1. Introduction
Lung cancer has emerged as the leading cause of cancer-related deaths globally [
1,
2]. The primary histological subtype of lung cancer, known as lung adenocarcinoma (LUAD), encompasses a spectrum of preinvasive lesions to invasive adenocarcinoma [
3,
4,
5]. This subtype is associated with higher mortality rates, primarily due to limited understanding of the molecular mechanisms underlying cancer development. Therefore, it is crucial to explore novel biomarkers and uncover the precise molecular targets of LUAD. Differentially expressed genes (DEGs) are identified as the result of genetic defects (e.g., EGFR and TP53 mutation) and epigenetic modifications (such as DNA methylation and RNA methylation), which play a vital role in promoting tumorigenesis, progression and metastasis of LUAD [
6,
7,
8].
With the continuous improvement of DNA microarrays and genechip technology, high throughput sequencing technology is now considered a promising tool for identifying epigenetic changes in cancer-related genes. This has led to the generation of a large number of open-access datasets in publicly accessible database repositories. The most widely used public database in the field of oncology are the Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) which contain thousands of clinical information and gene sequencing data. The database can provide a well-rounded analysis of aberrant gene expression in various cancers [
7,
9]. Lots of DEGs associated with LUAD were unequivocally identified by some gene expression profiling studies, based on reanalyzed gene expression data derived from different institutes. Usually, the analysis of aberrant gene expression can provide valuable information for the objective identification of novel biomarkers for LUAD. However, the traditional analysis of DEGs from a single study has its limitations. By obtaining overlapping genes from multiple available datasets and conducting comprehensive bioinformatics analysis, more representative and accurate results can be obtained.
DNA methylation is a common mechanism of epigenetic regulation that affects gene expression and various cellular processes. Several studies have demonstrated a close relationship between DNA methylation and the progression of cancer [
10,
11,
12,
13]. However, a comprehensive profile that includes both DEGs and DNA methylation in LUAD is still lacking. In this study, we analyzed the gene expression microarray dataset (GSE118370) and DNA methylation profiling microarray dataset (GSE139032) from GEO database. The methylation-regulated differentially expressed genes (MeDEGs) between human LUAD tissue and healthy lung tissue were identified from the previous two microarray datasets. Subsequently, gene-related functional enrichment was performed using gene ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) databases, and the protein–protein interaction (PPI) network construction was constructed to screen highly cancer-related genes for LUAD [
14]. Then, we identified eight hub genes, with the baculoviral IAP repeat containing the 5 (BIRC5) gene, which is considered a key biomarker. We further validated its regulatory effect on cell death in LUAD.
The BIRC5, also known as Survivin, was a member of the inhibitor of the apoptosis family. Survivin is tumor-specific gene, expressed only in tumors and embryonic tissues, and is closely related to differentiation, proliferation, invasion and metastasis of tumor cells [
15,
16,
17]. In our research, we observed an association between higher BIRC5 expression and unfavorable prognosis in LUAD. However, differing from the traditional view, our results show that the BIRC5 inhibition can mediate apoptosis to pyroptosis via caspase3/GSDME in LUAD; these findings provide a key biomarker and explore new therapeutic targets for lung adenocarcinoma diagnosis and treatment.
3. Discussion
Lung cancers are the main cause of cancer incidence and malignant tumor-related mortality worldwide, in which a 5-year survival rate is very low (approximately 19%). LUAD is a kind of non-small cell lung cancer, accounting for approximately 40% of lung malignancies. Numerous studies over recent decades have identified a number of oncologic dependencies with the most frequent somatic mutations in LUAD being TP53, KRAS, EGFR, ALK, MET, STK11 and KEAP1. Recently, targeted therapies for cancer have made remarkable progress. Nonetheless, a substantial proportion of LUAD patients still lack targeted treatment options, either because there is a lack of known or currently targetable mutations or the mortality of LUAD remains largely unchanged. To address this issue, proteogenomic profiling is needed to screen out more candidate therapeutic targets and provide deeper mechanical insights into LUAD [
21,
22].
Bioinformatic analysis of transcription profiles using microarrays is a novel approach to explore pathogenesis of LUAD, identify disease biomarkers and discover therapeutic targets. Unlike simple visualization of gene expression analysis, a more comprehensive deep-scale analysis of LUAD based on genomic and post-translational modifications (PTM) can allow us to find out more novel molecular target consequences of epigenetic and genomic alterations. In this study, we performed a comprehensive genomic and DNA methylation modifications analysis of paired (patient matched) LUAD tumors and normal adjacent tissues. We downloaded the gene expression microarray dataset (GSE118370) and DNA methylation profiling microarray dataset (GSE139032) from the GEO database and conducted a bioinformatics study based on these two datasets. We obtained a total of 1377 MeDEGs associated with the pathogenesis of LUAD. Among them, a total of 510 upregulated and hypormethylated genes and 867 downregulated and hypermethylated genes were identified in lung adenocarcinoma tissues compared to adjacent normal lung tissues. The GO process and KEGG enrichment analysis were performed to investigate the biological function and signaling pathway of MeDEGs. We carried out hub genes by studying the functional correlations of MeDEG-encoded proteins with the help of the STRING database. Then, to validate the hub genes in MeDEGs, we further testified the expression and assessed the influence of candidate key genes on patient survival using clinical information in the TCGA dataset. After analyzing the DNA methylation of hub genes using DiseaseMeth (
Table S1), we screened out eight novel hub genes including CCL20, MUC5B, ALDH3B2, TFF1, FA2H, BIRC5, ADRB2 and SLIT3 which were consistent with the sequencing data (GSE118370, GSE139032). Next, we verified the functionality of these hub genes through laboratory experiments. The results showed that BIRC5 is a new target for the diagnosis and treatment of lung adenocarcinoma, and that its expression was significantly elevated in LUAD and more positively associated with tumor progression [
9,
14,
23,
24,
25].
The BIRC5, which is localized in the nucleus, is known for its multifunctional role in cell differentiation, proliferation, migration, and invasion of various tumor cell types. It plays a crucial role in facilitating tumor progression [
26,
27]. BIRC5, a suppressor of apoptosis encoding the protein Survivin, is a mitotic spindle checkpoint gene located on chromosome 17 that is overexpressed in most cancer cells and its high expression is associated with worse survival [
28]. Moreover, research studies have shown that BIRC5 is widely used in the treatment of multiple cancers as a biomarker of chemotherapy resistance, which increases metastatic activity and tumor recurrence risk [
29]. Numerous studies have shown that the elevation of BIRC5 induced multi-drug resistance in various cancers. Conversely, silencing BIRC5 could inhibit the proliferation of tumor cells and promote apoptosis [
26,
30]. Recently, researchers have explored the combination of BIRC5 inhibitors with immune checkpoint inhibitors or monoclonal antibodies, which has shown promising results in inhibiting the growth of mouse xenograft tumors [
31]. It is generally believed in previous studies that the BIRC5 gene is a driving force for tumorigenesis and a known inhibitor of apoptosis [
28]. However, in our study, the flow cytometry results showed that the percentage of Annexin V-FITC and PI double-positive cells and PI single-positive cells was remarkably elevated after the BIRC5 inhibition compared with the control group in A549 cells. Therefore, these data suggested that the BIRC5 inhibition may have induced pyroptosis in A549 cells. Pyroptosis, also known as cell inflammatory necrosis, is a kind of programmed cell death that is characterized by cell size increase and formation of pores and plasma membrane disruption, which affects membrane integrity and increases the release of cell contents [
32]. This hypothesis was confirmed through an LDH release assay, which showed a significant increase in LDH release in the BIRC5 inhibition group. Mechanically, a key feature of pyroptosis is the involvement of gasdermin proteins as the main executive protein in cell death [
33]. The classic activation pathway of pyroptosis is triggered with the activation of caspase1, leading to the cleavage of the GSDMD [
34]. However, Western blot results revealed that the BIRC5 inhibition did not significantly affect the expression of cleaved caspase1 and GSDMD-N. The majority of BIRC5 was found to inhibit cell death by blocking the activation of caspase3, which aligns with our findings. Additionally, Shao et al. discovered that chemotherapy drugs can trigger pyroptosis through caspase3-mediated cleavage of Gasdermin E [
35]. Our results corroborate with this, as we observed a significant increase in the N-terminal fragments of Gasdermin E when BIRC5 was deleted. Furthermore, when GSDME was knocked down, it hindered the pyroptosis induced with BIRC5 inhibition.
In summary, our findings revealed the BIRC5 was identified and confirmed as a potential prognostic biomarker for LUAD. We found that inhibiting BIRC5 led to the induction of pyroptosis in lung adenocarcinoma cells through the caspase3/GSDME pathway. This study challenges the conventional understanding of programmed cell death, specifically, the role of caspase3 as the hallmark of apoptosis. Based on our findings, we hypothesize that the expression or expression level of GSDME plays a crucial role in determining the form of cell death in caspase3-activated cells. Pyroptosis was quickly induced after the cleavage of caspase3 in GSDME-enriched cells; however, apoptosis was dominant in low GSDME expression cells. While our study is perspective and efficient in identifying hub genes in lung adenocarcinoma, it is important to acknowledge the presence of certain limitations. For instance, it would be advantageous to incorporate a broader range of lung adenocarcinoma cell lines to explore the correlation between BIRC5 inhibition-induced pyroptosis and cellular GSDME expression levels. This would help to determine if the inhibition of BIRC5 induces pyroptosis via the caspase3-GSDME pathway in all lung adenocarcinoma cell lines. Furthermore, the increasing number of genetic screening results for clinical lung adenocarcinoma patients reveals a significant variability among individuals. However, it is important to note that our study evaluated only five clinical samples, which may limit the comprehensiveness of our assessment. In future research, we recommend increasing the sample size and include different types of samples. This will enable us to gain a more comprehensive understanding of the role of BIRC5 in lung adenocarcinoma and identify new personalized treatment targets for the clinical diagnosis and management of lung cancer. Finally, it is crucial to explore the influence of DNA methylation on BIRC5/Survivin expression. In future experiments, we plan to utilize sequencing and other techniques to identify key regulatory enzymes responsible for mediating methylation changes in BIRC5 [
36]. This study conducted a comprehensive bioinformatics analysis of MeDEGs and identified BIRC5 as a potential prognostic factor for LUAD. The findings of this study suggest that the inhibition of BIRC5 can induce cell death through caspase3/GSDME-mediated pyroptosis in lung adenocarcinoma. These findings offer potential therapeutic targets and candidate biomarkers and contribute to our understanding of the pathogenesis of lung adenocarcinoma.
4. Materials and Methods
4.1. RNA-Seq and Microarray Data
In the initiation of this study, we downloaded the lung adenocarcinoma methylation microarray and expression microarray datasets from publicly available Gene Expression Omnibus database (GEO) of NCBI (
https://www.ncbi.nlm.nih.gov/gds/; accessed on 9 January 2021). Finally, expression microarray datasets (GSE118370) and methylation microarray datasets (GSE139032) were used for biological information analysis. A total of 6 pairs of LUAD tumors and matched non-tumor tissues were enrolled in GSE118370 (Platform: GPL570, Affymetrix Human Genome U133 Plus 2.0 Array). The DNA methylation dataset (GSE139032), which was based on the GPL8490 platform (Illumina HumanMethylation27 BeadChip), included 77 primary lung adenocarcinoma cancer tissues and 77 adjacent normal lung tissues.
4.2. Identification of MeDEGs
The DEGs between tumor and non-cancerous samples in expression microarray datasets and DMGs in methylation microarray datasets were screened using GEO2R, which is an online analyzing tool that allows users to compare two groups of samples under the same experimental conditions in a GEO Series. Adj. p-value < 0.01 and |t| > 2 were set as the threshold values to screen aberrant expression genes. Overlapping genes from GSE1118370 and GSE139032 were identified with FunRich software (Functional Enrichment analysis tool). As a result, the overlapping genes were filtered out as methylation-regulated differentially expressed genes (MeDEGs), including upregulated and hypomethylated, and downregulated and hypermethylated genes.
4.3. Functional and Pathway Enrichment Analysis of MeDEGs
FunRich (Functional Enrichment analysis tool) is an online analysis tool consisting of an integrated biological knowledgebase and analytic tools providing a comprehensive set of functional annotation information from large gene/protein lists. In order to explain the function of the two gene lists, the Gene Ontology (GO) function and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were determined using FunRich, and a p value of <0.05 was considered statistically significant.
4.4. Construction of Protein–Protein Interaction (PPI) Network, Module Analysis and Hub Genes Screening
The PPI network of previously identified methylation-regulated differentially expressed genes was predicted and built using STRING online database (
http://string-db.org; accessed on 10 January 2021). An interaction with a combined score of >0.4 and
p < 0.05 were seen to be considered as the threshold value. Cytoscape software (version 3.7.0;
http://www.cytoscape.org/; accessed on 10 February 2021) is an open-source bioinformatics software platform that is utilized to integrate, analyze and visualize molecular interaction networks. The CytoHubba of Cytoscape is an application for screening the hub genes based on the total amount of degrees it possesses. A degree of more than 10 was used as cutoff criteria for hub gene identification.
4.5. Validation and Survival Analysis of Hub Genes in Other Datasets
Gene Expression Profiling Interactive Analysis (GEPIA;
http://gepia.cancer-pku.cn/; accessed on 16 February 2021) is a straightforward and easy-to-use online tool for the integration and visualization of gene expression based on The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx). This database was used to perform the expression level, tumor staging and overall survival analysis of hub genes in cancer. DNA methylation was validated in DiseaseMeth (
http://biobigdata.hrbmu.edu.cn/diseasemeth/; accessed on 19 February 2021), a database of DNA Methylation in human cancer based on The Cancer Genome Atlas (TCGA) database.
4.6. Cell Culture and Cell Transfection
Human A549 (lung carcinoma) cells were purchased from the American Type Culture Collection (ATCC) and maintained in RPMI1640 (Hyclone, Logan, UT, USA) medium containing 10% fetal bovine serum (Gibco, Waltham, MA, USA) and 1% penicillin-streptomycin-glutamine (Solarbio, Beijing, China). The cells were incubated at 37 °C in humidified incubator containing 5% CO2. Cell transfection was performed after the cells reached 60% of confluence. ShRNA knockdown for GSDME and negative controls were purchased from GENECHEM (Shanghai, China). Transfection of plasmids was carried out by mixing them with LipofectamineTM 2000 reagent (Invitrogen, Carlsbad, CA, USA). ShRNA knockdown lentivirus for BIRC5 and their controls were synthesized with GENECHEM (Shanghai, China). The A549 cell line was infected with lentiviral particles to knockdown BIRC5. For the negative control, only a lentiviral vector was used. A549 cells were cultured for 48 h following infection and selected via fluorescence using flow cytometry (BECKMAN COULTER MoFlo XDP, Brea, CA, USA) for subsequent experiments.
4.7. Cell Viability Assay
A549 cells were transfected with lentiviral particles to knockdown BIRC5. After 48 h, the cells were collected and seeded at a density of 5 × 103 cells per well into 96-well plates in DMEM containing 10% FBS and grown overnight. On the second day, 10 μL CCK-8 (APExBIO, Houston,TX, USA) solution was added to each well and incubated at 37 ℃ with 5% CO2 for 2 h. The absorbance values were measured at 450 nm wavelength using an enzyme-linked immunosorbent assay reader (BioTek, Thermo Fisher Scientific, Waltham, MA, USA). Results are presented as percentages relative to control cells.
4.8. LDH Release Assay
A549 cell culture medium was collected from various treatment groups and centrifuge at 1000 rpm for 5 minutes to eliminate impurities and cellular debris. The levels of released LDH in the A549 cell culture supernatants were measured with the LDH cytotoxicity assay kit (Solarbio, Beijing) according to the manufacturer’s instructions. The absorbance was measured at 490 nm with a microplate reader (BioTek, Thermo Fisher Scientific, Massachusetts, USA).
4.9. Western Blot Analysis
The tumor tissue and cells were lysed with RIPA buffer (Solarbio, Beijing, China) and contained a protease inhibitor. Total protein concentrations were quantified with BCA protein assay kit (Solarbio, Beijing, China). Equal amounts of protein samples were loaded in SDS-PAGE and transferred onto polyvinylidene fluoride membranes (PVDF, Millipore, Burlington, MA, USA). Then, the membrane was blocked with quick blocking solution (Epizyme, Shanghai, China) at room temperature for 15 min and incubated at 4 ℃ overnight with anti-BIRC5 rabbit monoclonal antibody (1:500; #2808S, Cell Signaling Technology, Danvers, MA, USA); anti-caspase1 rabbit monoclonal antibody (1:1000; ab207802, abcam, Cambridge, UK); anti-caspase3 rabbit monoclonal antibody (1:1000; ab32351, abcam, Cambridge, UK); anti-GSDMD rabbit monoclonal antibody (1:1000; ab209845, abcam, Cambridge, UK); anti-GSDM rabbit monoclonal antibody (1:1000; ab222408, abcam, Cambridge, UK) and anti-GAPDH rabbit monoclonal antibody (1:1000; TA309157, ZSGB-BIO, Beijing, China). Next, the membrane was incubated with HRP-conjugated secondary antibody for 2 h at room temperature (1:10,000; Epizyme, Shanghai, China). Subsequently, the membrane was incubated with an HRP-conjugated secondary antibody for 2 h at room temperature. The visualization of the membrane was achieved by using ECL detection reagent (BMU102-CN, abbkine, Wuhan, China) and the image was captured using the Bio-Rad ChemiDoc XRS+ chemiluminescence imaging system (Bio-Rad, Hercules, CA, USA).
4.10. RNA Extraction and qRT-PCR
Total RNA was extracted from tissues using the Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to manufacturer’s protocol, and 1 µg RNA was reversely transcribed into cDNA with PrimeScript™ RT reagent kit (RR047A, Takara, Otsu City, Shiga Prefecture, Japan) according to the manufacturer’s instructions. Quantitative RT-PCR was performed with SYBR Green (4913914001, Roche, Basel, Switzerland) using QuantStudio3 Real-Time PCR instrument (Thermo Fisher Scientific, Massachusetts, USA). PrimerDesign was performed with Invitrogen (Invitrogen, Carlsbad, CA, USA) and GAPDH was used as house-keeping gene. Fold changes were calculated using the 2
−ΔΔCT method. A full list of the primers used in qPCR analyses is shown in
Table S2.
4.11. Flow Cytometry Analysis
The treatment of A549 cells was transfected for 48 h and apoptosis assays were conducted using a BD Pharmingen Annexin V/PI Apoptosis Detection Kit I (BD Biosciences, Franklin Lake, NJ, USA). Briefly, A549 cells with different treatments were harvested, washed with PBS and diluted to a concentration of 1 × 104 in 100 μL 1 × binding buffer. A total of 5 μL FITC-conjugated annexin V were added to the cell suspension and incubated for 5 min in 5 μl propidium iodide (PI) for 15 min at 37 °C in darkness. The cells were analyzed with flow cytometry (BECKMAN COULTER MoFlo XDP, Brea, CA, USA).
4.12. Immunohistochemistry
Lung adenocarcinoma tumor samples and paracancer tissue samples were collected from patients with primary lung adenocarcinoma cancer from The Third Affiliated Hospital of Harbin Medical University. All samples were obtained with informed patients’ consent and approved by the Ethics Committee of Harbin Medical University Pharmacy College (No. IRB5004822). Formalin-fixed tissues were paraffin embedded and sectioned; then, the sections were deparaffinized in xylene, hydrated in graded alcohol solutions and heated (100 °C) for 15 min in antigen retrieval citra solution (pH 6.0) for antigen retrieval. A total of 3% hydrogen peroxide solution (Sigma-Aldrich, Saint Louis, MO, USA) in PBS with 0.3% Triton X-100 was used to block endogenous peroxidase activity for 10 min at room temperature, washed twice with PBS and blocked with goat serum for 15 minutes at 37 ℃. The sections were incubated with BIRC5 primary antibody (1:100; #2808S, Cell Signaling Technology, USA) overnight at 4 ℃, followed by 2 h incubation with the corresponding horseradish peroxidase-conjugated secondary antibody. Then, the images were obtained and analyzed with microscopy (X71, Olympus, Tokyo, Japan).
4.13. Animals and Animal Model
Six-week-old BALB/c nude female mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd (Vital River, Beijing, China). Mice were maintained in sterilized cages with 12/12 h light/dark cycle environment, 22 ℃ constant temperature and humidity of 55–60%, and were fed autoclaved food and water ad libitum. All animal experimental procedures and protocols conformed to the guide rules of the Animal Ethics Committee at Harbin Medical University. For xenografts, total 24 BALB/c nude mice were randomly divided into three groups (n = 8). The 2 × 107 of A549 cells (control, si-BIRC5, si-NC) in 200 uL cold PBS (Hyclone Laboratories, Logan, UT, USA) were subcutaneously injected into the right shoulder of mice to form a xenografts animal model. Tumor volumes were measured every day after the inoculation. Tumor volume was measured with a digital caliper and calculated with the following equation: V = L × W2 × 0.5 (V is the volume, L is the length, and W is the width). Tumor-bearing mice were sacrificed under anesthesia 21 days after tumor inoculation and tumor tissue was explanted for weight measurement. For the orthotopic mouse model, A549 cells were harvested and resuspended in PBS medium at a final concentration of 107 cells/mL. The animals were anesthetized with an intramuscular administration of 25 μL Zoletil™ 50. A LS-BA m blade assembly mouse laryngoscope (Penn-Century, Inc., Wyndmoor, PA, USA) was then used to expose the glottis, and a homemade syringe with a diameter of 0.4 mm was used to reach into the main trachea. Subsequently, tumor cells (50 μL) were inoculated into the lungs. After orotracheal intubation, the animals were kept in a stable environment with controlled temperature and humidity, as described above. The investigation conformed to the guide for the rules of the Declaration of Helsinki of 1975, revised in 2013, and was approved by the Institutional Animal Care and Use Committee of Harbin Medical University (No. IRB5004822).
4.14. Micro-PET/CT Imaging
FDG was produced routinely in GE TracerLab with a commercially available module following the standard protocols with quality control and radiochemical purity reaching 99%. Six hours before the imaging session, the mice were fasted, and water was given ad libitum. Approximately 200 μCi 18F-FDG was injected into tumor-bearing mice via the tail vein. A total of 15 min static scanning was performed at 1 h after injection with mice under isoflurane (R510-22-10, RWD, Shenzhen, China) anesthesia using a SuperArgus small-animal PET/CT scanner (Sedecal, Madrid, Spain). Images were reconstructed and regions of interest (ROI) were manually drawn around tumors.
4.15. Cell Migration Assays
For the migration studies, A549 cells were transfected with lentiviral particles to knockdown BIRC5. After 48 h, 1 × 105 A549 cells were resuspended in 100 µL serum-free 1640 RPMI medium and transferred into the upper chamber of 24-well transwell chamber (Corning, New York, NY, USA). Then, 500 μL complete 1640 RPMI with 10% FBS as a chemoattractant was added to the lower chambers. The plates were incubated at 37 °C with 5% CO2 for 24 h. Next, the cells adhering to the upper surface of the membrane were removed with a cotton swab. The migration cells, which adhered to the lower surface, were stained with 0.5% crystal violet for 15 min according to the manufacturer’s instructions. The images were captured with microscopy (X71, Olympus, Tokyo, Japan). The number of cells was counted randomly at ×200 magnification.
4.16. Statistical Analyses
Statistical analyses were performed using GraphPad Prism 8. Data are presented as mean ± SEM. Student’s t-test was performed to analyze data for experiments with two groups. One-way ANOVA was used to compare multiple experimental groups. A p value < 0.05 was considered as statistically significant.