TRPML2 Mucolipin Channels Drive the Response of Glioma Stem Cells to Temozolomide and Affect the Overall Survival in Glioblastoma Patients

The survival of patients with glioblastoma (GBM) is poor. The main cause is the presence of glioma stem cells (GSCs), exceptionally resistant to temozolomide (TMZ) treatment. This last may be related to the heterogeneous expression of ion channels, among them TRPML2. Its mRNA expression was evaluated in two different neural stem cell (NS/PC) lines and sixteen GBM stem-like cells by qRT-PCR. The response to TMZ was evaluated in undifferentiated or differentiated GSCs, and in TRPML2-induced or silenced GSCs. The relationship between TRPML2 expression and responsiveness to TMZ treatment was evaluated by MTT assay showing that increased TRPML2 mRNA levels are associated with resistance to TMZ. This research was deepened by qRT-PCR and western blot analysis. PI3K/AKT and JAK/STAT pathways as well as ABC and SLC drug transporters were involved. Finally, the relationship between TRPML2 expression and overall survival (OS) and progression-free survival (PFS) in patient-derived GSCs was evaluated by Kaplan–Meier analysis. The expression of TRPML2 mRNA correlates with worse OS and PFS in GBM patients. Thus, the expression of TRPML2 in GSCs influences the responsiveness to TMZ in vitro and affects OS and PFS in GBM patients.


Introduction
Glioblastoma (GBM) is the most aggressive malignant type of primary brain tumor with a high mortality rate. Generally, GBM has poor clinical outcomes and survival is rarely greater than 15 months after diagnosis [1,2]. It has been demonstrated that despite maximal resection of well-demarcated tumors combined with irradiation and chemotherapy, most tumors will recur due to resistance to therapy. The inter-and intra-heterogeneity of GBM, which inhibits proper treatment, is indicated to be at least partially responsible for poor patient outcomes [1,2]. Furthermore, a substantial body of evidence supports the existence of glioma-initiating or -propagating cells within GBMs. A subpopulation of tumorigenic cells exhibiting stem-like characteristics, the glioblastoma stem-like cells (GSCs), was shown to be responsible for relapse, resistance to therapy, and tumor maintenance. GSCs have been isolated from both human tumor tissues [3][4][5] and several glioma cell lines [4,[6][7][8]. In addition to self-renewing and proliferating, GSCs can also initiate tumors upon secondary transplantation and generate progeny from multiple lineages [8]. As a result, GSCs assure the heterogeneity of GBM. In order to provide significant and personalized therapeutic

TRPML2 mRNA Expression during Differentiation of Neural Stem/Progenitor Cells (NS/PCs)
We previously reported that TRPML2 mRNA is more highly expressed in NS/PC than in normal brains (NHBs) and normal human astrocytes (NHAs) [18]. Here, by quantitative RT-PCR, we evaluated the TRPML2 expression during the differentiation of two different NS/PC (NS/PC#1 and NS/PC#2) lines. We found that TRPML2 mRNA levels strongly and progressively increase by a factor of 4.3 and 7.5 for NS/PC#1; or of 6.3 and 11.5 for NS/PC#2, at 7 and 14 days of differentiation, respectively, in 5% FBS ( Figure 1A). The relative TRPML2 mRNA expression in human neural stem/progenitor (NS/ PC#1 and NS/PC#2) cells was evaluated by qRT-PCR during the differentiation. TRPML2 mRNA levels were normalized for GAPDH expression. Data are expressed as mean ± SD. * p < 0.01 vs. time 0. (B) The relative TRPML2 mRNA expression was evaluated by qRT-PCR in undifferentiated GSC lines and after 14 differentiation days. TRPML2 mRNA levels were normalized for GAPDH expression. Data are expressed as mean ± SD. * p < 0.05 vs. undifferentiated cells.
The analysis of the TRPML2 mRNA expression in the GSC positive lines evidenced lower levels compared with the NS/PCs, suggesting that transformation of NS/PC in GSC can be correlated with TRPML2 reduction. The relative TRPML2 mRNA expression in human neural stem/progenitor (NS/PC#1 and NS/PC#2) cells was evaluated by qRT-PCR during the differentiation. TRPML2 mRNA levels were normalized for GAPDH expression. Data are expressed as mean ± SD. * p < 0.01 vs. time 0. (B) The relative TRPML2 mRNA expression was evaluated by qRT-PCR in undifferentiated GSC lines and after 14 differentiation days. TRPML2 mRNA levels were normalized for GAPDH expression. Data are expressed as mean ± SD. * p < 0.05 vs. undifferentiated cells.
The analysis of the TRPML2 mRNA expression in the GSC positive lines evidenced lower levels compared with the NS/PCs, suggesting that transformation of NS/PC in GSC can be correlated with TRPML2 reduction.
In addition, in contrast with NS/PCs where the TRPML2 mRNA expression progressively increases during the NS/PC differentiation, the differentiation of GSCs further increases the TRPML2 heterogeneity (12.5% negative, 25% ex novo induced; 18.7% reduced and 43.7% increased TRPML2 mRNA expression).

TMZ Resistance Is Associated with TRPML2 mRNA Expression in Undifferentiated and Differentiated GSCs Lines
The 14 days' GSC differentiation is associated with induction (GSC#1) or increase (GSC#28 and GSC#70) of TRPML2 mRNA expression. Thus, the effect of TMZ treatment (125, 250 and 500 μM) was evaluated by an MTT assay in undifferentiated and differentiated D-GSC#1, D-GSC#28 and D-GSC#70 ( Figure 2). We found that differentiation of GSC at 14 days, which increases the TRPML2 mRNA expression, impairs the sensitivity to TMZ treatment (GSC#1; IC50 >500 vs. 210, GSC#28 >500 vs. 290 and GSC#70 >1000 vs. 540 μM). No significant changes in TMZ resistance were evidenced in undifferentiated TRPML2-positive D-GSC#30, D-GSC#83, D-GSC#61 and D-GSC#195 compared to undifferentiated cells. Thus, changes (whether increase or reduction) in the TRPML2 mRNA expression, happening during GSC differentiation, parallel the increase of TMZ-resistance in TMZsensitive lines. On the other hand, TRPML2-positive GSCs were resistant to TMZ treatment and no changes in TMZ sensitivity were evidenced upon differentiation. Overall, these results suggest the existence of a relationship between TRPML2 expression and TMZ response in GSCs.

Changes in the TRPML2 mRNA Expression Result in Modulation of the Responsivity of GSC Lines to TMZ Treatment
To further address the relationship between the TRPML2 mRNA expression and drug resistance or sensitivity to TMZ treatment, we transfected pCMV-TRPML2 in the TRPML2 negative GSC#1 line (pCMV-TRPML2 proneural GSC#1) and silenced by RNA interference the TRPML2 mRNA expression in TRPML2-positive GSC#83 line (siTRPML2 GSC#83 line) ( Figure S1). The TRPML2 mRNA and protein levels after transfection were confirmed by qPCR and western blot analysis ( Figure S1). Cell viability results revealed that TRPML2 overexpression conferred increased resistance (IC 50 700 µM) to TMZ compared with pCMV GSC#1 control cells (IC 50 215 µM) ( Figure 3A). In GSC#83 TRPML2, silencing reduced the TMZ-resistance even though the IC 50 value remains high (IC 50 1 mM in siTRPML2 vs. 50 mM in siGLO) ( Figure 3B). No major differences in the TMZ cytotoxic effects were evidenced comparing untreated vs. pCMV or siGLO control GSC lines (data not shown). assay in TRPML2 negative GSC (A) and in TRPML2 positive GSC (B). Data shown are expressed as mean ± SE of three separate experiments. * p < 0.05 vs. GSC.
Thus, changes (whether increase or reduction) in the TRPML2 mRNA expression, happening during GSC differentiation, parallel the increase of TMZ-resistance in TMZsensitive lines. On the other hand, TRPML2-positive GSCs were resistant to TMZ treatment and no changes in TMZ sensitivity were evidenced upon differentiation. Overall, these results suggest the existence of a relationship between TRPML2 expression and TMZ response in GSCs.

Changes in the TRPML2 mRNA Expression Result in Modulation of the Responsivity of GSC Lines to TMZ Treatment
To further address the relationship between the TRPML2 mRNA expression and drug resistance or sensitivity to TMZ treatment, we transfected pCMV-TRPML2 in the TRPML2 negative GSC#1 line (pCMV-TRPML2 proneural GSC#1) and silenced by RNA interference the TRPML2 mRNA expression in TRPML2-positive GSC#83 line (siTRPML2 GSC#83 line) ( Figure S1). The TRPML2 mRNA and protein levels after transfection were confirmed by qPCR and western blot analysis ( Figure S1). Cell viability results revealed that TRPML2 overexpression conferred increased resistance (IC50 700 μM) to TMZ compared with pCMV GSC#1 control cells (IC50 215 μM) ( Figure 3A). In GSC#83 TRPML2, silencing reduced the TMZ-resistance even though the IC50 value remains high (IC50 1 mM in siTRPML2 vs. 50 mM in siGLO) ( Figure 3B). No major differences in the TMZ cytotoxic effects were evidenced comparing untreated vs. pCMV or siGLO control GSC lines (data not shown). Given that different sensitivity to TMZ and also the different TRPML2 expression was found in the analyzed cell lines, we assessed whether specific molecular pathways related to TMZ resistance [23] could be influenced by TRPML2 expression. Using the STRING online database (https://string-db.org (accessed on 8 November 2022)), a Protein-Protein Interaction (PPI) network was constructed ( Figure 4A). The analysis showed that both PI3K/Akt and JAK/STAT pathways can be modulated. Thus, to analyze that effect in more detail, the key protein expression levels were evaluated by western blot. Data showed that both pathways are regulated by the different expressions of TRPML2. There was a significant increase in Akt signaling in the pCMV-TRPML2 GSC#1 line with respect to control cells ( Figure 4B). Immunoblots confirmed that phosphorylation levels of Akt Given that different sensitivity to TMZ and also the different TRPML2 expression was found in the analyzed cell lines, we assessed whether specific molecular pathways related to TMZ resistance [23] could be influenced by TRPML2 expression. Using the STRING online database (https://string-db.org (accessed on 8 November 2022)), a Protein-Protein Interaction (PPI) network was constructed ( Figure 4A). The analysis showed that both PI3K/Akt and JAK/STAT pathways can be modulated. Thus, to analyze that effect in more detail, the key protein expression levels were evaluated by western blot. Data showed that both pathways are regulated by the different expressions of TRPML2. There was a significant increase in Akt signaling in the pCMV-TRPML2 GSC#1 line with respect to control cells ( Figure 4B). Immunoblots confirmed that phosphorylation levels of Akt (pAKT) increase even though the total Akt form remains unchanged. In addition, the Akt-mediated anti-apoptotic target Bcl-2 and BIRC5 are upregulated in GSC#1 transfected cells. Moreover, the expression level of STAT3 and its phosphorylation status were investigated in control and transfected cells ( Figure 4C). The levels of STAT3 were unaffected, while the levels pSTAT-Ser727 were increased in the pCMV-TRPML2 GSC#1 cells as compared with the control cells. On the other hand, the level of pSTAT3-Tyr705 was decreased. In siTRPML2 GSC#83 cells, we detected only an increase in STAT3-Tyr705. TRPML2-related proteins  . Key candidate proteins in TMZ resistance related to TRPML2 expression. (A) The proteinprotein interaction network derived from STRING connects the selected markers (PPI enrichment, p value = 0.01). (B) Total cell lysates were subjected to western blot analysis to detect the expression levels of AKT, pAKT, STAT3, pSTAT3-Tyr705, pSTAT3-Ser727, ADAR1, Survivin (encoded by Birc5) and Bcl2 using the specific antibodies as indicated, with GAPDH as the loading control. Representative images are shown from one of three independent experiments, which produced similar results. Densitometric analysis assessed the relative protein expression levels in three independent experiments. The results are the mean ± SD. Akt, STAT3, ADAR1, Survivin and Bcl2 densitometry values were normalized to GAPDH. (C) The pSTAT3-Tyr705 and pSTAT3-Ser727 protein levels were determined with respect to STAT3 levels. The pAkt protein levels were determined with respect to Akt levels * p < 0.05 vs. control cells.

Drug Transporter Genes Correlated to TRPML2 Expression in GSC Lines
To deepen the study of drug transporters and their relationship with the TRPML2 channel modulation, we evaluated the drug transporter gene expression in pCMV-TRPML2 overexpressed GSC#1 and in siTRPML2 GSC#83 lines, with respect to their control cells (Table S3).
The analysis of genes belonging to the ABC, SLC and AQP families evidenced that among the modified ones, there are six genes whose expression is significantly modulated in the GSC#1 transfected model and seven genes in the GSC#83 silenced model ( Figure 6A). In particular, SLC16A2, SLC2A1, SLC5A1, ABCC3, ABCB4, and ABCC11 expression decreases and SLC5A4 increases in pCMV-TRPML2 GSC#1 with respect to pCMV cells. SLC15A2 decreases and ABCB1, ABCB11, SLC10A1, ATP7B, ABCB5, SLCO1B1, and SLC2A2 increase in siTRPML2 GSC#83 with respect to siGLO cells. In general, an opposite trend of gene modulation is visible.
Based on the information in the STRING protein query from public databases, we made the PPI network of the identified 13 DEGs, TRPML2 and proteins involved in TMZ resistance modulated in our models on the basis of the western blot analysis ( Figure 4). All proteins, except for SCL15A2, ATP7B, ABCC11 and ABCB5, are interconnected, which supports the hypothesis of a link between TRPML2 and resistance to TMZ ( Figure 6B).

Correlation between the Tumor and Clinical Characteristics and OS and PFS
We evaluated the correlation between the OS or PFS and the clinico-pathological parameters investigated at the time of diagnosis at the Institute of Neurosurgery, Catholic University School of Medicine, in Rome (Italy) (Figure 7, Table S1). Age (<60 vs. >60 years), sex (male vs. female), tumor localization (temporal vs. parietal vs. frontal), Ki67 (low vs.

Discussion
The glioblastoma cell population is heterogeneous, with tumor-differentiated cells coexisting with subpopulations displaying stem cell characteristics. It is thought that GSCs derived from the normal NS/PC compartment [27] are responsible for recurrence and clinical relapse of glioblastoma [28][29][30]. Given that pharmacological modulation of the TRP ion channel activity in cancer cells is linked to their sensitivity to chemotherapeutic drugs [31], our goal was to examine the TRPML2 expression in GSCs and its relationship to resistance to TMZ, the standard chemotherapy for newly diagnosed GBM since 2005 and the subsequent use of the Stupp regimen [32]. The treatment of GBM with TMZ is not successful in over 50% of patient cases; however, there are few predictive markers beyond MGMT status for GBM patients treated with TMZ [23].
Using two NS/PCs lines and 16 GSC lines, we evidenced a lower expression of TRPML2 in GCSs with respect to normal cells and, above all, a different regulation during the differentiation process. While the normal cells show an increase in the channel expression levels, the GSCs do not seem to follow a single trend, but display an aberrant multipotent differentiation along neuronal, astroglial and oligodendroglial cell maturation [33]. We then exploited the different regulation of TRPML2 during differentiation in different GSC lines to assess whether there was a correlation with TMZ resistance. In those GSC lines where TRPML2 is expressed at higher levels, the resistance to TMZ treatment is even more marked. Instead, the TRPML2 down-regulation did not significantly change drug response. Thus, these data support the idea of a role for TRPML2 in tumor transformation and also in TMZ resistance. Moreover, TRPML2 overexpressed and silenced models were used. The expression of TRPML2 in TRPML2-negative GSC#1 increased the resistance of GSC to TMZ treatment; by contrast, silencing of TRPML2 mRNA in TRPML2-positive GSC#83 cells increases the sensitivity. Therefore, in GSCs, the more TRPML2 is expressed, the more resistant the cells are to TMZ.
Several mechanisms of TMZ resistance have been described to-date [23]. With STRING analysis, we assessed which of the key molecular pathways could be connected with the TRPML2 channel, and western blot analysis confirmed that PI3K/Akt and JAK/STAT pathways are modulated in our models. Dysregulation of the PI3K/Akt pathway occurs in up to 88% of GBM tumors and Akt, also known as protein kinase B, is a serine/threonine kinase that plays a crucial role in promoting chemoresistance in GBM cells [34]. Several downstream targets of Akt have been found to be implicated in specific mechanisms of TMZ resistance, including apoptotic regulators such as Bcl-2. The balance between the expression level of anti-apoptotic proteins and pro-apoptotic proteins determines the fate of cancer cells and the development of chemotherapeutic resistance [35]. Additionally, Survivin, a member of the inhibitor of apoptosis family [36], confers TMZ resistance by blocking the effect of TMZ-induced apoptosis. Furthermore, it has been demonstrated that TMZ sensitivity can be increased by targeting the Survivin gene [37]. Signaling through the JAK/STAT pathway stimulates the stemness of glial cells. Specifically, activating STAT3 is known to affect the transition from proneural to mesenchymal GBM type [38,39] that is associated with more aggressive and multitherapy-resistant features [40]. In support of this, mesenchymal GSC#83 cells express more STAT3 to begin with, have more of it activated and are more resistant to TMZ treatment than proneural GSC#1 cells. However, STAT3 can be differentially activated to regulate cancer cell phenotype and control their fate. The phosphorylation of a serine at position 727 in the absence of Tyr705 phosphorylation correlates with the survival of neural stem cells [41] and it is also characteristic of TMZresistant glioma cells [42]. These reports are in agreement with our result in that the expression of pSTAT3-Ser727 was increased in pCMV-TRPML2 GSC#1 cells that are more resistant to TMZ as compared with the control cells.
The STRING analysis also showed a correlation between TRPML2 and ADAR1. This protein is a dsRNA-editing enzyme that catalyzes the conversion of adenosine to inosine. Even though many editing sites in the micro-RNA transcriptome have been discovered [43][44][45], the overall biological significance of ADARs is still largely unknown [46]. In most cases, increased ADAR promotes cancer generation and progression; while in a few cancers, low expression and/or activity of ADAR mediates cancer phenotypes [47,48]. In melanoma cells, an impairment of ADAR1 activity promotes cancer cell growth and metastasis [48]. In breast cancer, the migration and invasion ability are related to ADAR1 expression [47]. In GBM, Jiang et al. demonstrated that ADAR1 contributes to GSC selfrenewal [49]. Furthermore, ADAR1 was shown to be involved in the impairment of TMZ resistance in glioma stabilizing glutaminase 2 (GLS2) mRNA, involved in the ferroptosis pathway by lipid metabolism [50]. Our work for the first time relates TRPML2 mediated regulation of ADAR1 expression to drug resistance, by which the lower the expression of ADAR1, the greater the resistance of cells.
Drug-resistance can also be attributed to an altered expression of ATP-dependent drug efflux pumps and drug efflux mediated by ABC and SLC transporters leading to a decreased cellular accumulation of anticancer drugs. This is considered a major drawback of currently applied chemotherapy regimens, and abnormal low expression of drug transporters in GBMs has been associated with tumor malignancy. Dysregulation in the endolysosome compartment is involved in mediating drug resistance [51] and several drug resistance transporters have been found in the endolysosomal system [52]. Since drug efflux capacity has been associated with stem cells derived from neoplastic tissues [53], we analyzed its involvement in TRPML2-mediated TMZ resistance in GSCs. Data demonstrated that changes in the TRPML2 expression markedly affected drug transporter gene expression in GSCs. Thus, ABCC3/MRP3, ABCB4/MDR3, ABCC11/MRP8, SLC16A2 SLC2A1/GLUT-1, and SLC5A1 were downregulated, and SLC5A4 was upregulated by TRPML2 gene transfection in GSC#1 cells; on the other hand, silencing of TRPML2 in GSC#83 induced ABCB1/MDR1/P-gp, ABCB5/MDR5, ABCB11/MRP11, SLCO1B1/OATP1B1, SLC2A2/GLUT-2, and SLC10A1/NTCP, and reduced SLC15A2 mRNA expression. Among these transporters, according to [33], ABCB4 significantly correlates with the CD133 stem cell marker expression and poor OS in GBM patients with the CD133 + GSCs, contributing to TMZ-resistance by exhibiting reduced responsiveness to TMZ, compared to CD133 − GSCs. ABCB1, which encodes for Pgp, is highly expressed in GSCs but TMZ treatment reduces its transcriptional activation [54]. Moreover, ABCC3 can be regulated by TMZ administration [55]. Instead, ABCC11 and SLC10A1 has no documented interaction with TMZ. SLC5A1, which encodes a member of the sodiumdependent glucose transporter, promotes ferroptosis [56], an alternative pathway of cell death that can be targeted to reverse TMZ resistance in glioma [57]. Furthermore, the function of influx transporters, in particular the solute carriers (SLC) in cancer cells, has been recently reassessed regarding cancer therapy. Indeed, the SLC transporters also serve as the uptake mediators of essential nutrients for tumor growth and survival [58]. Given their role as glucose transporters and the importance of a highly efficient glucose uptake for brain tumor-initiating cell growth, these proteins play a significant role in glioblastoma survival [59]. In light of their well-established contribution in promoting metabolism, elimination and detoxification of chemotherapeutic drugs, the regulation of some of the analyzed genes may seem a little surprising given their well-established role in reducing therapeutic effectiveness and treatment failure. Nevertheless, emerging research has demonstrated that drug carriers can impart either drug resistance or drug sensitivity, depending on the context. Additionally, rather than considering the expression of each single protein, in the context of a drug resistant or sensitive phenotype, it is also critical to consider the ratio of efflux (ABC) to influx (SLC) transporters [60].
In conclusion, the limited effectiveness of TMZ in GBM can be correlated with deficits in apoptosis induction, activation of multiple signaling pathways, and extrusion of drugs through cell membrane. However, although new therapeutic targets have been identified, the overall survival of GBM patients remains dismal due to tumor recurrence followed by chemoresistance. Increasing evidence has identified TRPML2 as a potential biomarker [18,19]. Herein, Kaplan-Meier analysis, supported by our in vitro models, correlates TRPML2-positive expression in patients' derived GSCs with poor OS and PFS.
It is well demonstrated that the sensitivity to TMZ is significantly associated with MGMT methylation status. The work by [61] shows that GBM patients harboring methylated MGMT promoters had a longer OS compared to unmethylated MGMT, suggesting a positive predictive value of MGMT methylation status in clinical response to TMZ; moreover, MGMT promoter methylation has been found to be associated with better OS and PFS in IDH mutant GBM patients [61,62]. However, irrespective of MGMT status, patients with TRPML2-positive GSCs showed lower OS and PFS. A strict relationship between EGFRvIII and TRPML2 expression was also observed. Longer OS and PFS were evidenced in EGFRvIII-positive vs. EGFRvIII-negative TRPML2-positive GSC patients, and longer OS values were observed in EGFRvIII-negative vs. EGFRvIII-positive TRPML2negative GSC patients. Moreover, lower Ki67 levels and negative TRPML2 expression correlate with longer OS and PFS, compared to low Ki67 TRPML2-positive and to high Ki67 TRPML2-positive patients. Finally, in both normal and mutated PTEN, TRPML2-positive vs. TRPML2-negative patients showed shorter OS and PFS. In agreement with our results, EGFRvIII-negative GBM neurosphere cells were more resistant to TMZ than the positive ones. EGFRvIII expression is associated with prolonged OS and PFS of GBM patients treated with surgery and radio/chemotherapy. Depletion of EGFRvIII in recurrent GBMs, as well as differential sensitivity to TMZ in vitro, indicates that the EGFRvIII-negative cells are involved in resistance to radio/chemotherapy [63,64]. Finally, in normal or mutated PTEN expressed GBM patients, Day et al. demonstrated that GBM patients with PTEN mutations exhibited a significantly shorter OS than those without PTEN mutations [65]. A positive correlation between Ki67 staining percentage and OS in GBM patients with IDH-WT has also been reported, with Ki67 staining > 20% predicting poorer PFS [66].
In conclusion, data in this study support a relationship between TRPML2 and prognosis in GBM patients. Notably, higher TRPML2 expression levels were found to be strongly related to TMZ resistance in patients' derived GSCs. Experimental analysis was performed to clarify the role of this channel and we demonstrated that TRPML2 promoted the chemoresistance of GSCs to TMZ affecting PI3K/AKT and JAK/STAT pathways, and drug transporters. We need to acknowledge that there are limitations to our approach. The first is the relatively small number (16) of GSC-derived GBM patients studied whose TRPML2 protein patterns certainly may not recapitulate the full repertoire that exists among GBM patients. Moreover, data reported in GSC#1 and GSC#83 lines represent TMZ sensitivities under cell culture condition, and the in vitro assay may be different from those observed in in vivo clinical conditions. Therefore, the accurate identification of different GSC types in high-grade GBM must be the upcoming task in order to eventually provide significant and personalized therapeutic strategies, instead of applying a standard cure to all patients with GBM.

MTT Assay
Three × 10 4 /mL cells were plated in 96-well plates and treated with different doses of TMZ (125-500 µM), alone or in combination. After that, the samples were incubated for another 3 h with 0.8 mg/mL of MTT. A microtiter plate spectrophotometer (BioTek Instruments, Winooski, VT, USA) was used to measure the color of solutions after the formazan crystals had been dissolved with 100 µL of DMSO per well. Four replicates were used for each treatment. IC 50 values correspond to the drug concentration that induces 50% of cell growth inhibition compared to control cells. IC 50 values were calculated using GraphPad Prism ® 5.0a (GraphPad Software, San Diego, CA, USA).

Gene Expression Analysis
Total RNA was extracted with the RNeasy Mini Kit (Qiagen), and cDNA was synthesized using the High-Capacity cDNA Archive Kit (Applied Biosystems, Foster City, PA, USA) according to the manufacturer's instructions. Quantitative RT-PCR (qRTPCR) was performed by using the IQ5 Multicolor real-time PCR detection system (BioRad, Milan, Italy). The reaction mixture contained the Advanced Universal SYBR Green Supermix (BioRad). Human TRPML2 and GAPDH RT 2 qPCR Primer assay (Qiagen) were used. The PCR parameters were 10 min at 95 • C followed by 40 cycles at 95 • C for 15 s and 60 • C for 40 s. All samples were assayed in triplicate in the same plate. The relative amount of target mRNA was calculated by the 2 −∆∆Ct method, using GAPDH as a housekeeping gene.

RT-PCR Profiler Array
Total RNA was extracted from pCMV-TRPML2 and pCMV GSC#1 lines and siTRPML2 and siGLO GSC#83 lines with the RNeasy Mini Kit (Qiagen) and reverse transcribed using the Reaction Ready first strand cDNA kit (Superarray Bioscience Corporation, Frederick, MD, USA). qRT-PCR was performed using the IQ5 Multicolor Real-time PCR detection system (BioRad), the RT 2 real-time SYBR Green PCR Master Mix and the human ABC transporter plates (Superarray Bioscience Corporation) according to manufacturer's instructions. The ∆∆Ct-based fold change and statistical significance analysis was performed using the Integrated Web-based Software Package for the PCR Array System at the GeneGlobe Data Analysis Center on the Qiagen website.

Western Blot Analysis
GSC#1 and GSC#83 cell lines were lysed in a lysis-buffer containing the protease inhibitor cocktail (Sigma-Aldrich, Milan, Italy). Proteins were separated on 10% SDS polyacrylamide gel, in a Mini-PROTEAN Tetra Cell system (BioRad). Protein transfer from the gel to a nitrocellulose membrane was carried out using Mini Trans-Blot Turbo RTA system (BioRad). Non-specific binding sites were blocked with 5% low-fat dry milk in phosphate-buffered saline 0.1% Tween 20. Membranes were incubated with anti-TRPML2, anti-Bcl2, anti-Survivin or anti-GAPDH (Santa Cruz Biotechnology) primary Abs for 1h at room temperature or with anti-STAT3, anti-pSTAT3 Tyr705, anti-pSTAT3 Ser727, anti-Akt, and anti-pAkt Ser473 primary Abs overnight at 4 • C followed by HRP-conjugated secondary Abs for 1 h at room temperature. The detection was performed using the LiteAblot PLUS (EuroClone, Milan, Italy) kits, and densitometric analysis was carried out by a Chemidoc using the Quantity One software (BioRad). For quantification, GAPDH was used as loading control. One representative out of three independent experiments is shown.

Protein-Protein Interaction (PPI) Network Analysis
The search tool for retrieval of interacting genes (STRING) (https://string-db.org (accessed on 8 November 2022)) database, based on known and predicted PPIs, was employed to seek potential interactions between the markers [69]. Text mining, experiments, databases, co-expression, species limited to "Homo sapiens", and an interaction score > 0.7 were considered as active interaction sources and applied to construct the PPI networks. The PPI network was visualized by Cytoscape software version 3.9.1. (https://cytoscape.org/, accessed on 10 November 2022). In the networks, proteins are schematized as nodes and interactions as edges.

Statistical Analysis
All data are expressed as mean ± SD. Student's t-test was used to assess differences between groups. Survival rates were analyzed using the Kaplan-Meier method and differences between groups were compared using the Mantel-Cox tests. All statistical tests are two-tailed. ROC analysis was used to stratify patients according to Ki67 levels (≤20 or >20). A p value less than 0.05 was considered statistically significant.

Conclusions
The GSCs are resistant to conventional therapy, and strategies designed to specifically target them or the pathways involved in stemness characteristics might be useful in the clinic. Therefore, more in-depth studies on resistance mechanisms are needed.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.