The MAPK/ATF3/ASNS axis drives amino acid metabolic reprogramming to promote NSCLC survival under glucose deprivation

Abstract

Background

Glucose deprivation is a prevalent stressor within the tumor microenvironment. Nonetheless, the fundamental mechanism through which non-small cell lung cancer (NSCLC) cells orchestrate survival and progression through specific metabolic hubs under such conditions remains poorly understood.

Methods

This study utilized integrated multiomics analyses, encompassing transcriptomics and metabolomics, to identify pivotal targets. Functional validation was performed via in vitro assays, including CCK-8, colony formation, Transwell, and EdU assays, as well as in vivo models, such as subcutaneous xenografts and tail-vein lung metastasis models. The function and mechanism of the MAPK/ATF3/ASNS signaling axis were comprehensively investigated using gene knockdown/overexpression techniques, Western blotting, immunohistochemistry, and metabolomic analysis.

Results

Combined transcriptomic and metabolomic analyses revealed that glucose deprivation markedly upregulates ASNS expression in NSCLC cells, correlating with unfavorable patient outcomes. Functionally, both in vitro and in vivo experiments confirmed that ASNS significantly promotes the malignant behaviors of NSCLC cells under glucose-deprived conditions. Metabolomic analysis revealed that ASNS supports tumor cell survival during energetic stress by maintaining a dynamic balance of multiple amino acids. Mechanistically, glucose deprivation activates the MAPK signaling pathway, leading to the upregulation of the transcription factor ATF3, which binds directly to the ASNS promoter and transcriptionally activates its expression, thereby promoting malignant progression in NSCLC. Finally, the combined targeting of glucose metabolism using 2-DG, along with ASNS inhibition, demonstrated additive antitumor efficacy in vivo.

Conclusion

The results of this study revealed that, under glucose deprivation stress, the MAPK/ATF3/ASNS axis functions as a critical signaling–metabolic hub, promoting NSCLC progression by driving amino acid metabolic reprogramming. Targeting this axis offers a novel strategy for intervening in tumor metabolic adaptation and developing innovative combination therapies.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12967-026-08551-x.

Introduction

Non-small cell lung cancer (NSCLC) is among the malignancies with the highest global incidence and mortality rates [1]. The prognosis for lung cancer patients remains poor, resulting in a five-year survival rate of less than 20% [2]. Despite considerable advancements in targeted therapy and immunotherapy, tumor drug resistance and recurrence continue to present major clinical challenges [3]. Consequently, the need to identify potential therapeutic targets for lung cancer is urgent.

Owing to an aberrant vascular system, solid tumors often develop regional microenvironments characterized by nutrient deprivation, with glucose depletion being a prevalent form of energetic stress [4, 5]. To survive in this energy-deficient state, tumor cells frequently undergo intrinsic metabolic reprogramming. This reprogramming not only alters pathways for energy acquisition but also critically affects the amino acid metabolic network [6].

Amino acids are essential not only as building blocks for macromolecule synthesis but also for maintaining redox homeostasis, providing energetic substrates, and transducing stress signals [7, 8]. Recent extensive research has elucidated the complex interplay between amino acid metabolic reprogramming and the tumor microenvironment, revealing its multifaceted impact on tumor immune evasion, metastasis, and therapy resistance to therapy [9–13].Under the specific stress condition of glucose deprivation, the mechanisms through which tumor cells coordinate extensive amino acid metabolic reprogramming through core molecular hubs to facilitate their adaptive survival remain insufficiently elucidated.

Asparagine synthetase (ASNS) is the pivotal enzyme responsible for catalyzing asparagine synthesis. In tumors, ASNS is frequently overexpressed, promoting cancer cell proliferation, survival, and resistance to apoptosis by maintaining intracellular asparagine levels [14–16]. Its upregulated expression is also closely associated with resistance to chemotherapy, highlighting its potential as a therapeutic target [17, 18].

This study aims to systematically decipher the metabolic adaptation mechanisms of NSCLC cells in response to glucose deprivation. We initially identified ASNS as a potential key metabolic enzyme that is specifically upregulated and functionally essential under glucose deprivation. Through a series of experiments, we demonstrate the crucial role of ASNS in enabling NSCLC cells to cope with glucose deprivation and elucidate how the downstream metabolic network reprogramming mediated by ASNS collectively promotes malignant progression. Furthermore, we identify the upstream mitogen-activated protein kinase (MAPK)/activating transcription factor 3 (ATF3) signaling axis that regulates ASNS in the context of glucose deprivation.

Materials and methods

Tissue microarray and immunohistochemical staining

A tissue microarray was constructed using paired tumor and adjacent normal tissues from 80 pathologically confirmed NSCLC patients (AIfang Biological, Changsha, China). Immunohistochemical (IHC) staining was conducted utilizing antibodies against ASNS (92479 S, Cell Signaling Technology [CST], USA) and Ki-67 (9449T, CST, USA). All stained sections were digitally scanned, and the images were analyzed using the Aipathwell^® analysis software (Servicebio, Wuhan, China). Protein expression levels were quantified based on the Histochemistry score (H-Score), calculated using the following formula:

In this formula, ‘i’ represents the staining intensity score: 0 (negative), 1 (weak, light yellow), 2 (moderate, brownish yellow), and 3 (strong, brown). The variable ‘pi’ denotes the percentage of positive cells at each intensity level. The H-Score ranges from 0 to 300, with higher values indicating stronger overall protein expression, incorporating both the intensity and the proportion of positive cells. For patient stratification, the optimal cut-off value for ASNS expression was determined to be an H-Score of 183.03, which was utilized to divide patients into two subgroups.

Cell lines and cell culture

The NSCLC cell lines A549, H1299, H838, PC9, H1975, and 95D, as well as the human bronchial epithelial cell line (BEAS-2B), were all obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). The cells were maintained in either RPMI 1640 or Dulbecco’s Modified Eagle Medium (DMEM) (Procell, Wuhan, China), supplemented with 10% fetal bovine serum (FBS) (Gibco, Carlsbad, USA), 100 µg/mL streptomycin, and 100 U/mL penicillin G. All cultures were incubated at 37 °C in a humidified atmosphere containing 5% CO₂. For glucose deprivation studies, the cells were cultured in glucose-free RPMI 1640 or DMEM medium (Meilunbio, Dalian, China), similarly supplemented with 10% FBS and antibiotics.

Lentiviral vector construction and stable cell line establishment

To achieve stable knockdown of ASNS, three distinct short hairpin RNAs (shRNAs) specifically targeting ASNS were cloned into lentiviral vectors, employing a non-specific shRNA as a negative control. For the overexpression of ASNS, three lentiviral vectors encoding the full-length human ASNS sequence were synthesized, with an empty vector serving as the control. All lentiviral constructs were generated by GeneChem (Shanghai, China). The sequences for the shRNAs are provided in Supplementary Table 1. A549, H1299, and H838 cell lines were infected with the respective lentiviruses in the presence of 2 µl Polybrene (5 µg/ml, GeneChem) to enhance transduction efficiency. Seventy-two hours post-infection, the cells underwent selection using 5 µg/ml puromycin for two weeks to establish stably transfected polyclonal populations.

siRNA and plasmid transfection

For the knockdown of ATF3, three siRNAs specifically targeting ATF3 mRNA were designed and synthesized by GenePharma (Suzhou, China), along with a non-targeting control siRNA. To achieve ATF3 overexpression, an ATF3 overexpression plasmid containing the complete coding sequence was obtained from Genomeditech (Shanghai, China), with an empty vector serving as the control. The sequences for the siRNAs are detailed in Supplementary Table 2. Both siRNAs and plasmids were transfected into cells using liposome-based transfection reagents.

Quantitative reverse transcription–polymerase chain reaction (qRT–PCR) analysis

Total RNA was extracted from cells using the RNA-Quick Purification Kit (RN001, ESScience, China). Reverse transcription of RNA into complementary DNA was conducted with the PrimeScript RT Reagent Kit (RR036A, Takara, Japan). Quantitative RT–PCR was performed using SYBR Premix Ex Taq (RR420A Takara, Japan) on an Applied Biosystems Step One Plus Real-Time PCR System. The relative expression levels of target genes were quantified using the 2 − ΔΔCT method. Specific primer sequence information is provided in Supplementary Table 3. All experiments were conducted in triplicate.

Western blot analysis

Cells were lysed on ice using RIPA lysis buffer (Beyotime, China) for 30 min. The protein concentration of each lysate was quantified using a BCA assay kit (Thermo Scientific, USA). Subsequently, equal amounts of protein (30 µg per lane) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, USA).The membranes were blocked with QuickBlock™ Rapid Blocking Buffer (PS108P, Epizyme, China) for 15 min at room temperature and then incubated overnight at 4 °C with specific primary antibodies. Following the primary antibody incubation, the membranes were probed with corresponding horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence (ECL) substrate (Beyotime, China), and the images were captured using a chemiluminescence imaging system. Detailed information regarding all antibodies used is provided in the Supporting Table S4. Each experiment was independently repeated at least three times to ensure reproducibility.

Cell counting kit-8 (CCK-8) and colony formation assays

CCK-8 and Colony Formation Assays were conducted to evaluate cell proliferation and viability. For the CCK-8 assay, cells were seeded in 96-well plates at a density of 2–3 × 10³ cells per well and cultured for 5 days. Daily assessments of cell viability were performed by measuring the optical density at 450 nm using the CCK-8 reagent (Beyotime, China), in accordance with the manufacturer’s instructions. In the colony formation assay, 300–500 cells were plated into each well of 6-well plates and cultured for a period of 7–14 days. The resulting cell colonies were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet, photographed, and subsequently counted manually. All experiments were conducted in triplicate and repeated independently at least three times.

EdU assay

Cell proliferation was assessed using the EdU (5-ethynyl-2′-deoxyuridine) assay. Briefly, cells were seeded into confocal dishes and allowed to adhere and resume normal growth under standard culture conditions. The culture medium was subsequently replaced with fresh medium containing EdU, and the cells were incubated for an additional 2 to 4 h. Following the EdU labeling period, the cells were fixed with 4% paraformaldehyde and permeabilized with 0.3% Triton X-100 at room temperature for 10 to 15 min. The incorporated EdU was then detected by incubating the cells with a Click-iT reaction cocktail for 30 min in the dark. Cell nuclei were counterstained with Hoechst 33,342. Images were acquired using a laser scanning confocal microscope, and the percentage of EdU-positive cells was quantified. Each experiment was conducted in triplicate.

Wound healing assay

Cells were cultured in 12-well plates until they reached over 90% confluency. A uniform wound was created in the cell monolayer by scratching with a sterile 200 µL pipette tip, followed by washing with phosphate-buffered saline (PBS) to remove detached cells. The cells were then incubated in medium containing 1% fetal bovine serum (FBS) for 24 h. Wound widths were measured at 0 h and 24 h using microscopy, and the wound closure rate was calculated. Each experiment was conducted in triplicate.

Transwell migration and invasion assay

The abilities of cell migration and invasion were assessed using Transwell chambers (Corning, Cambridge, MA, USA). In the migration assay, 2–5 × 10⁴ cells suspended in serum-free medium were seeded into the upper chamber. The lower chamber was filled with 600 µL of medium containing 10% FBS to serve as a chemoattractant. After an incubation period of 12 to 24 h, non-migrated cells on the upper surface were carefully removed using a cotton swab. Cells that had migrated to the lower surface were subsequently fixed with 4% paraformaldehyde, stained with 0.1% crystal violet, and counted under a microscope. For the invasion assay, the Transwell membranes were pre-coated with Matrigel (BD Biosciences, Franklin Lakes, NJ, USA) 12 h prior to cell seeding. The subsequent steps were identical to those described for the migration assay. All experiments were performed in triplicate.

Luciferase reporter assay

The luciferase reporter assay was performed following the manufacturer’s instructions (Genomeditech, China). HEK293T cells were co-transfected with PGMLV-based luciferase reporter constructs containing either the wild-type (WT) or mutant (MUT) ASNS and ATF3 promoter sequences. A Renilla luciferase plasmid was included for normalization, and transfection was carried out using GMTrans Liposomal Transfection Reagent. At 48 h post-transfection, luciferase activity was measured using a dual-luciferase reporter assay system. Firefly luciferase activity was normalized to the corresponding Renilla luciferase activity in each sample.

Chromatin immunoprecipitation (ChIP)-qPCR

ChIP assays were performed using a commercial ChIP kit according to the manufacturer’s instructions. Briefly, cells were cross-linked with formaldehyde, and chromatin was extracted and sonicated into fragmented DNA. Chromatin complexes were immunoprecipitated using anti-ATF3 antibody, with normal IgG serving as a negative control. After reverse cross-linking, bound DNA was purified and subjected to qPCR amplification targeting the predicted ATF3-binding region of the ASNS promoter. The fold enrichment of target DNA was calculated and normalized to the input control. All experiments were repeated three times independently.

Subcutaneous and lung metastasis xenograft models

All animal experiments were conducted in accordance with protocols approved by the Animal Ethics Committee of Fudan University (Approval No.: 202411FD0007). Six-week-old male BALB/c nude mice were purchased from GemPharmatech (Beijing, China) and were randomly assigned to experimental groups with matched body weights. For the subcutaneous xenograft model, A549 cells (5 × 10⁶) suspended in 0.1 mL of sterile PBS were injected into the flank of each mouse. For the lung metastasis model, an equivalent number of GFP-labeled A549 cells in PBS were administered via tail vein injection. In drug intervention studies, one week after tumor cell inoculation, mice received intraperitoneal injections every other day of either 100 µL PBS (vehicle control) or 2-deoxy-D-glucose (2-DG, 500 mg/kg body weight) [19, 20]. All mice were euthanized 4–6 weeks post-inoculation. Subcutaneous tumors and lungs were harvested, fixed, and sectioned for histological analysis. Tissue sections were stained with hematoxylin and eosin (H&E) or immunohistochemically with specific antibodies against ASNS and KI-67. Upon completion of the experiment, subcutaneous tumors were measured, and their volumes were calculated using the formula: volume = length × width × height × 0.52.

Immunofluorescence assay

Transfected cells were seeded into confocal dishes and cultured for 24 h. Following the removal of the culture medium, the cells were fixed with 4% paraformaldehyde for 30 min and subsequently washed three times with PBS. Permeabilization was conducted using 0.3% Triton X-100 for 15 min, after which the cells were washed three additional times with PBS. The samples were then blocked with 5% bovine serum albumin (BSA) for 2 h at room temperature. The cells were incubated overnight at 4 °C with the primary antibodies: anti-ASNS (68004-1-Ig, Proteintech, China) and anti-ATF3 (18665 S, CST, USA). Following PBS washing, the cells were incubated with species-matched fluorescent secondary antibodies for 1 h in the dark. Finally, nuclei were stained with DAPI, and fluorescence signals were visualized using a confocal laser scanning microscope.

Immunohistochemistry

Tissue samples obtained from subcutaneous tumors or lung metastases in nude mice were fixed in 4% paraformaldehyde and subsequently embedded in paraffin blocks. The sections were then deparaffinized, rehydrated, and subjected to antigen retrieval. Following this, the samples were incubated overnight at 4 °C with specific primary antibodies targeting ASNS (92479, CST, USA) and Ki-67 (9449T, CST, USA). After thorough washing, the sections were incubated with the corresponding secondary antibodies for 2 h at room temperature. Immunostaining was developed using a DAB substrate, and the slides were counterstained with hematoxylin. All images were analyzed using Aipathwell^® analysis software (Servicebio, Wuhan, China).

RNA sequencing and transcriptome analysis

Total RNA was extracted from cancer cells using the RNA-Quick Purification Kit (RN001) in accordance with the manufacturer’s instructions. RNA sequencing was conducted by Novogene (Beijing, China). Libraries were constructed and sequenced using the Illumina platform. Read counts mapped to each gene were calculated and normalized as fragments per kilobase of transcript per million mapped reads (FPKM). Differential expression analysis was performed using DESeq2 (v1.30.0). Genes with |log2 fold change| > 1 and false discovery rate (FDR) < 0.05 were considered differentially expressed. Each experimental condition contained three biological replicates.

Targeted metabolomics analysis

Targeted metabolomics analysis was conducted by Biapply Biotech (Shanghai, China). In brief, A549 cells were collected and rapidly frozen in liquid nitrogen. The samples underwent liquid chromatography-tandem mass spectrometry (LC‒MS/MS) analysis. A standard curve was generated for each metabolite, allowing for the quantification of metabolite concentrations in each unknown sample based on the corresponding calibration curve. LC‒MS/MS data were normalized to total ion count and log2‑transformed. Missing values were imputed using half of the minimum detected value. Each experimental group contained 6 biological replicates. Metabolites with a coefficient of variation > 30% across replicates were excluded from further analysis.

ATAC-seq

The ATAC-seq assay was conducted by LC-Bio (Hangzhou, China) to assess genome-wide chromatin accessibility. In brief, 20,000 to 50,000 nuclei were freshly isolated from A549 cells. The isolated nuclei underwent tagmentation using the Tn5 transposase (Illumina, San Diego, USA). Following tagmentation, DNA fragments were purified using a specialized kit and subsequently amplified by PCR to construct sequencing libraries. The resulting ATAC-seq libraries were sequenced on the NovaSeq 6000 platform (Illumina, San Diego, USA), with the sequencing depth determined based on the experimental design.

Bioinformatics

The data utilized in this study were sourced from publicly accessible databases, including The Cancer Genome Atlas (TCGA) (https://portal.gdc.cancer.gov/), Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/), and Gene Expression Profiling Interactive Analysis (GEPIA) (http://gepia.cancer-pku.cn/index.html). The comparison of ASNS mRNA expression between tumor and adjacent normal tissues in patients with lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) was conducted using GEPIA. Kaplan-Meier survival analysis for overall survival in NSCLC patients was performed based on TCGA datasets, as well as the GSE30219 and GSE37745 datasets, via the Kaplan-Meier Plotter platform (https://kmplot.com/analysis/). Predictions of transcription factors were executed using hTFtarget (http://bioinfo.life.hust.edu.cn/hTFtarget#!/prediction) and KnockTF (http://www.licpathway.net/KnockTFv2/index.php).

Statistical analysis

Statistical analyses were conducted using SPSS 26 (IBM, USA) and GraphPad Prism 9 (GraphPad Software, USA). Data from at least three independent experiments are presented as mean ± standard deviation. Comparisons between two groups were analyzed using paired or unpaired Student’s t-tests, as appropriate. For comparisons among multiple groups, one‑way or two‑way analysis of variance (ANOVA) was performed, followed by Tukey’s or Dunnett’s post‑hoc test for multiple comparisons. Survival analysis was performed using the Kaplan–Meier method, with differences evaluated by the log-rank test. P‑values are reported throughout the manuscript. A p-value of less than 0.05 was deemed statistically significant.

Results

Glucose deprivation upregulates ASNS in NSCLC

Metabolic reprogramming serves as a critical adaptive strategy that facilitates tumor cell proliferation in nutrient-scarce microenvironments. To elucidate the molecular mechanisms underlying the response of NSCLC to glucose deprivation, we conducted an integrated analysis combining RNA-seq and LC‒MS/MS-based targeted metabolomics to systematically profile alterations in gene expression and metabolite levels under glucose-deprived conditions. The results from RNA-seq demonstrated that glucose deprivation significantly enriched pathways associated with aspartate metabolism in A549 and H1299 cells (Fig. 1A-B) and specifically upregulated the expression of ASNS, a key enzyme in this pathway (Fig. 1C). These findings were further supported by LC‒MS/MS metabolomics data obtained from A549 cells (Fig. 1D-G). To confirm the generalizability and sustainability of ASNS upregulation, we performed time‑course experiments under glucose deprivation for 24, 48, and 72 h in A549, H1299, H838, H1975 and PC9 cells. ASNS mRNA and protein levels progressively increased over time, peaking at 72 h, indicating a sustained rather than transient response (Fig. 1H-J).

Fig. 1.

Glucose deprivation upregulates ASNS in NSCLC. (A) Transcriptomic sequencing analysis of differentially expressed genes (DEGs) in A549 and H1299 cells after 24-hour glucose deprivation. (B) Gene Set Enrichment Analysis (GSEA) of DEGs showing significant enrichment of aspartate and glutamate metabolism pathways. (C) Venn diagram of commonly upregulated DEGs in A549 and H1299 cells, with ASNS highlighted as a key intersected gene. (D) Principal component analysis (PCA) score plot of metabolites in A549 cells after 24-hour glucose deprivation as detected by LC-MS/MS. (E) Hierarchical clustering heatmap of differential metabolites. (F) KEGG pathway enrichment analysis of differential metabolites. (G) Relative change in asparagine abundance under glucose deprivation. (H) ASNS mRNA levels detected by qRT-PCR under normal glucose or glucose deprivation for 24, 48 and 72 h; (I-J) ASNS protein levels detected by Western blot under normal glucose or glucose deprivation for 24, 48 and 72 h. All results are representative of three independent experiments. P < 0.05 was considered statistically significant

ASNS is significantly upregulated in NSCLC and correlates with poor patient prognosis

To investigate the expression pattern of ASNS in clinical specimens of NSCLC, we first assessed its protein levels through Western blot analysis in seven pairs of fresh NSCLC tumor tissues and adjacent nontumorous (NT) tissues. The results indicated that ASNS expression was significantly elevated in tumor tissues compared with that in NT tissues (P = 0.022; Fig. 2A, B). We subsequently performed IHC staining for ASNS on tissue microarrays from 80 NSCLC patients, which included tumor tissues and their corresponding NT tissues (lung tissue collected 2–3 cm from the tumor margin). As illustrated in Fig. 2C, ASNS was overexpressed in tumor tissues but was only faintly detectable in adjacent NT tissues. Representative IHC images demonstrating weak, moderate, and strong ASNS staining in tumor tissues are presented in Fig. 2D. A quantitative assessment utilizing the H-score (which integrates staining intensity and the percentage of positive cells) was conducted, and a heatmap visually represents the differential expression profile of ASNS between the 80 paired tumor and NT tissues (Fig. 2E). To elucidate the clinical significance of ASNS expression, we analyzed its correlation with patient prognosis. Kaplan‒Meier survival analysis based on the IHC scores of the 80 patients revealed that those with low ASNS expression exhibited significantly longer overall survival than those with high expression (P = 0.042; Fig. 2F). Additionally, we examined the relationships between ASNS expression levels and various clinicopathological characteristics of the patients (Fig. 2G, Supplementary Table 5). Univariate and multivariate Cox regression analyses were subsequently performed (Fig. 2H, Supplementary Table 6). High ASNS expression was identified as an independent prognostic factor (OR = 3.006, 95% CI: 1.037–8.710, P = 0.043), as were male gender (OR = 3.204;P = 0.045) and age ≥ 60 years (OR = 3.136; P = 0.041).To further validate these findings, we conducted supplementary analyses using public databases. Data from TCGA demonstrated that ASNS mRNA levels were significantly higher in tumor tissues than in NT tissues for both LUAD and LUSC (P < 0.05; Fig. 2I). Similarly, survival analyses based on the GSE30219 (Fig. 2J) and GSE37745 (Fig. 2K) datasets confirmed that NSCLC patients with low ASNS expression had significantly better overall survival than those with high expression (P < 0.05).

Fig. 2.

ASNS is significantly upregulated in NSCLC and correlates with poor patient prognosis. (A, B) Western blot analysis comparing ASNS protein expression levels between 7 paired fresh NSCLC tumor tissues and adjacent non-tumorous (NT) tissues. (C) Immunohistochemical analysis of representative tumor and corresponding NT tissue sections, showing ASNS expression. (D) Representative immunohistochemical images showing weak, moderate, and strong positive staining for ASNS in NSCLC tumor tissues. (E) Heatmap displaying the differential expression profile of ASNS between 80 paired NSCLC tumor and NT tissues. (F) Kaplan-Meier analysis of overall survival for the 80 NSCLC patients stratified by ASNS expression levels. (G) Comparison of clinicopathological characteristics between high and low ASNS expression groups in the 80 NSCLC patients. (H) Univariate and multivariate logistic regression analyses of clinicopathological characteristics associated with patient prognosis. (I) Comparison of ASNS mRNA expression between tumor and NT tissues from LUAD and LUSC patients in the TCGA database. (J) Kaplan-Meier analysis of overall survival for NSCLC patients in the TCGA datasets, stratified by ASNS expression levels. (K, L) Kaplan-Meier analysis of overall survival for NSCLC patients in the GSE30219 (K) and GSE37745 (L) datasets, stratified by ASNS expression levels. P < 0.05 was considered statistically significant

Knockdown of ASNS suppresses malignant phenotypes of NSCLC cells under glucose deprivation

To investigate the functional impact of ASNS inhibition under glucose deprivation, we first assessed ASNS protein levels in a panel of human NSCLC cell lines (A549, H1299, H838, PC9, 95D) as well as in normal human bronchial epithelial cells (BEAS-2B) using Western blot analysis. Compared with BEAS-2B cells, the majority of NSCLC cell lines exhibited elevated ASNS expression (Fig. 3A). Based on these findings, we selected two high-expressing lines, A549 and H1299, for ASNS knockdown experiments, and one low-expressing line, H838, for subsequent overexpression studies. Stable ASNS-knockdown A549 and H1299 cell lines were established through lentiviral infection (Fig. 3B; Supplementary Fig. 1A), and reduced ASNS expression was confirmed under both normal and glucose-deprived conditions (Fig. 3C; Supplementary Fig. 1B).

Fig. 3.

Knockdown of ASNS suppresses malignant phenotypes of NSCLC cells in vitro and in vivo. (A) Western blot analysis of ASNS protein levels in a panel of human NSCLC cell lines (A549, H1299, H838, PC9, 95D) and normal human bronchial epithelial cells (BEAS-2B). (B) Western blot showing ASNS knockdown efficiency in A549 cells. (C) ASNS expression in A549 cells under normal [Glu (+)] and glucose-deprived [Glu (-)] conditions following knockdown. (D) CCK-8 assay assessing proliferation of A549 cells under normal (C) and glucose deprived conditions (D). (E) Colony formation assay of A549 cells after ASNS knockdown. (F) EdU assay evaluating DNA synthesis in A549 cells. Scale bar, 100 μm. (G) Wound healing assay showing migratory ability of A549 cells. Scale bar, 100 μm. (H) Transwell assay showing invasive ability of A549 cells. Scale bar, 100 μm. (I) Representative images of subcutaneous tumors derived from control (shCtrl) and ASNS-knockdown (shASNS) A549 cells. (J) Tumor volume measurements. (K) Tumor weight measurements. (L) Representative immunohistochemical staining of Ki67 and ASNS in xenograft tissues. Scale bar, 50 μm. (M) Quantitative analysis of Ki67-positive cells. (N) Quantitative analysis of ASNS immunohistochemistry scores. (O) Western blot analysis of ASNS expression in five pairs of xenograft tumors. (P) Representative in vivo fluorescence images of mice in the tail vein lung metastasis model, showing reduced pulmonary metastatic signals in the shASNS group. (Q) Representative gross photographs of lung tissues (upper panel) and representative H&E-stained lung sections (lower panel) at the experimental endpoint. All results are representative of three independent experiments. P < 0.05 was considered statistically significant

Functional assays demonstrated that ASNS knockdown significantly suppressed the proliferative capacity of A549 and H1299 cells under both normal and glucose-deprived conditions. This conclusion was consistently supported by the results of the CCK-8 assays (Fig. 3D; S Figure. 1 C, D), colony formation assays (Fig. 3E; S Figure. 1E), and EdU assays (Fig. 3F; S Figure. 1 F).

Furthermore, wound healing assays (Fig. 3G; S Figure. 1G) and Transwell assays (Fig. 3H; S Figure. 1 H) revealed that ASNS knockdown markedly inhibited the migratory and invasive abilities of both cell lines under the respective culture conditions. Collectively, these results indicate that ASNS plays a crucial role in regulating the proliferation, migration, and invasion of NSCLC cells, particularly under conditions of glucose deprivation.

To further validate the tumor-promoting role of ASNS in vivo, we established subcutaneous xenograft models using A549 cells stably expressing control shRNA (shCtrl) or ASNS-targeting shRNA (shASNS). Representative images of excised tumors from seven paired mice are presented in Fig. 3I, indicating a trend toward reduced tumor size in the knockdown group. Quantitative analysis confirmed that both tumor volume (Fig. 3J) and tumor weight (Fig. 3K) were significantly lower in the shASNS group than in the shCtrl group. Immunohistochemical staining of xenograft tissues revealed decreased expression of Ki67 and ASNS in the shASNS group (Fig. 3L). Statistical analysis of the immunohistochemistry scores further revealed significantly lower Ki67-positive rates (Fig. 3M) and ASNS expression levels (Fig. 3N) in tumors from the knockdown group. Finally, Western blot expression analysis of tumor lysates confirmed the effective downregulation of ASNS protein in shASNS xenografts (Fig. 3O). These in vivo findings consistently support the conclusion that ASNS knockdown attenuates tumor growth and proliferation in NSCLC.

To further evaluate the impact of ASNS on the distant metastatic capacity of tumors, we established a tail-vein lung metastasis model. In vivo imaging revealed that, compared with the shCtrl group, mice injected with shASNS cells exhibited a significantly reduced fluorescence signal intensity in the lungs (Fig. 3P), indicating a diminished metastatic burden. Analysis of lung tissues at the experimental endpoint revealed fewer metastatic nodules on the lung surface in the shASNS group (Fig. 3Q, upper panel). Hematoxylin and eosin (H&E) staining of lung tissue sections further demonstrated that both the area and the number of intrapulmonary metastatic foci were significantly lower in the shASNS group than in the control group (Fig. 3Q, lower panel). Together, these in vivo data demonstrate that ASNS knockdown not only inhibits the growth of primary tumors but also effectively impairs the distant metastatic capacity of NSCLC cells.

Overexpression of ASNS enhances malignant phenotypes of NSCLC cells under glucose deprivation

To further validate the functional role of ASNS, we employed a gain-of-function approach by establishing an ASNS overexpression model in H838 cells, which exhibit low basal ASNS expression. Through lentiviral infection, we successfully generated stable ASNS-overexpressing cell lines (Fig. 4A) and confirmed a significant increase in ASNS expression under both normal and glucose-deprived conditions (Fig. 4B).

Fig. 4.

Overexpression of ASNS promotes malignant phenotypes of NSCLC cells in vitro. (A) Western blot showing the overexpression efficiency of ASNS in H838 cells. (B) ASNS expression in H838 cells under normal [Glu (+)] and glucose-deprived [Glu (-)] conditions following overexpression. (C, D) CCK-8 assay assessing the proliferation of H838 cells under normal (C) and glucose-deprived (D) conditions. (E) Colony formation assay of H838 cells after ASNS overexpression. (F) EdU assay evaluating DNA synthesis in H838 cells. Scale bar, 100 μm. (G) Wound healing assay showing the migratory ability of H838 cells. Scale bar, 100 μm. (H) Transwell assay showing the invasive ability of H838 cells. Scale bar, 100 μm. All results are representative of three independent experiments. P < 0.05 was considered statistically significant

Functional assays demonstrated that ASNS overexpression markedly enhanced the proliferative capacity of H838 cells in both normal and glucose-deprived environments. This conclusion was consistently supported by the results of the CCK-8 (Fig. 4C, D), colony formation (Fig. 4E), and EdU (Fig. 4F).

Furthermore, wound healing assays (Fig. 4G) and Transwell assays (Fig. 4H) revealed that ASNS overexpression significantly increased the migratory and invasive abilities of H838 cells under the appropriate culture conditions. Collectively, these gain-of-function results contrast with the findings from the aforementioned knockdown experiments, further confirming the critical role of ASNS in regulating the proliferation, migration, and invasion of NSCLC cells, particularly under conditions of glucose deprivation.

ASNS promotes tumor cell survival under glucose deprivation by regulating amino acid metabolic homeostasis

The aforementioned experiments demonstrated that tumor cells upregulate ASNS to sustain survival under the stress of glucose deprivation. While the primary function of ASNS is to catalyze asparagine synthesis, alterations in a single amino acid are insufficient to fully elucidate its significant prosurvival effects. We hypothesized that ASNS may function as a metabolic hub, facilitating cellular adaptation to energy crises through extensive metabolic reprogramming. To systematically elucidate the metabolic mechanisms by which ASNS drives tumor survival under energy stress, we conducted LC‒MS/MS-based metabolomic profiling on four groups of cells: normal glucose control (Glu(+)_Ctrl), glucose deprivation control (Glu(-)_Ctrl), ASNS knockdown under normal glucose (Glu(+)_shASNS), and ASNS knockdown under glucose deprivation (Glu(-)_shASNS). This analysis aimed to identify key metabolites and pathways dependent on ASNS. Principal component analysis (PCA) revealed a distinct separation in the metabolic profiles among the four groups (Fig. 5A), indicating that both glucose deprivation and ASNS knockdown induced metabolic reprogramming. To accurately identify ASNS-dependent metabolites, we screened for metabolites whose expression was upregulated in the Glu(-)_Ctrl group but subsequently downregulated in the Glu(-)_shASNS group (compared with that in the Glu(-)_Ctrl group) (Fig. 5B). These overlapping metabolites represent key metabolic nodes specifically sustained by ASNS upregulation under glucose deprivation stress. Hierarchical clustering analysis further demonstrated that these metabolites were enriched in the Glu(-)_Ctrl group but decreased in the Glu(-)_shASNS group, confirming their regulatory dependence on ASNS (Fig. 5C).

Fig. 5.

ASNS promotes tumor cell survival under glucose deprivation by regulating amino acid metabolic homeostasis. (A) Principal component analysis (PCA) score plot of targeted metabolomics data from cells under different treatment conditions. (B) Venn diagram showing overlapping metabolites that were upregulated in the glucose deprivation control group (Glu(-)_Ctrl) and downregulated in the glucose deprivation combined with ASNS knockdown group (Glu (-) shASNS). (C) Hierarchical clustering heatmap of the overlapping metabolites across three groups: Glu(+)_Ctrl, Glu(-)_Ctrl, and Glu (-)_shASNS. (D) KEGG pathway enrichment analysis of differential metabolites between the Glu(-)_Ctrl and Glu(+)_Ctrl groups. (E) KEGG pathway enrichment analysis of differential metabolites between the Glu(-)_shASNS and Glu(-)_Ctrl groups. (F–Q) Bar graphs showing the relative levels of key ASNS-dependent amino acids across the three cell groups. P < 0.05 was considered statistically significant

Pathway enrichment analysis indicated that, compared with the Glu (+) Ctrl group, the Glu (-) Ctrl group exhibited significant enrichment in pathways associated with amino acid metabolism, the tricarboxylic acid (TCA) cycle, and fatty acid oxidation (Fig. 5D), demonstrating a metabolic adaptation to energy crisis. Notably, in the comparison between the Glu(-)_shASNS and Glu(-)_Ctrl groups, several amino acid metabolism pathways—including arginine and proline metabolism, as well as alanine, aspartate, and glutamate metabolism—were significantly reversed (Fig. 5E). These findings strongly indicate that ASNS plays a crucial role in maintaining amino acid metabolic homeostasis during glucose deprivation. To validate these findings, we concentrated on key molecules involved in amino acid metabolism pathways. As illustrated in Fig. 5F–Q, the levels of various amino acids—including asparagine, proline, citrulline, and ornithine—were significantly elevated in the Glu(-)_Ctrl group than in the Glu (+) Ctrl group, while ASNS knockdown effectively impeded this increasing trend. These results suggest that under conditions of glucose deprivation, the upregulation of ASNS is essential for sustaining intracellular amino acid pool homeostasis and initiating adaptive metabolic reprogramming. Its deficiency results in the disruption of critical survival metabolic networks.

To further validate the functional contribution of these ASNS-dependent amino acids, we exogenously supplemented ASNS-knockdown A549 and H1299 cells with asparagine, proline, or citrulline under glucose deprivation. EdU and Transwell assays revealed that supplementation with each of these amino acids partially rescued the proliferation, migration and invasion defects caused by ASNS knockdown (Supplementary Fig. 2A–J), confirming that ASNS promotes tumor cell survival under glucose deprivation by maintaining the metabolic balance of multiple amino acids.

The transcription factor ATF3 transcriptionally activates ASNS under glucose deprivation and promotes the malignant progression of NSCLC

To investigate the upstream regulatory mechanisms controlling ASNS expression during glucose deprivation, we focused on its transcriptional regulation. First, to identify potential transcription factors whose binding accessibility may be influenced by ASNS expression, we assessed chromatin accessibility by comparing shCtrl and shASNS cells using ATAC-seq. ASNS knockdown induced extensive alterations in chromatin accessibility, particularly around the transcription start sites (TSSs) of numerous genes (Fig. 6A). Normalized read density plots centered on the ASNS TSSs confirmed altered accessibility upon ASNS knockdown (Supplementary Fig. 3A). Quantitative differential accessibility analysis of the ASNS promoter region revealed increased normalized read density in shASNS cells (Supplementary Table 7). To identify the key regulatory factors associated with these changes, we integrated multiple datasets: (1) results from transcription factor motif enrichment analysis conducted on differential ATAC-seq regions between shCtrl and shASNS cells; (2) genes upregulated under glucose-deprived [Glu(-)] conditions compared with normal glucose [Glu(+)] conditions from RNA-seq analysis; and (3) potential upstream regulators of ASNS predicted by the hTFtarget and KnockTF databases. Intersection analysis using a Venn diagram consistently revealed the expression of the stress-responsive transcription factor ATF3 (Fig. 6B). Further analysis of ATAC-seq signals revealed a significant decrease in chromatin accessibility at the ATF3 motif within the promoter region of the ASNS gene in shASNS cells compared with that in shCtrl cells (Fig. 6C). A specific ATF3 binding site in this region was predicted based on the ATAC-seq data (Fig. 6D).

Fig. 6.

The transcription factor ATF3 transcriptionally activates ASNS under glucose deprivation and promotes the malignant progression of NSCLC. (A) Heatmap of average ATAC-seq signal profiles centered on the nearest gene TSSs in ASNS-knockdown (shASNS) and control (shCtrl) cells. (B) Venn diagram illustrating the intersection of multiple datasets used to identify the upstream transcription factor of ASNS, including: ATAC-seq TF motif enrichment results, RNA-seq upregulated genes under glucose deprivation, and predictions from the hTFtarget and KnockTF databases, converging on ATF3. (C) ATAC-seq signal tracks in the promoter region of the ASNS gene in control and ASNS-knockdown cells, showing decreased chromatin accessibility in this region upon ASNS knockdown. (D) Schematic diagram of the predicted ATF3 binding site within the ASNS promoter region. (E) Western blot analysis of ATF3 and ASNS protein levels following ATF3 knockdown in A549, H1299, and H838 cell lines. (F) Dual-luciferase reporter assay detecting the effect of ATF3 on the activity of the wild-type and mutant ASNS promoters. (G) ChIP-qPCR analysis of ATF3 occupancy at the ASNS promoter in A549, H1299 and H838 cells. (H) Immunofluorescence (IF) detection of ATF3 (red) and ASNS (yellow) protein expression and localization in H838 cells after ATF3 knockdown, with DAPI (blue) staining indicating nuclei. (I-L) Functional rescue experiments verifying that ATF3 mediates its biological functions through ASNS. (I) Representative images and quantification of colony formation assays. (J) Representative images and quantification of EdU cell proliferation assays. Scale bar, 100 μm. (K) Representative images and quantification of wound healing assays. Scale bar, 100 μm. (L) Representative images and quantification of Transwell cell migration and invasion assays. Scale bar, 100 μm. All results are representative of three independent experiments. P < 0.05 was considered statistically significant

To validate the direct regulatory effect of ATF3 on ASNS, we knocked down ATF3 in three lung cancer cell lines: A549, H1299, and H838. Western blot analysis confirmed that the downregulation in ATF3 directly resulted in a reduction of ASNS protein levels (Fig. 6E). A dual-luciferase reporter assay revealed that ATF3 potentiated the transcriptional activity of the ASNS promoter under glucose deprivation (Fig. 6F). To further confirm the direct binding of ATF3 to the ASNS promoter, we performed ChIP‑qPCR in A549, H1299, and H838 cells. Compared with that in the siNC group, the enrichment of ATF3 at the predicted binding region of the ASNS promoter in the ATF3 knockdown group was significantly lower (Fig. 6G). Immunofluorescence staining provided direct visualization at the cellular level, confirming that ATF3 knockdown in H838 cells led to a concurrent decrease in both nuclear ATF3 and cytoplasmic ASNS protein signals (Fig. 6H).

Functionally, our results confirm that ATF3 is indispensable for sustaining malignant phenotypes in lung cancer cells. In H838 cells, ATF3 knockdown suppresses suppressed colony formation, cell proliferation, migration and invasion by 52.3%, 45.2%, 53.1% and 65.3% respectively. The corresponding reduction rate was determined via the formula (shCtrl minus siATF3) divided by shCtrl multiplied by 100%. Exogenous ASNS overexpression substantially reversed such phenotypic impairment with respective rescue efficiencies of 60.7%, 65.2%, 33.0% and 44.4%. The rescue efficiency was calculated as the value of siATF3 plus OEASNS minus siATF3 divided by shCtrl minus siATF3 and then multiplied by 100% (Fig. 6I–L).

Consistent rescue outcomes were validated in A549 and H1299 cells as documented in Supplementary Fig. 3. Analyses across these three NSCLC cell lines revealed that ATF3 silencing reduced colony formation by 24–52%, cell proliferation by 45–54%, migration by 41–53% and invasion by 50–65%. Additional ASNS overexpression restored these malignant capacities at partial recovery rates ranging from 45 to 92% for colony formation, 45–65% for proliferation, 66–84% for migration and 36–62% for invasion. These results confirm that ATF3 drives malignant progression of NSCLC largely although transcriptional upregulation of ASNS, though ASNS is not the sole downstream effector mediating ATF3-driven oncogenic phenotypes.

Glucose deprivation upregulates ATF3 and ASNS expression through activation of the MAPK signaling pathway

To elucidate the upstream signaling pathway responsible for ATF3 activation under glucose deprivation stress, we first analyzed RNA-seq data from A549 and H1299 cells subjected to glucose deprivation. Intersection analysis revealed a set of commonly upregulated genes (Fig. 7A), and KEGG pathway enrichment analysis revealed significant enrichment of the ‘MAPK signaling pathway’ among these genes (Fig. 7B). This finding suggests that MAPK pathway activation may serve as a critical link connecting glucose deprivation to the ATF3/ASNS axis. To test this hypothesis, we examined the activation status of the MAPK pathway in three lung cancer cell lines. Western blot analysis confirmed that glucose deprivation significantly increased the phosphorylation levels of p38, ERK, and JNK, and coordinately upregulated the protein expression of their downstream effector ATF3 and its target gene ASNS (Fig. 7C). To establish a causal relationship, we employed specific pharmacological inhibitors: Adezmapimod (a p38 inhibitor), U0126 (an ERK inhibitor), and SP600125 (a JNK inhibitor). Treatment of cells with each inhibitor (10 µM, 24 h) under glucose deprivation effectively reversed the upregulation of ATF3 protein expression. We further examined the phosphorylation levels of p38, ERK, and JNK under the same conditions. The results revealed that each inhibitor specifically suppressed the phosphorylation of its respective target without significantly affecting the other two MAPK branches, confirming the specificity of the inhibitors (Fig. 7D–F).

Fig. 7.

Glucose deprivation upregulates ATF3 and ASNS expression through activation of the MAPK signaling pathway. (A) Venn diagram of genes commonly upregulated in A549 and H1299 cells after glucose deprivation treatment, based on RNA-seq analysis. (B) Bubble plot of KEGG pathway enrichment analysis of the overlapping genes shown in (A). The “MAPK signaling pathway” was significantly enriched. (C) Western blot analysis of phosphorylation levels of key MAPK pathway proteins and expression of ATF3 and ASNS proteins in A549, H1299, and H838 cells after glucose deprivation treatment. (D–F) Western blot analysis showing the effects of MAPK inhibitors on protein expression and phosphorylation. Cells were treated with specific inhibitors (10 µM, 24 h): Adezmapimod (p38 inhibitor, D), U0126 (ERK inhibitor, E), or SP600125 (JNK inhibitor, F). (G–J) Functional rescue experiments in A549 cells. ATF3 was overexpressed with or without MAPK inhibitor treatment under glucose deprivation. ATF3 overexpression significantly enhanced colony formation (G), EdU‑positive cells (proliferation, H), wound healing (migration, I), and Transwell invasion (J), whereas treatment with Adezmapimod, U0126, or SP600125 effectively reversed these malignant phenotypes. Scale bar, 100 μm. All results are representative of three independent experiments. P < 0.05 was considered statistically significant

To determine whether the MAPK pathway mediates the protumorigenic effects of ATF3, we overexpressed ATF3 in A549 cells and then treated them with each MAPK inhibitor under glucose deprivation. The results showed that ATF3 overexpression significantly enhanced colony formation (Fig. 7G), proliferation (Fig. 7H), migration (Fig. 7I), and invasion (Fig. 7J), whereas treatment with the three inhibitors effectively reversed these ATF3‑induced malignant phenotypes. Consistent results were obtained in H1299 and H838 cells (Supplementary Fig. 3), indicating the generalizability of this regulatory mechanism.

To further confirm that MAPK‑driven malignant phenotypes are dependent on ASNS, we overexpressed ASNS in H838 cells treated with MAPK inhibitors. As shown in Supplementary Fig. 5A–E, ASNS overexpression significantly reversed the inhibitor-induced reduction in the number of EdU-positive cells and reversed the impairment of cell migration and invasion. These results demonstrate that the MAPK pathway promotes NSCLC cell proliferation, migration and invasion primarily through upregulating ASNS expression.

Collectively, these findings demonstrate that glucose deprivation activates the p38, ERK, and JNK branches of the MAPK pathway, leading to the upregulation of ATF3 and its downstream target ASNS, which ultimately drives the malignant progression of lung cancer.

Co-targeting glucose metabolism and ASNS additively inhibits NSCLC tumor growth in vivo

To validate the antitumor efficacy of co-targeting glucose and asparagine metabolism in vivo, we established subcutaneous xenograft models in mice using NSCLC cells with shASNS or shCtrl. Tumor-bearing mice were randomly allocated into four groups (n = 5/group) for treatment: shCtrl + PBS (control), shASNS + PBS (ASNS knockdown), shCtrl + 2-DG (glucose metabolism inhibition), and shASNS + 2-DG (combination therapy). 2-DG, a glycolytic inhibitor, was administered via intraperitoneal injection. The combination therapy demonstrated a significant additive antitumor effect (Fig. 8). At the experimental endpoint, tumors in the shCtrl + PBS group exhibited the largest volume and weight. Single interventions, either shASNS + PBS or shCtrl + 2-DG, partially inhibited tumor growth. In contrast, the shASNS + 2-DG combination group displayed the smallest tumor volume (Fig. 8B) and the lowest tumor weight (Fig. 8C), indicating the most potent suppression. To quantitatively evaluate the interaction between ASNS knockdown and 2‑DG treatment, we performed two‑way ANOVA and Bliss independence modeling. Two‑way ANOVA revealed no significant interaction (p = 0.382), and the Bliss combination index (CI) was calculated as 1.01 (95% confidence interval: 0.91–1.12). These results further confirm the additive rather than synergistic nature of the combination. IHC analysis of tumor sections corroborated these findings. The combination group (shASNS + 2-DG) had the lowest positivity rate for the proliferation marker Ki-67 (Fig. 8D, F) alongside downregulated ASNS protein expression (Fig. 8D, E). Together, these in vivo results demonstrate that simultaneous inhibition of glucose metabolism and asparagine synthesis exerts a powerful additive effect, effectively suppressing tumor growth and reducing tumor cell proliferation.

Fig. 8.

Co-targeting glucose metabolism and ASNS additively inhibits NSCLC tumor growth in vivo. (A) Representative photographs of subcutaneous xenograft tumors from each group at the experimental endpoint. (B) Statistical analysis of tumor volume at the endpoint. (C) Statistical analysis of tumor weight at the endpoint. (D) Representative immunohistochemical (IHC) staining images of tumor sections showing the expression of Ki-67 (proliferation marker) and ASNS. Scale bar, 50 μm. (E) Quantitative analysis of ASNS protein expression by IHC, presented as H-score. (F) Quantitative analysis of Ki-67 protein expression by IHC, presented as H-score. P < 0.05 was considered statistically significant

Discussion

This study identified and validated the MAPK/ATF3/ASNS signaling axis as a core regulatory pathway driving the adaptation of NSCLC to glucose deprivation. Under energetic stress, upstream MAPK signaling is activated and induces the expression of the key transcription factor ATF3, which in turn directly upregulates ASNS, triggering extensive reprogramming of amino acid metabolism. This adaptive metabolic remodeling supports tumor cell survival, proliferation, and progression in nutrient‑deprived environments.

Our study established the critical role of ASNS in the response of NSCLC to glucose deprivation. Although previous reports have indicated high ASNS expression in various cancers, including liver, bladder, pancreatic, and lung cancers, as well as hematologic malignancies [21–24], and its association with poor prognosis, chemoresistance, and immune evasion, its specific function and upstream regulatory mechanisms under energetic stress remained unclear. Here, we demonstrate that upregulation of ASNS under glucose deprivation is essential for maintaining clonogenicity, proliferation, and invasion, revealing its role as a key metabolic hub linking nutrient stress signals to malignant phenotypes. These findings align with the adaptive strategy of tumor cells to reprogram amino acid metabolism for survival under energy‑limiting conditions [25–28].

LC-MS/MS further elucidated the downstream metabolic network of ASNS. The knockdown of ASNS not only diminished asparagine synthesis but also induced extensive metabolic disturbances characterized by an imbalance in amino acid pools. Under glucose deprivation, the specific accumulation of multiple amino acids in ASNS-high cells suggests that the upregulation of ASNS establishes a “metabolic pool” that supports tumor cell survival through various pathways. Among these pathways, asparagine plays crucial roles beyond protein synthesis, including protein glycosylation and cellular signal transduction [29]; serine-linked glycine metabolism supports one-carbon metabolism, thereby sustaining nucleotide synthesis and redox balance [30]. The upregulation of arginine, ornithine, and citrulline may increase urea cycle flux, helping to detoxify ammonia generated by high metabolic rates and thus alleviate nitrogen toxicity [31, 32], and proline can be converted by proline oxidase into glutamate, which subsequently enters the TCA cycle or is transformed into other amino acids, providing carbon skeletons for energy metabolism [33]. From a network perspective, aspartate serves as a common precursor for several amino acids, including asparagine, methionine, lysine, and threonine, with its flux regulated by ASNS [34]. In contrast, glutamate functions as a central node in the metabolism of proline, arginine, and ornithine [35]. Consequently, ASNS dynamically reshapes the overall distribution of amino acid metabolic fluxes through utilization of the core metabolite aspartate and further facilitates cellular adaptation and survival upon energy shortage.

To investigate the upstream signaling pathways that drive ASNS expression, we conducted ATAC-seq, RNA-seq, and an integrated multiomics analysis of public databases and identified the transcription factor ATF3 as a key mediator that directly regulates ASNS transcription under conditions of glucose deprivation. ATF3, a basic leucine zipper (bZIP) transcription factor belonging to the ATF/CREB family [36], is a well-known stress-responsive element activated by various stresses, including nutrient deprivation and endoplasmic reticulum (ER) stress [36, 37]. Its downstream targets include not only ASNS [38, 39] but also genes involved in amino acid metabolism [40], autophagy [41], antioxidant responses [42], and DNA repair [43, 44]. In various solid tumors, such as breast, colon, and non-small cell lung cancers, ATF3 has been shown to promote tumor progression and immune evasion and is correlated with poor patient prognosis [45–48]. Our data demonstrate that ATF3 is significantly activated under glucose starvation and directly binds to specific stress-response elements in the ASNS gene promoter, thereby positively regulating its transcription [38].

Further upstream tracing revealed that the coordinated activation of the MAPK-p38/ERK/JNK signaling pathway links glucose deprivation to ATF3 expression [49]. This study confirmed that inhibitors targeting the three major MAPK branches effectively blocked ATF3 upregulation and the subsequent induction of ASNS [50, 51]. The coordinated response of these three MAPK signals is crucial for sustaining tumor cell survival under energy stress. Biologically, the p38 MAPK pathway serves as a central stress sensor [52] and, in conjunction with the ERK and JNK pathways, integrates glucose deprivation signals and transmits them to ATF3. ATF3 then directly drives the reprogramming of metabolic genes, including ASNS, thereby establishing an extensive metabolic buffer pool. In the pathological context of tumors, the activation of this metabolic axis under glucose deprivation represents a key mechanism that enables metabolic adaptation, sustained proliferation, and resistance to therapy.

Poorly structured vasculature generates persistent glucose-deprived niches within solid tumors, suggesting that glucose deficiency-induced ASNS upregulation is an intrinsic adaptive response to intratumoral metabolic stress [53, 54]. Consistent with these findings, our data demonstrate that compared with paired normal tissues, NSCLC tumor tissues have elevated ASNS expression and that high ASNS levels are correlated with unfavorable patient survival across independent cohorts. These results support therapeutic targeting of the MAPK/ATF3/ASNS cascade or asparagine metabolism in ASNS-high NSCLC. For patient stratification, integrating ASNS with metabolic biomarkers (GLUT1, HIF‑1α, or glucose deprivation signatures) may screen candidates susceptible to combined regimens of 2-DG and ASNS suppression [55, 56].

Through in vitro and in vivo experiments, our study preliminarily confirmed that ASNS upregulation is critical for maintaining tumor cell survival under glucose deprivation; its transcriptional regulation is mediated primarily by the stress-responsive factor ATF3, with upstream MAPK signaling acting as a major activator of ATF3. Notably, ASNS overexpression partially rescueed the impaired proliferation, migration and invasion induced by ATF3 depletion (Fig. 6I–L, Supplementary Fig. 3), demonstrating that ATF3 drives NSCLC progression predominantly via ASNS upregulation but also through ASNS‑independent pathways.

In vivo, combined ASNS knockdown and 2‑DG treatment yielded additive rather than synergistic tumor suppression (Fig. 8). A fixed 2‑DG dosage (500 mg/kg, q.o.d.) and constant ASNS silencing efficiency likely preclude optimal dosing ratios for synergy; the small cohort size further limits the statistical sensitivity to detect mild synergism. Nevertheless, co‑treatment outperformed single-agent therapy in inhibiting tumor growth, validating the prospect of dual-targeted intervention. However, the detailed molecular mechanisms by which ASNS-driven broad amino acid reprogramming fine-tunes protein synthesis efficiency, modulates signaling pathway activity, and influences epigenetic states remain to be systematically elucidated.

Conclusion

In summary, our study elucidates a crucial mechanism through which tumor cells survive under conditions of glucose deprivation through metabolic reprogramming. Specifically, energy stress activates MAPK signaling, which subsequently induces the expression of the transcription factor ATF3. ATF3 functions as a central molecular switch that directly upregulates the expression of ASNS. The increase in ASNS not only ensures a sufficient supply of asparagine but also drives the coordinated remodeling of multiple amino acid pathways, including proline and arginine-ornithine metabolism, thereby establishing a comprehensive metabolic buffer pool. This buffer pool systematically enhances tumor cell adaptation and survival during energy crises by maintaining substrates for protein synthesis, improving ammonia detoxification, and replenishing the TCA cycle. Furthermore, this metabolic adaptation program, driven by the MAPK/ATF3/ASNS axis, remains persistently activated within the tumor microenvironment, thereby facilitating malignant progression and resistance to therapy.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

NSCLC

Non-small cell lung cancer

ASNS

Asparagine synthetase

MAPK

Mitogen-activated protein kinase

ATF3

Activating transcription factor 3

IHC

Immunohistochemical

LC-MS/MS

Liquid chromatography-tandem mass spectrometry

TCGA

The Cancer Genome Atlas

GEO

Gene Expression Omnibus

GEPIA

Gene Expression Profiling Interactive Analysis

LUAD

Lung adenocarcinoma

LUSC

Lung squamous cell carcinoma

TCA

Tricarboxylic acid

TSSs

Transcription start sites

2-DG

2-deoxy-D-glucose

bZIP

Basic leucine zipper

ER

Endoplasmic reticulum

Author contributions

Fuzhi Yang: Writing-original draft, Visualization, Validation, Software, Methodology, Data curation, Conceptualization. Hao Qin: Data curation, Visualization, Conceptualization, Writing-review & editing. Xuelin Zhang: Visualization, Formal analysis, Conceptualization. Yun Zhao: Software, Formal analysis, Resources Shuai Jiang: Validation, Methodology, Formal analysis. Xiaoyu Chen: Validation, Software, Investigation. Zheng Li: Validation, Software, Conceptualization. Jing Wang: Validation, Software. Dongfang Tang: Conceptualization, Writing-review & editing. Wen Gao: Writing-review & editing, Supervision, Resources. Xunxia Zhu: Writing-review & editing, Resources, Project administration. Xiaoyong Shen: Writing-review & editing, Supervision, Conceptualization, Resources.

Funding

This research is supported by the National Natural Science Foundation of China (grant No. 82376211).

Data availability

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The animal study received approval from the Animal Ethics Committee of Fudan University and was conducted in accordance with the protocol sanctioned by the ethics committee (Approval No.: 202411FD0007).

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.