mTORC1-Driven Tumor Cells Are Highly Sensitive to Therapeutic Targeting by Antagonists of Oxidative Stress

Abstract

mTORC1 is a central signaling node in controlling cell growth, proliferation, and metabolism that is aberrantly activated in cancers and certain cancer-associated genetic disorders, such as tuberous sclerosis complex (TSC) and sporadic lymphangioleiomyomatosis. However, while mTORC1-inhibitory compounds (rapamycin and rapalogs) attracted interest as candidate therapeutics, clinical trials have not replicated the promising findings in preclinical models, perhaps because these compounds tend to limit cell proliferation without inducing cell death. In seeking to address this issue, we performed a high-throughput screen for small molecules that could heighten the cytotoxicity of mTORC1 inhibitors. Here we report the discovery that combining inhibitors of mTORC1 and glutamate cysteine ligase (GCLC) can selectively and efficiently trigger apoptosis in Tsc2-deficient cells but not wild-type cells. Mechanistic investigations revealed that coinhibition of mTORC1 and GCLC decreased the level of the intracellular thiol antioxidant glutathione (GSH), thereby increasing levels of reactive oxygen species, which we determined to mediate cell death in Tsc2-deficient cells. Our findings offer preclinical proof of concept for a strategy to selectively increase the cytotoxicity of mTORC1 inhibitors as a therapy to eradicate tumor cells marked by high mTORC1 signaling, based on cotargeting a GSH-controlled oxidative stress pathway.

Introduction

The mammalian or mechanistic target of rapamycin complex 1 (mTORC1) senses and integrates signals from growth factors, nutrients, energy, and oxygen to regulate a wide range of biologic processes including mRNA biogenesis, protein and lipid synthesis, and autophagy (1). Deregulation of mTORC1 has been associated with a number of human diseases including cancer, genetic tumor syndromes, diabetes, as well as obesity (2, 3). Therefore, drugs that selectively target mTORC1, such as rapamycin, are considered to have a broad impact on a number of diseases, particularly in treating cancer. Although mTORC1 inhibitors (rapamycin and rapalogs) promote tumor shrinkage, clinical studies showed that tumors returned to their original states when rapalogs were discontinued, underscoring the cytostatic and not cytotoxic effects of these agents (4, 5). Thus, there is a critical need to develop alternative and novel approaches that could render tumor cell death. In this study, we chose to focus on a distinct subset of mTORC1-driven tumor cells, which bear mutations in the tuberous sclerosis complex (TSC)-2 tumor suppressor gene. The TSC tumor suppressor is a heterodimer complex, which is composed of tuberin (TSC2), a GTPase-activating protein (GAP), and its activation partner hamartin (TSC1). TSC inhibits the activity of Ras homolog enriched in brain (Rheb) by stimulating the conversion of Rheb-GTP to Rheb-GDP to suppress mTORC1 signaling (6).

To explore the possibility of selectively killing tumor cells with high mTORC1 signaling, we used a high-throughput screening approach and identified a set of small molecules that collaborate with rapamycin to suppress cell metabolism, growth and/or survival in Tsc2-deficient mouse embryonic fibroblasts (Tsc2^−/− MEF). As described below, we selected one combination strategy for further analysis. Specifically, we found that inhibition of mTORC1 together with glutamate cysteine ligase (GCLC) selectively triggers cell death of Tsc2^−/− MEFs but not wild-type (WT) Tsc2^+/+ MEFs. At a molecular level, Tsc1/2-deficient cells have elevated reactive oxygen species (ROS; ref. 7), which is counterbalanced by an mTORC1-driven increase in antioxidants via NADPH and glutathione (GSH) production. This delicate balancing act is required for cell survival. However, this elevated equilibrium also creates a potential vulnerability. Thus treatment of Tsc2-deficient cells with rapamycin plus nontoxic doses of GSH synthesis inhibitor l-buthionine sulfoximine (BSO) potentiates the generation of ROS. Our results indicate that Tsc1/2-deficient cells are more susceptible to this further increase of ROS, resulting in acute cell death. Importantly, this combination promotes tumor regression of a Tsc2-deficient Eker rat uterine leiomoma-derived (ELT3)-xenograft tumor model. Overall, our data suggest that targeting oxidative stress may be a strategy for the treatment of mTORC1-driven tumors.

Materials and Methods

Cell lines and culture

Tsc2^−/− p53^−/− and Tsc2^+/+ p53^−/− MEFs were authenticated and kindly provided by Drs. Brendan Manning and David Kwiatkowski (Harvard Medical School, Boston, MA). ELT3 cells were authenticated and kindly provided by Dr. Cheryl Walker (University of Texas, Austin, TX). ELT3-luciferase cells were described previously (8). MEFs and ELT3 cells were cultured in DMEM supplemented with 10% FBS (Invitrogen) and dialyzed for experiments (Gibco). Cells derived from renal angiomyolipomas from lymphangioleiomyomatosis patients (621-101 TSC2-null cells) were used and maintained in IIA complete medium with 10% FBS as described previously (9). All extra energetic additives that are often added to some DMEM formulations such as sodium pyruvate and succinate were excluded.

Small-molecule screening

Tsc2^−/− p53^−/− MEFs were cultured in glutamine-containing media with 10% FBS and penicillin/streptomycin in white 384-well plates. After overnight culturing, pin transfer of small-molecule library was performed at ICCB-Longwood screening facility. At 48-hour postcompound addition, the plates were allowed to equilibrate to room temperature for 1 hour. Then 30-µL CellTiter Glo (Promega) was added to each well. The plates were allowed to sit for 1 minute before being read on the EnVision Multilabel Plate Reader (Perkin Elmer).

Cell viability measurements

All cell viability experiments were conducted with propidium iodide (PI) exclusion assay as described previously (10). For all phase images, the Nikon Eclipse TE300 camera was used, and images were taken at the indicated time points.

Determination of ROS using dihydroethidium

Cells growin in 6-well plates were treated with compounds. After treatment, media were removed and cells were washed with 1X PBS. Fresh media containing 10 µmol/L dihydroethidium were added. After 30-minute incubation, media were removed and cells were washed with 1X PBS and tyrpsinized with 300-µL tyrpsin. 1.2 mL of media without the probe was added to the cells. ROS levels were measured by FACS.

GSH measurement

Cells grown in 6- or 12-well plates were harvested and trypsinized. The trypsinized cells were then resuspended in 0.5 mL of PBS containing 1% FBS, and incubated with 40 µmol/L monobromobimane (Biochemika) for 10 minutes at room temperature. After incubation, cells were placed on ice and the fluorescence at 485 nm (blue spectra) was measured by flow cytometry.

MitoSOX staining for live cell imaging

Cells grown in 6-well plates were treated with compounds. After treatment, media were removed and cells were rinsed with 1X PBS. Mitocondiral superoxide level was then measured according to the manufacturer's instructions (Thermo Fisher Scientific) using Nikon Ti confocal micrsocopy.

Transmission electron microscopy

Cells were fixed with 2% glutaraldehyde/2% formaldehyde in cacodylate buffer, followed by 1% osmium tetroxide. The samples were embedded in epoxy resin and viewed with a FEI Tecnai 12 transmission electron microscope operated at 80 kV.

Metabolite profiling and analysis

Cells were harvested in triplicate, and intracellular metabolites were extracted using 80% (v/v) aqueous methanol. Targeted LC/MS-MS was performed using a 5500 QTRAP triple quadrupole mass spectrometer (AB/SCIEX) coupled to a Prominence UFLC HPLC System (Shimadzu) with Amide HILIC chromatography (Waters). Data were acquired in selected reaction monitoring (SRM) mode using positive/negative ion polarity switching for steady-state polar profiling of greater than 260 molecules. Peak areas from the total ion current for each metabolite SRM transition were integrated using Multi-Quant v2.0 software (AB/SCIEX). Informatics analysis was carried out using MetaboAnalyst.ca free online software.

siRNA transfections

siRNAs (25 nmol/L; SMARTpool: siGENOME GCLC siRNA M-043462- 00-0005) were transfected in cells right after being seeded at a density of 30% to 50% confluency depending on experiments using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's protocols. Cells were harvested 36 to 60 hours after transfection as described in the figure legends.

Quantitative RT-PCR analysis

Total cellular RNA was purified from cultured cells using the RNeasy Mini Kit (Qiagen) following the manufacturer's protocol. For quantitative real-time PCR (qRT-PCR), RNA was reverse-transcribed using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) according to the manufacturer's instructions. The resulting cDNA was analyzed by qRT-PCR using the QuantiTect SYBR Green qPCR System (Qiagen). A QuantiTect Primer Assay was used to amplify the target gene. All reactions were run on an ABI 7900HT Fast Real-Time PCR instrument with a 15-minute hot start at 95°C followed by 40 cycles of a three-step thermocycling program: denaturation, 15 seconds at 94°C; annealing, 30 seconds at 55°C; and extension, 30 seconds at 70°C. Melting curve analysis was performed at the end of every run to ensure that a single PCR product of the expected melting temperature was produced in a given well. A total of 3 biologic replicates × 3 technical replicates were performed for each treatment group. Data were analyzed using the comparative C_t method (ΔΔC_t method).

GSH/GSSG Glo assay

Cells were plated at 1,000 per well in white 384-well plates and were allowed to attach overnight. Cells were then treated with compounds for 24 hours. Media were removed and cells were washed with PBS. GSH level was then measured according to the manufacturer's instructions (Promega) with measurement of lumniscence performed using an EnVision multimode plate reader.

Animal studies

All animal work was performed in accordance with protocols approved by the Institutional Animal Care and Use Committee-Boston Children's Hospital. Female intact CB17-SCID mice, 6 to 8 weeks of age, were purchased from Taconic. Eker rat uterine leiomyoma–derived Tsc2-deficient cells (ELT3) were developed by Howe and colleagues (11, 12). ELT3 cells were transduced with luciferase tag for bioluminescent imaging. For xenograft tumor establishment, 2.5 × 10⁶ cells were inoculated bilaterally into the lower back region of mice.

Three weeks after cell inoculation, mice bearing subcutaneous tumors were randomized into four groups: vehicle control, rapamycin (1 mg/kg/day in 0.25% PGE300 and 0.25% Tween 80; i.p., Monday–Friday), BSO (450 mg/kg/day in PBS, i.p., Monday–Friday), and rapamycin plus BSO (1 mg/kg/day, i.p. and 450 mg/kg/day, i.p., Monday–Friday). Drug treatment was initiated 3 weeks after cell inoculation. Body weight was measured every week. Tumor volume (length × (width)²/2) was measured weekly using a Caliper.

Bioluminescent reporter imaging

Ten minutes prior to imaging, animals were injected with luciferin (Xenogen; 120 mg/kg, i.p.). Bioluminescent signals were recorded using the Xenogen IVIS System. Total photon flux of tumors was analyzed (12).

IHC

Histology sections were prepared from xenograft tumor harvested from mice treated with different drugs after 10% formalin fixation and cutting into five 1–2 mm sections in cassettes. IHC was performed on paraffin-embedded 10-µm sections using antibodies against proliferating cell nuclear antigen (PCNA; Cell Signaling Technology). Slides were deparaffinized and antigen retrieval was performed using sodium citrate buffer (pH = 6.0). Sections were stained followed the instruction of SuperPicture 3^rd Gen IHC Detection Kit (Life Technologies).

After staining, slides were viewed on a Nikon Eclipse E400 microscope, and images captured using Spot Insight digital camera with Spot software (Diagnostic Instruments).

TUNEL staining

Histology sections were prepared from xenograft tumor harvested from mice treated with different drugs after 10% formalin fixation and cutting into five 1–2 mm sections in cassettes. Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining was proceeded followed the instruction of In Situ Apoptosis Detection Kit (TREVIGEN, 4810-30-K).

Statistical analysis

Data were expressed as mean ± SD of at least three independent experiments. An unpaired, two-tailed Student t test was used to determine differences between two groups (*, P < 0.05; **, P < 0.01; ***, P < 0.001) ANOVA test was used for the analysis of tumor regression among treatment groups.

Results

Identification of rapamycin collaborators through small-molecule high-throughput screening

In an effort to identify small molecules that collaborate with rapamycin to induce death in tumor cells with activated mTORC1, we conducted a small-molecule high-throughput screen in Tsc2^−/− MEFs. The compound library was composed of known bioactives, which represent >50% coverage of FDA-approved compounds. The cells were treated with each test compound with or without rapamycin, in a 384-well plate format. After 2 days of incubation at 37°C, cell growth was determined using the CellTiter-Glo (Promega; Fig. 1A). We screened 8,247 known bioactives in duplicate, and the effect of each compound on the cellular response to rapamycin was determined on the basis of fold changes in cell growth (Fig. 1B and C). We then cherry-picked 82 "hits" and retested these compounds at 12 different doses. As a result, we identified 32 compounds (Table 1), and interestingly, several of which have been validated previously in the literature. For example, Goncharova and colleagues reported that the combination of simvastatin and rapamycin reduces TSC-null tumor growth in nude mice (13). Also, rapamycin has been shown to synergize with the EGFR inhibitor erlotinib in non–small cell lung, pancreatic, colon, and breast tumors (14). We next selected compounds for follow-up studies and we focused on the combination effect of BSO, an irreversible inhibitor of GCLC, in collaboration with rapamycin in mTORC1-driven tumor cells. GSH is synthesized by ligation of glutamate with cysteine to form γ-glutamylcysteine followed by a second ligation to glycine. The first and rate-limiting enzyme of GSH biosynthesis is composed of a catalytic (GCLC) and a modulator subunit (GCLM). The second reaction is catalyzed by GSH synthetase to produce the reduced form of GSH, which is a major and abundant cellular antioxidant (Fig. 1D; ref. 15). As shown in Fig. 1E, compared with single treatment with BSO alone, a significant decrease in cell growth was observed after treatment with BSO and rapamycin in a dose-dependent manner, whereas rapamycin alone caused about 44% decrease in cell growth. Synergism was assessed by applying the Chou–Talalay method. The combination of BSO and rapamycin showed synergistic effect with a combination index (CI) <1.0 (Supplementary Fig. S1A).

Figure 1.

Identification of rapamycin (Rapa) collaborators through small-molecule high-throughput screening. A, a schematic shows the small-molecule high-throughput screening. B, the plate-to-plate reproducibility of the assay is shown for all the small molecules alone from the screen. A linear fit of the data gives an R² of 0.9, indicating good agreement between the two replicate plates. The positive controls are shown in red; the negative controls are shown in blue; the test compounds are shown in turquoise. C, the plate-to-plate reproducibility of the assay is shown for all the small molecules plus rapamycin from the screen. A linear fit of the data gives an R² of 0.9, indicating good agreement between the two replicate plates. The positive controls are shown in red; the negative controls are shown in blue; the test compounds are shown in turquoise. D, a schematic shows the enzymes involved in GSH biosynthesis and the inhibitor used in this study (see text for more details). E, cell viability (measuring ATP levels) of Tsc2^−/− MEFs was measured via Cell TiterGlo. Relative luminescence was measured in Tsc2^−/− MEFs after 48 hours of treatment with increasing doses of BSO with or without rapamycin as indicated. The mean is shown; error bars, SD (n > 3).

Table 1.

Identification of rapamycin collaborators through small-molecule high-throughput screening

Compound	IC50 without Rapa (µmol/L)	IC50 with Rapa (µmol/L)
l-buthionine sulfoximine	Inactive	2.1
Cerivastatin	Inactive	0.03
Simvastatin	Inactive	3.0
Lovastatin	Inactive	7.7
Clothiapine	Inactive	1.5
Dopamine HCI	Inactive	10.1
Pioglitazone	23.4	0.5
Retinoic acid	Inactive	25.2
Capsaicin	Inactive	5.8
Mycophenolic acid	70.8	3.0
Fumagillone	Inactive	18.8
Erlotinib	Inactive	10.5
Quetiapine fumarate	Inactive	11.1
Glyburide	Inactive	19.7
Diclazuril	93.5	1.8
Meloxicam	Inactive	3.9
Indomethacin	Inactive	1.4
Gallic acid	Inactive	15.5
Sulindac	75.2	0.8
Captopril	Inactive	28.3
AL-8810	11.8	0.1
Tannic acid	66	1.6
L-670596	Inactive	7.7
Sulindac sulfone	96	1.8
6-α-methyl-11-β   hydroxyprogesterone	Inactive	4.5
19-Norethindrone	Inactive	5.5
Praziquantel	Inactive	12.4
Trovafloxacin mesylate	50	0.08
FPA124	Inactive	4.2
Tyrphostin 47	27.4	0.9
MDL 28170	39.4	1.0
Spironolactone	27.8	1.3

NOTE: IC₅₀ values for rapamycin collaborators alone or with rapamycin.

Abbreviation: Rapa, rapamycin.

Inhibition of mTORC1 and GCLC induces cytotoxicity in Tsc2^−/− MEFs

The assay used for the small-molecule screen measures the bioenergetic state of the cell. Therefore, to further examine the biologic consequences of targeting mTORC1 and GCLC, we first examined their effects on the proliferation of Tsc2^−/− MEFs. We observed BSO alone (2 µmol/L) to have little effect on cell proliferation. Rapamycin (20 nmol/L) suppressed proliferation significantly and the combination completely blocked cell proliferation by this assay (Fig. 2A). We next evaluated the effect of BSO, rapamycin, or BSO plus rapamycin on cell morphology. Consistent with the data in Figs. 1E and 2A, phase contrast imaging showed rapamycin (20 nmol/L, 48 hours) appeared to suppress proliferation, whereas BSO (0.5 µmol/L and 1 µmol/L, 48 hours) had no effect. Importantly, the combination resulted in a dramatic cell death phenotype (Fig. 2B). We therefore measured cell death with a PI-exclusion assay (Fig. 2C). FACS analysis showed that BSO induced dose-dependent cytotoxicity with rapamycin in Tsc2^−/− MEFs. Importantly, the RNAi-mediated knockdown of GCLC also induced cell death in Tsc2^−/− MEFs in the presence of rapamycin (Supplementary Fig. S2B) providing support for the on-target effect of BSO. Rapamycin is known to mildly induce autophagy in mammalian cells. Interestingly, we observed increased LC3-II levels in Tsc2^−/− MEFs treated with rapamycin (20 nmol/L) plus BSO (10 µmol/L) compared with WT MEFs after 24-hour culture (Fig. 2D). We also showed that autophagy inhibitors suppress BSO+rapamycin-induced cell death (Fig. 2E). In addition, similar to other published data (16, 17), we demonstrated that the addition of autophagy inhibitors to Tsc2^−/− MEFs treated with BSO+rapamycin, which block the last step of autophagy, led to an increased accumulation of LC3-II (Fig. 2F). When we compared the viability of Tsc2^−/− MEFs with Tsc2-WT MEFs treated with rapamycin (20 nmol/L) and BSO (1, 2, 10 µmol/L), the combination treatment was not toxic to WT MEFs (Fig. 2B and G and Supplementary Fig. S2A). We also examined the combination effect of BSO and rapamycin in Tsc2-deficient ELT3 (ELT3-V3) cells. Similarly, BSO dose-dependently reduced cell viability in combination with rapamycin in ELT3 cells (Supplementary Fig. S2C). Importantly, while decreased cell viability with BSO and rapamycin was observed in ELT3-V3 cells, this was not observed in TSC2-reexpressing ELT3 cells (Supplementary Fig. S2D). Similarly, the combination treatment induced cell death in lymphangioleiomyomatosis patient–derived 621-102 cells but not in TSC2-add back 621-103 cells (Supplementary Fig. S2E). Thus, although rapamycin but not BSO reduced cell proliferation, only rapamycin plus BSO-induced cell death by PI-exclusion assay.

Figure 2.

Inhibition of mTORC1 and GCLC induces cytotoxicity in Tsc2^−/− MEFs. A, cell proliferation was measured in Tsc2^−/− MEFs treated with BSO (2 µmol/L) and rapamycin (Rapa; 20 nmol/L) over 96-hour period of time by measuring cell number with a Coulter counter. B, Tsc2^−/− MEFs and WT MEFs were treated with DMSO, rapamycin (20 nmol/L), BSO (0.5 µmol/L, 1 µmol/L or 2 µmol/L) as indicated for 48 hours. Phase microscopy was used to observe cell viability. C, cell death of Tsc2^−/− MEFs after 48 hours of treatment with increasing concentrations of BSO with or without rapamycin (20 nmol/L) was measured via PI exclusion assay. The mean is shown; error bars, SD (n = 3). D, immunoblot analysis of LC3, p-S6, S6, and actin in Tsc2^−/− MEFs and WT MEFs treated with the indicated compounds for 48 hours. E and F, Tsc2^−/− MEFs were treated with DMSO, rapamycin (20 nmol/L), and BSO (10 µmol/L) for 24 hours, after which autophagy inhibitors, chloroquine (10 µmol/L, 100 µmol/L) or bafilomycin A1 (50 nmol/L, 100 nmol/L) were added. After 24 hours, phase microscopy was used to observe cell viability (E) and immunoblot analysis was performed (F). G, cell viability (measuring ATP levels) of WT MEFs was measured via Cell-Titer-Glo. Relative luminescence was measured in WT MEFs after 48 hours of treatment with increasing doses of BSO with or without rapamycin as indicated. The mean is shown; error bars, SD (n = 3).

Elevated levels of ROS are responsible for cell death in Tsc2^−/− MEFs

GSH is synthesized in the cytosol and then distributed into different compartments, such as mitochondria, endoplasmic reticulum (ER), and nucleus (18), and BSO inhibits the synthesis of the reduced form of GSH (Fig. 3A). We examined the GSH levels in Tsc2^−/− MEFs treated with BSO, rapamycin alone, or the combination. As expected, addition of BSO to Tsc2^−/− MEFs caused a decrease in GSH levels. Interestingly, cells treated with rapamycin also exhibited reduced the levels of GSH (Fig. 3B). Consistently, we observed decreased GSH levels in Tsc2^−/− MEFs treated with rapamycin by mass spectrometry (Supplementary Fig. S3A). Recently, our group reported that mTORC1 positively regulates glutaminase (GLS) and glutamine flux through this enzyme (19). As GLS converts glutamine to glutamate, which is a precursor for GSH synthesis, it is likely that rapamycin contributes to the decrease of GSH levels in Tsc2^−/− MEFs by suppressing glutamine–glutamate production through reduction of GLS production. Importantly, the combination treatment led to further decrease in GSH levels relative to single-agent treatment (Fig. 3B). It has been shown that mTORC1 stimulates the pentose phosphate pathway (PPP), and mTORC1 induces G6PD gene through the transcription factor sterol regulatory element-binding transcription factor 1 (SREBP1; ref. 20). G6PD is the first and rate-limiting enzyme of PPP, and plays a critical role in protection against oxidative stress (21). Oxidized glutathione (GSSG) is reduced to GSH by NADPH, generated by G6PD (Fig. 3A). Here we also show that rapamycin decreased the GSH/GSSG ratio (Supplementary Fig. S3B) in Tsc2^−/− MEFs. As GSH is a major intracellular antioxidant against oxidative stress, we then asked whether the elevated levels of ROS are responsible for the combination-induced cell death. As shown in Fig. 3C, BSO together with rapamycin showed an increase in ROS production levels compared with single-agent treatment in Tsc2^−/− MEFs. It has been reported that the cellular levels of ROS were significantly higher in TSC2-deficient cells than TSC2-reexpressing cells (7, 22). Thus, these observations suggest that increased ROS levels, as a result of BSO plus rapamycin, are contributing to death in Tsc2^−/− MEFs. To further examine this hypothesis, we used antioxidants, N-acetyl-cysteine (NAC) and vitamin C. The addition of NAC and vitamin C were able to reverse cell death induced by BSO and rapamycin as examined by phase contrast imaging and PI-exclusion assay (Fig. 3D and E). Moreover, the addition of GSH reduced ethyl ester (GSH-MEE), a membrane-permeable derivative of GSH, also rescued the viability of Tsc2^−/− MEFs treated with BSO and rapamycin (Fig. 3D and E).

Figure 3.

Elevated levels of ROS are responsible for cell death in Tsc2^−/− MEFs. A, a schematic shows the pathways that BSO and rapamycin are involved in reducing intracellular GSH levels. B, intracellular GSH levels were measured in Tsc2^−/− MEFs treated with DMSO, BSO (10 µmol/L), rapamycin (20 nmol/L), and BSO + rapamycin for 24 hours (n = 3). C, ROS levels were measured in Tsc2^−/− MEFs treated with DMSO, rapamycin (20 nmol/L), BSO (10 µmol/L), BSO + rapamycin, and AMA (positive control) for 24 hours (n = 3). D, Tsc2^−/− MEFs were treated with DMSO, rapamycin (20 nmol/L), BSO (10 µmol/L), BSO + rapamycin, NAC (10 mmol/L), GSH-MEE (2 mmol/L), and Vit C (100 µmol/L) as indicated for 48 hours. Phase microscopy was used to observe cell viability. E, cell death of Tsc2^−/− MEFs after 48 hours of treatment with indicated compounds was measured via PI exclusion assay. The mean is shown; error bars, SD (n = 3).

The combination of BSO and rapamycin induces mitochondrial ROS and alters mitochondrial morphology

ROS have essential roles in normal biologic functions. A moderate increase in ROS can promote cell growth, proliferation, and differentiation (23). Nonetheless, an excessive amount of ROS can cause oxidative damage to DNA, proteins, carbohydrates, and lipids (24). Thus, it is critical to maintain ROS homeostasis for normal growth and survival. Unlike normal cells, many types of tumor cells often display altered redox balance and have elevated basal levels of ROS relative to nontransformed cells (25). To cope with the high intracellular levels of ROS, tumor cells often express increased levels of antioxidant proteins to perform detoxification. As a result, tumor cells are more dependent on antioxidants for cell survival and proliferation than normal cells (26).

The generation of ROS can occur in several intracellular sites, such as cytosol, peroxisomes, plasma membrane, and ER; however, a majority of ROS is produced in the mitochondria when electrons, which are released from the mitochondrial respiratory chain due to partial reaction, react with molecular oxygen, resulting in the generation of the superoxide anion O₂⁻ (24, 27). It has been estimated that the steady-state concentration of superoxide in the mitochondrial matrix is 5- to 10-fold higher than that in the cytosol (28). As mitochondria are the main source of ROS, we asked whether the combination of BSO and rapamycin further potentiates the existing ROS levels in mitochondria. To address this question, we first tested the basal levels of mitochondrial-specific ROS (mtROS) in Tsc2^−/− MEFs versus WT MEFs. As shown in Supplementary Fig. S4A, the basal levels of mtROS were elevated in Tsc2^−/− MEFs compared with WT MEFs staining with the mitochondrial O₂⁻ indicator, MitoSOX Red. MitoSOX Red is nonfluorescent until oxidized by O₂⁻ and an increase in the fluorescence of MitoSOX Red indicates oxidation by mitochondrial O₂⁻. Then Tsc2^−/− MEFs were treated with BSO or rapamycin alone, or in combination for 12 hours and stained with MitoSOX Red. The samples were then counterstained for MitoTracker Green, which localizes to mitochondria regardless of mitochondrial membrane potential. Mitochondrial ROS levels were analyzed by confocal microscopy. Interestingly, Tsc2^−/− MEFs treated with BSO or rapamycin showed an increase in mtROS compared with DMSO control. More importantly, the combination treatment showed a further increase in mtROS levels (Fig. 4A). In addition, mtROS levels were also examined by FACS analysis, and we observed a similar effect with the combination treatment of BSO and rapamycin in Tsc2^−/− MEFs, but the combination treatment did not affect ROS levels in WT MEFs (Supplementary Fig. S4A).

Figure 4.

The combination of BSO and rapamycin induces mitochondrial ROS and alters mitochondrial morphology. A, representative confocal microscopic images of Tsc2^−/− MEFs after staining with MitoSOX (5 µmol/L) and Mitotracker Green (200 nmol/L) in cells treated with DMSO, BSO (10 µmol/L), rapamycin (20 nmol/L), and BSO + rapamycin for 16 hours. Scale bar, 10 µm. B, TEM images (6800×) of Tsc2^−/− MEFs after 24 hours of treatment with DMSO, BSO (2 µmol/L and 10 µmol/L), rapamycin (20 nmol/L), and BSO + rapamycin. Red arrows, mitochondria.

Next, we determined whether the drug treatment would affect mitochondrial morphology by performing transmission electron microscopy (TEM). Within 24 hours, the combination treatment of BSO and rapamycin-induced profound morphologic changes of subcellular components (Fig. 4B). Specifically, compared with single-agent or DMSO treatment, the combination treatment (BSO: 2 µmol/L, 10 µmol/L; rapamycin: 20 nmol/L) altered mitochondrial morphology by causing the remodeling of inner membrane structure, cristae. Several lines of evidence have indicated that increased oxidative stress and the associated mitochondrial dysfunction can induce cell death via autophagy (29). In addition, it has been reported that oxidative stress can induce autophagic cell death independent of apoptosis in transformed and cancer cells but not in nontransformed cells (30). Interestingly, our study suggested that BSO+rapamycin treatment cause cell death of Tsc2^−/− MEFs through a similar mechanism.

Targeting GCLC and mTORC1 with small molecules causes regression of Tsc2-deficient ELT3 cell xenograft tumors

To test the role of GCLC and mTORC1 inhibition in cells with mTORC1 hyperactivation in vivo, we used an ELT3 cell-xenograft model. Mice-bearing ELT3-luciferase–expressing xenograft tumors were treated with BSO and rapamycin as single treatments or in combination. We first evaluated drug toxicity by body weight changes, and we did not observe any obvious effect of drug treatment (Fig. 5A). We next assessed the effect of combination treatment on tumor development and found that xenograft tumor size for 5 weeks was smaller than single treatment of rapamycin by approximately 0.44-fold (Fig. 5B). Subsequently, compared with the vehicle control, single treatment of rapamycin for 5 weeks reduced tumor growth by approximately 0.69-fold in bioluminescence intensity (Fig. 5C and D). More interestingly, in mice treated with rapamycin alone, although tumor growth was slowed, they continued to grow after 5 weeks. Interestingly though, tumors in mice treated with the combination of BSO and rapamycin grew more slowly than with rapamycin alone for 5 weeks and then appeared to regress, consistent with suppression of xenograft tumor progression and cell death (Fig. 5D). The immunohistochemical staining showed that the combination treatment of BSO and rapamycin reduced cell proliferation using PCNA marker (Fig. 5E). Moreover, the combination treatment of BSO and rapamycin resulted in higher levels of TUNEL staining indicating the enhanced cell death in tumors relative to either single agent alone (Fig. 5F).

Figure 5.

Targeting GCLC and mTORC1 with small molecules causes regression of TSC2-deficient ELT3 cell xenograft tumors. Female CB17-scid mice were inoculated with ELT3-luciferase cells subcutaneously. Mice were treated with either vehicle, rapamycin, BSO or combined rapamycin and BSO for 6 weeks. A, body weight was measured every week. B, tumor volume was measured weekly using a digital caliper. The left y-axis indicated the relative fold growth of tumor size versus the baseline measurement prior to drug treatment. C and D, bioluminescent intensity in xenograft tumors was recorded and quantified weekly. The left y-axis indicated the relative tumor growth versus the baseline quantification prior to drug treatment. E, immunohistochemical staining of PCNA in different magnification. F, TUNEL staining in different magnification.

Discussion

Significant efforts have been made to identify inhibitors that effectively target the oncogenic signaling pathways. Such inhibitors have been proven efficacious in some cases as single agents. However, monotherapy often has its limitations in targeting multigenetic diseases, such as cancer. Therefore, drug combinations that impact multiple molecular targets can potentially provide improved therapeutic benefits, and drug combination approaches may also lead to the design of therapeutic strategies that reduce or overcome the development of drug resistance (31).

Combination therapy with rapamycin or rapalogs has been exploited as an alternative therapeutic strategy (32). A majority of the ongoing rapamycin-based clinical trials are testing the efficacy of rapamycin in combination with other agents in different disease settings (https://clinicaltrials.gov/ct2/results?term=rapamycin&pg=4). To identify new drugs that can synergistically collaborate with rapamycin to induce tumor cell death, we set out to screen against a large library of known bioactives, which represent 50% of the FDA-approved drugs in Tsc2^−/− MEFs and determine the sensitivity of these drugs to mTORC1 inhibition. Interestingly, we found that inhibition of GCLC and mTORC1 selectively induces death in Tsc2-deficient cells but not in Tsc2^+/+ cells. We demonstrated that the molecular basis for the selective toxicity of the combination of BSO and rapamycin was linked to a reduction of GSH levels and an increase in ROS in Tsc2-deficient cells. Thus, targeting oxidative stress holds great promise for the treatment of TSC and sporadic lymphangioleiomyomatosis as well as cancers.

The intrinsic ROS level is higher in many cancer cell types compared with their normal counterparts, which can result from different cellular processes including increased metabolic activity, mitochondrial dysfunction, and oncogenic activity (26). A mild increase of ROS levels may result in cellular alterations linked to an adaptive survival response, whereas a severe increase of ROS levels can be detrimental to cells. Thus, cancer cells are more vulnerable to further oxidative damage triggered by exogenous agents that increase ROS generation. As a result, shifting redox balance will cause cancer cells to go beyond their tolerability threshold, and induce cell death (24). Several strategies have been employed to induce ROS including ionizing radiation, inhibiting the ubiquitin-proteasome pathway as well as affecting GSH, thioredoxin, glucose, and glutamine metabolisms (33).

BSO is the currently only known inhibitor of de novo GSH biosynthesis (34). Increased GSH levels have been suggested to contribute to resistance to chemotherapeutic agents such as platinum-containing compounds, and melphan, doxorubicin, and bleomycin (35). Inhibition of GSH synthesis by BSO has been shown to enhance cytotoxicity with these agents (36–38). In addition, a new study reported that BSO enhanced the activity of l-Phenylalanine Mustard (l-PAM) against multiple myeloma. Moreover, a recent phase I study demonstrated that BSO and l-PAM were well tolerated in neuroblastoma patients (39). On the basis of previous studies and our findings, it is critical to identify novel therapeutic targets involved in redox systems and subsequently develop inhibitors for these targets.

One major concern associated with rapamycin-based therapy is its safety profile of long-term use as rapamycin suppresses the immune system, and such therapy may cause serious side effects including thrombocytopenia and hyperlipidemia, impaired wound healing, nephrotoxicity as well as altered insulin sensitivity (32). Thus, using low doses of rapamycin in combination with other agents may provide an alterative approach. In our in vivo study, we reduced the general doses of rapamycin, and showed that the combination treatment with BSO effectively reduced tumor growth. Future work exploring optimal and clinical doses of rapamycin and BSO remain to be completed.

Taken together, we reveal a new and promising therapeutic approach by targeting GCLC and mTORC1 in vitro and in a TSC-xenograft tumor model. In this current study, we used Tsc2-deficient cells as a model to study cells with hyperactive mTORC1. Future investigation will include testing the efficacy of this combination treatment in other mTORC1-driven tumor cells, such as those characterized by the loss of function of tumor suppressors, and/or gain-of-function mutations of upstream regulators of mTORC1, including PTEN and PI3KCA, respectively. Further study is warranted to evaluate the therapeutic efficacy of this drug combination.