GLMP promotes EGFR-TKI resistance by activating autophagy and RhoA pathway in non-small cell lung cancer Authors: Xiao Liang, Jiali Xu, Suhui Shu, Wei Lv, Dandan Yin, Sunan Miao, Xiyi Lu, Chen Zhang, Xinyin Liu, Jiali Dai, Jun Li, Weibing Wu, Erbao Zhang, Renhua Guo Published: 2025 ABSTRACT AbstractResistance to epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) restricts the long-term efficacy of treatment in patients with lung cancer. Here, we report that the EGFR-TKI resistance mechanism is mediated by lysosome-related regulation. The overexpression of glycosylated lysosomal membrane protein (GLMP), a lysosomal membrane protein, promotes resistance to Osimertinib both in vitro and in vivo. Mechanistically, GLMP could regulate the ubiquitination of RhoA and promote resistance by activating the epithelial-mesenchymal transition (EMT), which involves the RhoA pathway and activates the late stage of autophagy. Inhibition of the RhoA pathway alone enhances the initiation stage of autophagy. Lysosomal hyperactivity in TKI-resistant cells sustains the flow of autophagy. Therefore, the combined inhibition of the RhoA pathway and autophagy can effectively attenuate EGFR-TKI resistance. Our findings provide a potential therapeutic strategy to overcome resistance to EGFR-TKIs.Subject terms: Non-small-cell lung cancer, Oncogenes Introduction Lung cancer is one of the most common malignant tumors and a leading cause of cancer-related deaths worldwide1. Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs), such as Osimertinib, have significantly improved the prognosis of patients with lung cancer. Despite the initial effectiveness of EGFR-TKIs, resistance inevitably occurs, resulting in treatment failure and disease progression2–4. Hence, a comprehensive understanding of the mechanisms of EGFR-TKI resistance can provide greater benefits to patients. Lysosomes are essential organelles in eukaryotic cells that play key roles in multiple cellular processes, including signal transduction, endocytic degradation, and exocytosis5. Lysosomal membranes contain multiple membrane proteins that collectively regulate lysosomal functions6. Dysfunction of lysosomes can lead to multiple diseases, including cancer development7–9. However, the role of lysosomal hyperactivity in TKI resistance is not fully understood. Glycosylated lysosomal membrane protein (GLMP) is a lysosomal membrane protein previously reported to stabilize lysosomal membranes10,11. Several studies have found that GLMP could be involved in tumor progression12. However, its relationship with EGFR-TKI resistance remains unclear. Aberrant activation of epithelial-mesenchymal transition (EMT) and autophagy contributes to various pathological processes of tumor cells13. Previous studies have also identified EMT and autophagy as key factors in resistance to EGFR-TKIs12,14,15. In this study, we aimed to investigate new biomarkers and identify potential therapeutic targets for overcoming EGFR-TKI resistance. The evidence from in vitro and in vivo models confirmed the impact of GLMP on EGFR-TKI efficacy in lung cancer, which could provide a potential target for overcoming EGFR-TKI resistance in lung cancer. Results EGFR-TKI-resistant lung cancer cells exhibit a phenotype of lysosome hyperactivity To facilitate a comprehensive investigation of the key pathways and genes involved in EGFR-TKI resistance, the expression profiling data from four Osimertinib-resistant cell lines were downloaded from the GEO database (Fig. 1A). Functional enrichment analysis revealed that the upregulated genes were significantly associated with multiple lysosome-related pathways, including lysosome, autolysosome, autophagosome, lysosomal lumen and lysosomal membrane, suggesting a potential role of lysosomal function in Osimertinib resistance (Fig. 1B). EGFR-TKI-resistant cells exhibit a phenotype of lysosomal hyperactivity. A Differentially expressed genes were analyzed based on RNA-sequencing from four osimertinib-resistant cell lines. B Enrichment analysis of genes upregulated in the four Osimertinib-resistant cells. C Construction of drug-resistant cell lines. D IC50 of Osimertinib for Osimertinib-resistant cells and sensitive cells. E Immunofluorescence staining of sensitive and resistant cells with Lyso-Tracker Red and LysoSensor Green (scale bars: 10 μm). F The expression of lysosomal-related enzymes in Osimertinib-resistant cells and sensitive cells (two-tailed t-test, **p < 0.01). G Screening of potential core genes involved in Osimertinib-resistant cells. To further validate the relationship between lysosomal functional alterations and EGFR-TKI resistance, two Osimertinib-resistant cell lines (PC9OR and HCC827OR) based on PC9 and HCC827 were established. The detailed method for constructing the cell lines is shown in Fig. 1C. The IC50 values of Osimertinib in the two cell lines were assessed, revealing a substantial increase in the IC50 of the resistant cells (Fig. 1D). We subsequently assessed the lysosomal function of Osimertinib-resistant cells using Lyso-Tracker Red and LysoSensor Green, two different lysosomal probes. The fluorescence intensity of the two lysosomal probes in Osimertinib-resistant cells was higher than that in sensitive cells (Fig. 1E). PCR analysis of lysosomal enzyme expression also revealed the upregulation of CTSB, CTSD, and CTSL in resistant cells. These findings indicate that lysosomal hyperactivity is a prominent phenotype of EGFR-TKI-resistant cells (Fig. 1F). Furthermore, we aimed to screen potential core genes associated with lysosomes in EGFR-TKI resistance. 104 genes were selected across four Osimertinib-resistant cell line datasets (fold change >2, p value < 0.01) (Fig. 1A, G). Among these, 6 genes associated with lysosomes were identified, with GLMP being the only gene that was highly expressed in lung adenocarcinoma. Clinical relevance of GLMP We further investigated the clinical significance of GLMP in the tumorigenesis and drug resistance of lung cancer to EGFR-TKIs. The expression of GLMP could serve as a diagnostic biomarker for lung cancer in TCGA (Fig. 2A). After evaluating the expression of GLMP in vital organs and tumors, we found that GLMP expression exhibited strong tumor-specific characteristics (Supplementary Fig. S1A). We also observed the upregulation of GLMP in lung cancer across multiple GEO cohorts (Supplementary Fig. S1B–F). Higher expression of GLMP was also associated with more advanced stages of NSCLC (Fig. 2B). The survival analysis revealed that high GLMP expression was correlated with a poor prognosis for NSCLC patients (Fig. 2C). We further systematically analyzed the RNA expression of GLMP in different sensitive cell lines and resistant cell lines induced by the 1st generation (Gefitinib and Erlotinib) and 2nd generation (Dacomitinib) treatment and found that GLMP was also significantly upregulated in other EGFR-TKI resistant cells (Fig. 2D–F). We conducted an scRNA-seq analysis, which revealed the molecular profile of metastatic lung cancer biopsies before and after targeted therapy16, and found that the expression of GLMP in tumor cells was indeed increased over the course of Osimertinib treatment (Fig. 2G). We further validated the relationship between GLMP and TKI resistance by collecting a total of 40 pre-EGFR-TKI treatment biopsy samples (PTs) from patients with EGFR mutations, as well as 27 postresistance biopsy samples from patients with progressive disease (PD). The results of the qRT-PCR analysis showed that the expression of GLMP in the PD group was higher than that in the PT group (Fig. 2H). The ROC curve suggested that the GLMP expression can also serve as a predictive marker for drug resistance to EGFR-TKIs (Fig. 2H). The 40 patients in the NT group were subsequently stratified into a high group and a low group based on the expression of GLMP. Patients with high expression showed shorter progression-free survival (PFS) (Fig. 2I). We also collected three paired tissue samples before and after drug resistance (Supplementary Fig. S2) and validated the increased protein expression of GLMP in the drug-resistant samples through an IHC assay (Fig. 2J). Clinical relevance of GLMP. A The ROC curve of tumor tissues and adjacent non-cancerous tissues in lung cancer patients in TCGA. B Differential expression of GLMP in different pathological stages of patients with NSCLC in TCGA (wilcoxon test, *p < 0.05, **p < 0.01). C The correlations between the expression of GLMP and overall survival in Kaplan–Meier plotter database. D–F Comparison of GLMP expression between EGFR-TKIs-sensitive cells and EGFR-TKIs-resistant cells including first-generation (Gefitinib and Erlotinib) (D and E) and second-generation (Dacomitinib) (F) (two-tailed t-test, *p < 0.05, **p < 0.01). G The Single-cell sequencing revealed that GLMP expression increased with the progression of drug resistance to Osimertinib. H Comparison of the expression of GLMP between 40 pre-EGFR-TKI treatment (PT) biopsy samples and 27 post-resistance biopsy samples of patients with progressive disease (PD) and the ROC curve for GLMP in the PT and PD groups. (two-tailed t-test, **p < 0.01). I Kaplan–Meier survival curve for overall survival of patients stratified by the expression of GLMP. J Protein expression of GLMP in three patients (scale bars: 20 μm). Copy number amplification and TFEB transcriptional activation jointly contribute to the activation of GLMP in EGFR-TKI resistance Next, we detected the expression of GLMP in Osimertinib-resistant cells and sensitive cells by WB and confirmed that GLMP was highly expressed in drug-resistant cell lines (Fig. 3A). Upon further treatment of drug-resistant cells with 2 µmol of Osimertinib for 72 h, the expression of GLMP further increased (Fig. 3A). We subsequently explored the activation mechanisms of GLMP in EGFR-TKIs resistance. First, we analyzed the genomic alterations of GLMP in data from patients with tumors in TCGA and tumor cells in the CCLE database. GLMP exhibited a high frequency of copy number amplification in patients with lung cancer (Fig. 3B) and in cancer cells (Fig. 3C). The expression of GLMP was positively correlated with the copy number in lung cancer (Fig. 3D, E) and other types of malignant tumors (Supplementary Fig. S3). We then verified the copy number alterations of GLMP in sensitive and resistant cells and found that the copy number of GLMP was higher in resistant cells than in sensitive cells (Fig. 3F). Previous studies have revealed that the majority of lysosome-related genes are regulated by the microphthalmia (MiT/TFE) family, which comprises microphthalmia-associated transcription factor (MITF), transcription factor E3 (TFE3), transcription Factor EB (TFEB), and transcription factor EC (TFEC)17,18. These factors share a basic-helix-loop-helix-leucine zipper (bHLH-Zip) DNA binding motif and are known to bind to specific E-box sequences in the promoters of target genes (Fig. 3G). We further assessed the correlations between the expression of GLMP and these transcription factors. We individually knocked down these four transcription factors and assessed the expression of GLMP. As shown in Fig. 3H, knocking down TFEB expression downregulated GLMP expression in both the PC9OR and HCC827OR cell lines. Further experiments showed that knocking down TFEB could reduce the IC50 of Osimertinib in drug-resistant cells (p <0.05) suggesting that TFEB depletion partially attenuated resistance. (Fig. 3I). GEO data analysis confirmed that the overexpression of TFEB induced the expression of GLMP, whereas mutant TFEB had the opposite effect (Fig. 3J). Next, the luciferase reporter assay was performed, and the activity of the GLMP promoter in PC9OR cells was decreased in PC9OR cells with TFEB depletion (Fig. 3K). Furthermore, ChIP assays in PC9OR cells indicated that TFEB could directly bind the promoter region of GLMP and that TFEB knockdown could reduce this binding (Fig. 3L). Copy number amplification and TFEB transcriptional activation jointly contribute to the activation of GLMP in EGFR-TKIs resistance. A WB results of sensitive cells, Osimertinib-resistant cells, and Osimertinib-stimulated resistant cells (two-tailed t-test, **p < 0.01). Copy number variations (CNVs) of GLMP in patients (TCGA) (B) and cell lines (CCLE) (C). The correlations between the number of copies of GLMP and the expression of GLMP in lung cancer in TCGA (D) and CCLE (E). F CNVs between sensitive cells and Osimertinib-resistant cells (two-tailed t-test, **p < 0.01). G Structures of transcription factors in the microphthalmia (MiT/TFE) family. H Changes in the expression of GLMP after transcription factors of the MiT/TFE family were knocked down (two-tailed t-test, n.s. no significance, *p < 0.05, **p < 0.01). I Changes in the IC50 upon knockdown of TFEB in drug-resistant cell lines (two-tailed t-test, **p < 0.01). J GLMP expression after regulating the expression of TFEB in the GEO databases (two-tailed t-test, *p < 0.05, **p < 0.01). K Relative luciferase activity of the GLMP promoter in the PO9OR after TFEB was knocked down (two-tailed t-test, **p < 0.01). L ChIP assay in PC9OR (two-tailed t-test, **p < 0.01). GLMP promotes resistance to EGFR-TKIs in vitro Small interfering RNA (siRNA) and overexpression plasmids were used to regulate the expression of GLMP and explore its potential role in EGFR-TKI resistance (Fig. 4A). The IC50 values of Osimertinib were altered in response to the expression of GLMP in PC9OR (p <0.05) and HCC827OR (p <0.05) suggesting that GLMP depletion partially attenuated resistance (Fig. 4B). Two Gefitinib-resistant cell lines (PC9GR and HCC827GR) based on PC9 and HCC827 cells were also established (Supplementary Fig. S1G). Consistently, the IC50 values of Gefitinib also decreased after GLMP expression was knocked down in Gefitinib-resistant lines (Supplementary Fig. S1H). The role of GLMP in regulating Osimertinib resistance was further investigated by assessing its effects on cell proliferation through MTT, colony formation, and EdU assays (Fig. 4C, D and Supplementary Fig. S4A). Moreover, cellular apoptosis was induced in GLMP depletion-resistant cell lines treated with Osimertinib, as assessed by flow cytometry (Supplementary Fig. S4B). These results suggested that knocking down GLMP could also partially attenuated resistance to EGFR-TKIs in vitro by increasing apoptosis. GLMP promotes resistance to Osimertinib in vitro. A Western blot analysis to validate the efficiency of GLMP knockdown and overexpression. B Changes in IC50 of Osimertinib after regulating GLMP (two-tailed t-test, **p < 0.01). C The MTT assay was used to evaluate proliferation after regulating GLMP expression and Osimertinib treatment (two-tailed t-test, **p < 0.01). D Colony formation evaluating proliferation after regulating GLMP and Osimertinib treatment (two-tailed t-test, **p < 0.01). GLMP promotes resistance to EGFR-TKIs in vivo To further validate the relationship between GLMP and EGFR-TKIs resistance in vivo, PC9OR cells were injected into the left and right axillae of the nude mice and Osimertinib was administered every day. The procedures for the animal experiments were shown in Fig. 5A. After the mice were euthanized, the weights of tumors from each mouse were measured individually (Fig. 5B). The weights and volumes of the tumors in the sh-GLMP group were lower than those in the control group. Compared with those in the control group treated with Osimertinib, the weights and volumes of the tumors in the sh-GLMP group were further decreased after treatment of Osimertinib (Fig. 5B–D). Haematoxylin-eosin (HE) and immunohistochemical (IHC) staining for Ki67 were then performed to evaluate the proliferative activity of the tumors in these groups. The intensity of Ki67 staining was lower in the sh-GLMP group than in the control group. Compared with that in the control group treated with Osimertinib, the intensity of Ki67 staining was further decreased after treatment of Osimertinib and GLMP knockdown (Fig. 5E). GLMP regulates resistance to EGFR-TKIs in vivo. A PC9OR cells were injected into nude mice. The mice were treated with vehicle or Osimertinib (25 mg/kg) (n = 6). The procedures and protocols of mice were shown. B–D Comparison of tumor parameters between the sh-NC group and the sh-GLMP group (two-tailed t-test, *p < 0.05, **p < 0.01). E HE staining and immunohistochemical staining for Ki-67 expression in tumors (scale bars: 20 μm). GLMP regulates EGFR-TKI resistance by activating the RhoA pathway To further delineate the role of GLMP in the resistance to EGFR-TKIs, we performed RNA sequencing analysis after the knockdown of GLMP. The volcano plot of differentially expressed genes was shown in Fig. 6A. The gene enrichment analysis revealed that the downregulated genes were involved in a variety of pathways. The results showed that GLMP was associated with the actin pathway, GTPases, and autophagy (Fig. 6B). Moreover, the GSEA results revealed a strong correlation between GLMP expression and the RhoA pathway according to the RNA sequencing data and TCGA lung cancer data (Fig. 6C, D). RhoA, a member of the Rho family of small GTPases, plays a crucial role in actin cytoskeletal dynamics19. RhoA has been reported to be involved in various forms of drug resistance in cancers19,20. Therefore, we examined the effect of knocking down GLMP on RhoA. We found that when we knocked down GLMP, the RhoA protein levels decreased (Fig. 6E). Conversely, when we overexpressed GLMP, the RhoA protein level increased (Fig. 6E). A co-immunoprecipitation (Co-IP) assay confirmed the interaction between GLMP and RhoA. When RhoA-FLAG and GLMP-HA were overexpressed, GLMP was coprecipitated by RhoA, and conversely, RhoA was also coprecipitated by GLMP (Fig. 6F). The RNA-seq data in Fig. 6A revealed that RhoA mRNA expression was not regulated by GLMP (Fig. 6A). Therefore, we considered that posttranslational modifications occurred in RhoA. We examined the effect of GLMP on the stability of the RhoA protein and found that knocking down GLMP decreased the protein stability of RhoA, whereas the overexpression of GLMP reversed this change (Fig. 6G). RhoA has multiple ubiquitination modification sites and can potentially be modified by a variety of ubiquitin ligases21,22. Then, we examined the ubiquitination levels of RhoA, and observed that it increased after GLMP was knocked down, whereas the overexpression of GLMP had the opposite effect (Fig. 6H). Further experiments confirmed that RhoA played an important role in attenuating Osimertinib resistance (Supplementary Fig. S5A). We further clarified whether GLMP played a role in attenuating drug resistance by overexpressing GLMP followed RhoA knockdown (Supplementary Fig. S5B). The results indicated that knockdown of RhoA could attenuate the TKI resistance (Supplementary Fig. S5C) and proliferation phenotype (Supplementary Fig. S5D), which were induced by GLMP. GLMP regulates RhoA signaling through post-translational modulation. A Volcano plot of the RNA sequencing results following GLMP knockdown. B Gene enrichment analysis of genes that were downregulated upon GLMP knockdown. C GSEA analysis of differentially expressed genes after GLMP knockdown. D GSEA analysis of GLMP in lung cancer samples based on TCGA data. E Western blot analysis of RhoA protein expression upon GLMP modulation. F Co-immunoprecipitation (IP) and immunoblot analyses were performed to examine the interaction between GLMP and RhoA after overexpressing RhoA-FLAG and GLMP-HA. G Assessment of RhoA protein stability following GLMP modulation. The protein intensity was quantified by normalizing the target band intensity to that of the internal control (GAPDH), with the intensity at 0 h set as 1 for reference (two-tailed t-test, **p < 0.01). H Ubiquitination assay to evaluate the ubiquitination status of RhoA following GLMP modulation. GLMP regulates resistance by inhibiting RhoA-mediated EMT RhoA is a key regulator of actin cytoskeletal dynamics, and actin remodeling is a crucial component of the EMT23,24. The EMT could also contribute to the development of resistance to EGFR-TKIs15. Therefore, we hypothesize that GLMP may promote resistance by regulating the RhoA-mediated EMT. After knocking down the expression of GLMP, we found that E-cadherin expression increased, and N-cadherin expression decreased (Fig. 7A). When GLMP was overexpressed, the opposite regulatory trend was observed (Fig. 7A). IHC and immunofluorescence analyses also revealed that GLMP regulated the expression of E-cadherin and N-cadherin (Fig. 7B, C). The PC9OR cells stably transfected with sh-GLMP were injected into the tail vein of nude mice. After counting the metastatic nodules, we found that compared with the control treatment, GLMP knockdown significantly reduced the number of metastatic nodules (Fig. 7D). We subsequently selected the RhoA inhibitor Fasudil to overcome resistance to EGFR-TKIs through combination therapy with Fasudil and Osimertinib. We exposed PC9OR and HCC827OR cells to different concentrations of Fasudil and Osimertinib and assessed the synergistic effects using various models. We found that Fasudil could attenuate Osimertinib resistance at multiple concentrations (corresponding to the blue area of Fasudil in Fig. 7E). We further compared the difference in sensitization to Osimertinib caused by Fasudil between blank control drug-resistant cells (PC9OR and HCC827OR) and GLMP-depleted cells. Under GLMP depletion conditions, the sensitization effect of Fasudil on Osimertinib was weakened (Fig. 7F). These results suggested that the sensitization effect of Fasudil on Osimertinib depended on GLMP to some extent. Animal model experiments were subsequently conducted to explore the synergistic effects on long-term outcomes (Fig. 7G–J). The tumor weight and volume were significantly decreased in the mice in the groups treated with Fasudil and Osimertinib. During the first 15 days of the experiment, the combination therapy group effectively controlled the tumor development, and the tumors appeared to stop progressing. However, in the last few days of the experiment, the tumors continued to progress (Fig. 7H). Through the experiment, significant weight loss was not observed in the mice treated with Fasudil or Osimertinib, indicating that the toxicity of the drug was relatively low (Fig. 7J). GLMP regulates resistance by inhibiting the RhoA-mediated EMT. A WB analysis of changes in the expression of EMT markers (N-cadherin and E-cadherin) after GLMP modulation. B IHC for EMT markers after tumorigenesis in sh-NC cells and sh-GLMP cells (scale bars: 25 μm). C Changes in immunofluorescence for E-cadherin after GLMP knock down (scale bars: 20 μm). D Differences in tail vein tumorigenesis between sh-NC cells and sh-GLMP cells after the tail vein injection of PC9OR (two-tailed t-test, **p < 0.01). E Synergy scores of the combined treatment based on the HSA model and the LOEWE model are shown. F Evaluating the effect of Fasudil on sensitizing cell lines with and without GLMP knockdown to Osimertinib (one-way ANOVA, n.s. no significance, *p < 0.05, **p < 0.01). G–I Mice (n = 5) were subjected to different treatments (Fasudil: 50 mg/kg; Osimertinib 25 mg/kg). The tumor weight (I) (one-way ANOVA, n.s. no significance, *p < 0.05, **p < 0.01) and volume (H) (one-way ANOVA, n.s. no significance, *p < 0.05, **p < 0.01) were recorded. J Mouse weight monitored during treatment. Simultaneous inhibiting autophagy and EMT attenuates GLMP-mediated EGFR-TKI resistance We found that the combined effect was significant in the early stage but diminished over time intriguing. GLMP is localized in lysosomes11. Through an enrichment analysis, we found that the function of GLMP may be related to autophagy. We considered whether lysosomal autophagy also plays a role in EGFR-TKI resistance. Fasudil can occasionally be used to treat cardiovascular diseases and acts as a Rho-kinase inhibitor25. Fasudil has been reported to potentially have a cardioprotective effect by enhancing autophagy26. However, whether it regulates autophagy in tumors, especially in lung cancer, has not been clearly reported. Therefore, we separately knocked down RhoA. and administered Fasudil, and then assessed the expression of the autophagy markers P62 and LC3. We observed an increase in the LC3-II/I ratio and a decrease in P62 levels (Fig. 8A), indicating that autophagy was increased. When we knocked down GLMP, increases in both the LC3-II/I ratio and P62 levels were observed (Fig. 8B). Additionally, we examined the changes in autophagy after treatment of Osimertinib and found that the autophagy was also increased (Fig. 8C). Then, we used mCherry-GFP-LC3 to detect autophagy (Fig. 8D). Immunofluorescence results revealed that knocking down GLMP increased the accumulation of yellow LC3 puncta and that knocking down of RhoA only increased the accumulation of red LC3 puncta (Fig. 8E). These results indicate that GLMP can regulate the late stage of autophagy by influencing lysosomal function, whereas RhoA primarily affects the early stage of autophagy. This result explains why GLMP and RhoA, both of which regulate EMT signaling pathways, play different roles in autophagy. Compared with sensitive cells, the resistant cells inherently exhibit a phenotype of lysosomal hyperactivity compared to the sensitive cells. Lysosomes are the ultimate executors of the autophagic flux. We further compared autophagic flux dynamics between Osimertinib-resistant and sensitive cells using mRFP-GFP-LC3 dual-fluorescence assays. After stimulating the cells with EBSS, a common activator of autophagy, dual fluorescence detection revealed that the drug-resistant cells presented more red fluorescence spots and fewer yellow fluorescence spots than the sensitive cells (Supplementary Fig. S6A). These findings suggested that Osimertinib-resistant cells presented a sustained autophagic flux. The simultaneous inhibition of autophagy and EMT increases the vulnerability of EGFR-TKI-resistant cells. A WB analysis of changes in the expression of autophagy markers (LC3 and P62) in WB after the modulation of RhoA expression. B WB analysis of changes in the expression of autophagy markers after the modulation of GLMP expression. C WB analysis of changes in the expression of autophagy markers after treatment with Osimertinib. D The mechanism of autophagy detection using mCherry-GFP-LC3. E Changes in the immunofluorescence of mCherry-GFP-LC3 after the modulation of RhoA and GLMP expression (n = 6; two-tailed t-test, *p < 0.05, **p < 0.01; scale bars: 50 μm). F Changes in the IC50 of Osimertinib after inhibiting autophagy by Chloroquine (two-tailed t-test, **p < 0.01). G Mice (n = 5) were subjected to different treatments (Fasudil: 25 mg/kg; Chloroquine 25 mg/kg) in combination with Osimertinib (25 mg/kg) (one-way ANOVA, **p < 0.01). The tumor volume during treatment was recorded. H Mice (n = 5) were subjected to different treatments (Fasudil: 25 mg/kg; Chloroquine 25 mg/kg) in combination with Osimertinib (25 mg/kg) in GLMP-depleted cell lines. I Mechanistic diagram of GLMP-mediated Osimertinib resistance. Our further experiments also confirmed that Chloroquine, an autophagy inhibitor, could also attenuate resistance to Osimertinib to some extent (Fig. 8F). Following Chloroquine treatment, further treatment with fasudil further reduced the IC50 of Osimertinib (Supplementary Fig. S6B). Lower concentrations of Fasudil (25 mg/kg) and Chloroquine (25 mg/kg) were chosen for the animal experiments to better observe the sensitizing effect of the combination of Fasudil and Chloroquine on Osimertinib. Through animal experiments, we also evaluated the therapeutic efficacy of simultaneously inhibiting autophagy and the EMT to attenuate EGFR-TKI resistance. Compared with the administration of Fasudil or Chloroquine alone, the combined treatment with Fasudil and Chloroquine further attenuated resistance to Osimertinib (Fig. 8G). The tumor volume and weight in the mice from the combination therapy group were significantly smaller than those in the single-drug group (Fig. 8G, Supplementary Fig. S6C), and a significant decrease in body weight was not observed, indicating that the combined therapy attenuated drug resistance without significantly increasing toxicity (Supplementary Fig. S6D). Furthermore, we applied a combination of a RhoA inhibitor and an autophagy inhibitor in the GLMP-depleted cell to determine whether this benefit depended on GLMP. In GLMP-depleted cells, the combination of the RhoA inhibitor and the autophagy inhibitor with Osimertinib did not further reduce the tumor volume and weight significantly (Fig. 8H, Supplementary Fig. S6E, F). Discussion The lysosome is a dynamic intracellular organelle that is present in all eukaryotic cells. Its sizes, shapes, and functions vary significantly depending on the cell type to allow the cell to adapt to different locations and environments27. In this study, we observed a correlation between lysosomal hyperactivity and EGFR-TKIs resistance in lung cancer. Through further evaluation of lysosome-related genes, we found that GLMP is highly expressed in drug-resistant cells and is involved in the resistance phenotype. GLMP is a lysosomal membrane protein that was previously reported to stabilize lysosomal membranes. GLMP is also involved in fatty acid metabolism, glucose utilization, and dysregulation in the liver28,29. Our research was the first to discover that GLMP can promote EGFR-TKI resistance in lung cancer by simultaneously inducing autophagy and promoting the EMT, revealing a new function and mechanism of GLMP in lung cancer drug resistance. Because the understanding of lysosomes has drastically changed, they are no longer considered static degradative compartments but are now recognized as dynamic regulators of cellular homeostasis that adapt to changes in the cellular environment. Previous studies have shown that hyperactivated lysosomes are associated with increased tumor malignancy27. Lysosome-related inflammatory responses can drive proliferation and angiogenesis. Lysosome-derived cathepsins can affect the EMT of cancers and are involved in metastasis30. Lysosomes also play an important role in autophagy by serving as the endpoint for the degradation of autophagosomes31. Previous studies have explored the relationship between autophagy and EGFR-TKIs. Inhibiting autophagy can, to some extent, sensitize cells to TKIs14,32–34. In a phase I clinical trial, the combination of hydroxychloroquine and erlotinib was well tolerated in patients with EGFR-mutant non-small cell lung cancer, with an objective response rate of approximately 5% and a disease control rate of 26%35. These findings suggest that combining EGFR-TKIs with autophagy inhibition may hold some clinical promise. However, due to its overall suboptimal efficacy, this strategy has not been further pursued in clinical development, thereby underscoring the need for deeper mechanistic investigation and the design of more effective combinatorial approaches. The application of ROCK inhibitors in cancer therapy is also constrained by multiple factors, and existing clinical studies have all ended in failure36. Previous studies have indicated that combining low-dose ROCK inhibitors with other therapeutic approaches is an important strategy to enhance efficacy36. In this context, our study reveals that simultaneous inhibition of autophagy and RHOA signaling holds promise as a more effective therapeutic strategy. The MiT/TFE family of transcription factors has been shown to regulate lysosomal functions at the transcriptional level18,37. TFEB, the first identified member of the MiT/TFE family, was reported to promote the transcriptional activation of lysosome-related genes, including lysosomal hydrolases and lysosomal transmembrane proteins38. TFEB-mediated lysosomal biogenesis is involved in drug resistance39. In our research, we discovered that GLMP was transcriptionally activated by TFEB. This finding further expands the regulation of lysosomes mediated by TFEB. On the other hand, GLMP was found to exhibit a high frequency of copy number amplification in multiple malignant cancers. Genomic variation is a potential factor affecting the prognosis of patients receiving EGFR-TKI treatment40,41. Our research indicated that EGFR-TKIs resistant cells exhibit an increased copy number amplification of GLMP. The expression of GLMP was upregulated in samples with CNA. The CNA of GLMP has the potential to be an important biomarker and target for EGFR-TKI resistance, but more clinical data are needed to clarify this phenomenon. Taken together, these results suggest that TFEB, along with genomic copy number amplifications, are responsible for the activation of GLMP in EGFR-TKI-resistant lung cancer. In this study, we also revealed a novel RhoA-induced EMT pathway that is mediated by GLMP. Extensive evidence suggests that overall changes in the contractility of the actin cytoskeleton are among the primary driving forces for the EMT. RhoA plays a central role in this regulation24,42. Our findings indicated a synergistic sensitizing effect of combined treatment with the RhoA inhibition Fasudil and Osimertinib. By inhibiting RhoA, the simultaneous inhibition of autophagy effectively attenuated the resistance mediated by GLMP. The regulation of RhoA and treatment of Osimertinib promotes autophagy, whereas Osimertinib-resistant cells with lysosomal hyperactivity sustain a more unimpeded autophagic flux. When GLMP is inhibited, this potential drug-induced autophagy does not occur through the inhibition of autophagolysosome formation. Simultaneously, inhibiting autophagy and the EMT could become a promising approach for treating GLMP-mediated TKI resistance (Fig. 8I). Although our study shows that knockdown of GLMP can partially alleviate drug resistance, complete restoration of drug sensitivity remains unattainable. Given that EGFR-TKI resistance in lung cancer is a complex regulatory process involving multiple genes and multiple levels, further research is needed to clarify the entire resistance network in order to achieve more significant alleviation. In summary, we demonstrated that gene amplification-driven and transcription-activated GLMP participated in the EMT and autophagy to promote Osimertinib resistance. GLMP may serve as a novel biomarker for resistance and a potential target for reversing resistance. Methods Clinical samples and clinical information The sequencing data and clinical information of patients were obtained from TCGA dataset (https://portal.gdc.com). Data without corresponding clinical information were discarded. The sequencing data of the cell lines were obtained from the Cancer Cell Line Encyclopedia (CCLE) database (https://sites.broadinstitute.org/ccle/datasets). The RNA sequencing data of resistant cells were obtained from the GEO database (GSE272162, GSE262582, and GSE202859). The single-cell RNA sequencing data from a patient treated with Osimertinib were obtained from the NCBI BioProject (PRJNA591860)16. Clinical samples were obtained from Jiangsu Provincial People’s Hospital. The study was approved by the Ethics Committee of Jiangsu Provincial People’s Hospital and was performed in accordance with the Declaration of Helsinki. All patients provided written informed consent before participating. The approval number is 2020-SR-279-1. Cell culture and establishment of EGFR-TKI-resistant cell lines The PC9 and HCC827 cell lines used in the study were purchased from the Cell Library of the Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences (Shanghai, China). Both PC9 and HCC827 are cell lines that carry EGFR mutations. The PC9 cell line was cultured in DMEM, and HCC827 cells were cultured in RPMI 1640. EGFR-TKI-resistant cell lines were generated by exposing them to an increasing gradient of concentrations of the drug. The detailed methodology is shown in Supplementary Fig. S1B. Cell transfection Lipofectamine 2000 (Invitrogen) was used to transfect specific small interfering RNAs (siRNAs). siRNAs targeting the desired gene and a nontargeting control siRNA were designed and synthesized. X-tremeGENE™ HP DNA Transfection Reagent (Roche) was used to transfect the plasmids into the cells. Lentiviruses encoding shRNAs targeting GLMP were used for in vivo experiments. Cell Cloning Approximately 300–500 cells were plated. After 2 days of culture, the cells were treated with the corresponding drugs. Following a growth period of 7–14 days, when the colonies reached an appropriate size, they were fixed with paraformaldehyde and subsequently stained with crystal violet for visualization and quantification. Western blot assay and antibodies The protein lysates were separated by SDS–PAGE and transferred to nitrocellulose membranes (Sigma‒Aldrich). Antibodies used for WB were as follows: GLMP (Bioss, bs-15078R); RhoA (Proteintech, 10749-1-AP); E-cadherin (Proteintech, 20874-1-AP); N-cadherin (Proteintech, 22018-1-AP); P62 (Proteintech, 18420-1-AP); LC3B (Proteintech, 18725-1-AP); HA Tag (Proteintech, 66006-2-Ig); FLAG (Proteintech, 66008-4-Ig); GAPDH (Abclonal, AC001) and VeriBlot for IP Detection (ab131366). RNA isolation and qRT-PCR RNA was extracted by TRIzol reagent (Invitrogen) from cells of tissues from lung cancer and reverse-transcribed into complementary DNA (cDNA). The qRT-PCR was performed with SYBR Green, cDNA, and primers. Specific primer sequences are shown in Supplementary Table 1. MTT proliferation assays The cells were incubated in plates with 20 μl MTT (Roche, USA) for 4 h, and the crystals were dissolved with 150 μl dimethyl sulfoxide for 10 min. The plates then were measured at 490 nm using a spectrophotometer. 5-Ethynyl-2′-deoxyuridine (EdU) assay The treated cells were cultured in 96-well plates and 50 mΜ EdU (RiboBio, China) labeling medium was added. Then, the cells were treated with 4% formaldehyde and 0.5% Triton X-100. DAPI was added to stain the nucleus. Finally, fluorescence microscopy was used to calculate the number of EdU-positive cells. IC50 assays Approximately 3000 cells were seeded into a 96-well plate. After 24 h of culture, different concentrations of the drugs were added. Following 72 h of drug treatment, cell viability was assessed by measuring the absorbance in each well using the CCK-8 assay. To perform the CCK-8 assay, a 10% CCK-8 working solution was prepared by diluting the CCK-8 reagent in the corresponding culture medium. Then, 100 µL of the 10% CCK-8 working solution was added to each well, followed by a 2-hour incubation at 37°C. The absorbance was then measured at 450 nm using a microplate reader. The IC50 value, representing the concentration of the compound that inhibits 50% of cell viability, was determined using GraphPad Prism software. Three independent experiments were compared using a t-test. mCherry-GFP-LC3 for autophagy detection The mCherry-GFP-LC3 kit (C3011) was purchased from Beyotime. This recombinant adenovirus effectively expresses a fusion protein of red fluorescent protein (mCherry), green fluorescent protein (GFP), and LC3 in target cells, resulting in bright red and green fluorescence for autophagy detection. GFP is an acid-sensitive protein, whereas mCherry is a stable fluorescence marker that is unaffected by external factors. Due to the acidic environment inside the lysosome, GFP fluorescence is quenched. Since mRFP is stably expressed, the ratio of GFP to mRFP fluorescence intensity can be used to evaluate the autophagic flux. When the cell confluence reached approximately 50%, the cells were transfected with mCherry-GFP-LC3 at an MOI of 5, and then stimulated the cells with different treatments after 48 h. Finally, the fluorescence intensity was measured to assess changes in autophagy. Assessment of lysosomal function Lysosomal function was assessed using Lyso-Tracker Red (Beyotime, C1046) and LysoSensor Green (Yeasen, 40767ES50). A 50 nM solution of Lyso-Tracker Red was prepared in the appropriate culture medium and prewarmed to 37 °C, after which the cells were incubated with the dye for 10 min. Following the incubation, fluorescence images were captured using a fluorescence microscope. Similarly, a 2 µM solution of LysoSensor Green was prepared in the corresponding culture medium and pre-warmed to 37 °C, after which the cells were incubated with the dye for 30 min. Fluorescence images were subsequently obtained using a fluorescence microscope. Analysis of the efficacy for single drugs and combination treatments A total of 3000 cells were seeded into a 96-well plate. After 24 h of incubation, different concentrations of the drugs were added. Following a 72-h drug treatment, the cell survival rates treated with each concentration were assessed using the MTT assay. The combined efficacy of drugs was assessed by Combenefit, a platform that is frequently utilized to assess synergy scores in published studies43,44. Compounds used were as follows: Osimertinib (MCE, HY-15772), Fasudil (MCE, HY-10341), and Chloroquine phosphate (MCE, HY-17589). Chromatin immunoprecipitation assay A chromatin Immunoprecipitation (CHIP) assay was performed using the WAEZ-CHIP KIT (Millipore, USA) according to the manufacturer’s instructions. TFEB (CST, 37785) antibody was purchased from Cell Signaling Technology. The quantification of immunoprecipitated DNA was performed using quantitative PCR (qPCR) with SYBR Green Mix (Takara). Dual-luciferase reporter assays GLMP promoter-driven luciferase plasmids and Renilla luciferase vectors were cotransfected into different cells. A dual-luciferase assay was performed 24 h after transfection by using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA). Mouse model Five-week-old athymic BALB/c mice were used. The cells were injected into the mice in a 100 μL volume of PBS. The treatment was started on Day 9 after subcutaneous tumor cells injection, when a noticeable hard mass could be palpated, and the tumor had significantly increased in size compared with earlier stages. The volumes of the tumors and the weights of the mice were examined every 3 days. The tumor volume was measured as follows: volume = length × width2 × 0.5. In the experiment assessing the impact of GLMP on Osimertinib in vivo, PC9OR cells and stable GLMP knockdown cells were injected into the left and right axillae of athymic BALB/c mice, with six mice per group, for a total of twelve mice in the two experimental groups, to assess the effect of GLMP on Osimertinib resistance in vivo. PC9OR cells and stable GLMP knockdown cells were injected into the tail veins of the athymic BALB/c mice to induce metastasis to investigate the role of GLMP in tumor metastasis. At the conclusion of the experiment, the number of lung metastases was quantified. In the experiment evaluating the combination of inhibitors with Osimertinib, the mice were divided into groups according to the treatment plan, with 5 mice in each group. All in vivo studies were performed according to the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The study protocol was approved by the Committee on the Ethics of Animal Experiments of Nanjing Medical University. Transcriptome sequencing Total RNA was isolated from GLMP knockdown and control PC9OR cells. The quality of each sample was evaluated by an Agilent 2200 instrument (Agilent, Palo Alto, CA, USA). The sequencing library of each RNA sample was prepared using the Ion Proton Total RNA-Seq Kit version 2 based on the corresponding protocol. Bioinformatics data analysis Copy number variations were estimated by using the GISTIC2 method. The copy number data of GLMP for patients in TCGA were obtained from UCSC (https://xena.ucsc.edu). The copy number data for GLMP in cells were obtained from the CCLE database. The copy numbers for GLMP in resistant cells were determined using the threshold cycle values from the respective standard curves. The dataset of Osimertinib-resistant samples was analyzed for differentially expressed genes using the GEO2R tool. Statistical analysis Statistical analysis was performed by GraphPad Prism 7, SPSS 23, and the R 3.5.1 package. Correlation analysis was performed by the Spearman test. Survival analysis was performed by the Log-rank test and Kaplan–Meier survival curves. Differences between groups were performed by the Wilcoxon test, t-test, and one-way ANOVA. All cell experiments were independently performed in triplicate. p < 0.05 was considered statistically significant. Supplementary information Supplementary Information Acknowledgements This work was supported by grants from the National Natural Science Foundation of China (82272669 and 82203010), the Cancer Foundation of China (CFCSSSQ017), Natural Science Foundation of Jiangsu Province (Grants No BK20250029 and BK20250318), the Project of Xuzhou Medical University (XYFY202432) and the Project of Wuxi Health Commission (Q202417). Author contributions X.L., J.X., S.S., W.L., D.Y., S.M., X.L., C.Z., and X.L. got the data from experiments. J.D. and J.L. wrote the manuscript. W.W., E.Z., and R.G. conceived and designed the studies. All authors have approved the final manuscript. Data availability The RNA-seq data in this study have been deposited in GEO under accession number GSE298111. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Xiao Liang, Jiali Xu, Suhui Shu, Wei Lv, Dandan Yin. Contributor Information Weibing Wu, Email: wuweibing95@njmu.edu.cn. Erbao Zhang, Email: erbaozhang@njmu.edu.cn. Renhua Guo, Email: rhguo@njmu.edu.cn. Supplementary information The online version contains supplementary material available at 10.1038/s41698-025-01135-w. References Associated Data Supplementary Materials Supplementary Information Data Availability Statement The RNA-seq data in this study have been deposited in GEO under accession number GSE298111.