ARL13B is regulated by the ERK/P90 pathway and mediates TMZ resistance in glioblastoma via microvesicles Authors: Haichao Xing, Ying Chen, Haolin Li, Mingjie Wang, Yujing Cao, Xiaomin Liu Published: 2026 ABSTRACT AbstractTemozolomide (TMZ) resistance presents a serious challenge in glioma treatment. It has been revealed that high expression of ADP-ribosylation factor-like GTPase 13B (ARL13B) is an important risk factor for glioma patients and is strongly related to RHO GTPase. Here, we presented a unique mechanism by which ARL13B induces TMZ resistance. The TCGA and CGGA datasets were used to conduct survival analysis. Metascape was utilized to perform pathway enrichment analysis. Western blot (WB), nanoparticle tracking analysis (NTA), and liquid chromatography-mass spectrometry (LC-MS) were used to explore the mechanisms of drug vesicular exocytosis that leads to ARL13B-mediated TMZ resistance. WB, qRT-PCR, and ChIP were used to validate the extracellular-regulated kinase (ERK)/P90 transcriptional regulatory signaling pathways. The expression of ARL13B increased ectosome efflux and decreased intracellular drug concentration. Besides, the ERK/P90 pathway could transcriptionally regulate ARL13B with C/EBPβ being one of the transcription factors of ARL13B. Inhibiting the ERK/P90 pathway could enhance the efficacy of TMZ. ARL13B was regulated by the ERK/P90 pathway and might impair the efficacy of TMZ by increasing drug efflux through ectosomes. Therefore, our study provided a novel mechanism of TMZ resistance and identifies a new target for glioma therapy.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1038/s41598-026-58241-3.Subject terms: Cancer, Cell biology, Drug discovery, Molecular biology Introduction Glioblastoma multiforme (GBM) is the most aggressive primary central nervous system tumor1. There were 87,053 deaths attributed to malignant brain and other CNS tumors between 2017 and 2021. This represents an average annual mortality rate of 4.41 per 100,000 population and an average of 17,411 deaths per year. The five-year relative survival rate following diagnosis of a malignant brain or other CNS tumor was 35.7%. For a non-malignant brain or other CNS tumor the five-year relative survival rate was 92.0%2. GBM treatment after diagnosis includes maximal surgery followed by temozolomide (TMZ) and radiation. However, resistance to TMZ, the primary alkylating agent used in GBM treatment, presents a major therapeutic challenge, as evidenced by the lack of response in at least 50% of patients3,4. Extracellular vesicles (EVs) can be classified based on their production method into the following categories: exosomes, classical microvesicles, oncosomes, arrestin-domain-containing protein 1 (ARRDC1)-mediated microvesicles (ARMMs) and apoptotic bodies5. It has been demonstrated that EVs can transfer proteins, nucleic acids and metabolites associated with drug resistance to cancer cells, thereby inducing drug resistance. What’s more, EVs have also been shown to package and sequester drugs out of the cells, resulting in the development of drug resistance6. GBM-derived EVs can regulate tumor progression by transporting cargo including proteins, mRNAs and miRNAs, which drive tumor bioprocesses, such as invasion, proliferation and angiogenesis. Furthermore, GBM-derived EVs have been demonstrated to facilitate the transmission of drug resistance. However, the specific contribution of GBM-derived EVs remains to be elucidated7. ADP-ribosylation factor-like GTPase 13B (ARL13B) is a member of the RHO-GTPase family, and the protein is found in high concentration in the ciliary membrane8. It has been reported that ARL13B mutation is strongly associated with Joubert syndrome9. Recently, the implication of ARL13B in cancer especially in glioma has been validated10–14. Several studies report elevated ARL13B expression in glioma cell lines12 and demonstrate that ARL13B knockdown reduces TMZ resistance13. Furthermore, ARL13B regulates membrane trafficking, which is a process linked to the actin cytoskeleton15. However, the role of ARL13B in EVs remains unclear, though it may be functionally linked to drug resistance. Our bioinformatic analysis suggested that ARL13B potentially mediates EV-mediated efflux processes. In our study, we discovered that ARL13B influences the exocytosis of extracellular vesicles, thereby enhancing drug efflux and contributing to heightened drug resistance in glioma cells. Additionally, we observed that TMZ upregulates the expression of ARL13B by stimulating the ERK/P90 pathway, with ARL13B being transcriptionally regulated by the transcription factor C/EBPβ. In conclusion, our findings reveal a novel function of ARL13B in TMZ resistance and its new regulation pathway triggered by TMZ. Methods Data download and analysis The Cancer Genome Atlas (TCGA) (https://xenabrowser.net/datapages/) and Chinese Glioma Genome Atlas (CGGA) (http://www.cgga.org.cn/) databases were used for survival analysis and ARL13B expression level comparison. Metascape (https://metascape.org/gp/index.html) was used for pathway enrichment analysis, utilizing data obtained from the BioGRID database (https://thebiogrid.org/). Cell culture TBD0220 and LN18 cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% fetal bovine serum (FBS). TBD0220 as a glioma cell line, was derived from a GBM patient who underwent surgery at Hebei University Affiliated Hospital16. Human glioma cell lines used in this study were obtained from authenticated sources and used in accordance with relevant ethical guidelines16. Ethical Statement for the human-derived cell line used in this study: The human glioma cell line TBD0220 was originally established from a glioblastoma patient who underwent surgical resection at Hebei University Affiliated Hospital. The collection of human tumor tissue and establishment of this cell line were approved by the Ethics Review Committee of Hebei University Affiliated Hospital. The human glioma cell line TBD0220 was originally established from a glioblastoma patient who underwent surgical resection at Hebei University Affiliated Hospital. The collection of human tumor tissue and establishment of this cell line were approved by the Ethics Review Committee of Hebei University Affiliated Hospital. Written informed consent was obtained from the patient and/or their legal guardian(s) prior to surgery. These ethical approvals and consent procedures have been fully documented in a previous peer-reviewed publication16. Written informed consent was obtained from the patient and/or their legal guardian(s) prior to surgery. These ethical approvals and consent procedures have been fully documented in a previous peer-reviewed publication16. All experimental protocols involving the use of TBD0220 cell line in the present study were further reviewed and approved by the Ethics Committee of Tianjin Huanhu Hospital. All experimental protocols involving the use of TBD0220 cell line in the present study were further reviewed and approved by the Ethics Committee of Tianjin Huanhu Hospital. All methods involving human-derived cell lines were carried out in accordance with relevant guidelines and regulations, including the Declaration of Helsinki. All methods involving human-derived cell lines were carried out in accordance with relevant guidelines and regulations, including the Declaration of Helsinki. Cell transfection Lentiviral shRNA targeting ARL13B was obtained from GenePharma (Shanghai, China). The shRNA sequence was 5′-GCTGCTACATGTTATTGCAAG-3′, cloned into the LV3 (H1/GFP&Puro) vector. Western blot (WB) Cells were lysed by radioimmunoprecipitation assay (RIPA) buffer containing 1% phenylmethylsulfonyl fluoride (PMSF). Afterward, the lysate was sonicated on ice for 15 s. After centrifugation at 12,000g, the supernatant was aspirated and 5× loading buffer was added then heated at 100 °C for 10 min. Protein samples were separated by SDS-PAGE at room temperature, then transferred to PVDF membranes on ice for 1.5 h, blocked with 5% skim milk powder for 1 h, followed by overnight incubation at 4 °C with the following diluted primary antibodies: Next, the membranes were washed three times using phosphate buffer saline (PBS) with 0.1% Tween-20 for 10 min, then incubated with secondary antibodies for 1 h. Finally, a gel imaging system was used to detect images. EVs collection and analyzation The supernatant of cultured cells was collected, and centrifugated at 300 g (15 min) and 2000 g (15 min), then 36 mL of which was centrifugated at 100,000 g (1 h) and the precipitates were resuspended with PBS for NTA; another 36 mL was centrifugated at 20,000 g (1 h), then 100,000 g (1 h), collected separately using RIPA for WB. All WB original data can be found in the Supplementary Materials - WB -original Images. Nanoparticle tracking analysis (NTA) The extracted extracellular vesicles were diluted with PBS, and 1 mL was used in NTA tests by NanoSight System NS300 instrument. Five 60-second tracking video recordings were made for each sample and the number of particles and particle size were counted. Immunoprecipitation The immunoprecipitation (IP) Kit (Beyotime, Shanghai, China) was used according to the instructions. Briefly, cells were lysed by IP lysis buffer, then 20 µL protein A/G beads and 4 µg anti-ARL13B antibody were incubated with cell lysate overnight at 4 °C. The beads were collected and washed three times with IP lysis buffer, heated at 100 °C with 5×loading buffer, and the supernatant was collected for WB analysis. RNA extraction and qRT-PCR Cells were lysed with TRIzol and the lysate was collected for total RNA extraction. Chloroform in a volume of one-fifth of the lysate was added, mixed thoroughly, quiescence, and then centrifuged (12,000 g, 15 min). The clear supernatant was collected, an equal volume of isopropanol was added, mixed, quiescence at 4 °C for 30 min, and then centrifuged (12,000 g, 15 min). The white precipitate was washed once with 75% ethanol, and then the precipitate was dissolved in DEPC water to obtain the total RNA solution. The cDNA was obtained using the Reverse Transcription Kit according to the instructions. The qRT-PCR program was set up for pre-denaturation (95 °C for 30 s), cycling (95 °C for 10 s, 60 °C for 10 s), and melting curve (default melting curve of the instrument). The relative expression values for RNAs were calculated using the 2−ΔΔCt method. The primer sequences used for the qRT-PCR are listed in supplementary material-1. Chromatin immunoprecipitation (ChIP) The ChIP assay was performed using ChIP kit (9003 S, CST). TBD0220 and LN18 cells (1.2 × 107 each) were harvested and equally divided into IgG and C/EBPβ groups. The cells were lysed on ice using a lysis solution, and the DNA fragments were digested with Micrococcus spp. Following centrifugation of the resulting liquid, a portion of it was purified and submitted to agarose gel electrophoresis to confirm that the size of the DNA fragments ranged from 200 to 1000 bp. The resultant DNA solution was then mixed with the appropriate antibody (50-fold dilution) and incubated on an inverted shaker overnight at 4 °C. The following day, magnetic beads were added and incubated for 2 h. The magnetic beads were rinsed after separation by a magnetic rack, and the samples were filtered to yield DNA samples for PCR detection. The primer sequences used for the PCR are listed in supplementary material-1. Immunofluorescence Cells were washed with PBS three times, fixed with 4% paraformaldehyde (PFA) for 15 min, permeabilized with 0.2% Triton X-100 for 15 min, blocked with 5% BSA for 30 min at room temperature, incubated with anti-γ-H2AX (R) and ARL13B (M) antibody overnight at 4 °C. Next day, incubated with secondary antibodies of the corresponding species for 1 h, finally stained with DAPI and photographed with a laser confocal microscope. Flow cytometry To detect apoptosis, flow cytometry was performed using the Annexin V-FITC/PI Apoptosis Kit (Beyotime, Shanghai, China) according to the manufacturer’s instructions. For each sample, 1 × 10^5 cells were collected, washed twice with pre-cooled phosphate buffer saline (PBS), and resuspended in 190 µL Annexin V binding buffer. Then 5 µL Annexin V-FITC and 5 µL Propidium Iodide (PI) stain were added, mixed gently, and incubated for 15 min at room temperature in the dark. Green (FITC) and red (PI) fluorescence were detected by flow cytometry within 1 h. Intracellular drug concentration measurement Cells were washed with PBS three times, and then were collected using methanol at -20 °C. Samples were frozen and thawed three times repeatedly using liquid nitrogen, and the concentrations of the three drugs in the cells were detected using liquid chromatography- mass spectrometry (LC-MS). Patient-derived xenograft model building and immunofluorescence Animal experiments were performed according to the protocols approved by the Ethics Review Committee of Tianjin Huanhu Hospital (Approval Number: 202506181434000373866). All procedures were carried out in accordance with the ARRIVE guidelines and the AVMA Guidelines for the Euthanasia of Animals. Nu-nu nude mice of 4 weeks of age were used to construct an orthotopic GBM model by stereotactic injecting TBD0220-NC (TBD-NC) and TBD0220-ARL13B-KD (TBD-ARL13B-KD) cells (1.5 × 105 cells/mouse). TBD-NC and TBD-ARL13B-KD cells were transfected with luciferase lentivirus priorly. Bioluminescence images were performed on days 7, 14, and 21 after implantation of the tumor. Brain tissues were fixed in 4% PFA, dehydrated in 30% sucrose, embedded in optimal cutting temperature (OCT) compound, and sectioned at 10–20 μm thickness. Tissue sections were incubated overnight at 4 °C with primary antibodies against γ-H2AX, Ki-67, and ARL13B (all rabbit polyclonal). After washing, sections were incubated with the secondary antibody for 1 h at room temperature. Finally, sections were mounted with DAPI-containing antifade mounting medium under light-protected conditions. Mice were monitored daily for body weight, activity level, and neurological symptoms. At the experimental endpoint or when mice exhibited moribund conditions (body weight loss > 20%, severe neurological deficits, or inability to access food/water), humane euthanasia was performed in strict accordance with the AVMA Guidelines for the Euthanasia of Animals. Mice were first deeply anesthetized by inhalation of 5% isoflurane in 100% oxygen (flow rate 1 L/min) until complete loss of righting reflex and no response to toe pinch stimulation. Subsequently, rapid cervical dislocation was performed using a dedicated rodent cervical dislocator to ensure immediate and painless death. Death was verified by cessation of heartbeat and respiration for at least 2 min. Animal surgery Stereotaxic intracranial injection was performed using a mouse stereotaxic apparatus. The target injection site was the hippocampus, with coordinates defined relative to the bregma (0,0,0) as follows: anteroposterior (AP) = − 2.0 mm, mediolateral (ML) = ± 1.5 mm, dorsoventral (DV) = − 2.0 mm. Statistical analysis GraphPad Prism 9.0 and R (4.2.3) were used for all the data analysis. The P < 0.05 was considered statistically significant. Student’s t-test, one-way ANOVA, and Wilcoxon rank-sum test were used, Significance is indicated as ****P < 0.0001, ***P < 0.001, **P < 0.01, or *P < 0.05, and ns reflects non-significance. All antibodies, primers, drugs and kits etc. can be viewed in Supplementary Material 1. Results Elevated ARL13B expression predicts poor prognosis in high-grade glioma According to analysis of the data from TCGA-GDC-GBM, TCGA-GDC-LGG and CGGA-301 databases, patients with higher ARL13B expression had shorter overall survival (Fig. 1A,C). Furthermore, higher ARL13B expression predicted a poorer prognosis specifically in patients treated with TMZ (Fig. 1B,D), consistent with previous studies12,13. Furthermore, ARL13B expression levels increased with glioma malignancy. Specifically, GBM exhibited significantly higher ARL13B levels compared to low-grade glioma (LGG) (Fig. 1E,F). Moreover, within GBM samples, ARL13B expression progressively increased with higher pathological tumor grade (Fig. 1G,H). Elevated ARL13B expression predicts poor prognosis in high-grade glioma. (A,B) Kaplan-Meier survival analysis of the ARL13B expression level in full grade glioma patients (A) and patients treated with TMZ (B) based on TCGA. (C,D) Kaplan-Meier survival analysis of the ARL13B expression level in full grade glioma patients (C) and patients treated with TMZ (D) based on CGGA. (E,F) Differential expression of ARL13B in LGG and GBM based on TCGA (E) and CGGA (F). (G,H) Differential expression of ARL13B in three pathologic subtypes based on TCGA (G) and CGGA (H). (I) TOP5 pathway and their gene sets in the pathway enrichment. (J,K) Heatmap of gene expression matrices, in ascending order of expression correlation coefficient (absolute value > 0.3). (L)TOP20 pathway in pathway enrichment analysis. Moreover, we searched for gene groups from BioGRID that were predicted to have interaction with ARL13B, and then we used the gene groups for pathway enrichment analysis. Results showed that ARL13B was maximally enriched in the RHO-GTPase pathway (Fig. 1I), which indicated that ARL13B function may be related to extracellular vesicles. Additionally, we identified genes whose expression showed a significant correlation with ARL13B (absolute correlation coefficient > 0.3) in the TCGA and CGGA datasets. Heatmaps depicting the expression patterns of these correlated genes were generated (Fig. 1J,K). Subsequent survival analysis (Fig. S1) demonstrated that genes positively correlated with ARL13B expression were associated with poorer patient prognosis. Conversely, genes negatively correlated with ARL13B were associated with a better prognosis. Furthermore, the positively correlated genes exhibited higher expression in GBM compared to LGG, and their expression levels increased with higher GBM pathological grade. In contrast, the negatively correlated genes showed the opposite expression trends. In summary, ARL13B level had a negative association with glioma patients’ prognosis. Upregulation of ARL13B in TMZ-resistant GBM cells To validate the role of ARL13B in conferring TMZ resistance, we established cell lines with modulated ARL13B expression and their corresponding negative controls (NC) in TBD0220 and LN18 cell lines. Specifically, in TBD0220 cells, we generated ARL13B-overexpressing (TBD-ARL13B-OE), ARL13B-knockdown (TBD-ARL13B-KD), and NC cells (TBD-NC). In LN18 cells, we generated ARL13B-knockdown (LN18-ARL13B-KD) and NC cells (LN18-NC) (Figure S2C,D). To assess TMZ sensitivity, TBD-NC, TBD-ARL13B-OE, and TBD-ARL13B-KD cells were treated with increasing concentrations of TMZ for 72 h. We observed that TBD-ARL13B-OE cells exhibited a significantly higher IC50 value for TMZ compared to TBD-NC cells, indicating increased resistance. Conversely, TBD-ARL13B-KD cells showed a significantly lower IC50, indicating increased sensitivity (Fig. 2A). Similarly, LN18-ARL13B-KD cells demonstrated a lower TMZ IC50 compared to LN18-NC cells (Fig. 2B). Next, we evaluated TMZ-induced DNA damage by measuring levels of protein γ-H2AX, a marker of DNA double-strand breaks, using western blotting and immunofluorescence after 48 h of TMZ treatment. Results revealed that γ-H2AX levels were significantly lower in ARL13B-OE cells and higher in ARL13B-KD cells compared to their respective NC controls in cells (Fig. 2C-F), indicating that ARL13B overexpression attenuates, while its knockdown enhances, TMZ-induced DNA damage. Upregulation of ARL13B in TMZ-resistant GBM cells. (A) Cell viability of TBD-NC, TBD-ARL13B-KD and TBD-ARL13B-OE cells after 72 h treatment with gradient concentrations of TMZ. (B) Cell viability of LN18-NC and LN18-ARL13B-KD cells after 72 h treatment with gradient concentrations of TMZ. (C,D) Immunofluorescence (IF) images and Western blot (WB) analysis showing γ-H2AX levels after 48 h treatment with 400 µM TMZ. Scale bar: 20 μm. (E) Quantification of average γ-H2AX immunofluorescence intensity from (C,D). (F) WB analysis of γ-H2AX protein levels after 48 h treatment with 400 µM TMZ. (G,H) Flow cytometry analysis of apoptosis in TBD and LN18 cells with different ARL13B expression levels after 48 h treatment with 400 µM TMZ. Representative flow cytometry plots were generated using Annexin V-FITC/Propidium Iodide (PI) double staining. X-axis represents Annexin V-FITC fluorescence intensity (early apoptosis marker); Y-axis represents Propidium Iodide (PI) fluorescence intensity (late apoptosis/necrosis marker). (I) Quantification of total apoptotic rates from three independent biological replicates. Apoptotic cells were defined as the sum of early apoptotic (Q3: Annexin V-FITC positive, PI negative) and late apoptotic (Q2: Annexin V-FITC positive, PI positive) populations. Data are presented as mean ± SD (n = 3). Statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple comparisons test. A trend toward reduced apoptosis was observed in the OE + TMZ group compared with NC + TMZ in TBD cells, but this difference did not reach statistical significance, likely due to biological variability among replicates. *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. Furthermore, we analyzed apoptosis by flow cytometry in TBD-NC, TBD-ARL13B-OE, and TBD-ARL13B-KD cells following 48 h of TMZ treatment. Apoptosis levels were significantly increased in TBD-ARL13B-KD cells and decreased in TBD-ARL13B-OE cells compared to TBD-NC cells (Fig. 2G). Similarly, LN18-ARL13B-KD cells exhibited a higher level of apoptosis than LN18-NC cells after TMZ treatment (Fig. 2H). Altogether, these results demonstrated that ARL13B promotes TMZ resistance in GBM cells by reducing DNA damage and apoptosis. ARL13B knockdown enhances TMZ efficacy in vivo To further confirm ARL13B-mediated TMZ resistance, we established orthotopic xenograft models in nude mice by stereotactic injection of TBD0220 cells. Mice were randomized into TBD-NC, TBD-NC-TMZ, TBD-ARL13B-KD and TBD-ARL13B-KD-TMZ groups. Bioluminescence imaging was performed weekly for three weeks. TMZ (50 mg/kg) suspended in DMSO (0.5 mg/mL) was administered by oral gavage 5 days/week during weeks 2–3 (Fig. 3A). Kaplan-Meier analysis showed significantly different median survival: 22 days in TBD-NC, 32.5 days in TBD-NC-TMZ, 28.5 days in TBD-ARL13B-KD, and 40.5 days in TBD-ARL13B-KD-TMZ (Fig. 3D). Besides, bioluminescence imaging showed that the tumor size of the TBD-ARL13B-KD-TMZ group was smaller than the TBD-NC-TMZ group (Fig. 3B,C), Wilcoxon test was used to count the tumor size of 4 groups at the third imaging (Fig. 3H), which was identically confirmed by HE staining (Fig. 3E). Ki-67 staining showed that the TBD-ARL13B-KD-TMZ group has a slower-growing tumor than the TBD-NC-TMZ group. The γ-H2AX staining indicated that the TBD-ARL13B-KD-TMZ group has more DNA damage than the TBD-NC-TMZ group (Fig. 3F). Experiments in vivo showed that ARL13B knocking down can enhance TMZ efficacy and prolong the survival of nude mice xenograft models. ARL13B knockdown enhances TMZ efficacy in vivo. (A) Illumination of experiment. Tumor-bearing nude mice were treated with DMSO (10%inPBS) orTMZ (5 mg/kg) by gavage for 2 weeks, respectively. n = 10 per group. (B) Bioluminescence images of tumor growth on days 7, 14, and 21. (C) Quantification of bioluminescence intensity on days 7, 14, and 21 after implantation. (D) Kaplan–Meier survival curves showed the percentage survival of mice for each group. (E) H&E staining showed tumor size for each group. (F) IF of Ki-67, ARL13B and γ-H2AX. (G) Immunofluorescence intensity of FIG3 (F). (H) Wilcoxon text for tumor size on days 21. ARL13B increases ectosomes-mediated TMZ efflux Based on the result of pathway enrichment analysis (Fig. 1J,L), we collected particles at 100,000 g centrifugation grade after removing cell debris without removing 20,000 g particles, then performed NTA (Fig. 4A,B). NTA results showed that the abundance of large vesicles (about 250 nm diameter) significantly increased with elevated ARL13B expression (Fig. 4A), the portion of the difference indicated by the blue arrows is shown in Fig. 4B, a statistically significant difference in the number of vesicles within 100 nm of the main peak in the region of difference (Fig. 4C). Western blot analysis of differentially centrifuged pellets (20,000 g and 100,000 g) demonstrated ectosome markers (CD62, CD40, ARRDC1) were enriched in 20,000 g particles from high-ARL13B groups (TBD-ARL13B-OE and LN18-NC), while ARL13B itself was detected in both 20,000 g and 100,000 g fractions, indicating its potential role in extracellular vesicle formation or exocytosis (Fig. 4D). To further investigate the relationship among ARL13B, ectosomes and TMZ resistance, we detected intracellular drug concentration in three drugs in TBD0220 cells, including TMZ, 5-[(2E)-3-Methyl-2-triazen-1-yl]-4-imidazolidinecarboxamide (MTIC), 4-aminoimidazole-5-carboxamide (AICA), the intracellular drug concentration showed an opposite trend to the intracellular level of ARL13B (Fig. 4E), which indicates that ARL13B can promote TMZ efflux by improving ectosomes exocytosis. ARL13B increases ectosomes-mediated TMZ efflux. (A) Traces of Nanosight analysis for 100,000 g particles from equal number of TBD and LN18 cells. (B) The differential large-sized vesicle populations (blue arrows) in (A). (C) Statistical analysis of the more stable results of three out of five counts in the 100 nm range of the variance group. (D) WB analysis of ectosome and exosome marker in whole cell protein, 20,000 g particle and 100,000 g particle of equal number of TBD and LN18 cells. Green arrows indicate endogenous ARL13B bands, and yellow arrows indicate exogenous ARL13B bands. (E) Intracellular drug concentration in TBD-NC, TBD-ARL13B-OE and TBD-ARL13B-KD. (F) Co-IP of ARL13B and β-actin. To investigate why ARL13B promotes microvesicle exocytosis, we conducted preliminary validation experiments, performing Co-IP analyses in TBD and LN18 cells (Fig. 4F), which confirmed a direct interaction between ARL13B and β-actin. Based on this and previous literature reports, we hypothesize that ARL13B’s promotion of exocytosis may be related to β-actin. ARL13B can be transcriptionally regulated through the ERK/P90 pathway We noticed that TMZ treatment upregulated ARL13B expression (Fig. 2C-E)13. It has been reported that TMZ can activate extracellular-regulated kinase (ERK) phosphorylation17. Pathway enrichment analysis indicated that vascular endothelial growth factor A (VEGFA)/vascular endothelial growth factor receptor 2 (VEGFR2) and Ras signaling pathways rank in the top 20 pathways (Fig. 1L). ERK served as a shared node between the two pathways, suggesting potential ERK-P90-mediated ARL13B regulation. We found the ARL13B expression in GBM and LGG is highly correlated with ERK2 and its downstream molecule P90 by further GEPIA2 analysis (Fig. 5A,B). Next, time-course TMZ treatment revealed progressive phosphorylation of both ERK and P90 concomitant with ARL13B upregulation (Fig. 5C). To investigate whether ERK/P90 signaling regulates ARL13B expression, we treated GBM cells with SCH772984, an ERK1/2 inhibitor that blocks P90 phosphorylation18, either alone or combined with TMZ for 24 h. Subsequent analysis demonstrated significant reductions in ARL13B expression at both protein and mRNA levels following SCH772984 treatment (Figs. 5D,E, 6). ARL13B can be transcriptionally regulated through the ERK/P90 pathway. (A,B) Correlation analysis of ARL13B with ERK2 (A) and P90 (B) from GEPIA2. (C) WB analysis of changes in the ERK-P90 signaling pathway and ARL13B after time-gradient TMZ treatment. (D) WB analysis of changes in P90 and ARL13B after using TMZ (400 µM) and P90 inhibitor SCH772984 (2 μm). (E) Q-PCR analysis of the mRNA level of ARL13B after using TMZ (400µM) and SCH772984 (2µM). (F) ARL13B mRNA level after knocking down transcription factor. (G) The consensus sequence of the CEBPB binding sites. (H) Predicted binding sites for CEBPB in the ARL13B promoter binding region. (I) ChIP-Q-PCR result of TBD and LN-18. (J,L) Dose-response matrix for 48 H of TBD (J) and LN-18 (L). (K,M) 3D synergy distribution of TBD (K) and LN-18 (M). Mechanism of ARL13B resistant TMZ and ARL13B is transcriptionally regulated by ERK-P90 signaling pathway. Upregulation of ARL13B increases microvesicle secretion, reducing intracellular levels of TMZ metabolites MTIC and AICA. This diminishes TMZ-induced cytotoxicity, thereby enhancing TMZ resistance in glioma cells. Within cells, TMZ stimulates ERK phosphorylation, activating the ERK-P90 signalling pathway. This ultimately leads to increased levels of the transcription factor CEBPβ, which upregulates ARL13B expression at the transcriptional level. To further investigate how P90 regulates ARL13B expression at the transcriptional level, we performed siRNA knockdown experiments on a series of downstream transcription factors (TFs) of P90. Among these, knockdown of c-FOS and CEBP/β resulted in reduced ARL13B mRNA levels (Fig. 5F). Based on these results, we selected C/EBPβ as the putative transcription factor. Subsequently, we searched for potential transcription factor binding sites within the 2000 bp upstream region of the transcription start site using ChIP-seq data from vertebrates available on the JASPAR website. Figure 5G shows the nucleotide sequence of the predicted C/EBPβ binding site, and Fig. 5H displays the location of this predicted sequence within the 2000 bp upstream region of the transcription start site. Subsequently, ChIP experiments were performed in the TBD NC and LN18 NC groups (Fig. 5I), and the results showed that C/EBPβ was positive at the 10th site. These results indicate that ERK/P90 regulates ARL13B transcriptionally via C/EBPβ. Inhibition of P90 phosphorylation can enhance TMZ efficacy To investigate whether the ERK/P90 pathway inhibition could influence TMZ efficacy, we performed combinatorial drug screening in TBD0220 and LN18 cells at a dose-response matrix using SCH772984 and TMZ. Results showed that combinatorial treatment with SCH772984 and TMZ significantly suppress proliferation in TBD0220 and LN18 cells (Fig. 5J,L), with the synergy scores exceeding 10 (Fig. 5K-M), demonstrating that pharmacological inhibition of ERK/P90 signaling potentiates TMZ cytotoxicity. Discussion The standard management for GBM treatment involves maximal safe surgery and adjuvant chemotherapy19. However, the major limitation of GBM treatment is TMZ resistance3. In the last decades, many TMZ-resistance mechanisms have been reported, such as O6-methylguanine-DNA methyltransferase (MGMT) expression, Base excision repair (BER) pathway activation, Fanconi anemia (FA) pathway dysregulation, microRNA-mediated adaptive responses, Multidrug resistance (MDR) transporter overexpression20–24. It has been reported that ARL13B has a higher expression level in GBM than normal tissues and is negatively correlated with glioma patient survival25, consistent with our findings. Separately, ARL13B promotes TMZ resistance in GBM by affecting purine metabolism13. Our Metascape pathway enrichment analysis revealed that ARL13B is closely related to extracellular vesicles. As a result, we made a hypothesis that ARL13B may have an alternative mechanism to induce TMZ resistance via extracellular vesicle-mediated efflux mechanisms. Based on previous studies, ectosomes release required ARF6 and actin, which depend on ERK pathway activation26. Notably, ARL13B colocalized with ARF6 and interacted with β-actin15, while exhibiting strong expression correlation with ARF6 (Fig. S2). So, we expected that ARL13B may play a role in ectosomes shedding. Although the previous reports indicate that ectosomes collected by differential centrifugation can’t differ significantly from exosomes in NTA27, which means it couldn’t purely differentiate ectosomes and exosomes by centrifugation, western blot or Elisa. However, we observed the difference of large EVs in the NTA results and western blot simultaneously, which showed that ARL13B can promote large EV (20,000 g particles) exocytosis. In addition to CD62 and CD40, we also detected ARRDC1, another marker for ectosomes generated through plasma membrane budding. ARRDC1 levels in 20,000 g particles showed significant positive correlations with ARL13B expression, indicating potential involvement of ARL13B in ARMM biogenesis. In our results, LN18 has less ARL13B than TBD0220 cells (Fig. S2A,B), which may be the reason why LN18 has less ectosome than TBD0220 (Fig. 4A.B). In our experimental results, following siRNA-mediated knockdown of the c-FOS transcription factor, the mRNA levels of ARL13B also decreased. However, subsequent experiments did not yield positive ChIP results for c-FOS; therefore, we have only presented the results for the CEBP/β section. While the WNT pathway has been reported to transcriptionally regulate ARL13B via ZEB112. However, our enrichment analysis implicated the ERK/P90 pathway in ARL13B regulation. Experimental validation confirmed that ARL13B is regulated by the ERK/P90 signaling pathway. Notably, TMZ induces ARL13B upregulation13 and activates ERK phosphorylation17. These prior findings align with our results, demonstrating that TMZ transcriptionally upregulates ARL13B via ERK-dependent mechanisms. In the present study, we observed that the phosphorylation levels of ERK and P90 firstly increased and subsequently decreased to levels below the baseline after TMZ intervention. This dynamic change could be reasonably explained by the following molecular mechanisms. Mechanistically, TMZ-induced DNA damage initiates intracellular p53 activation, which further upregulates p21 and blocks cell cycle progression. The activated cell cycle checkpoints induce G1/S phase cell cycle arrest in surviving cells. Such cell cycle arrest weakens cellular reliance on the RAS signaling cascade and suppresses the activity of RAS as well as its downstream signaling molecules, ultimately restraining the phosphorylation of ERK and P9029. Apart from DNA damage response, TMZ treatment also elicits endoplasmic reticulum stress, which further activates the PERK/ATF4/CHOP signaling cascade and represses overall cellular protein translation process. This stress-related inhibitory effect is capable of directly or indirectly downregulating the phosphorylation status of ERK and P90, jointly driving their phosphorylation levels to drop below basal levels30. Conclusion ARL13B increased ectosomes exocytosis in TBD and LN18 cell lines, which increased TMZ resistance in cells. ARL13B was transcriptionally regulated by the ERK/P90 signaling pathway and one of the TF of ARL13B is C/EBPβ. TMZ could upregulate ARL13B expression by stimulating ERK phosphorylation. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Supplementary Material 5 Supplementary Material 6 Supplementary Material 7 Supplementary Material 8 Supplementary Material 9 Supplementary Material 10 Acknowledgements We extend our gratitude to the Tianjin Institute of Neurosurgery for providing laboratory facilities and technical support and we extend our gratitude to all personnel involved in this publication. Abbreviations 4-Aminoimidazole-5-carboxamide ADP ribosylation factor like GTPase 13B Arrestin-domain-containing protein 1-mediated microvesicle Arrestin domain containing 1 Chinese glioma genome atlas Chromatin immunoprecipitation Dulbecco’s modified Eagle’s medium Extracellular-regulated kinase Extracellular vesicle Immunoflurescence Immunoprecipitation Liquid chromatograph mass spectrometer 5-[(2E)-3-Methyl-2-triazen-1-yl]-4-imidazolidinecarboxamide Nanoparticle tracking analysis Optimal cutting temperature Paraformaldehyde Phenylmethylsulfonyl fluoride Radioimmunoprecipitation assay The cancer genome atlas Transcription factor Temozolomide Western-blot Author contributions conceptualization and the study design: XML. Carried out the experiments: HCX, YC, MJW, HLL, YJC. Analyzed the data, prepared the figures, and wrote the manuscript: HCX. Financially supported and supervised this work: XML. Funding Tianjin Key Medical Discipline Construction Project (Grant No. TJYXZDXK-3–002 A) Beijing-Tianjin-Hebei Basic Research Cooperation Special Projects (23JCZXJC00090/J230003); Tianjin Key Project of Health Science and Technology (TJWJ2021ZD008); Tianjin key Project of Municipal Education Commission Scientific Research Plan (2024ZD051). Data availability Mass spectrometry and proteomics data can be retrieved from the TCGA database, CGGA database, and Biogrid database. All analyses employed complete, unprocessed raw data. FIG1A-H present the full output results; FIG1I displays the top 5 signalling pathways (ranked by p-value in ascending order), while L shows the top 20 pathways (ranked by p-value in ascending order). FIG1 J and K illustrate sections where the absolute R value exceeds 0.3 in the final results. Specific data set and analytical methods can be requested by emailing xinghaichao1997@foxmail.com.The following are links to the data sources: TCGA: https://xena.ucsc.edu. CGGA: https://www.cgga.org.cn/. Biogrid: https://thebiogrid.org/128353/summary/homo-sapiens/arl13b.html. Declarations Competing interests The authors declare no competing interests. Ethical statement All animal experimental protocols were approved by the Ethics Review Committee of Tianjin Huanhu Hospital (Approval Number: 202506181434000373866). Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Supplementary Material 5 Supplementary Material 6 Supplementary Material 7 Supplementary Material 8 Supplementary Material 9 Supplementary Material 10 Data Availability Statement Mass spectrometry and proteomics data can be retrieved from the TCGA database, CGGA database, and Biogrid database. All analyses employed complete, unprocessed raw data. FIG1A-H present the full output results; FIG1I displays the top 5 signalling pathways (ranked by p-value in ascending order), while L shows the top 20 pathways (ranked by p-value in ascending order). FIG1 J and K illustrate sections where the absolute R value exceeds 0.3 in the final results. Specific data set and analytical methods can be requested by emailing xinghaichao1997@foxmail.com.The following are links to the data sources: TCGA: https://xena.ucsc.edu. CGGA: https://www.cgga.org.cn/. Biogrid: https://thebiogrid.org/128353/summary/homo-sapiens/arl13b.html.