We obtained transcriptome profiles and sample data from TCGA (https://portal.gdc.cancer.gov/), CCLE (http://www.sites.broadinstitute.org/ccle) and the Genotype-Tissue Expression (GTEx; (https://gtexportal.org/home/). Protein characterizations were obtained from The Cancer Proteome Atlas (TPCA; http://bioinformatics.mdanderson.org/main/TCPA: Overview) and CPTAC (https://pdc.cancer.gov/pdc/browse). Additionally, UCSC Xena databases (https://xenabrowser.net/datapages/) also provided most of these datasets we used.

CCLE RNA-seq data underwent TPM normalization and were filtered to retain genes expressed in >80% samples. Differential expression analyses were conducted via the limma package in R with empirical Bayes moderation. ACACA upregulation was defined by adj. p<0.001 and log₂FC>1.5. Optimal survival-based cutoff values for ACACA mRNA expression were identified using the “surv_cutpoint” R function, patient were categorized into two distinct subgroups, namely ACACA-High and ACACA-Low, according to the established threshold values. The limma package facilitated the analysis of differential gene expression, with significance defined as adjusted P < 0.05. Univariate Cox proportional hazards regression (using survival R package, v3.7.0) was used to calculate hazard ratios (HRs) for associations between ACACA expression and four survival endpoints: overall survival (OS), disease specific survival (DSS), disease free interval (DFI) and progression free interval (PFI). The “survfit” function was employed to construct Kaplan-Meier survival curves, and survival differences between groups were statistically evaluated via log-rank tests. The receiver operating characteristic (ROC) curves were computed using the pROC package (v1.18.0) to assess the predictive performance of the models, the evaluation of diagnostic performance was conducted by calculating the area under the curve (AUC).

To assess the Stromal and Immune Cells in Malignant Tumors (ESTIMATE) score, we utilized the “estimate” R package. Additionally, we explored associations between immune cell infiltration and gene expression patterns across various cancer types, we employed several computational deconvolution methods to analyze the correlation between these biological parameters (TIMER, xCell, MCP-counter, EPIC and CIBERSORT). Bubble plots generated through ggplot2 (v3.5.1) visualized associations between ACACA expression patterns and immune cell infiltration potential, with statistical significance defined by Benjamini-Hochberg adjusted p-values. TIDE (Tumor Immune Dysfunction and Exclusion) score is an algorithm for predicting responses to immunotherapy by analyzing tumor gene expression data, which calculating a comprehensive score by evaluating the two major mechanisms of tumor immune escape.

The scRNA-seq dataset was retrived from the Gene Expression Omnibus (GEO) under accession number GSE131907 (http://www.ncbi.nlm.nih.gov/geo). Tumor and matched normal samples were subjected to computational analysis via Seurat (v5.0) within R. The expression matrices were first normalized with the “NormalizeData” function, followed by identification of various features with the “FindVariableFeatures” function, and the “ScaleData” function was used for data scaling. Subsequently. Principal component analysis (PCA) and cell clustering were performed. Nonlinear manifold embedding was visualized through uniform manifold approximation and projection (UMAP) topology. Differential gene expression profiling across clusters was executed via the FindAllMarkers function employing a Wilcoxon rank-sum test framework. Cellular annotation leverages canonical lineage markers curated from peer-reviewed ontologies (Cell Marker database v2.0) and references carcinogenesis literature. Intercellular communication networks were deconvoluted using CellChat (v2.1.2) with the human ligand-receptor interaction repository (CellChatDB.human), which quantifies autocrine/paracrine signaling modalities including secreted factors, extracellular matrix interactions, and direct membrane contact pathways.

Information regarding drug sensitivity and gene expression was obtained from the Genomics of Drug Sensitivity in Cancer database (GDSC, https://www.cancerrxgene.org/). Drugs with an FDR < 0.05 were deemed statistically significant. Bubble plots were created via the R package ggplot2 to display associations between ACACA expression, drug half-maximal inhibitory concentrations (IC₅₀), and their FDR values(v3.5.1).

We collected 41 paraffin-embedded sections of tumor tissues and paired normal tissues from the Cancer Hospital of the Chinese Academy of Medical Sciences. These patients were diagnosed with LUAD and underwent surgical resection during 2015 and 2016. The project obtained approval from the Ethics Committee of the Cancer Hospital of the Chinese Academy of Medical Sciences and acquired patients’ informed consent. These paraffin-embedded samples were stained with anti-ACC1 antibody (1:200 dilution; Cell Signaling Technology, #3676). Immunohistochemistry staining was performed as previously described. We evaluated tissue protein expression levels using the Histochemistry score (H-score), specifically calculated as: ​​H-score = Σ(pi × i)​, i​​ represents the staining intensity grade: 0 (Negative), 1 (Weak positive), 2 (Moderate positive), 3 (Strong positive), and pi denotes the percentage of positively stained cells within each intensity category.

The experimental design included four human lung adenocarcinoma cell models (PC9, HCC827, A549, and H1299), with Beas-2B bronchial epithelial cells used as non-malignant controls. The osimertinib-resistant cells PC9OR and HCC827OR were established by the method of increasing the drug concentration step by step. The IC₅₀ values of Osimertinib in the parental cells versus resistant cells were quantified using the CCK-8 assay ( Supplementary Figure 5E ). Gene-specific siRNA duplexes (Shanghai GenePharma Co.) targeting ACACA were transfected via jetPRIME^® reagent following the manufacturer’s reverse transfection protocol. The siACACA siRNA sequences (5’-3’) were as follows: siRNA#1 (GCAGCUAUGUUCAGAGAAUTT), siRNA#2 (GCUCAUACACUUCUGAAUATT). Both preliminary and subsequent functional validation experiments demonstrated no significant difference in ACACA silencing efficiency between the two ACACA siRNAs.

Cellular samples were cultured via confocal imaging disher for 24 hours under standard growth conditions. After immobilization, permeabilization, and blocking, primary antibody incubation was performed with rabbit monoclonal anti-ACC1 (1:200 dilution; Cell Signaling Technology, #3676) and mouse anti-α-tubulin (1:200 dilution; Sigma-Aldrich, T6074) at 4°C for 16 hours. After washes with PBS-T (0.1% Tween-20), the samples were exposed to species-matched secondary antibodies conjugated to Alexa Fluor 488 (anti-rabbit) and Alexa Fluor 594 (anti-mouse) for 2 hours. Nuclear counterstaining employed DAPI for 5 min before mounting. Confocal imaging was performed, and images were captured.

Total RNA was isolated from the cellular samples using a DNA/RNA extraction kit (RK30153, ABclonal Biotechnology Co., Ltd., Wuhan, China) following the manufacturer’s protocol. Next, the RNA was reverse transcribed into complementary DNA (cDNA) using the ABScript II cDNA First-Strand Synthesis Kit (Takara).Then, the 7500 real-time PCR system from Applied Biosystems was employed for qPCR analysis, utilizing the SYBR Premix Ex Taq kit manufactured by Takara. Chemically synthesized primers were obtained from Generay (Shanghai, China), and their sequences are presented below. The forward primer for ACACA (5’->3’): is AGGAGCTGTCTATTCGGGGT, and the reverse primer (5’->3’) is GGTCGCTCAGCCTGTACTTT. The ACTB forward primer(5’->3’) is CTCGCCTTTGCCGATCC, and the reverse primer (5’->3’) is ATCCTTCTGACCCATGCCC.

Using the CCK-8 kit (RM02823, ABclonal Biotechnology Co., Ltd., Wuhan, China). In accordance with the experimental protocol, the cells were digested, resuspended, counted before and then plated into 96-well plates. At predefined time points (0, 24, 48, 72, and 96 hours), each well received 10 μl of CCK-8 solution. Following 2-hour incubation, absorbance at 450 nm was measured using a microplate reader. The average absorbance value of 5 wells was calculated, and each assay was replicated three times. Sterile PBS blanks and cell-free medium controls were included for background correction.

Metastatic potential was quantified through modified Boyden chamber assays using Corning BioCoat chambers (8-μm pore, #3422). In the migration experiments, a total of 5×10³ cells were plated into the upper compartment containning serum-free medium, while 600 μL of complete medium was added to the lower compartment. Prior to cell seeding in the invasion assay, 50 μL of diluted BD Matrigel matrix was applied to the upper chamber membrane. After the appropriate incubation period, the cells were incubated with 4% paraformaldehyde, and the fixed cells underwent coloration with 0.1% crystal violet solution. Migration patterns were observed, and image were acquired through microscopic examination. The Cells in five random fields per chamber were counted.

A protease inhibitor cocktail was added into the RIPA buffer. Following lysis, total protein concentration was quantified using the Pierce BCA kit (Thermo Scientific, #23227). Proteins were separated by 8% SDS-PAGE. All other steps are carried out in accordance with the standard instructions. The primary antibodies included rabbit monoclonal anti-ACC1 (Cell Signaling Technology, #3676), mouse anti-α-tubulin (Sigma, T6074).

To evaluate cell proliferation, the BeyoClick™ EdU Cell Proliferation Kit (Beyotime) was employed. The cells were cultured for 24 hours, followed by incubation with EdU labeling (10 μm) medium at 37°C. After fixation, permeabilization, and the Click reaction, nuclei were stained with DAPI. Fluorescence microscopy was employed for image capture, and the proportion of EdU-positive cell nuclei was quantified via ImageJ software to further calculate the cell proliferation rate.

Intracellular lipid content was assessed using a lipid fluorescent staining kit (Nile Red Method) (Solarbio, China) in strict accordance with the manufacturer’s protocol. Briefly, the culture medium was aspirated from the cells, which were then rinsed twice with PBS. Subsequently, the cells were fixed with a fixation solution for 10 minutes, followed by an additional PBS rinse. The fixed cells were incubated with the staining solution for 15 minutes under light-protected conditions. After 2–3 rounds of PBS washing, the cells were visualized and imaged using a fluorescence microscope. The fluorescence intensity of the stained cells was further quantified and analyzed using Image J software.

Data analysis was performed using R version 4.3.1 and GraphPad Prism 10 software. For intergroup comparative analyses, either unpaired or paired Student’s t-tests were applied, with the selection based on the intrinsic data structure. Correlation assessments were executed through Spearman’s rank correlation coefficient (for non-parametric data) or Pearson’s correlation coefficient (for parametric data). Survival analyses were implemented via Kaplan-Meier Kaplan-Meier curves with log-rank tests and multivariate Cox regression. All experiments were performed in triplicate, with quantitative data presented as mean ± standard deviation (SD). The threshold for statistical significance was defined as two-tailed P < 0.05.

We first conducted a systematic analysis using the CCLE database, and found significant upregulation of ACACA in prostate, lung, stomach, pancreas and liver cell lines, which suggested possible aberrant lipid biosynthesis in these tumors ( Figure 1A ). Subsequent validation in TCGA data also demonstrated significant overexpressions across multiple tumors compared with matched normal controls ( Figure 1B ). Integrated RNA sequencing analysis of TCGA cancer specimens and Genotype-Tissue Expression (GTEx) normal controls confirmed consistent upregulation of ACACA in diverse tumors ( Figure 1C ). Specifically, elevated ACACA expression pattern was observed in hepatocellular carcinoma (LIHC), stomach adenocarcinoma (STAD), pancreatic ductal adenocarcinoma (PDAC), prostate adenocarcinoma (PAAD), head and neck squamous cell carcinoma (HNSC), non-small cell lung cancer (NSCLC), esophageal carcinoma (ESCA), and cholangiocarcinoma (CHOL). Utilizing transcriptomic and proteomic data from the TCGA and THPA datasets coupled with detailed clinical features, we systematically profiled ACACA expressions across diverse TNM stages ( Supplementary Figures 1A–C ). In LUAD, ACACA mRNA expression was significantly elevated in the M1-stage tumors, indicating its potential involvement in tumor metastasis.

The expression and catalytic activity of ACACA exert direct regulatory control over cellular lipogenesis, which is modulated through multilevel regulation involving transcriptional, post-translational, and metabolic feedback mechanisms. Leveraging proteogenomic data from CPTAC, we identified significant oncogenic dysregulation of ACC1 protein expression across diverse malignancies. Specifically, compared with the corresponding normal controls, quantitative proteomic profiling revealed a substantial increase in protein levels in lung adenocarcinoma, uterine cancer, kidney cancer, colon cancer, and pancreatic ductal adenocarcinoma patients. Conversely, significant downregulation was observed in brain cancer, breast cancer, and ovarian cancer ( Figure 1D ).

To explore the role of ACACA expression in prognostic prediction, Cox regression analyses were carried out to assess its association with different survival endpoints. Integrated analysis revealed that ACACA transcript levels were significantly correlated with adverse prognostic indices, with elevated hazard ratios (HRs) for OS, DSS, DFI, and PFI in several tumor types, including LIHC, adrenocortical carcinoma (ACC), mesothelioma (MESO), and uveal melanoma (UVM) ( Figure 1C ). Subsequently, Kaplan-Meier (KM) survival analyses were performed utilizing prognostic data from the TCGA database to visualize survival trends. KM analyses revealed that elevated ACACA expression was significantly associated with poorer OS in patients with LIHC, HNSC, ACC, ovarian serous cystadenocarcinoma (OV), sarcoma (SARC), kidney renal clear cell carcinoma (KIRC), cervical squamous cell carcinoma (CESC), kidney chromophobe (KICH) and thyroid carcinoma (THCA) ( Figure 2A ).

Interestingly, in both colon adenocarcinoma (COAD) and lower-grade glioma (LGG) patients ( Supplementary Figures 2A,B ), patients with high ACACA expression demonstrated better OS, suggesting a potential context-dependent role for ACACA in different cancer types. K-M curves derived from CPTAC clinical proteomic datasets demonstrated significantly worse overall survival in patients with elevated ACC1 protein levels, particularly in those with kidney cancer, liver cancer, uterine cancer, and lung adenocarcinoma ( Supplementary Figures 2C–F ).

Integrative analysis of TCGA transcriptomic data revealed the diagnostic potential of ACACA across malignancies, which was validated through receiver operating characteristic (ROC) curve analysis ( Figure 2B ). ACACA demonstrated superior discriminatory capacity in LIHC (AUC = 0.891, 95%CI: 0.846-0.921), COAD (AUC = 0.887, 95%CI: 0.828-0.931), CESC (AUC = 0.873, 95%CI: 0.833-0.886) and CHOL (AUC = 0.918, 95%CI: 0.766-0.948). Additionally, significant tumor discrimination potential was observed in HNSC (AUC = 0.748, 95% CI: 0.671-0.810), KIRC (AUC = 0.777, 95% CI: 0.715-0.839), PRAD (AUC = 0.794, 95% CI: 0.742-0.841), and STAD (AUC = 0.799, 95% CI: 0.715-0.870).

The tumor microenvironment, composed of stromal elements and immune cells, occupies a central position in governing tumor progression and facilitating immune evasion. Pan-cancer analysis across 33 distinct tumor types, employing ESTIMATE algorithm-derived metrics, revealed a consistently negative association between ACACA expression and TME characteristics. Specifically, ACACA expression was significantly negatively correlated with stromal, immune and ESTIMATE scores ( Figure 3A ), suggesting that elevated ACACA expression may facilitate the development of an immunosuppressive TME.

Additionally, comprehensive multi-algorithmic analysis (TIMER 2.0, xCell, MCPCOUNTER, EPIC, CIBERSORT) revealed a dual role of ACACA in shaping the tumor immune landscape across malignancies. Using the TIMER 2.0 database, we carried out a Spearman correlation analysis to link the ACACA transcriptional profiles to the immune cell infiltration data ( Figure 3B ). Systematic interrogation of tumor immune landscapes across 33 malignancies revealed that ACACA expression showed significant inverse associations with the infiltration of CD8⁺ cytotoxic T cells and activated NK cells, whereas it positively correlated with resting memory CD4⁺ T cells and M2 macrophage abundance. These associations are particularly apparent in breast cancer, renal cancer, lung adenocarcinoma, and prostate cancer. Further analysis using the xCell algorithm confirmed a significant negative relationship between ACACA expression levels and the infiltration of CD8⁺ T cells ( Figure 3C ). EPIC and MCPCOUNTER analyses displayed a significant positive correlation between elevated ACACA mRNA expression and increased numbers of endothelial cells, neutrophils, and CD4⁺ T cells ( Figure 3C ). These results indicated that ACACA may modulate the populations of diverse immune and stromal cells within the tumor microenvironment, possibly fostering conditions that support tumor development.

Immune checkpoints are pivotal in governing tumor immune evasion, and therapies targeting these axes have markedly transformed cancer therapy. In order to elucidate the interplay between ACACA-driven lipid metabolism and immune checkpoint regulation, we performed correlation analyses across 33 cancer types. Strikingly, ACACA expression exhibited significant negative correlations with the immune checkpoint-related genes (PDCD1, LAG3, LAGLS9, IDO1, CD244, CTLA4 and TIGIT), while showing positive associations with other genes (TGFBR1, KDR, IL10RB, CD160 and CD274) across most cancer types ( Figure 3D ). Furthermore, the results of TIDE algorithm demonstrated that increased ACACA expression was related to higher TIDE scores in LGG, THCA, SARC, cholangiocarcinoma (CHOL), glioblastoma (GBM) and breast invasive carcinoma (BRCA), which suggested impaired cytotoxic T-cell infiltration and increased immune evasion ( Figure 3E ). In contrast, inverse correlations were observed for UVM, testicular germ cell tumor (TGCT) and diffuse large B-cell lymphoma (DLBC), suggesting tumor-specific regulatory mechanisms. Validation in independent immunotherapy cohorts (GSE126044 and GSE91061) further demonstrated that non-responders (SD/PD) exhibited significantly higher ACACA levels than responders (PR/CR) ( Figures 3F, G ). KM survival analysis of the GSE91061 cohort indicated that the patients with high ACACA expression patients had an unfavorable prognosis ( Figures 3H, I ). These findings identify ACACA as a metabolic orchestrator of immune checkpoint networks and propose its utility as a predictive biomarker for immunotherapy resistance, thereby providing new mechanistic insights into the metabolic-immune interplay in cancer.

To delineate the oncogenic networks modulated by ACACA, we performed GSEA on transcriptomic datasets spanning 33 TCGA cancer types, by stratifying patients into high (top 25%) and low (bottom 25%) ACACA expression cohorts. Hallmark pathway analysis revealed striking differential enrichment patterns, the high ACACA group exhibited significant activation of cell cycle regulatory pathways, including mitotic spindle formation, the G2/M checkpoint, E2F transcriptional targets, and MYC-driven signaling pathway ( Figure 4A ). These findings suggest that ACACA may enhance tumor cell proliferation by promoting cell cycle progression and DNA replication. Intriguingly, high ACACA expression was inversely correlated with key immunomodulatory pathways, as evidenced by the suppression of interferon-α responses, the inflammatory signaling cascades, and complement activation, particularly in LGG, SARC, PAAD, and lung squamous cell carcinoma (LUSC). These findings imply that ACACA might facilitate tumor immune evasion by attenuating critical antitumor immune responses.

The GDSC database provides data on the sensitivity of a broad range of antineoplastic drugs across various cancer cell lines and is widely used for drug-gene association analysis. We next examined the associations between ACACA expression levels and dug the IC₅₀ values ( Figure 4B ). The efficacy of drugs such as TGX221, CHIR-99021, temsirolimus, and dabrafenib was decreased in patients with higher ACACA expression. In contrast, drugs (e.g., mitomycin C, thapsigargin, MLN4924, and epothilone B) showed a negative correlation with ACACA expression, suggesting that in pancancer analysis, increased ACACA gene expression may contribute to drug sensitivity.

We subsequently conducted an in-depth analysis of the mRNA and protein expression profiles retrieved from TCGA lung cancer database. Our findings demonstrated that lung cancer patients with elevated ACACA mRNA expression had a worse prognosis ( Figure 5A ). Concordantly, KM survival analysis of proteomic data from the TCPA database revealed a significant correlation between high ACC1 protein levels and poorer OS ( Figure 5B ). Interestingly, analysis of ACC1-S79 phosphorylation status suggested that patients with elevated ACC1-S79 expression levels experienced an better prognosis ( Figure 5C ), which is congruent with previous reports indicating that phosphorylation of ACC1 at serine 79 inhibits its enzymatic activity (18, 19). Immunohistochemical analysis via the Human Protein Atlas (HPA) database confirmed significant ACC1 protein upregulation in LUAD tumor tissues ( Figures 5D, E ).

Single-sample gene set enrichment analysis (ssGSEA) of TCGA-LUAD datasets validated significant associations between ACACA expression and pathway activity scores ( Figure 5F ). Specifically, ACACA upregulation was positively linked to the activation of pathways related to fatty acid synthesis, tumor proliferation, TGF-β signaling, and the PI3K-AKT-mTOR pathway. Conversely, it was negatively correlated with oxidative phosphorylation, apoptosis, and immune-related pathways like the p53 pathway, the tumor inflammatory signature, and the IL-10 anti-inflammatory signaling pathway. The results indicate that ACACA may drive proliferation and survival of tumor cells in lung adenocarcinoma by modulating multiple oncogenic and metabolic pathways while inhibiting apoptosis and immune responses. KEGG and GO analyses were conducted to examine biological pathways linked to ACACA-related DEGs ( Figure 5G ), the results revealed that ACACA expression had associations with key pathways, including fatty acid metabolism, intercellular tight junctions, protein digestion, cytokine-receptor interactions, and the AMPK signaling pathway. Collectively, these findings imply that ACACA potentially has a multifaceted role in orchestrating energy balance, modulating immune responses, and driving cancer progression.

Through a systematic evaluation of the relationship between common driver gene mutations in lung cancer and the expression level of ACACA, we observed frequent upregulation of ACACA expression in patients harboring TP53 mutations or EGFR mutations ( Supplementary Figure 3 ). To further elucidate the therapeutic implications of ACACA expression, we analyzed drug sensitivity using IC₅₀ value as a quantitative pharmacodynamic parameters. Our analysis demonstrated that the high ACACA expression group exhibited markedly reduced IC₅₀ values (p < 0.05) for both chemotherapeutic agents (docetaxel, gemcitabine, vinorelbine) and the ALK-inhibitor crizotinib when compared with the low expression group, indicating enhanced therapeutic sensitivity to these treatments ( Figure 5H ). Interestingly, by integrating mRNA expression data of EGFR-mutated lung cancer patients from TCGA and the drug sensitivity data in the GDSC database, we found that for small molecule EGFR-TKI drugs (e.g., gefitinib, erlotinib), high ACACA expression in the EGFR-mutated group was associated with increased IC₅₀ value. These findings suggest that ACACA-mediated metabolic reprogramming may contribute to the development of resistance to EGFR-TKI therapy in LUAD.

To characterize the role of ACACA in the LUAD TME, we integrated single-cell transcriptomic datasets from normal lung (nLung), early-stage (tLung), and advanced tumor (tL/B) tissues derived from the GSE131907 dataset (20) ( Figure 6A ). Following rigorous quality control and batch correction, unsupervised clustering analysis identified seven major cell types: B cells, T/NK cells, myeloid cells, mast cells, fibroblasts, endothelial cells, and epithelial cells ( Figure 6B ), with annotation based on known marker genes from previous article ( Supplementary Figure 4D ). ACACA expression was predominantly localized to epithelial cells, with comparable expression levels across tumor and normal tissues ( Figure 6C ). In advanced tumors (tL/B), subclustering analysis of the epithelial cell compartment was carried out, which further stratified the epithelial cells into ACACA-high and ACACA-low subsets ( Figures 6D–F ). Subsequent CellChat analysis revealed enhanced interactions between ACACA-high tumor cells and CD8⁺ T cells, primarily via secreted signaling pathways like the macrophage migration inhibitory factor (MIF) axis (MIF-CD74+CXCR4, MIF-(CD274+CD44)) ( Figure 6G ). The LAMB3-DAG1 pathway is dominant in ECM-receptor interactions ( Figure 6H ), and the APP (APP-CD74) and CD99 (CD99-CD99) pathways were the major contributors to cell-cell contact pathways. These findings collectively suggest that ACACA may orchestrate fatty acid metabolic reprogramming to potentiate tumor-immune crosstalk through specific signaling networks, thereby fostering the establishment of an immunosuppressive niche and metastatic progression ( Supplementary Figure 4 ).

Immunohistochemical analysis of ACC1 protein expression was performed on paraffin-embedded sections from 41 LUAD patients ( Figure 7A ). The expression of ACC1 in tumor tissues was significantly elevated compared to the adjacent normal lung tissues ( Figure 7B ). Using the median H-score as a cutoff, patients were divided into high (n =28) and low (n=13) ACACA expression groups. Kaplan Meier (K-M) survival analysis indicated that high levels of ACC1 suffered poorer OS ( Figure 7C ). And then, immunofluorescence staining of LUAD cell lines (HCC827 and PC9) revealed that ACC1 protein is predominantly localized within the cytoplasm ( Figure 7D ), which is consistent with its central role in fatty acid metabolism. Bioinformatic analyses revealed the significant influence of ACACA on cell proliferation across a variety of malignant tumors. For a deeper investigation of its biological function in LUAD, ACC1 protein levels were quantified in lung cancer cell lines, revealing significantly higher protein expression in most LUAD cell lines relative to normal bronchial epithelial cells ( Figure 7E ).

Then we employed RT-qPCR and western blotting to validate the silencing efficiency of siRNAs targeting of ACACA ( Figure 8A ). The results of the CCK-8 assays and EdU assays showed that ACACA silencing decreased the proliferative ability of HCC827 and PC9 cells ( Figures 8B, C ). Additionally, cells transfected with ACACA siRNA notably impaired migration and invasion capacities ( Figure 8D ). Nile red staining further demonstrated that cellular lipids are significantly lowered by ACACA silencing, compared with control cells ( Figure 8E ). Osimertinib-resistant cells (HCC827OR and PC9OR) were established by exposing parental cells to gradually escalating doses of osimertinib over an extended period. CCK8 assays showed that the IC₅₀ values of osimertinib in HCC827OR and PC9OR were 10.87 μm and 5.509 μm, respectively, both significantly higher than those in sensitive cells ( Supplementary Figure 5E ). Then we found that ACACA expression was upregulated in Osimertinib-resistant cells through qPCR and WB ( Figure 8F ). Using ACACA-specific siRNA markedly inhibit cell proliferation of HCC827 OR and PC9 OR cells ( Figures 8G, H ), and might increase the sensitivity of drug-resistant cells to Osimertinib ( Figure 8I ). These findings suggest that ACACA may be a key factor regulating the biological behavior of tumor cells, and a similar effect of ACACA silencing effect were observed in sarcoma cells. According to the CCK8, transwell and Nile Red staining assays, ACACA notably increased the proliferation, migration and lipid synthesis capabilities of MG63 and U2OS cells ( Supplementary Figures 5A–D ). Inhibiting ACACA effectively suppressed fatty acid synthesis and tumor growth.