Lapatinib induces ferroptosis in osteosarcoma via the SLC1A5-GPX4 axis

Abstract

Background

Osteosarcoma (OS) continues to have a poor prognosis, underscoring the urgent need for novel therapeutic strategies. Ferroptosis, an iron-dependent form of regulated cell death, offers a promising alternative to circumvent apoptosis resistance. Lapatinib (Lap), a well-known EGFR/HER2 inhibitor, exhibits potential beyond its canonical targets.

Methods

The anti-OS effects of Lap were evaluated in U2OS and HOS cell lines using CCK-8, colony formation, Transwell, and flow cytometry assays. Its potential target was identified through molecular docking, molecular dynamics simulations, CETSA, and DARTS. Ferroptosis induction was assessed by measuring key markers (MDA, GSH, Fe²⁺, lipid ROS) and related proteins (SLC1A5, GPX4). An in vivo xenograft mouse model was used to confirm ferroptosis induction.

Results

Lap significantly inhibited OS cell proliferation, migration, and invasion, and induced cell death predominantly via ferroptosis. It elevated lipid peroxidation and triggered characteristic ferroptotic events (Fe²⁺ accumulation, lipid ROS, GSH depletion), all of which were reversed by the ferroptosis inhibitor DFO. Moreover, SLC1A5 knockdown attenuated lapatinib-induced ferroptosis, whereas SLC1A5 overexpression rescued it, confirming that SLC1A5 is a functional target. Molecular docking, CETSA, and DARTS supported a potential interaction between lapatinib and SLC1A5. Mechanistically, lapatinib disrupted glutamine uptake and GSH synthesis, leading to GPX4 suppression and ferroptosis. In vivo, lapatinib suppressed tumor growth and downregulated SLC1A5/GPX4, effects that were reversible by DFO.

Conclusion

This study reveals a novel mechanism by which lapatinib inhibits OS via the SLC1A5-GPX4 axis to induce ferroptosis, providing a preclinical rationale for further evaluation of lapatinib repurposing in osteosarcoma.

Highlights

• •Lapatinib induces ferroptosis in osteosarcoma through interaction with the glutamine transporter SLC1A5, suggesting a novel mechanism distinct from its canonical targets.
• •Mechanistically, lapatinib modulates the SLC1A5-GPX4 axis, which contributes to ferroptosis induction.
• •The SLC1A5-mediated ferroptosis pathway suppresses tumor growth in vivo, providing a preclinical rationale for further exploration of lapatinib repurposing in osteosarcoma.

1. Introduction

Osteosarcoma (OS), the most common primary malignancy of bone, accounts for approximately 35% of all primary bone tumors [1], [2]. Its clinical management is severely hampered by aggressive local growth, early metastatic dissemination, and frequent therapy resistance, culminating in poor survival rates and adverse patient outcomes [1], [3]. The efficacy of conventional chemotherapy is often limited by intrinsic or acquired drug resistance, leading to a high incidence of relapse [3], [4], [5]. This pressing clinical reality underscores the urgent need for novel therapeutic paradigms.

Ferroptosis, an iron-dependent, non-apoptotic form of regulated cell death [6], [7], is mechanistically and morphologically distinct from apoptosis, necrosis, and autophagy [8], [9]. Its execution is driven by the iron-catalyzed accumulation of lipid peroxides, which overwhelms cellular antioxidant defenses and disrupts redox balance [10], [11], [12], [13]. As the anti-tumor activity of many standard chemotherapeutics relies on the engagement of apoptotic pathways, evasion of apoptosis represents a common mechanism of drug resistance [14]. Harnessing ferroptosis thus presents a compelling alternative strategy for eliminating resistant cancer cells [15], [16], [17]. Growing evidence indicates that pharmacological induction of ferroptosis effectively suppresses tumor cell viability across various musculoskeletal cancers, including OS [18], highlighting the need to define the precise regulators of this pathway in OS.

Lapatinib (Lap), an orally bioavailable small-molecule inhibitor, selectively targets the tyrosine kinase activities of EGFR and HER2 [19], [20]. Its primary mechanism involves suppressing downstream pro-survival signaling cascades, notably the MAPK and PI3K/AKT pathways [21], [22], leading to its initial approval in breast cancer. Emerging evidence suggests that lapatinib may also induce ferroptosis in tumor cells [23]. Although its anti-proliferative efficacy has been established in several cancers [24], its potential to trigger non-canonical cell death in OS, and the underlying mechanisms, remain largely undefined.

The Solute Carrier Family 1 Member 5 (SLC1A5), a primary plasma membrane transporter for glutamine (Gln), is a critical gatekeeper for cellular Gln metabolism [25], [26], [27], [28]. Its overexpression is frequently associated with tumor progression in diverse malignancies including pancreatic [29], lung [30], and breast cancers [31], as well as leukemia [32]. Notably, SLC1A5 is implicated in regulating ferroptosis [33]. For example, SLC1A5 knockdown has been shown to suppress glioma proliferation and invasion while increasing ferroptosis susceptibility in a GPX4-dependent manner [34]. Furthermore, in gastric cancer, the flavonoid quercetin promotes ferroptosis by targeting SLC1A5 and downregulating the XCT/GPX4 axis [35]. These observations posit SLC1A5-mediated Gln flux as a crucial node linking tumor metabolism to ferroptosis, making it an attractive yet unexplored target in OS.

Here, we investigated the mechanism by which lapatinib inhibits OS progression and its potential link to SLC1A5. We hypothesized that lapatinib might interact with SLC1A5, thereby inducing ferroptosis. To our knowledge, this is the first study to identify SLC1A5 as a potential molecular target of lapatinib and to demonstrate that the SLC1A5-GPX4 axis mediates lapatinib-induced ferroptosis in osteosarcoma. To test this, we performed molecular docking, CETSA, DARTS, and functional assays to determine whether SLC1A5 is a potential target and whether the SLC1A5-GPX4 axis mediates lapatinib-induced ferroptosis in OS cells and in a xenograft model.

2. Materials and methods

2.1. Cell culture

Human osteosarcoma cell lines U2OS (CL-0236) and HOS (CL-0360) were obtained from Procell (Wuhan, China). U2OS cells were cultured in McCoy's 5 A medium (PM150710; Procell, Wuhan, China), and HOS cells were maintained in Minimum Essential Medium (MEM) (PM150410; Procell, Wuhan, China), both supplemented with 10% fetal bovine serum (FBS). All cells were incubated at 37 °C in a humidified atmosphere with 5% CO₂ and routinely tested for mycoplasma contamination.

2.2. Cell viability assay

Cell viability was assessed using the Cell Counting Kit-8 (KTA1020; Abbkine, Wuhan, China). Briefly, cells were seeded into 96-well plates at a density of 5000 cells per well. After 24-h treatment with lapatinib (0, 0.625, 1.25, 2.5, 5, 10, 20, 30 μM), the medium was replaced with 100 μL of fresh medium containing 10% CCK-8 reagent. Plates were incubated at 37 °C in the dark for 2 h, and the absorbance at 450 nm was measured using a microplate reader.

2.3. Live/dead staining

Cell viability was qualitatively assessed using a Calcein-AM/PI double-staining kit (G1707, Servicebio, Wuhan, China). Cells were seeded in 96-well plates (10,000 cells/well) overnight. Following 24-h treatment with lapatinib (0, 5, 8 μM), the staining solution was added and incubated at 37 °C for 30 min. Cells were visualized using an inverted fluorescence microscope; live cells (green, Calcein-AM) and dead cells (red, PI).

2.4. Colony formation assay

OS cells treated with lapatinib (0, 5, 8 μM) for 24 h were plated at low density in 6-well plates and cultured for 7–10 days under standard conditions, with medium refreshed every 2–3 days. Colonies were then fixed with 4% paraformaldehyde (P1110, Solarbio, Beijing, China), stained with 0.1% crystal violet (G1014, Servicebio, Wuhan, China), and imaged. Colonies containing more than 50 cells were counted.

2.5. Wound healing assay

Cells were seeded in 6-well plates (8 × 10⁵ cells/well) and grown to confluence. A straight scratch was created using a sterile 200 μL pipette tip. After washing with PBS, serum-free medium containing lapatinib (0, 5, 8 μM) was added. Images of the wound area were captured at 0, 12, and 24 h post-scratching. The migration rate was quantified by measuring the wound area.

2.6. Iron、GSH、 MDA、and Gln assays

Intracellular iron levels were measured using the FerroOrange fluorescent probe (Dojindo, Japan, Cat# F374) according to the manufacturer's protocol. Briefly, cells treated with lapatinib (0, 5, 8 μM) for 24 h were incubated with 1 μM FerroOrange in culture medium for 30 min at 37 °C, washed twice with PBS, and then analyzed by flow cytometry (Ex/Em = 488/585 nm).

Glutathione (GSH) and malondialdehyde (MDA) were quantified using the GSH Assay Kit (Beyotime, China, Cat# S0053) and the MDA Assay Kit (Beyotime, Cat# S0131), respectively. For GSH, cell lysates were mixed with DTNB and the absorbance at 412 nm was measured. For MDA, cell lysates were reacted with thiobarbituric acid at 95 °C for 30 min, and the absorbance at 532 nm was recorded.

Glutamine (Gln) levels were determined using the Glutamine Assay Kit (Abcam, UK, Cat# ab197011). Cells were lysed and deproteinized, followed by enzymatic reaction producing NADH, which was measured at 340 nm.

2.7. Flow cytometry

Apoptosis Analysis: After 24-h lapatinib (0, 5, 8 μM) treatment, cells were collected, stained using an Annexin V-FITC/PI Apoptosis Detection Kit (A211, Vazyme, Nanjing, China), U2OS and HOS cells (5 × 10⁵ cells/well) were seeded in 6-well plates, cultured overnight, and then treated with lapatinib for 24 h. After treatment, cells were harvested, washed with PBS, and stained with Annexin V-FITC and PI according to the manufacturer's protocol. For apoptosis analysis, cells were gated based on forward scatter (FSC) and side scatter (SSC) to exclude debris. Annexin V-FITC positive and PI positive populations were quantified using FlowJo software. A minimum of 10,000 events were acquired per sample.

Lipid ROS Detection: Lipid peroxidation was assessed using the C11-BODIPY 581/591 probe (S0043S, Beyotime, Beijing). Cells were seeded into a six-well plate at a density of 5 × 10⁵ cells per well and treated with lapatinib (5 μM), DFO (100 μM), or both. The cells were then incubated with 10 μM BODIPY 581/591 C11 at 37 °C for 30 min. Subsequently, the cells were washed with PBS, and lipid ROS levels were measured using a fluorescence microscope or flow cytometer. For lipid ROS detection, the mean fluorescence intensity of C11-BODIPY was measured, and the gating strategy was set based on untreated control cells.

2.8. Molecular docking

The 3D structure of Lapatinib was retrieved from PubChem. The crystal structure of SLC1A5 was obtained from the RCSB PDB. Water molecules and native ligands were removed using PyMOL. Molecular docking simulations were performed using AutoDockTools (version 1.5.7), and the resulting poses were visualized and analyzed using Discovery Studio Visualizer (version 2019).

2.9. Molecular dynamics simulation

Molecular dynamics simulations of the SLC1A5-Lapatinib complex were performed using the GROMACS 2025.2 package. Solvation was performed using the CHARMM36 force field and the TIP3P water model, with ions added to physiological concentration (150 mM NaCl). Following energy minimization and stepwise NVT/NPT equilibration under positional constraints (100 ps each), a 100 ns production simulation was conducted (with an integration step size of 2 fs). Trajectories were analyzed for root mean square deviation (RMSD), root mean square fluctuation (RMSF), radius of gyration (Rg), solvent accessible surface area (SASA), and intermolecular hydrogen bonds.

2.10. Protein-ligand complex prediction

The structure of the SLC1A5-Lapatinib complex was predicted using AlphaFold3. The predicted model was evaluated using the fraction ordered, has no clash, interface predicted TM-score (iptm), and predicted TM-score (ptm) metrics. A composite ranked score was calculated.

2.11. Cellular thermal shift (CETSA) assay

Cells treated with/without lapatinib (5 μM) were harvested and resuspended in PBS, aliquoted into PCR tubes, and heated at a gradient of temperatures (37, 40, 43, 46, 49, 52 °C) for 3 min, followed by cooling on ice. Lysis buffer was added, and samples were incubated on ice, followed by centrifugation. The supernatant was subjected to Western blot analysis to detect remaining SLC1A5 protein.

2.12. In vivo xenograft study

All animal procedures were approved by the Animal Ethics Committee of China Medical University (Approval No. CMU20250035). Four-week-old female BALB/c nude mice (n = 4 per group) were purchased from Beijing Huafukang Laboratory Animal Technology Co., Ltd. A xenograft model was established by subcutaneously injecting 5 × 10⁶ HOS cells (resuspended in 100 μL PBS) into the right flank of each mouse. When tumors reached approximately 100 mm³, mice were randomly assigned to three groups (n = 4/group) as follows: (1) Vehicle control: intraperitoneal saline (daily); (2) Lapatinib alone: intraperitoneal lapatinib (25 mg/kg, daily) [36]; (3) Lapatinib + DFO: intraperitoneal lapatinib (25 mg/kg, daily) + intraperitoneal DFO (10 mg/kg, daily) [37], [38]. All treatments were continued for 21 days. Tumor volumes and body weights were measured every 3 days using calipers and calculated as (length × width²)/2. At the endpoint, mice were euthanized by cervical dislocation. Tumors were excised and weighed, and organs were dissected. Tumors and blood samples were collected for further analysis.

2.13. Immunohistochemistry and HE staining

Immunohistochemical staining protocol: Tumor tissue was fixed with 4% formaldehyde solution; following tissue dehydration and clearing, the tissue was embedded in paraffin, and 5-μm-thick sections were prepared. The sections were oven-dried for 20 min; the dried sections were then dewaxed and rehydrated, followed by antigen retrieval. Treat the sections with 3% hydrogen peroxide solution for 10 min, followed by blocking with 5% BSA at room temperature for 30 min. Incubate with the primary antibody (Ki67, Servicebio, China) and secondary antibody, followed by development and counterstaining. Finally, dehydrate and mount the sections for storage and microscopic examination.

HE Staining: After dewaxing and rehydrating the baked sections, stain with hematoxylin and eosin. Proceed with dehydration and clearing. Remove the sections, allow them to dry thoroughly, mount them, and after air-drying, examine under a microscope and take photographs.

2.14. Western blot analysis

Protein expression levels were evaluated using Western blotting. U2OS and HOS cells treated with different groups were collected, and total protein was extracted using RIPA lysis buffer containing 1% PMSF. Protein concentrations were determined using a BCA protein assay kit, and the protein concentration of each sample was adjusted to 2 μg/μL with 5 × SDS loading buffer. The samples were then heated at 100 °C for 10 min and cooled for later use. Separating gel (10%) and stacking gel were prepared. An aliquot of 5-10 μL of each protein sample was loaded, corresponding to 10–20 μg of protein per lane. SDS-PAGE was performed, and the proteins were transferred onto PVDF membranes using a wet transfer system. The membranes were blocked with 5% non-fat milk at room temperature for 1 h, then incubated overnight at 4 °C with primary antibodies against GPX4 (1:5000, Abmart, T56959), SLC1A5 (1:2500, Selleck, F0619), and β-actin (1:10000, Servicebio, GB11001). After incubation, the membranes were washed three times with TBST (10 min each) and then incubated with HRP-conjugated secondary antibody (1:5000) for 60 min. The membranes were washed three times with TBST, and protein bands were visualized using an enhanced chemiluminescence (ECL) detection system. Finally, the relative protein expression levels were quantified using ImageJ software.

2.15. siRNA transfection

siRNAs targeting human SLC1A5 were purchased from GenePharma (Suzhou, China). The sequences were siRNA#1 (5′-GAAGCACAGAGCCTGAGTT-3′); siRNA#2 (5′-GCCUUGGCAAGUACAUUCUTT-3′); siRNA#3 (5′-GUCGACCAUAUCUCCUUGATT-3′). A negative control siRNA (5′-UUCUUCGAACGUGUCACGUTT-3′) was used as control. Cells were seeded in 6-well plates (5 × 10⁵ cells/well) and transfected with 50 nM siRNA using Lipofectamine 3000 (Invitrogen, USA) for 8 h. After 48 h of transfection, cells were treated with or without lapatinib (5 μM) for 24 h. Knockdown efficiency was confirmed by Western blot (Fig. 5J).

Fig. 5.

The SLC1A5-GPX4 axis mediates lapatinib-induced ferroptosis. (A-B) Western blot analysis of SLC1A5 protein expression in OS cells treated with Lap or Lap+DFO (100 μM). (C) Evaluation of the interaction between Lap and SLC1A5 using CETSA. (D) DARTS assay confirming the interaction between Lap and SLC1A5. (E) Viability assessment by calcein-AM (live, green)/PI (dead, red) staining in cells overexpressing SLC1A5. (F) Detection of intracellular Fe²⁺ by FerroOrange probe. (G-H) Measurements of Gln and GSH levels in OS cells under indicated conditions. (I) Western blot analysis of SLC1A5 and GPX4 expression following Lap treatment in SLC1A5-overexpressing cells. (J) Western blot analysis of SLC1A5 protein expression in OS cells transfected with control siRNA or siRNA SLC1A5, showing knockdown efficiency. (K) GPX4 protein expression after lapatinib treatment for 24 h in control and SLC1A5 knockdown cells. (L) Cell viability determined by CCK-8 assay. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

2.16. Statistical analysis

All quantitative experiments were performed with at least three independent replicates. Data are presented as mean ± standard deviation (SD). Statistical significance was determined using GraphPad Prism 9 software by unpaired two-tailed Student's t-test or one/two-way ANOVA with Tukey's post-hoc test. A P-value <0.05 was considered statistically significant.

3. Results

3.1. Lapatinib suppresses proliferation, migration, invasion and induces cell death

To evaluate the anti-OS activity of lapatinib, we first examined its effects on malignant phenotypes. Lapatinib significantly inhibited the viability of U2OS and HOS cells in a dose-dependent manner, with IC50 values of 6.26 μM and 8.3 μM, respectively (Fig. 1A-C). Bright-field microscopy revealed a decrease in cell number and morphological alterations indicative of cell death (Fig. 1D). Time-course experiments confirmed a time-dependent suppression of cell growth (Fig. 1E, F). The colony formation assay demonstrated that lapatinib substantially impaired the long-term proliferative capacity of OS cells (Fig. 1G, H). Wound healing and Transwell assays collectively showed that lapatinib effectively inhibited cell migration and invasion (Fig. 1I, J; Fig. S1A, B). Furthermore, Annexin V/PI staining revealed that lapatinib treatment significantly increased the rate of cell death (Fig. 1K, L). These results indicate that lapatinib effectively counteracts the malignant phenotype of OS cells in vitro.

Fig. 1.

Lapatinib suppresses proliferation, migration, invasion and induces cell death. (A) Chemical structure of Lapatinib (Lap). (B—C) Dose-response curves (IC50) of OS cell viability following 24-h Lap treatment, as determined by CCK-8 assay. (D) Phase-contrast micrographs showing morphological alterations in OS cells treated with indicated concentrations of Lap for 24 h. (E-F) Time-course analysis of OS cell viability upon Lap treatment. (G-H) Representative images and quantification of colony formation in OS cells treated with Lap. (I-J) Wound healing assay assessing the migratory capacity of OS cells after Lap treatment and corresponding quantitative analysis. (K-L) Flow cytometric analysis of apoptosis in OS cells stained with Annexin V-FITC/PI after 24-h Lap treatment and its quantification. **, p < 0.01; ***, p < 0.001.

3.2. Lapatinib-induced ferroptosis is rescued by DFO

To delineate the mode of cell death induced by lapatinib, we employed pharmacological inhibitors of different death pathways. The ferroptosis inhibitor deferoxamine (DFO, 100 μM), most effectively rescued the lapatinib-induced loss of cell viability, with a significantly higher rescue rate compared to the apoptosis inhibitor Z-VAD-FMK (50 μM) and the necroptosis inhibitor Nec-1 (50 μM) (Fig. 2A, B). Transmission electron microscopy (TEM) of lapatinib-treated cells revealed shrunken mitochondria with condensed membranes and diminished cristae—ultrastructural hallmarks of ferroptosis—which were ameliorated by DFO co-treatment (Fig. 2C). The clonogenic assay further confirmed that DFO rescued the anti-proliferative effect of lapatinib (Fig. 2D). Biochemically, lapatinib triggered characteristic ferroptotic events: intracellular Fe²⁺ accumulation (Fig. 2E), lipid ROS accumulation (Fig. 2F, G), elevated MDA levels (Fig. 2H), and GSH depletion (Fig. 2I). All these perturbations were reversible upon DFO administration. Together, these data establish ferroptosis as the predominant mechanism of lapatinib-induced cell death in OS, although apoptosis is also observed.

Fig. 2.

Lapatinib-induced ferroptosis is rescued by DFO. (A-B) CCK-8 assay evaluating the effects of Lap co-treatment with various inhibitors on cell viability, demonstrating the most significant rescue by the ferroptosis inhibitor (DFO (100 μM), Z-VAD(50 μM), Nec-1(50 μM)). (C) Representative TEM images of mitochondria. Red arrows highlight characteristic ferroptotic mitochondrial morphology, such as shrinkage and increased membrane density. (D) Colony formation assay of OS cells treated with Lap alone or in combination with DFO for 7 days. (E) Fluorescence imaging of intracellular Fe²⁺ using FerroOrange probe. (F-G) Flow cytometric analysis of lipid reactive oxygen species (ROS) using C11-BODIPY probe. (H—I) Measurements of malondialdehyde (MDA) and reduced glutathione (GSH) levels. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.3. Lapatinib induces ferroptosis by suppressing GPX4

We next investigated the molecular effector downstream of lapatinib. Western blot analysis showed that lapatinib treatment significantly suppressed the protein level of GPX4, a key guardian against ferroptosis, and this suppression was negated by DFO (Fig. 3A-D). To establish a causal link, we enforced GPX4 expression. GPX4 overexpression conferred significant protection against lapatinib-induced cell death (Fig. 3E, F), and effectively mitigated the surge in total ROS (Fig. 3G, H) and lipid peroxides (Fig. 3I, J). These results identify the suppression of GPX4 as a critical event in lapatinib-triggered ferroptosis.

Fig. 3.

Lapatinib induces ferroptosis by suppressing GPX4. (A-B) Western blot analysis of GPX4 protein expression in OS cells following Lap treatment. (C—D) Western blot analysis of GPX4 expression in OS cells treated with Lap and DFO (100 μM). (E-F) Assessment of cell death by propidium iodide (PI) staining following GPX4 overexpression. (G-H) Measurement of intracellular ROS levels using DCFH-DA probe. (I-J) Flow cytometric analysis of lipid ROS using C11-BODIPY probe.

3.4. Lapatinib interacts with SLC1A5

Given the established role of the glutamine transporter SLC1A5 in ferroptosis, we postulated it as a potential target of lapatinib. Molecular docking simulations positioned lapatinib snugly within the substrate-binding pocket of SLC1A5, with a favorable binding free energy of −8.6 kcal/mol and multiple specific hydrogen-bond interactions (Fig. 4A-D; Fig. S2). Subsequent molecular dynamics simulations attested to the complex's robustness: the system achieved stable RMSD values (Fig. 4E), exhibited low residual fluctuation (RMSF) at the binding interface (Fig. 4F), maintained a compact conformation (Rg, Fig. 4G), and sustained an average of four intermolecular hydrogen bonds throughout the simulation (Fig. 4I). The free energy landscape, characterized by a deep global minimum (Fig. 4J, K), reinforced the stability of the binding mode. Finally, AlphaFold3-predicted complex structures exhibited high local quality, were sterically plausible, and achieved a high composite confidence score (Fig. 4O, P). These in silico analyses support a potential interaction between lapatinib and SLC1A5, providing a computational basis for subsequent experimental validation.

Fig. 4.

Lapatinib interacts with SLC1A5. (A-D) Three-dimensional and two-dimensional representations of the predicted binding mode of Lap within the SLC1A5 binding pocket. (E) Time evolution of the root mean square deviation (RMSD) for the SLC1A5 protein backbone. (F) Root mean square fluctuation (RMSF) per residue of the SLC1A5 protein. (G) Time-dependent change in the radius of gyration (Rg) of the Lap-SLC1A5 complex. (H) Temporal profile of the solvent accessible surface area (SASA) for the complex. (I) Time evolution of the number of hydrogen bonds between SLC1A5 and Lap. (J) Three-dimensional free energy landscape of the complex. (K-L) MM-PBSA-based calculation of the binding free energy and its decomposition. (M) Electrostatic surface potential of the Lap-SLC1A5 complex. (N) Predicted structure of the SLC1A5-Lap complex generated by AlphaFold3. (O) Quality assessment radar chart for the AlphaFold3 prediction.

3.5. The SLC1A5-GPX4 axis mediates lapatinib-induced ferroptosis

We next validated the Lapatinib-SLC1A5 interaction experimentally. Lapatinib reduced SLC1A5 protein abundance, an effect reversible by DFO (Fig. 5A, B). Cellular Thermal Shift Assay (CETSA) demonstrated a pronounced stabilization of SLC1A5 against thermal denaturation in the presence of lapatinib (Fig. 5C). A physical interaction was further supported by the Drug Affinity Responsive Target Stability (DARTS) assay (Fig. 5D). Crucially, SLC1A5 overexpression effectively rescued cells from Lapatinib-induced cytotoxicity (Fig. 5E), iron overload (Fig. 5F), and GPX4 downregulation (Fig. 5I). To molecularly dissect the pathway, we interrogated the metabolic consequences. Lapatinib treatment profoundly depleted intracellular Gln and its downstream product, GSH. These deficits were rectified by restoring SLC1A5 function (Fig. 5G, H). These findings delineate a coherent signaling axis: lapatinib interacts with SLC1A5, thereby limiting Gln import and crippling GSH synthesis, leading to functional impairment of GPX4 and ferroptosis.

To further confirm that lapatinib's effects depend on SLC1A5, we knocked down SLC1A5 expression using siRNA (Fig. 5J). Western blot analysis confirmed efficient knockdown of SLC1A5 protein expression following siRNA transfection (Fig. 5J). As shown in Fig. 5K and L, SLC1A5 knockdown significantly attenuated lapatinib-induced GPX4 downregulation and partially rescued cell viability. These genetic loss-of-function data further support that SLC1A5 is a functional target of lapatinib.

3.6. Lapatinib suppresses tumor growth by activating ferroptosis in vivo

We established a xenograft mouse model to validate the in vivo efficacy of lapatinib (Fig. 6A). Lapatinib monotherapy significantly inhibited tumor growth, as reflected by reduced tumor volume and weight, an effect substantially attenuated by DFO co-treatment (Fig. 6B-E). Immunohistochemical analysis confirmed decreased proliferation (Ki-67 staining) in lapatinib-treated tumors (Fig. 6H). Critically, lapatinib reduced the protein levels of both GPX4 and SLC1A5 in tumor tissues, and these effects were reversed by DFO (Fig. 6I). Furthermore, plasma analysis from lapatinib-treated mice showed elevated MDA and decreased GSH, changes that were similarly normalized by DFO (Fig. 6J, K). These in vivo data robustly demonstrate that lapatinib constrains OS growth in association with activation of the ferroptosis pathway. (See Fig. 7.)

Fig. 6.

Lapatinib suppresses tumor growth by activating ferroptosis in vivo. (A) Experimental timeline for in vivo administration of Lap (25 mg/kg, i.p., daily) and DFO (10 mg/kg, i.p., daily). (B—C) Representative photographs of resected xenograft tumors from each group (n = 4) (Scale bar: 1 cm). (D-E) Tumor weight and volume at the study endpoint. (F-G) Curves depicting body weight changes and tumor growth over time. (H) Hematoxylin and eosin (H&E) staining and immunohistochemical (IHC) analysis of tumor sections (scale bar: 50 μm). (I) Western blot analysis of SLC1A5 and GPX4 protein levels in tumor tissues. (J-K) Measurement of GSH and MDA levels in mouse plasma.

Fig. 7.

The mechanism of Lapatinib-induced ferroptosis in osteosarcoma.

4. Discussion

This study elucidates a novel anti-cancer mechanism of lapatinib in OS, distinct from its canonical role as an EGFR/HER2 inhibitor. Our data consistently demonstrate that lapatinib induces ferroptosis by interacting with the Gln transporter SLC1A5. While lapatinib has been reported to influence ferroptosis in other cancer types (e.g., colon cancer, breast cancer), the underlying molecular mechanism remained unclear. Here, we provide the first evidence that SLC1A5 is a potential functional target of lapatinib, and that the SLC1A5-GPX4 axis is the critical mediator of lapatinib-induced ferroptosis in OS. The interaction of the drug with SLC1A5 initiates critical metabolic disruption, culminating in GSH depletion, subsequent inactivation of GPX4, and ultimately, lethal lipid peroxidation. This newly identified SLC1A5-GPX4 signaling axis underlies the potent anti-OS activity of lapatinib, both in vitro and in vivo. This finding not only expands the potential clinical applications of lapatinib but also establishes SLC1A5 as a promising novel therapeutic target in OS.

The core of our discovery lies in establishing the mechanistic link between SLC1A5 inhibition and ferroptosis activation. Ferroptosis is characterized by iron-driven accumulation of lipid peroxides, a process kept in check by the cellular antioxidant defense system, with GPX4 serving as its central guardian [39], [40]. We detected a potential interaction between lapatinib and SLC1A5 using multiple complementary techniques, including molecular docking, dynamics simulations, CETSA, and DARTS. As SLC1A5 is the primary transporter for glutamine (Gln), and glutamate (Glu) produced from Gln degradation serves as a key precursor for GSH synthesis [27], our functional experiments verified that its inhibition by lapatinib led to a significant reduction in intracellular Gln and GSH levels. The depletion of GSH, an essential cofactor for GPX4, incapacitates this key enzyme, thereby crippling the cell's ability to reduce lipid hydroperoxides and setting the stage for ferroptosis [41].

Although lapatinib is a well-known dual EGFR/HER2 tyrosine kinase inhibitor, several lines of evidence suggest that the ferroptosis-inducing effects observed in our study are independent of canonical EGFR/HER2 signaling. First, both U2OS and HOS osteosarcoma cell lines have been reported to express functional EGFR [42], [43]. However, previous studies have demonstrated that EGFR is not a major driver for osteosarcoma cell growth under standard culture conditions, but rather contributes to stress survival under starvation or chemotherapy-induced conditions [43]. Moreover, clinically achievable concentrations of the EGFR-specific inhibitor gefitinib showed only limited cytotoxic activity against osteosarcoma cells [44], suggesting that EGFR inhibition alone is insufficient to induce robust cell death. Second, previous studies have reported that lapatinib can exert anti-tumor effects through EGFR/HER2-independent off-target mechanisms in other cancer types. Dolloff et al. demonstrated that lapatinib enhances TRAIL sensitivity in colorectal cancer cells via JNK activation and DR5 up-regulation, independent of EGFR and HER2 inhibition [45]. Zhang et al. further showed that the in vitro activity of lapatinib is not dependent on EGFR expression level in HER2-overexpressing breast cancer cells [46]. Third, and most importantly, our comprehensive target identification studies support that SLC1A5 is a functional interactor of lapatinib. Molecular docking and molecular dynamics simulations predicted a stable interaction between lapatinib and the substrate-binding pocket of SLC1A5 (Fig. 4). This computational prediction was experimentally validated by CETSA and DARTS assays, which indicated potential physical engagement (Fig. 5C-D). Furthermore, our bidirectional genetic evidence confirmed the functional dependency: SLC1A5 overexpression rescued lapatinib-induced ferroptosis (Fig. 5E-I), while SLC1A5 knockdown attenuated lapatinib's effects on GPX4 expression and cell viability (Fig. 5J-L). Taken together, these data suggest that lapatinib induces ferroptosis in osteosarcoma through SLC1A5 targeting, independent of canonical EGFR/HER2 inhibition. We acknowledge that our conclusion regarding EGFR/HER2 independence relies primarily on literature evidence and indirect observations. Direct experimental comparison, such as genetic knockout of EGFR in OS cells followed by lapatinib treatment, would provide stronger support.

Taken together, while we acknowledge that lapatinib is an EGFR/HER2 inhibitor and that U2OS and HOS cells express functional EGFR, the preponderance of evidence—including target identification, bidirectional genetic validation, and literature support—indicates that the ferroptosis-inducing effects of lapatinib are mediated through interaction with SLC1A5 rather than through inhibition of EGFR/HER2 signaling.

Multiple rescue experiments in our study solidify this causal relationship. The ferroptosis inhibitor DFO consistently reversed lapatinib-induced cell death, lipid ROS production, and the downregulation of both SLC1A5 and GPX4. More compellingly, overexpression of either SLC1A5 or GPX4 alone was sufficient to mitigate the cytotoxic effects of lapatinib, confirming their pivotal roles within this pathway. Furthermore, we acknowledge that lapatinib also induced apoptosis in OS cells, as evidenced by Annexin V/PI staining (Fig. 1K, L). However, several lines of evidence support that ferroptosis is the predominant mechanism of lapatinib-induced cell death. First, DFO rescued viability to a significantly greater extent than apoptosis or necroptosis inhibitors (Fig. 2A, B). Second, TEM showed mitochondrial shrinkage and membrane condensation–hallmark features of ferroptosis, not apoptosis. Third, lapatinib triggered ferroptosis-specific biochemical events (iron, lipid ROS, GSH depletion), all reversed by DFO. Therefore, while apoptosis may contribute to lapatinib's anti-tumor effects, ferroptosis plays a dominant role. Our in vivo data further corroborated these findings, where lapatinib treatment not only significantly suppressed tumor growth but also elicited changes in established ferroptosis markers, effects that were similarly reversible by DFO co-administration.

Our findings carry significant implications for improving the therapeutic landscape of OS. The prognosis for patients, particularly those with metastatic or chemoresistant disease, remains poor, highlighting an urgent need for novel strategies [47], [48], [49], [50]. Inducing ferroptosis has emerged as a promising approach to eliminate cancer cells resistant to conventional apoptosis-inducing chemotherapy [14], [51]. In this context, our research positions lapatinib as a compelling candidate for drug repurposing. By revealing its capacity to trigger ferroptosis via SLC1A5, we provide a preclinical rationale for further evaluation of lapatinib repurposing in osteosarcoma.

Furthermore, the role of ferroptosis in modulating the tumor microenvironment and anti-tumor immunity adds another layer of therapeutic potential [52], [53]. Ferroptosis is recognized as an immunogenic cell death modality; dying cells release damage-associated molecular patterns that can stimulate dendritic cell maturation and enhance T-cell-mediated tumor killing [54]. Consequently, future studies should investigate whether Lapatinib-induced ferroptosis in OS can synergize with immune checkpoint inhibitors, potentially yielding a more powerful combination therapy.

Despite the compelling evidence, we acknowledge the limitations of our study. While our computational docking and molecular dynamics simulations suggest a stable binding mode between lapatinib and SLC1A5, these predictions require further experimental validation, such as co-crystallization or site-directed mutagenesis, to definitively confirm the binding interface. Although preliminary safety assessment showed no significant histopathological changes in major organs (heart, liver, lung, kidney) at the tested dose (Fig. S1G—H), this assessment was limited to a single dose and short treatment duration, and comprehensive toxicity and pharmacokinetic studies are needed before clinical translation. Another limitation of this study is the use of only two osteosarcoma cell lines (U2OS and HOS). While these are well-characterized and commonly used OS models, future studies should validate our findings in additional OS cell lines (e.g., MG63, Saos-2) and, ideally, in patient-derived xenograft (PDX) models to account for tumor heterogeneity. The current study is entirely preclinical and does not address pharmacokinetics, long-term toxicity, or therapeutic index. Therefore, cautious interpretation is warranted when considering clinical translation.

5. Conclusion

In conclusion, this study delineates a novel mechanism of lapatinib that operates independent of its canonical targets. Lapatinib interacts with SLC1A5, thereby disrupting glutamine metabolism and GSH synthesis, which inactivates GPX4 and ultimately triggers ferroptosis. Ferroptosis plays a dominant role in lapatinib-induced cell death, although apoptosis may also contribute. Collectively, our findings position the SLC1A5-GPX4 axis as a therapeutically targetable vulnerability in OS, providing a preclinical rationale for further evaluation of lapatinib repurposing.

CRediT authorship contribution statement

Haichuan Miao: Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Xijia Fu: Formal analysis, Data curation, Conceptualization. Baolong Liu: Formal analysis, Data curation, Conceptualization. Xinfang Pan: Methodology, Data curation. Yahao Han: Resources, Methodology, Data curation. Zhongshen Yu: Writing – review & editing, Validation, Supervision, Project administration.

Institutional review board statement

This research received approval from the Medical Animal Research Ethics Committee of China Medical University (CMU20250035).

Funding

This work was supported by the Natural Science Foundation of Liaoning Province (2024-MS-259).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material 1