Cigarette smoke extract preparation followed a previously reported method with some modifications (Higashi et al., 2014, 2016). One piece of cigarette (Septwolves, Longyan Cigarette Factory, Fujian, China) was burned in a 500 ml bottle with a modified 50-ml syringe-driven apparatus. The smoke was dissolved in 20 ml of RPMI-1640 medium. To remove large particles and bacteria, the resulting solution was filtered through a 0.22-μm pore-size filter and adjusted to a pH of 7.4. The solution was identified as 100% CSE. Concentrations of 5% CSE were obtained by diluting 100% CSE in RPMI-1640 complete medium. Within 30 min of preparation, 5% CSE was used for subsequent experiments.

Bronchoalveolar epithelial cells (BAECs) were obtained from Biyuntian Biological Technology Co., Ltd., (Shanghai, China). Complete medium was made from RPMI-1640 (HyClone Laboratory Inc., United States) supplemented with 10% fetal bovine serum (Gibco Life Technologies Inc., United States). Cells were cultured in the incubator at 37°C (Thermo Fisher Scientific, Waltham, MA, United States) with a humidified atmosphere of a 5% CO2. In CSE group, once BAECs cells reached 70–80% confluency, replaced complete medium with 5% CSE. CSE exposure was applied for 24 h. In CSE + ferrostatin-1 group, co-treated BAECs with 5% CSE and 1 μM ferroptosis inhibitor, ferrostatin-1 was applied for 24 h.

After extracting total RNA from BAECs by TRIzol Reagent (Invitrogen, United States), the quality and quantity of the total RNA were quantified via DNA/RNA concentration assay (Pharmathea Inc., China). Poly-T oligo-attached magnetic beads were then used to purify the mRNA. First-strand cDNA and second-strand cDNA was synthesized with random Primers and dUTP respectively. End-repair and ligation with “A-tailing” base adaptors were performed in all synthetic cDNA. After that, polymerase chain reaction (PCR) was used to amplify suitable fragments and construct the cDNA library. The quantity and quality of the sample library were verified by qPCR quantification methods and Agilent 2100 Bioanalyzer (Agilent, United States). The libraries were sequenced by NovaSeq 6000 (Illumina, United States).

After quality control and removal of ribosomal RNA, the raw sequence data were detected with adaptor sequences, removed street sequences and short fragments in Read. The trimmed data thus obtained were compared to the reference genome. Fragments per kilobase of exons per million fragments mapped (FRKM) were used to calculate transcript levels. The genes with a FRKM > 0.5 were included for follow-up analysis. Genes with p-value < 0.05 and a CSE vs. control fold change > 1.50 (or <1/1.5) were defined as DEGs.

Volcano graph, scatter plots, and hierarchical clustering were used to visualize the differential gene expression profiles. Functional enrichment of DEGs based on the DAVID database was performed via Gene Ontology (GO) enrichment analysis¹. Signal pathways involving DEGs were obtained via Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis². Significant GO terms and pathways were identified via Fisher’s exact test. A p value < 0.05 was regarded as significant.

Total RNA was reverse transcribed into cDNA by PrimeScript RT reagent Kit (Takara Bio Inc., Japan). To determine the mRNA expression levels of Glutathione peroxidase 4 (GPX4), acyl-CoA synthetase long-chain family member 4 (ACSL4), Recombinant Solute Carrier Family 7, Member 11 (SLC7A11), ferritin heavy chain 1 (FTH1), interleukin-6 (IL6) and tumor necrosis factor-alpha (TNF-α), the SYBR Green PCR Master Mix (Takara Bio Inc., Japan) was used and RT-qPCR was performed on an ABI 7500 thermocycler (Applied Biosystems, Foster City, CA, United States). Relevant primers are listed in Table 1. β-Actin was used as an internal control. Fold changes were calculated by the 2^–ΔΔCT method.

A Cell Counting Kit-8 assay (Beyotime Bio Inc., China) was used to detect cell viability. Five thousand cells per well were added to a 96-well plate, with five replicate wells in each group. phosphate-buffered saline (PBS) was added to surrounding wells to reduce errors. After 12 h, 5% CSE was added in CSE group. After modeling, 10 μL CCK-8 solution was added to each well and the 96-well plate was incubated at 37°C for 1 h. The absorbance at 450 nm was detected with a microplate reader. Cell viability was calculated as (experimental group-blank control) / (negative control-blank control) × 100%.

Cells were double stained with Annexin V-fluorescein isothiocyanate (FITC) and propidium iodide (PI) (Beyotime Bio Inc., China). The cell apoptosis rate was then assessed by flow cytometry analysis. In brief, 1 × 10⁵ cells were collected per well, with three replicate wells in each group. After washing in PBS, cells were resuspended in 200 μl binding buffer, mixed with 10 μl of PI and 5 μl of Annexin V-FITC. The suspensions were analyzed by a C6 flow cytometer (Becton Dickinson, United States).

Intracellular reactive oxygen species (ROS) were detected by the Reactive Oxygen Species Assay Kit (Beyotime Bio Inc., China). The DCFH-DA fluorescent probe and serum-free culture medium were diluted at a ratio of 1:1,000. One milliliter of diluted DCFH-DA solution was added to one well of the 12-well plates, with three replicate wells in each group. After incubation for 20 min in a 37°C incubator, cells were collected for flow cytometry analysis.

When the data were normally distributed, the mean ± standard deviation was used; otherwise, the interquartile range was used. An independent t test was used to compare the mean of continuous variables in normally distributed data; the Mann–Whitney test was used to compare non-normally distributed data. A p value < 0.05 was considered statistically significant. Statistical analysis was performed by Graphpad Prism 7.0 (GraphPad Software Inc., United States).

After 24-h exposure to 5% CSE, BAEC viability decreased significantly compared with the control group, (78.7 ± 4.8)% vs. (100.0 ± 0.0)%, p < 0.05. The rate of early apoptotic BAECs increased in the CSE group compared with the control group, (5.78 ± 0.61)% vs. (4.47 ± 0.32)%, p < 0.05. These data support our CSE-induced injury model in BAECs, as shown in Figure 1.

Differentially expressed genes were identified in CSE-treated BAECs by RNA-seq. With a fold change cutoff of 1.5 (or less than 1/1.5) and p < 0.05, 369 dysregulated genes (210 upregulated genes and 159 downregulated genes) were identified in the BAECs exposed to 5% CSE compared with controls. To visualize the DEGs between the CSE group and the control group, volcano plots were performed by fold change values and p values; Gene expression differences are also presented as scatter plots (Figures 2A,B). Hierarchical clustering was conducted to predict the function of DEGs (Figure 2C). SERPINB2, SHISA2, NQO1, PRDM2, CYP1B1, STC2, AHRR, TRIM16L, EREG and SLC7A11 were the top 10 upregulated genes, as shown in Table 2. JAG1, S100A2, NOTCH3, RPL36A-HNRNPH2, MEGF6, SULF1, EPPK1, SLIT3, CD14 and CTGF were the top 10 downregulated genes, as shown in Table 3.

The DEGs were classified by the GO database on the basis of the categories of biological process (BP), cellular component (CC), and molecular function (MF). The top ten GO terms for upregulated genes and downregulated genes are shown in Figure 3. The dysregulated genes were enriched in BP terms including response to oxidative stress and developmental process, CC terms including nucleoplasm and extracellular region, and MF terms including metal ion binding, oxidoreductase activity, and protein binding.

Kyoto encyclopedia of genes and genomes pathway analysis was conducted to find potential biological functions of the significant DEGs. Twenty signaling pathways, including ferroptosis, glutathione metabolism and cancer, were associated with the upregulated DEGs and 12 signaling pathways, including protein digestion and absorption, the Notch signaling pathway, Th1 and Th2 cell differentiation and the Mitogen-activated protein kinase (MAPK) signaling pathway, were associated with the downregulated DEGs. The top 10 enriched pathways sorted by enrichment score are showed in Figure 4.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was performed to verify the differential expression of ferroptosis genes in the CSE group compared with the control group. Compared with the control group, ACSL4, FTH1, and SLC7A11 mRNA levels were significantly upregulated in CSE group, while GPX4 mRNA expression was significantly downregulated in CSE group, p < 0.05 (Figure 5A).

TNF-α and IL6 are biomarkers of inflammation. The qRT-PCR results suggested that CSE treatment increased mRNA levels of TNF-α and IL6, which reflected CSE treatment induced inflammation in BAECs. Therefore, we co-treated BEACs with CSE and the ferroptosis inhibitor, ferrostatin-1. Compared with CSE treatment alone, incubation with CSE + ferrostatin-1 resulted in decreased mRNA expression of TNF-α and IL6, p < 0.05 (Figure 5A). Ferrostatin-1 also changed the mRNA levels of GPX4 and ACSL4 in CSE treated BAECs, p < 0.05 (Figure 5A).

The results of flow cytometry analysis suggested that CSE treatment increased the apoptosis rate of BAECs, while co-treated BEACs with CSE and ferrostatin-1 alleviated apoptosis of BAECs (Figures 5B,D). The DCFH-DA fluorescent probe assay showed increased ROS levels in CSE-treated BAECs. Compared with the CSE group, ROS levels were lower in the CSE + ferrostatin-1-treated group (Figures 5C,E).