Toll-Interacting Protein Down-Regulation by Cigarette Smoke Exposure Impairs Human Lung Defense against Influenza A Virus Infection

Abstract

Cigarette smoking is a primary cause of chronic obstructive pulmonary disease (COPD). Smokers have a higher risk of influenza-related mortality, but the underlying mechanisms remain unclear. Toll-interacting protein (TOLLIP), an immune regulator, inhibits influenza A virus (IAV) infection, but its regulation in COPD has not been well understood. This study was designed to determine whether cigarette smoke (CS) exposure down-regulated TOLLIP expression via epigenetic mechanisms, including histone methylation. TOLLIP and histone-methylating enzymes enhancer of zeste homolog 1/2 (EZH1/2) were measured in healthy and COPD human lungs, human airway epithelial cells cultured under submerged and air-liquid interface conditions, and precision-cut lung slices (PCLSs) exposed to CS with or without IAV infection. EZH1/2 siRNA and inhibitors were used to investigate their effects on TOLLIP expression. In patients with COPD, TOLLIP levels decreased, whereas EZH1 and EZH2 expression increased. Repeated CS exposure decreased TOLLIP and increased EZH1, EZH2, trimethylation of histone H3 at lysine 27 (H3K27me3), and IAV levels in human airway epithelial cells and PCLSs. EZH1/2 siRNA or their pharmacologic inhibitor valemetostat tosylate, in part, restored TOLLIP and reduced IAV levels in CS-exposed airway epithelial cells and PCLSs. These findings suggest that repeated CS exposure during viral infection reduced TOLLIP levels and increased viral load in part through EZH1/EZH2-H3K27me3–mediated epigenetic mechanisms. Targeting EZH1 and EZH2 may serve as one of the potential therapeutic strategies to restore TOLLIP expression and host defense against viral infections in patients with COPD.

Graphical abstract

Chronic obstructive pulmonary disease (COPD), a prevalent lung disease associated with smoking, affects approximately 480 million individuals worldwide, with 4.2 million new cases diagnosed in 2021.¹ Cigarette smokers exhibit structural and functional abnormalities in the lung, resulting in a gradual and irreversible restriction of airflow, a primary issue in COPD.² Patients with COPD experience exacerbations associated with viral infections.² So far, mechanisms underlying viral exacerbations of COPD remain to be determined. Toll-interacting protein (TOLLIP) is an intracellular adaptor protein with diverse functions related to host defense against pathogens, including regulation of inflammation, intracellular vacuole trafficking, and autophagy.³ Genetic variations in TOLLIP have been associated with a range of respiratory conditions, including idiopathic pulmonary fibrosis, asthma, primary graft dysfunction following lung transplantation, as well as susceptibility to bacterial and viral infections.⁴ TOLLIP inhibits influenza A virus (IAV) and rhinovirus infection in human airway epithelial cells and mouse lungs.^5,6 Although TOLLIP's involvement in viral infections is evident, its regulation and function in smoking-related lung diseases, such as COPD, remain largely unknown. Cigarette smoke (CS) inhalation disrupts the host defense system.⁷ However, whether CS down-regulates TOLLIP and subsequently impairs host defense function has not been explored.

Epigenetic modifications, including histone methylation, histone acetylation, aberrant DNA methylation, and dysregulated miRNAs, have been reported in COPD.⁸ Polycomb repressive complex 2 (PRC2), a vital chromatin regulator, features two key catalytic enzymes: enhancer of zeste homolog 1 (EZH1) and enhancer of zeste homolog 2 (EZH2) methyltransferases. EZH1 and EZH2 are responsible for methylating the lysine 27 residue on histone proteins, thereby repressing transcriptional activation.⁹ Inappropriate trimethylation of histone H3 at lysine 27 (H3K27me3) deposition due to elevated EZH1 and EZH2 activity and associated abnormal transcriptome have been observed in various cancers and lymphomas,¹⁰ where H3K27me3 is associated with transcriptional repression of tumor suppressor genes and cancer progression. A recent study by Anzalone et al¹¹ showed that patients with COPD exhibited more H3K27me3-positive bronchiolar epithelial cells than normal controls.

In this study, we hypothesized that EZH1 and EZH2 are up-regulated by CS exposure, which subsequently suppresses TOLLIP expression and impairs host defense against IAV in human lungs. To test this hypothesis, we analyzed the levels of TOLLIP, EZH1, and EZH2 in the lung tissue of both healthy individuals and patients with COPD. To investigate the molecular mechanisms of TOLLIP down-regulation in the context of EZH1 and EZH2 up-regulation by CS, we performed cell and tissue culture studies using human primary airway epithelial cells and human precision-cut lung slices (PCLSs).

Materials and Methods

Ethical Approval and Consent to Participate

Institutional review board (IRB) statement: IRB approval statement for healthy lungs: The study protocol for generating data in human donor lung slices/tissues was approved by the National Jewish Health IRB (protocol number HS-3209) as nonhuman subjects research.

To isolate primary human tracheobronchial epithelial cells, human tracheas were obtained from de-identified organ donors whose lungs were not suitable for transplantation and were donated for medical research through the International Institute for the Advancement of Medicine (Edison, NJ), or Donor Alliance of Colorado. The IRB at National Jewish Health deemed this research as nonhuman subjects research.

IRB approval statement for lungs of patients with COPD: This research project (protocol number HS-4190: Role of Tollip in the pathogenesis of chronic obstructive pulmonary disease) does not constitute human subjects research as defined by the federal regulations. As such, IRB review is not required.

Human Lung Tissue

The lung tissues from normal subjects and patients with COPD were obtained from Donor Alliance of Colorado (Denver, CO) or International Institute for the Advancement of Medicine (Philadelphia, PA) and National Heart, Lung, and Blood Institute–sponsored Lung Tissue Research Consortium. Detailed demographic information is given in Table 1. The IRB at National Jewish Health approved the present work.

Table 1.

Characteristics of Human Subjects for COPD Lung Tissue Study

Groups	Age, years	N	Sex, M:F	Severity	Smoking, pack-years	Smoking, pack-day	FEV1, % predicted
Normal	49.1 ± 17.0	15	9:15	NA	NA	NA	NA
COPD	54.2 ± 9.1	12	6:12	Mild (n = 1) Moderate (n = 2) Severe (n = 7) Not known (n = 2)	49.1 ± 18.9	1.7 ± 0.9	33.8 ± 25.6

F, female; M, male; COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in 1 second; NA, not applicable.

Isolation and Culture of HTBEs

Human tracheobronchial epithelial cells (HTBEs) were isolated as previously published from deidentified human donor lungs under a protocol approved by the IRB at National Jewish Health.¹² Lungs from donors with a diagnosis of brain death and without a history of lung disease, who were nonsmokers, were used. At the time of lung harvest for donation, the donor had to have absence of lung injury as indicated by an arterial oxygen tension/inspiratory oxygen fraction ratio >300, a chest radiograph without changes to indicate an acute process, and mechanical ventilation of <5 days. The details about the donors are given in Table 2. HTBEs were isolated from the distal region of the trachea and proximal parts of the main bronchi using enzymatic digestion method. HTBEs were cultured on collagen-coated 60-mm dishes with BronchiaLife culture medium (Lifeline Cell Technology, Frederick, MD) and subsequently passaged for respective experiments.¹³

Table 2.

Characteristics of Human Subjects for Airway Epithelial Cell Culture Study

Subject	Age, years	Sex	Smoking status	Cause of death
1	55	Male	Nonsmoker	Stroke
2	67	Female	Nonsmoker	Cardiac arrest during dialysis
3	62	Female	Nonsmoker	Intracranial hemorrhage
4	32	Female	Nonsmoker	Drug overdose
5	75	Male	Nonsmoker	Cardiac arrest

Preparation of PCLSs from Human Donor Lungs

The upper lobes of the right lung from the donors without history of smoking and lung diseases (Table 3) were inflated with 1.5% low-melting agarose (42°C) and subsequently sliced into sections (450 μm thick) using a Compresstome VF-300 vibratome (Precisionary Instruments, Natick, MA). PCLSs were moved into 24-well plates containing 0.5 mL Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific, Waltham, MA) supplemented with antibiotics.¹⁴ The plates were incubated in 5% CO₂ incubator at 37°C until further treatments.

Table 3.

Characteristics of Human Subjects for PCLS Culture Study

Subject	Age, years	Sex	Smoking status	Cause of death
1	55	Male	Nonsmoker	Stroke
2	67	Female	Nonsmoker	Cardiac arrest during dialysis
3	62	Female	Nonsmoker	Intracranial hemorrhage
4	32	Female	Nonsmoker	Drug overdose
5	75	Male	Nonsmoker	Cardiac arrest
6	60	Female	Nonsmoker	Head trauma
7	65	Female	Nonsmoker	Stroke

PCLS, precision-cut lung slice.

CSE Preparation

Cigarette smoke extract (CSE) was prepared according to a previous publication.¹⁵ Briefly, one research-grade cigarette (2R4F; University of Kentucky, Lexington, KY) was combusted using a Master Flex L/S Economy Drive minipump (Fisher Scientific, Pittsburgh, PA) in 25 mL of culture media [BronchiaLife (Lifeline Cell Technology) for HTBEs and DMEM for PCLSs]. The combustion continued until the whole cigarette was burned. Once the cigarette was burned out, the medium was mixed thoroughly. This solution was considered as 100% CSE, which was further filter sterilized using the 0.22-μm syringe filter for culture experiments. The 100% CSE was diluted to 10% using media and added to HTBEs or PCLSs and incubated in 5% CO₂ incubator at 37°C. A 10% dose of CSE was used in a previous publication,¹⁵ which in this study successfully reduced TOLLIP levels after 6 days of exposure.

IAV Preparation

The pandemic influenza A/CA/07/2009 virus H1N1 was kindly provided by Dr. Kevin Harrod (University of Alabama at Birmingham). The virus was passaged in Madin-Darby canine kidney (ATCC, Manassas, VA) cells, as previously published.¹⁶ Madin-Darby canine kidney cells were cultured in DMEM supplemented with 10% fetal calf serum (MilliporeSigma, Burlington, MA), L-glutamine, and antibiotics. Virus was propagated in Madin-Darby canine kidney cells in DMEM supplemented with L-glutamine, antibiotics, and 1.5 μg/mL of N-tosyl-l-phenylalanine chloromethyl ketone–treated trypsin (Thermo Fisher Scientific) and harvested at 72 hours after infection and titered by plaque assay using Madin-Darby canine kidney cells.16, 17, 18, 19

CSE Treatment and IAV Infection in Human HTBEs and PCLSs

HTBEs and PCLSs were treated with 10% CSE for 6 days (CSE refreshed every other day), followed by IAV infection. HTBEs (submerged culture) in 12-well plates were infected with IAV at 500 plaque-forming units/well for 2 hours and incubated at 37°C and 5% CO₂. In control cells, 1.5 μg/mL of N-tosyl-l-phenylalanine chloromethyl ketone–treated trypsin was added (Thermo Fisher Scientific). After 2 hours, IAV-containing medium was removed from the wells, followed by washing with 1× sterile phosphate-buffered saline for three times to remove unbound virus. Then, 1 mL of BronchiaLife was added to the wells, and cells were harvested at 48 hours after viral infection.

PCLSs in 24-well plates were challenged with IAV (3 × 10⁵ plaque-forming units/well) or DMEM with 1.5 μg/mL of N-tosyl-l-phenylalanine chloromethyl ketone–treated trypsin alone in 250 μL of DMEM media for 2 hours at 37°C and 5% CO₂.¹⁴ After 2 hours, IAV-containing medium was removed from the wells, followed by washing with 1× sterile phosphate-buffered saline for three times to remove unbound virus. Then, 500 μL of DMEM was added to the wells before harvesting PCLSs at 48 hours. The dose of IAV was selected on the basis of a previous publication.¹⁴

Air-Liquid Interface Culture

HTBEs were isolated and seeded on collagen-coated 24-well transwells (Corning Inc., Corning, NY) at a seeding density of 5 × 10⁴ cells per well in Pneumacult air-liquid interface (ALI) medium (Stemcell Technologies, Vancouver, BC, Canada). Briefly, cells on the transwells were under submerged culture for 7 to 10 days to form a monolayer and then cultured at ALI for 21 days to promote mucociliary differentiation. To switch from submerged to ALI culture, medium in the upper chamber was reduced from 100 μL for submerged culture to 20 μL for ALI culture to keep cells hydrated and promote cell mucociliary differentiation.²⁰ The details about the donors are given in Table 2.

WCS Exposure to ALI Cultured Human Airway Epithelial Cells

One research cigarette (2R4F; University of Kentucky, Lexington, KY) was combusted using a Master Flex L/S Economy Drive minipump (Fisher Scientific, Pittsburgh, PA), which drew the smoke into a 1000-mL flask and then into a smoking chamber. In this chamber, differentiated HTBEs in transwells were exposed to either whole cigarette smoke (WCS) or air (control) for 5 minutes each day for 5 consecutive days.²¹ At the end of each exposure, cells were washed with 1× phosphate-buffered saline to remove smoke condensate, including tar.

IAV Infection in HTBEs Grown at ALI

After WCS exposure, HTBEs grown at ALI were infected with IAV at 100 plaque-forming units/well for 2 hours and incubated at 37°C and 5% CO₂.

siRNA-Mediated Gene Knockdown of EZH1 and EZH2

HTBEs were seeded on collagen-coated 12-well plates and treated with CSE for 4 days. siRNA-mediated knockdown of EZH1 and EZH2 was performed 48 hours before the final CSE treatment. Briefly, cells were treated with 10 nmol/L scrabbled control siRNA, EZH1 siRNA, or EZH2 siRNA using Lipofectamine RNAiMax transfection reagent (Thermo Fisher Scientific) as per the manufacturer's protocol. After 24 hours of transfection, siRNA-containing medium was replaced with BronchiaLife medium with antibiotics. Cells were then treated with the final dose of CSE, followed by IAV infection for 48 hours and collection of cells.

Treatment with EZH1/EZH2 Inhibitor VTT

Valemetostat tosylate (VTT; MedChemExpress, Monmouth Junction, NJ) was prepared in dimethyl sulfoxide, according to the manufacturer's instructions. VTT effectively removes H3K27me3 and prevents unexpected gain of H3K27me3.²² The final VTT concentration used for culture of HTBEs and PCLSs was 500 nmol/L. During CS treatment of HTBEs or PCLSs, VTT was added from day 4 until the completion of the experiment. In control groups, dimethyl sulfoxide alone (0.1%) was added. After 48 hours of IAV infection, cells and PCLSs were collected in radioimmunoprecipitation assay and RNA lysis buffer.

Western Blot Analysis

Total proteins were extracted from HTBEs, PCLSs, and human lung tissues using radioimmunoprecipitation assay buffer containing protease and phosphatase inhibitors. Equal amount of protein was separated onto SDS-PAGE, followed by transfer onto the polyvinylidene difluoride membranes (Sigma-Aldrich, St. Louis, MO). The polyvinylidene difluoride membranes were further blocked with blocking buffer and then incubated in primary antibodies overnight at 4°C. Next day, membranes were washed with 1× wash buffer (phosphate-buffered saline with 0.1% Tween-20) and incubated in appropriate horseradish peroxidase–linked secondary antibodies and developed using a Fotodyne imaging system. ImageJ software version 1.53t (NIH, Bethesda, MD; https://imagej.net/software/imagej) was used to perform densitometry and calculate the fold change values of treatment groups versus control groups.

Co-Immunoprecipitation Assay

Immunoprecipitated proteins isolated from HTBEs were separated onto 10% SDS-PAGE for Western blot analysis using rabbit anti-H3K27me3, rabbit anti-EZH1, and EZH2 antibodies. HTBEs without any treatment were harvested in a denaturing lysis buffer, followed by heating samples to 95°C for 5 minutes. The samples were diluted using a 10× lysis buffer, and protein levels were quantified. Antibody binding to protein A/G magnetic beads (MedChemExpress) was performed for 2 hours at 4°C. Rabbit anti-EZH1 or anti-EZH2 antibody containing protein A/G magnetic beads with the precleared lysate was incubated for 24 hours at 4°C. A non-specific IgG-negative control was used to ensure the specificity of the interaction being studied. Followed by magnetic separation, 50 μL of 2× Laemmli sample buffer was added to pelleted beads, and samples were heated at 95°C for 5 minutes.

Real-Time PCR

Total RNA from HTBEs were extracted using MiniSpin Columns from GenCatch Total RNA Extraction System (Epoch Life Science Inc., Missouri City, TX). For PCLSs, homogenization using the TRIzol reagent was performed followed by using MiniSpin Columns (Epoch Life Science Inc.) for RNA extraction, according to the manufacturer's instructions. Total RNA from human lung tissues was extracted using the TRIzol reagent method.¹⁴ Real-time PCR was performed using a probe-based method, where 18S rRNA (Thermo Fisher Scientific) was used as a housekeeping gene. Custom-made primers and probe (Integrated DNA Technologies, Coralville, IA) for IAV used in this study were as follows: 5′-GACCRATCCTGTCACCTCTGAC-3′ (forward), 5′-AGGGCATTYTGGACAAAKC-3′ (reverse), and 5′-TGCAGTCCTCGCTCACTGGGCACG-3′ (probe). Also, custom primers for human interferon-λ1 were as follows: 5′-GGGAACCTGTGTCTGAGAACGT-3′ (forward), 5′-GAGTAGGGCTCAGCGCATAAATA-3′ (reverse), and 5′-CTGAGTCCACCTGACACCCCACACC-3′ (probe). TaqMan primers for TOLLIP, EZH1, and EZH2 were purchased from Thermo Fisher Scientific.¹⁴

Assay for Transposase-Accessible Chromatin Using Sequencing

HTBEs were seeded on 100-mm plates and treated with CSE for 6 days. On the sixth day, cells were treated with DNaseI for 30 minutes, followed by disassociation of cells using trypsinization. Centrifuged live cells (10⁶ cells) were stored in 1.5-mL Eppendorf tubes in MACS cell storage solution (Miltenyi Biotec, Inc., Auburn, CA). Assay for transposase-accessible chromatin using sequencing (ATAC-seq) libraries were built, and sequencing was performed on an Illumina NovaSEQ6000 platform by Novogene Corp. Inc. (Sacramento, CA). In brief, trimmomaticPE version 0.33 (http://www.usadellab.org/cms/index.php?page=trimmomatic) was used to remove adaptors and low-quality sequences from the raw sequencing reads. The trimmed reads were then aligned to the hg38 reference genome using Bowtie2 version 2.3.4.1 (https://bowtie-bio.sourceforge.net/bowtie2/index.shtml) with very-sensitive and -x 2000 parameters. SAMtools version 1.7 (https://samtools.sourceforge.net) was used to remove PCR duplicates. Narrow peaks were called by MACS2 version 2.1.2 (https://anaconda.org/bioconda/macs2/files?sort=length&sort_order=desc&type=&version=2.1.2) with the q-value cutoff of 0.05. Reads were presented as the read counts per million uniquely mapped reads. The analyzed sequence data were visualized using Integrative Genomics Viewer (https://igv.org).

ATAC Data Analysis and Protein-Protein Interaction Network Analysis by Search Tool for the Retrieval of Interacting Genes/Proteins

The ATAC-seq data were analyzed as previously described.²³ Multiple em for motif elicitation–chromatin immunoprecipitation (MEME-ChIP) version 5.5.5 was used to identify the transcription factor binding motifs enriched in accessible region.²⁴ Novogene Corp. Inc. used Homer software version 4.9.1 (http://homer.ucsd.edu/homer) for motif enrichment analysis based on the genome region of peak. As two different (MEME-ChIP and Homer) software programs were used to analyze the motif binding outcomes, the significance is expressed as E value and P value in Results.

To predict the interaction patterns of enriched motifs between control and CSE exposure groups, a protein-protein interaction network was constructed using the Search Tool for the Retrieval of Interacting Genes/Proteins database (https://string-db.org, last accessed November 11, 2024). A combined score of >0.4 was used to visualize reliable interactions.²⁵

Bulk RNA Sequencing and Analysis

Total RNA was extracted from CSE-exposed HTBEs. Briefly, preparation of RNA library, transcriptome sequencing, and downstream analysis were conducted by Novogene. The RNA library was generated by using the NEBNext Ultra II RNA library preparation kit (New England BioLabs Inc., Ipswich, MA) with a poly-A mRNA selective workflow (nondirectional, no rRNA depletion step). RNA samples were then sequenced using Illumina PE150 technology with a target output of 6 Gb PE150 data (20 million PE150 read pairs/40 million individual 150-bp reads). ClusterProfiler R package was used by Novogene Co, Ltd., to test the statistical enrichment of differential expressed genes in Kyoto Encyclopedia of Genes and Genomes pathways. Additionally, genes with a P < 0.05 and log2 fold change >0 between the groups were assigned as differentially expressed and were included for pathway analysis using the functional enrichment analysis of the Database for Annotation, Visualization and Integrated Discovery Bioinformatic Resources, which was developed by the Laboratory of Human Retrovirology and Immunoinformatics in collaboration with the National Institute of Allergy and Infectious Diseases.26, 27, 28, 29

Statistical Analysis

GraphPad Prism software version 10.0 (GraphPad Software, Boston, MA) was used for all statistical analyses. The Shapiro-Wilk test was used to assess the normality of the data distribution. For comparing two groups of parametric data in cell or tissue culture experiments, a paired t-test was applied, whereas the Wilcoxon signed-rank test was used for nonparametric data. For multiple group comparisons of parametric data, one-way analysis of variance with Tukey post hoc test was performed. For nonparametric data, two-group comparisons were conducted using the U-test, and multiple group comparisons were performed using the Friedman test, followed by the Dunn post hoc test.^30,31 P < 0.05 was considered statistically significant (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001).

Results

Reduced TOLLIP, but Increased EZH1/EZH2 and H3K27me3 Expression in Lungs of Patients with COPD

TOLLIP protein and mRNA levels in COPD lung tissue were significantly lower than those in normal lung tissue (Figure 1, A and B). To determine the potential mechanisms of TOLLIP down-regulation in COPD lung tissue, the expressions of EZH1 and EZH2, two key regulators of H3K27me3 involved in the suppression of gene transcription, were measured. Unlike TOLLIP, EZH1 (Figure 1C) and EZH2 (Figure 1D) protein levels were elevated in COPD lung tissue. H3K27me3, the catalytic product of EZH1/2, increased, although the increase was not statistically significant (Figure 1E).

Figure 1.

Reduced TOLLIP and increased EZH1/2 and H3K27me3 levels in lung tissues from patients with chronic obstructive pulmonary disease (COPD). Healthy and COPD lung tissues were used to measure TOLLIP protein (A) and mRNA (B), EZH1 protein (C), EZH2 protein (D), and H3K27me3 protein (E). Each data point (blue = healthy; red = COPD) represents data from an individual of healthy subjects or patients with COPD. Horizontal lines indicate the medians. Each specimen was measured twice. Data were analyzed using the nonparametric U-test. n = 9 to 15 healthy subjects (A–E); n = 9 to 10 patients with COPD (A–E). ∗P < 0.05, ∗∗P < 0.01. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; NS, not significant.

Repeated CSE Exposure Reduces TOLLIP and Increases EZH1, EZH2, and H3K27me3 Levels in Human PCLSs with IAV Infection

To determine whether CSE exposure was responsible for TOLLIP reduction in human lungs, the human PCLS culture model was used, as it preserves human distal lung architecture, encompassing small airways, respiratory parenchyma, and structural and immune cell populations.¹⁴ CSE alone induced EZH1, but not EZH2, levels. However, the combination of CSE and IAV increased both EZH1 and EZH2 protein (Figure 2, A and B) and mRNA levels (Supplemental Figure S1). CSE, IAV, and combination of both increased H3K27me3 protein levels (Figure 2C).

Figure 2.

A–C: Repeated cigarette smoke extract (CSE) exposure reduced TOLLIP and induced H3K27me3, EZH1, and EZH2 in human precision-cut lung slices (PCLSs). Human PCLSs were treated with CSE for 6 days, followed by influenza A virus (IAV) infection for 48 hours. A–C: Total protein was extracted, and Western blot analysis for EZH1 (A), EZH2 (B), and H3K27me3 (C) was performed. D–G: Valemetostat tosylate (VTT) restored TOLLIP with decreased IAV levels and increased antiviral genes in human PCLSs. Human PCLSs were treated with CSE for 6 days and then infected with IAV for 48 hours with or without valemetostat tosylate for measuring TOLLIP mRNA (D), TOLLIP protein (E), IAV RNA (F), and interferon (IFN)-λ mRNA (G). Each colored data point represents individual data from healthy donors. Horizontal lines indicate the medians. A–C, E, and G: Nonparametric Wilcoxon matched pairs signed rank test was used for two group comparisons. D and F: The Friedman test, followed by Dunn post hoc test, was used for multiple group comparisons. n = 7 healthy donors (A–G). ∗P < 0.05, ∗∗P < 0.01. DMSO, dimethyl sulfoxide; NS, not significant.

EZH1 and EZH2 Inhibitor VTT Restores TOLLIP, Reduces IAV Levels, and Increases Antiviral Genes in CSE-Treated Human PCLSs

VTT, an inhibitor targeting both EZH1 and EZH2, represents a novel therapeutic avenue for treating relapsed or refractory adult T-cell leukemia/lymphoma.³² TOLLIP protein level was restored when VTT was added to PCLSs treated with CSE alone or CSE + IAV. Similar trend was also observed at mRNA levels where TOLLIP was restored after VTT treatment (Figure 2, D and E). Moreover, IAV levels in CSE-treated PCLSs were reduced (Figure 2F) by VTT.

EZH1/2 inhibitors induce a cellular antiviral state, effectively suppressing infections caused by both DNA (human cytomegalovirus, adenovirus) and RNA (Zika virus) viruses.³³ Herein, interferon-λ was increased following VTT treatment in CSE-exposed PCLSs with IAV infection (Figure 2G).

Repeated CSE Exposure in Submerge-Cultured HTBEs Up-Regulates EZH1, EZH2, and H3K27me3 Expression, but Down-Regulates TOLLIP

As large airway epithelial cells are the initial site exposed to CS, whether they responded to CS in a manner similar to PCLSs was studied next. A submerged cell culture model was used to determine how basal cells respond to CS, as they represent a state of airway epithelial injury due to CS exposure and are critical to epithelial cell repair and remodeling in COPD.³⁴ After 6 days of CSE exposure, HTBEs demonstrated a significant increase of EZH1 (Figure 3A), EZH2 (P = 0.05) (Figure 3B), and H3K27me3 (Figure 3C) and reduction in TOLLIP protein and mRNA expression (Figure 3D and Supplemental Figure S2A).

Figure 3.

A–E: EZH1, EZH2, and H3K27me3 were up-regulated and TOLLIP was reduced by repeated cigarette smoke extract (CSE) treatment in human tracheobronchial epithelial cells (HTBEs). A–D: HTBEs were treated with CSE for 6 days for measuring EZH1 (A), EZH2 (B), H3K27me3 (C), and TOLLIP (D). E: Altered genes (bulk RNA-sequencing data) in the toll-like receptor (TLR)–TOLLIP pathway by CSE. F–H: EZH1 and EZH2 knockdown restored TOLLIP and reduced EZH1 and EZH2 levels in human tracheobronchial epithelial cells after repeated CSE exposure. siRNA-mediated gene silencing of EZH1 and EZH2 restored TOLLIP (F) with decreased EZH1 (G) and EZH2 (H). Each colored data point represents individual donors. Horizontal lines indicate the medians. A paired t-test was used for statistical analysis. n = 3 subjects (E); n = 5 donors (A–D and F–H). ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗∗P < 0.0001. IRAK2, interleukin-1 receptor-associated kinase-like 2 protein; NS, not significant; SC, Scrabble control.

Differentially Expressed Gene Analysis Shows Dysregulated Toll-Like Receptor–TOLLIP Signaling after 6 Days of CSE Exposure

The bulk RNA-sequencing data indicated that CSE treatment resulted in the up-regulation of 231 genes and down-regulation of 188 genes in HTBEs. Notably, after a 6-day exposure to CSE, altered gene expression of toll-like receptor–TOLLIP pathway genes, including TOLLIP, interleukin-1 receptor-associated kinase-like 2 (IRAK2), toll-like receptor 2, and NF-κB, was observed (Figure 3E).

Knockdown of EZH1 and EZH2 Restores TOLLIP Expression in CSE-Treated HTBEs

A recent study suggested that EZH1 suppressed TOLLIP transcription in dendritic cells and macrophages through the methylation of its promoter.³⁵ However, there is no report indicating whether EZH1 and/or EZH2 regulate TOLLIP expression in CS-exposed airways. Silencing EZH1 and EZH2 restored TOLLIP levels (Figure 3F) in cells with repeated CSE treatment. Similar trends of TOLLIP, EZH1, and EZH2 expressions at the mRNA levels were observed (Supplemental Figure S2, A–C). Notably, silencing EZH1 led to increased EZH2 expression and vice versa, a compensatory effect reported before (Figure 3, G and H).³⁶

IAV Infection in CSE-Exposed HTBEs Induces H3K27me3, EZH1, and EZH2 and Reduces TOLLIP and EZH Inhibitor Protein (EZHIP) Levels

The effects of IAV infection on TOLLIP regulation in CSE-exposed cells were determined. IAV infection in CSE-exposed cells demonstrated the highest levels of EZH1 (Figure 4A) and EZH2 (Figure 4B) and reduced EZHIP (Figure 4C), with increased levels of H3K27me3 (Figure 4D) and reduced TOLLIP (Figure 4, E and F). EZHIP is known to inhibit EZH1/EZH2 methyltransferase activity and expression.³⁷ mRNA levels of EZH1 and EZH2 were also altered after CSE and IAV exposure (Supplemental Figure S3).

Figure 4.

A–F: Cigarette smoke extract (CSE) and influenza A virus (IAV) infection induced EZH1 (A) and EZH2 (B), reduces EZH inhibitor protein (EZHIP) (C), induces H3K27me3 (D), and reduces TOLLIP (E and F) in human tracheobronchial epithelial cells. Each colored data point represents individual donors. Horizontal lines indicate the medians. Data were analyzed using a paired t-test for parametric data (A, B, and D) and Wilcoxon matched-pairs signed rank test for nonparametric data (C, E, and F). Interaction of H3K27me3 with EZH1 or EZH2 in human tracheobronchial epithelial cells at the baseline. G: Co-immunoprecipitation using an anti-EZH1 or EZH2 antibody, followed by Western blot analysis of H3K27me3, EZH1, and EZH2 was performed. n = 5 donors (A–F). ∗P < 0.05, ∗∗∗∗P < 0.0001. IB, immunoblotting; NS, not significant.

EZH1 and EZH2 Interact with H3K27me3 in HTBEs without CSE Treatment

A recent study suggests the interaction of EZH2 with H3K27me3 in alveolar macrophages.³⁸ Whether EZH1 and EZH2 interact with H3K27me3 in primary human airway epithelial cells was studied next. H3K27me3 was observed in EZH1 or EZH2 pulldown, but not the IgG control samples (Figure 4G). Successful pulldown of EZH1 and EZH2 was further confirmed by immunoblotting of EZH1 or EZH2, but not β-actin (Supplemental Figure S4). These findings demonstrate that both EZH1 and EZH2 interact with H3K27, which may induce histone trimethylation in human airway epithelial cells.

EZH1/2 siRNA or Inhibitor VTT Restores TOLLIP Levels with Reduced H3K27me3 Expression and IAV Levels

EZH1 or EZH2 siRNA was used to understand their individual effects on TOLLIP regulation. EZH1 or EZH2 siRNA similarly increased TOLLIP expression (Figure 5A and Supplemental Figure S5, A–C). To explore the therapeutic potential of EZH1 and EZH2 inhibitor, CSE-exposed cells were treated with VTT. VTT restored TOLLIP reduction by CSE and IAV (Figure 5B). The levels of H3K27me3 were reduced in both knockdown and inhibitor groups (Figure 5, C and D). In addition, TOLLIP mRNA levels were restored in cells treated with VTT, which supported the protein data (Supplemental Figure S5D). CSE treatment increased IAV load (Figure 5E), which was reduced by VTT (Figure 5F). Notably, VTT increased the expression of antiviral gene interferon-λ (Figure 5G).

Figure 5.

EZH1 and EZH2 inhibition restores TOLLIP and antiviral signaling with reduced H3K27me3 and influenza A virus (IAV) levels in human tracheobronchial epithelial cells (HTBEs). HTBEs treated with cigarette smoke extract (CSE) and/or IAV infection in the presence or absence of EZH1/2 inhibitor valemetostat tosylate (VTT) or siRNA were used for measuring TOLLIP (A and B), H3K27me3 (C and D), and IAV and interferon (IFN)-λ (E–G). Each colored data point represents individual donors. Horizontal lines represent the medians. One-way analysis of variance with Tukey multiple comparisons test was applied to analyze parametric data (A and B), whereas the Friedman test, followed by Dunn post hoc test, was used for nonparametric data (C). For comparisons between two groups, a paired t-test was used for parametric data (D and E), and the Wilcoxon matched-pairs signed-rank test was applied for nonparametric data (F and G). n = 5 donors (A–G). ∗P < 0.05, ∗∗P < 0.01. DMSO, dimethyl sulfoxide; SC, Scrambled control.

Reduced Chromatin Accessibility at TOLLIP and EZHIP Locus in CSE-Exposed HTBEs

ATAC-seq data revealed that after 6 days of CSE exposure, chromatin accessibility (indicated by the height of peaks) was reduced significantly at the TOLLIP locus, suggesting limited accessibility for transcription activators/repressors to bind and initiate the transcription process. Interestingly, the locus of EZH inhibitor protein (EZHIP), which primarily inhibits EZH1- and EZH2-mediated H3K27me3, also showed reduced accessibility in CSE-exposed cells (Supplemental Figure S6).

Motif Enrichment Analysis at the TOLLIP Locus Reveals Altered Transcription Factor Binding after CSE Exposure in HTBEs

Motif enrichment analysis identifies recurring sequence patterns in DNA or RNA that serve as binding sites for regulatory proteins. ATAC-seq peak motif enrichment analysis after CSE exposure was performed (Supplemental Figure S7) to compare motif binding changes³⁹ at the TOLLIP locus upstream and downstream 500 kb from the transcription start site.⁴⁰ On the basis of the present analysis, a potential association between the zinc finger 384 (ZNF384) and E74 Like ETS transcription factor (ELF) 5 binding motifs following CS exposure was observed (Supplemental Figure S8A). Given that ELF5, a member of the E26 transformation-specific (ETS) family-like ELF1 (which negatively regulates TOLLIP), plays a role in gene regulation and is induced during viral infections, ELF5 may also have role in the cellular response to CS exposure.^41,42 The MEME-ChIP motif results match the mouse Elf 5, which has high similarity with human ELF5 in terms of structure and function.⁴³ Motif enrichment (using Homer software) at whole genome after CSE treatment was also analyzed, which showed motifs involved in innate immune responses, inflammation, airway remodeling, and TOLLIP signaling altered (Supplemental Figure S8B).

WCS Exposure Reduces TOLLIP and Increases H3K27me3 and IAV Levels in HTBEs Grown at ALI Culture

The ALI culture model, which mimics the exposure of intact differentiated airway epithelium to CS, was used to confirm whether CS exposure reduces TOLLIP, as seen in human PCLSs and airway epithelial cells under submerged culture conditions. WCS exposure reduced TOLLIP levels, but increased H3K27me3 expression. Additionally, IAV levels were higher in WCS-exposed cells than the control (air) cells, further supporting the findings from the PCLS data (Figure 6, A, B, and D). The protein levels of EZH1 and EZH2 were unaltered despite the increase of H3K27me3 (data not shown).

Figure 6.

Valemetostat tosylate (VTT) restores TOLLIP with decreased influenza A virus (IAV) levels and increased antiviral genes in whole cigarette smoke (WCS)–exposed human tracheobronchial epithelial cells (HTBEs). HTBEs were cultured at air-liquid interface culture for 21 days, followed by treatment with WCS for 5 days. A: IAV infection was performed for 48 hours with or without valemetostat tosylate after WCS treatment. B–G: TOLLIP (B and C), H3K27me3 (D), IAV (E and F), and interferon (IFN)-λ (G) were measured. Each colored data point represents individual donors. Horizontal lines indicate the medians. A–G: Wilcoxon matched-pairs signed rank test was used for nonparametric data (A–D), and paired t-test was used for parametric data (E–G). n = 5 (A–G). ∗P < 0.05. DMSO, dimethyl sulfoxide; NS, not significant.

EZH1 and EZH2 Inhibitor VTT Restores TOLLIP, Reduces IAV Levels, and Increases Antiviral Genes in HTBEs Grown at ALI

Reduced TOLLIP protein was restored when VTT was added to ALI culture exposed to WCS and IAV. In addition, IAV levels were also reduced in VTT added groups. VTT increased the expression of antiviral gene interferon-λ, similar to PCLSs and submerge culture models (Figure 6, C and E–G).

Discussion

The present study reveals that repeated CS exposure reduces TOLLIP expression and exacerbates IAV infection. Mechanistically, the EZH1/EZH2/H3K27me3 axis may serve as one of the pathways involved in TOLLIP regulation in CS-exposed and IAV-infected human distal lungs and proximal airway epithelial cells.

TOLLIP deficiency due to pathologic conditions or genetic variation (polymorphisms) can increase susceptibility to infections.⁵ So far, the role of TOLLIP in smoking-related pathologies, particularly during COPD exacerbations, remains largely unexplored. The current study has advanced the research of TOLLIP regulation in several key aspects. It, for the first time, showed a reduction in TOLLIP levels in lung tissues from patients with COPD compared with healthy individuals. Although CS exposure is a major cause of COPD, whether CS directly decreases TOLLIP has not been examined thus far. Here, CS in the absence or presence of IAV infection decreased TOLLIP expression in both human distal lung tissue and proximal airway basal cells or fully differentiated epithelial cells. IAV infection alone reduced TOLLIP expression, which was worsened by CS exposure.

These findings implicate a possible role of epigenetic regulation in CS-mediated suppression of TOLLIP expression. EZH1 and EZH2 are methyltransferases that modify H3K27me3 that serves as a repressor for gene transcription.⁹ Histone methylations are commonly observed in COPD, and recently, EZH2-mediated H3K27 trimethylation has been found to be elevated in human bronchial epithelial cells from COPD smokers and nicotine-exposed patients with cancer.^11,44 Increased levels of H3K27me3 have also been observed during viral infections and are associated with poor clinical outcomes in smokers.⁴⁵ DNA hypermethylation at the TOLLIP locus is associated with its down-regulation. Notably, demethylating agents, such as 5′-aza-2′-deoxycytidine, can induce TOLLIP expression.⁴⁶ However, the role of histone methylations in regulating TOLLIP expression remains an area of active investigation. One study suggests that EZH1 inhibits TOLLIP expression by inducing H3K27me3 modification at the TOLLIP promoter, which can silence gene transcription through chromatin remodeling mechanisms. This finding points to an epigenetic mechanism involving histone modifications as a possible regulatory pathway for TOLLIP dysregulation.³⁵ In this study, the up-regulation of EZH1, EZH2, and H3K27me3 was consistently observed in the lungs of patients with COPD, as well as in CS-exposed PCLSs and airway epithelial cells. Additionally, IAV infection further elevated H3K27me3 levels. Gene silencing and pharmacologic inhibition approaches demonstrated the inhibitory role of EZH1 and EZH2 on TOLLIP expression and antiviral response. Moreover, the role of EZH1 and EZH2 in regulating H3K27me3 expression was supported by the fact that EZH1/EZH2 knockdown/inhibition reduced H3K27me3 levels in airway epithelial cells exposed to both CSE and IAV. This study offers initial insights into the regulation of TOLLIP through histone methylations, particularly focusing on the role of EZH1 and EZH2 in altering chromatin states. Although current findings suggest that these enzymes may play a role in down-regulating TOLLIP, the precise mechanism by which EZH1 and EZH2 affect TOLLIP expression remains unclear. It is possible that EZH1 and EZH2 may not be the only regulators of TOLLIP. Other transcriptional and post-transcriptional mechanisms, including miRNAs, RNA-binding proteins, and additional epigenetic modifications, could also contribute to TOLLIP regulation in both healthy and disease states.^3,4

An interesting finding in this study is that EZH2 inhibition had a stronger impact on H3K27me3 reduction compared with EZH1 inhibition. The PRC2/EZH2 complex had higher catalytic activity, whereas the PRC2/EZH1 complex exhibited greater nucleosome binding activity, aiding in chromatin compaction and maintenance after development. A recent study found that knocking down EZH2 in embryonic stem cells significantly reduces H3K27me2/3 levels, whereas H3K27 methylation is not fully abolished in EZH1 knockdown groups.⁴⁷ The current data align with previous findings, suggesting that EZH2 may be the primary catalytic component of PRC2, whereas EZH1 is likely active in the absence of EZH2.⁴⁸ The current findings also suggest that EZH2 and EZH1 compensate each other when one is absent. This compensatory role was also observed in the presence of CS exposure. EZH2 has been implicated in various lung injury models, including abnormal epithelial remodeling after injury and acute respiratory distress syndrome–associated pulmonary fibrosis.⁴⁹ It will be interesting to determine how these compensatory properties of EZH1 and EZH2 during COPD affect TOLLIP and subsequent IAV infections.

Inhibition of EZH1 and EZH2 suppresses a broad spectrum of viral infection,³³ but the underlying mechanisms have not been clear. The current study suggests TOLLIP restoration or up-regulation in CS-exposed human lung tissue or cells as a new mechanism of EZH1- and EZH2-mediated antiviral defense. TOLLIP utilizes various pathways (eg, IL-33, autophagy)⁵ to exert its host defense function. Thus, targeting EZH1 and EZH2 in CS-exposed lungs of patients with COPD may be a new promising approach to attenuate viral infection involved in disease exacerbations. The primary role of EZH proteins involves modulating chromatin accessibility, which, in turn, may regulate TOLLIP transcription. However, as EZH1/2 regulates H3K27me3, it may also affect other gene expression and subsequently influence TOLLIP promoter activity and transcription.

How TOLLIP is regulated at the transcriptional levels has been a challenge in the research field. This pilot ATAC-seq study demonstrated reduced chromatin accessibility at the TOLLIP locus following CS exposure. CS reduced the chromatin accessibility at the locus of EZHIP, an inhibitor of H3K27me3, suggesting an additional mechanism whereby EZH1 and EZH2 were regulated. Future work is needed to better understand how CS-mediated reduction of chromatin accessibility affects TOLLIP-dependent or TOLLIP-independent mechanisms in driving various pathways (eg, metabolism, transcription dysregulation, and chemokine signaling) that eventually increase the severity of viral infection in CS-exposed lungs.

Multiple small-molecule inhibitors targeting the PRC2 complex have been developed, with several in early clinical trials showing promising responses and acceptable tolerability. Interim analyses from a phase 1 study in the United States and Japan indicated that a 200-mg daily oral dose of valemetostat had an acceptable safety profile and efficacy in relapsed or refractory non-Hodgkin lymphomas, including adult T-cell leukemia/lymphoma. A phase 2 single-arm study showed valemetostat's promising response rates in Japanese patients with relapsed or refractory adult T-cell leukemia/lymphoma. EZH2 inhibitors, like CPI-1205, PF-06821497, SHR2554, and tazemetostat, are in clinical trials primarily for cancer. Although valemetostat is mostly used in the treatment of advanced cancers, it has been also tested in preclinical models to suppress viral infections.⁵⁰ Given increased EZH1/2 in COPD, the potential use of these inhibitors in COPD needs to be evaluated.⁵¹

Several limitations were noticed in this study. First, the ATAC-seq data only highlighted the impact of repeated cigarette smoking on chromatin accessibility in airway epithelial cells. Other conditions, including viral infection, will be considered in future experiments. The sample size (n = 3) for ATAC-seq was small. Nonetheless, the peaks at the TOLLIP locus were significantly reduced by CS exposure. Moreover, mRNA levels of TOLLIP, as measured by both real-time PCR and bulk RNA sequencing, were significantly reduced by CS exposure, further supporting the ATAC-seq data. Second, the human lung or PCLS contains multiple types of immune and structural cells. Although the study focused on airway epithelial cells, the impact of CS exposure on EZH1/EZH2 and H3K27me3 was not examined in other types of cells, such as macrophages. Future experiments using cell-specific approaches could address this limitation. Third, the relative contribution of EZH1 and EZH2 to TOLLIP regulation and antiviral response needs to be further investigated. Fourth, the current study used the short-term cigarette smoke exposure models. Although the short-term in vitro models did not recapture all the pathologic features of human COPD, they replicated the reduction of TOLLIP seen in human COPD lungs. These models may provide a platform to test the mechanisms of TOLLIP reduction to provide research directions toward the development of potential therapeutic targets to alleviate acute exacerbations of COPD associated with viral infections. Last, the current study focused on the use of human COPD samples and cell and tissue culture models but did not utilize the murine model of CS exposure to explicitly reveal the in vivo role of TOLLIP deficiency or EZH1/EZH2 in viral infection. As mouse COPD models have significantly advanced the research of viral infection in COPD exacerbations,^52,53 they will be considered in future experiments to provide more in-depth mechanisms.

Conclusion

The current study revealed the reduction of TOLLIP by CS exposure and its impact on viral infection. TOLLIP down-regulation may be partially mediated by EZH1/EZH2-dependent H3 lysine 27 methylation. Targeting EZH1/EZH2 to restore TOLLIP expression in smoking-induced lung pathologies, such as COPD, holds potential for reestablishing host defense and mitigating exacerbations triggered by viral infections.

Disclosure Statement

None declared.

Supplemental Data

Supplemental Figure S1.

Repeated cigarette smoke extract (CSE) exposure increases EZH1 and EZH2 levels in human precision-cut lung slices (PCLSs) with influenza A virus (IAV) infection. A: Human PCLSs were treated with CSE for 6 days and then infected with IAV for 48 hours. B–D: Total protein and RNA were extracted from tissue, and Western blot analysis for TOLLIP (B) and RT-PCR for EZH1 and EZH2 (C and D) were performed. Each data point (colored dot) represents individual donors. Horizontal lines indicate the median. Data were analyzed using the nonparametric Wilcoxon matched-pairs signed rank test. n = 5 donors (B–D). ∗P < 0.05. ns, not significant.

Supplemental Figure S2.

Gene knockdown of EZH1 and EZH2 restores cigarette smoke extract (CSE)–reduced TOLLIP levels in human tracheobronchial epithelial cells (HTBEs). HTBEs were treated with CSE, and siRNA-mediated knockdown of EZH1 and EZH2 was performed for 48 hours before the last CSE treatment. Total RNA was extracted, followed by RT-PCR for TOLLIP (A), and EZH1 and EZH2 (B and C). Each data point (colored dot) represents individual donors. Horizontal lines indicate the medians. A paired t-test and Wilcoxon matched-pairs signed rank test was used for statistical analysis. n = 5 donors (A–C). ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. ns, not significant; SC, Scrambled control.

Supplemental Figure S3.

Influenza A virus (IAV) infection induces EZH1/2 and reduces TOLLIP in cigarette smoke extract (CSE)–exposed human tracheobronchial epithelial cells (HTBEs). HTBEs were treated with CSE and infected with IAV. Total RNA was extracted, and RT-PCR for EZH1 (A) and EZH2 (B) was performed. Each data point (colored dot) represents individual donors. Horizontal lines indicate the medians. A paired t-test was used for statistical analysis. n = 5 donors (A and B). ∗P < 0.05, ∗∗P < 0.01.

Supplemental Figure S4.

H3K27me3 interacts with EZH1 and EZH2 in human tracheobronchial epithelial cells (HTBEs). HTBEs were cultured and harvested in lysis buffer, followed by performing co-immunoprecipitation assay using anti-EZH1 and EZH2 antibody containing protein A/G magnetic beads. Successful pulldown of EZH1 and EZH2 was confirmed by immunoblotting (IB) of EZH1 or EZH2, but not β-actin.

Supplemental Figure S5.

Gene knockdown/inhibition of EZH1 and EZH2 restores cigarette smoke extract (CSE)–reduced TOLLIP levels in human tracheobronchial epithelial cells (HTBEs). HTBEs were treated with CSE, and siRNA-mediated knockdown of EZH1 and EZH2 or valemetostat tosylate (VTT)–mediated inhibition was performed for 48 hours before the last CSE treatment. Total protein and RNA were extracted, followed by Western blot analysis for EZH1 and EZH2 (A and B) and TOLLIP (C and D). Each data point (colored dot) represents individual donors. Horizontal lines indicate the medians. A paired t-test and Wilcoxon matched-pairs signed rank test was used for statistical analysis. n = 5 donors (A–D). ∗P < 0.05, ∗∗P < 0.01. DMSO, dimethyl sulfoxide; IAV, influenza A virus; ns, not significant; SC, Scrambled control.

Supplemental Figure S6.

Reduced chromatin accessibility at the TOLLIP and EZH inhibitor protein (EZHIP) gene loci. Human tracheobronchial epithelial cells with repeated cigarette smoke extract (CSE) treatment were subjected to assay for transposase-accessible chromatin sequencing. Representative Integrative Genomics Viewer tracks (A), and TOLLIP and EZHIP peak signal value data (B; corresponding to orange arrows in A). Data represent three biological samples. Data were analyzed using Wilcoxon matched-pairs signed rank test.

Supplemental Figure S7.

Assay for transposase-accessible chromatin using sequencing (ATAC-seq) peak motif enrichment analysis within 500 kb upstream and downstream from the TOLLIP transcription start site. Human tracheobronchial epithelial cells with repeated cigarette smoke extract (CSE) treatment were subjected to ATAC-seq. A and B: MEME-ChIP motif enrichment analysis at TOLLIP. Elf5, E74 like ETS transcription factor 5 binding motifs; ZNF384, zinc finger 384.

Supplemental Figure S8.

A: Search Tool for the Retrieval of Interacting Genes/Proteins analysis demonstrated the interaction between the ZNF384 motif and ELF1. B: Homer motif enrichment analysis on whole genome. Data represent three biological samples. AMMECR1L, Alport syndrome mental retardation midface hypoplasia and elliptocytosis chromosomal region gene 1; AP-1, activator protein 1; CEBPE, CCAAT enhancer binding protein epsilon; COPD, chronic obstructive pulmonary disease; CPSF7, cleavage and polyadenylation specific factor 7; CSN1S1, alpha-S1-casein; ELF1/2, E74 like ETS transcription factor 1/2; ELF5, E74 like ETS transcription factor 5 binding motifs; GCM1, glial cells missing transcription factor 1; KLF5, Krüppel-like factor 5; NAB1, NGFI-A-binding protein 1; PAX5, paired box 5; TEAD4, TEA domain transcription factor 4; TLR, toll-like receptor; WFDC5, WAP four-disulfide core domain 5; ZNF, zinc finger.