Cell culture

We sourced the mouse lewis lung cancer (LLC) cell line and the human embryonic kidney cell line HEK293T from the Cell Bank of the Chinese Academy of Sciences located in Shanghai, China. These cells were cultured at 37 °C with 5% CO2 in DMEM medium supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 0.1 mg/mL streptomycin. We also obtained the human NSCLC cell lines H1299 and A549, the human monocytic cell line THP-1, and the mouse macrophage RAW264.7 cell line from ATCC and cultured them according to recommended guidelines. We regularly conducted mycoplasma contamination tests on all cell lines used in this study, and we are pleased to report that all results were negative. The cell lines underwent authentication through cellular morphology and short tandem repeat analysis conducted by Microread Inc. in Beijing, China.

Mouse CD8+ T cells were isolated from spleens using the EasySep™ Mouse CD8+ T Cell Isolation Kit (Stemcell Technologies, Cambridge, MA). These isolated T cells were then stimulated with anti-CD3 (Clone 145-2C11, BD Biosciences, San Jose, CA) and anti-CD28 (Clone 37.51, BD Biosciences) antibodies for a duration of 2 days. Primary bone marrow-derived macrophages (BMDMs) were prepared from mice as previously described. To induce polarization of macrophages into the M1 subset, BMDMs, RAW264.7 cells, or THP-1 cells (pretreated with 200 ng/mL PMA) were exposed to 100 ng/mL LPS and 20 ng/mL IFNγ. For polarization into the M2 subset, BMDMs, RAW264.7 cells, or THP-1 cells (pretreated with 200 ng/mL PMA) were incubated with 20 ng/mL IL4 and 20 ng/mL IL-13. Supplementary Table S1 provides a summary of the key resources used in this process.

Animals

The Sftpc-creERT2; KrasLSL-G12D/+; Cpt1aFlox/Flox transgenic mice were used in this study to specifically knockout the Cpt1a gene in type II alveolar epithelial cells of the lung. This was achieved through the Cre-FloxP system, where loxP fragments were inserted using the CRISPR system and homology-directed repair. The deletion of exon 9 resulted in a frame-shift mutation, leading to the inactivation of the Cpt1a gene. To initiate tumor formation in these mice, Tamoxifen was administered twice a week for 2 months via intraperitoneal injection. Tamoxifen was dissolved in corn oil at a concentration of 20 mg/ml, and each injection contained 75 mg/kg of body weight. This treatment activated the oncogenic KrasG12D and facilitated the conditional deletion of Cpt1a.

In this study, male C57BL/6J or NOD/SCID mice aged 6-8 weeks were used for the LLC tumor graft model. The mice were randomly assigned to each group with a minimum of 5 mice per group. To form subcutaneous tumors, 1 × 106 LLC cells expressing shCtrl, shCPT1A#1, and shCPT1A#2 separately were injected subcutaneously into each mouse. Tumor volume was calculated using the formula: (width2 × length)/2, and tumors were analyzed histologically, immunologically, and for survival at different time points. In order to test the efficacy of anti-PD-1 antibody treatment, BioXcell’s mouse anti-PD-1 antibody or mouse anti-IgG2a antibody was administered via intraperitoneal injections at a dosage of 100 μg/injection on days 7, 10, 13, 16, and 19 after tumor inoculation. When the tumors reached a size of approximately 80 mm3, the mice were randomly divided into four groups for erastin treatment. Mice were intraperitoneally injected with 20 mg/kg erastin every day for 15 days. For L-carnitine treatment, once the tumors reached about 80 mm3, the mice were orally treated with L-carnitine at a dosage of 100 mg/kg per day (Sigma-Aldrich) or saline (200 μL) for 21 days (n = 6-7). Animal studies were conducted in accordance with institutional guidelines and with the authorization of the Experimental Animal Welfare and Ethics Committee of Basic Medical Sciences Institute of Chinese Academy of Medical Sciences (ACUC-A02-2022-114).

Stable cell strain establishment

To achieve stable knockdown of mouse Cpt1a in LLC cells or homo CPT1A in H1299/A549 cells, we utilized the pLV2N-U6-Puro vector (GenePharma Co., Ltd., Shanghai, China) to insert control or CPT1A-targeting shRNA templates. These shRNA templates were named shNC, shCPT1A#1, and shCPT1A#2. Lentiviruses expressing these shRNAs were used to infect tumor cells, and subsequently, clone selection was performed using puromycin at a concentration of 2 μg/mL. This process allowed us to establish stable cell lines with CPT1A knockdown. Supplementary Table S2 provides the specific sequences of the shRNAs.

Tumor sphere formation assay

After preparing single cell suspensions, the cells were enumerated and seeded onto 6-well ultra-low attachment culture plates (Corning, USA) at an optimal density. These plates were filled with complete Mammocult medium (Stem Cell Technologies, Canada). Following a 7-day incubation period, the tumorspheres with a diameter of at least 70 μm were quantified.

ATP production measurement

To determine the level of ATP production, we utilized the cell titer-blue cell viability assay kit (Promega, USA) in accordance with the manufacturer’s instructions. The resulting values were normalized to the protein concentration.

RNAi/overexpression experiment

c-Myc siRNA was transiently transfected using Lipo2000 according to the manufacturer’s instructions. The siRNAs were synthesized by Beijing Qingke Biological Company. pCDNA3.1-3FLAG-MYC plasmid (Tsingke Biotechnology Co., Ltd.) was transiently transfected using Lipo2000 according to the manufacturer’s instructions.

Quantitative PCR analysis

The Trizol reagent (Invitrogen) was employed for the extraction of total RNA from tissue or cells. The concentration of RNA was determined using a NanoDrop spectrometer. Subsequently, 1000 ng of total RNA was subjected to reverse transcription using the Prime Script® RT reagent Kit Perfect Real Time kit (TaKaRa) to generate cDNA. For RNA-time PCR analysis, SYBR-Green fluorescent dye (Biorad) was utilized, and the reactions were performed on a Biorad CFX 96 instrument. The specific primer sequences used can be found in supplementary Table S2.

Histology and immunohistochemistry

Tissue chips containing 75 pairs of lung adenocarcinoma and para-cancerous tissues or 80 lung adenocarcinoma tissues were purchased from SHANGHAI OUTDO BIOTEC CO., LTD (LUC1505, LUC1601). Lung adenocarcinoma tissues of two NSCLC patients were obtained from Tianjin Medical University General Hospital. Tumor tissue from patients or mice was fixed in 4% paraformaldehyde. Tissues were embedded with paraffin and sectioned by microtome. The slides were stained with hematoxylin and eosin (H&E) using standard protocol. For CPT1A or c-Myc immunohistochemistry, slides of various tissue were blocked with goat serum for 1 h. Subsequently, the slides were incubated with anti-CPT1A (1:200; 15184-1-AP, Proteintech) anti-c-Myc (1:200; 10828-1-AP, Proteintech), anti-FBXW7 (1:200; 28424-1-AP, Proteintech), anti-GPX4 (1:200; ab125066, Abcam) or anti-ACSL4 (1:200; ab155282, Proteintech) overnight at 4 °C followed by detection with the microscope (Leica). Hematoxylin (ZSGB-BIO) was used as counterstain. The human samples were obtained with informed consent, and the study was approved by Ethics Committee of Basic Medical Sciences Institute of Chinese Academy of Medical Sciences (2019029). The study was conducted in accordance with recognized ethical guidelines. The clinicopathological characteristics of NSCLC patients are listed in Supplementary Tables S3 and S4.

Immunofluorescence (IF)

Immunofluorescence (IF) was performed following a previous protocol47 and using primary antibodies raised against CPT1A (1:50; 15184-1-AP, Proteintech) or c-Myc (1:50; 10828-1-AP, Proteintech).

Western bolt

The whole cell lysates were prepared following established protocols. In summary, cells were lysed using M-PER Mammalian Protein Extraction Reagent (Thermo Scientific Technologies, USA) supplemented with a protease inhibitor cocktail (Thermo Scientific Technologies, USA). The resulting lysates contained total protein. The protein concentration in the lysates was measured using a Bradford assay kit (Beyotime Biotechnology, Shanghai). Antibodies for CPT1A, c-Myc, SOX2, OCT4, NANOG, GPX4, ACSL4, NRF2, Ubiqutin, FBXW7, GAPDH and β-actin (Cell Signaling Technology; 1:1000), were used for western bolt.

Flow cytometry

To detect cells with high ALDEFLUOR activity, we conducted the ALDEFLUOR assay (Stemcell Technologies) in accordance with the manufacturer’s instructions. The assay involved incubating cells in the presence or absence of the ALDH inhibitor diethylaminobenzaldehyde for 40 min at 37 °C. Following this, single cells were exposed to CD44 and other primary antibodies in PBS with 1% FBS for 30 min at 4 °C. Finally, cells were stained with 7-AAD, which was used to exclude nonviable cells.

BODIPY-C11 staining

For BODIPY-C11 staining, cells were suspended in 1 mL of Hanks Balanced Salt Solution (HBSS, Gibco 14-025-092) containing either 3 mM BODIPYTM 581/591 C11 or BODIPYTM 665/676. The suspension was then incubated for 30 min at 37 °C in a tissue culture incubator. After incubation, the cells were washed and resuspended in 200 mL of fresh HBSS. Subsequently, the stained cells were immediately analyzed using a flow cytometer (LSR II, BD Biosciences). In the case of BODIPYTM 581/591 C11 staining, both non-oxidized C11 (PE channel) and oxidized C11 (FITC channel) signals were monitored. The ratio of the mean fluorescence intensity (MFI) of FITC to the MFI of PE was calculated for each sample. To account for variations, the data were normalized to control samples, as indicated by relative lipid ROS levels.

Chromatin immunoprecipitation (ChIP)

ChIP was carried out using the EZ-Zyme Chromatin Prep Kit (Millipore, Billerica, MA, USA) according to the manufacturer’s protocol. Anti-c-Myc antibody was used to precipitate DNA crosslinked with c-Myc, and anti-mouse IgG was also used as a negative control. qPCR was performed to detect DNA fragments of the CPT1A promoter region. The primers used are listed in Supplementary Table S2.

Luciferase reporter assay

We co-transfected 0.5 mg of pGL4.10 vector expressing CPT1A-BS1 or CPT1A-BS2 (or the indicated mutant), along with 0.5 mg of pRL-TK plasmid (as a reference) and 50 ng of Renilla luciferase reporter into H1299 cells using Lipofectamine 2000. This transfection was performed in triplicate. Luciferase activity was measured 24 hours later using the Dual-Luciferase Reporter Assay System from Promega. Firefly luciferase activities were normalized to Renilla luciferase control values and presented as an average of the triplicates.

Seahorse XF24 respirometry

XF24 plates were seeded with 2 × 104 per well of H1299 shNC / shCPT1A sphere cells and allowed to stabilize overnight. The extracellular acidification rate (ECAR) was measured using the XF24 extracellular flux analyzer, utilizing the glucose stress fuel flex test kits from Agilent. The ECAR measurements were carried out following the manufacturer’s instructions. Subsequently, the data obtained were analyzed using Wave software provided by Seahorse/Agilent.

Transmission electron microscopy (TEM)

The samples were prepared as follows: first, the samples were embedded in 10% gelatin and fixed in glutaraldehyde at 4 °C; subsequently, the fixed samples were sectioned into blocks of less than 1 mm in size. The blocks were then dehydrated using a series of alcohol concentrations (30%, 50%, 70%, 90%, 95%, and 100%) for 10 min each. Following dehydration, the samples were infiltrated with Quetol-812 epoxy resin mixed with propylene oxide at increasing concentrations (25%, 50%, 75%, and 100%) for 3 h each step. Finally, the samples were embedded in pure, fresh Quetol-812 epoxy resin and polymerized at different temperatures (35 °C for 12 h, 45 °C for 12 h, and 60 °C for 24 h). Ultrathin sections (100 nm) were obtained using a Leica UC6 ultramicrotome and stained with uranyl acetate for 10 min and lead citrate for 5 min at room temperature. The samples were observed using an FEI Tecnai T20 TEM transmission electron microscope.

Quantification of L-carnitine or NADPH

For analysis of intracellular nucleotides, cells (1 × 106) or cell suspension were homogenized in 100 μL of Assay Buffer and centrifuged at 13,000 g for 10 min to remove insoluble materials. The resulting sample was brought to 50 μL/well with Assay Buffer in a 96-well plate, and the manufacturer’s instructions (Biovision, K642-100) were then followed. To detect NADP, NADPH, and their ratio, the NADP/NADPH Assay Kit (Abcam, ab65349) was used.

LC-MS analysis of metabolites

H1299-shNC or H1299-shCPT1A cells were cultured in 100 mm dishes until they reached approximately 60–70% density. Metabolites were then extracted from the cells using cold 80% methanol, followed by two washes with PBS. The collected supernatants were spun at 13,300 rpm for 10 min at 4 °C and subsequently subjected to vacuum freeze-drying. The metabolites were then analyzed using a Dionex Ultimate 3000 UPLC system coupled to a TSQ Quantiva Ultra triple-quadrupole mass spectrometer (Thermo Fisher, CA) equipped with a heated electrospray ionization (HESI) probe. Separation of the extracts was performed using a synergi Hydro-RP column (2.0 × 100 mm, 2.5 mm, Phenomenex) with a binary solvent system consisting of 10 mM tributylamine adjusted with 15 mM acetic acid in water (mobile phase A) and methanol (mobile phase B). A 25-min gradient from 5% to 90% mobile phase B was used for the analysis. Metabolites were detected using selected reaction monitoring (SRM) in positive-negative ion switching mode with a resolution of 0.7 FWHM for both Q1 and Q3. The source voltage was set to 3500 v for positive ion mode and 2500 v for negative ion mode. Source parameters included a capillary temperature of 350 °C, heater temperature of 300 °C, sheath gas flow rate of 35, and auxiliary gas flow rate of 10. Metabolite identification and peak integration were performed using Tracefinder 3.2 (Thermo, USA).

Bioinformatics analysis

RNA sequencing data from NSCLC patients who received anti-PD-1/PD-L1 treatment were obtained from the GEO accession GSE135222 and GSE126044. The CIBERSORTx online analysis platform (https://cibersortx.stanford.edu/) provided a means to evaluate the relative proportions of immune cells infiltrating the tumors. To assess mRNA expression, the GDC TCGA Lung Adenocarcinoma (LUAD) database (HiSeqV2) was accessed through the UCSC Xena platform (https://xena.ucsc.edu). Z-scores were then calculated based on all tumor samples for further analysis. Different gene signatures were determined by averaging the z-scores of specific genes. For example, the apoptosis gene signature was derived from the average z-scores of BIM, PUMA, BID, BMF, BAD, HRK, BIK, and NOXA. Necroptosis gene signature was determined by averaging the z-scores of RIPK1, RIPK3, and MLKL. Autophagy gene signature was derived from the average z-scores of MAP1LC3B, BECN1, and SQSTM1, while the ferroptosis gene signature was determined by averaging the z-scores of ALOX5, LPCAT3, ACSL4, TFRC, and SLC11A2. Pearson’s correlation analysis was performed using RStudio. Overall survival (OS) data along with the expression levels of CPT1A, c-Myc, and FBXW7 genes were obtained from the UCSC Xena platform (https://xena.ucsc.edu). Survival curves were generated using the R packages “survival” and “survminer”.

Statistical analysis

The data were analyzed using GraphPad Prism 8.0. Statistical significance was determined by performing an unpaired two-tailed Student’s t-test to compare two groups or by conducting one-way or two-way ANOVA with Tukey’s correction for multiple comparisons when comparing three or more groups. A P value less than 0.05 was considered statistically significant. The survival of mice in the experiments was represented using Kaplan-Meier curves and statistical significance was estimated using a log-rank test.