Association of ITPKB with glioma malignancy

TMZ, in conjunction with radiotherapy, followed by surgical resection, has become the standard therapy for glioma globally, where drug resistance poses a significant challenge in GBM. To reveal potential biomarkers, novel drug targets, and the underlying mechanisms of TMZ resistance, six pairs of GBM patients who underwent standard-of-care treatment and surgery for both primary and recurrent diseases were enrolled. The tandem mass tag (TMT)-based proteomic analysis identified 7,524 proteins, with 266 proteins significantly upregulated and 382 downregulated between the recurrence and primary groups (p < 0.05 and fold change > 1.2) (Fig. 1a and Supplementary Table 1). Functional enrichment analysis of differentially expressed proteins revealed that BIOLOGICAL_OXIDATIONS and METABOLISM_OF_LIPIDS pathways were overrepresented in proteins upregulated in recurrent tumors (Fig. 1b and Supplementary Fig. 1a).

Fig. 1

Differential expression analysis of the protein ITPKB in primary and recurrent glioma patients. a Six pairs of GBM tissues from glioma patients who received standard TMZ treatment and underwent primary and recurrent surgical treatment, were analyzed by mass spectrometry. A volcano plot represented the differential expression of proteins with a condition of P < 0.05 and an absolute fold change > 1.2. b GSEA pathway enrichment analysis was performed based on the identified differentially expressed proteins between the primary and recurrent groups in the tumors. c Lipid kinases and differentially expressed proteins depicted in a Venn diagram. d–g Expression profiles of differentially expressed lipid kinases (PIP4K2A, ITPKB, ITPK1, and PRKDC) in the proteomic analysis of primary and recurrent GBM tissue from six patients. h The prognostic significance of ITPKB in recurrent glioma patients was analyzed using the CGGA database (http://www.cgga.org.cn/). i, j IHC staining of ITPKB was performed in 19 pairs of primary/recurrent GBM patient specimens who received TMZ chemotherapy at Xiangya Hospital. The represented images are shown and quantification was analyzed by t-test. k, l IHC staining of ITPKB was analyzed in human glioma tissue arrays to assess the ITKPB survival values. The high or low group was reduced by the median total score of 4. The association of ITPKB with progression-free survival and overall survival of glioma patients was assessed using the log-rank test. m, n Survival subgroup analysis was performed using the log-rank test in HGG and LGG patients. Statistical significance is shown as: *p < 0.01, **p < 0.001, ns nonsignificant

During the past four decades, the lipid peroxidation process has been shown to play crucial roles in cell biology and human health, wherein lipid kinases catalyze the production of lipids.27 By comparing the proteomic data of our study and the lipid kinases of KiNativ,28 a Venn diagram reveals four intersected kinases (PIP4K2A, ITPKB, ITPK1, and PRKDC) (Fig. 1c), which may be associated with glioma recurrence. Further validation showed that only ITPKB and ITPK1 were consistently highly expressed in recurrent samples compared to primary ones (Fig. 1d–g). Considering the oncological relevance, ITPKB was selected for further investigation. ITPKB, an essential cellular enzyme, catalyzes the inositol phosphate phosphorylation of the second messenger inositol 1,4,5-trisphosphate (IP3) to inositol 1,3,4,5-tetrakisphosphate (IP4), regulating intracellular calcium fluxes.17 Data from the Chinese Glioma Genome Atlas database (CGGA, http://www.cgga.org.cn/)29 demonstrated that high ITPKB expression was correlated with a poorer survival probability (Fig. 1h). Western blotting of the six pairs of GBM patients validated that ITPKB was highly expressed in recurrent samples compared to primary samples (Supplementary Fig. 1b). Pancancer analysis using the UALCAN platform30 revealed that ITPKB was expressed at significantly higher levels in pancreatic and GBM tumors compared to normal tissues (Supplementary Fig. 1c). To validate these findings, an additional 19 pairs of glioma patient tissues from Xiangya Hospital were analyzed, confirming the higher expression of ITPKB in the recurrent group compared to the primary group (Fig. 1i, j). Furthermore, tissue arrays collected from 151 glioma cases revealed that the group with higher expression of ITPKB had poorer overall survival (median OS: 86 vs. 99 months, p < 0.001) and PFS (median PFS: 85 vs. 99 months, p < 0.001) than the lower expression group (Fig. 1k, l). The stratified analysis showed that higher ITPKB expression predicted poorer OS in patients with high-grade glioma (HGG) and low-grade glioma (LGG) (Fig. 1m, n). These findings suggest a strong association of ITPKB with GBM recurrence and patient survival.

Upregulation of ITPKB in TMZ-resistant GBM

In our study, patients diagnosed with pathologically confirmed GBM underwent standard first-line treatment, which included maximal surgical resection followed by TMZ-based chemotherapy with concurrent radiotherapy. Clinically, treatment resistance is identified as a primary factor contributing to glioma recurrence. To explore the role of ITPKB in the radioresistance of GBM, a pair of radiation-sensitive and resistant GBM cells (U251 and U251-IR) were used. The expression of the ITPKB protein did not differ significantly between radiosensitive and radioresistant GBM cells (Supplementary Fig. 2a). Therefore, we propose a direct association between elevated ITPKB expression in relapsed patients and TMZ resistance.

Using two TMZ-resistant cell lines (U118-R, T98G-R) (Supplementary Fig. 2b, c), we observed a higher expression of the ITPKB protein in TMZ-resistant cells compared to sensitive glioma cells (U118, T98G) (Fig. 2a). In TMZ-sensitive T98G and U118 cells, TMZ treatment induced significant expression of ITPKB in a time-dependent manner (Fig. 2b). In contrast, this induction was not observed in TMZ-resistant T98G-R and U118-R cells (Fig. 2c, Supplementary Fig. 2d), suggesting that TMZ resistance may involve an adaptive process. Using two different ITPKB- shRNAs, we established T98G-R and U118-R cell lines with stable ITPKB knockdown (Fig. 2d). The colony formation assay revealed that ITPKB depletion significantly increased drug sensitivity in TMZ-resistant cell lines (Fig. 2e, f). Reexpression of ITPKB-WT, but not of the kinase-dead ITPKB-DN mutant, led to the reversion of survival inhibition in ITPKB-depleted cells (Fig. 2g, h).

Fig. 2

ITPKB-mediates TMZ sensitivity in GBM cells dependent on its kinase activity. a The expression of ITPKB was assessed in TMZ-sensitive (T98G, U118) and resistant (T98G-R, U118-R) cells using Western Blot. b T98G and U118 cells were treated with TMZ for indicated durations, and the cell lysates were blotted with specified antibodies. c After 48 h of TMZ treatment, lysates were collected from both TMZ-sensitive (T98G, U118) and resistant (T98G-R, U118-R) cells. Subsequently, these cell lysates were blotted with the indicated antibodies. d Construction of ITPKB-depleted TMZ-resistant glioma cell lines involved the transduction of T98G-R and U118-R cells with lentivirus encoding control (Ctrl) or ITPKB shRNAs. Cell lysates were then blotted with the specified antibodies. e, f Glioma cells from d were subjected to treatment with indicated doses of TMZ (0, 25, 50, 75, 100 μM) for 10-15 days. Cell survival was determined by colony formation assay. Error bars represent ± SD from three independent experiments. g TMZ-resistant glioma cells stably expressing Ctrl or ITPKB shRNA-1 were transiently transfected with the wild-type (WT) and kinase-dead mutant D897N of ITPKB. Subsequently, the cells were lysed, and Western blot analysis was performed using the specified antibodies. h Glioma cells from g were treated with 100 μM TMZ for 10–15 days, and cell survival was determined by colony formation assay. Error bars represent ± SD from three independent experiments. i TMZ-resistant glioma cells were treated with 20 μM GNF362 and/or 25 μM TMZ, and cell survival was determined by colony formation assay. Error bars represent ± SD from three independent experiments. j The expression of ITPKB in glioma cell lines (U251, U87, U118, U373, U138, LN299, HS683, and T98G) was assessed by western blot, and IC50 of each cell line was calculated after the addition of gradient concentration of TMZ for 96 h. The correlation between ITPKB expression and IC50 was analyzed by Pearson’s test in a panel of glioma cell lines. Statistical significance is indicated as: *p < 0.01, **p < 0.001, ns nonsignificant

GNF362 is a selective, potent, and orally bioavailable inhibitor of ITPKB.17 In TMZ-resistant cell lines, the combined use of GNF362 with TMZ significantly inhibited cell survival compared to the GNF362 or TMZ used alone (Fig. 2i). Furthermore, a positive correlation was validated between ITPKB protein expression and TMZ IC50 using a panel of eight glioma cell lines (Pearson r = 0.8412, p = 0.0013) (Fig. 2j). Considering the high heterogeneity of human glioblastoma, we further validated our findings in another TMZ-resistant U87-R GBM cell line, established in our previous study23 (Supplementary Fig. 2e). We showed that downregulation of ITPKB increased sensitivity to TMZ in U87-R cells (Supplementary Fig. 2f). Wild-type ITPKB could rescue enhanced sensitivity, but not the kinase-dead ITPKB-DN mutant control (Supplementary Fig. 2g). The combined use of GNF362 and TMZ significantly inhibited U87-R cell survival compared to GNF362 or TMZ used alone (Supplementary Fig. 2h). Collectively, these observations underscore the crucial role of ITPKB in TMZ resistance through its kinase activity.

ITPKB-mediated TMZ resistance through regulation of ROS homeostasis

To unravel the underlying mechanism of ITPKB-mediated TMZ resistance, we used the Linkedomics database (http://www.linkedomics.org/login.php)31 to analyze the TCGA glioma dataset based on ITPKB expression. Analysis revealed a positive correlation between ITPKB expression and peroxisomal lipid metabolism and ROS-related signaling pathways, including Oxidative Stress Response, Biological Oxidations, and RHO GTPases Activate NADPH Oxidases (Supplementary Fig. 3a). Our proteomic analyses of six pairs of GBM patients also identified the enrichment of the BIOLOGICAL_OXIDATIONS pathway in recurrent tumors (Fig. 1b), further suggesting the involvement of ITPKB in regulating ROS homeostasis.

As illustrated in Fig. 3a, b and Supplementary Fig. 3b, c, TMZ-resistant cells exhibited significantly higher basal ROS levels than sensitive glioma cells. Although TMZ treatment increased ROS levels in sensitive cells (T98G and U118), it did not affect resistant cells (T98G-R and U118-R). Knocking down ITPKB in resistant cells restored TMZ-induced ROS production and apoptosis (Fig. 3c–e and Supplementary Fig. 3d–k). Furthermore, the continued expression of ITPKB WT, but not of the kinase-dead ITPKB-DN mutant, in ITPKB-depleted T98G-R, U118-R, and U87-R cells reversed the phenotype to resistance, including decreased ROS production and apoptosis rates (Fig. 3f, g and Supplementary Fig. 3l–r). These findings suggest that the impact of ITPKB on the chemotherapy response in TMZ-resistant glioma cells depends on ROS homeostasis.

Fig. 3

ITPKB participates in TMZ sensitivity through ROS homeostasis. a, b TMZ-sensitive and resistant cells were exposed to 100 μM TMZ for 48 h, and ROS levels were assessed using the DCFDA assay. Data, derived from three independent samples, are presented as mean fold change to control ± SD. c–e T98G-R cells stably expressing Ctrl or ITPKB shRNAs were treated with or without 100 μM TMZ for 48 h. ROS levels were measured using the DCFDA assay, and cell apoptosis was assessed after 500 μM TMZ treatment using the Annexin V-FITC apoptosis kit. Data, from three independent samples, are displayed as mean fold change to control ± SD. f, g TMZ-resistant glioma cells stably expressing Ctrl or ITPKB shRNA-1 were transiently transfected with the wild-type (WT) and kinase-dead mutant D897N of ITPKB. Following 100 μM TMZ treatment, ROS levels were determined by DCFDA assay, and cell apoptosis was assessed after 500 μM TMZ treatment using the Annexin V-FITC apoptosis kit. h, i T98G-R cells stably expressing Ctrl or ITPKB shRNAs were treated with 100 μM TMZ and/or 1000 U/ml antioxidant enzyme Catalase. ROS levels were measured using the DCFDA assay, and relative cell survival was determined by CCK8 assay. Error bars represent ± SD from three independent experiments. j T98G-R cells stably expressing Ctrl or ITPKB shRNAs were treated with or without 100 μM TMZ for 48 h, and NADH Oxidase Activity Assay Kit was used to analyze NOX activity. k, l T98G-R cells stably expressing Ctrl or ITPKB shRNAs were treated with 100 μM TMZ and/or 10 μM NOX1/4 inhibitor GKT137831. ROS was measured using the DCFDA assay. For cell apoptosis, T98G-R cells were treated with 500 μM TMZ and/or 10 μM NOX1/4 inhibitor GKT137831, and assessed by the Annexin V-FITC apoptosis kit. m The level of IP4 was measured in T98G-R cells stably expressing control or ITPKB shRNAs using an ELISA assay. Statistical significance is shown as: *p < 0.01, **p < 0.001

To validate this hypothesis, we used catalase, an enzyme detoxifying hydrogen peroxide, to decrease ROS levels in ITPKB-depleted cells. Catalase treatment reversed the cell survival inhibition caused by ITPKB knockdown (Fig. 3h, i and Supplementary Fig. 3s, t). The NOX family of NADPH oxidases is a key enzyme responsible for ROS production and critical determinants of chemosensitivity in multiple cancers.32,33 In T98G-R and U118-R cells, ITPKB depletion significantly induced NOX activity after TMZ treatment (Fig. 3j and Supplementary Fig. 3u). Furthermore, the elevated ROS level was reduced, and apoptosis induced by stable ITPKB knockdown was rescued when the NOX1/4 inhibitor GKT137831 was used to inhibit NOX activity (Fig. 3k, l and Supplementary Fig. 3v, w). Previous studies have reported that the ITPKB product, IP4, can effectively inhibit NOX4 activity in cisplatin-resistant cancer cells.20 In TMZ-resistant cells, we observed a significant decrease in IP4 levels after ITPKB knockdown (Fig. 3m and Supplementary Fig. 3x). These data indicate that the downregulated expression of ITPKB increases TMZ sensitivity by inducing NOX activity and ROS accumulation.

E3 ligase Trim25 is an ITPKB-binding protein

Given the observed increase in ITPKB protein levels, but not mRNA levels, in six pairs of primary and recurrent tissues (Supplementary Fig. 1b and Fig. 4a) and the finding that TMZ induction did not alter ITPKB mRNA levels (Fig. 4b and Supplementary Fig. 4a), we hypothesized that the elevated ITPKB protein level in TMZ-resistant cells occurred through a posttranscriptional mechanism. To verify this, we conducted coimmunoprecipitation and mass spectrometry experiments after overexpressing Flag-tagged ITPKB in HEK293T and T98G-R cells. Among the three main potential binding partners, AP2A1 or CLIP1 was not related to ITPKB stability, as evidenced by data showing that AP2A1 or CLIP1 knockdown did not affect ITPKB levels (Supplementary Fig. 4b–d). Our focus then turned to Trim25, and we validated its binding to ITPKB in HEK293T and T98G-R cells (Fig. 4c). Immunoprecipitation assays using a Trim25 antibody confirmed the interaction between Trim25 and endogenous ITPKB in glioma cells, with a weaker interaction observed in resistant cell lines compared to sensitive cells (Fig. 4d).

Fig. 4

Trim25 is an ITPKB-binding protein. a Relative ITPKB mRNA levels in six pairs of GBM tissues from Fig. 1a were determined by RT-PCR. b Transcriptional mRNA levels of ITPKB were analyzed in TMZ-sensitive and resistant cells after TMZ treatment by RT-PCR. c A list of ITPKB-binding proteins was identified through mass spectrometric analysis. HEK293T and T98G-R cells expressing Flag-ITPKB were generated, and ITPKB complexes were subjected to mass spectrometric analysis. d Glioma cell lysates from TMZ-sensitive (T98G, U118) and resistant (T98G-R, U118-R) cells were immunoprecipitated with control IgG or anti-Trim25 antibody, followed by blotting with indicated antibodies. e T98G and U118 cells treated with 100 μM TMZ for 48 h had their cell lysates subjected to immunoprecipitation with control IgG or anti-Trim25 antibody. f HEK293T cells expressing wild-type Flag-ITPKB were transfected with V5-tagged Trim25. Cells were pretreated with 100 μM TMZ, then co-treated with 50 μM MG132 for an additional 3 h. Cell lysates were immunoprecipitated with Anti-FLAG® M2 Magnetic Beads, followed by blotting with the indicated antibodies. g, h Representative images of merged Proximity Ligation Assay (PLA) and nuclei (DAPI) channels from PLA experiments. In situ PLA was utilized to assess the interaction between Trim25 and endogenous ITPKB in TMZ-sensitive and resistant glioma cells. Each red dot represents the detection of the Trim25-ITPKB interaction complex, and the graphs representing mean ± SD are shown in (g). The scale bar in the bottom left is 20 μm. i, j Trim25 was found to bind to the kinase domain of ITPKB. HEK293T cells were transiently transfected with V5-Trim25 along with wildtype (WT) and truncated mutants (1–768, 1–800, 768–946, and 800–946 aa) of Flag-ITPKB. The protein interaction was assayed by immunoprecipitation with Anti-FLAG® M2 Magnetic Beads, followed by blotting with the indicated antibodies. Schematic representation of Flag-ITPKB and its deletion mutants is shown in (j)

The binding ability of ITPKB and Trim25 decreased after TMZ treatment in T98G and U118 cells (Fig. 4e). Furthermore, V5-tagged Trim25 was detected in immunoprecipitation assays using anti-Flag magnetic beads from cells expressing Flag-ITPKB, and treatment with TMZ reduced the interaction between Trim25 and ITPKB (Fig. 4f). Further confirmation of the close interaction between Trim25 and ITPKB was obtained by proximity ligation assay (PLA), revealing more dots representing the interaction between Trim25 and ITPKB in TMZ-sensitive cells than in resistant cells (Fig. 4g–h). To identify the region(s) of ITPKB that mediates its interaction with Trim25, we constructed a series of Flag-tagged ITPKB truncation mutants. Co-IP analysis showed that the ITPKB kinase domain was sufficient and necessary for its binding to Trim25 (Fig. 4i–j). Overall, these results suggest that Trim25 interacts with ITPKB in cells, and this interaction is related to the cellular response to TMZ.

Trim25 mediates K48-linked ubiquitination of ITPKB at K793 and K818 sites

To explore the biological significance of the interaction between Trim25 and ITPKB in glioma cells, we further examined cellular responses upon loss or gain of Trim25 function. In T98G-R and U118-R cells, the knockdown of Trim25 by two different siRNAs increased the level of ITPKB (Fig. 5a). At the same time, Trim25 overexpression reduced the ITPKB protein level (Fig. 5b). The decreased ITPKB level after Trim25 overexpression was reversed by adding MG132, a proteasome inhibitor, indicating that Trim25 reduces ITPKB stability through the proteasome pathway (Fig. 5c and Supplementary Fig. 5a). We performed cycloheximide (CHX) chase assays to determine the protein stability of ITPKB. We found that Trim25 knockdown increased the half-life of the ITPKB protein in both T98G-R and U118-R cells (Fig. 5d, e and Supplementary Fig. 5b, c). In contrast, ITPKB stability was significantly attenuated in glioma cells with Trim25 overexpression (Fig. 5f, g and Supplementary Fig. 5d, e).

Fig. 5

ITPKB interaction with Trim25 dependent on K48 ubiquitination at sites K793 and K818. a Depletion of Trim25 induces ITPKB protein levels. T98G-R and U118-R cells were transfected with control (Ctrl) or Trim25 siRNAs, and cell lysates were immunoblotted with the indicated antibodies. b T98G-R and U118-R cells were transiently transfected with Trim25 plasmid, and the cell lysates were immunoblotted with the indicated antibodies. c T98G-R cells transfected with Ctrl or Trim25 plasmid were treated with vehicle or MG132 (50 μM) for 3 h. Cell lysates were then immunoblotted with the indicated antibodies. d T98G-R cells transfected with Ctrl or Trim25 siRNAs were treated with CHX (0.1 mg/mL) and harvested at the indicated times. Cells were lysed, and cell lysates were then immunoblotted with the indicated antibodies. e Quantification of the ITPKB protein levels relative to Actin. f T98G-R cells transfected with Trim25 plasmid were treated with CHX (0.1 mg/mL) and harvested at the indicated times. g Quantification of the ITPKB protein levels relative to Actin. h Cells transfected with Ctrl or Trim25 siRNAs were treated with MG132 for 3 h before harvest. Immunoprecipitation with control IgG and ITPKB was performed, followed by immunoblotting with the indicated antibodies. i, j HEK293T cells expressing Flag-ITPKB were transiently transfected with V5-tagged Trim25 and HA-tagged ubiquitin/K48 ubiquitin. After 48 h, cells were treated with MG132 (50 μM) for 3 h. Cell lysates were immunoprecipitated with Anti-FLAG® M2 Magnetic Beads, and then immunoblotted with the indicated antibodies. k HEK293T cells expressing Flag-ITPKB were transiently transfected with indicated HA-K48 lysine-specific mutant constructs. After 48 h, cells were treated with MG132 for 3 h before collection. Cell lysates were immunoprecipitated and immunoblotted with the indicated antibody. l, m Identification of the ubiquitination sites of ITPKB for its K48-specific polyubiquitination. ITPKB stably knockdown HEK293T cells were transiently transfected with indicated constructs. After 48 h, cells were treated with MG132 for 3 h before collection. Cell lysates were immunoprecipitated and then immunoblotted with the indicated antibodies. n–q Trim25 plasmids were co-transfected with a control vector, ITPKB wildtype, or 2KR mutant in T98G-R cells. ROS levels were measured using the DCFDA assay. Cell apoptosis was assessed by the Annexin V-FITC apoptosis kit. Relative cell survival was determined by CCK8 assay. Statistical significance is shown as: *p < 0.01, **p < 0.001

To test whether Trim25-mediated ubiquitination of ITPKB caused a change in ITPKB stability, ubiquitination assays were performed under various conditions. Trim25 knockdown in T98G-R and U118-R cells significantly decreased the endogenous polyubiquitination of ITPKB compared to control cells (Ctrl) (Fig. 5h). Trim25 overexpression increased ITPKB polyubiquitination in HEK293T cells expressing Flag-tagged ITPKB (Fig. 5i). To further elucidate the characteristics of Trim25-mediated ITPKB ubiquitination, we determined the types of ubiquitin linkage using K48-specific and K63-specific ubiquitin plasmids in which only one specific lysine was retained. The ubiquitination assays revealed that ITPKB was extensively ubiquitinated with Ub-K48 but not Ub-K63 (Fig. 5j and Supplementary Fig. 5f). Using Ub-K48 and its mutant Ub-K48R, we found that the K48 polyubiquitination of ITPKB was abolished by the Ub-K48R mutation (Fig. 5k). To examine which lysine residues in ITPKB were ubiquitinated by Trim25, we used BioGRID (https://thebiogrid.org)34 and the web-based GPS-Uber databases (http://gpsuber.biocuckoo.cn/online.php),35 predicting that ITPKB K793 and K818 might be potential ubiquitination sites near their binding domain with Trim25. The ITPKB K793R and K818R mutants were then generated, and the single-site mutation (K793R or K818R) showed little effect on polyubiquitination, while the two K to R mutations (2KR) almost completely abolished K48-linked polyubiquitination in ITPKB (Fig. 5l–m). To further verify whether the Trim25-ITPKB-ROS target axis is associated with TMZ resistance, the Trim25 plasmid was co-transfected with a control vector, wild-type ITPKB, or 2KR mutant plasmids in TMZ-resistant GBM cells. In Trim25 overexpression, the ITPKB wildtype, but not the 2KR mutant, increased the survival of T98G-R and U118-R cells (Fig. 5n and Supplementary Fig. 5g, h), accompanied by decreased ROS levels (Fig. 5o–p and Supplementary Fig. 5i–k) and a reduced rate of apoptosis (Fig. 5q and Supplementary Fig. 5l, m). These data demonstrate that the interaction between ITPKB and Trim25 leads to the K48 polyubiquitination of ITPKB, decreasing ITPKB stability.

To investigate the reasons for the functional differences of Trim25 in primary and recurrent GBM, we analyzed quantitative proteomic data, showing that there was no difference in Trim25 level between primary and recurrent samples (Supplementary Fig. 6a). Since post-translational modifications, such as phosphorylation, have been shown to play an important role in modulating Trim25 activities,26,36,37 we performed a phosphoproteomics analysis of Trim25 phosphorylation in the six pairs of samples from primary and recurrent GBM tumors. Among the five potential phosphorylation sites in Trim25, only phosphorylation at the S100 site was significantly reduced in the recurrent samples (Fig. 6a, b), suggesting that the decreased level of S100 phosphorylation in Trim25 might disrupt the activity of the Trim25 E3 ligase, leading to increased stability of ITPKB. To test this hypothesis, in vitro ubiquitination assays were conducted and showed that ITPKB ubiquitination was significantly alleviated when Trim25 was dephosphorylated by alkaline phosphatase (ALP) (Fig. 6c). Furthermore, through phosphomimetic and dephosphorylation approaches, we found that the binding ability of Trim25 to ITPKB was enhanced in the Trim25 S100D mutation (which mimics a phosphomimetic form) but reduced in the Trim25 S100A mutation (which mimics a dephosphorylated form) (Fig. 6d). In contrast, the Trim25 S100D mutant further increased the ubiquitination level of ITPKB, while the Trim25 S100A mutant decreased the ubiquitination level of ITPKB compared to wild-type Trim25 (Fig. 6e). These observations suggest that phosphorylation of the S100 site is necessary for Trim25-mediated ubiquitination of the ITPKB protein.

Fig. 6

Phosphorylation of Trim25 at S100 is required for ITPKB ubiquitination. a Phosphorylation proteomics analysis of Trim25 in primary and recurrence GBM. b The phosphorylation level of Trim25 S100 site in GBM tissue. c Dephosphorylation of the E3 ligase Trim25 resulted in a more pronounced alleviation of ITPKB ubiquitination in a cell-free system compared to mock-treated Trim25. Dephosphorylated or mock-treated Trim25 was incubated with immunopurified Flag-ITPKB, ubiquitin, recombinant E1 (Uba1), and E2 (UbcH5b). ITPKB ubiquitination was then determined using a ubiquitination assay. d The interaction between ITPKB and Trim25 was affected by the Trim25 S100D phosphomimetic mutant and S100A dephosphorylation mutant. e The ubiquitination of ITPKB was altered when the S100 site was mutated to Asp (S100D) or Ala (S100A). HEK293T cells were co-transfected with the respective plasmids and treated with MG132 for 3 h after 48 h of transfection. Cell lysates were immunoprecipitated and blotted with specific antibodies for analysis. Statistical significance is shown as: *p < 0.01

ITPKB deficiency increased TMZ sensitivity in vivo

To investigate the relevance of ITPKB to TMZ resistance in vivo, we subcutaneously implanted ITPKB-depleted T98G-R cells in nude mice. We monitored tumor growth for five weeks (Fig. 7a). Thirty-five days after tumor cell implantation, we observed that ITPKB knockdown, in combination with TMZ treatment, significantly decreased tumor growth and tumor weight compared to the control group (Fig. 7b–d). The mice’s body weight did not change significantly during the experiment (Fig. 7e). To validate the impact of ITPKB on ROS homeostasis in vivo, we assayed H2O2 concentration in tumors from in vivo tumorigenesis experiments. The H2O2 level increased dramatically in the ITPKB-depleted and TMZ-treated samples (Fig. 7f). The expression of ITPKB and the proliferative marker Ki67 in the above experiment was validated by IHC staining. The results indicated that Ki67 staining decreased in the ITPKB knockdown groups treated with TMZ (Fig. 7g–i).

Fig. 7

In vivo regulation of glioma TMZ sensitivity by ITPKB through ROS homeostasis. a Schematic diagram of tumor xenograft experiments using ITPKB knockdown glioma cells. 1 × 107 cells were subcutaneously injected into nude mice. Tumor volumes were measured at indicated days. Mice were sacrificed after 5 weeks. b Tumor images were acquired (n = 5). c Tumor weights were measured and represented as mean tumor weight ± SD. d, e Tumor volume and mice body weight measured on the indicated day, represented as mean ± SD. f Tissues from a were homogenized and centrifuged. The H2O2 concentration of each tissue was analyzed by the H2O2 assay kit at 560 nm absorbance. g Representative IHC staining image of tissues from a using ITPKB and Ki67 antibodies. h, i Quantifications of ITPKB and Ki67 IHC results. j Schematic diagram of tumor xenograft experiments using combined treatment with GNF362 and TMZ. Tumor images of each group were shown in k, and tumor weight represented as mean ± SD was shown in (l). m, n Tumor volume and mice body weight measured on the indicated day represented as mean ± SD. o Tissues from j were homogenized and centrifuged. The H2O2 concertation of each tissue was analyzed by the H2O2 assay kit at 560 nm absorbance. p A schematic representation of ITPKB degradation by E3 ligase Trim25, participating in TMZ resistance of glioma through ROS homeostasis. Statistical significance is indicated as: **p < 0.001

To further confirm the tumor-promoting function of ITPKB in TMZ-resistant glioma, we examined tumor growth in vivo after treatment with the ITPKB inhibitor GNF362 (Fig. 7j). After 15 days of TMZ and/or GNF362 treatment, the combined treatment group exhibited a significantly better therapeutic response in xenograft tumors. Throughout the experiment, mice receiving combination treatment showed an expected decrease in tumor growth with no change in body weight (Fig. 7k–n). Additionally, the combination treatment group showed a significantly increased H2O2 level similar to that of ITPKB depletion (Fig. 7o). These data support the notion that ITPKB plays a vital role in regulating the chemotherapy response of TMZ-resistant glioma through ROS homeostasis.