Aerosolized vitamin E acetate causes oxidative injury in mice and in alveolar macrophages

Abstract

Although vitamin E acetate (VEA) is suspected to play a causal role in the development of electronic-cigarette, or vaping, product use-associated lung injury (EVALI), the underlying biological mechanisms of pulmonary injury are yet to be determined. In addition, no study has replicated the systemic inflammation observed in humans in a murine EVALI model, nor investigated potential additive toxicity of viral infection in the setting of exposure to vaping products. To identify the mechanisms driving VEA-related lung injury and test the hypothesis that viral infection causes additive lung injury in the presence of aerosolized VEA, we exposed mice to aerosolized VEA for extended times, followed by influenza infection in some experiments. We used mass spectrometry to evaluate the composition of aerosolized VEA condensate and the VEA deposition in murine or human alveolar macrophages. Extended vaping for 28 days versus 15 days did not worsen lung injury but caused systemic inflammation in the murine EVALI model. Vaping plus influenza increased lung water compared with virus alone. Murine alveolar macrophages exposed to vaped VEA hydrolyzed the VEA to vitamin E with evidence of oxidative stress in the alveolar space and systemic circulation. Aerosolized VEA also induced cell death and chemokine release and reduced efferocytotic function in human alveolar macrophages in vitro. These findings provide new insights into the biological mechanisms of VEA toxicity.

INTRODUCTION

Electronic-cigarette (e-cigarette), or vaping, product use-associated lung injury (EVALI) is a clinical syndrome with gastrointestinal and constitutional symptoms followed by the subacute onset of respiratory failure (1, 2). Unlike extrinsic lipoid pneumonia caused by aspiration of large oil globules, the small lipid droplets generated by heating the hydrophobic e-liquid can reach the distal airway and alveoli uniformly across all lung areas. EVALI is a biologically novel exposure and fundamentally different from lipoid pneumonia and now is considered a new category of disease, although it may overlap with coexisting infections or other etiologies of lung injury (3). EVALI impacted more than 2,800 adolescents and young adults, resulting in at least 68 deaths, during the outbreak from 2019 to 2020 (4). Although the number of cases has decreased rapidly since the emergence of the COVID-19 pandemic, there are ongoing reports of cases of EVALI (5, 6).

Notably, vitamin E acetate (VEA) was recovered in bronchoalveolar lavage (BAL) fluid from nearly all patients with EVALI (7). Furthermore, VEA was exclusively detected in illicit tetrahydrocannabinol (THC)-containing products from 2019 and beyond, but not in those from 2018 before the EVALI outbreak (8). Although VEA has been highly associated with EVALI, the presence of VEA in BAL is not a sufficient proof of causation. To address this question, recent studies from our research group and others have shown increased inflammation in the airspaces in a mouse model of EVALI using aerosolized VEA (9–11). However, none of the studies replicated the systemic inflammation leading to gastrointestinal and constitutional symptoms observed in humans with EVALI, and they did not define the possible pathways of VEA-induced lung injury. We have also reported the direct toxicity of aerosolized VEA on primary human alveolar type II (ATII) cells grown in air-liquid interface cultures (11). However, several questions remain unanswered to date, such as the potential additive toxicity of infection in addition to VEA itself, the impact and contribution of resident alveolar macrophages, and the underlying biological mechanisms that cause lung injury (12).

Therefore, this study was designed to address the following questions: 1) Does prolonged exposure to VEA cause systemic inflammation? 2) Does VEA exposure have an additive injurious effect in the presence of pulmonary viral infection? 3) What are the underlying biological mechanisms of systemic and pulmonary toxicity? 4) What is the impact of aerosolized VEA on alveolar macrophages in mice and humans?

MATERIALS AND METHODS

Aerosol Generation and Exposure Systems

VEA was aerosolized as previously described (11). Briefly, VEA (all rac α-tocopheryl acetate, >96%, Sigma) was aerosolized using a 1.5-Ω ceramic coil (Shenzhen Ocity Times Technology) at 9.4 W with a Gram Universal Vaping Machine (Gram Research). The VEA was aerosolized with an atomizer and aspirated into a syringe (80 mL). Then the aerosol was then injected into the chamber. This process mimics the equivalent of one 80 mL puff of smoke or vape by a person.

Following an initial fill in 1 min of the vaping chamber (25 L) with 10 puffs of aerosolized VEA (80 mL/puff), one puff was injected into the chamber every 30 s for 1 h. The chamber was evacuated at a constant rate of 2.0 L/min during the exposure using a calibrated flowmeter (Dwyer) to draw a mixture of fresh aerosol and ambient air into the chamber.

Experimental Animals

All experimental procedures were performed under protocols approved by the University of California San Francisco (UCSF) Institutional Animal Care and Use Committee. Group size was determined to ensure adequate statistical power based on our extensive experience with models of acute lung injury (11, 13). Adult 9- to 11-wk-old female C57BL6 mice were purchased from the National Cancer Institute and exposed to aerosolized VEA in the vaping chamber for 1 h twice daily for up to 28 days, and euthanized either 12 h or 6 days after the last exposure. In some experiments, following 6 days of exposure to VEA, the mice were intranasally inoculated with either 400 foci-forming units (FFU) of influenza Puerto Rico/8/34 (PR8) or PBS control 12 h after the last exposure, then euthanized 7 days after the inoculation.

Following overdose of ketamine, mice underwent bilateral thoracotomy and were exsanguinated by right ventricle puncture. Bilateral lungs were harvested and excess extravascular lung water (ELW, pulmonary edema in the interstitial and air spaces above the level in normal mice of the same size) was measured as previously described (14). Separate animals underwent tracheal cannulation and serial bronchoalveolar lavage (BAL) with 250 µL of saline twice (total 500 µL). To count BAL total cells, 10 µL of BAL fluid was mixed 1:1 with trypan blue and cell counts were measured using Corning cell counter (Cytosmart, Eindhoven, the Netherlands). Another 100 µL of BAL was cytocentrifuged with CytoSpin 3 (Thermo Fisher Scientific, Waltham, MA) and stained with Hema 3 solution (Fisher Scientific, Pittsburgh, PA), then differential cell counts were performed at ×40 magnification by an observer blinded to treatments. Macrophages were imaged using Nikon Eclipse 80i microscope (Tokyo, Japan) equipped with SPOT RT CMOS camera 5.03 (Sterling Heights, MI) and the diameters of at least 150 cells from each group were measured manually using SPOT imaging software (SPOT imaging). BAL protein concentration was measured with the BCA protein assay (Thermo Fisher Scientific, Waltham, MA). Lungs were intratracheally fixed with 700 µL of 4% paraformaldehyde (PFA) and submerged in 4% PFA for ∼24 h, embedded in paraffin, and the left lung and right inferior lobe were sectioned for histological evaluation by a blinded observer.

Primary Human Alveolar Macrophages

Primary human alveolar macrophages (AMs) were isolated from adult human lungs (Asian female, 31-yr old, nonsmoker; White male, 66-yr old, nonsmoker) declined for transplantation (15). All experiments using cadaver human lung tissue were approved by the UCSF Biosafety Committee. BAL fluid recovered from a human lung lobe free of gross consolidation or injury was passed through Corning 70-µm cell strainer (Corning, NY). Five hundred thousand AMs were cultured in RPMI with 10% fetal bovine serum (FBS) in 24-well plate (Corning) for more than 48 h. The purity of freshly isolated AMs from recovered BAL was 94.6 ± 2.9% (n = 4 subjects). We have also verified that cultured BAL cells were almost exclusively AMs by two methods: flow cytometry (99.4 ± 0.4%, n = 3 subjects) and cytospin differential cell count (100%, n = 1 cytospin sample). For lactate dehydrogenase (LDH) and chemokine assays, medium was changed to serum-free before starting exposure to mitigate the effects of serum on measurements. For the morphology and functional assays, cells were incubated with media with 10% FBS. AMs were exposed to VEA aerosol inside an exposure chamber in the incubator set at 37°C and 5% CO₂ for 1 h daily for three consecutive days. Culture media were not exchanged throughout the course of exposure. Following 3 days of exposure, culture supernatant was collected for LDH and chemokine measurements. For morphological evaluation, cells were detached by Accutase (Sigma), cytocentrifuged, and imaged using Nikon Eclipse 80i microscope with SPOT camera.

Efferocytosis Assay

All experiments using whole blood from human donors were approved by the institutional review board of University of California San Francisco (UCSF 10-04691). Efferocytotic function of alveolar macrophages was analyzed as described previously (16). Whole blood (5 mL) donated by a healthy adult volunteer (Asian male, 42-yr old, nonsmoker) was loaded onto 5 mL of Polymorphprep (Abbott Diagnostics Technologies AS, Oslo, Norway) and centrifuged at 500 g for 30 min at room temperature. Polymorphonuclear leukocytes (PMNs) were harvested from the lower leukocyte band, resuspended in 0.45% saline to restore normal osmolarity, and washed once with PBS. Red blood cells (RBCs) were lysed with Lysing buffer (BD Bioscience, San Jose, CA), PMNs were washed once again with PBS, resuspended in RPMI, and counted to check purity (>90%). Then PMNs were isolated by centrifugation (200 g, 10 min) and resuspended at 4 × 10⁶ cells/mL in RPMI with 10% FBS. Four milliliters of PMNs in solution were mixed with 15 µg of CellTracker Deep Red (Thermo Fisher Scientific) dissolved in 20 µL of DMSO. After incubation for 45 min at 37°C and 5% CO₂, labeled PMNs were centrifuged, resuspended in serum-free RPMI at concentration of 2 × 10⁶/mL, and incubated for 24 h to induce apoptosis. AMs exposed to VEA or control were detached using Accutase (Sigma), washed with PBS, counted, and resuspended in RPMI with 10% FBS at a concentration of 250 × 10³ cells. Some AMs were pretreated with 5 µg/mL Cytochalasin D (Thermo Fisher Scientific) for 30 min. AMs were centrifuged, resuspended in 500 µL of CellTracker-labeled PMN solution (AM:apoptotic PMN = 1:4), and incubated for 2 h at 37°C and 5% CO₂ to induce efferocytosis. Following incubation, cells were placed on ice, then 5,000 AMs counted with LSR Fortessa Flow Cytometer (BD Bioscience), and analyzed using FlowJo software. Controls containing “AM only” or “PMN only” were used to set gates on scatter plot and the “PMN only” control was also used to set a gate on APC channel to detect labeled PMNs. The efferocytosis index was calculated as the proportion of APC-positive macrophages in AM + PMN subtracted by those in AM + PMN + cytochalasin D (i.e., PMNs attached to the surface of AMs).

Oil Red-O Staining

Primary human AMs cytocentrifuged on a charged slide were briefly rinsed with 60% isopropanol and stained with freshly prepared Oil Red O working solution (Sigma). Cell nuclei were counterstained with HEMA 3 solution II (Fisher Scientific) and briefly rinsed with deionized water. Macrophages were imaged using Nikon Eclipse 80i microscope equipped with SPOT RT CMOS camera 5.03.

ELISA and Multiplex Protein Analysis

LDH in culture supernatant was measured with CyQUANT LDH cytotoxicity assay (Invitrogen, Thermo Fisher Scientific) using an Epoch microplate spectrophotometer (BioTek, Winooski, VT). Protein biomarkers in mouse BAL and plasma were measured with Luminex (Luminex Corp.) using a ProcartaPlex 26 plex kit [Eotaxin, granulocyte-macrophage colony-stimulating factor (GM-CSF), keratinocyte chemoattractant (KC), interferon (IFN)-γ, interleukin (IL)-1β, IL-10, IL-12p70, IL-13, IL-17A, IL-18, IL-2, IL-22, IL-23, IL-27, IL-4, IL-5, IL-6, IL-9, IFN-inducible protein (IP)-10, monocyte chemotactic protein (MCP)-1, MCP-3, macrophage inflammatory protein (MIP)-1α, MIP-1β, MIP-2, regulated upon activation, normal T cell expressed and secreted (RANTES), and tumor necrosis factor (TNF)-α] from Thermo Fisher Scientific. Chemokines in primary human AMs culture media were measured with Luminex using a ProcartaPlex multiplex kit [MIP-1α, stromal cell-derived factor (SDF)-1α, C-X-C ligand (CXCL)-10, IL-8, Eotaxin, RANTES, MIP-1β, MCP-1, growth-related gene product (GRO)-α] as per manufacture’s protocol (Thermo Fisher Scientific).

Measurement of VEA, α-Tocopherol, and Malondialdehyde

Murine BAL fluid was centrifuged and cell pellets were flash frozen at −80°C. Cells were extracted using one-phase solvent system (25:10:65, vol/vol/vol of methylene chloride:isopropanol: methanol, with 20 µg/mL butylated hydroxyl toluene). Cells were homogenized with 0.5-mm zirconium oxide beads (Next Advance Inc., Troy, NY) using a counter-top bullet blender for 1 min, and the homogenates were centrifuged at 4°C at 15,000 g for 10 min. Supernatant was transferred to an autosampler vial and run by Sciex Tri-TOF 5600 MS/MS mass spectrometer. Samples were delivered to the source at 50 µL/min using isocratic flow acetonitrile:water (6:4) with 10 mM ammonium formate and 0.1% formic acid. A second isocratic pump delivered a solution of isopropanol:acetonitrile (9:1) with 10 mM ammonium formate and 0.1% formic acid. The parameters for mass spectrometry were set as described (17). The condensate in the tubing in the vaping system after 1 h of exposure was collected to analyze the breakdown products generated during the vaping process. α-Tocopherol (Vit. E), α-tocopheryl quinone (VEQ), and α-tocopheryl acetate (VEA) as well as cellular cholesterol were quantitated by comparison with their external standards. Cellular α-tocopherol-related compounds are reported per cholesterol. Plasma malondialdehyde (MDA) was measured following alkaline hydrolysis; after cooling, samples were acidified and thiobarbituric acid (TBA) reagent was added. The MDA(TBA)₂ adduct was analyzed using an isocratic mobile phase, a 2695 HPLC system (Milford, MA), and detected by fluorescence (excitation/emission 532/533 nm), as described previously (18). Quantitation was done using an external standard of 1,1,3,3-tetraethoxypropane (Sigma) prepared using the same method as the samples.

Statistics

All data were tested for normality with Shapiro–Wilk tests using Prism 9.0 (GraphPad, La Jolla, CA). Statistical significance was determined by one-way ANOVA with Tukey’s post hoc correction for multiple comparisons or unpaired t tests for normally distributed parameters, or by Kruskal–Wallis test with Dunn’s multiple-comparison test or Mann–Whitney test for skewed parameters. All data are presented as means ± standard deviation (SD).

RESULTS

Prolonged VEA Exposure Induced Both Pulmonary and Systemic Inflammation in Mice

To investigate the effects of prolonged exposure to VEA similar to what might be the case for human users, we exposed mice to aerosolized VEA for 1 h twice daily for up to 28 days. Pulmonary edema measured by excess extravascular lung water was significantly increased following 15 days of exposure and remained elevated after 28 days (Fig. 1A). Alveolar-capillary barrier permeability, as measured by BAL protein concentration, increased by day 6, remained elevated, then decreased by 28 days of exposure (Fig. 1B). BAL macrophage/monocyte counts (Fig. 1D) increased gradually through 15 days of exposure and then decreased significantly by day 28. By contrast, BAL neutrophils (Fig. 1C) remained elevated through 28 days. Of note, the diameters of the macrophage/monocytes were significantly increased by day 6, and remained increased at 15 and 28 days (Fig. 1J) with vacuolation observed in substantial proportion of cells (Fig. 1I). BAL monocyte chemokine MCP3, neutrophil chemokine KC, and proinflammatory cytokine IL6 were uniformly elevated at 28 days relative to control (Fig. 1, E–G). IL10 was nonsignificantly elevated at 28 days, suggesting a trend toward an increased anti-inflammatory compensatory response by day 28 (Fig. 1H). Representative histology photomicrographs (Fig. 1K) showed mixed acute and chronic inflammation in the bronchovascular bundle of distal bronchioles in a bronchiolocentric pattern and diffuse foamy macrophages filling the distal alveolar spaces at 15 days and 28 days. Higher magnification showed neutrophil infiltration (arrows) in thickened septa (arrowheads), suggesting that extended exposure leads to prolonged inflammation with lung injury. Histological analysis (Fig. 2A) showed increasing amounts of proximal perivascular cuffing by lymphocytes (arrowhead).

Figure 1.

Dose-dependent pulmonary toxicity by aerosolized vitamin E acetate (VEA) reaches plateau during prolonged exposure and does not lead to more severe lung injury. Nine-to-eleven-week-old C57BL/6 mice were exposed to aerosolized vitamin E acetate (VEA) for 1 h twice daily for up to 28 days. Following 6, 15, or 28 days of exposure, mice were euthanized and assessed for endpoints. A: excess extravascular lung water (ELW) was dose-dependently increased up to 15 days and did not increase further at 28 days (n = 5/time point, P < 0.0001 ANOVA). In contrast, bronchoalveolar lavage (BAL) protein concentrations increased significantly by 6 days (B), plateaued at 15 days, then decreased at 28 days (P < 0.0001 ANOVA). BAL neutrophils numbers were increased significantly at 15 days (P = 0.014 ANOVA) (C), BAL monocyte/macrophage had an increasing trend relative to control but significantly decreased at 28 days (P = 0.054 ANOVA) (D). E–H: BAL monocyte chemokine monocyte chemotactic protein-3 (MCP3), neutrophil chemokine keratinocyte chemoattractant (KC), and proinflammatory cytokine interleukin (IL) 6 were uniformly elevated at 28 days relative to control (E–G). Although not statistically significant, IL10 had trend to be elevated at 28 days, suggesting increased anti-inflammatory compensation mechanism by day 28 (H). I: representative cytology photomicrograph (×600 magnification) showed vacuolated macrophages after 28 days of exposure with (J) significant increase in diameter for all time points relative to the control, P < 0.0001 by ANOVA. K: representative histology photomicrographs revealed similar degrees of mixed acute and chronic inflammation in the bronchovascular bundle in a bronchiolocentric pattern (top, ×40 magnification) and thickened septa, foamy macrophages, and proteinaceous debris filling alveolar spaces (bottom, ×600 magnification) at 15 and 28 days. n = 10/treatment groups, 5 for ELW measurement and 5 for BAL and histology. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by Tukey’s multiple-comparison test (MCT) following ANOVA.

Figure 2.

Prolonged exposure to vitamin E acetate (VEA) aerosol leads to perivascular cuffing by mononuclear cells and increased markers of endothelial injury and systemic inflammation. A: representative photomicrographs (control, left column; VEA 15 days, middle; VEA 28 days, right column) demonstrated mononuclear cell aggregates around blood vessels (arrowheads, bottom) following 28 days of exposure (top, ×40 magnification; bottom, ×400 magnification). B–D: plasma angiopoietin (Ang) 2, plasma intercellular adhesion molecule (ICAM) 1, and matrix metalloproteinase (MMP) 8 were all significantly increased following 28 days of exposure to VEA. n = 5/treatment groups. P = 0.0001, = 0.0003, and = 0.0059, respectively, by ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001 by Tukey’s MCT. MCT, multiple-comparison test.

By day 28, VEA exposure caused a significant increase of 1) plasma angiopoietin 2 (Ang2), an established biomarker of endothelial injury in human acute respiratory distress syndrome (ARDS) (19) (Fig. 2B), 2) intercellular adhesion molecule 1 (ICAM1) (20) (Fig. 2C), and 3) matrix metalloproteinase 8 (MMP8), a biomarker that predicts mortality in systemic inflammation by sepsis (21) (Fig. 2D). Thus, the inflammatory effects of aerosolized VEA exposure were not limited to the pulmonary compartment and mirrored the systemic inflammation that has been reported in human EVALI.

VEA Predisposes Murine Lungs to Greater Injury upon Influenza Infection

Experimental mice were inoculated with 400 FFU of Puerto Rico/8 (PR-8) influenza following 6 days of VEA exposure, whereas control mice were inoculated with PR-8 without prior VEA exposure (Fig. 3A). Compared with mice inoculated with influenza alone, mice exposed to 6 days of aerosolized VEA before infection lost significantly more body weight (Fig. 3B) and had significantly higher ELW (Fig. 3C). BAL protein (Fig. 3D) and total cell counts (Fig. 3E) were similar between the groups.

Figure 3.

Exposure to vitamin E acetate (VEA) aerosol augment lung injury caused by influenza infection. A: both mice exposed to VEA for 6 days and unexposed mice were inoculated with 400 foci-forming units (FFUs) of influenza Puerto Rico/8/34 (PR8). On 7 days postinfection (dpi), mice were euthanized and assessed for endpoints. B–E: body weight loss from inoculation (0 dpi; B) and extravascular lung water (ELW; C) were significantly higher in VEA-exposed mice, whereas bronchoalveolar lavage (BAL) protein (D) and total cell (E) counts were not different between groups. n = 10/treatment groups; 5 for ELW measurement and 5 for BAL. **P < 0.01 by unpaired t test.

Morphologic Alterations of Alveolar Macrophages in the Resolution Phase with Reduced Inflammation

It is not known whether lungs have the ability to remove hydrophobic compounds such as VEA from the airspaces. To answer this question, we next investigated lung injury resolution kinetics. Mice were exposed to VEA for 1 h twice daily for six consecutive days, then euthanized 12 h or 6 days after the last exposure (Fig. 4A). After 6 days of resolution time, there was persistent lung edema (Fig. 4B). By contrast, BAL protein (Fig. 4C) and total cells (Fig. 4D) tended to decrease; the numbers of BAL neutrophils (Fig. 4E) returned nearly to baseline levels. Although the number of macrophages remained elevated (Fig. 4F), their microscopic appearance at 6 days suggests that VEA was partially cleared from cytoplasm, as shown in representative photomicrographs (Fig. 4G) and by the significantly smaller macrophage diameters after 6 days (Fig. 4H). In addition, chemokines for neutrophils [keratinocyte-derived chemokine (KC) Fig. 4I], and monocytes [monocyte-chemotactic protein (MCP)-3, Fig. 4J] as well as the potent inflammatory cytokine IL-6 (Fig. 4K) were significantly lower after 6 days of resolution.

Figure 4.

Inflammation caused by vitamin E acetate (VEA) attenuates over time with morphological changes in alveolar macrophage. A: following 6 days of exposure to VEA aerosol, some mice were euthanized and analyzed 12 h after the last exposure for endpoints while the others were allowed to recover for 6 days to study the resolution kinetics. B: extravascular lung water (ELW) remained elevated after the 6-day recovery period (P < 0.0001, ANOVA), whereas bronchoalveolar lavage (BAL) protein (P < 0.0001, ANOVA) (C) and BAL total cells (P = 0.0185, ANOVA) (D) tended to decrease. By marked contrast, BAL neutrophils were decreased after 6 days (E), (*P = 0.0091 by Kruskal–Wallis test). Despite no changes in BAL macrophage/monocyte counts (F), representative photomicrographs (×600 magnification) (G) show that their diameters were significantly decreased (P < 0.0001 by Kruskal–Wallis test) (H) suggesting that phagocytosed VEA was hydrolyzed. I–K: BAL keratinocyte-derived chemokine (KC) (P < 0.0001, ANOVA), monocyte-chemotactic protein (MCP)-3 (P < 0.0001 by Kruskal–Wallis test), and interleukin (IL)-6 (P < 0.0001 by Kruskal–Wallis test) were uniformly decreased following recovery period. n = 10/treatment groups, 5 for ELW measurement and 3–5 for BAL. *P < 0.05, **P < 0.01, ****P < 0.0001 by Tukey’s MCT for normally distributed data or Dunn’s MCT for skewed data. MCT, multiple-comparison test.

VEA Was Hydrolyzed and Oxidized in Airspace Cells with Evidence of Alveolar and Systemic Oxidative Stress

Mass spectrometry of aerosolized VEA condensate demonstrated a similar pattern to the VEA standard, suggesting that VEA was not pyrolyzed to break down compounds by standard vape settings (Fig. 5A) and was incorporated into cells in its original form. Next, to study the intracellular metabolism of VEA in airspace cells, we collected the cells from murine BAL. Following 3 days of exposure to aerosolized VEA, mice were euthanized soon after the last exposure, and BAL was recovered. We measured cellular VEA, its hydrolysis product α-tocopherol (Vit. E), and the α-tocopherol oxidation product α-tocopheryl quinone (VEQ) (22). Notably, the most abundant tocopherol-related compound was not VEA (Fig. 5B) but Vit. E (Fig. 5C), indicating that airspace cells actively hydrolyzed the VEA to release α-tocopherol. Interestingly, we also found a considerable quantity of VEQ (Fig. 5D), the oxidized form of Vit. E, the presence of which indicates increased oxidative stress in the airspaces (23). In accordance with evidence of increased oxidative stress in the airspaces, we also found a dose-dependent increase of plasma malonaldehyde (MDA) (Fig. 5E), an end product of lipid peroxidation, indicating systemic oxidative stress in the mice exposed to VEA aerosol.

Figure 5.

Vitamin E acetate (VEA) was hydrolyzed within airspace cells with evidence of systemic oxidative stress. A: mass spectrometry (time of flight) spectra from 70 to 1,700 mass units of aerosolized VEA condensate demonstrated a similar mass/charge (m/z) and mass transition pairs as seen in the authentic VEA standard (m/z, 473.4→207.1). Following 3 days exposure to VEA aerosol, mice were lavaged and cell pellets from bronchoalveolar lavage (BAL) fluid were analyzed for VEA (α -tocopheryl acetate) (B), α-T (α-tocopherol, Vit. E) (C), and α-TQ (α-tocopheryl quinone VEQ) (D). Shown are the ratios of the indicated compound to cholesterol (mmol/mol) to account for variable cells counts per BAL sample; column heights equal means ± SD with individual samples indicated (****P < 0.0001 by unpaired t test of log-transformed data). E: plasma malondialdehyde (MDA), a lipid peroxidization marker, was time-dependently increased following 6 or 15 days of exposure (means ± SD, ANOVA: P < 0.0001, ***P < 0.001, ****P < 0.0001 by Tukey’s MCT). n = 5/treatment groups. MCT, multiple-comparison test.

Human AM Exposure to Aerosolized VEA Induced Cell Death, Altered Morphology and Function, and Increased Inflammation

We next investigated the physiological and pathological impact of aerosolized VEA on human alveolar macrophages. Primary human alveolar macrophages were isolated from human lungs, cultured for 48 h, and exposed to VEA aerosol for 1 h daily for three consecutive days. VEA was directly cytotoxic to macrophages, measured by culture supernatant LDH (Fig. 6A). Efferocytotic ability was also significantly decreased in VEA-exposed macrophages (Fig. 6B). Although VEA was the most abundant tocopherol-related product, Vit. E was increased significantly relative to the control (Fig. 6C), suggesting that human alveolar macrophages are also capable of hydrolyzing VEA. Representative photomicrographs showed vacuolated multinuclear macrophages following 3-day exposure to VEA aerosol with positive Oil Red-O staining (Fig. 6D). VEA exposure also increased the release of monocyte, neutrophil, and lymphocyte chemokines relative to control conditions (Fig. 6E).

Figure 6.

Vitamin E acetate (VEA) exposure to human primary alveolar macrophages induced cell death, impaired function, and increased chemokine release. Primary human alveolar macrophages cultured in media were exposed to aerosolized VEA for an hour daily for three consecutive days, which increased lactate dehydrogenase (LDH) concentrations in culture supernatant (A) and impaired efferocytotic function (B), (P = 0.0065 unpaired t test). C: VEA and Vit. E were significantly increased in human alveolar macrophages exposed to aerosolized VEA. D: representative photomicrographs (×600 magnification) revealed vacuolated, multinucleated, and Oil Red O (ORO) positive macrophages after VEA exposure. E: after 3 days of VEA exposure, the culture supernatant contained significantly increased concentrations of monocyte chemotactic protein (MCP)-1, macrophage inflammatory protein (MIP)-1α (but not MIP-1β), stromal cell-derived factor (SDF)-1α, growth-regulated oncogene (GRO)-α, regulated on activation, normal T cell expressed and secreted (RANTES), and eotaxin. n = 3–6/treatment groups. *P < 0.05, **P < 0.01, ***P < 0.001 by unpaired t test.

DISCUSSION

The main findings of the current study are 1) Extended VEA exposure results in systemic inflammation in mice, 2) Exposure to VEA predisposes lungs to additive injury from viral infection, 3) VEA aerosol is phagocytosed by alveolar cells, then hydrolyzed to free tocopherol, likely resulting in oxidative stress in mice, and 4) VEA aerosol is cytotoxic to primary human alveolar macrophages, impairs efferocytotic function, and induces AM chemokine release.

Clinically, 100% of patients with EVALI present with systemic/constitutional symptoms such as fever, chills, and weight loss, whereas 97% present with respiratory symptoms (24). Though rare, some patients with EVALI present exclusively with constitutional symptoms in the absence of pulmonary complaints (25). However, to date, no study has shown systemic inflammation with constitutional findings (e.g., weight loss) in a murine model of EVALI. To recapitulate both the pulmonary and systemic inflammation seen in human patients with EVALI, we extended the VEA exposure up to 28 days in the current study. Although extended exposure did not worsen the lung injury relative to 15 days, we found evidence of endothelial injury and systemic inflammation that mirrors the clinical presentation of human patients. Interestingly, we also found increased systemic oxidative stress measured by plasma MDA following 6 and 15 days of VEA exposure, suggesting that this pathway may partly be responsible for the systemic inflammation in murine model of EVALI. The systemic findings in this model strengthen its translational relevance by reflecting a constellation of symptoms consistent with the clinical picture of EVALI.

Amid the COVID-19 pandemic, the question as to whether aerosolized VEA has additive effect on pulmonary viral infection is of particular interest. A primary goal of this study was to assess the impact of viral infection at a very early stage of lung injury caused by aerosolized VEA. We exposed mice to VEA for 6 days before infection because it is the minimum exposure time for detecting lung edema in these studies (11). We found that mice exposed to aerosolized VEA in addition to a mouse-adapted influenza strain (PR-8) experienced significantly greater body weight loss as compared with mice that were not exposed to VEA. VEA plus viral infection also increased lung injury as measured by lung edema compared with viral infection alone. These results may not reflect the true injurious potential of aerosolized VEA in the presence of viral infection as mice were only exposed to VEA for 6 days in these experiments. Extended exposure to e-cigarettes can cause significant morphological and functional changes in murine alveolar macrophages and predispose to higher mortality in influenza A infection (26), suggesting that chronic exposure may have even more harmful effects than we found with this relatively limited exposure time.

Since the 2019 outbreak of EVALI, several clinical and experimental studies investigated the putative toxins produced by heating VEA. Recent studies (27, 28) demonstrated that ketene can be generated by heating VEA either with extremely high power (50 W) or with dry-hit conditions in which the atomizer is dried out and produces extremely high temperatures (29). These conditions, however, do not reflect typical use. Another study identified numerous respiratory toxicants detected in the liquid and vapor phases of substances in counterfeit THC-containing cartridges collected from patients with EVALI, suggesting that exposure to a combination of chemicals might contribute to the development of EVALI (30). Interestingly, Jiang et al. (31) showed that some small chemicals such as acetone, trimethyl dodecanol, and duroquinone or durohydroquinone can be generated by aerosolizing VEA with standard vape settings (9.3 W). Duroquinone and durohydroquinone are significant contributors to reactive oxygen species generation (31). Because our mass spectrometry analyses studied the condensate accumulated in the tubing of a semiclosed plexiglass chamber evacuated by vacuum and did not analyze gas-phase substances, our results cannot determine the responsible substances from the inhalant. However, the current study demonstrates that VEA remained in its original form and was not pyrolyzed to vitamin E plus ketene under a standard vape setting, and that intact VEA caused significant lung injury in mice and had biological impacts on human alveolar macrophages, suggesting that VEA itself is toxic.

Intracellular enzymatic hydrolysis of VEA in alveolar cells is not well studied. In the digestive tract, VEA is absorbed as tocopherol, secreted in chylomicrons, and transported via the lymph to the liver where hepatocyte α-tocopherol transfer protein (α-TTP) mediates the secretion into the plasma. VEA is not absorbed enterally without hydrolysis, thus protecting against excessive systemic absorption of ingested VEA (32). In contrast, there is no such protective system in the alveolar space. Nonspecific esterases that can hydrolyze α-naphthyl acetate have been located ultrastructurally outside human alveolar macrophages, suggesting that they function as ectoenzymes (33). However, it is not known if these esterases can hydrolyze VEA. To our knowledge, our result is the first evidence that alveolar macrophages have the ability to hydrolyze VEA to vitamin E.

Vitamin E is the most potent lipid-soluble antioxidant in human plasma and tissue. When carbon-centered radicals react with oxygen, a peroxyl radical (reactive oxygen species) is generated. The peroxyl radical is scavenged by vitamin E, resulting in a tocopheroxyl radical, preventing further lipid peroxidation. This tocopheroxyl radical is restored to its unoxidized form by other antioxidants such as vitamin C, ubiquinol, or thiols such as glutathione. Alternatively, the tocopheroxyl radical can be further oxidized to the tocopheryl quinone under oxidative stress (34). It is well known that some of the health effects of cigarette smoke and e-cigarette aerosol are driven by oxidative stress (35, 36). For example, patients who receive oxygen therapy have greater quantities of tocopheryl quinone in BAL than those who do not receive oxygen therapy (37). Furthermore, alveolar macrophages from cigarette smokers have a higher amount of tocopheryl quinone than nonsmokers (38). The presence of MDA in murine plasma in this study indicates oxidative damage not only in the alveolar space but also in the systemic circulation. Collectively, our results indicate that EVALI is mediated at least in some part by alveolar and systemic oxidative stress.

Alveolar macrophages play an important role in maintaining the local environment in the airway lumen and alveolar space by orchestrating the innate and adaptive immune responses, regulating lipid and iron homeostasis, and clearing apoptotic neutrophils in lung injury (efferocytosis) (39). Although the precise biological mechanism is not clear, our study identified that VEA aerosol induced cell death, increased chemokine release, and attenuated efferocytosis in primary alveolar macrophages. Numerous studies have demonstrated that apoptosis of primary human alveolar macrophages is induced by oxidative stress from cigarette and e-cigarette smoke exposure (40, 41). Cell death may be also attributed to ferroptosis, a form of caspase-independent regulated cell death, as it is characterized by lipid peroxidation and disruption of plasma membrane that shares same biological pathways seen in VEA-induced injury (42). Recent studies have described the role of ferroptosis in pulmonary disease (43, 44) and cigarette smoke toxicity (45). Additional studies have shown decreased efferocytotic ability of primary macrophages in human patients with chronic obstructive pulmonary disease (COPD) (46) and murine bone marrow-derived macrophages exposed to e-cigarette (47). The mechanisms of reduced efferocytosis are multifactorial and may be attributed to altered extracellular matrix proteins (48), defects in phagosome and lysosome fusion (49), and increased oxidative stress (47). A recent study reported that efferocytosis was significantly impaired in patients with sepsis with ARDS as compared with those without ARDS, suggesting that macrophages polarized to a proinflammatory phenotype have reduced efferocytosis (16). Taken together, reduced efferocytosis in macrophages exposed to aerosolized VEA may be mediated by both oxidative stress and inflammation. Moreover, it is possible that decreased efferocytotic activity of alveolar macrophages results in the accumulation of apoptotic neutrophils and further augments inflammation, resulting in a proinflammatory cascade in EVALI.

Our study has several limitations. First, we could not perform THC experiments due to federal research restrictions, which limits our ability to study illicit products vaped by patients with EVALI. Second, as mentioned previously, we could only analyze the condensate of aerosolized VEA and could not study the gas phase, which limits the assessment of additional potentially responsible chemicals via mass spectrometry. Third, we did not directly measure the level of oxidative stress in AMs, although the presence of α-tocopheryl quinone in AMs indicates intracellular oxidative stress. Fourth, only female mice were used in the studies and gender difference was not tested. Finally, antioxidants to inhibit oxidative stress were not tested in this study; therefore the hypothesis that oxidative stress mediates pathological consequences requires further testing.

In conclusion, extended exposure to VEA aerosol induced systemic inflammation in mice, similar to that seen in human patients with EVALI, and produced evidence of systemic oxidative stress. In addition, murine alveolar macrophages have the ability to hydrolyze VEA, leading to increased vitamin E that is subsequently oxidized to vitamin E quinone under oxidative stress. VEA aerosol induces morphological and functional changes in human alveolar macrophages and induces chemokine release, which may in part explain the underlying biological mechanism in EVALI. To our knowledge, these findings provide the first concrete evidence of both systemic inflammation and oxidative stress as possible pathways for the injurious effect of VEA as a causative agent in EVALI.

DATA AVAILABILITY

The data that support the findings of this study will be made available upon reasonable request from the corresponding author.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute (NHLBI) Grants U54 HL147127 and R01 HL134828. Research reported here was supported in part by the Diabetes Research Center Grant NIH P30 DK063720.

DISCLOSURES

C. S. Calfee has received grants and personal fees from Bayer and Roche-Genentech, and personal fees from Quark Pharmaceuticals, Vasomune, Gen1e Life Sciences; and Janssen. M. A. Matthay has received grants from Roche-Genentech and Bayer Pharmaceuticals, personal fees from Boehringer Ingelheim, Novartis Pharmaceuticals, Citius Pharmaceuticals, Pliant Therapeutics, Johnson & Johnson Pharmaceuticals, and Gilead Pharmaceuticals. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

S.M., M.G.T., C.S.C., J.E.G., and M.A.M. conceived and designed research; S.M., S.W.L., J.C., X.F., and J.E.G. performed experiments; S.M., M.G.T., S.W.L., X.F., M.M., K.D.W., K.D.J., C.S.C. J.E.G., and M.A.M., analyzed data; S.M., M.G.T., X.F., M.M., K.D.W., K.D.J., C.S.C., J.E.G., and M.A.M. interpreted results of experiments; S.M., M.G.T., and M.A.M. prepared figures; S.M. drafted manuscript; S.M., M.G.T., X.F., K.D.W., C.S.C. J.E.G., and M.A.M., edited and revised manuscript; S.M., M.G.T., S.W.L., J.C., X.F., M.M., K.D.W., K.D.J., C.S.C., J.E.G., and M.A.M. approved final version of manuscript.