Glioblastoma multiforme is the most aggressive malignant primary brain tumor in adults. Several studies have shown that glioma cells upregulate the expression of xCT (SLC7A11), the catalytic subunit of system xc−, a transporter involved in cystine import, that modulates glutathione production and glioma growth. However, the role of system xc− in regulating the sensitivity of glioma cells to chemotherapy is currently debated. Inhibiting system xc− with sulfasalazine decreased glioma growth and survival via redox modulation, and use of the chemotherapeutic agent temozolomide together with sulfasalazine had a synergistic effect on cell killing. To better understand the functional consequences of system xc− in glioma, stable SLC7A11-knockdown and -overexpressing U251 glioma cells were generated. Modulation of SLC7A11 did not alter cellar proliferation but overexpression did increase anchorage-independent cell growth. Knockdown of SLC7A11 increased basal reactive oxygen species (ROS) and decreased glutathione generation resulting in increased cell death under oxidative and genotoxic stress. Overexpression of SLC7A11 resulted in increased resistance to oxidative stress and decreased chemosensitivity to temozolomide. In addition, SLC7A11 overexpression was associated with altered cellular metabolism including increased mitochondrial biogenesis, oxidative phosphorylation, and ATP generation. These results suggest that expression of SLC7A11 in the context of glioma contributes to tumorigenesis, tumor progression, and resistance to standard chemotherapy.
Implications: SLC7A11, in addition to redox modulation, appears to be associated with increased cellular metabolism and is a mediator of temozolomide resistance in human glioma, thus making system xC− a potential therapeutic target in glioblastoma multiforme. Mol Cancer Res; 14(12); 1229–42. ©2016 AACR.
This article is featured in Highlights of This Issue, p. 1171
Glioblastoma multiforme, a grade IV astrocytoma, is the most common and aggressive primary brain tumor in adults; patient survival averages only 14 months after diagnosis (1). The standard-of-care for patients with newly diagnosed glioblastoma multiforme includes aggressive safe tumor resection followed by radiotherapy with concomitant systemic chemotherapy using the alkylating agent temozolomide (2). Unfortunately, patients with high-grade gliomas inevitably progress or relapse an average of only 6.9 months after treatment (3). Therapeutic options for recurrent glioblastoma multiforme are limited and generally not effective. One of the main causes of treatment failure in patients with glioblastoma multiforme is resistance to postoperative radiation and chemotherapy. Many mechanisms contribute to the development of drug resistance, including DNA repair, drug uptake and efflux, apoptosis, and glutathione-mediated cellular detoxification pathways (4). Thus, an improved understanding of the molecular mechanisms involved in glioma progression and survival, as well as mediators of temozolomide resistance, could lead to development of more effective therapeutic strategies.
One of the mechanisms for chemo- and radiotherapy is to disproportionately increase intracellular reactive oxygen species (ROS) to induce cell-cycle arrest, senescence, and apoptosis (5). Accumulation of ROS can trigger apoptosis due to oxidative damage to DNA, macromolecules, lipids, and mitochondria. However, upregulation of antioxidant systems is observed in various tumors making them more resistant to chemotherapy (6). Approaches to maximally exploit ROS-mediated cell death by combining drugs that induce ROS generation with compounds that suppress cellular antioxidant capacity have been proposed years ago (7).
System xc−, a sodium-independent membrane transporter, couples the influx of extracellular cystine to the efflux of glutamate (8). Expression of system xc−, specifically the catalytic domain xCT, is upregulated in gliomas, and several studies have shown overexpression confers a growth advantage, either through increased extracellular glutamate levels that promote neuronal cell death or through increased import of cystine that is converted to cysteine (8–11). Cysteine is the rate-limiting precursor for generating the major antioxidant glutathione. Glutathione can neutralize intracellular ROS or be conjugated by glutathione S-transferases to xenobiotic agents, which are then exported out of the cell (12, 13).
There is accumulating evidence that glutathione and system xc− may mediate resistance of cancer to cytotoxin-based therapies (5). Radiation therapy–resistant glioma cells exhibited a 5-fold increase in the expression of antioxidant enzymes such as superoxide dismutase, glutathione peroxidase, and glutathione reductase that maintain redox balance (14). High intracellular glutathione levels in cancer cells have also been associated with drug resistance and detoxification of alkylating agents. System xc− has been shown to maintain intracellular glutathione levels in ovarian cancer cells, resulting in cisplatin resistance (15). Inhibition of system xc− in pancreatic cancer cells led to growth arrest and overexpression led to gemcitabine resistance (16). Recent findings also support the hypothesis of a correlation between adaption to oxidative stress, low mitochondrial ROS, enhanced mitochondrial respiration, and resistance to chemotherapy drugs (17).
In this study, we show that xCT (SLC7A11) was highly expressed in established glioma cell lines and that treating cells with sulfasalazine, an inhibitor of system xc−, decreased glioma growth and increased ROS-mediated cell death. Inhibiting system xc− in U251 glioma cells with sulfasalazine concomitant with temozolomide treatment had a synergistic killing effect. However, stable knockdown of SLC7A11 in U251 glioma cells did not alter cell growth or viability under basal conditions, despite changes in redox balance. Overexpression of SLC7A11 resulted in anchorage independence and resistance to oxidative stress, whereas SLC7A11 knockdown decreased anchorage-dependent cell growth and resistance to oxidative stress. Knocking down expression of SLC7A11 in U251 glioma cells also increased their sensitivity to temozolomide. In contrast, U251 cells in which SLC7A11 was overexpressed had decreased sensitivity to temozolomide and reduced apoptosis. Overexpression of SLC7A11 in U251 glioma cells also resulted in an upregulation of genes involved in cellular metabolism, increased mitochondrial biogenesis, oxidative phosphorylation, and ATP production while maintaining low cytoplasmic and mitochondrial ROS levels. These results suggest that high expression of SLC7A11/system xc− activity may confer resistance of glioma to temozolomide treatment by increasing glutathione production for redox balance and promoting cellular metabolism.
Materials and Methods
Human glioma cell lines (U251 and U87) and normal primary human astrocytes (pNHA) were purchased from ATCC and cultured as previously described (18). The primary high-grade glioma line PBT017 was obtained and cultured as previously described (18). For genotoxic or oxidative stress studies, cells were treated 24 hours after plating with either 300 μmol/L temozolomide for 72 hours or 100 μmol/L H2O2 for 6 hours.
Orthotopic transplantation and histopathologic analysis
In vivo studies were carried out in an orthotopic U251, U87, and PBT017 human glioma mouse model. Tumors were established by stereotactic, intracranial injection of 2 × 105 cells into the frontal lobe of NOD/SCID mice. At 4 weeks, mice were perfused transcardially with 4% paraformaldehyde in PBS. Brains were harvested, and formalin-fixed, paraffin-embedded sections were stained with hematoxylin and eosin to confirm the presence of tumors. To assess xCT protein expression in vivo, sections were incubated overnight with a goat anti-human polyclonal xCT antibody (LS-B4345, LifeSpan Biosciences, Inc.) followed by incubation with secondary antibodies conjugated to horseradish peroxidase (HRP) and detected using a DAB Peroxidase Substrate Kit (Vector Laboratories, Inc.)
RNA isolation, cDNA synthesis, and quantitative real-time PCR
Total RNA was extracted using TRIzol reagent. Synthesis of cDNA was performed using the Bio-Rad cDNA synthesis kit. SYBR green PCR master mix (Life Technologies) was used for quantitative real-time PCR (qRT-PCR) monitored with a C1000 Thermal Cycler (Bio-Rad) as previously described (19). Reaction conditions for qRT-PCR were as follows: 1 cycle of 3 minutes at 95°C; 39 cycles of 10 seconds at 95°C, 10 seconds at 55°C, 30 seconds at 72°C; 1 cycle of 10 seconds at 95°C; and a melting curve of 5 seconds at 65°C to 95°C. A standard linear curve was generated using pooled sample DNA, and the threshold exponential amplification cycle (CT) was calculated by system software. The following primer sequences were used: human SLC7A11: 5′-CTGAGGAGCTGCTGCTTTCAAA-3′ and 5′-AGGAGAGGGCAACAAAGATCGGAA-3′ and human GAPDH: 5′-ACCAAATCCGTTGACTCCGACCTT-3′ and 5′-TTCGACAGTCAGCCGCATCTTCTT-3′.
Western blot analysis
Total protein extraction and Western blot analysis were performed as previously described (9). The following primary antibodies were used: polyclonal goat anti-xCT (GTX89082; GeneTex) and rabbit monoclonal anti-active caspase-3 (ab32042; Abcam). Immunoreactivity was detected with a polyclonal rabbit-anti goat HRP-conjugated secondary antibody and a polyclonal rabbit-anti goat HRP-conjugated secondary antibody, respectively. Mouse β-actin (A1978; Sigma-Aldrich) was used as a loading control and detected with polyclonal goat- anti mouse HRP- conjugated secondary antibody (Cat. 1706516; Bio-Rad).
Cells were fixed with 4% paraformaldehyde and stained with a rabbit polyclonal antibody to xCT (NB-300-318; Novus Biologics). Immune complexes were detected with an AlexaFlour-488–conjugated secondary antibody (Molecular Probes). Nuclei were counterstained with DAPI (Vectastain). Images were acquired with an LSM 510 Meta Inverted 2-photon confocal microscope.
Production of shSLC7A11 and SLC7A11 U251 glioma cell lines
Lentivirus particles were produced by transfection of HEK 293T cells with either 15 μg of human TRC-pLKO.1-SLC7A11 shRNA (TRCN0000043123, TRCN0000043125, TRCN0000043126, TRCN0000288865, or TRCN0000380471), 15 μg of pLK01-non-targeting shRNA (Mission shRNA, Sigma-Aldrich), or 15 μg of a human SLC7A11-pLX304 plasmid (DNASU Plasmid Repository) using calcium phosphate co-precipitation. The culture medium was replaced with fresh 10% FBS in 1× DMEM after 8 hours and supernatant was collected 48 hours after transfection. After determination of viral titers, U251 cells were incubated with a viral vector containing the appropriate overexpressing RNA, shRNA, or control shRNA, using a multiplicity of infection of 0.5. Blasticidin (1.0 μg/mL; Sigma-Aldrich) or puromycin (10 μg/mL; Sigma-Aldrich) selection were used to obtain stable recombinant SLC7A11-overexpressing and shSLC7A11-knockdown U251 cells, respectively. Parental U251 cells served as controls for SLC7A11-overexpressing cells whereas cells transduced with an empty vector served as controls for the SLC7A11-knockdown cells.
Production of intracellular ROS under basal and treatment conditions was measured using the cell-permeant 2′,7′-dichlorodihydrofluorescein diacetate (H2DCFDA; Invitrogen). To evaluate the direct production of mitochondrial ROS in cells, MitoTracker Red CM-H2XRos, which is dependent on the mitochondria membrane potential (ΔΨm), was used. At the indicated time-points, cells were incubated (6% CO2, 37°C, 30 minutes) with either 5 μmol/L H2DCFDA or 500 nmol/L MitoTracker Red. Media were aspirated; cells collected with Accutase and spun down at 1,200 × g for 5 minutes. Cells were then resuspended in flow buffer (1% FBS in PBS) and analyzed using a BD Accuri C6 Flow Cytometer.
Glutamate was measured in media samples using the BioProfile 100 Plus (Nova Biomedical). SLC7A11-modified and control U251 cells were cultured in glutamate-free DMEM for 24 hours. After culture, 600 μL of media was removed from each sample dish and analyzed according to the manufacturer's instructions.
Assessment of mitochondria function
Mitochondrial function was examined by staining with the mitochondrial membrane potential (ΔΨm)-sensitive fluorochrome MitoTracker Red CMXRos. Cells (200,000 cells/well) were plated (12-well plate), cultured overnight, and then incubated (60 minutes, 37°C) with 500 nmol/L of MitoTracker Red CMXRos. After 2 washes with PBS, cells were fixed with methanol:acetone (3:1) for 10 minutes. Cells were washed twice in PBS, mounted in Dako Fluorescent Mounting Medium, and imaged on an LSM 510 Meta Inverted 2-photon confocal microscope.
Measurement of apoptosis
Apoptosis ratios were analyzed using the Alexa Fluor 488 AnnexinV/Dead Cell Apoptosis Kit (Invitrogen) according to the manufacturer's instructions. Samples were analyzed on a BD Accuri C6 Cytometer, and Annexin V−/propidium iodide (PI)− cells were used as unstained controls.
Quantification of total cellular ATP
To measure intracellular ATP, cells were lysed in buffer (200 mmol/L Tris, 2 mmol/L EDTA, 150 mmol/L NaCl, 0.5% Triton X-100) and CellTiter-Glo Luminescence Viability Assay (Promega) was performed according to the manufacturer's protocol. An ATP standard curve was generated by serial dilutions of a 1-mg ATP stock (Sigma-Aldrich). Luminescence was measured using a SpectraMax M3 (Molecular Devices).
At the indicated time points, cells were lysed with 200 μL of MES buffer [0.4 mol/L of 2-(N-morpholino) ethanesulphonic acid, 0.1 mol/L phosphate, 2 mmol/L EDTA, pH 6.0] and sonicated. Protein concentrations were quantified using the BCA Protein Assay (Thermo Scientific). A Glutathione Assay Kit (Cayman Chemical) was used to quantify total glutathione and glutathione disulfide (GSSG) according to the manufacturer's protocol. Absorbance was measured at 405 nm using a SpectraMax M3.
Cell viability and proliferation assays
Cell counting kit-8 (CCK-8; Dojindo Molecular Technologies) was used to measure cell viability according to the manufacturer's protocol. Absorbance was measured at 450 nm using a SpectraMax M3. A colorimetric immunoassay (Roche Diagnostics) was used according to the manufacturer's protocol to quantify cell proliferation on the basis of the measurement of bromodeoxyuridine (BrdUrd) incorporation during DNA synthesis. Absorbance was measured at 370 nm using a SpectraMax M3. Cell numbers at each time point were determined by flow cytometry (Guava EasyCyte, Millipore) using Guava Viacount (Millipore).
Soft-agar assay for anchorage-independent cell growth
Anchorage-independent growth was determined by seeding 2.5 × 104 cells per 12 well in 0.35% agar on top of a base layer containing 0.4% agar. Plates were incubated at 37 °C at 5% CO2 in a humidified incubator for 30 days and stained with 0.005% crystal violet for 1 hours. Colonies more than 0.1 mm in diameter were counted under a microscopic field at 10× magnifications.
Transmission electron microscopy
Cultured cells were pelleted and cryo-fixed in a Leica EM PACT2 high-pressure freezer (∼2,000 bars). In a Leica automated freeze substitution system AFS2, cryo-fixed specimens were freeze-substituted in anhydrous acetone containing 2% osmium tetroxide. The temperature progression was 8 hours at −90°C, −90°C to −60°C at 5°C/h, −60°C for 16 hours, −60°C to 0°C at 5°C/h. Cells were held at 0°C until time for further processing, when they are warmed to room temperature, rinsed in pure acetone, infiltrated, and embedded in Epon812 at 60°C for 48 hours. Ultra-thin sections (∼70-nm-thick) were cut using a Leica Ultra cut UCT ultramicrotome with a diamond knife, picked up on 200-mesh nickel EM grids. For morphology, grids were stained with 2% uranyl acetate in 70% ethanol for 1 minute followed Reynold lead citrate staining for 1 minute. Electron microscopy was done on an FEI Tecnai 12 transmission electron microscope equipped with a CCD camera.
Oxygen consumption and extracellular acidification rate
Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were determined using XF24 Extracellular Flux analyzer (Seahorse Bioscience). Briefly, cells (20,000/well) were plated into XF24 polystyrene cell culture plates (Seahorse Bioscience) and incubated for 24 hours in a humidified 37°C incubator with 10% CO2 (DMEM with 10%FBS). The following day, cells were washed, fresh assay media (2 mmol/L l-glutamine + 2 mmol/L pyruvate + 25 mmol/L glucose in XF Base Media; pH to ∼7.35) were added and cells incubated in a 37°C/non-CO2 incubator for 60 minutes prior to the start of an assay. Sensor cartridges were calibrated and loaded to dispense 3 metabolic inhibitors sequentially at specific time points: oligomycin (inhibitor of ATP synthase, 8 μmol/L/port), followed by FCCP (a protonophore and uncoupler of mitochondrial oxidative phosphorylation, 18 μmol/L/port), followed by the addition of rotenone (mitochondrial complex I inhibitor, 25 μmol/L/port). Basal OCR and ECAR were measured, as well as changes in oxygen consumption caused by the addition of the metabolic inhibitors described above. Cells were treated with trypsin and then counted to determine the cell number in each well after the assay. OCR and ECAR were reported as normalized rates (pmoles/cell for OCR and mpH/cell for ECAR) or expressed as a percentage of the baseline oxygen consumption. Each datum was determined minimally in replicates of 5. Several parameters were deducted from the changes in oxygen consumption including ATP turnover, proton leak, maximal respiratory capacity, and mitochondrial reserve capacity [= (maximum mitochondrial capacity) – (basal OCR)].
RNA sequence data generation and analysis
Sequencing libraries were prepared with TruSeq RNA Sample Preparation Kit V2 (Illumina, San Diego) according to the manufacturer's protocol, starting from 500 ng of total RNA from each sample, from which polyadenylated RNA was isolated, with the minor modification of using 10 cycles of PCR for library amplification. Approximately 40 million single-end raw reads were generated for each sample. The refSeq annotation for the hg19 version of the human genome was used to create a transcriptome Bowtie (20) index (version 0.12.7), to which reads were aligned with the following settings: “-v 3 -a”. Gene expression levels were estimated using eXpress (21) (version 1.4.1), and DESeq (22) was used for evaluating differential expression.
Experiments were performed in at least triplicate and repeated for at least 3 independent times. Statistical significance was determined by difference of means between 2 groups and was calculated using Student t test. All reported P values were 2-sided; P < 0.05 was considered significant.
SLC7A11/xCT is upregulated in glioma cell lines and inhibition induces ROS-mediated cell death
The expression of SLC7A11 and xCT in human astrocytoma cell lines was assessed by qRT-PCR and Western blot analysis, respectively. Both the U87 and U251 established astrocytoma cell lines had greater SLC7A11 gene (Fig. 1A) and xCT protein expression (Fig. 1B) as compared with pNHAs. Protein expression was also retained when U87 and U251 cells or a primary high-grade glioma line, PBT017, were orthotopically transplanted into NOD/SCID mice (Fig. 1C).
To determine whether system xc− transport activity is necessary for the survival of glioma cells, we treated U251 glioma cells with increasing doses of the pharmacologic inhibitor sulfasalazine. BrdUrd incorporation, which is a relative measure of cellular proliferation, showed a dose-dependent decrease with higher sulfasalazine doses (1,000 and 1,500 μmol/L) at multiple time points (Supplementary Fig. S1A). After 72 hours of sulfasalazine treatment, there was a dose-dependent decrease in U251 cell numbers (Fig. 1D). A majority of the cells were lost at the higher doses for the cell count assay; therefore, we chose to use 500 μmol/L sulfasalazine for subsequent assays. To determine whether the decrease in cell growth was due to apoptosis, Annexin V/PI staining was performed. Treatment of U251 cells with 500 μmol/L sulfasalazine resulted in a large increase in apoptotic cells (Supplementary Fig. S1B). We also reasoned that inhibiting system xc− transport activity should result in decreased amounts of intracellular antioxidants (e.g., cysteine and glutathione) and a subsequent increase in pro-oxidants. Intracellular ROS levels were measured by DCF staining to determine whether there was a more pro-oxidant state that could contribute to the increased cell death observed after sulfasalazine treatment. Indeed, treatment of the glioma cells with 500 μmol/L sulfasalazine resulted in significantly greater amounts of intracellular ROS compared with nontreated cells (Fig. 1E).
Synergistic cytotoxicity in glioma between the chemotherapeutic agent temozolomide and an inhibitor of system Xc−
We evaluated whether inhibition of system xc− could change the chemosensitivity or chemoresistance of U251 cells to temozolomide, the standard chemotherapeutic agent used to treat newly diagnosed patients with glioblastoma multiforme. We first examined whether pharmacologically inhibiting system xc− with sulfasalazine would increase the toxicity of temozolomide. Parental U251 cells were treated for 3 days with increasing doses of sulfasalazine (50, 100, 200, 400, 800, or 1,600 μmol/L) and increasing doses of temozolomide (12.5, 25, 50, 100, 200, 400, or 800 μmol/L). A combination index (CI) isobologram equation was used to quantitatively determine drug interactions, where CI < 1 indicates synergism, 1 indicates an additive effect, and >1 indicates antagonism (23). A synergistic effect was seen with all doses of temozolomide and 50 μmol/L of sulfasalazine as well as other lower doses of both drugs (Fig. 1F), indicating that inhibition of system xc− may sensitize glioma cells to chemotherapy. Sulfasalazine doses at 800 μmol/L or higher showed additive or antagonistic effects.
SLC7A11 overexpression confers resistance to oxidative stress
To gain insight in the function of SLC7A11 in human glioma, we established stable U251-overexpressing and -knockdown lines using lentiviral vector–mediated gene transfer and RNA silencing technology. The SLC7A11-overexpressing cells expressed 10-fold more SLC7A11 mRNA than did control cells (Fig. 2A), and this was accompanied by greater xCT protein expression, as shown by Western blot analysis (Supplementary Fig. S2A) and immunocytochemistry (Fig. 2B). Five different shRNA clones to target different regions of SLC7A11 were tested to obtain the most efficient knockdown compared with empty vector control. Because the shSLC7A11_ TRCN0000043126 construct suppressed gene and total protein expression the most in the transduced U251 cells (Supplementary Fig. S2B), which was confirmed by immunocytochemistry (data not shown), we chose it for further analysis. The SLC7A11-knockdown cell expressed significantly less SLC7A11 mRNA than did the cells transduced with an empty vector (Fig. 2A). Knockdown of xCT (SLC7A11) protein expression was confirmed by immunocytochemistry (Fig. 2B).
To assess whether modification of SLC7A11 expression affects system xc− transport activity, we measured glutamate release into the media. SLC7A11-overexpressing cells released significantly more glutamate compared with control cells, whereas SLC7A11-knockdown cells released significantly less, indicating the cystine influx:glutamate efflux transporter was functional (Fig. 2C). We found that enhancement or suppression of SLC7A11 expression in established cell lines did not influence cell proliferation (Fig. 2E) or cell viability (Fig. 2F). However, a soft agar assay revealed that the SLC7A11-overexpressing cells had higher anchorage-independent cell growth, whereas the SLC7A11-knockdown cells had lower anchorage-independent cell growth, indicating that the SLC7A11-overexpressing cells may be more tumorigenic (Fig. 2D).
Because system xc−, or more specifically its catalytic subunit xCT, plays an important role in regulating glutathione levels, we evaluated intracellular glutathione and ROS in the SLC7A11-modified U251 lines. Under basal conditions, the SLC7A11-knockdown cells had significantly lower intracellular glutathione levels than their respective controls (Fig. 3A) and this correlated with significantly higher intracellular ROS levels (Fig. 3B). Although the SLC7A11-overexpressing U251 cells did not have higher total intracellular glutathione levels than the control, despite increased xCT expression, they did have significantly lower intracellular ROS levels under basal conditions (Fig. 3B). However, under oxidative stress conditions, mimicked by treatment with 100 μmol/L H2O2, the SLC7A11-overexpressing cells had significantly increased intracellular glutathione levels as compared with under basal conditions and to control cells treated with H2O2 (Fig. 3C). Consistent with increased glutathione levels, H2O2-treated SLC7A11-overexpressing cells showed markedly less H2O2-induced intracellular ROS (Supplementary Fig. S3) and less H2O2-induced apoptosis than control cells (Fig. 3D).
SLC7A11 overexpression confers resistance to genotoxic stress induced by temozolomide
Various types of cancer exhibit elevated glutathione levels, which makes these cancers more resistant to chemotherapy (6). The combinatorial drug studies indicated that inhibition of system xc− with sulfasalazine could increase cell killing with temozolomide under low doses. We next examined whether genetic manipulation of SLC7A11 could modulate chemosensitivity to temozolomide. The SLC7A11-overexpressing cells had greater viability than did control cells treated with increasing doses of temozolomide, whereas the SLC7A11-knockdown cells were less viable than control cells (Fig. 4A). Consistent with this, the SLC7A11-overexpressing cells had a 6-fold higher IC50 value for temozolomide (419 μmol/L) as compared with control cells (64 μmol/L). The increased resistance to temozolomide upon SLC7A11 overexpression was further confirmed by higher cell proliferation rates compared with the control cells (Fig. 4B). In addition, the SLC7A11-overexpressing cells exhibited much less temozolomide-induced apoptosis when treated than did control cells, whereas the SLC7A11-knockdown cells showed a large increase in the percentage of apoptotic cells (Supplementary Fig. S4A). To further test whether SLC7A11 confers resistance to temozolomide by inhibiting apoptosis, we analyzed expression of cleaved caspase-3, a marker of apoptotic cells, by Western blot analysis. Expression of cleaved caspase-3 was markedly greater in all temozolomide-treated cell lines but was lower in the SLC7A11-overexpressing cells than in the SLC7A11-knockdown cells (Fig. 4C). To determine whether increased glutathione levels contributed to the increased chemoresistance of the SLC7A11-overexpressing cells, we measured intracellular glutathione levels cells exposed to 300 μm temozolomide. Consistent with the effect of oxidative stress on increasing glutathione levels (Fig. 3C) in SLC7A11-overexpressing cells, the genotoxic stress induced by temozolomide resulted in a significant increase in intracellular glutathione levels (Fig. 4D). Temozolomide treatment of the SLC7A11-knockdown cells did not alter glutathione levels as compared with untreated cells (Supplementary Fig. S4B).
Differential gene expression in SLC7A11-overexpressing and -knockdown glioma cells
RNA-sequence analysis was performed to assess gene expression changes upon SLC7A11 overexpression and knockdown in U251 glioma. Differential expression analysis identified more than 4,000 differential genes between SLC7A11 overexpression and respective control cells (Fig. 5A and B) and more than 3,000 differential genes when SLC7A11 is knocked down in U251 cells (Fig. 5A and Supplementary Fig. S5A). Gene Ontology (G) enrichment analysis revealed the functional categories enriched in the SLC7A11-overexpressing glioma (Fig. 5C) and in the SLC7A11-knockdown glioma compared with respective controls (Supplementary Fig. S5B), with a complete list of enriched categories included in the Supplementary Files. Representative GO terms indicate that a number of genes involved in mitochondrial biogenesis and mitochondrial function are upregulated in the SLC7A11-overexpressing glioma (Fig. 5C), suggesting that SLC7A11-overexpressing U251 cells may have altered metabolic function.
SLC7A11-overexpressing cells have increased mitochondrial biogenesis
Morphologic examination of mitochondria was performed utilizing a transmission electron microscope (TEM). The electron micrographs revealed relatively well-preserved mitochondria in the SLC7A11-modified and control cells (Fig. 6A). Magnification of the mitochondria (right of each set) shows typical double membranes, the intermembrane space, cristae, and a matrix. Quantification of the mitochondria in the cell lines reveals that the SLC7A11-overexpressing glioma has significantly higher number of mitochondria present than control cells (Fig. 6B), indicating that there is increased mitochondrial biogenesis. To further confirm increased mitochondria number and assess mitochondrial function, cells were stained with the mitochondrial membrane potential ΔΨm-sensitive fluorochrome, MitoTracker Red CMXRos. Fluorescent microscopic examination revealed a stronger MitoTracker Red CMXRos stain in the SLC7A11-0overexpressing glioma (Fig. 6C). Analysis of the mean fluorescence intensity showed significantly higher MitoTracker Red CMXRos stain in the SLC7A11-overexpressing glioma compared with control cells (Fig. 6D). The increased accumulation and retention of the probe indicates higher mitochondrial membrane potential (ΔΨm) in the SLC7A11-overexpressing glioma.
SLC7A11-overexpressing cells have increased oxidative phosphorylation
To assess cellular bioenergetics in the SLC7A11-overexpressing U251 cells, extracellular flux analysis was used to determine the oxygen consumption rate (OCR) which is a measurement of mitochondrial respiration. Basal respiration was significantly higher in the SLC7A11-overexpressing U251 glioma than in control cells (Fig. 7A). Basal respiration is usually controlled by ATP turnover and only partly by substrate oxidation and proton leak (24). Therefore, we next examined ATP turnover by inhibiting ATP synthase using oligomycin. Treatment with oligomycin revealed that the SLC7A11-overexpressing cells had higher ATP-linked respiration (oligomycin-sensitive fraction) compared with control cells (Fig. 7B). The maximal respiratory capacity was measured in the presence of carbonyl cyanide-p-trifluoromethoxyphenyl-hydrazon (FCC), an uncoupler that causes dissipation of the proton gradient by carrying protons across the inner mitochondrial membrane (17). FCCP causes rapid depolarization of mitochondria and acceleration of electron flux through the electron transport chain which resulted in significantly higher mitochondrial oxidative capacity in the SLC7A11-overexpressing glioma than in parental U251 cells (Fig. 7C). The spare respiratory capacity is the ability of substrate supply and electron transport to respond to an increase in energy demand, and, therefore, maintenance of some spare respiratory capacity is a major factor defining cell survival (24). Although not statistically significant, the SLC7A11-overexpressing U251 glioma had increased reserve capacity (P = 0.058) compared with the control cells (Fig. 7D). Together, these data suggest that the SLC7A11-overexpressing glioma, which display higher basal OCR, ATP-linked respiration, mitochondrial oxidative capacity, and spare respiratory capacity, have higher oxidative phosphorylation. Because mitochondria are one of the main producers of ROS and the primary producers of ATP, we performed analysis of ATP generation in the SLC7A11-overexpressing cells. Consistent with the OCR data, SLC7A11-overexpressing cells generate more ATP compared with control cells (Fig. 7E). Despite the increased ATP generation, the SLC7A11-overexpressing cells had lower mitochondrial ROS compared to the SLC7A11-knockdown cells or the control cells (Fig. 7F).
It has been suggested that system xc− may be a promising cancer target, as it may sensitize tumors to conventional chemo/radiation-based therapies by lowering glutathione levels (25). One of the causes of postoperative radiation and chemotherapy treatment failure in patients with glioblastoma multiforme is an increase in glutathione levels that may decrease chemotherapy-associated oxidative stress and play a role in glutathione-mediated cellular detoxification pathways. In this work, we have expanded on the role system xc− plays in temozolomide resistance in human glioma and have identified a novel mechanism that may contribute to glioma progression, via metabolic alteration. Inhibitors of system xc−, such as sulfasalazine, have been put forward as possible effective therapeutic options. Sulfasalazine has been shown to deplete glutathione levels by inhibiting the uptake of cystine, which lowers levels of intracellular cysteine, the rate-limiting precursor for glutathione synthesis (8).
Supporting use of system xc− inhibitors, we show that sulfasalazine treatment of U251 glioma resulted in ROS-mediated cell death. In addition, sulfasalazine treatment in combination with temozolomide treatment had a synergistic killing effect in U251 glioma cells. By inhibiting system xc− using sulfasalazine, the efficacy of temozolomide was increased at lower doses in vitro. Other pharmacologic inhibitors of xCT, such as erastin, have been shown to sensitize glioma cells to temozolomide (26). This suggests that use of specific system xc− inhibitors could have promise for avoiding or greatly reducing temozolomide-associated toxicity in patients with glioblastoma multiforme. However, other factors contributing to the decreased cell growth and viability cannot be ruled out, as sulfasalazine has been noted to target several other pathways (27). Sulfasalazine has several immunomodulatory effects, including inhibition of NF-κB and inhibition of glutathione S-transferase, which is responsible for conjugating glutathione to xenobiotics for detoxification (28). These off-target effects limit the conclusions drawn from sulfasalazine inhibition studies of system xc− in glioma cells. Clinical trials studying the effects of sulfasalazine for the treatment of malignant glioma in adults were terminated because of adverse events and toxicity (29).
We circumvented the problems associated with sulfasalazine inhibition of system xc− by generating stable SLC7A11-knockdown and -overexpressing U251 cell lines to better understand the role system xc− plays in glioma progression. In contrast to reports that inhibition of system xc− with sulfasalazine led to decreased growth and cell-cycle progression in glioma (8), we did not observe any changes in proliferation in the SLC7A11-modified lines. One possibility is that system xc− does not directly modulate cell growth and previous reports of transporter inhibition did not take into account the off-target effects associated with sulfasalazine use. In addition, it is possible that the SLC7A11-knockdown cells are able to take up enough cystine to support proliferation, whereas a complete SLC7A11 knockout would decrease glioma viability and growth. Although proliferation of the SLC7A11-knockdown cells was not affected, their cystine uptake was impaired as evidenced by the significant decrease in glutamate release and glutathione generation and significant increase in intracellular ROS. In terms of resistance to oxidative stress, overexpression of SLC7A11 conferred increased resistance to apoptosis, which we attributed to a significant increase in glutathione generation. This suggests that the primary role of system xc− is in redox regulation rather than protein synthesis.
Because elevated glutathione levels in cancer cells have been associated with drug resistance, we hypothesized that system xc− may play a role in modulating genotoxic stress induced by temozolomide. We found that overexpression of SLC7A11 reduced the sensitivity of glioma to temozolomide and decreased temozolomide-induced apoptosis. Disruption of system xc− function by knocking down SLC7A11 expression to low levels increased the sensitivity to temozolomide, resulting in pronounced apoptosis. The increased sensitivity may be partly due to the low levels of glutathione because only the SLC7A11-overexpressing cells exhibited increased glutathione production after treatment with temozolomide. In addition, marked activation of caspase-3 in knockdown cells confirmed that suppression of SLC7A11 facilitated temozolomide-induced activation of the apoptotic pathway.
It may be that the SLC7A11-knockdown cells exhibit the so-called “threshold concept for cancer therapy,” whereby an additional increase in ROS levels by ROS-generating agents, such as chemotherapy drugs, pushes the endogenous levels of ROS past a cellular tolerability threshold (30). Temozolomide has also been shown to generate ROS in human glioblastoma cell lines, including superoxides, cytosolic H2O2, and mitochondrial H2O2, which could be suppressed by pretreatment with antioxidants resveratrol, vitamin C, and iron (31). Concurrent treatment of glioblastoma cell lines with temozolomide and valproic acid (VPA), an anticonvulsant and mood-stabilizing drug, showed an increase in ROS and glutathione depletion resulting in higher apoptosis than in temozolomide or valproic acid alone (32). These data suggest that redox regulation and temozolomide-triggered ROS bursts can contribute to the sensitivity of glioma cells to chemotherapy. Indeed, the SLC7A11-knockdown cells exhibited a significantly greater amount of endogenous ROS under basal conditions, and temozolomide treatment augmented these levels because of impaired glutathione production.
Recently, it has been shown that glioma chemoresistance to temozolomide is linked to tighter mitochondrial coupling and low ROS production, indicating that chemoresistance is related to a remodeling of the electron transport chain (17). Because mitochondrial respiration generates ROS and system xc− has been shown to increase the antioxidant defense mechanisms to maintain redox balance, we hypothesized that SLC7A11 levels in glioma may have an impact on mitochondrial health and function. Overexpression of SLC7A11 in U251 glioma cells resulted in increased mitochondrial biogenesis and enhanced mitochondrial functions, indicated by increased mitochondrial membrane potential (ΔΨm). This was further confirmed by increased mRNA levels of several genes involved in mitochondrial biogenesis and function. While knockdown of SLC7A11 did not impair mitochondrial respiration, overexpression of SLC7A11 significantly increased the OCR which is reflective of increased OXPHOS and was accompanied by an increase in ATP generation. However, SLC7A11-overexpressing U251 glioma generated less cytosolic and mitochondrial H2O2 compared with SLC7A11-knockdown and control cells.
These results suggest that high expression of SLC7A11 may play a role in mitochondrial biogenesis and energetics. How SLC7A11 increases OXPHOS and the molecular signaling pathway involved have yet to be determined. The ability of cells to respond to stress is influenced by the energetic capacity of mitochondria, especially under conditions of increased energy demand. One possibility is that the SLC7A11-overexpressing cells rely more on oxidative phosphorylation due to an increased ATP demand. The generation of glutathione relies on 2 ATP-dependent steps: synthesis of γ-glutamylcysteine and then subsequent addition of cysteine to the γ-glutamylcysteine. Thus, the SLC7A11-overexpressing cells may rely on higher respiratory rates for glutathione production. Alternatively, it may be that the increased glutathione generation in the SLC7A11-overexpressing cells prevents the accumulation of ROS-mediated defects in the mitochondria, thereby promoting enhanced mitochondrial respiration. Mitochondrial DNA has a high mutation rate and, as mtDNA mutations increase, the energy capacity of the cell declines until there is insufficient energy to sustain cellular function, indicating that the bioenergetic threshold of the cell has been reached (33). The molecular mechanisms involved and how the model of chemoresistance and mitochondria metabolism can be integrated warrants further examination. Whether other cancer cells that overexpress SLC7A11 have increased mitochondrial respiration and whether OXPHOS is dependent on system xc− remains to be explored. However, our observations provide a potential new link between system xc−, redox balance, and OXPHOS to promote glioma progression and survival.
In conclusion, we have shown that overexpression of system xc− in glioma confers a survival advantage suggesting worse prognosis in patients with glioblastoma multiforme. Indeed, strong xCT/SLC7A11 expression in patients with glioblastoma multiforme correlated with an infiltrative phenotype on MRI and has been shown to be significantly associated with shorter progression-free and overall survival (34, 35). Microarray gene analysis of 60 human cancer cell lines used by the National Cancer Institute for drug screening (NCI-60) showed that SLC7A11 expression was negatively correlated with sensitivity of tumor cells to anticancer drugs (4). In addition, cytotoxic drugs have been shown to activate SLC7A11 expression in various cancer cells (36). Our data show that overexpression of SLC7A11 not only promoted resistance to oxidative stress but also implicated system xc− in temozolomide resistance and altered metabolism in glioblastoma multiforme. In U251 cells, overexpression of SLC7A11 promoted increased glutathione production under oxidative stress and genotoxic stress. The glutathione neutralized intracellular ROS and increased the survival of the cells once stressed. Overexpression of SLC7A11 was also correlated with an increased mitochondrial metabolism, which may contribute to the increased chemoresistance in these cells to temozolomide. Additional studies are warranted to explore the molecular mechanisms involved in the SLC7A11-mediated drug resistance and determine whether SLC7A11 may be a therapeutic target. Therefore, further investigation of manipulating the activity of this transporter either alone or in combination with other treatment modalities may lead to improved therapies and clinical outcomes of patients with glioblastoma multiforme.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: M.D. Polewski, R.F. Reveron-Thornton, G.A. Cherryholmes
Development of methodology: M.D. Polewski, R.F. Reveron-Thornton, G.A. Cherryholmes
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.D. Polewski, R.F. Reveron-Thornton, K.S. Aboody
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.D. Polewski, R.F. Reveron-Thornton, G.A. Cherryholmes, G.K. Marinov, K. Cassady
Writing, review, and/or revision of the manuscript: M.D. Polewski, R.F. Reveron-Thornton, G.A. Cherryholmes, K.S. Aboody
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.D. Polewski, R.F. Reveron-Thornton, K.S. Aboody
Study supervision: M.D. Polewski, K.S. Aboody
We acknowledge the technical support of the City of Hope RNAi Core (Dr. Claudia M. Kowolik), the Light Microscopy Digital Imaging Core (Dr. Brian Armstrong and Tina Patel), the Electron Microscopy Core, Megan Gilchrist for staining the glioma orthotopic xenograft sections, and Dr. Keely L. Walker for critical reading and editing of the article. Research reported in this publication included work performed in the Integrative Genomics Core supported by the National Cancer Institute of the National Institutes of Health under award number P30CA033572. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
This work was supported by funding from the California Institute of Regenerative Medicine (TG2-01150), the Rosalinde and Arthur Gilbert Foundation, STOP Cancer, and the Cancer Center Support Grant (P30CA033572).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.