Kaplan-Meier plotter (kmplot.com), which compiles publicly available datasets for analysis, was used to examine the effect of GPM6B and HPGD expression on the overall survival (OS) of patients with LUAD. The survival analysis included multiple regression analyses considering pathological type, age, gender, and smoking history, with the median expression level serving as the cutoff value. The Gene Expression Profiling Interactive Analysis (GEPIA; gepia.cancer-pku.cn/) and Gene Expression Omnibus (GEO; ncbi.nlm.nih.gov/geo/) (31) databases were used to determine the relative expression of GPM6B and HPGD in LUAD and normal tissues. The online software UALCAN (ualcan.path.uab.edu/analysis.html) was used to analyze the differential expression of GPM6B with respect to tumor, node, metastasis stage and nodal metastasis status, as well as to assess the levels of DNA methylation in the promoter region of GPM6B.

Spliced Transcripts Alignment to a Reference (STAR)-count data and clinical information relevant to LUAD were obtained from The Cancer Genome Atlas (TCGA) database (portal.gdc.cancer.gov) (32). LUAD dataset (portal.gdc.cancer.gov/projects/TCGA-LUAD) was used for analysis. The data were converted into transcript per million (TPM) format, and log2 (TPM + 1) normalization was performed.

R package ‘rms’ was used to construct a nomogram featuring clinical node stage, sex, age, tumor stage and the expression of GPM6B or HPGD. The analysis was performed using R version 4.0.0 (R Core Team, 2020). Through Cox regression analysis, a nomogram model was developed to evaluate the prognostic significance of these features across 476 samples and predict the survival rates of patients at 1, 3 and 5 years. LUAD samples was acquired from TCGA-LUAD dataset (portal.gdc.cancer.gov/projects/TCGA-LUAD) through the GDC Data Portal (portal.gdc.cancer.gov).

LUAD and adjacent normal tissues were collected from the Pathology Department of Chaohu Hospital, Chaohu, China. The adjacent normal tissues were 1 cm from the margin of the LUAD tissue. Samples were obtained during the preliminary diagnosis of patients between January 2023 and January 2025, all of whom were diagnosed with LUAD based on pathological examination. Verbal informed consent was obtained from all participants. The study included six patients (four males and 2 female; age, 54–77 years) with a pathological diagnosis of lung adenocarcinoma were enrolled in the study. All procedures were approved by the Ethics Committee of Chaohu Hospital of Anhui Medical University (approval no. KYXM-202410-010). Tissue samples were used for immunohistochemistry (IHC) analysis.

Human LUAD A549 and PC9 and 293T cells were obtained from Procell Life Science and Technology Co., Ltd. All cells were cultured in DMEM (Biosharp Life Sciences) supplemented with 10% (v/v) fetal bovine serum (FBS; Shanghai ExCell Biology, Inc.) and 100 U/ml penicillin0streptomycin (Biosharp Life Sciences) at 37°C with 5% carbon dioxide.

GPM6B and HPGD coding sequences were downloaded from National Center for Biotechnology Information (NCBI; ncbi.nlm.nih.gov) and cloned into the pLJM1-EGFP vector (cat. no. 19319; Addgene, Inc.) via AgeI and EcoRI enzyme sites. Coding sequences of GPM6B coupled with HA tag sequences were cloned into pLVX vector (cat. no. 632164; Clontech) via BamH1 and EcoRI enzyme sites. The lentivirus were packaged with 2nd generation system. Briefly, the transfection of vectors into 293T cells was conducted at room temperature with a mixture consisting of 5.0 overexpression construct, 5.0 p8.9 packaging plasmid and 2.5 µg VSVG envelope plasmid utilizing the Hieff Trans™ Liposomal Transfection Reagent (Shanghai Yeasen Biotechnology Co., Ltd.). The lentiviral particles were collected following 48 h transfection. A549 or PC9 cells were infected with in a 6-well plate with a multiplicity of infection of 0.5. After 10 h of infection, antibiotic selection with 2 µg/ml puromycin dihydrochloride (Biosharp Life Sciences) was used to generate the stable cell lines. Following selection for 72 h, the exogenous GPM6B-overexpressing stable A549 and PC9 cells were used for subsequent experiments. The negative control for GPM6B-overexpressing stable A549 cells was the A549 cell line transfected with the pLJM1 backbone plasmid. Similarly, the negative control for GPM6B-overexpressing stable PC9 cells was the PC9 cell line transfected with the pLJM1 backbone plasmid. Stably transfected A549 and PC9 cells were maintained with 1 µg/ml puromycin.

Total RNA was extracted from 1×10⁶ A549 or PC9 cells using the SPARKeasy Tissue/Cell RNA kit (Shandong Sparkjade Scientific Instruments Co., Ltd.) and cDNA was synthesized using the MonScript™ RTIII All-in-One Mix with dsDNase (Monad Biotech Co., Ltd.), according to the manufacturer's instructions. qPCR was performed using the ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotech Co., Ltd.). Thermocycling conditions were as follows: Initial denaturation for 30 sec at 95°C for initial denaturation, followed by 40 cycles of 10 sec at 95°C and 30 sec at 60°C. Fold-changes in mRNA levels were determined using the comparative 2^−ΔΔCq method (33), with GAPDH or β-actin as the internal reference gene. The primer sequences used were as follows: GAPDH: forward, 5′-CAACTGCTTAGCACCCCTGG-3′ and reverse, 5′-GTCAAAGGTGGAGGAGTGGG-3′; β-actin: forward, 5′-CATGTACGTTGCTATCCAGGC-3′ and reverse, 5′-CTCCTTAATGTCACGCACGAT-3′; creatine kinase, mitochondrial 1A (CKMT1A): Forward, 5′-TCTCCTCCAGCACAGGAACT-3′ and reverse, 5′-AAATGGGGGTCGGATTGGAG-3′; DSTN pseudogene 2 (DSTNP2): Forward, 5′-CATCAGATGCCTTTGAGACATACA-3′ and reverse, 5′-TCTGGTGTGGAGCATTTACGAAC-3′; TNFSF12-TNFSF13 readthrough (TNFSF12-TNFSF13): Forward, 5′-GAAGCCAGAATCAACAGCTCC-3′ and reverse, 5′-GCATCGGAACTCTGACAGTACAG-3′; eukaryotic translation initiation factor 4E binding protein 3 (EIF4EBP3): Forward, 5′-TGTCAACGTCCACTAGCTGC-3′ and reverse, 5′-CTCCAGCAGGAACTTTCGGT-3′; Polyamine modulated factor 1 binding protein 1 (PMFBP1): forward, 5′-CACGCCTTTGACAAGAAGCTA-3′ and reverse, 5′-TTGAGGTCATTCTGGTGTTGC-3′; Protocadherin alpha 12 (PCDHA12): forward, 5′-AGAGAGCAAACGCCAAAACTC-3′ and reverse, 5′-CACATCCAGGACGGTTATTTGA-3′; Oxidized low density lipoprotein receptor 1 (OLR1): 5′-ACTCTCCATGGTGGTGCCTGG-3′ and reverse, 5′-GCTTGTTGCCGGGCTGAGATCT-3′; Ankyrin repeat domain 1 (ANKRD1): Forward, 5′-AGCGCCCGAGATAAGTTGCT-3′ and reverse, 5′-CACCAGATCCATCGGCGTCT-3′; Inhibin subunit βA (INHBA): Forward, 5′-CAACAGGACCAGGACCAAAGT-3′ and reverse, 5′-GAGAGCAACAGTTCACTCCTC-3′; GPM6B: Forward, 5′-CGTGGCGATTCTTGAGCAAC-3′ and reverse, 5′-CAGGCCACTCCAAGCACATA-3′; HPGD: Forward, 5′-AGCAGCCGGTTTATTGTGCTT-3′ and reverse, 5′-GCGTGTGAATCCAACTATGCC-3′ and EZH2: Forward, 5′-GGACCACAGTGTTACCAGCAT-3′ and reverse, 5′-GTGGGGTCTTTATCCGCTCAG-3′.

Protein extracts from 1×10⁶ A549 or PC9 cells were prepared on ice using RIPA (cat. no. PR20035; Proteintech Group, Inc.) lysis buffer containing proteinase inhibitor. The total protein concentration was determined using a BCA Protein Quantification kit (Vazyme Biotech Co., Ltd.). Proteins were separated using 8% SDS-PAGE with 2 mg protein/well and transferred onto PVDF membranes. After 1 h blocking in 5% skimmed milk at room temperature, the PVDF membranes were incubated with primary antibodies against GAPDH (cat. no. GB15002; Servicebio; 1:10,000), GPM6B (cat. no. AB03180; Qizhidao Technology Co., Ltd.; 1:1,000 dilution), HPGD (cat. no. 66798-1-Ig; Wuhan Sanying Biotechnology, 1:5,000 dilution), P53 (cat. no. HY-P80257; MedChemExpress; 1:1,000 dilution), Cyclin D1 (CCND1; cat. no. 380999), Bcl2 (cat. no. R23309; Chengdu Zen-Bioscience Co., Ltd.; 1:1,000 dilution) and Bax (cat. no. 200958; all Chengdu Zhengneng Biotechnology Co., Ltd.; all 1:1,000) overnight at 4°C. The membranes were incubated the following horseradish peroxidase (HRP)-conjugated secondary antibodies: goat anti-rabbit IgG-HRP (cat. no. BL003A, White Shark Biotechnology Co., Ltd.; 1:10,000) and goat anti-mouse IgG-HRP (cat. no. BL001A, White Shark Biotechnology Co., Ltd.; 1:10,000 dilution) at room temperature for 2 h. The protein signals were detected using the SuperKine™ West Femto Maximum Sensitivity Substrate (Abbkine Scientific Co., Ltd.) and images were acquired using a multi-function imaging system (Vilber Lourmat).

Cell proliferation was evaluated using the Cell Counting Kit-8 (CCK-8) assay (Biosharp Life Sciences). Briefly, 2×10³ cells were seeded into 15 wells of a 96-well plate. A total of 100 µl medium containing 10 µl CCK-8 regent was added at 1–5 days after seeding and the optical density at 450 nm was measured after 2 h.

Ki67 (cat. no. 350504; BioLegend, Inc.) is a commonly utilized marker for assessing the tumor proliferation index (34–36). was. Following 72 h infection, the A549 and PC9 cells were harvested and subjected to two washes with PBS, fixed for 3 h at 4°C using cold 70% ethanol (Chinasun Specialty Products Co., Ltd.) and washed twice more with PBS. The cells were stained with PE anti-mouse/human Ki-67 (cat. no. 350504; BioLegend, Inc.; 1:20) at room temperature for 15 min. Flow cytometry analysis was performed using the phycoerythrin detection channel on a Beckman cytometer (Beckman Coulter, Inc.) to quantify the Ki67 signals. The results were analyzed using CytoExpert 2.4 software (Beckman Coulter, Inc.).

For the wound healing assay, 5×10⁵ A549 or PC9 cells were seeded in a six-well plate. A scratch was made on the surface of the cell layer when the cell confluence reached 100%. The cells were maintained in DMEM containing 1% FBS after scratching. The cells were imaged every 24 h under a light microscope (ECLIPSE Ts2, Nikon), and the migration distance was calculated. For the Transwell migration assay, 5×10⁴ A549 or PC9 cells were seeded in the upper chambers (8 µm) and maintained in serum-free DMEM, whereas DMEM supplemented with 10% FBS and 100 U/ml penicillin/streptomycin was added to the lower chambers. For the Transwell invasion assay, the bottom surface of each upper chamber was coated with Matrigel (BD Biosciences) for 1 h at 37°C. A total of 5×10⁴ cells was seeded in the upper chambers. The cells were cultured at 37°C for 14–16 h and fixed with 100% methanol at 4°C for 20 min, stained the cells with 0.1% crystal violet at room temperature for 20 min and the imaged with inverted light microscope (ECLIPSE Ts2, Nikon).

A total of 10 ml of lentivirus packaged from the pLVX overexpression construct, along with 5.0 µg of the p8.9 packaging plasmid and 2.5 µg of VSVG, were used to infect 3×10⁶ 293T cells seeded in a 10-cm dish. After 72 h of infection, 1×10⁷ cells were harvested and lysed with 1 ml lysis buffer (cat. no. HY-K1000; MedChemExpress) at 4°C with rotation for 4 h. The cell lysate was then centrifuged at 15,000 g for 15 min at 4°C to collect the supernatant. A total of 1 µg of the primary antibody against IgG (cat. no. B900620; Proteintech) or HA (cat. no. 66006-2-Ig; Proteintech) as well as 50 µl Protein A/G agarose beads to the cell supernatant for incubation at 4°C overnight. Following incubation, the Protein A/G agarose beads (cat. no. PR40025; Proteintech) were washed three times with PBS at 4°C. The supernatant was then discarded, and 50 µl of 2×SDS (cat. no. P0015B; Beyotime) was added to the agarose beads. After vortexing at 500 × g for 10 sec at room temperature, the mixture was placed in a metal bath at 95°C for 10 min. Finally, 10 µl of each sample was taken for Western blot experiments.

Numerous studies have used the A549 cell line to establish tumor-bearing models (37–39). A total of 12 BALB-c nude mice were divided into a control group and a GPM6B-overexpressing group, each group containing six 5-week-old female mice (Biomart). All mice exhibited comparable body weights (~15 g). Mice were housed at 20–26°C and the relative humidity is kept at 50–60%, 12/12-h light/dark cycle and ad libitum food and water. Mice were subcutaneously injected with ~2×10⁶ cells into the right forelimb 1 week after purchase. During the experiment, the mice were observed daily to detect any signs of illness or adverse reactions. Tumor volume was measured every 3 days. At 33 days post-injection, the mice exhibited signs of anxiety, along with a noticeable decrease in appetite. Therefore, euthanasia was performed by cervical dislocation following anesthesia by inhalation of 2% isoflurane (cat. no. R510-22-10, RWD Life Science Co., Ltd.) in oxygen until the loss of eyelid reflexes and the onset of muscle relaxation were observed. Death was confirmed by cessation of pulse and breathing. Tumors were dissected and weighed. All animal care and experimental procedures adhered to ARRIVE guidelines and were approved by the Ethics Committee of Animal Experiments of Anhui Medical University (approval no. LLSC20242301).

Paraffin-embedded clinical samples and tumor tissue from mice were sliced into 5-µm thick sections and rehydrated using descending alcohol concentrations. Antigen retrieval was performed using a sodium citrate retrieval solution with a pH of 6.0 at 100°C then allowed to cool. This heating and cooling cycle was repeated twice. Slides were washed three times with PBS at room temperature. Sections were sealed with 5% bovine serum albumin (Wuhan Servicebio Technology Co., Ltd.) at room temperature for 45 min. The clinical samples were incubated with antibodies against GPM6B (cat. no. PC11847S; Abmart Pharmaceutical Technology Co., Ltd.; 1:400) and HPGD (cat. no. 66798-1-Ig; Proteintech Group, Inc.; 1:400), while the mouse tissue samples were incubated with Ki67 primary antibody (cat. no. A11390; ABclonal Biotech Co., Ltd.; 1:500) overnight at 4°C. The sections were washed three times with PBS for 5 min each and incubated with goat anti-rabbit IgG-HRP (cat. no. BL003A; Hefei White Shark Biotechnology Co., Ltd.; 1:300) or Goat anti-mouse IgG-HRP (cat. no. BL001A; Hefei White Shark Biotechnology Co., Ltd.; 1:300) at 25°C for 50 min. The samples were washed with PBS, stained with diaminobenzidine and counterstained with hematoxylin (Wuhan Servicebio Technology Co., Ltd.) at room temperature for 3 min. Finally, ascending ethanol dehydration and resin sealing were performed. For H&E staining, paraffin-embedded mouse tissue sections were stained with hematoxylin at room temperature for 5 min. This was followed by eosin at room temperature for 15 sec. Finally, sections were visualized using a light microscope (ECLIPSE Ts2, Nikon).

Human EZH2 siRNA was produced by Beijing Tsingke Biotech Co., Ltd. and transfected into LUAD cells. In brief, 20 µM siRNA was transfected into 3×10⁵ A549 or PC9 cells at room temperature using Lipofectamine 2000 (cat. no. CN2541156; Invitrogen; Thermo Fisher Scientific, Inc.). After 8 h of transfection, the medium was replaced with fresh complete DMEM and the cells were cultured at 37°C for an additional 48 h before proceeding with subsequent experiments. The siRNA sequences were as follows: Negative control, 5′-UUCUCCGAACGUGUCACGU-3′; siEZH2#1, 5′-GAGGGAAAGUGUAUGAUAATT-3′ and siEZH2#2, 5′-GAGGUUCAGACGAGCUGAU-3′. Decitabine (DCA; cat. no. HY-A0004) and trichostatin A (TSA; cat. no. HY-15144; both MedChemExpress) stocks were prepared in dimethyl sulfoxide. LUAD cells were treated with 1, 3 and 5 TSA for 24 h, and 100, 500 and 1,000 µM DCA at 37°C for 48 h.

Total RNA from control and GPM6B overexpressing PC9 cells was extracted using the SPARKeasy Tissue/Cell RNA kit (cat. no. AC0205-B; Shandong Sparkjade Scientific Instruments Co., Ltd.). Library construction and RNA-sequencing (RNA-Seq) were performed by Shanghai Meiji Biopharmaceutical Technology Co., Ltd. with Illumina NovaSeq 6000 (Illumina Inc.). The sequencing library was prepared using the Stranded Total RNA Prep with Ribo-Zero Plus kit (cat. no. 20040529; Illumina, Inc.). RNA samples meeting the following criteria were selected for library preparation: ≥1 µg total RNA at a concentration of ≥30 ng/µl, with an RNA Quality Number >6.5 (Agilent 5300) and A260/A280 ratios of 1.8–2.2 (Nanodrop 2000). RNA integrity was verified by agarose gel electrophoresis. Libraries were prepared using the Illumina^® Stranded mRNA Prep Kit (Illumina) and quantified via Qubit fluorometry (cat. no. Q33327; Thermo Fisher), adjusting the final working concentration to 3 ng/µl. Finally, paired-end sequencing with a read length of 150 bp was performed.

After acquiring gene read counts, a differential expression analysis was conducted across multiple samples (≥2) to identify differentially expressed genes (DEGs) between sample groups, thereby facilitating subsequent functional studies of these DEGs. The analysis was carried out using DESeq2 (version: 1.44.0: http://bioconductor.org/packages/stats/bioc/DESeq2/). The criteria for significant differential expression were set as follows: P<0.05 and |log2FC|≥1.

GO enrichment analysis was conducted using GOATools software (https://github.com/tanghaibao/GOatools), employing the Fisher's exact test as the statistical method. To control for false positives, four multiple testing correction methods-Bonferroni, Holm, Sidak and false discovery rate-were applied to adjust the p-values. Generally, a Gene Ontology (GO) term is deemed significantly enriched when the adjusted P-value is less than 0.05.

The Python SciPy package (version: 3.7.0; scipy.org/install/) was used for KEGG pathway enrichment analysis. To control the false positive rate, the Benjamini-Hochberg (HR) method was implemented for multiple testing correction. A corrected P-value threshold of 0.05 was established, with KEGG pathways meeting this criterion classified as significantly enriched.

The GSEA analysis was conducted using the Majorbio Cloud Platform with standardized protocols (40). The platform's built-in preprocessing ensured proper gene ID conversion and low expression filtering prior to analysis. Gene expression data were analyzed against MSigDB Hallmark gene sets. Significant pathways were identified using the following thresholds: |NES|>1, FDR (q-value) <0.25, and P<0.05.

Chromatin immunoprecipitation-sequencing data for H3K27ac in IMR90 (ENCFF694JJX) and A549 (ENCFF701DZR) cells were obtained from the ENCODE database (encodeproject.org/).

Data are presented as the mean ± standard deviation. All experiments were repeated at least twice. Data were analyzed using GraphPad Prism 9 (Dotmatics). One-way analysis of variance or unpaired two-tail student t-test was used to compare data. Tukey's Honestly Significant Difference test was performed as the post hoc analysis P<0.05 was considered to indicate a statistically significant difference.

Survival outcomes associated with GPM6B in patients with LUAD were analyzed using the Kaplan-Meier plotter. Higher levels of GPM6B expression were associated with improved OS (Fig. 1A). A nomogram was developed to assess the impact of GPM6B on the prognosis of patients with LUAD (Fig. 1B). Clinical node and tumor stages are the primary predictors of survival, while age, sex, and GPM6B expression levels contributed smaller risks. In the GEPIA database and two GEO datasets (accession nos. GSE10072 and GSE115002), expression of GPM6B was revealed to be decreased in LUAD tissues when compared with normal tissues (Fig. 1C and D). In addition, GPM6B expression levels were negatively associated with cancer stages and nodal metastasis status (Fig. 1E and F). IHC staining of GPM6B revealed decreased expression of GPM6B in LUAD tissues compared with adjacent normal tissue (Fig. 1G).

To elucidate the role of GPM6B in tumor pathogenesis, stable A549 and PC9 cell line with GPM6B overexpression were generated, as confirmed by RT-qPCR and western blot analysis (Fig. 2A and B). CCK-8 assay revealed that GPM6B overexpression significantly suppressed the proliferation of both LUAD cell lines at day 5 (Fig. 2C). Ki67 is a nuclear protein associated with cell cycle progression and proliferation (34). Flow cytometry revealed that Ki67 staining signal was lower in GPM6B-overexpressing LUAD compared with control cells (Fig. 2D). Additionally, Transwell assay demonstrated that GPM6B overexpression significantly suppressed migration and invasion of LUAD cells (Fig. 2E). Consistently, the wound healing assays confirmed the inhibitory role of GPM6B to cell migration in LUAD cells (Fig. 2F).

Transcriptome sequencing and data analysis were performed on GPM6B-overexpressing PC9 cells to identify the downstream genes and signaling pathways that are regulated by GPM6B. GPM6B affected the expression of 228 genes, including 120 upregulated and 108 downregulated genes (Fig. 3A). RT-qPCR was performed to identify the downstream genes, which revealed HPGD as a target gene (Fig. 3B). Furthermore, the protein levels of HPGD were increased in both LUAD cell lines overexpressing GPM6B (Fig. 3C). Survival analysis revealed that higher HPGD expression was associated with improved survival rates among patients with LUAD (Fig. 3D). Furthermore, the nomogram demonstrated an association between HPGD and clinical variables in predicting survival of patients with LUAD (Fig. 3E). Analysis of TCGA and GEO databases indicated that HPGD was downregulated in LUAD tissue (Fig. 3F and G). The expression of HPGD was lower in LUAD tissues than in normal tissue (Fig. 3H). Co-immunoprecipitation was used to further explore the relationship between GPM6B and HPGD, however, no interaction between GPM6B and HPGD was detected, suggesting that GPM6B regulated HPGD at the transcriptional level (Fig. 4A). Moreover, there was a positive association between HPGD and GPM6B expression levels (Fig. 4B and C). These findings collectively suggest that HPGD may be a target gene of GPM6B.

The aforementioned results indicated GPM6B serves as a tumor suppressor and promotes HPGD expression in LUAD cells. Therefore, the effects of HPGD on cell proliferation and migration were investigated using gain-of-function studies in A549 and PC9 cells overexpressing HPGD. RT-qPCR and western blot analyses confirmed the increased expression of HPGD in both LUAD cell lines (Fig. 5A and B). Cells with increased HPGD levels exhibited decreased proliferation and migration capacities (Fig. 5C-F), similar to the effects observed in cells with elevated GPM6B expression levels. Overall, these findings suggested that GPM6B exerts its tumor-suppressive effects by transcriptionally regulating HPGD.

Gene Ontology analysis demonstrated that the differentially expressed genes were associated with ‘regulation of apoptotic process’, ‘regulation of cell death’, ‘regulation of cell population proliferation’, ‘negative regulation of biological process’ and ‘tissue development’ (Fig. 6A). Kyoto Encyclopedia of Genes and Genomes analysis indicated that GPM6B expression was associated with ‘cellular senescence’, ‘cGMP-PKG signaling pathway’, ‘cytokine-cytokine receptor interaction’, ‘IL-17 signaling pathway’, ‘retinol metabolism’ and ‘p53 signaling pathway’ (Fig. 6B). In addition, Gene Set Enrichment Analysis revealed that DNA replication, glycolysis, gluconeogenesis and glutathione metabolism were positively enriched in both GPM6B-overexpressing cell lines (Fig. 6C). Notably, GPM6B was associated with the p53 signaling pathway, which is frequently dysregulated in a number of human cancer types (41). Western blot analysis demonstrated that GPM6B overexpressing cells exhibited changes in expression of proteins involved in the p53 signaling pathway, including decreased expression of Bcl-2 and CCND1 and increased expression of p53 and Bax (Fig. 6D). This suggested that GPM6B may inhibit tumor development in LUAD by activating the p53 signaling pathway.

Nude mouse xenografts with A549 cells overexpressing GPM6B were used to investigate the oncogenic role of GPM6B in LUAD progression (Fig. 7A). Tumor volume at day 33 and weight in the GPM6B-overexpressing mice were significantly lower compared with control mice (Fig. 7B-D). Compared with control mice, H&E staining in GPM6B-overexpressing mice revealed that dissected tissue cells had enlarged nuclei and the intercellular stroma was decreased, which were consistent with characteristics of tumor cells (Fig. 7E). Moreover, IHC analysis of the excised tumors revealed decreased Ki67 staining intensity in the GPM6B-overexpressing group (Fig. 7F). Taken together, these findings indicated that elevated GPM6B expression contributes to the decreased proliferative capacity of LUAD tumor cells.

As GPM6B expression was decreased in LUAD, the effect of DNA and H3K27 methylation and histone deacetylation on GPM6B levels was assessed. Disrupting expression of EZH2 did not alter expression of GPM6B, suggesting that EZH2-mediated H3K27 methylation was not responsible for the reduction in GPM6B expression levels (Fig. 8A). DNA methylation was subsequently evaluated at the GPM6B promoter region by exploring the UALCAN database. DNA methylation was significantly greater in LUAD compared with normal tissues (Fig. 8B). Histone acetylation at the H3K27ac site, which can be inhibited by HDAC proteins, promotes chromatin relaxation and gene expression (42,43). Chromatin immunoprecipitation-seq data of H3K27ac in normal human embryonic lung fibroblasts (IMR90 cells) and LUAD (A549 cells) were downloaded from the ENCODE database. Enrichment of H3K27ac at the GPM6B promoter region was decreased in A549 cells compared with IMR90 cells (Fig. 8C). To investigate the effects of DNA methylation and histone deacetylation on GPM6B expression, their specific inhibitors DCA and TSA were used to treat LUAD cells. Based on the CCK-8 assay results, 100, 500 and 1,000 µM DCA was selected for treatment for 48 h and 1, 3 and 5 µM TSA was selected for treatment for 24 h in both LUAD cell lines (Fig. 8D and E). RT-qPCR results revealed that the expression level of GPM6B increased in a dose-dependent manner in response to treatment with DCA or TSA (Fig. 8F and G). These observations indicated that both DNA methylation and histone deacetylation are associated with the downregulation of the tumor suppressor gene GPM6B in LUAD cells.