Primary Nasal Epithelial Cell Donors

Collection of primary hNECs from adults was performed as previously described (21). Superficial nasal epithelial scrape biopsies were obtained from healthy, nonsmoking male and female adults 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 (protocol numbers 05–2528, 09–0716, 11–1363). Written informed consent was obtained from all study participants. hNECs from an equal number of male and female donors were used for each experiment. Demographic information about the donors used for each exposure including age, body mass index (BMI), and race is provided in Table 1. Nasal biopsies were stored in RPMI-1640 medium (Gibco, Cat. No. 11875-093) on ice until further processing.

Expansion and Culture of hNECs

Culture of hNECs was performed as previously described (21, 26). Cells from nasal biopsy were expanded at passage 0 on a 12-well, PureCol-coated (Advanced Biomatrix, 5005-100ML) cell culture plate (Costar 3512) in PneumaCult-Ex Plus Medium (StemCell Technologies, Inc., Cat. No. 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). Cells were passaged and further expanded in 25 cm² tissue culture flasks (Corning, Cat. No. 430639) until passage 2. hNECs were then 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 203,000–333,000 cells per well and maintained in PneumaCult-Ex Plus Medium. Once confluency was reached on the transwells, the cultures were taken to air-liquid interface (ALI) and the apical medium was permanently removed, whereas the basolateral medium was switched for PneumaCult ALI Medium (StemCell Technologies, Inc., Cat. No. 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). After this point, three times per week, the basolateral medium was changed and the apical surfaces of the cultures were washed with 37°C HBSS + CaCl₂, + MgCl₂ (HBSS⁺⁺; Gibco, Cat. No. 14025-092). Mucociliary differentiation of the cultures was achieved after 4–6 wk of ALI conditions. At the time of exposure, cultures were at ALI for 5.29–9.14 wk.

Diesel Exhaust Particle (DEP) Suspension Preparation

Whole diesel exhaust particle material from an automobile engine was collected as described by Sagai et al. (38). The DEPs were generated using an Isuzu Automobile Co. 4JB1-type light duty 4 cylinder diesel engine (2,740 mL). The engine was operated under a load of 6 kg·m of torque at 2,000 rpm. Particles were collected “cold” at a sampling temperature of 50°C from glass fiber filters and the stainless-steel walls of the collection duct (38). Twenty-five milligrams of the DEPs was diluted in 5 mL of warmed (37°C) phenol red-free MEM basal medium (Gibco, Cat. No. 51200-038). The suspension was sonicated with a Fisher Sonic Dismembrator Model 500 with a microprobe tip for two 1-min cycles. During each cycle, the probe was moved up and down in the suspension, and sonication alternated between 30% output for 0.5 s and 0% output for 0.5 s. After each cycle, the suspension was mixed by inversion. An additional 20 mL of warmed (37°C) medium was then added to the suspension to achieve a final concentration of 1 mg/mL. Aliquots of the suspension were snap frozen in liquid nitrogen and stored at −80°C for future use.

Woodsmoke Particle Suspension Preparation

Woodsmoke generated from eucalyptus (Eucalyptus globulus) and red oak (Quercus rubra) were each collected as previously described by Kim et al. (11). Briefly, eucalyptus or red oak was burned in a quartz tube furnace at 640°C and smoke was collected in a series of cryogenic traps. The resulting woodsmoke particle condensates were then collected in acetone and concentrated with a rotary evaporator. Finally, the particles were dried and the solid PM was resuspended in Dulbecco’s PBS (Gibco, Cat. No. 14200-075) at 2 mg/mL and frozen at −20°C. Prior to exposure, aliquots were sonicated in a water bath sonicator (Sinosonic Industrial Co. Ltd., Taiwan, Model B200) at 40 kHz for 4.75 min.

Particle Size Measurements

Chemical composition analyses of particles used here from previously published studies are presented in Supplemental Table S1 (all Supplemental material is available at https://doi.org/10.6084/m9.figshare.16413261.v1). Particle size distributions of the three particle suspensions were determined by diluting an aliquot of each to 50 μg/mL in ddH₂O. The diluted suspensions were run through a BD FACSVerse 2013 Flow Cytometer for size measurement and compared with size calibration standards (Thermo Fisher, Cat. No. F13838) of 1.0, 2.0, 4.0, 6.0, 10.0, and 15.0 μm in diameter. Graphs of particle size distribution overlaid with the standard sizes are shown in Supplemental Fig. S1.

Exposure of hNECs to DEPs or WSPs

A pictorial depiction of the exposure and infection scheme is provided in Fig. 1. Prior to exposure, the apical surface of each culture was washed with 100 μL of warmed (37°C) HBSS⁺⁺ and basolateral medium was replaced with 1.0 mL of 37°C PneumaCult ALI Medium. Warm ALI Medium was used as the control exposure and as the vehicle for particle exposures. hNECs from three male and three female donors were used for each type of exposure (DEPs, eucalyptus WSPs, and red oak WSPs). Separate cultures from the same donor were used in multiple groups in some cases. Particle stock aliquots were diluted in ALI medium and applied to the apical surface of the experimental wells at a concentration of 165.7 μg/mL in 150-μL apical volume. This corresponds to a dosage of 22 μg/cm², which we have studied previously (23). Control wells received 150-μL ALI medium apically. Cultures were then returned to the incubator (37°C, 5% CO₂) for 2 h.

Infection with SARS-CoV-2 or Mock

At the end of the 2-h exposure, half the wells exposed to particle and half the control wells were apically infected with SARS-CoV-2 derived from clinical isolate WA1 (19) in high-glucose DMEM (Gibco, Cat. No. 11995-065) with 5% heat-inactivated fetal bovine serum, 1% l-glutamine diluted in ALI medium at a multiplicity of infection (MOI) of 0.5 in 100 μL. The other half of the cultures were mock infected with 100 μL of high-glucose DMEM with 5% heat-inactivated fetal bovine serum and 5% l-glutamine diluted in ALI medium. To avoid damaging or disturbing the cell monolayer, particle suspensions were not removed before addition of viral inoculum or vehicle. Total apical volume during viral infection was thus 250 μL. Cultures were then returned to the incubator (37°C, 5% CO₂) for 2 h.

Sample Collection

After the 2-h infection, cells were checked under the microscope for signs of cell death. The apical liquid was carefully removed from every well. Cultures were then washed with 200-μL 37°C HBSS⁺⁺ and returned to the incubator until collection. At the time of collection (0, 24, or 72 h p.i.), 100-μL 37°C HBSS⁺⁺ was added to the apical surface of each culture, and cells were returned to the incubator for 15 min. Apical washes were then carefully collected and analyzed for viral titer. Cells were lysed using 350-μL cold TRIzol reagent (Life Technologies, Cat. No. 15596018) for subsequent gene expression analysis.

Determination of Viral Titer

Fifty microliters of the apical wash were mixed with 450 μL of medium (DMEM + 5% FBS + 1% l-glutamine) followed by 10-fold serial dilutions resulting in a dilution series of 10⁻¹ to 10⁻⁶. Two hundred microliters of each dilution was added to plated Vero E6 cells (C1008, ATCC) and incubated at 37°C. Plates were rocked every 15 min to ensure even distribution of the virus over the surface of the well. After 1 h, 2 mL of overlay (50:50 mixture of 2.5% carboxymethylcellulose and 2× MEM α containing 6% FBS + 2% penicillin/streptomycin + 2% l-glutamine + 2% HEPES) was added to each well. Plates were incubated at 37°C, 5% CO₂ for 4 days, and then fixed with 2 mL of 4% paraformaldehyde left on overnight. Following removal of the fixative, wells were rinsed with water to remove residual overlay and then stained with 0.25% crystal violet. Visible plaques were counted and averaged between two technical replicate wells. Viral titers were calculated as plaque-forming units (pfu) per mL. The limit of detection for the assay was determined to be 12.5 pfu/wash, and samples that yielded no plaques were assigned a value of 6.25, half of the limit of detection.

RNA Extraction from Whole Cell Lysates in TRIzol

Whole cell lysates in TRIzol reagent were thawed on ice. An additional 650-μL cold TRIzol was added to each sample to facilitate RNA collection. Two hundred microliters of chloroform was added to each tube, and tubes were shaken vigorously and incubated at room temperature for 3 min. Samples were then centrifuged for 15 min at 12,000 g at 4°C. The aqueous phase containing RNA was then carefully removed from each sample and transferred to new microcentrifuge tubes. One volume of 100% ethanol was added per volume of aqueous phase removed, and samples were vortexed. Samples were further processed with the Zymo RNA Clean and Concentrator Kit (Zymo, Cat. No. R1016) according to the manufacturer’s instructions. Eluted RNA was stored at −80°C until use.

Generation of cDNA and Quantification of Gene Expression of 48 Genes by qPCR

RNA concentration and purity were measured using a CLARIOstar plate reader and an LVis Plate (BMG LABTECH). For each sample, 800 ng of RNA was used to generate cDNA in a reaction volume of 25 μL. The final concentrations of reagents in each reaction were as follows: 0.50 mM dNTPs (Promega, Cat. No. U151B), 1.00 U/μL RNasin Ribonuclease Inhibitor (Promega, Cat. No. N211A), 10.0 U/μL M-MLV Reverse Transcriptase (Invitrogen, Cat. No. 28025-013), 0.10 μg/μL Random Primers (Invitrogen, Cat. No. 58875), 50.0 mM KCl, 0.25 mM MgCl₂, 20.0 mM Tris-HCl, and 0.01 mg/mL BSA. PCR was performed in 96-well plates (Thermo, Cat. No. AB-0600 and AB-0851) for one cycle (25°C for 10 min, 37°C for 50 min, 70°C for 15 min, followed by 4°C infinite hold). Samples were submitted to the UNC School of Medicine Center for Gastrointestinal Biology and Disease Advanced Analytics Core for high-throughput qPCR gene expression analysis. Gene expression of a panel of 48 genes (including two reference genes) was assayed in a Fluidigm BioMark HD system using TaqMan primers and probes (Table 2). Duplicate Ct values were measured for each sample/gene combination and averaged for further analysis. Gene expression was calculated using the ΔΔCt method with normalization to the geometric mean of expression of the two reference genes (ACTB and GAPDH). Two samples (out of 216) showing poor amplification across the panel (i.e., comparable to the no-template controls) were excluded from the data set and not further analyzed.

Quantification of SARS-CoV-2 N1 and N2 Gene Expression by qPCR

Expression of viral SARS-CoV-2 N1 and N2 genes was also quantified and normalized to human RNase P gene expression using the Integrated DNA Technologies 2019-nCoV RUO Kit (IDT 10006713). For a single reaction, 6.5-μL nuclease-free water, 1.5 μL of one primer/probe mix, and 10 μL of TaqMan Universal Master Mix II, with UNG (Thermo Fisher, Cat. No. 4440038) were mixed and added to every well of a Sapphire 96-well PCR Microplate (Greiner Bio-One, Cat. No. 652260). cDNA was then added to each well (2 μL) for a total volume of 20 μL per reaction. The plate was sealed with a plate film (Thermo Fisher, Cat. No. 4311971) and centrifuged for 5 min at 500 g at room temperature. RT-qPCR was performed on a QuantStudio 3 using the following reaction conditions: hold 50°C for 2 min then hold 95°C for 10 min and cycle through 95°C for 15 s and 60°C for 1 min for 40 cycles. Transitions between temperatures occurred at 1.6°C/s. The two samples excluded from the Fluidigm PCR data were also excluded here. Results were collected as Ct and analyzed with the ΔΔCt method, normalized to expression of human RNase P.

Statistical Analysis

Analysis was carried out using SAS PROC MIXED as a full factorial design, with sex (male, M, or female, F), particle treatment (control, DEPs, eucalyptus WSPs, or red oak WSPs), virus or no virus, and duration (0, 24, or 72 h), as well as all their interactions. Donor was fit as a random effect. Preplanned hypothesis tests for differences between marginal means were carried out as t tests with the LSMESTIMATE command. Sex-specific means were calculated for each combination of particle treatment, virus, and duration and differences were tested using a t test with the LSMESTIMATE command. Correction for multiple comparisons was performed across all statistical tests for the entire experiment using the “qvalue” R package (v. 2.22.0), with a false-discovery rate q value threshold of 0.05, assuming pi0 = 1 (equivalent to Benjamini-Hochberg correction). The resultant P value for statistical significance was P ≤ 0.00369. Viral titer data were analyzed using GraphPad Prism, v. 8.4.0. Unpaired t tests (with Welch’s correction when appropriate) were used to evaluate differences in log₁₀-transformed data.

Particle Exposure Alone Has Modest Effects on Expression of Antiviral Host Response Genes

First, we assessed how exposure to particles alone without subsequent viral infection would affect expression of host response genes in our panel. hNECs from male and female donors were exposed to one of three particle suspensions (DEPs, eucalyptus WSPs, or red oak WSPs) or control for 2 h, followed by a “mock” infection for 2 h. Results are shown graphically in Fig. 2 and statistically significant results are reported in Table 3. At 0 h after mock infection, exposure to both types of WSPs increased expression of IL6 and eucalyptus WSPs also upregulated expression of IL1B. Further, DEPs and red oak WSPs significantly decreased expression of IFNG at 0 h p.i. (data for eucalyptus WSPs not shown due to missing data points). By 24 h p.i., both eucalyptus WSPs and red oak WSPs further upregulated IL1B expression, whereas IL6 was no longer upregulated. Overall, by 24 and 72 h p.i., particle treatment in the absence of infection had little effect on expression of the genes in our panel.

SARS-CoV-2 Infection Greatly Affects Expression of Antiviral Host Response Genes in hNECs

To assess how particle exposure affects expression of antiviral host defense genes in the presence of an infection, we next needed to measure the independent effects of SARS-CoV-2 infection on gene expression. Thus, hNECs from male and female donors that were not exposed to any particles were infected with SARS-CoV-2 (or mock infected with vehicle). Virus-induced changes in gene expression in hNECs at 0, 24, and 72 h p.i. are shown in Fig. 3 and fold-inductions and P values are tabulated in Table 4. By 24 h p.i., the type III IFNs (IFNL1 and IFNL2) were upregulated in hNECs from male and female donors, with statistically significant upregulation of both genes in males. Expression of IFNL1 and IFNL2 was even more highly upregulated at 72 h p.i. and reached statistical significance in both sexes. In addition, by 72 h p.i., infection had upregulated mRNAs of type I interferons, ISGs, chemokines, transcription factors, and viral recognition receptors. In most instances, gene expression in hNECs from males was more highly induced by infection than in hNECs from females, suggesting an overall more robust epithelial response to SARS-CoV-2 in hNECs from male donors. For each gene that was differentially expressed in infected cells from both sexes, the ratio of expression in males and females was calculated. Indeed, on average, the level of virus-induced gene expression in hNECs from males was 2.08 times (95% CI: ± 0.57) that of hNECs from females. We also assessed whether baseline differences in gene expression existed between the sexes in uninfected cells. There were no statistically significant differences in baseline gene expression between hNECs from males and females at 24 and 72 h after mock infection (data not shown).

Woodsmoke Particles Affect Expression of Virus-Induced Genes in hNECs Infected with SARS-CoV-2

We hypothesized that exposure to particles would dampen expression of crucial antiviral host response genes on subsequent SARS-CoV-2 infection. To test this, hNECs from male and female donors were exposed to control or DEPs, eucalyptus WSPs, or red oak WSPs for 2 h, followed by infection with SARS-CoV-2. Overall, red oak WSPs caused more statistically significant changes in virus-induced gene expression than the other particles (Table 5). DEPs had very few effects on virus-induced gene expression at all time points. In general, the number of statistically significant effects on gene expression increased with duration of infection in the WSP-exposed groups. More specifically, at 0 h p.i., both types of WSPs increased IL1B and IL6 expression compared with unexposed, infected cells, with red oak WSP exposure generating more potent upregulation of IL6. By 24 h p.i, all three types of particles upregulated IL1B to similar degrees and red oak WSPs downregulated MX1 and STAT2. At 72 h p.i., WSP exposures, especially from red oak, decreased expression of several genes, including IFNB1, CCL3, CCL5, CXCL10, and CXCL11 (Fig. 4). Red oak WSPs also decreased expression of IFNL1 and IFNL2, albeit not statistically significantly. Other genes that are important for the antiviral response group were also downregulated by WSPs, such as IFIT1, IFITM3, MX1, IRF7, STAT1, STAT2, DDX58, and MMP7. Thus, exposure to WSPs prior to infection with SARS-CoV-2 suppressed IFN-dependent immune gene expression.

Woodsmoke Particle Effects on Gene Expression in Infected hNECs Are Sex-Specific

Because the virus-induced effects on gene expression were sex-dependent (Fig. 3), we next assessed whether gene expression changes in cells exposed to particles prior to infection were also sex-dependent. Few sex-specific changes were observed at 0 and 24 h p.i. (Table 6). However, at 72 h p.i., WSPs from eucalyptus and red oak modified virus-induced expression of more genes in hNECs from female donors than from male donors (Fig. 5). At this time point, WSPs from both eucalyptus and red oak caused statistically significant downregulation of IFITM3, MX1, IRF7, and STAT1 in hNECs from females. In addition, red oak WSPs caused a statistically significant decline in MX1 expression in hNECs from females versus males. These results suggest that WSP exposure, especially from red oak, dampens expression of antiviral genes in hNECs from females during SARS-CoV-2 infection, with lesser effects on hNECS from males.

Particle Exposure Does Not Affect Viral Load in hNECs

Previously, we found that exposing hNECs and other airway epithelial cells to DEPs prior to infection with influenza A enhanced viral replication and susceptibility to viral infection (23). Because WSP exposure altered expression of antiviral genes in the present study, we assessed whether viral replication and release were also altered by WSP exposure. Apical viral loads for the hNECs exposed to particles and their respective controls at 0, 24, and 72 h p.i. are shown in Fig. 6, A–C. The amount of infectious virus recovered from apical washes increased with duration of infection (Fig. 6D), suggesting increased viral replication and apical secretion over time, consistent with our previous study (19). However, exposure to particles regardless of type had no effect on viral loads in apical washes (Fig. 6, A–D).

The relationship between the expression level of each gene (relative to reference genes) and the viral titer recovered from respective samples is shown in Fig. 7. As expected, expression levels of SARS-CoV-2 N1 and N2 genes are highly correlated with viral load recovered (Pearson’s r = 0.91 for both). This indicates that apical release of infectious viral particles is highly correlated with viral mRNA levels. The following genes are also correlated with viral titer, with a statistically significant Pearson’s r > 0.70: ACE2, IFIT1, IFITM3, IFNB1, IFNL1, IFNL2, MX1, CCL5, CXCL10, IRF7, STAT1, DDX58, and TLR9. In contrast, TMPRSS2 and IL1B both appear to be negatively correlated with viral titer.