LPIN2 contributes to tyrosine kinase inhibitor resistance via activation of PI3K pathway

Introduction

Lung cancer is the leading cause of cancer-related mortality worldwide, with an estimated 2.48 million new cases and 1.81 million deaths annually. In China, lung cancer accounts for approximately 1.066 million new cases each year (1,2). Histologically, lung cancer is classified into small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), with NSCLC comprising approximately 85% of all cases (3-5). The 5-year overall survival rate of NSCLC patients remains below 20%, highlighting the urgent need for effective therapeutic strategies (6).

The advent of tyrosine kinase inhibitor (TKI)-targeted therapy has significantly improved survival duration and quality of life for eligible patients. Currently, TKIs targeting driver gene mutations such as epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK), including osimertinib (for EGFR-mutant NSCLC) and alectinib (for ALK-fusion NSCLC) have significantly improved objective response rates and progression-free survival, substantially enhancing patients’ quality of life (7-9). However, acquired resistance remains a major challenge, leading to disease progression in most patients within approximately one year of treatment (7-10). To date, studies on TKI resistance mechanisms have mainly focused on two categories: target-dependent mechanisms (e.g., secondary mutations in EGFR such as T790M and C797S, or ALK kinase domain mutations) and target-independent mechanisms. Among the latter, research has primarily centered on bypass signaling pathway activation [e.g., mesenchymal-epithelial transition factor (MET) or human epidermal growth factor receptor 2 (HER2) amplification], tumor microenvironment (TME) remodeling, cell phenotypic transition (e.g., epithelial-mesenchymal transition, small cell transformation), and metabolic reprogramming (11-14), while metabolic reprogramming has emerged as a key contributor to TKI resistance—with abnormalities in glucose, fatty acid, and amino acid metabolism reported to participate in resistance progression (15-22). Importantly, target-independent resistance mechanisms are a common phenomenon shared across different driver gene mutations (EGFR and ALK) (23,24). The specific role of triglyceride (TG) synthesis, a key branch of metabolic reprogramming, remains unclear. This critical knowledge gap hinders the development of effective strategies to overcome TKI resistance.

Our preliminary analysis confirmed that the glycerophospholipid metabolism pathway, which is closely linked to TG synthesis, is significantly upregulated in TKI-resistant cells, suggesting its potential role in mediating resistance. Glycerophospholipids are core components of cell membranes and play crucial roles in signal transduction, membrane trafficking, and energy metabolism (25-28). Aberrant activation of glycerophospholipid metabolism has been implicated in progression and therapy resistance in multiple malignancies (29-33). Moreover, elevated serum glycerophospholipid levels have been observed in NSCLC patients following EGFR-TKI or immunotherapy, indicating a potential association with treatment adaptation (34). Notably, most previous studies utilized resistant cell lines maintained under long-term high-concentration TKI conditions, where gene expression may be influenced by drug stress responses. In this study, by identifying pathways consistently altered in resistant cells regardless of TKI withdrawal, we revealed that upregulation of the glycerophospholipid metabolism pathway represents a drug stress-independent feature of TKI resistance.

Based on these findings, we hypothesize that the glycerophospholipid metabolic pathway serves as a critical metabolic axis in acquired TKI resistance of NSCLC. Integrated metabolomic and transcriptomic analyses identified significant upregulation of the LPIN2 gene within this pathway (35). LPIN2 encodes the lipid phosphatase LPIN2, which catalyzes the conversion of phosphatidic acid (PA) to diacylglycerol (DAG) and acts as a key rate-limiting enzyme in the glycerophospholipid/triacylglycerol synthesis pathway (36-39). Although LPIN family members are known to play important roles in lipid homeostasis maintenance, endoplasmic reticulum stress response, and inflammation regulation (40-43), the specific function and regulatory mechanism of LPIN2 in tumor resistance, especially its role in NSCLC TKI resistance, remain largely unexplored. This study aims to systematically elucidate the role of LPIN2-mediated glycerophospholipid metabolic reprogramming in NSCLC TKI resistance, thereby providing new perspectives for understanding resistance mechanisms. We present this article in accordance with the MDAR and ARRIVE reporting checklists (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1471/rc).

Methods

Cell lines

PC-9, NCI-H3122, and HEK293T cell lines were procured from the Shanghai Academy of Science (Shanghai, China). PC-9 and NCI-H3122 cells were cultured in Roswell Park Memorial Institute Medium-1640 (RPMI-1640; Gibco, Grand Island, NY, USA), while HEK293T cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Gibco). The complete culture medium contained 10% fetal bovine serum (FBS; Gibco) and 1% penicillin-streptomycin (Gibco). All cell lines were cultured at 37 ℃ in a humidified atmosphere with 5% CO₂.

Antibodies

The following primary antibodies were used: LPIN2 (Santa Cruz, Dallas, TX, USA, sc-514353); ETNK2 (ABclonal, Woburn, MA, USA, A14628); PLPP3 (ABclonal, A15743); PLA2G4C (ABclonal, A7753); ACHE (ABclonal, A2806); Bcl2 (Selleck, Houston, TX, USA, H20N17); Bax (Selleck, L6B9); BCL-xL (Selleck, D4A3); PI3K-α (CST, Danvers, MA, USA, #4255); PI3K-β (CST, #3011); p-AKT (CST, #4060); AKT (CST, #9272); p-mTOR (CST, #5536); mTOR (CST, #2983); Ki67 (Abcam, Cambridge, UK, ab16667); β-Actin (ZSGB-BIO, Beijing, China, TA-09); β-Tubulin (ZSGB-BIO, TA-10).

Reagents

Osimertinib (S7297), alectinib (S2762), wortmannin (S1105), MK-2206 2HCl (S1078), PF-06424439 (S9921), and A 922500 (S2674) were purchased from Selleckchem. Rapamycin (HY-10219), Nile Red (HY-D0718), and 1,2-dioleoyl-3-arachidoylglycerol (AOO) (HY-157739) were obtained from MedChemExpress (Monmouth Junction, NJ, USA). AOO was freshly solubilized in 100% anhydrous DMSO to prepare a 50 mM stock solution. For all in vitro experiments, this stock was further diluted in culture medium to a final concentration of 100 µM (resulting in a final DMSO concentration of 0.2% v/v), ensuring no visible precipitate or crystal formation under microscopic observation.

sgRNA and knockout plasmids

LPIN2 knockout was performed using the CRISPR/Cas9 system. Two sgRNAs targeting distinct exons of the human LPIN2 gene were designed and cloned into the lentiCRISPR v2-puro vector (Addgene). The target sequences are: sgLPIN2-1: ACCCCTTTGGTGAAATCGGG; sgLPIN2-2: AGAAATCAACGGCAGTGCAG.

Animal studies

Mice were housed in facilities accredited by the Animal Care and Use Committee of West China Hospital, Sichuan University. All in vivo animal experiments were performed under a project license (approval No. 20240315033) granted by the Animal Ethics Committee of West China Hospital, Sichuan University, in compliance with institutional animal welfare guidelines for the care and use of animals. The study used six-week-old female nude mice purchased from GemPharmatech Biotechnology Co., Ltd. (Jiangsu, China). Mice were subcutaneously inoculated with 2×10⁶ cells and monitored until visible tumor formation.

For PC-9, PC-9OR, and PC-9ORsgLPIN2 xenograft studies, when tumor volume reached 200 mm³, mice (n=5 per group) were treated daily via intraperitoneal injection with either vehicle or osimertinib (5 mg/kg/day) for 27 days.

For combination drug studies in PC-9 and PC-9OR xenografts, when tumor volume reached 200 mm³, mice (n=5 per group) received daily intraperitoneal injections of vehicle, osimertinib (5 mg/kg, once daily), or osimertinib combined with wortmannin (1 mg/kg, once daily) for 27 days.

Tumor volume was calculated using the formula: V (mm³) = L × W²/2, where L and W represent the longest longitudinal diameter and transverse diameter, respectively.

Western blotting (WB)

Fresh frozen tissue or cell samples were lysed using radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime, Shanghai, China) supplemented with protease and phosphatase inhibitor cocktails (Beyotime). Protein concentrations were measured using a bicinchoninic acid (BCA) Protein Assay Kit (Beyotime). Equal amounts of cell lysates (15–30 µg) were loaded onto polyacrylamide gels of appropriate concentrations containing 0.1% sodium dodecyl ulfate (SDS), followed by transfer to polyvinylidene difluoride (PVDF) membranes (Merck Millipore, Cork, Ireland). The membranes were then blocked at room temperature for approximately 1 hour in 1× phosphate-buffered saline (PBS) containing 0.1% Tween-20 and 5% skim milk, and subsequently incubated overnight at 4 ℃ with specified antibodies diluted in either 1× PBS or commercial antibody dilution buffer (Beyotime). Densitometric quantification of Western blot bands was performed using ImageJ software. For normalization, the densitometry values of total proteins were calculated relative to the controls (β-actin or β-tubulin). To accurately determine the extent of protein phosphorylation, the densitometric signals of phosphorylated proteins were first normalized to the corresponding loading control, and the resulting values were then divided by the normalized relative expression levels of their respective total proteins.

Immunohistochemistry (IHC)

Tumor sections were deparaffinized, rehydrated, and treated with 0.3% hydrogen peroxide in methanol for 20 minutes to inhibit endogenous peroxidase activity. Subsequently, the sections were blocked with goat serum and incubated overnight with corresponding primary antibodies. After washing, the sections were incubated with horseradish peroxidase-conjugated anti-rabbit antibody (Zhongshan Golden Bridge, Beijing, China). Antigen-antibody complexes were detected using an HRP-conjugated secondary antibody and visualized with DAB Substrate Kit (CST, #8059). Sections were counterstained with hematoxylin, dehydrated, and mounted. For the quantitative evaluation of protein expression, the open-source ImageJ plugin IHC Profiler was utilized (44).

Cell viability assay

Cells were seeded in 96-well plates with three replicate wells per condition. After 24 hours, cells were treated with specified compounds at varying concentrations and incubated for 3 additional days. The cells were then stained with Cell Counting Kit-8 (CCK-8) reagent. Following 1 hour of incubation, absorbance was measured at 450 nm using an Epoch2 multi-volume spectrophotometer system. Dose-response curves were generated by normalizing the absorbance signals after blank subtraction, and half-maximal inhibitory concentration (IC₅₀) values were calculated accordingly.

Cell morphology observation and colony formation assay

Cell morphology, adherence, and density were examined under an inverted microscope (10× objective). For the colony formation assay, 3,000 cells were seeded per well in 6-well plates and treated with specified compounds at different concentrations 24 hours later, with control groups cultured in standard medium. The medium was replaced every three days. After 7 days of culture, cells were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet. The crystal violet-stained pictures were analyzed using the software ImageJ following the automated processing protocol previously described (45). Briefly, digital images were thresholded and analyzed to determine the percentage of surface coverage in each well. After extracting the area fraction from ImageJ, all values were normalized against the mean of the untreated wells to account for inter-experimental variations.

Lipid droplet staining (Nile Red)

A 1 mM Nile Red stock solution was prepared in dimethyl sulfoxide (DMSO) and diluted to a 100 nM working solution using serum-free medium or PBS. After washing with PBS, cells were incubated with 1 mL of the working solution in the dark for 30 minutes. After washing, lipid droplet fluorescence signals were then examined under a fluorescence microscope (46). All fluorescence images were acquired using the same microscope under identical settings (e.g., exposure time, laser intensity) to ensure comparable intensity scaling. The mean fluorescence intensity (MFI) was measured using ImageJ software. For each group, data were collected from three independent replicate images. To standardize the data, the MFI of the experimental groups was normalized against that of the wild-type control cells.

Real-time-polymerase chain reaction (RT-PCR) and quantitative polymerase chain reaction (qPCR)

Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA), followed by reverse transcription into first-strand cDNA using the PrimeScript^TM RT Reagent Kit with gDNA Eraser (Takara, Kusatsu, Shiga, Japan). Quantitative real-time polymerase chain reaction (qRT-PCR) was subsequently performed on the synthesized cDNA using gene-specific primers and TB Green II (Takara) on a Real-Time PCR system (Bio-Rad, Hercules, CA, USA). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (human) was used as the internal control for normalization. Primer sequences were: LPIN2 forward: 5'-TCTACAAGGGCATTAACCAGGC-3'; LPIN2 reverse: 5'-AACGTGAAAAGGTGAACACTGA-3'; GAPDH forward: 5'-TGTGGGCATCAATGGATTTGG-3'; GAPDH reverse: 5'-ACACCATGTATTCCGGGTCAAT-3'.

Lentivirus production and infection

Lentivirus production involved co-transfection of HEK293T cells with lentiviral backbone constructs (8 µg), the packaging plasmid psPAX2 (6 µg), and the envelope plasmid pMD2.G (4 µg) using polyethylenimine. The viral supernatant was collected 48 hours after transfection. To remove cellular debris, the supernatant was either filtered through a 0.45 µm filtration system or centrifuged at 1,500 ×g to obtain clarified supernatant. Target cells were exposed to the lentiviral supernatant supplemented with 2 µg/mL polybrene for 24 hours, followed by washing. Fresh medium was then added and maintained until cells reached confluence. The transduced cells were subsequently selected using an appropriate concentration of antibiotics.

RNA-seq data processing

Paired-end sequencing (150 bp) was performed on the Illumina NovaSeq 6000 platform. For each experimental condition (PC-9, PC-9 OR, PC-9 OR-R1W, PC-9 OR-sgLPIN2), three independent biological replicates were included. No batch correction was needed as all samples were processed in a single batch. The raw reads were trimmed and aligned to the GRCh38/hg38 human genome reference using HISAT2 (v2.1.0) (47). Transcript assembly and quantification were performed with StringTie (v2.2.1) (48) against GENCODE v38. The gene-level counts were transformed into transcripts per million (TPM) and log2-transformed. Differential expression analysis was performed using DESeq2 (v1.38.3), with genes meeting the thresholds of |log₂ fold change (FC)| >1 and false discovery rate (FDR) <0.05 considered significantly differentially expressed. Pathway enrichment analysis was conducted using GSVA (v1.46.0) and the clusterProfiler package (v4.7.1.3) (49) with gene sets from MSigDB and KEGG. Pathways with FDR <0.05 were considered significantly enriched. The VennDiagram package (v1.7.3) was used to identify overlapping up- and down-regulated genes across the datasets. The network and box plot were visualized using Cytoscape (v3.7.2) and the ggpubr package (v0.5.0), respectively.

Metabolomics data analysis

Metabolomics data were acquired using LC-MS with three independent biological replicates per condition, all processed in a single batch to avoid inter-batch variability. Data were analyzed using the MetaboAnalystR package (v3.0.0) (50) and were normalized by median and log-transformed. Identification of significant metabolites was performed via t-tests and orthogonal partial least squares-discriminant analysis (OPLS-DA). Metabolites with fold change >2, FDR <0.05, and variable importance in projection (VIP) >1 were considered significantly differential. The VennDiagram package was used to visualize overlapping differential metabolites across groups. Metabolic pathway enrichment analysis was performed using MetaboAnalyst (51); pathways with FDR <0.05 were considered significantly enriched. The metabolic volcano plot was generated using ggVolcano (v0.0.2).

Integration of metabolomic and transcriptomic data and candidate gene selection

To identify metabolic pathways commonly involved in TKI resistance, metabolomic pathway enrichment analysis was first performed as described in “Metabolomics data analysis” section. Pathways were prioritized if they were significantly enriched (FDR <0.05) in resistant cells compared to sensitive cells. Subsequently, within these prioritized metabolic pathways, we examined the expression levels of genes encoding key metabolic enzymes. Candidate genes were selected based on the following criteria: (I) significantly upregulated at the transcriptomic level (|log₂FC| >1, FDR <0.05) in resistant versus sensitive cells, regardless of drug withdrawal conditions; (II) corresponding metabolite intermediates or end products were significantly accumulated (FC >2, FDR <0.05, VIP >1); and (III) consistent upregulation at the protein level validated by Western blot. Genes meeting all three criteria were prioritized for subsequent functional validation.

Statistical analysis

Statistical analysis was performed using GraphPad Prism software (version 9.0, GraphPad). For cell proliferation assays, differences among various treatment groups over time were evaluated using a two-way analysis of variance (ANOVA) followed by Šídák’s multiple comparisons test. Survival data were analyzed using the Kaplan-Meier method coupled with the log-rank test for between-group comparisons, and the Bonferroni correction was applied for pairwise comparisons among multiple groups. All data are expressed as the mean ± standard error of the mean (SEM). Statistical significance was defined as P<0.05. Significance levels are denoted as follows: *, P<0.05; **, P<0.01; ***, P<0.001; and ****, P<0.0001.

Results

Result 1: LPIN2 overexpression induces TKI resistance in NSCLC cells

To elucidate the intrinsic mechanisms of acquired resistance to EGFR-TKIs, we established osimertinib-resistant (PC-9 OR) and alectinib-resistant (H3122 ALR) cell lines using a dose-escalation strategy. These models were derived from parental PC-9 cells carrying an activating EGFR mutation and NCI-H3122 cells harboring the EML4-ALK fusion (Figure 1A, Figure S1A,S1B). STR profiling and morphological examination confirmed that the genetic background and cellular morphology of the resistant cells remained largely unchanged compared to their parental counterparts (Figure S1C,S1D).

Figure 1 LPIN2 is upregulated in TKI-resistant NSCLC cells and drives resistance through triglyceride synthesis. (A) Schematic diagram illustrating the construction strategy and sequencing analysis of PC-9 and H3122 sensitive cells and their corresponding osimertinib-resistant (PC-9 OR) and alectinib-resistant (H3122 ALR) cells. (B) Untargeted metabolomics analysis of PC-9 and PC-9 OR cells. The figure shows that the TG synthesis pathway is significantly enriched in resistant cells (PC-9 OR versus PC-9). (C) Integrated metabolomic and transcriptomic analysis of PC-9 and PC-9 OR cells. The left panel shows upregulated metabolites in the TG pathway, and the right panel displays genes significantly upregulated in PC-9 OR cells, including LPIN2, a key gene in the TG synthesis pathway. (D) Lipid droplet content detection by Nile Red staining (red fluorescence) in PC-9, PC-9 OR, H3122, and H3122 ALR cells. Scale bar, 500 µm. The right panels show the quantification of relative fluorescence intensity. (E) Western blot analysis of protein expression levels of five TG synthesis pathway-related genes in PC-9 and PC-9 OR cells. (F,G) Western blot (F) and qRT-PCR (G) analysis of LPIN2 protein and mRNA expression levels in PC-9/H3122 sensitive cells and their corresponding resistant cells. Data are presented as mean ± SEM. (H,I) Cell proliferation curves measured by CCK-8 assay over 4 days for PC-9 OR (H) and H3122 ALR (I) sensitive cells, resistant cells, and LPIN2-knockout cells under treatment with 1 µM TKI. (J,K) Dose-response curves of PC-9 OR (J) and H3122 ALR (K) cells before and after LPIN2 knockout, treated with the corresponding TKI for 3 days. (L,M) Representative images of colony formation for PC-9 OR (L) and H3122 ALR (M) cells before and after LPIN2 knockout under 500 nM TKI treatment. (N) Tumor growth curves in nude mouse CDX models inoculated with PC-9, PC-9 OR, or PC-9 OR-sgLPIN2 cells (n=5). Mice received osimertinib treatment (5 mg/kg/day) after tumor formation. Data are presented as mean ± SEM. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001 by Student’s t-test or two-way ANOVA. AL, alectinib; OS, Osimertinib; ANOVA, analysis of variance; BF, bright field; NSCLC, non-small cell lung cancer; qRT-PCR, quantitative real-time polymerase chain reaction; SEM, standard error of the mean; TG, triglyceride; TKI, tyrosine kinase inhibitor.

Recent studies have gradually revealed the important role of metabolic reprogramming in TKI resistance (15,16,18,20). To systematically investigate the metabolic basis of TKI resistance, we performed metabolomic profiling using liquid chromatography-mass spectrometry (LC-MS). This analysis identified glycerophospholipid metabolism as the most significantly upregulated pathway, common to both osimertinib-maintained (PC-9 OR) and withdrawn one week (PC-9 OR-R1W) resistant cells, compared to sensitive cells (Figure 1B). Integrated transcriptomic analysis further identified five core genes within this pathway (LPIN2 ≈2.70, ETNK2 ≈3.84, PLPP3 ≈3.27, PLA2G4C ≈59.07, ACHE ≈4.70) whose expression levels and corresponding metabolites were significantly upregulated compared to sensitive cells (Figure 1C, Figure S1E). Nile Red staining confirmed increased TG levels in both PC-9 OR and H3122 ALR cells. These data collectively indicate that the TG synthesis metabolic pathway is upregulated in TKI-resistant cells (Figure 1D).

Notably, at the protein level, only LPIN2 was consistently overexpressed in the resistant cells (Figure 1E). Although the magnitude of response varied across different oncogenic backgrounds, the overall upregulation trend was consistent. Western blot analysis demonstrated a significant increase in LPIN2 protein levels in PC-9 OR cells, accompanied by an approximate 2-fold upregulation in LPIN2 mRNA. Meanwhile, H3122 ALR cells also exhibited an upward trend in both LPIN2 protein and mRNA expression, albeit to a more moderate extent compared to the PC-9 model (Figure 1F,1G).

To investigate the functional role of LPIN2 in resistant cells, LPIN2-knockout (KO) cell lines were established in both the PC-9 OR and H3122 ALR cell lines using CRISPR/Cas9 (Figure S1F,S1G). Although LPIN2 knockout did not significantly affect basal cell proliferation in the absence of drug treatment (Figure S1H,S1I), the growth of PC-9 OR LPIN2-KO cells was inhibited by over 50% compared to PC-9 OR controls (Figure 1H). Consistently, LPIN2 knockout also markedly suppressed the growth of H3122 ALR cells exposed to alectinib (Figure 1I).

While both models exhibited restored sensitivity, the magnitude of the phenotypic response displayed oncogene-specific differences. LPIN2 knockout in PC-9 OR cells caused a pronounced 7.5- to 15.6-fold reduction in the IC₅₀ of osimertinib (from 3,585 nM to 477.5–229.1 nM) (Figure 1J). Conversely, in the alectinib-treated EML4-ALK H3122 ALR model, the IC₅₀ decreased by a more moderate 4.6- to 6.9-fold (from 2,264 nM to 325.7–488.9 nM) (Figure 1K). Colony formation assays revealed that LPIN2 knockout significantly suppressed clonogenic ability in PC-9 OR cells under 500 nM osimertinib treatment, with a consistent phenotype observed in the H3122 ALR model (Figure 1L,1M). Quantitative analysis of the colony numbers further confirmed a significant reduction in clonogenic efficiency upon LPIN2 ablation in both models (Figure S1J,S1K). Based on these findings, we conclude that LPIN2 and its mediated TG synthesis pathway are critical for TKI resistance in NSCLC.

This sensitizing effect was recapitulated in vivo. In the cell-derived xenograft (CDX) mouse model, tumors formed by LPIN2-KO PC-9 OR cells showed significantly inhibited tumor growth upon osimertinib treatment compared to the non-knockout group (Figure 1N, Figure S1L). IHC revealed a dramatic reduction in the proliferation marker Ki67 in the LPIN2-KO group relative to the control group (Figure S1M). Together, these functional experiments in cellular and animal models demonstrate that LPIN2 and its regulated TG synthesis pathway are significantly upregulated in TKI-resistant cells. LPIN2 knockout restores drug sensitivity in NSCLC resistant cells, indicating that LPIN2 is a key metabolic gene essential for maintaining the TKI-resistant phenotype.

Result 2: LPIN2 drives TKI resistance in NSCLC cells by promoting the accumulation of TG

To determine whether LPIN2-mediated TKI resistance involves specific metabolites, we further analyzed the metabolomics data. Metabolomic analysis confirmed that TGs, which represent the end products of the LPIN2-involved TG synthesis pathway, were substantially accumulated in TKI-resistant cells (Figure S2A,S2B). Nile Red staining confirmed that knockout of LPIN2 effectively reversed this abnormal TG accumulation in resistant cells (Figure 2A, Figure S2C).

Figure 2 LPIN2-mediated TKI resistance depends on triglyceride accumulation. (A) Nile Red staining and quantitative fluorescence intensity analysis showing the effect of LPIN2 knockout on lipid droplet content in PC-9, PC-9 OR, and PC-9 OR sgLPIN2 cells. Scale bar, 500 µm. (B) Quantitative biochemical assay of intracellular TG levels in PC-9 cells following treatment with different concentrations of AOO. (C,D) Dose-response curves showing the effect of triglyceride AOO supplementation (100 µM) on TKI sensitivity in PC-9 (C) and H3122 (D) sensitive cells after 3 days of co-treatment with the respective TKI. (E,F) Colony formation assays evaluating the effect of AOO addition (100 µM) on the clonogenic ability of PC-9 (E) and H3122 (F) cells under TKI treatment. (G,H) Dose-response curves to the respective TKIs in LPIN2-knockout PC-9 OR (G) and H3122 ALR (H) cells following supplementation with AOO (100 µM). (I,J) Representative images and quantitative colony area analysis of LPIN2-knockout PC-9 OR (I) and H3122 ALR (J) cells under 500 nM TKI pressure after supplementation with AOO (100 µM). (K,L) Dose-response curves of PC-9 OR (K) and H3122 ALR (L) cells after 3 days of combined treatment with DGAT1/2 inhibitors (10 nM A922500/10 nM PF-06424439) and the respective TKI. (M,N) Representative images and quantitative analysis of the effect of combined treatment with 10 nM DGAT1/2 inhibitors and 500 nM TKI on the colony formation ability of PC-9 OR (M) and H3122 ALR (N) cells. Data are presented as mean ± SEM. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001 by Student’s t-test or two-way ANOVA. ANOVA, analysis of variance; OS, osimertinib; SEM, standard error of the mean; TG, triglyceride; TKI, tyrosine kinase inhibitor.

We further investigated the direct impact of TGs on TKI resistance utilizing AOO as an exogenous TG supplement. Dose-response viability assays initially confirmed that 100 µM AOO did not induce non-specific cytotoxicity independent of TKI treatment (Figure S2D, S2E). Furthermore, both qualitative Nile Red fluorescence imaging and quantitative biochemical assays robustly validated that this concentration effectively induced intracellular TG accumulation in the tested models (Figure 2B, Figure S2F-S2H). Following these method validations, we observed that treatment with 100 µM AOO increased the IC₅₀ to osimertinib in PC-9 cells from 18.5 to 201.9 nM (Figure 2C). Similarly, treatment with 100 µM AOO elevated the IC₅₀ to alectinib in H3122 cells from 70.75 to 706.7 nM (Figure 2D), while LPIN2 expression remained unaffected (Figure S2I, S2J). Furthermore, exogenous lipid supplementation with 100 µM AOO effectively sustained the colony-forming ability of PC-9 cells under 500 nM osimertinib pressure. A comparable pro-survival phenotype was concurrently observed in the H3122 model under alectinib treatment, which further reinforces the role of lipid accumulation in supporting TKI tolerance across distinct oncogenic contexts (Figure 2E,2F). Meanwhile, treatment with 100 µM AOO increased the IC₅₀ to osimertinib in PC-9 OR LPIN2-KO cells from 929.7/962.5 to 3,052/4,010 nM (Figure 2G) and elevated the IC₅₀ to alectinib in H3122 ALR LPIN2-KO cells from 170.8/165.8 to 635.5/380 nM (Figure 2H). In addition, under 500 nM osimertinib or alectinib treatment, treatment with 100 µM AOO significantly enhanced the colony-forming capacity (Figure 2I,2J) and proliferation ability (Figure S2K,S2L) of LPIN2-KO cells. These results indicate that LPIN2 likely mediates TKI resistance through TGs.

Next, we inhibited the TG synthesis pathway using diacylglycerol O-acyltransferase 1/2 (DGAT1/2) inhibitors (A922500/PF-06424439) and evaluated their effects on TKI resistance. Treatment with 10 nM A922500 and 10 nM PF-06424439 significantly reduced the proliferation and viability of resistant cells under TKI pressure. In line with the oncogene-specific differences observed previously, the magnitude of sensitization varied between the two models: the IC₅₀ in EGFR-mutant PC-9 OR cells decreased by approximately 5.8-fold (from 4,011 to 692.7 nM), whereas ALK-fusion H3122 ALR cells exhibited a slightly more pronounced 7.8-fold reduction (from 2,208 to 284 nM) (Figure 2K,2L, Figure S2M,S2N). Colony formation assays demonstrated that combined treatment with 10 nM A922500, 10 nM PF-06424439, and TKIs markedly suppressed the clonogenic ability of both PC-9 OR and H3122 ALR resistant cell lines (Figure 2M,2N), without affecting LPIN2 expression (Figure S2O,S2P). These results indicate that LPIN2 primarily mediates TKI resistance in NSCLC through its metabolite TG.

Result 3: TG accumulation is associated with PI3K-AKT pathway modulation to promote TKI resistance in NSCLC

To elucidate the downstream mechanisms by which LPIN2 regulates drug resistance, we performed transcriptomic analysis of LPIN2-KO cells. The results indicated that phosphoinositide 3-kinase (PI3K)-related signaling pathways were significantly suppressed upon LPIN2 depletion (Figure 3A). Western blot analysis further confirmed that the phosphorylation levels of key proteins in the PI3K-protein kinase B (AKT) pathway were markedly elevated in PC-9 OR cells, whereas LPIN2 knockout effectively inhibited this activation (Figure 3B). Flow cytometry analysis of apoptosis revealed that osimertinib treatment induced apoptosis in LPIN2-KO PC-9 OR cells (Figure S3A). In the CDX model, immunohistochemical staining showed that LPIN2 knockout reduced p-AKT levels in tumor tissues (Figure 3C).

Figure 3 LPIN2 activates the PI3K-AKT survival pathway to confer TKI resistance. (A) Transcriptomic analysis of PC-9 OR-sgLPIN2 versus control cells. GSEA enrichment plot shows significant downregulation of the PI3K signaling pathway following LPIN2 knockout. (B) Western blot analysis of key PI3K-AKT pathway proteins and their phosphorylation levels in PC-9 OR cells before and after LPIN2 knockout. (C) Immunohistochemical staining for p-AKT (S473) levels in PC-9 OR and PC-9 OR-sgLPIN2 tumor tissues from the nude mouse CDX model. Scale bar, 100 µm. (D) Western blot analysis of apoptosis-related proteins (BCL2, BCL-XL, BAX) in PC-9 OR cells before and after LPIN2 knockout. (E,F) Western blot analysis of phosphorylation and total protein levels of PI3K-AKT pathway proteins in PC-9 (E) and H3122 (F) sensitive cells, resistant cells, and LPIN2-knockout cells. (G,H) Western blot analysis showing the effect of AOO addition on PI3K-AKT pathway activation in PC-9 (G) and PC-9 OR-sgLPIN2 (H) cells. (I,J) Western blot analysis showing the effect of AOO addition on PI3K-AKT pathway activation in H3122 (I) and H3122 ALR-sgLPIN2 (J) cells. (K,L) Dose-response curves and IC50 values for PC-9 OR (K) and H3122 ALR (L) cells treated with 10 nM of the PI3K inhibitor Wortmannin or the AKT inhibitor MK-2206 in combination with the respective TKI. (M,N) Colony formation assays showing the effect of 10 nM Wortmannin on the colony-forming ability of PC-9 OR (M) and H3122 ALR (N) cells under TKI pressure. (O) Proliferation inhibition curves of PC-9 OR cells treated with osimertinib in the presence or absence of 10 nM Wortmannin, as measured by CCK-8 assay. ns, P≥0.05; *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. CCK-8, Cell Counting Kit-8; CDX, cell-derived xenograft; GSEA, Gene Set Enrichment Analysis; NC, negative control; OS, osimertinib; TKI, tyrosine kinase inhibitor.

We further sought to verify whether LPIN2-mediated lipid accumulation regulates apoptosis and drug resistance through an association with the PI3K-AKT signaling pathway. We found that LPIN2 knockout significantly reduced the level of the anti-apoptotic protein B-cell lymphoma 2 (BCL2) and enhanced apoptosis in resistant cells [decreased B-cell lymphoma-extra large (BCL-xL), increased BCL2-associated X protein (BAX)] (Figure 3D). The PI3K-AKT pathway was significantly activated in both PC-9 OR and H3122 ALR cells, while LPIN2 knockout suppressed this activation (Figure 3E,3F). Notably, assessment of the MAPK pathway revealed that ERK phosphorylation (p-ERK) was markedly upregulated in TKI-resistant PC-9 OR cells, whereas the phosphorylation levels of JNK and p38 exhibited no obvious changes. Furthermore, this elevated p-ERK level remained unaffected regardless of LPIN2 ablation (Figure S3B). Treatment with 100 µM AOO enhanced the phosphorylation of PI3K-AKT pathway components in PC-9 cells and counteracted osimertinib-induced apoptosis (Figure 3G,3H, Figure S3C). Similarly, activation of the PI3K-AKT pathway was also observed in H3122 cells (Figure 3I,3J).

To clarify the role of the PI3K-AKT pathway in cellular drug resistance, we respectively treated resistant cells with the PI3K inhibitor Wortmannin and the AKT inhibitor MK-2206. Both inhibitors significantly enhanced the sensitivity of resistant cells, displaying oncogene-specific differences in the magnitude of response. In EGFR-mutant PC-9 OR cells, the IC₅₀ of osimertinib was reduced by approximately 3.1- to 4.1-fold (from 2,473 to 798.7–605.3 nM) (Figure 3K). In contrast, the ALK-fusion H3122 ALR cells exhibited a more moderate 2.0- to 2.4-fold decrease in the IC₅₀ of alectinib (from 995.8 to 411.1–493.7 nM) (Figure 3L). Such variation in sensitization efficacy suggests that distinct primary oncogenic drivers may confer varying degrees of metabolic dependency on the LPIN2 pathway under therapeutic stress. Colony formation and cell proliferation assays further supported these findings, demonstrating that both Wortmannin and MK-2206 markedly increased the drug sensitivity of resistant cells to their corresponding TKIs (osimertinib or alectinib) (Figure 3M-3O). These results indicate that LPIN2-mediated lipid accumulation promotes TKI resistance via PI3K-AKT activation.

Result 4: targeting the LPIN2-associated PI3K signaling network overcomes TKI resistance in vivo

To validate the critical role and potential therapeutic value of the LPIN2-associated PI3K signaling network in TKI resistance at the in vivo level, we conducted a series of animal studies. First, we performed pharmacological intervention experiments in a CDX model established with PC-9 OR cells. Tumor-bearing mice were randomized into four groups receiving: osimertinib monotherapy (5 mg/kg/d), Wortmannin monotherapy (1 mg/kg/d), combination therapy, or vehicle control (PC-9 tumor-bearing mice). The result revealed that the combination of Wortmannin and osimertinib exhibited the strongest tumor growth inhibition compared with either monotherapy group (Figure 4A). Furthermore, the combination therapy significantly prolonged the overall survival of tumor-bearing mice (Figure 4B). Immunohistochemical analysis of the tumors aligned with this functional synergy: while PC-9 OR xenografts exhibited strong activation of the PI3K-AKT pathway (high p-AKT), combination therapy significantly suppressed this activation and concurrently reduced Ki67 expression (Figure 4C).

Figure 4 In vivo validation of the LPIN2-PI3K axis and a proposed mechanism model for TKI resistance. (A,B) In CDX models established with PC-9 sensitive cells and PC-9 OR cells, tumor growth curves (A) and survival curves (B) of mice (n=5) treated with osimertinib monotherapy, Wortmannin monotherapy, or combination therapy during the treatment period. (C) Immunohistochemical staining for Ki67 and p-AKT (S473) expression levels in tumor tissues from each treatment group after the end of therapy. Scale bar, 100 µm. The lower panels display the corresponding quantitative analysis of the Ki67 and p-AKT H-scores. (D) Tumor growth curves of PC-9 OR CDX models (n=5 per group) treated with osimertinib monotherapy, DGAT1/2 inhibitors (A922500 and PF-06424439) monotherapy, or combination therapy. (E) A schematic model illustrating how LPIN2-driven lipid metabolic reprogramming modulates the PI3K-AKT signaling network to confer TKI resistance in NSCLC. ns, P≥0.05; *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. CDX, cell-derived xenograft; NSCLC, non-small cell lung cancer; OS, osimertinib; TKI, tyrosine kinase inhibitor.

Furthermore, to more directly address the therapeutic potential of targeting TG biosynthesis in vivo, we evaluated the combination of osimertinib and DGAT1/2 inhibitors in the PC-9 OR CDX model. Consistent with our in vitro findings and the Wortmannin intervention data, the in vivo pharmacological blockade of TG synthesis effectively sensitized the resistant tumors to osimertinib, resulting in a marked suppression of tumor growth compared to the osimertinib monotherapy group (Figure 4D), which is consistent with the enhanced apoptosis observed in vitro upon combination treatment (Figure S3D).

These in vivo findings indicate that LPIN2 contributes to TKI resistance in NSCLC through the PI3K-AKT pathway activation.

This study reveals a central metabolic-signaling axis that contributes to acquired TKI resistance in NSCLC. In resistant cells, the lipid-metabolizing enzyme LPIN2 is significantly upregulated, catalyzing the conversion of PA to DAG and thereby driving TG biosynthesis and lipid droplet deposition. This lipid metabolic reprogramming subsequently may modulate PI3K-AKT-mTOR cell survival signaling, ultimately inhibiting apoptosis and conferring TKI resistance (Figure 4E).

Discussion

Although targeted therapies can treat cancer with precision, the tumor develop acquired resistance through various mechanisms such as secondary mutations, bypass signaling activation, and histological transformation, leading to reduced efficacy and patient relapse (52). Recent studies have revealed that tumor cells adapt to drug pressure and sustain survival through metabolic reprogramming, which has become a core driver of resistance (53). Unlike most studies conducted under sustained high-concentration TKI pressure (19), we employed a TKI-withdrawal resistant cell model to exclude interference from cellular drug stress responses. Integrated metabolomic and transcriptomic analysis revealed that the TG synthesis pathway and its key enzymes are significantly upregulated following TKI resistance. Through multiple rounds of functional experiments, LPIN2 was pinpointed as the core regulator from among five candidate genes.

Further research demonstrates that LPIN2 plays a pivotal role in acquired TKI resistance in NSCLC by driving TG metabolic reprogramming and promoting the modulation of the PI3K-AKT survival signaling pathway. The genomic landscape of NSCLC is characterized by diverse driver mutations, with EGFR mutations and ALK rearrangements representing two major therapeutic targets. Our findings suggest that LPIN2-driven lipid remodeling potentially serves as a bypass mechanism that may facilitate the development of TKI resistance. Although quantitative variations in response magnitude were observed between the two models, these differences likely reflect the inherent wiring of EGFR versus ALK signaling. Such metabolic convergence under therapeutic stress underscores the potential of this pathway as a shared adaptive strategy that transcends specific genomic subtypes in TKI resistance.

Considering that the LPIN family comprises multiple members, we specifically observed that only LPIN2 is selectively upregulated during resistance and participates in TKI resistance of tumor cells. We speculate that while LPIN1 is considered essential for basal lipid metabolism in normal cells and LPIN3 is expressed at very low levels in the lung and most tissues, this selective upregulation of LPIN2 indicates its significant role in adapting to targeted drug therapy (54,55). Such metabolic reprogramming may provide a “backup” pathway for the persistence of drug tolerance. When the primary oncogenic driver signal is blocked by TKIs, the upregulation of LPIN2 and the consequent lipid remodeling may support survival through a dual mechanism: first, by storing TGs as a dense energy reserve to fuel rapid proliferation upon recovery or to aid in surviving metabolic stress (56); second, by potentially generating lipid second messengers, altering membrane composition, or modulating downstream signaling cascades to reactivate critical pro-survival pathways such as PI3K-AKT, thereby achieving “metabolic escape” from drug pressure. We propose that the TG accumulation facilitated by induced LPIN2 expression is not merely a passive energy reservoir. Instead, it constitutes an active, inducible survival strategy, enabling tumor cells to reshape their metabolic landscape when oncogenic pathways are inhibited. Elucidating the exact transcriptional mechanisms driving the specific induction of LPIN2 upon TKI treatment, and identifying the transcription factors responsible for this upregulation, would not only clarify the upstream signals connecting TKI target inhibition to metabolic adaptation but may also reveal new therapeutic nodes.

Based on the above mechanisms, targeting LPIN2 or its downstream pathways demonstrates significant clinical translation potential. LPIN2, a key lipid metabolic enzyme, represents a highly attractive potential therapeutic target. This study has confirmed that both knockout of LPIN2 and pharmacological inhibition of its downstream enzymes DGAT1/2 can effectively reverse drug resistance. Considering that direct inhibitors targeting LPIN2 are not yet available, a more immediately feasible strategy is to utilize TGs as a biomarker to identify patient populations resistant to TKIs and guide them toward combination therapy regimens. Our in vitro and in vivo experiments both indicate that combining PI3K/AKT pathway inhibitors with TKIs produces significant synergistic anti-tumor effects, which is highly consistent with recent findings that AKT or PI3K blockade can successfully overcome resistance to osimertinib and ALK inhibitors in specific NSCLC subtypes (57,58). Future clinical studies should focus on exploring the feasibility of LPIN2 as a predictive biomarker for resistance and evaluating the safety and efficacy of combination metabolic therapy in patients with advanced NSCLC.

Conclusions

Based on the above findings, this study systematically elucidates the critical role of LPIN2 in TKI resistance in NSCLC and its underlying molecular mechanism. We demonstrated that LPIN2 expression is significantly upregulated in resistant cells, where it promotes TG synthesis and lipid droplet accumulation, indirectly modulates the PI3K-AKT signaling pathway, and subsequently inhibits apoptosis, ultimately leading to TKI resistance. In both in vitro and in vivo models, knockout of LPIN2 effectively reversed lipid accumulation, suppressed PI3K-AKT pathway activity, and restored tumor TKI sensitivity, albeit with oncogene-specific differences in the magnitude of response. Notably, exogenous TG supplementation restored the resistance even in the absence of LPIN2, while inhibition of key TG synthesis enzymes DGAT1/2 or direct suppression of the PI3K-AKT pathway effectively mimicked the sensitization effect of LPIN2 knockout. These results not only reveal the important role of lipid metabolic reprogramming in TKI resistance, but also provide new potential targets and combination therapy strategies for overcoming targeted therapy resistance in NSCLC.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the MDAR and ARRIVE reporting checklists. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1471/rc

Data Sharing Statement: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1471/dss

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1471/prf

Funding: This work was supported by the 1.3.5 Project for Disciplines of Excellence from West China Hospital of Sichuan University (No. ZYJC21015, to Y.C.; No. ZYGD23010, to L.L.), the National Natural Science Foundation of China (No. 82203153, to K.Z.), Natural Science Foundation of Sichuan Province (No. 2024NSFSC0746, to Y.C.), and “RongPiao” innovated Program of Chengdu (No. TJZ20244902, to Y.C.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1-1471/coif). All authors report grants from the 1.3.5 Project for Disciplines of Excellence from West China Hospital of Sichuan University, the National Natural Science Foundation of China, the Natural Science Foundation of Sichuan Province, and the “Rongpiao” Innovated Program of Chengdu during the conduct of the study. The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All in vivo animal experiments were performed under a project license (approval No. 20240315033) granted by the Animal Ethics Committee of West China Hospital, Sichuan University, in compliance with institutional animal welfare guidelines for the care and use of animals.

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

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Han B, Zheng R, Zeng H, et al. Cancer incidence and mortality in China, 2022. J Natl Cancer Cent 2024;4:47-53. [Crossref] [PubMed]
3. Lai B, Xu C, Lai S, et al. From EGFR PTM network to TKI resistance: spatial subtypes and targeting in lung cancer. Cancer Drug Resist 2026;9:7. [Crossref] [PubMed]
4. Herbst RS, Morgensztern D, Boshoff C. The biology and management of non-small cell lung cancer. Nature 2018;553:446-54. [Crossref] [PubMed]
5. de Castro J, Rodríguez MC, Martínez-Zorzano VS, et al. Erythrocyte fatty acids as potential biomarkers in the diagnosis of advanced lung adenocarcinoma, lung squamous cell carcinoma, and small cell lung cancer. Am J Clin Pathol 2014;142:111-20. [Crossref] [PubMed]
6. Torrente M, Sousa PA, Franco F, et al. Understanding prognosis and survival outcomes in patients with early-stage non-small-cell lung cancer. Clin Med (Lond) 2022;22:38-40. [Crossref] [PubMed]
7. Cho BC, Kim DW, Park K, et al. Real-world use of osimertinib in non-small cell lung cancer: ASTRIS study Korean subgroup analysis. Curr Med Res Opin 2020;36:477-82. [Crossref] [PubMed]
8. Masuda K, Fujiwara Y, Shinno Y, et al. Efficacy and safety of crizotinib in patients with ROS1 rearranged non-small cell lung cancer: a retrospective analysis. J Thorac Dis 2019;11:2965-72. [Crossref] [PubMed]
9. Cao Y, Qiu X, Xiao G, et al. Effectiveness and safety of osimertinib in patients with metastatic EGFR T790M-positive NSCLC: An observational real-world study. PLoS One 2019;14:e0221575. [Crossref] [PubMed]
10. Solomon BJ, Liu G, Felip E, et al. Lorlatinib Versus Crizotinib in Patients With Advanced ALK-Positive Non-Small Cell Lung Cancer: 5-Year Outcomes From the Phase III CROWN Study. J Clin Oncol 2024;42:3400-9. [Crossref] [PubMed]
11. Chmielecki J, Gray JE, Cheng Y, et al. Candidate mechanisms of acquired resistance to first-line osimertinib in EGFR-mutated advanced non-small cell lung cancer. Nat Commun 2023;14:1070. [Crossref] [PubMed]
12. Li Z, Liu F, Wu S, et al. Research Progress on the Drug Resistance of ALK Kinase Inhibitors. Curr Med Chem 2022;29:2456-75. [Crossref] [PubMed]
13. Westover D, Zugazagoitia J, Cho BC, et al. Mechanisms of acquired resistance to first- and second-generation EGFR tyrosine kinase inhibitors. Ann Oncol 2018;29:i10-9. [Crossref] [PubMed]
14. Wu J, Lin Z. Non-Small Cell Lung Cancer Targeted Therapy: Drugs and Mechanisms of Drug Resistance. Int J Mol Sci 2022;23:15056. [Crossref] [PubMed]
15. You Q, Wang J, Yu Y, et al. The histone deacetylase SIRT6 promotes glycolysis through the HIF-1alpha/HK2 signaling axis and induces erlotinib resistance in non-small cell lung cancer. Apoptosis 2022;27:883-98. [Crossref] [PubMed]
16. Fang Y, Li X, Cheng H, et al. ANGPTL4 Regulates Lung Adenocarcinoma Pyroptosis and Apoptosis via NLRP3\ASC\Caspase 8 Signaling Pathway to Promote Resistance to Gefitinib. J Oncol 2022;2022:3623570. [Crossref] [PubMed]
17. Bach DH, Luu TT, Kim D, et al. BMP4 Upregulation Is Associated with Acquired Drug Resistance and Fatty Acid Metabolism in EGFR-Mutant Non-Small-Cell Lung Cancer Cells. Mol Ther Nucleic Acids 2018;12:817-28. [Crossref] [PubMed]
18. Kim S, Jeon JS, Choi YJ, et al. Heterogeneity of glutamine metabolism in acquired-EGFR-TKI-resistant lung cancer. Life Sci 2022;291:120274. [Crossref] [PubMed]
19. Lin Z, Li J, Zhang J, et al. Metabolic Reprogramming Driven by IGF2BP3 Promotes Acquired Resistance to EGFR Inhibitors in Non-Small Cell Lung Cancer. Cancer Res 2023;83:2187-207. [Crossref] [PubMed]
20. Song H, Liu D, Wang L, et al. Methyltransferase like 7B is a potential therapeutic target for reversing EGFR-TKIs resistance in lung adenocarcinoma. Mol Cancer 2022;21:43. [Crossref] [PubMed]
21. Fang S, Zhu Y, Chen W, et al. NAT10-mediated lipid metabolic reprogramming drives EGFR-TKI resistance in non-small cell lung cancer via ac4C-dependent mRNA stabilization. Exp Hematol Oncol 2025;14:134. [Crossref] [PubMed]
22. Ma L, Han Y, Xu L, et al. PKC-iota drives EGFR-TKI resistance in EGFR-mutated NSCLC by phosphorylating FASN to reprogram lipid metabolism. Transl Lung Cancer Res 2026;15:25. [Crossref] [PubMed]
23. Chen Y, Bai M, Liu M, et al. Metabolic Reprogramming in Lung Cancer: Hallmarks, Mechanisms, and Targeted Strategies to Overcome Immune Resistance. Cancer Med 2025;14:e71317. [Crossref] [PubMed]
24. Wang J, Wang J, Chen J. Precision navigation through the labyrinth: overcoming EGFR resistance in non-Small cell lung cancer. Ann Med 2025;57:2574526. [Crossref] [PubMed]
25. Shao W, Yu C, Xu R, et al. Targeted lipidomics meets transcriptomics: how cinobufagin rewires fatty acid, sphingolipid, and glycerophospholipid metabolism to combat hepatoma cell growth. Front Pharmacol 2025;16:1664915. [Crossref] [PubMed]
26. Hermansson M, Hokynar K, Somerharju P. Mechanisms of glycerophospholipid homeostasis in mammalian cells. Prog Lipid Res 2011;50:240-57. [Crossref] [PubMed]
27. Farooqui AA, Horrocks LA, Farooqui T. Glycerophospholipids in brain: their metabolism, incorporation into membranes, functions, and involvement in neurological disorders. Chem Phys Lipids 2000;106:1-29. [Crossref] [PubMed]
28. Ecker J, Liebisch G. Application of stable isotopes to investigate the metabolism of fatty acids, glycerophospholipid and sphingolipid species. Prog Lipid Res 2014;54:14-31. [Crossref] [PubMed]
29. Dickinson A, Saraswat M, Joenväärä S, et al. Mass spectrometry-based lipidomics of oral squamous cell carcinoma tissue reveals aberrant cholesterol and glycerophospholipid metabolism - A Pilot study. Transl Oncol 2020;13:100807. [Crossref] [PubMed]
30. Henderson F, Johnston HR, Badrock AP, et al. Enhanced Fatty Acid Scavenging and Glycerophospholipid Metabolism Accompany Melanocyte Neoplasia Progression in Zebrafish. Cancer Res 2019;79:2136-51. [Crossref] [PubMed]
31. Yang T, Hui R, Nouws J, et al. Untargeted metabolomics analysis of esophageal squamous cell cancer progression. J Transl Med 2022;20:127. [Crossref] [PubMed]
32. Xu X, Lv J, Guo F, et al. Gut Microbiome Influences the Efficacy of PD-1 Antibody Immunotherapy on MSS-Type Colorectal Cancer via Metabolic Pathway. Front Microbiol 2020;11:814. [Crossref] [PubMed]
33. Triantafyllou EA, Georgatsou E, Mylonis I, et al. Expression of AGPAT2, an enzyme involved in the glycerophospholipid/triacylglycerol biosynthesis pathway, is directly regulated by HIF-1 and promotes survival and etoposide resistance of cancer cells under hypoxia. Biochim Biophys Acta Mol Cell Biol Lipids 2018;1863:1142-52. [Crossref] [PubMed]
34. Yan C, Wu D, Gan L, et al. Significant metabolic alterations in non-small cell lung cancer patients by epidermal growth factor receptor-targeted therapy and PD-1/PD-L1 immunotherapy. Front Pharmacol 2022;13:949745. [Crossref] [PubMed]
35. Kappelt F, Maniak M. LPIN2 is the phosphatase dominating the penultimate step of neutral lipid biosynthesis in Dictyostelium. MicroPubl Biol 2024;
36. Carman GM, Han GS. Roles of phosphatidate phosphatase enzymes in lipid metabolism. Trends Biochem Sci 2006;31:694-9. [Crossref] [PubMed]
37. Donkor J, Sariahmetoglu M, Dewald J, et al. Three mammalian lipins act as phosphatidate phosphatases with distinct tissue expression patterns. J Biol Chem 2007;282:3450-7. [Crossref] [PubMed]
38. Coleman RA, Mashek DG. Mammalian triacylglycerol metabolism: synthesis, lipolysis, and signaling. Chem Rev 2011;111:6359-86. [Crossref] [PubMed]
39. Ali A, Levantini E, Teo JT, et al. Fatty acid synthase mediates EGFR palmitoylation in EGFR mutated non-small cell lung cancer. EMBO Mol Med 2018;10:e8313. [Crossref] [PubMed]
40. Zhang P, Csaki LS, Ronquillo E, et al. Lipin 2/3 phosphatidic acid phosphatases maintain phospholipid homeostasis to regulate chylomicron synthesis. J Clin Invest 2019;129:281-95. [Crossref] [PubMed]
41. Ryu D, Seo WY, Yoon YS, et al. Endoplasmic reticulum stress promotes LIPIN2-dependent hepatic insulin resistance. Diabetes 2011;60:1072-81. [Crossref] [PubMed]
42. Valdearcos M, Esquinas E, Meana C, et al. Lipin-2 reduces proinflammatory signaling induced by saturated fatty acids in macrophages. J Biol Chem 2012;287:10894-904. [Crossref] [PubMed]
43. Lordén G, Sanjuán-García I, de Pablo N, et al. Lipin-2 regulates NLRP3 inflammasome by affecting P2X7 receptor activation. J Exp Med 2017;214:511-28. [Crossref] [PubMed]
44. Varghese F, Bukhari AB, Malhotra R, et al. IHC Profiler: an open source plugin for the quantitative evaluation and automated scoring of immunohistochemistry images of human tissue samples. PLoS One 2014;9:e96801. [Crossref] [PubMed]
45. Guzmán C, Bagga M, Kaur A, et al. ColonyArea: an ImageJ plugin to automatically quantify colony formation in clonogenic assays. PLoS One 2014;9:e92444. [Crossref] [PubMed]
46. Gong Z, Zhu J, Chen J, et al. CircRREB1 mediates lipid metabolism related senescent phenotypes in chondrocytes through FASN post-translational modifications. Nat Commun 2023;14:5242. [Crossref] [PubMed]
47. Kim D, Paggi JM, Park C, et al. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol 2019;37:907-15. [Crossref] [PubMed]
48. Kovaka S, Zimin AV, Pertea GM, et al. Transcriptome assembly from long-read RNA-seq alignments with StringTie2. Genome Biol 2019;20:278. [Crossref] [PubMed]
49. Xu S, Hu E, Cai Y, et al. Using clusterProfiler to characterize multiomics data. Nat Protoc 2024;19:3292-320. [Crossref] [PubMed]
50. Pang Z, Chong J, Li S, et al. MetaboAnalystR 3.0: Toward an Optimized Workflow for Global Metabolomics. Metabolites 2020;10:186. [Crossref] [PubMed]
51. Xia J, Sinelnikov IV, Han B, et al. MetaboAnalyst 3.0--making metabolomics more meaningful. Nucleic Acids Res 2015;43:W251-7. [Crossref] [PubMed]
52. Zhao J, Xu W, Zhou F, et al. Navigating the landscape of EGFR TKI resistance in EGFR-mutant NSCLC - mechanisms and evolving treatment approaches. Nat Rev Clin Oncol 2026;23:63-83. [Crossref] [PubMed]
53. Pulido I, García-Cañaveras JC, Rodríguez ML, et al. NNMT-driven metabolic reprogramming creates a NAMPT druggable vulnerability and reveals liquid biopsy biomarkers for TKI resistance in EGFR-mutant NSCLC. Cancer Lett 2025; Epub ahead of print. [Crossref]
54. Gan G, Luo Y, Zeng Y, et al. Gut microbiota dysbiosis links chronic apical periodontitis to liver fibrosis in nonalcoholic fatty liver disease: Insights from a mouse model. Int Endod J 2024;57:1608-22. [Crossref] [PubMed]
55. Chen Y, Rui BB, Tang LY, et al. Lipin family proteins--key regulators in lipid metabolism. Ann Nutr Metab 2015;66:10-8. [Crossref] [PubMed]
56. Gonzalez MA, Lopez JL, Schenkel K, et al. A Role for Lipid Metabolism in Tyrosine Kinase Inhibitor (TKI) Resistance of Chronic Myeloid Leukemia (CML). Blood 2021;138:2542.
57. Grazini U, Markovets A, Ireland L, et al. Overcoming Osimertinib Resistance with AKT Inhibition in EGFRm-Driven Non-Small Cell Lung Cancer with PIK3CA/PTEN Alterations. Clin Cancer Res 2024;30:4143-54. [Crossref] [PubMed]
58. Talwelkar SS, Mäyränpää MI, Schüler J, et al. PI3Kβ inhibition enhances ALK-inhibitor sensitivity in ALK-rearranged lung cancer. Mol Oncol 2023;17:747-64. [Crossref] [PubMed]