Catalase Overexpression Drives an Aggressive Phenotype in Glioblastoma

Glioblastoma remains the deadliest form of brain cancer, largely because these tumors become resistant to standard of care treatment with radiation and chemotherapy. Intracellular production of reactive oxygen species (ROS) is necessary for chemo- and radiotherapy-induced cytotoxicity. Here, we assessed whether antioxidant catalase (CAT) affects glioma cell sensitivity to temozolomide and radiation. Using The Cancer Genome Atlas database, we found that CAT mRNA expression is upregulated in glioma tumor tissue compared with non-tumor tissue, and the level of expression negatively correlates with the overall survival of patients with high-grade glioma. In U251 glioma cells, CAT overexpression substantially decreased the basal level of hydrogen peroxide, enhanced anchorage-independent cell growth, and facilitated resistance to the chemotherapeutic drug temozolomide and ionizing radiation. Importantly, pharmacological inhibition of CAT activity reduced the proliferation of glioma cells isolated from patient biopsy samples. Moreover, U251 cells overexpressing CAT formed neurospheres in neurobasal medium, whereas control cells did not, suggesting that the radio- and chemoresistance conferred by CAT may be due in part to the enrichment of glioma stem cell populations. Finally, CAT overexpression significantly decreased survival in an orthotopic mouse model of glioma. These results demonstrate that CAT regulates chemo- and radioresistance in human glioma.


Introduction
Glioblastoma (GBM) is the most aggressive form of brain cancer, with a mean survival of only 14 months [1], even with the current standard of care treatment, which includes temozolomide (TMZ) combined with radiotherapy [2,3]. Although there is a significant initial benefit, treatment eventually becomes ineffective due to the development of tumor cell resistance. Indeed, resistance to radio-chemotherapy presents the most challenging barrier in the successful treatment of cancer and is one of the main phenomena underlying the failure to achieve a sustainable clinical benefit for patients with GBM [4,5].
Many factors have been associated with GBM resistance to radiation and chemotherapy, including tumor hypoxia, efficient and redundant DNA repair capacities [6], glioma stem cells (GSCs) [7], and increased expression of antioxidant enzymes that reduce the accumulation of reactive oxygen species (ROS) [8,9].
ROS comprises free radicals (superoxide, hydroxyl radical, singlet oxygen) and non-radical species (e.g., hydrogen peroxide [H 2 O 2 ]) that activate signaling pathways necessary for cell growth, proliferation, and differentiation [10,11]. However, excess cellular levels of ROS cause damage to important macromolecules, such as DNA, proteins, and lipids [12], which can lead to the activation of cell death signaling pathways [13,14]. Intracellular ROS levels are controlled by a complex network of antioxidant enzymes (e.g., superoxide dismutase and glutathione peroxidase) [15].
CAT is a key enzyme in the metabolism of H 2 O 2 . The molecular mechanisms regulating the expression of catalase (CAT) are not completely understood. Altered expression levels of CAT have been reported in tumor tissues. Specifically, the downregulation of CAT expression has been shown in pancreatic cancer cells and mouse skin tumors [16,17], but upregulation of CAT expression has been reported in melanoma [18], colon cancer [19], gastric adenocarcinoma [20], and GBM [21] cells. Importantly, altered CAT expression has been correlated with resistance to therapy in GBM [22,23], pancreatic cancer [9], and breast cancer cells [24]. In our previous studies, we demonstrated that glioma cell resistance to TMZ-induced oxidative stress is mediated by a reduction in the levels of mitochondrial ROS and enhanced antioxidant production and is associated with increased CAT activity [21,25,26].
Although studies by us and others have indicated an inverse correlation between the level of CAT expression or activity and resistance to therapy in different cancer cells, other studies have shown divergent results. In particular, controversy remains as to whether CAT is up-or downregulated in gliomas [22,[25][26][27][28]. Thus, it is important to understand the precise roles of CAT in GBM biology and resistance to therapy. In this study, we investigated the association between CAT expression in GBM tumors and patient survival, as well as the cellular and molecular mechanisms by which CAT influences therapeutic resistance in GBM.

Gene Expression Analyses
For comparisons between high-grade glioma and normal brain samples, CAT read counts of non-tumor (28 samples), mixed glioma (11 samples), oligodendroglioma (67 samples), astrocytoma (147 samples), and GBM (219 samples) were obtained from TCGA. Evaluation of TCGA glioma patient gene expression data and survival analysis were performed with the TCGA Rembrandt dataset. For Kaplan-Meier survival analysis, median gene expression was selected as the cutoff to split high and low expresser populations. The TCGA database can be downloaded from the GlioVis data portal (http://gliovis.bioinfo.cnio.es/, date accessed 8 September 2021).

Cell Culture and Electroporation
Glioma cells were cultured as we previously described [21,29,30]. Cells were electroporated using a Gene Pulser Xcell Electroporation System (BioRad, Hercules, CA, USA) under the following conditions: square wave pulse, 25 msec, and 140V. U251 cells were electroporated with CMV6 plasmids containing Myc-DDK-epitope-tagged human CAT or the pCMV6-Entry mammalian vector with C-terminal Myc-DDK Tag as the control (Catalog # RC210763 and PS100001, respectively; OriGene Technologies, Rockville, MD, USA). To generate cell lines stably overexpressing CAT, cells were selected with G418 (800 µg/mL) for 15 days. The stable lines isolated were characterized for the level of CAT by Western blot analysis and enzymatic activity.

CAT Activity
CAT activity was determined as previously described [28]. Briefly, cells were plated in 100-mm tissue culture plates and grown to 70-75% confluence. Cells were washed twice in phosphate buffered saline (PBS) and then collected and lysed in 50

Acquisition of Tissue Specimens and Tumor Dissociation
The protocol for this study was approved by the Institutional Review Board for Human Use at the University of Iowa (IRB 201103721). Patients with high-grade gliomas were identified by the University of Iowa, Tissue Procurement Core for inclusion in the Glioma Library Project. All patients provided written informed consent for the surgical procedures and gave permission for the use of resected tissue specimens, and all samples were de-identified to maintain confidentiality. Tumor tissues were dissociated into singlecell suspensions using a Brain Tumor Dissociation Kit (Miltenyi Biotec Inc., Auburn, CA, USA) according to the manufacturer's protocol. Single-cell suspensions were cultured in a neurobasal medium as we previously described [21,31]. Cell stocks were stored in liquid nitrogen until use.

Cell Proliferation and Anchorage-Independent Clonogenic Assays
For cell proliferation, cells were seeded into 6-well plates (2 × 10 4 cells/well). Cell number was assayed every 24 h for 5 days using a TC20 automated cell counter (BioRad, Hercules, CA, USA). For crystal violet cell proliferation assays [32], cells were fixed with ice-cold 3.7% paraformaldehyde, washed twice with PBS, and stored at −80 • C overnight. Cells were then stained with 0.05% CV solution and air-dried. The dye was solubilized with 10% acetic acid, and absorbance was measured at 590 nm. Anchorage-independent clonogenic assays were performed as we previously described [21,33].

Xenograft GBM Tumors
All surgical and experimental procedures and animal care practices were performed in compliance with the policies approved by the Institutional Animal Care and Use Committee of the University of Iowa. Establishment of intracranial tumors was performed as we previously described [21,34]. Any animal exhibiting signs of neurological deterioration was killed and the brain removed for examination. Paraffin-embedded tumor tissues were serially sectioned (5 µm), and sections were counterstained with hematoxylin and eosin (H&E, Abcam, Waltham, MA, USA, Cat. # ab245880), as we previously described [21].

Clonogenic Survival Assays
Clonogenic survival assays were performed as previously described [35]. Briefly, cells were plated and allowed to attach for 24 h, then treated with 50-250 µM TMZ or 0.5% DMSO (vehicle control) for 48 h followed by irradiation (2-8 Gy) at a dose rate of 0.65 Gy/min −1 using a 6000 Ci 137 Cs source. Cells were trypsinized immediately after treatment, counted, and seeded into 6-well plates at varying densities. The dishes were maintained in an incubator at 37 • C for 10 days to allow colony formation. The colonies were then fixed with 70% ethanol, stained with Coomassie blue, and counted (colonies containing > 50 cells were scored). Plating efficiency was determined by the following formula: (number of colonies formed/number of cells inoculated) × 100.

Cell Cycle Analysis
Cell cycle analysis was performed as we previously described [34]. Briefly, cells were treated with DMSO (control) or 300 µM TMZ for 48 h, then washed with PBS, trypsinized, and resuspended in PBS with 0.1% TritonX-100. PI and ribonuclease A were added at a final concentration of 10 µg/mL. Cells were analyzed by flow cytometry with a BD FACScalibur for DNA content, and the percentage of cells in the G1, S, and G2/M phases was established using ModFit LT (Verity Software House, Topsham, ME, USA).

Determination of Apoptosis
The apoptotic response after treatment with 250 µM TMZ or 10 µM staurosporine (positive control) was measured by flow cytometry as we described previously [29,30]. Treated and untreated cells were harvested, washed once with PBS, and stained with an Annexin V-PE Apoptosis Detection Kit (BD Pharmigen, San Diego, CA, USA) according to the manufacturer's instructions. Samples were analyzed by flow cytometry using a BD LSR II flow cytometer using λ Ex. 492-495 nm, λ Em. 517-527 nm. For each sample, 10,000 cells were analyzed, and the number of apoptotic cells was calculated with FlowJo Software Version 10 (FlowJo, LLC, Ashland, OR, USA).

In Vitro Limiting Dilution Assay
In vitro dilution assays were performed as we previously described [21,34,36]. Briefly, cells were plated at 1, 2, 5, 10, 25, 50, 100, and 200 cells per well in 96-well plates in the presence or absence of 3-AT (25 mM). Ten days after plating, the number of neurospheres in each well and the percentage of positive wells were quantified by manual counting. Extreme limiting dilution assay analyses were performed on the data as we previously described [21].

Statistics
Data were evaluated using GraphPad. Differences among the groups were tested using either unpaired two-tailed t-test or one-or two-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test, and probability (p) values were reported, with p < 0.05 indicating statistical significance. Experiments were performed in duplicate or triplicate and repeated twice or more to verify results. Data are expressed as the mean ± standard deviation (SD) and significance indicated as follows: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****), respectively.

Overexpression of CAT in Malignant Brain Tumors
We previously reported that CAT activity is elevated in TMZ-resistant glioma cells [21]. To investigate if CAT expression is clinically relevant in gliomas, we interrogated The Cancer Genome Atlas (TCGA) data accessed via GlioVis [37]. We compared CAT mRNA levels in glioma tumor tissue to levels in control brain tissues in datasets from Rembrandt cohorts comprising results from 537 samples. Compared with the expression in non-tumor tissue, CAT mRNA expression was significantly upregulated in astrocytoma, oligodendroglioma, mixed glioma, and GBM samples ( Figure 1A, Supplementary Table S1). Analysis of CAT mRNA in glioma samples of different grades showed that the level of CAT mRNA expression in the tumor correlated inversely with overall survival (OS) in patients (high median CAT expression: OS = 18.0 months; low median CAT expression: OS = 23.4 months; p = 0.0107 by log-rank test) ( Figure 1B). Together, these results confirm that human glioma tumors express high levels of CAT, and higher tumor expression of CAT is associated with poor prognosis. expression: OS = 23.4 months; p = 0.0107 by log-rank test) ( Figure 1B). Together sults confirm that human glioma tumors express high levels of CAT, and hig expression of CAT is associated with poor prognosis.

Stable Overexpression of CAT Results in Decreased Levels of Intracellular H2O2
To evaluate the effect of CAT overexpression on glioma phenotype, we sta fected U251 glioma cells with a vector encoding CAT cDNA. U251 cells stably t with an empty vector were used as the control. CAT protein expression was v Western blot analyses ( Figure 2A) and assays for enzymatic activity (H2O2 con ( Figure 2B). CAT expression in parental and vector-transfected cells is often w detectable by western blot under the tested conditions, confirming the low abu CAT in U251 cells. We selected U251 clone 3, which exhibited CAT activity 15 tim than that in control cells, for further analysis. Treatment with an irreversible in CAT [38], 3-amino-1,2,4-triazole (3-AT; 25 mM for 24 h), completely abrogated C ity in the CAT-overexpressing cells, demonstrating that the difference in CAT a tween the two cell types is due to CAT overexpression ( Figure 2C). To asses CAT overexpression contributes to ROS scavenging, cellular levels of H2O2 w ured by monitoring the oxidation of Amplex Red using a spectrofluorometer (Fl 3; Horiba Jobin Yvon, Edison, NJ, USA). As shown in Figure 2B, overexpressio induces a 10-fold decrease in H2O2 production compared with H2O2 formation transfected cells (1.61 ± 0.07 and 15.97 ± 0.65 pmoles/min/10 5 cells, respective ment with 3-AT (25 mM, 24 h) increases H2O2 production in CAT cells (12 pmoles/min/10 5 cells), indicating that H2O2 depletion is due to CAT overexpres ure 2D).

Stable Overexpression of CAT Results in Decreased Levels of Intracellular H 2 O 2
To evaluate the effect of CAT overexpression on glioma phenotype, we stably transfected U251 glioma cells with a vector encoding CAT cDNA. U251 cells stably transfected with an empty vector were used as the control. CAT protein expression was verified by Western blot analyses ( Figure 2A) and assays for enzymatic activity (H 2 O 2 consumption) ( Figure 2B). CAT expression in parental and vector-transfected cells is often weak or undetectable by western blot under the tested conditions, confirming the low abundance of CAT in U251 cells. We selected U251 clone 3, which exhibited CAT activity 15 times higher than that in control cells, for further analysis. Treatment with an irreversible inhibitor of CAT [38], 3-amino-1,2,4-triazole (3-AT; 25 mM for 24 h), completely abrogated CAT activity in the CAT-overexpressing cells, demonstrating that the difference in CAT activity between the two cell types is due to CAT overexpression ( Figure 2C). To assess whether CAT overexpression contributes to ROS scavenging, cellular levels of H 2 O 2 were measured by monitoring the oxidation of Amplex Red using a spectrofluorometer (FluoroMax-3; Horiba Jobin Yvon, Edison, NJ, USA). As shown in Figure 2B, overexpression of CAT induces a 10-fold decrease in H 2 O 2 production compared with H 2 O 2 formation by vector-transfected cells (1.61 ± 0.07 and 15.97 ± 0.65 pmoles/min/10 5 cells, respectively). Treatment with 3-AT (25 mM, 24 h) increases H 2 O 2 production in CAT cells (12.63 ± 0.58 pmoles/min/10 5 cells), indicating that H 2 O 2 depletion is due to CAT overexpression ( Figure 2D).

CAT Overexpression Promotes Cell Proliferation
Since ROS activates signaling pathways that contribute to the regulation of cell proliferation [25], we investigated whether CAT overexpression in glioma cells is associated with a change in cell proliferation. When adherent cell growth was assessed, the doubling time did not differ between CAT overexpressing cells and control U251 cells (18.0 ± 5.7 h and 21.5 ± 4.1 h, respectively; p = 0.68) ( Figure 3A). However, CAT overexpression markedly promoted colony formation in the anchorage-independent soft agar condition. In fact, colonies were present only in CAT-overexpressing cell cultures (209 ± 23) ( Figure 3B), supporting the notion that CAT can facilitate anchorage-independent growth. Treatment with 3-AT caused a 70% reduction in the number of colonies (63 ± 21) ( Figure 3B), suggesting that the decrease in intracellular ROS levels caused by CAT overexpression is tightly associated with the increase in soft agar colonies. To assess the clinical relevance of this finding, we investigated the effect of CAT inhibition on the proliferation of GSCs freshly isolated from human GBM samples. As shown in Figure 3C, treatment with 3-AT significantly inhibited cell proliferation in all patient-derived GSC lines tested. To determine if CAT expression influences host survival, CAT-overexpressing or control U251 cells were implanted orthotopically into the brains of nude mice. Median survival was significantly shorter in mice bearing CAT-overexpressing tumor cells (33 days versus 48 days in mice bearing control tumor cells [p = 0.0027]) ( Figure 3D). Notably, mice bearing CAT cells developed invasive tumors characterized by multifocal lesions throughout the brain parenchyma. In comparison, brains with vector-U251 tumors displayed only single lesions ( Figure 3E). Overall, these results suggest the pivotal role of CAT in promoting glioma tumorigenicity and anchorage-independent growth, a hallmark of the aggressive GBM phenotype. Graphs represent mean ± SEM from duplicate determinations from three independent experime * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 calculated by Student's t-test.

CAT Overexpression Promotes Cell Proliferation
Since ROS activates signaling pathways that contribute to the regulation of cell p liferation [25], we investigated whether CAT overexpression in glioma cells is associa with a change in cell proliferation. When adherent cell growth was assessed, the doubl time did not differ between CAT overexpressing cells and control U251 cells (18.0 ± 5. and 21.5 ± 4.1 h, respectively; p = 0.68) ( Figure 3A). However, CAT overexpression ma edly promoted colony formation in the anchorage-independent soft agar condition. fact, colonies were present only in CAT-overexpressing cell cultures (209 ± 23) (Figure 3 Graphs represent mean ± SEM from duplicate determinations from three independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 calculated by Student's t-test.

CAT Promotes Resistance to TMZ and Radiation
Because the standard of care treatment for GBM involves TMZ as a chemotherapy agent, we used the clonogenic assay to investigate the in vitro long-term survival of glioma cells after TMZ treatment. TMZ had a clear dose-dependent cytotoxic effect on control U251 cells ( Figure 4A); these doses were comparable in efficacy to TMZ's known cytotoxicity [29]. However, the cytotoxic effect of TMZ at any dose was mostly abolished in CAT-overexpressing U251 cells (cell survival rate with 250 µM TMZ, 93% and 10% for CAT-overexpressing and control cells, respectively) ( Figure 4A). To evaluate the effect of TMZ on cell proliferation, we used flow cytometry to measure cell cycle progression in CAT-overexpressing and control cells treated with 300 µM TMZ for 48 h. TMZ treatment led to S phase accumulation that reached 72.7% (p < 0.0001) at the expense of G 1 accumulation (16.6%) in control cells but did not affect cell cycle distribution in CAT-overexpressing cells ( Figure 4B). We also examined the induction of apoptosis using annexin V/propidium iodide (PI) staining and flow cytometry. Treatment with 300 µM TMZ for 48 h caused a 30% increase in the number of annexin V/PI-positive control cells but did not affect the number of annexin V/PI-positive CAT-overexpressing cells ( Figure 4C). Interestingly, CAT-overexpressing cells were also resistant to treatment with staurosporine (10 µM) ( Figure 4C), a well-known inducer of apoptosis [39], suggesting that CAT-overexpressing glioma cells are resistant not only to TMZ but to apoptotic stimuli in general.

CAT Promotes Resistance to TMZ and Radiation
Because the standard of care treatment for GBM involves TMZ as a chemotherapy agent, we used the clonogenic assay to investigate the in vitro long-term survival of glioma cells after TMZ treatment. TMZ had a clear dose-dependent cytotoxic effect on control U251 cells ( Figure 4A); these doses were comparable in efficacy to TMZ's known cytotoxicity [29]. However, the cytotoxic effect of TMZ at any dose was mostly abolished in CAT-overexpressing U251 cells (cell survival rate with 250 µM TMZ, 93% and 10% for CAT-overexpressing and control cells, respectively) ( Figure 4A). To evaluate the effect of TMZ on cell proliferation, we used flow cytometry to measure cell cycle progression in CAT-overexpressing and control cells treated with 300 µM TMZ for 48 h. TMZ treatment led to S phase accumulation that reached 72.7% (p < 0.0001) at the expense of G1 accumulation (16.6%) in control cells but did not affect cell cycle distribution in CAT-overexpressing cells ( Figure 4B). We also examined the induction of apoptosis using annexin V/propidium iodide (PI) staining and flow cytometry. Treatment with 300 µM TMZ for 48 h caused a 30% increase in the number of annexin V/PI-positive control cells but did not There is unequivocal pre-clinical and clinical evidence that ROS influence the genotoxic stress caused by ionizing radiation [40]. Given the important role of ROS production in mediating glioma cell sensitivity to radiation, we next examined the effect of radiation on clonogenic survival in the context of CAT overexpression. Treatment with radiation (2, 4, 6, or 8 Gy) dose-dependently decreased the survival fraction in control cells, but the effect at 4, 6, and 8 Gy was markedly diminished in CAT-overexpressing cells (p < 0.0001) ( Figure 4D). Pretreatment of CAT-overexpressing cells with 3-AT reduced the clonogenic survival after treatment with radiation (p < 0.0001) ( Figure 4D), further suggesting that CAT may protect glioma cells against radiation-induced toxicity.
in mediating glioma cell sensitivity to radiation, we next examined the effect of radiation on clonogenic survival in the context of CAT overexpression. Treatment with radiation (2, 4, 6, or 8 Gy) dose-dependently decreased the survival fraction in control cells, but the effect at 4, 6, and 8 Gy was markedly diminished in CAT-overexpressing cells (p < 0.0001) ( Figure 4D). Pretreatment of CAT-overexpressing cells with 3-AT reduced the clonogenic survival after treatment with radiation (p < 0.0001) ( Figure 4D), further suggesting that CAT may protect glioma cells against radiation-induced toxicity.

CAT Promotes Neurosphere Formation
Because we previously demonstrated that GSCs display higher CAT expression than non-tumor brain cells [21], we investigated whether CAT-overexpressing U251 cell cultures are enriched in GSCs when cultured in a neurobasal medium supplemented with epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF). CAT-overexpressing cells formed neurospheres ranging from 0.1 to 1.0 mm in diameter over the course of 15 days. In contrast, control cells failed to form neurospheres and attached to the culture dish ( Figure 5A). Furthermore, when plated in an in vitro limiting dilution assay, CAT-overexpressing cells formed neurospheres, but control cells did not ( Figure  5B). Treatment with 3-AT significantly reduced the frequency of GSCs in CAT-overexpressing cells (1 in 1.3 cells under control conditions versus 1 in 4.6 cells and 1 in 33.5 in cells treated with 5 mM and 25 mM 3-AT, respectively) ( Figure 5B). These results suggest a critical role for CAT in mediating glioma neurosphere formation.

CAT Promotes Neurosphere Formation
Because we previously demonstrated that GSCs display higher CAT expression than non-tumor brain cells [21], we investigated whether CAT-overexpressing U251 cell cultures are enriched in GSCs when cultured in a neurobasal medium supplemented with epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF). CAT-overexpressing cells formed neurospheres ranging from 0.1 to 1.0 mm in diameter over the course of 15 days. In contrast, control cells failed to form neurospheres and attached to the culture dish ( Figure 5A). Furthermore, when plated in an in vitro limiting dilution assay, CAToverexpressing cells formed neurospheres, but control cells did not ( Figure 5B

Discussion
Multiple mechanisms have been described to explain therapy resistance in g including drug inactivation, drug efflux, DNA damage repair, cell death inhibition, hypoxia, and increased expression of antioxidant enzymes [41,42]. We previously d strated that TMZ-resistance in glioma cells is due to more efficient mitochondrial co and reduced ROS production. Specifically, under conditions of oxidative stress, T sistant glioma cells generate substantially less ROS and more antioxidant enzyme TMZ-sensitive glioma cells [30]. Our previous studies also revealed increased exp of CAT, superoxide dismutase 2 (SOD2), and BMI1, a protein related to stemness an apy resistance, in TMZ-resistant glioma cells [21]. In this study, we further assess role of CAT upregulation in glioma cell resistance to therapy and demonstrate chemo-and radioresistance relies upon the regulation of hydrogen peroxide due, in part, to altered expression of CAT.
We found that human glioma cells stably overexpressing CAT express lower of intracellular H2O2 than control cells and are resistant to TMZ and radiation. In ingly, we did not observe differences in the rate of proliferation of CAT and contr in adherent culture conditions. However, CAT overexpression is associated with a icant increase in anchorage-independent proliferation, a hallmark of the aggressive phenotype, and enhanced capacity for self-renewal. These effects were abrogated b ment with 3-AT. Similar differences between adherent cell growth and anchorag pendent growth were previously observed in T98G [43], U87 and LN229 [44], an [45] glioblastoma cells.
In agreement with our results, Smith et al. demonstrated that CAT activity is

Discussion
Multiple mechanisms have been described to explain therapy resistance in gliomas, including drug inactivation, drug efflux, DNA damage repair, cell death inhibition, tumor hypoxia, and increased expression of antioxidant enzymes [41,42]. We previously demonstrated that TMZ-resistance in glioma cells is due to more efficient mitochondrial coupling and reduced ROS production. Specifically, under conditions of oxidative stress, TMZ-resistant glioma cells generate substantially less ROS and more antioxidant enzymes than TMZ-sensitive glioma cells [30]. Our previous studies also revealed increased expression of CAT, superoxide dismutase 2 (SOD 2 ), and BMI1, a protein related to stemness and therapy resistance, in TMZ-resistant glioma cells [21]. In this study, we further assessed the role of CAT upregulation in glioma cell resistance to therapy and demonstrated that chemo-and radioresistance relies upon the regulation of hydrogen peroxide due, at least in part, to altered expression of CAT.
We found that human glioma cells stably overexpressing CAT express lower levels of intracellular H 2 O 2 than control cells and are resistant to TMZ and radiation. Interestingly, we did not observe differences in the rate of proliferation of CAT and control cells in adherent culture conditions. However, CAT overexpression is associated with a significant increase in anchorage-independent proliferation, a hallmark of the aggressive cancer phenotype, and enhanced capacity for self-renewal. These effects were abrogated by treatment with 3-AT. Similar differences between adherent cell growth and anchorageindependent growth were previously observed in T98G [43], U87 and LN229 [44], and A172 [45] glioblastoma cells.
In agreement with our results, Smith et al. demonstrated that CAT activity is constitutively elevated in 36B10 rat glioma cells compared with normal astrocytes and has an important role in the resistance to oxidative stress and ionizing radiation [26]. Nanjaiah et al. also reported that activation of N-methyl-D-aspartate receptor (NMDAR) attenuated the intracellular ROS production and significantly enhanced the cell viability in LN18 and U251MG glioblastoma cells by increasing CAT activity. These results suggest that enhanced CAT activity contributes to glioma cell survival under exogenous oxidative stress [46].
Significant overexpression of CAT has been observed in a radioresistant variant clone (RRC) of U251 glioma cells. RRC showed an activation of up to 5-fold of antioxidant enzymes, such as SOD, CAT, glutathione peroxidase (GPX), and glutathione reductase (GR). Compared with parental U251 cells, RRC cells increase the activity of major antioxidant enzymes at times soon after radiation, suggesting a rapid scavenge of ROS to minimize the deleterious effects of radiation [23]. As Lee et al. demonstrated, exposure to radiation upregulated CAT and other oxidant enzymes and increased radio and chemoresistance. Our study demonstrated that overexpression of CAT is, at least in part, responsible for the resistant phenotype.
Significant overexpression of CAT has also been observed in human melanoma biopsy samples [18], colon cancer tissue [19], and gastric carcinoma cells [20]. In contrast, it was reported that the expression of antioxidant enzymes decreases during the progression from the normal pancreas to chronic pancreatitis to pancreatic cancer [16]. Similarly, CAT downregulation has been reported during skin cancer progression [17]. Interestingly, CAT overexpression in MCF-7 human breast cancer cells impaired cell proliferation and sensitized the cells to paclitaxel, etoposide, and arsenic trioxide [47]. The discrepancy between results may be due to cell type-specific responses or variations in experimental conditions. Even though our results clearly support a role for CAT in the mechanism of chemo-and radioresistance, we cannot exclude the possibility that the switch to a therapyresistant phenotype could be the result of a coordinate effect of multiple antioxidant enzymes. For example, it was recently reported that SOD 2 expression critically regulates tumorigenesis and the acquisition of TMZ resistance in GBM [48]. Additionally, it has been recently suggested that ROS regulates nuclear epigenetic modifications to promote cell survival and stemness [49][50][51][52]. Even though our results clearly support a role for CAT in chemo-and radioresistance, additional research is required to fully understand the underlying mechanism.
Here we also found that a shorter survival time was associated with orthotopic xenograft mice bearing glioma cell overexpression of CAT. Notably, CAT tumors are invasive and multifocal, while vector cells develop single lesions. Multifocal GBMs are rare and usually associated with worse outcomes compared with single lesion GBMs [53,54], supporting an important role of CAT in the development of more aggressive tumors.

Conclusions
Overall, this study provides strong evidence suggesting that overexpression of CAT in glioma cells leads to increased resistance to TMZ and radiation, suggesting an essential role of H 2 O 2 in the molecular mechanism by which these treatments lead to tumor cytotoxicity. As a crucial factor in the acquisition of TMZ and radiation resistance, pharmacological inhibition of CAT activity is a promising strategy for the treatment of malignant gliomas, including highly aggressive GBM tumors.