DNA hypomethylation-activated PRAME drives early-stage lung adenocarcinoma recurrence via ZNF740 and PI3K/AKT/mTOR signaling

Introduction

Non-small cell lung cancer (NSCLC) accounts for approximately 85% of all lung cancers, with lung adenocarcinoma (LUAD) as the predominant subtype (1). For patients with early-stage LUAD (pathological stages IA-IB), surgical resection remains the standard curative treatment and is associated with favorable 5-year survival rates of 70–90% (2,3). However, despite complete tumor removal, 13–30% of patients experience recurrence within 2 years post-surgery, predominantly in the form of distant metastases, such as to the bone, brain, and liver, rather than local relapse (4,5). Recurrence is linked to a dismal prognosis, with 5-year survival rates plummeting to <20% (6), highlighting an urgent need to understand the molecular mechanisms underlying early metastatic progression.

Early-stage LUAD (stage I) represents a critical clinical window: while most patients are cured by surgery, a subset experiences rapid recurrence due to occult micro-metastatic disease. The drivers of this aggressive phenotype remain poorly defined. In our recent analysis of early-stage LUAD samples, we identified that preferentially expressed antigen in melanoma (PRAME) was specifically upregulated in the recurrence group of stage I lung cancer (7), suggesting its potential role in early metastasis. Therefore, elucidating how PRAME contributes to tumor dissemination may enable improved risk stratification and inform early intervention strategies.

Emerging evidence underscores epigenetic dysregulation, particularly DNA methylation alterations, as a key contributor to cancer recurrence (8-10). PRAME, originally identified as a cancer-testis antigen capable of eliciting cytotoxic T lymphocyte (CTL) responses in melanoma (11), has since been recognized for its oncogenic functions beyond immune modulation. It is aberrantly overexpressed in diverse malignancies, including melanoma, acute myeloid leukemia, breast cancer, and LUAD (12-14), while exhibiting highly restricted expression in normal tissues, primarily confined to testicular germ cells (15). Functionally, PRAME promotes tumor progression, inhibits apoptosis, and facilitates immune evasion (16,17). In multiple cancers, elevated PRAME expression correlates with advanced stage, metastasis, and poor survival, supporting its role as a prognostic biomarker and potential therapeutic target.

Our previous epigenetic profiling identified that PRAME is one of the most significantly hypomethylated and transcriptionally upregulated genes in LUAD (7), implicating the loss of promoter DNA methylation in its aberrant activation. However, the precise regulatory mechanisms linking epigenetic remodeling to PRAME expression, and ultimately to metastatic recurrence, remain unclear. In this study, we systematically investigated the functional significance of PRAME in early-stage LUAD recurrence and elucidated its upstream epigenetic regulation and downstream signaling pathways. We demonstrate that DNA hypomethylation enhances chromatin accessibility at the PRAME promoter, enabling the recruitment of the zinc finger transcription factor ZNF740, which directly binds to and activates PRAME transcription. Furthermore, we show that PRAME overexpression drives malignant phenotypes associated with metastasis, including enhanced cell migration, invasion, and induction of epithelial-to-mesenchymal transition (EMT). Mechanistically, PRAME activates the PI3K/AKT/mTOR signaling pathway, a well-established driver of tumor progression and therapeutic resistance.

We propose the following mechanistic model: epigenetically driven DNA hypomethylation facilitates ZNF740 binding to the PRAME promoter, leading to transcriptional activation of PRAME, which in turn promotes early metastatic recurrence via activation of the PI3K/AKT/mTOR pathway and induction of EMT. Collectively, this axis provides novel insights into the molecular basis of early relapses in LUAD and identifies PRAME and its associated regulatory network as potential targets for precision intervention in high-risk patients. We present this article in accordance with the MDAR and ARRIVE reporting checklists (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1417/rc).

Methods

Patients and specimens

Patient samples were derived from a previously described cohort (7). Briefly, we prospectively enrolled patients diagnosed with LUAD who underwent curative surgical resection at West China Hospital, Sichuan University, between 2014 and 2020. None of the patients received neoadjuvant therapy prior to surgery. During surgery, primary tumor tissues and matched adjacent non-tumor lung tissues, collected from sites at least 5 cm away from the tumor margin, were obtained. All pathological diagnoses and tumor cellularity assessments were independently confirmed by two experienced pathologists. Cancer staging was performed according to the 8th edition of the American Joint Committee on Cancer (AJCC) tumor-node-metastasis (TNM) classification system. The current analysis included 28 patients with stage I LUAD (pathological stage IA-IB). Inclusion criteria were: (I) complete surgical resection without neoadjuvant treatment; (II) availability of paired tumor and normal tissue specimens; and (III) a minimum of 2 years of follow-up. Detailed clinical characteristics are provided in Table S1. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Institutional Review Board of West China Hospital, Sichuan University (No. 20251133), and individual consent for this retrospective analysis was waived.

Sample collection and preparation

All tissues were obtained from surgically resected specimens. Primary tumor tissues and matched adjacent non-tumor lung tissues (>5 cm from tumor margin) were prospectively collected for research purposes. To ensure consistency, both formalin-fixed paraffin-embedded (FFPE) [for pathology and immunohistochemistry (IHC)] and fresh-frozen (FF) samples (for molecular assays) were derived from the same tumor region. Adjacent non-tumor tissues were collected from the same lobe as the primary tumor. Tissue collection and processing followed a standardized protocol to minimize bias. Immediately after resection, tissues designated for FF storage were rapidly frozen in liquid nitrogen and transferred to −80 ℃ for long-term preservation. Those intended for FFPE processing were fixed in 10% neutral buffered formalin for 24–48 hours to ensure complete tissue penetration and optimal fixation, then dehydrated and embedded in paraffin wax to maintain structural integrity.

Library preparation for transcriptome sequencing and data processing

Total RNA was extracted from FF tissues using the AllPrep DNA/RNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. RNA quantity and quality were assessed using the Qubit^® 3.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) and the Agilent 4200 TapeStation system (Agilent Technologies, Santa Clara, CA, USA). For library preparation, 2 µg of total RNA per sample was used as input material. Strand-specific RNA sequencing libraries were constructed using the NEBNext^® Ultra™ RNA Library Prep Kit for Illumina^® (New England Biolabs, Ipswich, MA, USA) with unique index adapters incorporated for sample multiplexing. Libraries were sequenced on an Illumina Novaseq platform, generating 150 bp paired-end reads. Raw reads were then aligned to GRCh38 using BWA-MEM (version 0.7.17).

Transcriptomic analysis

Raw sequencing data were filtered using fastp (version 0.23.4) to remove low-quality reads, including those with adapter contamination, more than 1% unknown bases (N), or over 40% of bases with a quality score below 15. Clean reads were subsequently aligned to the human reference genome (GRCh38.p14, Ensembl release 104) using STAR (version 2.7.10b) under default parameters. Transcript-level read counts were quantified using featureCounts (version 2.0.4). The expression matrix was normalized using the Trimmed Mean of M-values (TMM) method implemented in the R package edgeR (version 4.0.16). Differentially expressed genes (DEGs) were identified using a generalized linear model (GLM) framework in edgeR, with significance defined as P<0.05 after Benjamini-Hochberg correction and an absolute log₂ fold change (|log₂FC|) >1. Functional enrichment analysis of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways was performed on the DEGs using the R package clusterProfiler (version 4.10.1).

Cell culture and reagents

PC9 (RRID: CVCL_B260) and A549 (RRID: CVCL_0023) parental cell lines were obtained from the American Type Culture Collection (ATCC); A549-PRAME-knockdown cells were generated in house, and knockdown efficiency has been previously validated and described (7). All cell lines were authenticated by short tandem repeat (STR) profiling upon receipt and generation. All cells were cultured in RPMI-1640 medium (Gibco, USA) supplemented with 10% fetal bovine serum (Gibco), and maintained at 37 ℃ in a humidified 5% CO₂ atmosphere. All cells were regularly tested for mycoplasma contamination throughout the study and confirmed to be mycoplasma-negative.

The PI3K inhibitor LY294002 (50 µM) and AKT inhibitor MK2206 (20 µM) were purchased from MedChemExpress (New Jersey, USA). Inhibitors were used for treatment of LUAD cells for all functional assays and Western blot (WB) analysis. Additionally, PC9 cells were treated with 10 µM 5-azacytidine (5-azaC) prior to Chromatin Immunoprecipitation (ChIP) and WB analysis.

RNA extraction and quantitative real-time polymerase chain reaction assay (qRT-PCR)

Total RNA was isolated from LUAD clinical specimens, PC9 cells and A549 cells using TRIzol^® Reagent (Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s protocol. RNA was reverse transcribed to cDNA using the iScript™ cDNA Synthesis Kit (Cat#1708891, Bio-Rad, Hercules, CA, USA). qRT-PCR was performed using iTaq™ Universal SYBR^® Green Supermix (Cat#1725124, Bio-Rad). Relative gene expression was calculated via the 2^−ΔΔCt method (GAPDH as endogenous reference). Full primer sequences are provided in Table S2. All assays were performed in three independent biological triplicates, with three technical replicates per biological replicate. All qPCR runs included no template and no reverse transcriptase negative controls.

WB analysis

Total cellular proteins from primary tumor tissues and PC9 and A549 cells were extracted using RIPA lysis buffer supplemented with a complete protease and phosphatase inhibitor cocktail. Following centrifugation, the supernatant was collected and protein concentration was quantified. Equal amounts of protein (30 µg in 20 µL) were mixed with loading buffer prior to electrophoresis.

Target proteins were separated on 4–20% gradient MOPS gels (electrophoresis: 80 V for 30 minutes, then 140 V for 60 minutes) and transferred to PVDF membranes using a semi-dry blotter (2.5 A, 25 V for 7 minutes). Membranes were blocked and then incubated overnight at 4 ℃ with the respective primary antibodies. Antibodies used were: PRAME, E-cadherin, N-cadherin, PI3K, Phospho-PI3K, AKT, Phospho-AKT, mTOR, Phospho-mTOR, Caspase 8, Cleaved Caspase 8, Caspase 3, Cleaved Caspase 3, Bax, Bcl2, Vimentin, Snail, ZNF740 (to detect endogenous ZNF740 protein in cells transfected with ZNF740-Flag construct) and GAPDH. Full details including antibody dilutions, catalogue numbers and RRIDs are provided in Table S3. Membranes were washed three times with Tris-buffered saline with Tween 20 (TBST), then incubated for 2 hours with species-matched HRP-conjugated secondary antibodies (anti-rabbit or anti-mouse). Following further TBST washes to remove unbound secondary antibody, protein bands were visualized using UltraSignal ECL substrate. Band densitometry was performed, and total protein levels were normalized to GAPDH as the loading control.

IHC

A total of 14 pathologically confirmed LUAD FFPE specimens from patients diagnosed between 2017 and 2019 were included in this study. The cohort comprised 7 recurrent and 7 recurrence-free cases, 1:1 matched for age, sex and TNM stage to minimize confounding. Paraffin sections (3 µm thick) were baked at 70 ℃ for 4 h, dewaxed in xylene, rehydrated through graded ethanol, blocked with goat serum for 40 minutes at room temperature, then incubated overnight at 4 ℃ with PRAME antibody (1:1,000, ab219650, Abcam). Following three washes in PBS, sections were incubated with goat anti-rabbit IgG secondary antibody for 1 h at 37 ℃. The target protein was visualized with 3,3'-diaminobenzidine, and nuclei were counterstained with hematoxylin. The identical IHC protocol was used for all animal tissue specimens. Lung sections from the nude mouse metastatic model were stained for PRAME and E-cadherin (both diluted 1:1,000, E-cadherin: 14472S, CST). Sections from the subcutaneous tumor model were stained for PRAME and Ki67 (1:10,000, 28074-1-AP, Proteintech).

IHC results were scored semi-quantitatively using the Immunoreactive Score (IRS) method, by two independent observers blinded to all sample groupings and clinical data. Areas of necrosis, hemorrhage, or edge artefact were excluded under low magnification prior to scoring. Five random high-power fields (×400) per slide were selected for evaluation of tumor cells only. IRS was calculated as the product of nuclear staining intensity score (0 = negative; 1 = weak; 2 = moderate; 3 = strong) and percentage positive cells score (0 = <5%; 1 = 5–25%; 2 = 26–50%; 3 = 51–75%; 4 = >75%), giving a possible total score between 0 and 12. The final IRS for each sample was defined as the average score across the five fields. Any discrepancies between observers were resolved by consensus joint review. The Mann-Whitney U test was used for comparison of IRS between groups.

Vector construction

Functional studies of PRAME in PC9 and A549 cells were performed using both overexpression and knockdown approaches. For overexpression, cells were transfected with pcDNA3.1-PRAME and the corresponding empty control vector (Youbao Biotechnology, Changsha, China). For knockdown, cells were transfected with siRNA targeting PRAME and non-targeting control siRNA (GenePharma, Shanghai, China). The sequences used were: siRNA-PRAME: 5'-GGUCAUGCUGACCGAUGUATT-3' and 5'-UACAUCGGUCAGCAUGACCTT-3'; siRNA-NC: 5'-UUCUCCGAACGUGUCACGUTT-3' and 5'-ACGUGACACGUUCGGAGAATT-3'.

Interaction studies between ZNF740 and PRAME used a Flag-tagged ZNF740 vector (Youbao Biotechnology) and siRNA targeting ZNF740 (GenePharma). The same non-targeting siRNA control described above was used for these experiments. The sequence of siRNA-ZNF740 was: 5'-GCAAAGAUGAUGACAGCUUTT-3' and 5'-AAGCUGUCAUCAUCUUUGCTT-3'.

For investigation of the methylation status of the PRAME promoter region, CRISPRoff v2.1 and TETV4 vectors were purchased from Miaoling Biotechnology (Wuhan, China). All dCas9 and sgRNA vectors, including pU6-sgRNA-mCherry-Puro(dCas9)-PRAME, pU6-sgRNA-Puro-ccdb-PRAME, and their respective empty control vectors, were constructed by Genomics Corporation (Beijing, China). Full details of all target and insert sequences are provided in Table S4.

Gene transfection

For the wound healing assay: PC9 and A549 cells were seeded in 24-well plates (60–80% confluence) and transfected with 0.5 µg pcDNA3.1-PRAME or siRNA-PRAME using Lipofectamine^® 3000 (pcDNA3.1 or siRNA-NC served as controls; Thermo Fisher Scientific). For qRT-PCR and WB experiments: cells were seeded in 6-well plates, transfected with 2 µg plasmid or siRNA, and harvested at 48 hours. For Cell Counting Kit-8 (CCK-8) assays: cells were seeded in 96-well plates and transfected with 0.1 µg plasmid or siRNA. For PRAME methylation editing: cells were co-transfected with 1.2 µg CRISPRoff + 0.6 µg pU6-sgRNA-mCherry-Puro(dCas9)-PRAME; for demethylation: 1.0 µg TETV4 + 0.6 µg pU6-sgRNA-Puro-ccdb-PRAME (corresponding control plasmids were used similarly). Cells subjected to methylation editing were harvested 7 days post-transfection for WB analysis.

Wound healing assay

PC9 and A549 cells (8×10⁴) were seeded in 24-well plates and cultured until 60–80% confluence. A straight wound was made with a 1,000 µL pipette tip; the supernatant was discarded and replaced with transfection medium containing overexpression plasmid or siRNA, and cells were then incubated. Images at the same well position were captured at 0, 24, 48, and 72 h under a fluorescence microscope (CKX53, Olympus, Japan), and the relative scratch distance was measured. Cell migration was calculated based on the measured migration distance. The experiment was repeated in triplicate (n=3).

CCK-8 assay

PC9 and A549 cells (6×10³) were seeded in 96-well plates and incubated for 16–18 h. For PRAME overexpression, cells were transfected with pcDNA3.1-PRAME or pcDNA3.1; for knockdown, cells were transfected with siRNA-PRAME or siRNA-NC. At 0, 24, 48, 72, and 96 h post-transfection, 10 µL CCK-8 (HY-K0301, MCE, Junction, NJ, USA) was added to each well, followed by incubation for 0.5 h. Absorbance at 450 nm was measured using a Gen5 microplate reader (BioTek, USA) to assess cell viability. Each data point represents the mean of triplicate replicates (n=5).

Cell invasion assay

Cell invasive capacity was evaluated via the transwell assay (Corning Costar, USA). The Transwell membrane was pre-coated with 80 µL Matrigel (1:8 dilution in FBS-free medium). After cell attachment, 1×10⁵ cells in 100 µL FBS-free medium were seeded into the upper chamber, with the lower chamber containing 600 µL medium with 10% FBS. After 16 h incubation, plasmid or siRNA complexed with Lipofectamine^® 3000 was added to the upper chamber, and cells were incubated at 37 ℃ for 48 h. Post-incubation, chambers were rinsed with PBS, fixed with methanol for 30 minutes, stained with 0.1% crystal violet for 15 minutes, and invaded cells were quantified by counting 4 random microscopic fields.

Colony formation assay

PC9 cells (1×10³) were seeded into a 6-well plate and incubated for 16–18 h. Cells were transfected with Flag-ZNF740 or siRNA-ZNF740 and cultured under the same conditions. Ten days post-transfection, cell colonies were rinsed 3 times with PBS, fixed with methanol (RT, 30 minutes), then stained with 0.1% crystal violet for 30 minutes. Excess stain was washed away with distilled water, colony images were captured, and colony formation efficiency was assessed by counting visible colonies (representative images collected for further evaluation).

Assay for transposase-accessible chromatin using sequencing (ATAC-seq)

PC9 cells were cultured in 1640 medium and treated with 5-azaC (10 µM) for 24 h. Post-treatment, cells were harvested, washed, and resuspended in ATAC buffer (10 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl₂). Nuclei were then incubated with Tn5 transposase (Illumina) at 37 ℃ for 30 minutes for chromatin fragmentation and adapter insertion. The reaction was terminated with 0.5 volume of 0.5 M EDTA, and DNA was purified via magnetic beads, followed by PCR for sequencing library construction. Library quality was assessed using a Bioanalyzer, and sequencing was performed on an Illumina platform (paired-end reads). Raw data were processed via standard pipelines (reference genome alignment, peak calling for open chromatin regions).

ChIP

ChIP analysis was conducted using the SimpleChIP^® Plus Enzymatic Chromatin IP Kit (9005, CST, USA). For endogenous ChIP, PC9 cells were treated with 10 µM 5-azaC for 2 days, and an anti-ZNF740 antibody (25411-1-AP, Proteintech, USA) was used. The immunoprecipitated DNA was analyzed by PCR using the following primer pair: forward 5'-CTCTTAGCCACCATGCCCAT-3' and reverse 5'-AAGTACTCCGCCTCCACAAA-3'. ChIP efficiency was calculated using the following formula:

%Input=2(CtInput−CtChIP)×100%[1]

Methylated DNA immunoprecipitation (MeDIP)

Before MeDIP, genomic DNA from PC9 and A549 cells was purified using the Genomic DNA Extraction Kit (DP304, TIANGEN, Beijing, China). Genomic DNA (20 µg) was diluted in 300 µL TE Buffer and sonicated at 4 ℃ using a non-contact ultrasonic processor (30 s sonication/30 s pause, 4 cycles) to fragment into 300–1,000 bp. Sonicated DNA (4 µg) was used for MeDIP as previously described (18). The DNA was denatured at 95 ℃ for 10 min and immediately cooled on ice for 10 min, then immunoprecipitated (4 ℃, overnight) with 3.5 µg anti-5-methylcytidine antibody (MA5-31475, Invitrogen, USA) in 500 µL IP buffer. The mixture was rotated with 40 µL Protein A/G Magnetic Beads (4 ℃, 2 h), then washed 3 times with 700 µL IP buffer. The beads were treated with proteinase K (50 ℃, 3 h), and methylated DNA was recovered via ethanol precipitation, resuspended in TE Buffer, and subjected to qPCR (forward and reverse primer sequences below). Positive control H19ICR: 5'-GAGCCGCACCAGATCTTCAG-3' and 5'-TTGGTGGAACACACTGTGATCA-3'; negative control HIST1H3B: 5'-CCCACACTTCTTATGCGACA-3' and 5'-CTGTGCCTGGTTGCAGATTA-3'; PRAME: 5'-GTGCTCGTAGACCTGTTCCT-3' and 5'-CTGGCCCAGGTAAGGAGAAA-3'.

Animal study

A protocol was prepared before the study without registration. All animal experiments were performed under a project license (No. 20250313010) granted by the Laboratory Animal Committee of West China Hospital, Sichuan University, in compliance with the Laboratory Animal Management Measures of Sichuan Province for the care and use of animals.

BALB/c nude mice (RRID: IMSR_CRL:490) were purchased from GemPharmatech and used for xenograft and metastasis models because their immunodeficiency allows for stable engraftment and growth of human-derived lung cancer cells, effectively recapitulating patterns observed in recurrent LUAD patients. All animals were housed in a controlled environment at 21–22 ℃ and 50–60% humidity with a 12-h light:12-h dark cycle and fed on standard laboratory chow.

To evaluate PRAME’s effects on LUAD metastasis and growth, two BALB/c nude mouse models were employed: a tail vein injection model for lung metastasis and an axillary tumorigenesis model for ectopic tumor growth. Each model included two groups: the experimental group (A549-PRAME-KD) and the control group (A549-NC), with 8 mice per group (16 mice per model), for a total of 32 mice. A single animal was considered the experimental unit. Sample size was determined to account for biological variability while adhering to the Reduction principle of the 3R framework in animal research. Mice were enrolled based on pre-established inclusion and exclusion criteria. Inclusion criteria: female, 4–6 weeks old, body weight 18–22 g, specific pathogen-free (SPF) grade, no prior experimental manipulation, and normal activity and mental state assessed daily. Exclusion criteria were (predefined before the experiment): (I) failure of cell injection (no detectable tumor formation 14 days post-inoculation); (II) non-procedural death (e.g., infection, accidental injury) during the experiment; (III) abnormal tumor growth (e.g., tumor ulceration, distant metastasis beyond the axillary region) that interfered with outcome assessment. Humane endpoints were defined as weight loss >20% compared to the start of the experiment, tumor diameter exceeding 15 mm, or signs of severe respiratory distress. If any of these criteria were met, animals were immediately euthanized to minimize suffering.

For the tail vein injection (lung metastasis) model, 16 female BALB/c nude mice were randomly assigned to the A549-PRAME-KD and A549-NC groups (8 mice per group). Randomization was performed using a computer-generated random number sequence in Excel. Each mouse received 2×10⁶ corresponding cells (100 µL cell suspension) via the tail vein within 10 seconds. The order of cell injections and tumor measurements was alternated between the A549-PRAME-KD and A549-NC groups. Additionally, cages were assigned to different shelf levels randomly. Two mice died due to technical issues during tail vein injection; the remaining 14 mice completed the 30-day experimental cycle. Under isoflurane anesthesia, mice were euthanized by cervical dislocation, and their lungs were immediately dissected and fixed in 4% paraformaldehyde for 48 h. H&E staining and IHC staining were performed to assess PRAME expression and its association with metastatic burden. The researchers (P.Y.) responsible for animal injections and tumor measurements were blinded to the treatment groups. Histopathological evaluations and quantitative analyses of tumor burden were performed by two independent investigators (Q.Z., X.F.H.) blinded to the group allocation.

For the subcutaneous axillary tumor model, another 16 age-, weight-, and sex-matched female BALB/c nude mice were randomly divided into two groups (8 mice per group: A549-PRAME-KD and A549-NC). Each mouse received a subcutaneous injection of 2×10⁶ cells (200 µL cell suspension) into the right axilla within 10 seconds. Tumors formed approximately 7 days post-inoculation; thereafter, body weight and tumor dimensions (length, width, and height) were measured every 2 days, and tumor volume was calculated using the formula: (length × width × height) × 0.5236. After 19 days of observation, mice were euthanized; tumors were excised for measurement of fresh weight and volume, fixed in 4% paraformaldehyde, and processed for H&E and IHC staining as described above. All stained slides were scanned using an Olympus VS200 whole-slide scanning system to capture high-resolution images.

Primary outcomes were the number of lung metastatic nodules (metastasis model) and tumor volume or weight (subcutaneous model). Secondary outcomes included the Ki67 proliferation index and expression of EMT markers.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 8.0 software and results are expressed as means ± standard deviation. Data normality was assessed using the Shapiro-Wilk test. For datasets with normal distribution, two-group comparisons were conducted using the Student’s t-test. Multiple group comparisons were analyzed using one-way analysis of variance (ANOVA) followed by Tukey’s Honestly Significant Difference post-hoc test. For non-normally distributed datasets, the Mann-Whitney U test was applied for two-group comparisons. Variance equality was assessed using the F-test for t-tests or Brown-Forsythe test for ANOVA before proceeding with parametric analyses. Statistical significance was defined as P<0.05.

Results

PRAME is upregulated in the early stage of LUAD and associated with metastatic recurrence

To investigate the role of PRAME in metastatic recurrence of early-stage LUAD, we designed an integrated study combining clinical, molecular, and functional approaches. As illustrated in Figure 1A, the overall workflow encompasses three key components: (I) RNA sequencing of primary tumor tissues from patients with or without postoperative recurrence; (II) validation of PRAME expression at the mRNA and protein levels and functional characterization through in vitro assays; and (III) in vivo evaluation using subcutaneous xenograft and metastasis mouse models. Initial analysis of 28 stage I LUAD specimens revealed that PRAME was significantly upregulated in recurrent cases compared to non-recurrent controls. Notably, higher PRAME expression was strongly associated with poorer recurrence-free survival (RFS) (Figure 1B,1C, Figure S1A). At the molecular level, WB and qRT-PCR analyses confirmed elevated PRAME expression in tumors from patients who experienced metastatic recurrence (Figure 1D, Figure S1B). IHC further demonstrated specific enrichment of PRAME protein within tumor cells of recurrent specimens (Figure S1C,S1D), supporting its tumor-intrinsic role.

Figure 1 PRAME expression in stage I LUAD cases was negatively correlated with prognosis. (A) Schematic overview of this study. (B) RNA-seq analysis showing elevated PRAME expression in recurrent and non-recurrent LUAD cases (box plots, DESeq2). (C) Kaplan-Meier curves for recurrence-free survival of PRAME expression in LUAD patients (high, n=13; low, n=15). (D) Representative western blotting images of PRAME expression in recurrent and non-recurrent early-stage primary tumor tissues (n=7 per group); GAPDH as a loading control. (E) Lung metastasis mice model by tail vein injection with A549-control cells (Con) or A549-PRAME-knockdown cells (PRAME-KD) (H&E staining; whole slide imaging). Black arrowheads indicate metastatic foci. (F) Quantification of lung tumor area percentage in control (n=4) vs. PRAME-KD groups (n=4). (G) Body weight comparison between control (n=6) vs. PRAME-KD (n=8) groups at initial 2×10⁶ cells injection and 3 weeks after. Statistical significance was determined by Mann-Whitney U test (B,F), log-rank test (C), and two-way ANOVA with Bonferroni correction (G). *, P<0.05; **, P<0.01; ***, P<0.001. ANOVA, analysis of variance; EMT, epithelial-mesenchymal transition; LUAD, lung adenocarcinoma; PRAME, preferentially expressed antigen in melanoma; TPM, transcripts per million.

Considering the limitation of the small sample, we employed the TCGA-LUAD cohort for validation. The overall survival (OS) analysis was performed on all stages (n=428) and early-stage (stage IA/IB, n=226), respectively. The results for full stages showed a significant association between PRAME gene expression level and prognosis (P=0.026), which was not observed in early-stage (P=0.81) (Figure S2A,S2B). However, we further explored disease-free survival (DFS) in early-stage (n=162), and the result suggested that higher expression of PRAME was marginally significant (P=0.067) related to higher recurrence (Figure S2C). Notably, the Gene Set Enrichment Analysis (GSEA) of KEGG results based on the DEGs between high PRAME expression group and low PRAME expression group revealed significant enrichment of the PI3K-AKT signaling pathway (normalized enrichment score =1.49, P<0.01, q<0.001), which was consistent with our research cohort (Figure S2D).

To functionally validate the pro-metastatic capacity of PRAME in vivo, we established an experimental lung metastasis model through tail vein injection of tumor cells, with metastatic burden quantified at 3 weeks post-injection (19). Two mice in the negative control (NC) group died due to procedural reasons. Subsequently, four mice were randomly selected from both the NC group and the PRAME-knockdown (PRAME-KD) group for H&E staining and tumor burden analysis. Compared with the NC group, the PRAME-KD group exhibited significantly fewer metastatic nodules and reduced tumor burden (n=4 per group) (Figure 1E,1F). The PRAME-KD group also maintained better body weight, indicating reduced systemic tumor effects (NC group=6, PRAME-KD group=8) (Figure 1G). Successful knockdown of PRAME was confirmed by WB and qRT-PCR analyses (Figure S2E,S2F). Mechanistically, histopathological examination via H&E staining and E-cadherin IHC demonstrated that, compared with control tumors, PRAME-KD tumors displayed elevated E-cadherin expression and attenuated mesenchymal morphology (n=4 per group) (Figure S2G,S2H). These multi-modal findings demonstrate that PRAME plays an essential role in promoting metastatic progression of LUAD in vivo.

PRAME promotes the migration, proliferation, and invasion of LUAD cells

To further explore the potential role of PRAME in LUAD progression, we evaluated its effects on key malignant behaviors using wound healing, transwell invasion, and CCK-8 assays. PRAME overexpression significantly enhanced the migration and invasiveness of both PC9 and A549 cells compared to the control group (Figure 2A-2C). CCK-8 assays further revealed that PRAME overexpression markedly promoted cell proliferation (Figure 2D). Consistent with the established association between invasion and EMT, PRAME-overexpressing cells displayed reduced E-cadherin expression and increased expression of the mesenchymal markers N-cadherin, Snail, and Vimentin at both mRNA and protein levels (Figure 2E-2G). Conversely, PRAME knockdown via siRNA significantly suppressed migration, invasion, and proliferation (Figure S3A-S3D) and reversed EMT marker expression, with E-cadherin upregulated and N-cadherin, Snail, and Vimentin downregulated (Figure S3E-S3G).

Figure 2 PRAME promotes the progression of LUAD via regulating EMT. (A,B) Images of cell scratch assays (magnification, 100×) and migration rates in PC9 and A549 cells after overexpression of PRAME for 48 hours (mean ± SD; n=3 biological replicates). (C) Invasive capability of A549 cells assessed by transwell assays after overexpressing PRAME (mean ± SD; n=3) (0.1% crystal violet, magnification, 200×). (D) Proliferation efficiency curves of PC9 and A549 cells after overexpressing PRAME, assessed using CCK-8 assays over a 96-hour period. (E-G) Following overexpression of PRAME, the mRNA expression levels (E,F) and protein expression levels (G) of the epithelial markers E-cadherin and the mesenchymal markers N-cadherin, Vimentin and Snail in PC9 and A549 cells were assessed using qRT-PCR and Western blotting, respectively. (H) The subcutaneous tumor at endpoint in nude mice injected with A549 (Con) vs. A549-PRAME-KD (PRAME-KD) cells (n=8 per group, 2×10⁶ per mouse). (I,J) Tumor volume curve and final tumor weight in (H) (mean ± SD; n=8 per group; unpaired t-test). Statistical significance was determined by Kruskal-Wallis H test with post-hoc Dunn’s test (B,C,E,F), unpaired two-tailed Student’s t-test (I,J), and two-way ANOVA with Bonferroni correction (D). *, P<0.05; **, P<0.01; ***, P<0.001. ANOVA, analysis of variance; CCK-8, Cell Counting Kit-8; EMT, epithelial-mesenchymal transition; LUAD, lung adenocarcinoma; NC, negative control; OD, optical density; OE, overexpression; PRAME, preferentially expressed antigen in melanoma; qRT-PCR, quantitative real-time polymerase chain reaction; SD, standard deviation.

To further validate the functional impact of PRAME on tumor proliferation and growth in vivo, we established a subcutaneous xenograft model in nude mice using either control or PRAME-knockdown (PRAME-KD) A549 cells. Tumors formed by PRAME-KD cells exhibited significantly smaller volume and weight compared to the control group, indicating that PRAME depletion impairs tumor growth (n=8 per group) (Figure 2H-2J). Consistently, IHC staining showed markedly lower Ki67 immunoreactivity scores (IRS) in PRAME-KD xenografts (Figure S3H-S3K). Raw tumor growth data are provided in Table S5.

ZNF740 and DNA methylation synergistically regulate PRAME transcription

DNA methylation is a fundamental epigenetic mechanism that regulates gene expression by adding methyl groups to cytosine residues, frequently resulting in transcriptional silencing (20,21). To determine whether PRAME expression in LUAD is epigenetically regulated, we first performed methylated DNA immunoprecipitation followed by quantitative PCR (MeDIP-qPCR) using an anti-5-methylcytosine antibody in A549 and PC9 cells. Significant methylation enrichment was observed within the PRAME promoter region (−500 and +1,500 bp relative to the transcription start site), with markedly higher methylation levels in PC9 cells compared with A549 cells (Figure 3A).

Figure 3 Zinc finger protein 740 (ZNF740) promotes PRAME expression by binding to its hypomethylated promoter and enhances tumor cell proliferation by inhibiting apoptosis signaling pathway. (A) Methylated DNA immunoprecipitation (MeDIP) tested DNA methylation at the PRAME promoter in PC9 and A549 cells with anti-5mC antibody (mean ± SD; n=3). (B) Assessment of chromatin accessibility at the PRAME promoter region in PC9 cells treated with 5-azaC using assay for transposase-accessible chromatin using sequencing (ATAC-seq). (C) Detection of PRAME protein levels in PC9 and A549 cells treated with 5-azaC using WB. (D,E) JASPAR database prediction and schematic of the conserved ZNF740-binding motif within the PRAME promoter. Sequence analysis revealed that the core binding motif of ZNF740 is a highly conserved cytosine-enriched sequence at positions 2-8, suggesting a significant DNA-binding preference for GC-rich regions. (F) Examination of ZNF740 binding on the PRAME promoter region in PC9 cells after treatment with 5-azaC using chromatin immunoprecipitation followed by quantitative polymerase chain reaction (n=3). (G) Assessment of PC9 cells colony formation ability following overexpression or knockdown of ZNF740 using a clonogenic assay (0.1% crystal violet). (H,I) Detection of PRAME and endogenous ZNF740 protein expression levels by WB analysis after treatment of PC9 cells with 5-azaC while simultaneously overexpressing (H) or knocking down (I) ZNF740. (J) WB analysis showing the protein levels of PRAME, ZNF740, and apoptosis-related proteins in PC9/A549 cells after overexpression of ZNF740. (K) Quantification of cleaved Caspase 8 (p41)/Caspase 8, Bax/Bcl2 and cleaved Caspase 3 (p19)/Caspase 3 protein expression in PC9 and A549 cells was measured using ImageJ. All experiments were repeated three times. Statistical significance was determined by Mann-Whitney U test (A,F), unpaired two-tailed Student’s t test (K). *, P<0.05; **, P<0.01; ***, P<0.001. 5-azaC, 5-azacytidine; ATAC-seq, assay for transposase-accessible chromatin using sequencing; GC, guanine-cytosine; NC, negative control; PRAME, preferentially expressed antigen in melanoma; SD, standard deviation; WB, Western blotting.

To further investigate whether DNA demethylation affects PRAME activation, PC9 cells were treated with 10 µM 5-azaC, a DNA methylation inhibitor, for 24 h. ATAC-seq analysis revealed a 2.0-kb chromatin region spanning the PRAME promoter with significantly increased accessibility, with peak signal overlapping a canonical ZNF740-binding motif (Figure 3B). Consistent with enhanced chromatin openness, 5-azaC treatment significantly upregulated PRAME protein expression in both PC9 and A549 cells (Figure 3C).

Given the increased chromatin accessibility observed upon demethylation, we next sought to identify transcription factors potentially involved in PRAME activation. Motif enrichment analysis using the JASPAR database identified ZNF740 as a high-confidence candidate regulator of the PRAME promoter. A highly conserved ZNF740 binding motif was predicted within the proximal PRAME promoter region (−500 bp upstream to +1,500 bp downstream relative to the TSS). Sequence logo analysis demonstrated that the core motif comprises a cytosine-enriched sequence at positions 2–8, indicating preferential binding to GC-rich DNA regions. The enrichment of CC dinucleotides is consistent with the canonical DNA-binding properties of C2H2 zinc finger proteins (Figure 3D,3E). To validate this prediction, ChIP-qPCR confirmed that ZNF740 binds to the PRAME promoter region, with this binding significantly enhanced upon 5-azaC treatment, indicating that DNA demethylation facilitates ZNF740 recruitment by increasing chromatin accessibility (Figure 3F). Functionally, we assessed the impact of ZNF740 on PRAME expression and LUAD cell growth. ZNF740 overexpression in PC9 cells markedly promoted colony formation, whereas ZNF740 knockdown exerted the opposite effect (Figure 3G).

To explore the potential synergy between ZNF740 and DNA demethylation in regulating PRAME, we compared PRAME expression under ZNF740 overexpression, 5-azaC treatment, or their combination. The highest PRAME expression was observed when ZNF740 overexpression and 5-azaC treatment were combined, whereas ZNF740 knockdown significantly attenuated PRAME induction even in the presence of 5-azaC (Figure 3H,3I). Furthermore, WB analysis revealed that ZNF740 overexpression suppressed apoptosis, as evidenced by reduced levels of cleaved Caspase-8, cleaved Caspase-3, and Bax, coupled with increased expression of the anti-apoptotic proteins Bcl-2 and mTOR (Figure 3J,3K).

CRISPRoff and CRISPRon systems targeted editing promoter methylation regulate PRAME expression

CRISPR-based epigenetic editing was performed using the CRISPRon (TETV4) and CRISPRoff systems (Figure S4A,S4B) to directly interrogate the role of promoter methylation in regulating PRAME expression (22,23). Specific sgRNAs targeting the PRAME promoter region, along with corresponding non-targeting controls, were designed (Figure 4A, Figure S4C-S4F; sgRNA sequences are provided in Table S4). In PC9 cells, which exhibit high baseline PRAME promoter methylation, CRISPRon-mediated demethylation significantly reduced methylation levels compared with sgRNA-NC controls, as confirmed by bisulfite sequencing (Figure 4B). Functionally, this targeted demethylation markedly enhanced both migration and invasion of PC9 cells (Figure 4C-4E). Conversely, in hypomethylated A549 cells, CRISPRoff V2.1-mediated methylation of the PRAME promoter significantly suppressed cell migration (Figure 4F,4G).

Figure 4 Epigenetic editing of PRAME promoter by CRISPRon/off regulates the methylation status and affects cell invasion, migration and proliferation in LUAD cells. (A) A schematic of sgRNA-plasmid for the CRISPRon system. A trans-activator ribonucleoprotein complex mediated by a sgRNA targeting PRAME transcription factor binding region was fused with MS2 loops, MCP, Rta-AD and NLS, which removes PRAME promoter methylation and recruits transcriptional machinery. (B) Bisulfite PCR analysis was performed to assess the DNA methylation status of the PRAME promoter region in PC9 cells transfected with TETV4 + sgRNA-PRAME (sg-PRAME) vs. TETV4 + sgRNA-negative control (sg-NC). TETV4 is a CRISPRon system designed for demethylation editing, sg-NC serves as the control vector, and sg-PRAME is the guide-RNA vector targeting the PRAME promoter region. White circles, unmethylated CpG; black circles, methylated CpG. Each row represents one sequencing read. (C) The invasive capacity of PC9 cells following PRAME demethylation treatment (TETV4 + sg-PRAME) via transwell assay (0.1% crystal violet, magnification, 100×). (D-G) Wound healing assay (magnification, 100×) to assess the cell migration ability of PC9 cells following demethylation editing (D) and A549 cells following methylation editing (F), along with quantitative analysis of the migration rate (E,G) (mean ± SD). (H,I) After demethylation editing of PC9 cells using the CRISPRon system (H) and methylation editing of A549 cells using the CRISPRoff system (I), WB was used to assess the protein expression levels of mTOR, Vimentin, and Rap1, which are associated with the survival and migration capabilities of tumor cells. Statistical significance was determined by two-way ANOVA with Tukey’s post-hoc test (E,G). *, P<0.05; **, P<0.01; ***, P<0.001. ANOVA, analysis of variance; CpG, Cytosine-phosphate-Guanine; LUAD, lung adenocarcinoma; MCP, MS2 capsid protein; MS2 loops, MS2 RNA hairpin loops; NC, negative control; NLS, nuclear localization signal; PCR, polymerase chain reaction; PRAME, preferentially expressed antigen in melanoma; Rta-AD, rta activation domain; SD, standard deviation; sgRNA, single-guide RNA; WB, Western blotting.

WB analysis showed that CRISPRon-induced demethylation in PC9 cells upregulated mTOR, Vimentin, and Rap1, while downregulating the epithelial marker E-cadherin (Figure 4H). In contrast, CRISPRoff-induced methylation in A549 cells produced the opposite trends in protein expression (Figure 4I). Notably, Rap1 has been implicated in tumor progression through the activation of the MAPK/ERK and Src/FAK signaling pathways (24,25), further emphasizing the relevance of PRAME in regulating pathways that facilitate tumor progression.

PRAME promotes metastatic recurrence through PI3K/AKT/mTOR signaling pathway

KEGG enrichment analysis was performed on DEGs between recurrent and non-recurrent early-stage primary tumor tissues and revealed significant activation of multiple oncogenic pathways, including the PI3K-AKT signaling pathway (Figure 5A), which is known to play a central role in tumor progression (26).

Figure 5 PRAME influences EMT markers through PI3K/AKT/mTOR signaling pathway. (A) Annular bar plot illustrating significantly enriched Kyoto Encyclopedia of Genes and Genomes pathways in recurrent LUAD patients compared with non-recurrent LUAD patients. Bar height corresponds to −log₁₀(P value), and color gradient represents the z-score. (B) The expression levels of PI3K/AKT/mTOR pathway-related proteins were detected in PC9 and A549 cells after PRAME overexpression via WB. (C,D) WB analysis of EMT-related expressions (E-cadherin, N-cadherin, Vimentin, and Snail) in PC9 and A549 cells upon PRAME overexpression combined with PI3K inhibitor LY294002 (C) or AKT serine/threonine kinase (AKT) inhibitor MK2206 (D). (E) The p-PI3K (p85) and p-mTOR expression in PC9 and A549 cells after treatment with the PI3K inhibitor LY294002. (F) The p-AKT and p-mTOR expression in PC9 and A549 cells after treatment with the AKT inhibitor MK2206. All experiments were performed in 3 biological replicates (n=3). EMT, epithelial-mesenchymal transition; LUAD, lung adenocarcinoma; PRAME, preferentially expressed antigen in melanoma; WB, Western blotting.

WB analysis showed that PRAME overexpression substantially increased the phosphorylation levels of PI3K (p-PI3K), AKT (p-AKT), and mTOR (p-mTOR) in both PC9 and A549 cells (Figure 5B). To determine whether PRAME promotes EMT through activation of the PI3K/AKT/mTOR pathway, PC9 and A549 cells were transfected with pcDNA3.1-PRAME followed by treatment with the PI3K inhibitor LY294002 (50 µM) or the AKT inhibitor MK2206 (20 µM). PRAME overexpression significantly upregulated the expression of the EMT-associated proteins, including N-cadherin, Vimentin, and Snail; treatment with either inhibitor markedly reversed these PRAME-induced changes (Figure 5C,5D). Concurrently, both inhibitors significantly attenuated PRAME-induced phosphorylation of PI3K, AKT, and mTOR (Figure 5E,5F).

Discussion

PRAME, initially identified as a cancer-testis antigen capable of eliciting cytotoxic T-lymphocyte responses in melanoma (27,28), has increasingly been recognized as a clinically relevant biomarker across multiple solid and hematologic malignancies (17,29-34). Beyond its immunogenic properties, accumulating evidence suggests that PRAME contributes to tumor progression, metastatic potential, and adverse prognosis. In the present study, we extend these observations to LUAD, particularly in the context of early recurrence.

In this study, we demonstrated that PRAME is significantly upregulated in recurrent primary LUAD tissues, and that high PRAME expression is an independent negative predictor of both RFS and OS. Notably, PRAME expression was associated with promoter hypomethylation, suggesting that epigenetic deregulation underlies its aberrant activation in early recurrent disease. These findings are consistent with prior reports in acute myeloid leukemia and ovarian cancer, in which promoter hypomethylation directly correlated with PRAME overexpression (35,36), indicating that epigenetic control of PRAME may represent a conserved mechanism across tumor types.

Functional assays and in vivo models confirmed that PRAME promotes LUAD cell proliferation, migration, and invasion, and tumor growth. PRAME overexpression induced EMT, as evidenced by downregulation of E-cadherin and upregulation of N-cadherin, Snail, and Vimentin (37). These findings align with previous findings in triple-negative breast cancer, where PRAME was shown to promote EMT (38), and with pan-cancer transcriptomic analysis linking high PRAME expression to EMT signatures and metastatic phenotypes, particularly in lung cancer and melanoma (39). In contrast, prior reports suggested a metastasis-suppressive role of PRAME in certain lung cancer contexts (40), underscoring the context-dependent nature of PRAME function and highlighting the influence of tumor-specific genetic and epigenetic backgrounds.

Mechanistically, our data demonstrate that PRAME exerts its oncogenic effects, at least in part, through activation of the PI3K/AKT/mTOR signaling pathway. PRAME overexpression significantly increased phosphorylation of PI3K, AKT, and mTOR, accompanied by increased expression of EMT-associated proteins. Pharmacological inhibition of PI3K and AKT using LY294002 and MK2206 attenuated pathway activation and partially reversed PRAME-induced EMT changes (Figure 6), supporting a functional link between PRAME and this canonical oncogenic axis. Given the established role of PI3K/AKT/mTOR signaling in tumor cell survival, metabolic adaptation, and metastasis, these findings provide a mechanistic explanation for the aggressive phenotype observed in PRAME-high LUAD.

Figure 6 Schematic diagram summarizing the function of PRAME and PRAME-induced PI3K/AKT/mTOR signaling pathway. PRAME, preferentially expressed antigen in melanoma.

In addition to downstream signaling, we investigated the upstream regulatory mechanisms governing PRAME expression. Differential methylation patterns between PC9 and A549 cells suggest that the epigenetic status contributes to PRAME heterogeneity in LUAD. Chromatin accessibility profiling identified a promoter region with enhanced accessibility overlapping a predicted binding motif for ZNF740, which was previously implicated in hepatocellular carcinoma progression (41) and was confirmed to directly bind the PRAME promoter and promote its transcriptional activation. Importantly, promoter hypomethylation facilitated ZNF740 recruitment, indicating that epigenetic deregulation enhances transcription factor accessibility. This ZNF740–PRAME interaction is consistent with the concept of epigenetic-mediated positive regulatory circuits in LUAD progression (42). In this framework, promoter hypomethylation increases chromatin accessibility, enabling transcription factor binding and sustained oncogene activation. Treatment with the demethylating agent 5-azaC further increased chromatin accessibility and upregulated PRAME expression, corroborating previous reports that PRAME is frequently activated through promoter demethylation across multiple malignancies (40,43-45).

To directly validate the causal role of promoter methylation in regulating PRAME, we employed CRISPRoff and CRISPRon systems for locusspecific epigenetic editing, which avoids inducing DNA double-strand breaks, minimizing genotoxicity (46,47). CRISPRon-mediated demethylation in hypermethylated PC9 cells reactivated PRAME expression and enhanced cell migration and invasion, along with upregulation of mTOR, Vimentin, and Rap1. In contrast, CRISPRoff-mediated methylation in hypomethylated A549 cells suppressed PRAME expression and reversed malignant phenotypes (48). These findings provide functional evidence that promoter methylation dynamically regulates PRAME expression and tumor behavior in LUAD.

Collectively, our results delineate an integrated mechanistic axis in LUAD: promoter hypomethylation enhances chromatin accessibility, facilitating ZNF740 binding and transcriptional activation of PRAME; activated PRAME subsequently promotes EMT and metastatic progression via PI3K/AKT/mTOR signaling. This epigenetically regulated ZNF740–PRAME–PI3K/AKT/mTOR cascade provides a coherent framework linking molecular deregulation to clinical recurrence.

Despite these compelling findings, several limitations of this work should be acknowledged. First, the limited sample size in the IHC cohort, with only 7 patients per group. Due to the low number of included cases, the conclusions drawn from this comparison should be considered preliminary and require validation in larger, independent datasets. Second, although the tail-vein metastasis model is technically simple and highly reproducible and effectively mimics hematogenous dissemination leading to lung metastasis-thereby reflecting the clinical pattern of distant recurrence, it does not fully recapitulate the complete sequence of metastatic progression from primary tumor development to distant spread. Moreover, this model differs from clinical recurrence scenarios following curative resection and does not account for the contributions of the primary tumor microenvironment. Third, the in vivo experiments were not independently replicated due to experimental time, animal housing costs, ethical constraints and adherence to the 3R Reduction principle, which may limit the robustness of these findings. To address these limitations, future work will use orthotopic LUAD transplantation models to confirm these findings and explore interactions between the primary tumour microenvironment and PRAME methylation. Additionally, other potential mechanisms, such as chromosomal copy-number alterations (e.g., 8q gain) (49), may also contribute to PRAME overexpression and should be explored. The upstream drivers responsible for PRAME promoter hypomethylation in recurrent LUAD remain to be elucidated.

Conclusions

In summary, our study delineates a regulatory axis in which DNA hypomethylation enhances ZNF740 binding to the PRAME promoter, leading to PRAME upregulation and activation of the PI3K/AKT/mTOR pathway. This molecular cascade contributes to LUAD progression and metastatic recurrence. These findings support the potential of PRAME as a therapeutic target in recurrent LUAD, and the mechanistic insights presented here may inform the development of targeted strategies aimed at attenuating PRAME-driven oncogenic signaling. In addition, the application of CRISPR-based epigenetic editing offers a proof-of-concept framework for precision modulation of PRAME expression and its downstream tumor-promoting effects.

Acknowledgments

The authors are grateful to Sisi Wu, Yi Zhang, Yue Li, Mingjie Xu from Core Facilities, and Jinghong Xian from State Key Laboratory of Respiratory Health and Multimorbidity at West China Hospital of Sichuan University, for their support of the experiment.

Footnote

Reporting Checklist: The authors have completed the MDAR and ARRIVE reporting checklists. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1417/rc

Data Sharing Statement: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1417/dss

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1417/prf

Funding: This work was supported by Noncommunicable Chronic Diseases-National Science and Technology Major Project (Nos. 2024ZD0529502 and 2024ZD0529500, to Z.F.W.); National Natural Science Foundation of China (No. 32370628 to Z.F.W.); the Science and Technology Project of Sichuan (No. 2023NSFSC0041, to Z.F.W. and No. 2024NSFSC0402, to W.M.L.); State Key Laboratory Special Fund (No. 2060204, to W.M.L.); Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences (No. 2023-12M-2-001, to W.M.L.); Key R&D Support Plan of Chengdu Science and Technology Bureau (No. 2023-YF09-00039-SN, to W.M.L.); and 1.3.5 Project of State Key Laboratory of Respiratory Health and Multimorbidity, West China Hospital, Sichuan University (No. RHM24101, to W.M.L. and No. RHM25210, to Z.F.W.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1417/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of West China Hospital of Sichuan University (No. 20251133), and individual consent for this retrospective analysis was waived. All animal experiments were performed under a project license (No. 20250313010) granted by the Laboratory Animal Committee of West China Hospital, Sichuan University, in compliance with the Laboratory Animal Management Measures of Sichuan Province for the care and use of animals.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Zhou C, Jing Z, Liu W, et al. Prognosis of recurrence after complete resection in early-stage lung adenocarcinoma based on molecular alterations: a systematic review and meta-analysis. Sci Rep 2023;13:18710. [Crossref] [PubMed]
3. Goldstraw P, Chansky K, Crowley J, et al. The IASLC Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groupings in the Forthcoming (Eighth) Edition of the TNM Classification for Lung Cancer. J Thorac Oncol 2016;11:39-51. [Crossref] [PubMed]
4. Guerrera F, Lococo F, Evangelista A, et al. Risk of recurrence in stage I adenocarcinoma of the lung: a multi-institutional study on synergism between type of surgery and type of nodal staging. J Thorac Dis 2019;11:564-72. [Crossref] [PubMed]
5. Ekeke CN, Mitchell C, Schuchert M, et al. Early Distant Recurrence in Patients With Resected Stage I Lung Cancer: A Case Series of "Blast Metastasis". Clin Lung Cancer 2021;22:e132-5. [Crossref] [PubMed]
6. Mahvi DA, Liu R, Grinstaff MW, et al. Local Cancer Recurrence: The Realities, Challenges, and Opportunities for New Therapies. CA Cancer J Clin 2018;68:488-505. [Crossref] [PubMed]
7. Wang C, Li J, Chen J, et al. Multi-omics analyses reveal biological and clinical insights in recurrent stage I non-small cell lung cancer. Nat Commun 2025;16:1477. [Crossref] [PubMed]
8. Su PL, Chen JY, Chu CY, et al. The impact of driver mutation on the treatment outcome of early-stage lung cancer patients receiving neoadjuvant immunotherapy and chemotherapy. Sci Rep 2022;12:3319. [Crossref] [PubMed]
9. Offin M, Rizvi H, Tenet M, et al. Tumor Mutation Burden and Efficacy of EGFR-Tyrosine Kinase Inhibitors in Patients with EGFR-Mutant Lung Cancers. Clin Cancer Res 2019;25:1063-9. [Crossref] [PubMed]
10. Yu X, Zhao H, Wang R, et al. Cancer epigenetics: from laboratory studies and clinical trials to precision medicine. Cell Death Discov 2024;10:28. [Crossref] [PubMed]
11. Chemla Y, Itzhaki O, Melamed S, et al. HLA export by melanoma cells decoys cytotoxic T cells to promote immune evasion. Cell 2026;189:233-251.e29. [Crossref] [PubMed]
12. Kessler JH, Beekman NJ, Bres-Vloemans SA, et al. Efficient identification of novel HLA-A(*)0201-presented cytotoxic T lymphocyte epitopes in the widely expressed tumor antigen PRAME by proteasome-mediated digestion analysis. J Exp Med 2001;193:73-88. [Crossref] [PubMed]
13. Xu Y, Zou R, Wang J, et al. The role of the cancer testis antigen PRAME in tumorigenesis and immunotherapy in human cancer. Cell Prolif 2020;53:e12770. [Crossref] [PubMed]
14. Pujol JL, De Pas T, Rittmeyer A, et al. Safety and Immunogenicity of the PRAME Cancer Immunotherapeutic in Patients with Resected Non-Small Cell Lung Cancer: A Phase I Dose Escalation Study. J Thorac Oncol 2016;11:2208-17. [Crossref] [PubMed]
15. Zhang Y, Liu T, Meyer CA, et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol 2008;9:R137. [Crossref] [PubMed]
16. Lee YK, Park UH, Kim EJ, et al. Tumor antigen PRAME is up-regulated by MZF1 in cooperation with DNA hypomethylation in melanoma cells. Cancer Lett 2017;403:144-51. [Crossref] [PubMed]
17. Field MG, Durante MA, Decatur CL, et al. Epigenetic reprogramming and aberrant expression of PRAME are associated with increased metastatic risk in Class 1 and Class 2 uveal melanomas. Oncotarget 2016;7:59209-19. [Crossref] [PubMed]
18. Borgel J, Guibert S, Weber M. Methylated DNA immunoprecipitation (MeDIP) from low amounts of cells. Methods Mol Biol 2012;925:149-58. [Crossref] [PubMed]
19. Oh BY, Hong HK, Lee WY, et al. Animal models of colorectal cancer with liver metastasis. Cancer Lett 2017;387:114-20. [Crossref] [PubMed]
20. Liu XS, Wu H, Ji X, et al. Editing DNA Methylation in the Mammalian Genome. Cell 2016;167:233-247.e17. [Crossref] [PubMed]
21. Lei Y, Zhang X, Su J, et al. Targeted DNA methylation in vivo using an engineered dCas9-MQ1 fusion protein. Nat Commun 2017;8:16026. [Crossref] [PubMed]
22. Nuñez JK, Chen J, Pommier GC, et al. Genome-wide programmable transcriptional memory by CRISPR-based epigenome editing. Cell 2021;184:2503-2519.e17. [Crossref] [PubMed]
23. Konermann S, Brigham MD, Trevino AE, et al. Genome-scale transcriptional activation by an engineered CRISPR-Cas9 complex. Nature 2015;517:583-8. [Crossref] [PubMed]
24. Li H, Liang J, Wang J, et al. Mex3a promotes oncogenesis through the RAP1/MAPK signaling pathway in colorectal cancer and is inhibited by hsa-miR-6887-3p. Cancer Commun (Lond) 2021;41:472-91. [Crossref] [PubMed]
25. Shah S, Brock EJ, Jackson RM, et al. Downregulation of Rap1Gap: A Switch from DCIS to Invasive Breast Carcinoma via ERK/MAPK Activation. Neoplasia 2018;20:951-63. [Crossref] [PubMed]
26. Yu L, Wei J, Liu P. Attacking the PI3K/Akt/mTOR signaling pathway for targeted therapeutic treatment in human cancer. Semin Cancer Biol 2022;85:69-94. [Crossref] [PubMed]
27. Wadelin F, Fulton J, McEwan PA, et al. Leucine-rich repeat protein PRAME: expression, potential functions and clinical implications for leukaemia. Mol Cancer 2010;9:226. [Crossref] [PubMed]
28. Kurtenbach S, Sanchez MI, Kuznetsoff J, et al. PRAME induces genomic instability in uveal melanoma. Oncogene 2024;43:555-65. [Crossref] [PubMed]
29. Takata K, Chong LC, Ennishi D, et al. Tumor-associated antigen PRAME exhibits dualistic functions that are targetable in diffuse large B cell lymphoma. J Clin Invest 2022;132:e145343. [Crossref] [PubMed]
30. Nettersheim D, Arndt I, Sharma R, et al. The cancer/testis-antigen PRAME supports the pluripotency network and represses somatic and germ cell differentiation programs in seminomas. Br J Cancer 2016;115:454-64. [Crossref] [PubMed]
31. Hemminger JA, Toland AE, Scharschmidt TJ, et al. Expression of cancer-testis antigens MAGEA1, MAGEA3, ACRBP, PRAME, SSX2, and CTAG2 in myxoid and round cell liposarcoma. Mod Pathol 2014;27:1238-45. [Crossref] [PubMed]
32. Traynor S, Jakobsen MK, Green TM, et al. Single-cell sequencing unveils extensive intratumoral heterogeneity of cancer/testis antigen expression in melanoma and lung cancer. J Immunother Cancer 2024;12:e008759. [Crossref] [PubMed]
33. Ercolak V, Paydas S, Bagir E, et al. PRAME Expression and Its Clinical Relevance in Hodgkin's Lymphoma. Acta Haematol 2015;134:199-207. [Crossref] [PubMed]
34. Partheen K, Levan K, Osterberg L, et al. Four potential biomarkers as prognostic factors in stage III serous ovarian adenocarcinomas. Int J Cancer 2008;123:2130-7. [Crossref] [PubMed]
35. Zhang W, Barger CJ, Eng KH, et al. PRAME expression and promoter hypomethylation in epithelial ovarian cancer. Oncotarget 2016;7:45352-69. [Crossref] [PubMed]
36. Ortmann CA, Eisele L, Nückel H, et al. Aberrant hypomethylation of the cancer-testis antigen PRAME correlates with PRAME expression in acute myeloid leukemia. Ann Hematol 2008;87:809-18. [Crossref] [PubMed]
37. Yun H, Han GH, Wee DJ, et al. Loss of E-cadherin Activates EGFR-MEK/ERK Signaling, Promoting Cervical Cancer Progression. Cancer Genomics Proteomics 2025;22:271-84. [Crossref] [PubMed]
38. Al-Khadairi G, Naik A, Thomas R, et al. PRAME promotes epithelial-to-mesenchymal transition in triple negative breast cancer. J Transl Med 2019;17:9. [Crossref] [PubMed]
39. Kaczorowski M, Lasota J, Dudek K, et al. Expression of Immunotherapy Target PRAME in Cancer Correlates with Histone H3 Acetylation and Is Unrelated to Expression of Methylating (DMNT3A/3B) and Demethylating (TET1) Enzymes. J Clin Med 2024;13:1554. [Crossref] [PubMed]
40. Huang Q, Wei H, Wu Z, et al. Preferentially Expressed Antigen of Melanoma Prevents Lung Cancer Metastasis. PLoS One 2016;11:e0149640. [Crossref] [PubMed]
41. Zhang H, Han B, Tian S, et al. ZNF740 facilitates the malignant progression of hepatocellular carcinoma via the METTL3/HIF‑1A signaling axis. Int J Oncol 2024;65:105. [Crossref] [PubMed]
42. Cheng Y, Luo Y, Hu T, et al. Epigenetic-mediated positive feedback loop facilitates the progression of lung adenocarcinoma. iScience 2025;28:113339. [Crossref] [PubMed]
43. Yang J, Chen M, Ye J, et al. Targeting PRAME for acute myeloid leukemia therapy. Front Immunol 2024;15:1378277. [Crossref] [PubMed]
44. Hosseininia Z, Dehghani H. PRAME-AS lncRNA, regulated by MZF1, modulates PRAME expression and cell stemness. PLoS One 2025;20:e0331190. [Crossref] [PubMed]
45. Fang R, Vallius T, Zhang A, et al. PRAME Expression in Melanoma is Negatively Regulated by TET2-Mediated DNA Hydroxymethylation. Lab Invest 2025;105:104123. [Crossref] [PubMed]
46. Jost M, Santos DA, Saunders RA, et al. Titrating gene expression using libraries of systematically attenuated CRISPR guide RNAs. Nat Biotechnol 2020;38:355-64. [Crossref] [PubMed]
47. Knott GJ, Doudna JA. CRISPR-Cas guides the future of genetic engineering. Science 2018;361:866-9. [Crossref] [PubMed]
48. Rosenblum D, Gutkin A, Kedmi R, et al. CRISPR-Cas9 genome editing using targeted lipid nanoparticles for cancer therapy. Sci Adv 2020;6:eabc9450. [Crossref] [PubMed]
49. Onken MD, Worley LA, Harbour JW. A metastasis modifier locus on human chromosome 8p in uveal melanoma identified by integrative genomic analysis. Clin Cancer Res 2008;14:3737-45. [Crossref] [PubMed]