Emissions from plastic incineration induce inflammation, oxidative stress, and impaired bioenergetics in primary human respiratory epithelial cells

Abstract

Inhalation exposure to plastic incineration emissions (PIEs) is a problem of increasing human relevance, as plastic production and waste creation have drastically increased since mainstream integration during the 20th century. We investigated the effects of PIEs on human nasal epithelial cells (HNECs) to understand if such exposures cause damage and dysfunction to respiratory epithelia. Primary HNECs from male and female donors were cultured at air–liquid interface (ALI), and 16HBE cells were cultured on coverslips. Smoke condensates were generated from incineration of plastic at flaming (640°C) and smoldering (500°C) temperatures, and cells were subsequently exposed to these materials at 5–50 μg/cm² concentrations. HNECs were assessed for mitochondrial dysfunction and 16HBE cells for glutathione oxidation in real-time analyses. HNEC culture supernatants and total RNA were collected at 4-h postexposure for cytokine and gene expression analysis, and results show that PIEs can acutely induce inflammation, oxidative stress, and mitochondrial dysfunction in HNECs, and that incineration temperature modifies biological responses. Specifically, condensates from flaming and smoldering PIEs significantly increased HNEC secretion of cytokines IL-8, IL-1β, and IL-13, as well as expression of xenobiotic metabolism pathways and genes such as CYP1A1 and CYP1B1 at 5 and 20 μg/cm² concentrations. Only 50 μg/cm² flaming PIEs significantly increased glutathione oxidation in 16HBEs, and decreased respiration and ATP production in HNEC mitochondria. Impact Statement: Our data reveal the impact of incineration temperatures on biological outcomes associated with PIE exposures, emphasizing the importance of temperature as a factor when evaluating respiratory disease associated with PIEs exposure.

Graphical Abstract

Open in new tabDownload slide

The vast increase of plastic waste in the 20th and 21st century poses enormous potential human and environmental hazards. Since its introduction in the 1950s, 9.1 billion tons of plastic has been produced, with only approximately 9% of it being recycled, resulting in over 8 billion tons of plastic waste introduced into the environment (Geyer et al., 2017; Walker and Fequet, 2023). The United Nations reports that over 440 million tons of plastic waste is produced every year globally; a number that is expected to increase as global plastic production is forecasted to double in the next 20 years (Geyer et al., 2017; Lebreton and Andrady, 2019; Walker and Fequet, 2023). As plastic is resistant to biodegradation, nonrecycled plastic has historically been disposed of by landfill storage (79%) or incineration (12%). Although landfill storage and breakdown of plastics is the source for microplastics entering the environment (Amato-Lourenço et al., 2020; Dong et al., 2020; Xu et al., 2019; Zimmermann et al., 2019), over 70 million tons of plastic were incinerated without regulation in 2016 alone, introducing over 500 million tons of toxic aerosols into the air (Geyer et al., 2017; Walker and Fequet, 2023; Wu et al., 2021). In most cities and developed countries, incineration techniques have been adopted to permanently eliminate plastic waste (Han et al., 2018; Johnson, 2016; Nanda and Berruti, 2021; Wang et al., 2020). However, in many underdeveloped countries and rural communities, waste collection and disposal services are not available, making incineration of household trash the primary means of plastic and other waste disposal (Bardales Cruz et al., 2023; Christian et al., 2010; Reyna-Bensusan et al., 2019; Sharma et al., 2019; Velis and Cook, 2021). Such incineration of plastics releases toxins such as dioxins, furans, and mercury (Verma et al., 2016), along with large amounts of respirable particulate matter (PM) (Islam et al., 2022; Simoneit et al., 2005; Wu et al., 2021) and high concentrations of polycyclic aromatic hydrocarbons (PAHs) (Cabanes et al., 2020; Conesa et al., 2021; Li et al., 2001), which have been established as harmful to human respiratory health (Lelieveld et al., 2015; Wu et al., 2021; Yang et al., 2021).

Another example of human exposure to large amounts of plastic incineration emissions (PIEs) is military burn pits. Burn pits are designated areas on military sites for open-air combustion of waste without any standard waste management systems. This open-air waste management system was prevalent in Afghanistan and Iraq from 2001 to 2009 (Kim et al., 2018; McLean et al., 2021) where it was estimated that over 60 000 pounds of solid waste (Long-Term Health Consequences of Exposure to Burn Pits in Iraq and Afghanistan, 2015), including an estimated 17 000 pounds of plastic was burned. Air monitoring data near burn pits have identified airborne chemicals associated with respiratory toxicity, morbidity, and respiratory cancers (Long-Term Health Consequences of Exposure to Burn Pits in Iraq and Afghanistan, 2015; Taylor et al., 2008), and associations between burn pit exposure during deployment and development of self-reported chronic respiratory conditions such as asthma, bronchitis, and sinonasal disease have also been documented (Hill et al., 2022; Long-Term Health Consequences of Exposure to Burn Pits in Iraq and Afghanistan, 2015; McLean et al., 2021; Olsen et al., 2022). This is especially relevant as military personnel have self-reported upper respiratory diseases such as rhinitis and sinusitis after deployment (Hill et al., 2022), which are now recognized by the VA as presumptive conditions of burn pit exposure. However, causal relationships between the inhalation of PIEs and adverse health effects of the nasal mucosa have not been determined. Two temperatures common to burn pit incineration conditions, smoldering (500°C), and flaming (640°C), were utilized to generate PIEs condensates and respiratory exposures (Kim et al., 2021).

Chemical analyses of condensates of PIEs generated under flaming and smoldering conditions revealed high concentrations of toxic nitro- and oxy-PAHs and other EPA priority toxicants, with higher PAH concentrations in flaming condensate PIEs (Kim et al., 2021). Respiratory exposures to these PAHs and other toxicants are associated with adverse pulmonary outcomes and are known to contribute to inflammation, oxidative stress and redox imbalance, DNA damage, and dysregulated development in respiratory cells (Kermani et al., 2021). Human nasal epithelial cells (HNECs) line the nasal cavity and compromise the first line of respiratory defense against respirable toxins, generating a physical barrier function (Scherzad et al., 2019). Compromised nasal mucosa is associated with various inflammatory sinonasal diseases, and PM is associated with the initiation and aggravation of upper respiratory diseases including rhinitis and sinusitis (Mo, 2019). We hypothesized that exposure to PIEs would induce inflammation, oxidative stress, and cellular bioenergetics impairment in human respiratory epithelia, and that PIEs incineration temperature (smoldering and flaming) would modify biological responses. In this study, we examined various cellular effects of acute exposures to PIEs on HNECs to model impacts of incinerated plastic emissions on the nasal mucosa.

Materials and methods

Cell culture of HNECs

Primary HNECs were collected from adults as previously described (Müller et al., 2013). Written consent was acquired from healthy, nonsmoking male and female adult study participants for superficial nasal epithelial scrape biopsies with a Rhino-Pro curette (Arlington Scientific, Inc. 96-0900) per protocols approved by the University of North Carolina School of Medicine Institutional Review Board for Biomedical Research. Nasal biopsies were stored on ice in RPMI-1640 medium (Gibco, CAT No. 11875-093) until further processing. Demographic information on donor HNECs used in each exposure including age, body mass index (BMI), and race is provided in Table 1.

Differentiation of HNECs was performed as previously described (Escobar et al., 2021; Müller et al., 2013). Cells from nasal scrape biopsies were expanded in PneumaCult-Ex Plus Medium (StemCell Technologies, Inc., CAT Nos 05041 and 05042) supplemented with hydrocortisone (StemCell Technologies, Inc., CAT No. 07925), antibiotic antimycotic solution (Sigma, CAT No. A5955), and gentamicin reagent solution (Gibco, CAT No. 15750-060) on 12 well PureCol-coated (Advanced Biomatrix, 5005-100ML) cell culture plates (Costar 3512) at passage 0. Cells were then passaged and expanded until passage 2 in 25-cm² tissue culture flasks (Corning, CAT No. 430639). At passage 3, cells were seeded on 12-mm transwell inserts with 0.4-μm pores (Costar, CAT No. 3460) coated with human placental collagen (Sigma, CAT No. C7521-10MG) at a density of 268 000–350 000 cells per well and maintained in PneumaCult-Ex Plus Medium on the apical and basolateral side until confluency in transwells. Once confluency was achieved, cells were then taken to air–liquid interface (ALI), at which apical media was permanently removed, and basolateral media was changed to PneumaCult ALI Medium (StemCell Technologies, Inc., CAT Nos 05002, 05003, and 05006), supplemented with 1% pen strep (Gibco, CAT No. 15140-122), hydrocortisone (StemCell Technologies Inc., CAT No. 07925), and heparin (StemCell Technologies Inc., CAT No. 07980) to promote nasal cell differentiation, which was achieved within 4- to 6-week post-ALI. Fully differentiated HNECs grown at ALI exhibited ciliation, mucus production, and heterogeneous cell presence, therefore mimicking normal pseudostratified respiratory epithelium. Cells were then fed 3 times a week by changing the basolateral media and washed with HBSS⁺⁺ (Gibco, CAT No. 14025-092) until exposure.

Cell culture of 16HBE cells

For all live-cell imaging experiments, the commercially available 16HBE14o- (human bronchial epithelial) cell line was originally obtained from a 1-year-old male and immortalized with SV40 plasmid (Callaghan et al., 2020; Cozens et al., 1994) was utilized from passages 36 to 48. 16HBEs expressing genetically encoded fluorescent reporters for reduction-oxidation sensitive green fluorescent protein (roGFP) previously described here(Masood et al., 2022) were gifted by the laboratory of Dr James Samet (UNC Chapel Hill, North Carolina), and cultured in MEM with Glutamax with 10% FBS and 1% penicillin-streptomycin complete media to 80% confluency in submerged culture. Prior to live-cell imaging, cells were cultured in 96-well plates and then starved in Locke’s Buffer, a minimal buffered salt solution containing no organic compounds prior to imaging. All culture plates and flasks for 16HBE culture were coated with 30 μg/ml bovine type 1 collagen, 0.01% BSA, and 1% human fibronectin in LHC Basal Medium.

PIE condensate preparation

PIE condensates were created through incineration of a mixture of plastics: low-density polyethylene, high-density polyethylene, polyethylene terephthalate, and polystyrene pellets, representative of the prevalent plastic types from the U.S. military waste analyses as well as contributors to domestic municipal solid waste (Aurell et al., 2019; USARCENT AOR Contingency Base Waste Stream Analysis: An Analysis of Solid Waste Streams at Five Bases in the U.S. Army Central (USARCENT) Area of Responsibility, 2013). Plastic incineration utilized a quartz tube automated furnace system described in detail (Kim et al., 2018), and PIEs were collected by a multi-stage cryotrap system of 3 sequential impingers at −10, −50, and −70°C at smoldering (500°C) and flaming (640°C) temperatures representative of open-burn pit incineration conditions. Smoldering emissions used in this study model slower, lower temperature plastic combustion that can release higher levels of carbon monoxide and PM, whereas temperatures used to generate flaming emissions represent more complete combustion at high temperatures of plastic incineration, releasing higher levels of PAHs into the environment. The cryo-trapped PIEs condensate samples were stored under acetone at −80°C to limit degradation until use. PIE samples in acetone were then solvent exchanged into PBS through acetone evaporation and subsequent resuspension in PBS for a PIEs stock concentration of 1 mg/ml for cell toxicity exposures. This smoke sampling method allows for collection of soluble and insoluble fractions in resuspension material, including smoke particles, ions, inorganic elements, and semivolatile organic compounds representative of the chemical components in PIEs in nature.

Dosing justifications

Exposure dosing parameters are based on previous PM dosing justifications described here (Kim et al., 2021) and the following information: The selection of the exposure dosing parameter is based on calculations of peak PM levels at the U.S. military base of 299 μg/m³, and a more realistic estimation of acute burn pit smoke PM exposure at 2.8 mg/m³ which is based of exposure levels of first responders, such as firefighters (Swiston et al., 2008). Using these data with the 160-cm² surface area of the nasal passage (Gizurarson, 2012), dosing calculations were executed as shown in Table 2, although we recognize that a host of factors, including particle size, density, and chemistry as well as individual respiratory volumes and nasal cavity morphology make it difficult to perfectly model such exposures. The PIE condensate dosing used in the studies described here considered both calculated concentrations for a single 4-h exposure, with doses ranging from 5 to 50 μg/cm² for relevant exposures, well under the peak 84 μg/cm² dose and centered around the calculated 8.9 μg/cm² for burn pit data reference relevance. Calculations are based on particle mass which includes both soluble and insoluble fractions contained in the condensate, representing the complex mixtures PIEs create in the environment.

Vitamin E α-tocopherol dosing of 20 µM for pre-PIE exposure was based on measured Vitamin E serum levels in humans, with α-tocopherol serum measurements in adults ranging from <20 to >50 µM (Jiang, 2017).

In vitro PIE exposures

Fully differentiated HNEC cultures at the ALI were exposed to PIE condensates as follows: 1 h prior to exposure, cells were apically washed with 150 µl of warmed (37°C) HBSS⁺⁺, and basolateral media was exchanged for 1 ml of fresh warmed (37°C) media. Condensates from smoldering and flaming PIEs were sonicated, vortexed, and immediately diluted in PBS for cell exposure. For exposure, 150 µl diluted stock PIE condensates achieving doses of 5–20 μg/cm² previously described or 150 µl of PBS vehicle control were added to the apical side and returned to the incubator (37°C, 5% CO₂) for 4 h. Following exposure, desired samples for each experiment were collected, including apical media, basolateral media, and RNA. All collected samples were stored for long-term storage at −80°C. 16HBEs and HNECs in submerged culture experiments were exposed to PIE condensates directly added to culture media for 25 or 50 µg/cm² exposure conditions, with culture media as the vehicle control. 16HBEs in submerged culture for roGFP analyses were directly exposed to PIE condensates while concurrently recording roGFP measurements for the duration of approximately 1 h. HNECs in submerged culture for bioenergetics analyses were similarly exposed to PIEs condensates and then immediately analyzed with the Seahorse Extracellular Flux Modified Cell Mito Stress Test, which took approximately 3 h. For experiments determining the effects of antioxidant exposure prior to PIE exposure, Vitamin E α-tocopherol (Sigma-Aldrich, CAT No. 258024) was diluted from stock solution to 20 µM in PBS (4.03 µg/cm² assuming 100% deposition of Vitamin E in solution on HBE cells) and apically applied to 16HBE cells for 3-h pre-PIE exposure and roGFP measurements. Supplemented PBS was then removed, and cells were washed twice with PBS to ensure removal of all Vitamin E before exposure to the diluted PIE condensates.

Cytokine analysis

IL-8 cytokine levels in basolateral media were measured according to kit instructions using the BD OptEIA Human IL-8 ELISA Set (CAT No. 555244), LLOD: 0.8 pg/ml. For all other cytokine measures, we used a multi-plex ELISA platform (MSD, Rockville, Maryland) chosen for cytokines implicated in respiratory inflammation. Specifically, 20 cytokines were measured per manufacturer instructions using the V-Plex Proinflammatory Panel 1 and Human Cytokine Panel Human Kit (MSD, catalog numbers K15049D-1 and K15054D-2) at a 1:2 dilution. Of the 20 cytokines measured, 10 (VEGF-A [LLOD: 1.12 pg/ml], IL-1β [LLOD: 0.05 p/ml], IL-2 [LLOD: 0.09 pg/ml], IL-4 [0.02 pg/ml], IL-6 [0.06 pg/ml], IL-8 [LLOD: 0.8 pg/ml], IL-10 [LLOD: 0.04 pg/ml], IL-7 [LLOD: 0.12 pg/ml], IL-13 [LLOD: 0.24 pg/ml], and TNF-α [LLOD: 0.04 pg/ml]) had adequate signal measurements and passed quality control, which was defined as having at least 50% of measured samples within the detection signal range. Cytokine signals and concentrations were automatically determined using a MESO QuickPlexSQ 120 instrument and DISCOVERY WORKBENCH 4.0 software, and visualized in Graphpad.

Generation of cDNA and gene expression quantification by qPCR and RNA-seq

RNA was isolated and reverse transcribed as described by us before (Brocke et al., 2022; Escobar et al., 2020; Speen et al., 2016). Samples were submitted to the UNC Center for Gastrointestinal Biology and Disease for high-throughput qPCR gene expression analysis. Gene expression of 24 target genes of interest was profiled using the Fluidigm Biomark HD (Fluidigm Corporation, South San Francisco, California) 192.24 Array system using Taqman primers and probes (see Table 3). Ct values were measured for each sample and gene for further analysis. Differences in gene expression data were generated using the ^ΔΔCt method with normalization to the house-keeping gene (TBP), which was stable across all samples. Housekeeping gene ACTB was also measured but was not stable in all samples and thus was excluded for sample ^ΔΔCt normalization. To compare each pairwise set of samples, the difference in ΔCt was calculated for each gene of interest between samples to determine ^ΔΔCt, and gene expression fold-change was calculated as 2^−ΔΔCt. One gene expression analyte, IL-13, had poor or undetectable amplification across all samples, and was excluded from analysis. Data visualization was done using Graphpad.

For unbiased identification of gene expression profiles, isolated RNA samples were submitted to Azenta Life Sciences (Morrisville, North Carolina) for bulk RNA-seq using the Illumina hiSeq platform. Sequencing configuration utilized the Illumina hiSeq 2×150 bp, producing 350 million read pairs per lane. QC was conducted by measuring RIN: >6.0, ≥50 ng/l and A260/A280: 1.8–2.2. All submitted samples passed such QC metrics. Standard DESeq2 analysis was carried out and provided by Azenta Life Sciences. Sample raw counts from DESeq2 data analyses were downloaded, analyzed, and visualized in R using the EnhancedVolcano package, and differentially expressed genes (DEGs) were selected with a false discovery rate (FDR)-adjusted p-value cutoff of .05 and a log2-fold change of 1. Data visualization was done in R (https://www.r-project.org) using the ComplexHeatmap package, and Quiagen Ingenuity Pathway Analysis (IPA) top canonical pathways function.

16HBE glutathione oxidation live cell imaging analysis

All live cell imaging exposures were conducted as previously described (Masood et al., 2022). 16HBE cells encoding oxidation-sensitive roGFP were imaged over an approximately 1-h time course and glutathione oxidation was measured in nontreated control cells as well as cells treated with either 5, 25, or 50 μg/cm² concentrations of flaming or smoldering PIEs in real time, with PBS as a negative control. Each exposure experiment consisted of 4 distinct intervals: (1) an untreated baseline for 5 min, (2) exposure to PIEs for 25–35 min, (3) exposure to 1 mM H₂O₂ for 5 min as a maximal control to check roGFP integrity, and (4) exposure to 5 mM DTT for 5 min to return the sensor back to baseline. All experiments involved treatment with paired controls within 1 h of exposure. 16HBE cells were monitored and imaged on the Nikon Eclipse C1si spectral confocal imaging system (Nikon Instruments Corporation, Melville, New York) during experiments at excitation wavelengths of 488 and 405 nm, and emissions wavelength at 525 nm. Laser power and exposure time were kept constants for all experiments, and optimization of detector gain was accomplished prior to each experiment and kept constant for the experiment duration. Regions of interest (ROI) were drawn for 10 individual cells within a single well and monitored for the duration of the experiment using the NIS-Elements AR software (Nikon Instruments Corporation, Melville, New York). Data obtained from each experiment were calculated as ratios of the emission intensities at 510 nm for laser excitation at 488 and 405 nm. Ratios were calculated for each cell observed and normalized to the baseline and maximal response for each individual cell. The normalized ratios for individual ROIs (cells) were averaged to compile as 1 sample. The results presented in each graph are the mean±SEM of 3 independent experiments.

Cellular bioenergetics analysis

Cellular bioenergetic parameters on HNECs were assessed using a Seahorse Extracellular Flux Modified Cell Mito Stress Test (Agilent Technologies, Santa Clara, California) based on existing literature and manufacturer guidelines as previously described (Lavrich et al., 2018). Briefly, HNECs were grown to confluence on a Seahorse XF24 cell culture microplate. For PIEs exposure, culture media was replaced with Seahorse XF RPMI media without phenol red (2 mM L-glutamine, 1 mM sodium pyruvate, 1 mM HEPES, and 10 mM glucose, pH 7.4) containing diluted PIEs, and were allowed to rest for 20 min in a non-CO₂ incubator prior to the start of the assay. Final treatment concentrations and injection order are as follows: Port A—1 µM oligomycin; Port B—1.25 µM FCCP; Port C—0.5 µM antimycin A and 0.5 µM rotenone. The assay was run with 3 baseline readings and then mix/wait/measure cycles of 6 min/6 min/6 min for each compound injection on a Seahorse XFe24 Extracellular Flux Analyzer (Agilent Technologies) at 37°C, with a total assay duration time of approximately 3 h. OCR was automatically calculated by the Seahorse system. Immediately after the assay, Hoechst 33342 (Thermo Fisher Scientific) was used to stain nuclei, and fluorescence was measured on a CLARIOstar plate reader. OCR data were normalized by dividing measurements by Hoechst fluorescence intensity for each well.

Statistical analyses

Statistical analysis of differentiated HNEC data from cytokine and qPCR were analyzed by Friedman’s test comparing PIE exposures to PBS controls in a single analysis, and corrected for multiple comparisons with Benjamini Hochburg FDR of 0.05. Bioenergetics data were analyzed with Friedman’s test comparing PIE exposures to PBS controls, with post hoc Dunn’s test. Data visualization was done in Graphpad PRISM 9.1 (Graphpad Software, La Jolla, California), with matched donor exposures averaged. Exposures were analyzed independently against PBS controls, as incineration temperatures (smoldering vs flaming) were considered separate mixtures with distinct chemical and physical characteristics. All data are presented as means±SEM, with significance determined by discovery q- and p-values under .05 and a minimum of 5 biological replicates (n = 5–6). For cytokine data under the threshold of detection, the median LLOD in pg/ml was used as provided in kit instructions. 16HBE live-cell imaging data are expressed as percentage of maximum of oxidized GSH (GSSG) to reduced GSH over 3 independent experiments per exposure (n = 3). Maximum oxidation responses of each exposure were plotted and analyzed using a 1-way ANOVA with post hoc Sidak’s multiple comparisons test to control. Bulk RNA-seq results were analyzed for comparisons between smoldering and flaming PIE exposures as well as PIE exposure conditions versus vehicle controls to determine DEGs related to PIE exposure. Analyses were done in R with the EnhancedVolcano package and Quiagen Ingenuity Pathway Analysis, with DEGs selected by a log2-fold change threshold of 1 and an FDR (0.05) adjusted p-value minimum threshold of .05.

Results

Exposure to PIE condensate increases inflammatory cytokine secretion

To evaluate inflammatory signal changes in response to PIE exposure, fully differentiated HNECs from 6 donors (3M, 3F) were apically exposed to either a PBS control or smoldering or flaming PIE condensates for 4 h at 5 and 20 μg/cm². PIE exposures at both low and high concentrations caused an inflammatory response as marked by increased cytokine levels in basolateral supernatants. Multiplex ELISA analysis for proinflammatory cytokines of basolateral media revealed 4 measured cytokines were significantly increased by PIE exposure in comparison to control. Proinflammatory cytokines IL-8 and IL-1β were increased by high smoldering and all flaming PIE exposures (Figs. 1A and 1C). Immunoregulatory cytokine IL-13 was significantly increased at both low and high concentrations of smoldering PIE condensates (Figure 1D), and anti-inflammatory cytokine IL-4 was significantly increased at high flaming concentrations of PIE condensate exposures. IL-7, VEGF-A, IL-2, IL-6, TNF-α, and IL-10 levels were not significantly changed with PIE exposure in comparison to control (Figs. 1B, 1E–1H, and 1J), although several exposures approached statistical significance (q < 0.1). All p- and q-values for each cytokine and exposure are reported in Supplementary Table 2.

PIE condensate exposures increase gene expression of electrophile and inflammatory response genes

To investigate targeted gene expression changes induced by PIEs, differentiated HNECs from 6 donors (3M, 3F) were apically exposed to either a PBS control or flaming or smoldering PIE condensates at either 5 or 20 μg/cm² for 4 h. RNA was reverse transcribed to cDNA, and targeted genes associated with inflammation, electrophile responses, mucin formation, and aldehyde metabolism were assessed on a high-throughput RT-qPCR platform. Targeted genes and functions are listed in Table 3. Of the 22 genes that were measured, 9 gene expression levels were significantly increased by PIE exposure in comparison to a PBS vehicle control. Genes associated with inflammatory responses (IL6, IL-10, and IL-1β) were significantly increased at low (5 μg/cm²) flaming concentrations (Figs. 2G and 2H), and CSF2 was increased at a low (5 μg/cm²) smoldering PIE concentration (Figure 2E). Expression of CYP1A1 and CYP1B1, xenobiotic metabolism genes known to contribute to intracellular oxidative stress during PAH metabolism (Androutsopoulos et al., 2009; Mescher and Haarmann-Stemmann, 2018; Zhou et al., 2017), were significantly increased by all flaming PIE exposures (Figs. 2A and 2B), and expression of an oxidative stress response gene, HMOX1, was increased only at the high smoldering concentration (Figure 2F). Expression of aldehyde metabolism genes was increased by flaming exposures, as ALD1A3 was significantly increased by all flaming concentrations, and ALD3A1 was significantly increased by the high flaming concentration only (Figs. 2C and 2D). All p- and q-values for each gene and exposure are reported in Supplementary Table 3.

Flaming PIE condensate exposure increases glutathione oxidation

Considering the ability of flaming PIEs to upregulate expression of enzymes associated with oxidative stress, we evaluated the effects of smoldering and flaming PIEs on glutathione oxidation in real time, as a marker of cellular oxidative stress. To do so, we used a model of 16HBEs expressing oxidation-sensitive roGFP, which has been previously described(Masood et al., 2022) Cells in submerged culture were exposed to (1) PBS (negative control), (2) 100 µM peroxide (H₂O₂) (positive control), (3) 25 or 50 μg/cm² concentrations of flaming PIEs, or (4) 25 or 50 μg/cm² of smoldering PIEs. Live cell imaging of exposed 16HBEs demonstrated concentration-dependent increases in ratio of oxidized (GSSG) to reduced (GSH) glutathione in response to flaming PIEs exposure at both 25 and 50 μg/cm² concentrations, suggesting increased cellular oxidative stress (Figure 3A). In contrast, exposure to smoldering PIEs demonstrated no glutathione oxidation increases at 25 or 50 μg/cm² (25 μg/cm² not shown), indicating that flaming but not smoldering PIEs rapidly cause glutathione oxidation in respiratory epithelium. Maximum GSSG/GSH oxidation responses showed a significant increase in glutathione oxidation between the vehicle control and 50 μg/cm² flaming PIE condensate exposure (Figure 3B).

Vitamin E pre-exposure reduces glutathione oxidation response after exposure to PIE

We investigated if pretreatment with an antioxidant could modulate the glutathione oxidation response. Prior to the start of the live-cell imaging assay, a subset of 16HBEs were treated with 20 µM Vitamin E α-tocopherol in PBS for 3 h, after which all Vitamin E-supplemented PBS was removed, cells were exposed to PIE, and live-cell imaging of glutathione oxidation was assessed in submerged culture. Exposure to 25 μg/cm² flaming PIEs increased glutathione oxidation in comparison to the negative control. 16HBEs that were treated with Vitamin E did not exhibit such increases in glutathione oxidation, but instead mimicked PBS control glutathione oxidation levels (Figure 4A). Maximum GSSG/GSH oxidation responses showed a significant increase in glutathione oxidation between Vitamin E supplemented cells and those without supplementation when exposed to 25 μg/cm² flaming PIE condensate exposure (Figure 4B), suggesting that treatment with Vitamin E provides a protective effect against the cellular oxidative stress effects from PIEs exposure.

Flaming and smoldering PIE condensate exposures impair mitochondrial functions/efficiency

As glutathione functions as the main protective antioxidant in mitochondria, we examined whether PIEs exposures could perturb cellular bioenergetics based on the effects of flaming PIEs on glutathione oxidation. HNECs from 5 donors (3M, 2F) cultured in submerged conditions were exposed to a PBS control, or 25 or 50 μg/cm² concentrations of either flaming or smoldering PIEs prior to performing the Seahorse Extracellular Flux Modified Cell Mito Stress Test followed by subsequent normalization of the readings in each well by Hoechst nuclei staining, similar to our previous studies (Hickman et al., 2019; Lavrich et al., 2018). Exposure to flaming PIEs at 50 μg/cm² significantly decreased basal respiration (Figure 5A), ATP production (Figure 5B), maximum respiration (Figure 5C), and coupling efficiency (Figure 5D) in comparison to cells treated with a PBS control. Exposure to 25 μg/cm² smoldering PIEs significantly decreased maximum respiration, and 50 μg/cm² smoldering PIEs significantly decreased coupling efficiency (Figure 5D). Statistically significant decreases in mitochondrial parameters were only observed in cells exposed to 50 and not 25 μg/cm² flaming PIEs. All p- and q-values for each gene and exposure are reported in Supplementary Table 4.

Exposure to both flaming and smoldering PIEs upregulates expression of xenobiotic metabolism enzymes and proteins (RNA-seq)

To better examine which cellular pathways were perturbed by exposures to PIEs, we conducted bulk RNA-seq analysis and subsequent QIAGEN Ingenuity Pathway Analysis on RNA collected from 6 (3M, 3F) HNEC donors achieving full differentiation at ALI apically exposed to 20 μg/cm² of smoldering or flaming PIE condensates, with PBS vehicle as the control. RNA-seq analysis revealed that in comparison to the control, 20 μg/cm² smoldering exposures significantly increased expression of 13 genes and decreased 1 gene based on FDR-adjusted p-value cutoffs. Genes involved in the metabolism of PAHs and other toxic chemicals through the aryl hydrocarbon receptor (AhR) signaling pathway, including AHRR, ALDH1A3, CYP1A1, CYP1B1, and IL1A were upregulated, as well as genes involved in oxidative stress responses OSGIN1 and TXNRD1. Smoldering exposures also upregulated genes involved PXR/RXR signaling pathways AlAS1, CYP1A2, and UGT1A1, and S100 family signaling pathways C5AR2, ESR1, IGHG1, IGHG2, ITPR1, PLA2G4E, PLAT, and S1PR3. Flaming PIE exposure analysis at 20 μg/cm² exhibited a more pronounced effect than smoldering exposures, increasing expression of 54 genes and decreasing expression of 24 genes in comparison to control. Genes involved in PAH and other xenobiotics metabolism through the AhR signaling pathway were upregulated such as AHRR, ALD1A3, CYP1A1, CYP1A2, CYP1B1, UGT1A1, and UGT1A6, as well as genes involved in PXR/RXR activation AlAS1, CYP1A2, and UGT1A1. Genes associated with the IL-13 signaling pathway, FGR, IL24, SPDEF, were also upregulated. Genes were selected as significantly up- or downregulated by a fold-change >1, and a p-value cutoff of .05. All DEGs in comparison to control within these parameters are included in Supplementary Table 1 along with corresponding exposure. Plotting base mean count reads for all DEGs for PBS control, flaming, and smoldering exposures highlight similar genetic responses in HNECs to flaming exposure responses in comparison to control and smoldering exposures (Figure 6A). Using QIAGEN Ingenuity Pathway Analysis for RNA-seq data, top activated signaling pathways for 20 μg/cm² flaming and smoldering exposures were determined. Smoldering exposures activated pathways including the xenobiotic metabolism AHR signaling pathway (Figure 6B). Flaming exposures similarly activated the xenobiotic metabolism and AHR signaling pathways with higher significance than in smoldering exposures and had top hits for nicotine and melatonin degradation pathways, which use similar metabolism genes such as CYP1A1, CYP1A2, CYP1B1, and UGT1A1 for degradation (Figure 6C). Overall, strong upregulation presence of genes involved in xenobiotic metabolism and degradation drove signaling pathways analysis matching.

Discussion

Inhalation of emissions derived from incineration of plastic materials poses a significant upper respiratory health risk. Despite a growing body of research on the physical and chemical characterization of incineration emissions from various sources and the potential hazards associated with exposure to such emissions, biochemical and cellular signaling effects in the upper respiratory, in particular, are not well understood. As the first line of defense against respiratory toxicants, understanding the human nasal epithelial response is important to evaluate acute exposure to PIEs such as those sustained from burn pit emissions, especially since sinonasal disease is associated with these types of exposures (Hill et al., 2022; Lim and Shin, 2021; Mo, 2019). Throughout this study, our findings addressed the knowledge gap regarding the impact of PIEs exposure on human respiratory epithelium and the resulting downstream events triggered by PIE exposure. Our data demonstrate that acute PIE exposure causes inflammation, oxidative imbalance, impaired mitochondrial function, and transcriptomic alterations in human respiratory cells, which are implicated in the development of upper respiratory disease. Novel insights from our work indicate that incineration temperatures of PIE exposures should be considered as a factor when evaluating PIE-associated respiratory injury, as PIEs from flaming temperatures cause higher degrees of cell injury than their smoldering counterparts. An important factor to consider is that incineration of plastic materials at smoldering temperatures creates more respirable PM than their flaming counterparts (Kim et al., 2021), which could modulate our observed effects in a real-world scenario, as PIEs concentrations were kept constant in this study regardless of incineration temperature.

Recent studies have characterized the PIEs used in this study, demonstrating distinct chemical and physical profiles between smoldering and flaming PIEs. Specifically, PAH levels in flaming PIEs were measured to be over an order of magnitude higher than smoldering PIE emissions for 16 EPA priority PAHs and 10 nitro- and oxy-PAHs (Kim et al., 2021). Exposure to PAHs is associated with respiratory toxicity, mutagenicity, and upregulation of detoxifying and metabolizing enzymes (Kermani et al., 2021; Kim et al., 2013, 2022; Mallah Manthar et al., 2022; O’Driscoll et al., 2018; Shimada and Fujii-Kuriyama, 2004; Shimada and Guengerich, 2006). Genes identified through high-throughput RT-qPCR analysis of HNECs exposed to flaming and smoldering PIEs identified several significantly upregulated genes associated with oxidative and xenobiotic metabolism responses, including CYP1A1, CYP1B1, and HMOX-1 (Figs. 2A, 2B, and 2F). These data were supported by whole transcriptomic analysis of PIEs exposed HNECs, showing upregulation of genes associated with oxidative responses and PAH metabolism through the AhR pathway (AHRR, ALD1A3, CYP1A1, CYP1A2, CYP1B1, UGT1A1, and UGT1A6; Figure 6A). QIAGEN Ingenuity Pathway Analysis of whole transcriptomic data identified the AhR metabolism signaling pathway as a major pathway associated with PIE exposure (Figure 6C). As critical enzyme components of the AhR signaling metabolism pathway, CYP1A1, CYP1A2, and CYP1B1 are activated to metabolize PAHs and other environmental pollutants. It is commonly accepted that Phase I enzyme activity in the metabolism of PAHs can generate oxidative metabolites, potentially leading to DNA damage and inflammation in cells and tissues (O’Driscoll et al., 2018; Shimada and Fujii-Kuriyama, 2004; Shimada and Guengerich, 2006; Vogel et al., 2020). Hence, activation of such metabolic pathways raises a point of concern and aligns with chemical analysis identification of high PAH concentrations in PIEs.

The potential of PM and PAHs to generate reactive oxygen species (ROS) in respiratory epithelia has been shown in several studies (Guo et al., 2017; Halimu et al., 2022; Shi et al., 2022). Increases in intracellular ROS (iROS) and oxidative stress have been implicated to cause cellular inflammation, mitochondrial dysfunction, DNA damage, and posttranslational modifications in the respiratory system (Annangi et al., 2023; Guo et al., 2017; Mittal et al., 2014; Shi et al., 2022; Xiong et al., 2011). Oxidative stress is also implicated as a contributor to the pathogenesis of upper airway diseases such as sinusitis and impairment of the nasal epithelial barrier (Tai et al., 2023) however these relationships have not been definitively established. Although intracellular increases in ROS from plastic microparticle exposure have been shown in respiratory epithelia (Halimu et al., 2022; Shi et al., 2022; Wang et al., 2021), an understanding of biological timing and downstream effects from PIEs exposure in respiratory epithelia have not been examined. Glutathione (GSH) is the most abundant intracellular antioxidant, where it functions to defend cells from redox imbalance due to oxidative stress (Biswas and Rahman, 2009). At normal conditions, reduced GSH represents >99% of total glutathione in a cell, with increases in oxidized glutathione (GSSG) signaling cellular oxidative stress. GSSG/GSH redox status therefore is a critical indicator for redox homeostasis in the cell, and an important factor for proinflammatory gene transcriptional regulation. Our data demonstrate that exposure to flaming PIEs causes direct, immediate oxidation of glutathione, which continually increased over the time of exposure (Figs. 3A and 3B). We additionally found that 3-h pretreatment with 20 μM Vitamin E (α-tocopherol) significantly negated the glutathione oxidative effect of flaming PIEs in real-time (Figs. 4A and 4B). Differences in glutathione oxidation responses for flaming exposures in Figures 3 and 4 were observed and could be attributed to variance within selected ROIs for analysis, as well as variabilities within our cell culture model, such as cell passage number and cell confluency differences. As a line of defense against ROS in the cell and cellular membranes, α-tocopherol, the main form of Vitamin E, inhibits ROS production by donating electrons to break oxidative lipid chain reactions and neutralize unstable free radicals, thus preventing cellular membrane damage and inflammation (Nazrun et al., 2012; Singh and Devaraj, 2007). Therefore, the ability of α-tocopherol pretreatment to successfully protect cells from glutathione oxidation responses induced by PIEs treatment suggests that Vitamin E or other antioxidants could aid in protection of respiratory epithelia oxidative effects caused by inhalation of toxicant mixtures derived from incineration of plastics.

Numerous studies have found that PM exposure can cause inflammatory responses in respiratory epithelia (Arias-Pérez et al., 2020; Fujii et al., 2001; Kim et al., 2018, 2022; Pryor et al., 2022; Van Eeden et al., 2001). Acute exposure of HNECs to smoldering and flaming PIEs increased levels of 9 of the 10 measured cytokines at either smoldering or flaming exposures (Figure 1). Exposure to PIEs significantly increased concentrations of cytokines that function to promote inflammation (IL-8, IL-1β) as well as regulate (IL-13) and repress (IL-4) innate immune responses, with the understanding that numerous cytokines exhibit pleiotropic effects. Cytokines associated with canonical NF-κB pathway activation (IL-1β and IL-8) were significantly increased, similar to published data of NF-κB activation by ambient PM exposure (Hoesel and Schmid, 2013; Liu et al., 2017; Zhang et al., 2021). More importantly, cytokines IL-8 and IL-1β are specifically associated with acute sinusitis (Lim and Shin, 2021; Min and Lee, 2000), and were increased with PIE exposure (Figs. 1A and 1C), with expression of IL-10 and IL-1β also seen to be up-regulated with exposure (Figs. 2G and 2I). Discrepancies between inflammatory gene expression and protein concentration were observed and could be due to transcriptional regulation or posttranslational modifications that could occur with electrophilic PIEs exposures. The singular 4-h timepoint utilized may not have been optimal for multiple analytical assays, and our discordance between transcriptional and protein measurements is in agreement with previous reports (Chen et al., 2002; Shebl et al., 2010; Tian et al., 2004). Overall, the results support our previous data showing increased cytokine secretion in bronchoalveolar lavage fluid of mice exposed to PIEs by oropharyngeal aspiration (Kim et al., 2021), and expands the knowledge of cytokines that may play a role in upper respiratory disease progression after PIEs exposure. Whether our data represent an inflammatory response due to oxidative stress, direct toxicity, or a combination of the 2 is unknown, but overall, the results show an acute inflammatory response mounted by nasal epithelia with exposure to PIEs.

Pro-oxidative respiratory exposures such as PIEs could cause mitochondrial dysfunction. Mitochondria are susceptible to changes in the redox environment, as elevated ROS can lead to impairment of metabolic functions through damage to proteins, DNA, and membranes. Such damage can impair cellular health and functions such as ciliary beating (Burkhalter et al., 2019), and is implicated in various upper respiratory pathologies such as rhinosinusitis (Yoon et al., 2020). As current literature indicates oxidative stress as a major driver of mitochondrial dysfunction (Murphy, 2009; Ott et al., 2007) we expected that the previously demonstrated oxidative effects from PIEs exposure could also induce mitochondrial dysfunction in HNECs. Our study is the first to evaluate cellular metabolic changes in response to emissions from incinerated plastics in primary nasal epithelium from multiple donor cell populations. The data demonstrate that basal respiration, ATP production, maximum respiration, and coupling efficiency were all decreased with exposure to PIEs and caused inefficient energy metabolism in the nasal mucosa. These findings indicate that the oxidative environment introduced by PIEs exposure induced mitochondrial dysfunction, which could cause cellular metabolic insufficiency and impairment, and be a contributor to upper respiratory pathologies associated with PIEs exposure (Murphy, 2009; Ott et al., 2007; Yoon et al., 2020). Although immediately relevant to upper respiratory disease, our observations likely have relevance for lower respiratory tract epithelium and associated pathologies as well, as HNEC cell models have been established as surrogates for bronchial epithelial studies (Alves et al., 2016; McDougall et al., 2008; Thavagnanam et al., 2014).

Throughout this study, we investigated the ability of PIEs to induce various detrimental effects on human nasal epithelia using 2 different temperatures of plastic incineration. In previous murine studies analyzing respiratory responses to incinerated waste emission condensates, including the plastic emissions used here, found that emissions from plastic incineration are more proinflammatory than their cardboard or plywood counterparts, and flaming plastic PM exhibited higher long toxicity than smoldering PM (Kim et al., 2021). The 2 temperatures of plastic incineration, smoldering (500°C) and flaming (640°C), have been previously discussed as resulting in emission mixtures with distinct chemical and physical components, and therefore were studied separately to elucidate the biological changes induced by each. The results indicate that respiratory injury from PIEs exposure differed depending on the temperature of incineration, although general dose-response outcomes within smoldering and flaming exposures were not established. The lack of dose-response consistency could be due to donor response variability within cultured primary cells, or threshold PIE responses from exposures seen in cytokine, gene expression, and bioenergetics data. Since 2 PIEs concentrations used for exposure were chosen to represent calculated possible nasal deposition in Table 2, true dose-response relationships could not be captured. Both smoldering and flaming PIEs induced a significant inflammatory response as evidenced by elevated cytokine secretion in HNECs (Figure 1). Interestingly however, 3 cytokines IL-6, IL-10, and VEGF-A, were only significantly increased by the smoldering plastic exposure, suggesting different types of inflammatory responses caused by smoldering versus flaming PIE exposures. Greater differences in incineration temperature responses were seen in DEGs indicating oxidative and PAH metabolism with PIE exposure. RT-qPCR analysis of PIEs-exposed HNECs showed that expression of CYP1A1 and CYP1B1 was noticeably higher in exposure to flaming PIEs versus their smoldering counterparts. Similarly, transcriptomic analysis of PIE-exposed HNECs at 20 μg/cm² induced more DEGs in flaming (54) versus smoldering (14) PIE exposures, and identified a cluster of DEGs most up-regulated by flaming PIE exposure (Figure 6A). These included oxidative and xenobiotic metabolism signaling pathway genes AHRR, ALD1A3, CYP1A1, CYP1A2, CYP1B1, UGT1A1, and UGT1A6. Smoldering exposures also up-regulated AHRR, CYP1A1, and CYP1B1, but at lower levels (Supplementary Figure 1). These results are consistent with the changes in GSSG/GSH, where, in a short-time (40 min) exposure, flaming PIEs induced glutathione oxidation while smoldering exposures did not. Lastly, flaming PIEs exposures significantly decreased basal respiration and ATP production in HNEC mitochondria, whereas smoldering exposures did not reach statistical significance (Figure 5). Taken together, these results expand previous studies and indicate that both smoldering and flaming PIEs can induce inflammation and oxidative stress in upper respiratory epithelia, but flaming PIEs cause greater oxidative effects than smoldering PIE exposures.

Limitations

Our study had certain limitations. First, we only monitored acute (up to 4 h) exposures to determine toxicity of PIEs in HNECs from a single PIE exposure insult. In ambient or occupational scenarios, human nasal epithelia could be exposed to PIEs repeatedly and for much longer periods, especially as exposed individuals, such as military personnel, have been deployed to sites with continuous burn pit and therefore incinerated plastic exposures for weeks to months at a time. In addition, all exposures utilized PIEs condensates as previously described (Kim et al., 2021). The use of condensates does not take into account gases or volatile organics that were not collected in the cryotrap system. Hence, these additional emission mixtures components may potentially amplify the observed effects of PIEs exposure on HNECs described here. Additionally, discrepancies were observed among dosing between experiments, however, we included a comparable dose across all experiments at 20 or 25 μg/cm² concentrations and stayed within our calculated relevant dose range shown in Table 2. Live cell imaging of glutathione oxidation experiments was done using a bronchial epithelial cell line instead of nasal epithelial cells, due to 16HBEs being the available cell line to express the fluorogenic sensor to emit during oxidation. Lastly, in the experiments examining the effects of Vitamin E glutathione oxidation, we cannot not completely exclude the possibility that Vitamin E remaining on the cell surface after pretreatment directly reacted with PIEs flaming condensates. Although this is unlikely due to washing 16HBE cells twice to remove residual Vitamin E before PIEs exposure, we cannot completely exclude this possibility.

Conclusions

Observations from our current study indicate that respiratory exposure to PIEs causes toxicity in respiratory epithelium, as it induces oxidative stress, inflammation, mitochondrial dysfunction, and activation of AhR signaling. Robust increases in oxidative stress were associated with PIEs exposures and the observed toxicity and dysfunction in HNECs. Flaming PIE exposures induce greater toxicity than their smoldering counterparts. Therefore, temperature of incineration of plastics at the time of exposure as well as more detailed chemistry is needed to better understand emission-induced toxicity. Overall, the data show that short-term inhalation of PIEs causes significant toxicity in HNECs, suggesting toxicity caused by inhaled emissions from plastic incinerations could contribute to the development of sinonasal and other respiratory diseases associated with burn pits and other environmental PM exposures.

Supplementary data

Supplementary data are available at Toxicological Sciences online.

Declaration of conflicting interests

The author/authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Acknowledgments

The graphical abstract of this manuscript was created with BioRender.com. Opinions, interpretations, conclusions, and recommendations are those of the author and are not necessarily endorsed by the Department of Defense. This article has been reviewed and approved for release by the Center for Public Health and Environmental Risk Assessment, U.S. Environmental Protection Agency. Approval does not signify that the contents necessarily reflect the views and policies of the U.S. Environmental Protection Agency, nor does mention of trade names of commercial products constitute as endorsement or recommendation for use. We thank Dr James Samet and Dr Colette Miller for their careful review of this manuscript.

Funding

U.S. Army Medical Research Acquisition Activity, through the Department of Defense Peer-Reviewed Medical Research Program under Award No. WI1XWH-18-1-0731 of the United States Department of Defense (I.J.); Intramural Research Program of the Office of Research and Development of the U.S. Environmental Protection Agency (M.I.G.); National Institute of Environmental Health Sciences (R03ES032539 and T32 ES007126 to Y.H.K. and K.R.). The U.S. Army Medical Research Acquisition Activity, 820 Chandler Street, Fort Detrick, MD 21702-5014, is the awarding and administering acquisition office.

Digital repository

Data are available at https://dataverse.unc.edu/privateurl.xhtml?token=b52bb21b-8650-4ea3-8a80-713475608254. Github code for RNA-seq DEGs visualization available at: https://github.com/UNC-CEMALB/Emissions-from-plastic-incineration-induce-toxicity-in-primary-nasal-epithelial-cells/blob/main/Github%20RNASeqDEGs%20Heatmap%20Code.docx.

References