1A, online only) while the percent SA of dead PAECs differed significantly at 500 μM NM compared with PBS controls (21.7 ± 4.0 % versus 2.8 ± 1.7 %; n = 3 images per condition; n = 4/group; an ANOVA with Dunnett’s; ****P < 0.0001; Fig. SE1B, online only). When adjusted for percent surface area of live cells, the percent dead-to-live cells remained elevated at 500 μM NM (19.8 ± 5.0% versus 3.6 ± 3.4%; n = 3 images per condition; n = 4/group; an ANOVA with Dunnett’s; ****P < 0.0001; Fig. 4B). Thus, at 500 μM NM, a significant decrease in cellular viability of PAECs occurred in the basolateral compartment of the coculture system.

Increased VEGF expression after nm exposure

At 24 h after exposure, the coculture media was collected and analyzed for a panel of growth factors as potential biomarkers of airway injury. Results from all eleven growth factors tested in the media of exposed cocultures are listed in Table 1. Of the 11 growth factors tested, only two factors, VEGF-A and PlGF increased significantly compared with PBS controls (n = 4/group; ANOVA; *P = 0.025 and *P = 0.011; Fig. 5A and 5B, respectively). VEGF-A levels in coculture media exposed to 500 μM NM increased compared with PBS controls (5,559 ± 2,771 versus 572 ± 288 pg/mL; an ANOVA with Dunnett’s; *P = 0.023; Fig. 5A). PlGF levels in coculture media exposed to 500 μM NM were also significantly increased compared with PBS controls (179.8 ± 94.4 versus 26.9 ± 2.6 pg/mL; an ANOVA with Dunnett’s; *P = 0.016). The other nine growth factors tested did not differ significantly from PBS controls at any of the NM concentrations tested (Table 1). Thus, a rise in both VEGF-A and PlGF expression occurred concurrently with increased cellular cytotoxicity when exposed to 500 μM NM.

Table 1.

Cytokine expression levels of eleven human-specific growth factors tested in conditioned media collected at 24 h after exposure (n = 4/group)

	PBS	50 μM	100 μM	250 μM	500 μM	ANOVA
BDNF	OOR	OOR	OOR	OOR	OOR	N/A
NGF-β	OOR	OOR	OOR	OOR	OOR	N/A
EGF	922 ± 743	1006 ± 695	2685 ± 2647	3597 ± 1717	2678 ± 2702	P = 0.29
FGF2	688 ± 448	310 ± 435	1866 ± 1943	1311 ± 797	1850 ± 864	P = 0.19
HGF	17.5 ± 15.9	5.6 ± 5.2	34.4 ± 62.8	27.6 ± 31.6	63.5 ± 53.4	P = 0.36
LIF	4.9 ± 6.4	17.0 ± 9.0	20.6 ± 22.8	14.7 ± 13.1	35.3 ± 12.2	P = 0.08
PDGF-BB	OOR	OOR	OOR	OOR	OOR	N/A
PIGF-1	26.9 ± 2.6	17.1 ± 5.4	112.6 ± 108.2	137.7 ± 25.7	179.8 ± 94.4#	*P = 0.011
SCF	0.20 ± 0.13	0.23 ± 0.08	0.78 ± 0.77	0.31 ± 0.18	0.78 ± 0.55	P = 0.17
VEGF-A	572 ± 288	500 ± 375	3836 ± 3935	3183 ± 1389	5559 ± 2771#	*P = 0.025
VEGF-D	OOR	OOR	OOR	OOR	OOR	N/A

Note: Specific cytokines: BDNF, beta brain–derived neurotrophic factor; NGF-β, nerve growth factor; EGF, epidermal growth factor; FGF2, fibroblast growth factor 2; HGF, hepatocyte growth factor; LIF, leukemic inhibitory factor; PDGF-BB, platelet-derived growth factor-BB; PIGF, placental growth factor; SCF, stem cell factor; VEGF-A and VEGF-D, vascular endothelial growth factors A and D (Growth Factor 11-plex Human ProcartaPlex™ Panel; Invitrogen; Carlsbad, CA).

^*denotes an ANOVA with P < 0.05;

^#denotes an ANOVA with Dunnett’s comparison of P < 0.05.

Figure 5.

(A) Vascular endothelial growth factor A (VEGF-A) and (B) placental growth factor (PlGF), cytokine levels in coculture conditioned media at 24 h after exposure. VEGF-A levels increased significantly in coculture media exposed to 500 μM NM compared with PBS controls (5559 ± 2771 versus 572 ± 288 pg/mL; ANOVA; n = 4/group; *P = 0.025). (B) PlGF levels increased significantly in coculture media exposed to 500 μM NM compared with PBS controls (179.8 ± 94.4 versus 26.9 ± 2.6 pg/mL; ANOVA; n = 4/group; * P = 0.011).

Discussion

The current study is novel in its use of the coculture model of primary HAECs and primary human endothelial cells exposed to NM vapor. This in vitro coculture system was developed to model the known in vivo pulmonary complications of high-concentration mustard vapor exposure with airway epithelial cell death ^7,15, adjacent pulmonary vasculature injury ^11,16,18, and subsequent vascular leak ¹⁷. Different from prior in vitro mustard exposure studies ^19,24-27, the model system uses vaporized NM in place of topical mustard exposure for exposures. This model system is also unique in its use of primary, differentiated human cell cultures for studying the interaction of AECs with PAECs after mustard exposure. When cocultured and exposed to high-concentration NM vapor (500 μM NM), both VEGF-A and PlGF expression increased and may be used as potential biomarkers of acute mustard-induced airway injury.

The model system developed here uses a more realistic exposure of mustard vapor than that of topical mustard exposure. We adapted a vapor cup model for vapor exposure in place of topical mustard exposures ²¹. In previous studies using topical exposure to the NM derivative 4-[bis(2-chloroethyl)amino]-L-phenylalanine (melphalan; L-Pam), Pohl et al.²⁸ demonstrated decreased viability and decreased transepithelial electrical resistance (TEER) measurements in a differentiated airway epithelial culture. Consistent with prior topical exposures, our results demonstrate increased LDH expression and epithelial cellular detachment in airway epithelial cultures at 24 h postexposure after a 1-h 50 μM NM exposure. In contrast to prior studies using undifferentiated transformed airway cell lines exposed to topical NM ^20,29, the cytotoxic effects seen here were experienced at lower concentrations of NM in our primary, differentiated culture model. Thus, differentiated HAECs exposed to NM vapor demonstrated similar cytotoxicity to that of differentiated AECs exposed to topical NM, but differentiated epithelial cells were more sensitive to NM cytotoxicity than undifferentiated, transformed airway cultures exposed to NM.

To the best of our knowledge, this is also the first study to evaluate the effects of NM exposure on the coculture of primary AECs with primary endothelial cells. Using transformed cell lines, Pohl et al.²⁸ evaluated the effects of topical melphalan on a coculture alveolar model of a lung adenocarcinoma cell line NCI H441 with an angiosarcoma cell line Iso-Has-1. This culture model is most reflective of the distal lung parenchyma and less reflective of airway pathology. NCI H441 cells were cultured apically while Iso-Has-1 cells were cultured on the basolateral surface. Surprisingly, the Iso-Has-1 cells were both more sensitive to topical L-Pam exposure in this model with a lower half-maximal effective concentration (EC₅₀) than the apical NCI H441 cultures. In our model, the primary endothelial cells demonstrated increased cell death only at the highest NM concentration of 500 μM. The most likely explanation for the higher LC₅₀ of pulmonary endothelial cells in our model compared with the prior model is differences in in vitro design. In our model, the airway epithelial layer consisted of a pseudostratified epithelium with pulmonary endothelial cells cultured on the bottom of a 6-well culture plate (see schematic Fig. 1). This design provided greater barrier protection to the underlying endothelial layer than a single monolayer culture used in the prior study. In Pohl et al.²⁸, endothelial cells were cultured directly on the basolateral surface of the permeable cell insert. Thus, the current study is likely more reflective of the airway considering that PAECs do not grow directly on the basement membrane of the airway while the prior study may be a more reflective model of the distal lung parenchyma where epithelial and endothelial cells grow adjacent to one another at the alveolar–capillary interface.

The second purpose of the study was to assess for potential biomarkers of mustard-induced airway injury. Of the eleven growth factors analyzed, only vascular endothelial growth factors, specifically VEGF-A and PlGF, increased significantly compared with PBS controls. Changes in VEGF-A and PlGF expression occurred concurrently with increased pulmonary endothelial cell cytotoxicity following the highest NM vapor concentration (500 μM). The VEGF-A expression seen in coculture following NM exposure is equivalent in magnitude to VEGF-A concentration known to stimulate human endothelial cell proliferation in culture³⁰. Both VEGF and PlGF are potent angiogenesis-stimulating cytokines. VEGF signaling is known to increase pulmonary endothelial cell migration ³¹, endothelial proliferation ^32,33, and vasodilation ^34,35. VEGF-A signaling is also necessary for proper wound healing ^36,37 but can also lead to abnormal vessel permeability ³⁸ or chronic neovascularization ^32,39,40. In previous studies of mustard-induced ocular complications, increased VEGF-A signaling is strongly associated with the development of corneal neovascularization ^41–43. Furthermore, treatment with anti-VEGF therapy for mustard induced corneal injury significantly reduced corneal neovascularization ^44,45. Hence, VEGF-A signaling appears to promote abnormal vascular remodeling in corneas exposed to mustard vapor.

In the lung, few studies have evaluated for VEGF signaling after mustard injury ^46,47. Karami et al.⁴⁶ compared VEGF-A expression levels in bronchoalveolar lavage (BAL) of subjects with previous SM exposure who developed hemoptysis compared with BAL of subjects with SM exposure and without hemoptysis. VEGF-A levels in BAL were not significantly different between the two groups. Of note, Karami et al.⁴⁶ did not recruit unexposed controls for proper comparison of BAL VEGF-A levels to mustard-exposed subjects. More recently, Nejad-Maoghaddam et al. ⁴⁷ assessed for serum gene expression levels of VEGF in subjects with documented long-term pulmonary symptoms following prior SM exposure. Subjects underwent intravenous mesenchymal stem cell (MSC) treatment, and VEGF gene expression was measured for up to 150 days after treatment. Serum VEGF expression was reduced 20 days after MSC treatment, and gradually increased over the 150 days assessed after treatment. Similar to Karami et al.,⁴⁶ control subjects were not included for assessing change in treated versus untreated mustard-exposed subjects. Thus, VEGF measurements in humans both acutely and chronically following mustard exposure remains an unexplored future area of investigation specific to chronic mustard lung pathology.

PlGF is a second proangiogenic cytokine increased in the coculture model after NM vapor exposure. PlGF is a member of the VEGF family and is minimally expressed in the mature lung under homeostatic conditions ^48,49. This cytokine has been most commonly studied in pathologic lung diseases such as bronchopulmonary dysplasia ^50,51, lung cancer ^52–55, and hyperoxia-induced lung injury ^51,56. More recently, investigators have identified PlGF as a key signaling cytokine in hepatopulmonary syndrome ^57,58. Unlike VEGF-A which regulates angiogenesis in homeostasis as well as during lung development, PlGF exclusively regulates angiogenesis in pathologic conditions ⁴⁹. PlGF promotes angiogenesis, or new vessel formation from previous vessels, through two mechanisms: (1) direct activation of VEGF receptor 1 (VEGFR1) and (2) competitive displacement of VEGF-A from VEGF receptor 1 (VEGFR1), thereby promoting signaling of VEGF-A through VEGFR2 ⁵⁸. Different from other VEGF family isomers, anti-PlGF antibodies have been shown to inhibit pathologic angiogenesis with minimal off-target toxicity and without induction of a rescue angiogenic program ⁵⁸. To the best of our knowledge, ours is the first set of studies to identify PlGF as a potential biomarker of NM vapor exposure. Additionally, we speculate that PlGF inhibition may prevent pathologic angiogenesis after high-concentration NM inhalation exposure.

There are some limitations to the current studies. First, real-time NM vapor concentrations were not measured in our in vitro culture system. The vapor exposure assumes a fixed NM concentration for exposure. In vitro NM exposures concentrations can be calculated, knowing the volume within the 24-well culture (3.5 cm³) and assuming all NM evaporated during the 1-h exposure. Second, the NM concentrations chosen for exposure were high (>300 mg/m³), as these concentrations are known to cause severe short and long-term mustard lung morbidity and mortality from previously published in vivo rodent modeling ^6,9,59,60. At such concentrations, the primary cause of acute mortality is airway injury with vascular leak and airway cast formation as supported by our prior in vivo exposures ^9,11,17,61. Lower, noncytotoxic concentrations were not evaluated as lower concentrations are less associated with both acute and chronic morbidity after mustard exposure. Third, NM was used in place of SM due to local and federal restrictions on SM use. Studies are currently underway to validate our findings of increased expression of angiogenic cytokines using a similar in vitro model of SM vapor exposure. Lastly, a mesenchymal or smooth muscle layer of cells was not included in the model design to limit complexity. Future model development may demonstrate the need for additional cellular layers to recapitulate an in vivo exposure of mustard-induced airway injury.

In conclusion, the in vitro model of primary airway epithelial and endothelial cells at the ALI closely models the injury seen after an acute mustard inhalation injury with acute cytotoxicity at increasing NM concentrations. Two important growth factors, VEGF-A and PlGF, related to pulmonary vascular biology and angiogenesis were increased at 24 h after a single-hour NM vapor exposure. Future in vivo and in vitro studies are required to validate the use of VEGF-A or PlGF as a potential biomarker or drug therapy for either acute or chronic vascular pathogenesis after high-concentration mustard inhalation exposure.

Supplementary Material

Supp FigS1

Figure SE1. (A) Total percent (%) surface area of live pulmonary artery endothelial cells (PAECs) at 24 h after NM vapor exposure. (B) Total percent (%) surface area of dead PAECs at 24 h after NM vapor exposure.