Human lung epithelial adenocarcinoma cell line A549 (cat. no. GDC0063) and human lung epithelial normal cell line BEAS-2B (cat. no. GDC0139) were provided by The China Center For Type Culture Collection. All cells were subjected to short tandem repeat analysis and tested negative for Mycoplasma. BEAS-2B cells were cultured in keratinocyte-SFM (cat. no. 17005; Invitrogen; Thermo Fisher Scientific, Inc.) with supplements [30 µg/ml bovine pituitary extract (cat. no. abs9119; Absin Bioscience, Inc.) and 0.2 ng/ml rat epidermal growth factor (cat. no. CYT-556; Ameijie Technology Co., Ltd.)]. A549 cells were maintained in RPMI-1640 medium (cat. no. PM150110; Procell Biotechnology Co., Ltd.) containing 10% FBS (cat. no. C0234; Beyotime Biotechnology), 100 U/ml penicillin and 0.1 mg/ml streptomycin (cat. no. 15140122; Gibco; Thermo Fisher Scientific, Inc.). Cell culture was performed in a moist incubator at 37°C with 5% CO₂.

Cell culture plates [60 mm; cat. no. CCD06N-060; Belamb Biotechnology (Hangzhou) Co., Ltd.] were coated with 250 µl poly-2-hydroxyethylmethacrylate (poly-HEMA; 50 mg/ml in 95% ethanol; cat. no. P3932; Sigma-Aldrich, Merck KGaA) and dried in a laminar flow hood at room temperature overnight. A549 and BEAS-2B cells were trypsinized into single cell suspensions and 8×10⁵ cells were plated on poly-HEMA-coated dishes. After 24 h, the cells were collected by centrifugation at 300 × g for 5 min at room temperature and processed for cell viability, flow cytometry and protein analyses (22).

ABCA8/NOX4/FOXA2 was inserted into the plasmid cytomegalovirus (pCM)-V6-Entry vector (cat. no. PS100001; OriGene Technologies, Inc.) after amplification to construct ABCA8/NOX4/FOXA2 overexpression (oe-) plasmids (oe-ABCA8/oe-NOX4/oe-FOXA2) and the empty vector served as the negative control (NC). In addition, short hairpin RNA (shRNA) targeting ABCA8 (sh-ABCA8; cat. no. C02001; 5′-AAGTCAAATAGTTAAAGCAAATT-3′) and shRNA targeting NC (sh-NC, 5′-TTCTCCGAACGTGTCACGT-3′) were purchased from Shanghai GenePharma Co., Ltd. These shRNA constructs were cloned into the pGPU6 vector backbone. The pGPU6 vector contains a U6 promoter for shRNA expression and a neomycin resistance gene for selection in mammalian cells. The plasmid sequence is listed in Supplementary Data 1. Following the manufacturer's instructions of Lipo6000™ transfection reagent (Lipo6000; cat. no. C052; Beyotime Biotechnology), 3×10⁵ A549 cells were seeded in a 96-well plate (Biovision, Inc.). Simultaneously, Lipo6000 (0.2 µl) was diluted with 5 µl DMEM (cat. no. PM150210; Procell Biotechnology Co., Ltd.) without serum. A total of 100 ng oe-ABCA8/oe-NOX4/oe-FOXA2 or sh-ABCA8/sh-NC was diluted with 5 µl serum-free DMEM (1:1 ratio). After 20 min, Lipo6000 and oe-ABCA8/oe-NOX4/oe-FOXA2 and/or sh-ABCA8/sh-NC was introduced to the A549 cell culture medium for co-culture at 37°C for 6 h. Subsequently, the whole culture medium was replaced with a fresh complete medium for additional 24 h of culture. Transfection efficiency was detected through reverse transcription quantitative PCR (RT-qPCR) and western blotting.

First, A549 and BEAS-2B cells were divided into attachment (Att) and suspension (Sus) groups and received attachment or suspension culture as aforementioned. Second, A549 cells were assigned to Control (without treatment), oe-NC (transfected with oe-NC) or oe-ABCA8 (transfected with oe-ABCA8) groups. Third, A549 cells were distributed to Control, oe-NC and oe-NOX4 groups to validate the transfection efficiency. Fourth, A549 cell suspension was assigned to oe-NC, oe-NOX4, oe-ABCA8 and oe-NOX4 + oe-ABCA8 groups, where transfection with NC, oe-NOX4, oe-ABCA8 and oe-NOX4 + oe-ABCA8 was performed. Fifth, A549 cells were divided to Control, oe-NC (transfected with oe-NC) and oe-FOXA2 (transfected with oe-FOXA2) groups. Sixth, A549 cells were distributed into Control, sh-NC (transfected with sh-NC) and sh-ABCA8 (transfected with ABCA8) groups. Seventh, A549 cell suspension was divided into Control (no treatment), sh-ABCA8, oe-FOXA2 and sh-ABCA8 + oe-FOXA2 groups and cells in the last 3 groups were separately transfected with sh-ABCA8, oe-FOXA2 and sh-ABCA8 + oe-FOXA2.

Prediction of FOXA2 and ABCA8 expressions in LUAD based on the sample types (normal n=59; primary tumor n=515) was achieved by searching the The University of Alabama at Birmingham Cancer data analysis portal (UALCAN) (23). P<0.05 was considered statistically significant. GEPIA (http://gepia.cancer-pku.cn/; version 2.0) was used for predicting the association between ABCA8 and NOX4 or between ABCA8 and FOXA2 in LUAD. In addition, FOXA2 was identified as an upstream binding transcription factor of ABCA8 from the hTFtarget database (guolab.wchscu.cn/hTFtarget/#!/; version 1.0) (24). Subsequently, the University of California Santa Cruz (http://genome.ucsc.edu/; version 1.0) and JASPAR database (https://jaspar.genereg.net/; 2024 update) were employed to obtain the predicted binding sites of FOXA2 in the ABCA8 promoter region. Default parameters were used for transcription factor binding site prediction in JASPAR, with a relative profile score threshold of >80% for high-confidence predictions.

Enrichment of FOXA2 in ABCA8 promoter region was evaluated with ChIP kit (cat. no. P2078; Beyotime Biotechnology). After fixation in formaldehyde (1%; cat. no. BTN160220; Beijing Bio-Lab Technology Co., Ltd.) at 37°C for ~10 min, the A549 cells were treated with glycine (125 mM; cat. no. HY-Y0966; MedChemExpress) and lysed in 1 mM PMSF (cat. no. 10837091001; Sigma-Aldrich; Merck KGaA) added with SDS lysis buffer (cat. no. P0013G; Beyotime Biotechnology). To obtain the genome at the size of 200–1,000 bp, the lysate was sonicated on ice (10 sec pulse and 50 sec break, six times; 20% amplitude). A total of 20 ml sonicated lysate was taken as input for subsequent testing and 2 ml sonicated lysate was centrifuged at 10,000 × g (5 min; 4°C) with 70 ml Protein A + G Agarose/Salmon Sperm DNA. The supernatant was removed for measurement. Next, ChIP dilution buffer containing 1 mM PMSF was prepared and added to dilute the sonically processed lysate. The cross-linked chromatin lysates were incubated with FOXA2 antibody (1:5; cat. no. ab256493; Abcam) and HRP-goat anti-rabbit IgG (H+L; 1:500; cat. no. ab205718; Abcam) at 4°C overnight. A total of 60 µl Protein A + G Agarose/Salmon Sperm DNA was added and slowly mixed for 1 h to precipitate the primary antibody-recognized protein or corresponding complex. The mixture was then centrifuged again as before and was washed with the following buffers: Low-salt immune complex washing buffer (×1); high-salt immune complex washing buffer (×1); LiCl immune complex washing buffer (×1); Tris-EDTA buffer (×2). Endogenous DNA-protein compounds were precipitated by protein agarose/agarose and de-crosslinked overnight at 65°C. Finally, DNA components were extracted by the standard phenol-chloroform method (25). RT-qPCR was performed to detect the enrichment of FOXA2 in the promoter region of ABCA8. The primer sequences were as follows: Forward (5′-GCTTTCTATGTCCACCGCCT-3′); and reverse (5′-AGCCCATCCAGCATTCAACA-3′). The results were analyzed with Image J software (version 1.53; National Institutes of Health).

Wild-type (WT-pcDNA-FOXA2) and mutant FOXA2 (MUT-pcDNA-FOXA2) were cloned into the pGL3-basic plasmid (E1751; Promega Corporation) to construct the luciferase reporter vectors, with the empty pGL3-Basic plasmids (WT/MUT-Empty vector) serving as the NC. The Renilla luciferase reporter vector pRL-SV40 (cat. no. E2231; Promega Corporation) were regarded as an internal control. The recombinant plasmids were co-transfected into A549 cells for 36 h. Then, the cells were lysed in 100 µl passive lysis buffer (cat. no. E1941; Promega Corporation). Luciferase activity in cell lysates was detected with a Dual Luciferase^® Reporter Assay System (cat. no. E1910; Promega Corporation) (26).

Total RNA from BEAS-2B/A549 cells were extracted with NucleoSpin^® RNA Plus kit (cat. no. 740984.50; Machery-Nagel GmbH). To determine the quantity of the isolated RNA, a NanoDrop™ 2000 spectrophotometer (cat. no. ND-2000; Thermo Fisher Scientific, Inc.) was employed. RT and amplification of the RNAs were performed utilizing the One-Step RT-qPCR Kit (cat. no. P4305; Beijing Genenode Biotech Co., Ltd), according to the manufacturer's instructions. The reaction was performed on a CFX384 Touch PCR detection system (cat. no. 1855195; Bio-Rad Laboratories, Inc.) under the following conditions: 42°C for 30 min, 94°C for 2 min, followed by 40 cycles of 94°C for 20 sec, 60°C for 20 sec, and 72°C for 20 sec. The expression level was calculated using 2^−ΔΔCq method (27) with GAPDH as the internal reference. The relative primers are listed in Table I.

A549 cell suspension was prepared, dripped onto a microslide and covered with a coverslip. Confocal microscopy imaging was performed on a Zeiss LSM 510 laser scanning microscope (magnification, ×100; Zeiss GmbH).

A549 cells in suspension culture were seeded into 96-well plate at a density of 3×10⁵/well and were incubated at 37°C for 24 h. The culture medium was then discarded and the cells were incubated with 20 µl MTT reagent (5.0 mg/ml; cat. no. ab211091, Abcam) for 4 h at 37°C. The medium solution was removed. Next, 100 µl DMSO (cat. no. ST038; Beyotime Biotechnology Co., Ltd.) was added to the wells. A microplate reader (cat. no. 1681130; Bio-Rad Laboratories, Inc.) was exploited to evaluate absorbance at 590 nm (28).

A549 apoptosis in suspension culture was monitored using the Annexin V-FITC/PI Apoptosis Detection Kit (cat. no. 40302ES50; Shanghai Yeasen Biotechnology Co., Ltd.). Following the instructions of the kit, A549 cells were centrifuged at 300 × g for ~5 min (at 4°C) and the cell culture medium was removed. Cells were washed with PBS and counted. Subsequently, 100 µl binding buffer was added to the wells, followed by addition of 5 µl Annexin V-FITC and 10 µl PI staining solution. The cells were then cultured in an incubator at 25°C for 10 min (in the dark). Next, 400 µl binding buffer was added to the cells, which were then placed in an ice bath. The results were assessed using a FACSCalibur flow cytometer (BD Biosciences) with FlowJo software (version 10; FlowJo, LLC).

With lysis buffer (cat. no. 20118ES60; Shanghai Yeasen Biotechnology Co., Ltd.), the protein was isolated from A549 cells. Then, the extracted lysates underwent 5 min centrifugation (1,000 × g) at 4°C and the supernatant was obtained for following detection. A BCA assay kit (cat. no. 23225; Thermo Fisher Scientific, Inc.) was exploited to determine the protein concentration. Then, SDS-PAGE gel electrophoresis (8–12%; cat. no. M00659; GenScript Biotech Corporation) was applied for electrophoresis of equal amount of protein (30 µg/lane), which was immediately loaded onto PDVF membranes (cat. no. MSPVDF022BX, Membrane Solutions). Next, the membranes were blocked (10 min) with 10% blocking buffer (cat. no. YT067; Beijing Biolab Technology Co., Ltd.) at room temperature and probed with primary antibodies for NOX4 (cat. no. ab154244; 67 kDa; 1:500; Abcam), cleaved-caspase3 (cat. no. ab32042; 17 kDa; 1:500; Abcam), cleaved caspase-8 (cat. no. 9496; 1:1,000; 25 kDa; Cell Signaling Technology, Inc.), caspase3 (cat. no. ab32351; 31 kDa; 1:5,000; Abcam), caspase8 (cat. no. ab25901; 55 kDa; 1 µg/ml; Abcam), ABCA8 (cat. no. ab230896; 179 kDa; 1:1,000; Abcam) and GAPDH (cat. no. 5174; 37 kDa; 1:1,000; Cell Signaling Technology, Inc.) at 4°C overnight. Subsequently, the membranes were exposed to HRP-secondary antibody goat anti-rabbit IgG H&L (cat. no. ab6721; 1:2,000; Abcam) for 1 h at room temperature. The membranes with proteins were then visualized using ECL Plus (cat. no. WBKlS0100; MilliporeSigma; Merck KGaA) and exposed using the Kodak In-Vivo Imaging System FX Pro (Kodak). The results were analyzed with a semi-quantification software (Image J; version 1.53; National Institutes of Health) and GAPDH served as the internal reference.

All measurement data are presented as mean ± SD. Comparisons between two groups were performed by independent sample t-tests, while comparisons among multiple groups were made through ANOVA tests. The normality of data distribution was verified using the Shapiro-Wilk test and the homogeneity of variances was assessed with the Brown-Forsythe test. Statistical analyses were achieved using GraphPad software (version 8.0; Dotmatics) and P<0.05 was considered to indicate a statistically significant difference.

Given that ABCA8 expression is downregulated in NSCLC (13), a series of experiments were conducted to reveal its associated pathways in LUAD mechanisms. Through retrieval and assessment of the UALCAN website, expression levels of FOXA2 and ABCA8 in LUAD were predicted based on sample type (normal n=59; primary tumor n=515). FOXA2 expression was decreased in LUAD tumor tissues, compared with normal tissues (P=8.250×10⁻⁹; Fig. 1A) and this was also true for ABCA8 expression in LUAD tumor tissues (P=1.624×10⁻¹²; Fig. 1B). These findings indicated that both FOXA2 and ABCA8 expression levels were decreased in LUAD.

Anoikis resistance has been established as a key process of cancer cell metastasis in LUAD (29). NOX4 can mediate anoikis resistance in lung cancer cells (7). Considering the existing results and the established downregulation of ABCA8 and FOXA2 in LUAD, the association between ABCA8 and NOX4 in the anoikis resistance of LUAD was determined. Suspension culture was carried out for A549 cells, H1299 cells, H1975 cells and BEAS-2B cells to induce the anoikis resistance, with an attachment culture serving as the control. Subsequently, RT-qPCR was conducted to detect the mRNA expression of FOXA2 and ABCA8 in A549 cells and BEAS-2B cells undergoing attachment culture. Compared with BEAS-2B cells, A549 cells, H1299 cells and H1975 cells presented levels of FOXA2 and ABCA8 decreased by ~50% (P<0.001; Fig. 2A and B). Then, mRNA levels of FOXA2, ABCA8 and NOX4 in A549 cells, H1299 cells and H1975 cells experiencing attachment and suspension culture were measured. The results revealed that ABCA8 and FOXA2 levels in A549 cells, H1299 cells and H1975 cells were lower in suspension culture compared with attachment culture (P=0.0007, P=0.0003, P=0.005, P=0.0004, P=0.001 and P=0.0008; Fig. 2C, D, F, G, I and J). Conversely, A549 cells, H1299 cells and H1975 cells in suspension culture exhibited higher NOX4 mRNA level compared with those in attachment culture (P=0.0011, P=0.0011 and P=0.0013; Fig. 2E, H and K). Subsequently, RT-qPCR data reflected that oe-ABCA8 significantly increased ABCA8 expression by 1-fold in A549 cells (P=0.0002; Fig. 2L), which demonstrated successful transfection. Before experiments, a negative correlation was predicted between ABCA8 and NOX4 in GEPIA (P=0.014; R=−0.11; Fig. 2M), which was further determined by western blotting results. Furthermore, oe-ABCA8 decreased NOX4 protein expression by ~70% (P=0.0001; Fig. 2N). Collectively, it was concluded that FOXA2 and ABCA8 levels were downregulated in A549 cells and a further reduction could be induced in A549 cells under suspension suspension; correspondingly, NOX4 expression was upregulated and exhibited a negative correlation with ABCA8 level.

Subsequently, oe-NOX4 plasmid was transfected into A549 cell to ascertain the negative association between ABCA8 and NOX4. From the RT-qPCR results, it was noted that oe-NOX4 successfully increased the NOX4 mRNA level in A549 cells (P=0.0002; Fig. 2O). Then, confocal microscopy images substantiated that oe-NOX4 promoted the aggregation of A549 cells, which was reversed by oe-ABCA8 (Fig. 3A). In addition, MTT assay results indicated that oe-ABCA8 attenuated the stimulatory effect of NOX4 on A549 cell viability by ~20% (P=0.002; Fig. 3B). According to the outcomes of flow cytometry, it was observed that oe-NOX4 reduced apoptosis rate of A549 cells by ~63% (P<0.0001; Fig. 3C and D), which was offset by oe-ABCA8 (P<0.0001; Fig. 3C and D). In addition, the expressions of cleaved-caspase3 and cleaved-caspase8, established apoptosis-associated proteins (30), as well as NOX4 expression were determined using western blotting. The relative protein expression of cleaved-caspase3 and cleaved-caspase8 in A549 cells was reduced by ~50% after transfection of oe-NOX4 (P=0.0028 and P<0.0001; Fig. 3E-G), whereas oe-ABCA8 reversed this inhibitory effect of oe-NOX4 (P=0.0028 and P<0.0001; Fig. 3E-G). NOX4 expression in A549 cells was upregulated by oe-NOX4, yet downregulated by oe-ABCA8 (P=0.0018 and P=0.0018; Fig. 3E-H). Overall, oe-NOX4 strengthened anoikis resistance in LUAD (A549) cells, which was counteracted by oe-ABCA8.

FOXA2 was considered as the upstream transcription factor of ABCA8 after analysis and experimental validation of their association. Transfection efficiency of oe-FOXA2 was examined by RT-qPCR and the data revealed FOXA2 mRNA level was increased 1-fold by oe-FOXA2 (P=0.0003; Fig. 4A). By searching GEPIA, it was determined that ABCA8 was positively correlated with FOXA2 (P=5.6×10⁻⁷; R=0.23, Fig. 4B). Western blotting was used for determination of ABCA8 expression in A549 cells and the results elucidated that relative protein expression of ABCA8 was significantly increased by 0.8-fold after transfection with oe-FOXA2 (P=0.0013, Fig. 4C). Furthermore, JASPAR was employed to predict the binding sites of FOXA2 to ABCA8 (Fig. 4D). ChIP and dual-luciferase reporter assays were conducted to test the binding of FOXA2 and ABCA8. It was observed that FOXA2 was enriched in the promoter of ABCA8 (P<0.0001; Fig. 4E). In addition, the outcomes of dual-luciferase reporter assay revealed that relative luciferase activity was ~3-fold higher in WT-pcDNA-FOXA2 compared with that of the WT-empty vector (P=0.0004; Fig. 4F), while no significant difference was found in MUT groups (Fig. 4F). Sh-ABCA8 reduced ABCA8 expression in A549 cells by ~50% (P=0.0003; Fig. 4G), indicating successful transfection. The aggregation of A549 cells was promoted after transfection of sh-ABCA8, whereas oe-FOXA2 weakened the effect of sh-ABCA8 (Fig. 4H). In addition, oe-FOXA2 counteracted the promoting role of sh-ABCA8 in A549 cell viability (P=0.0023; Fig. 4I). Flow cytometry data indicated that the apoptosis rate was decreased after transfection with sh-ABCA8 (P<0.0001; Fig. 5A and B), which was reversed by oe-FOXA2 (P<0.0001; Fig. 5A and B). Western blotting results further illustrated that sh-ABCA8 inhibited cleaved-caspase3 and cleaved-caspase8 protein expressions and promoted NOX4 expression (P=0.0029, P=0.0116 and P=0.0025; Fig. 5C-F), which was offset by oe-FOXA2 (P=0.0035, P<0.0001 and P=0.0026; Fig. 5C-F). Overall, these results indicated that FOXA2 promoted anoikis of A549 cells by upregulating the expression of ABCA8. Collectively, these findings support a model whereby FOXA2 upregulates ABCA8, which suppresses NOX4 and alleviates anoikis resistance (Fig. 6).