Free Radical Biology and Medicine 68 (2014) 110–121

Contents lists available at ScienceDirect

Free Radical Biology and Medicine journal homepage: www.elsevier.com/locate/freeradbiomed

Original Contribution

A novel small molecule that induces oxidative stress and selectively kills malignant cells Francesca R. Šalipur a,b, E. Merit Reyes-Reyes b,c, Bo Xu d, Gerald B. Hammond d, Paula J. Bates a,b,c,n a

Department of Biochemistry and Molecular Biology, University of Louisville, Louisville, KY 40202, USA Molecular Targets Program of the James Graham Brown Cancer Center, University of Louisville, Louisville, KY 40202, USA c Department of Medicine, University of Louisville, Louisville, KY 40202, USA d Department of Chemistry, University of Louisville, Louisville, KY 40202, USA b

art ic l e i nf o

a b s t r a c t

Article history: Received 5 April 2013 Received in revised form 7 November 2013 Accepted 2 December 2013 Available online 7 December 2013

We have synthesized a novel molecule named XB05 (1-bromo-1,1-difluoro-non-2-yn-4-ol) and evaluated its effects in a variety of human cell lines. XB05 displayed potent antiproliferative activity against cell lines derived from leukemia or solid tumors, but had less effect on nonmalignant cells. To identify factors that contribute to the cancer selectivity of XB05, we chose three cell lines that had high sensitivity to XB05 (U937 leukemia), moderate sensitivity (A549 lung cancer), or low sensitivity (Hs27 nonmalignant skin fibroblasts), and proceeded to assess cell death and oxidative stress in these cells. XB05 was found to induce cell death via both apoptotic and nonapoptotic mechanisms in U937 and A549 cells, whereas it had no cytotoxicity against Hs27 cells at comparable concentrations. Treatment with XB05 caused an increase in reactive oxygen species in all cell lines tested, but levels were higher in malignant compared to nonmalignant cells. XB05 treatment also induced DNA damage exclusively in the malignant cells. Differences in antioxidant responses were observed between cell lines. For example, XB05 caused a decrease in levels of glutathione and nuclear Nrf2 in the most sensitive cells (U937), whereas the least sensitive cells (Hs27) displayed increased glutathione levels and no change in nuclear Nrf2. XB05 could react in vitro with cysteine and glutathione, but had much lower reactivity compared to typical thiolreactive electrophiles, diethyl maleate and maleimide. In summary, XB05 is a novel compound that selectively kills malignant cells, most likely by disrupting cellular redox homeostasis, making it a promising candidate for development as a chemotherapeutic agent. & 2013 Elsevier Inc. All rights reserved.

Keywords: XB05 Cell death Oxidative stress Glutathione Nrf2 Cancer Leukemia

Introduction XB05 is a synthetic small molecule that was originally generated as a building block to introduce fluorine-containing groups into organic molecules [1]. Intrigued by its novel chemical structure and resemblance to marine-derived natural products with antitumor activity [2], we screened XB05 for biological activity in human cell lines. Although XB05 was not originally intended as an anticancer agent, our studies have revealed significant activity

Abbreviations: AA, L-ascorbic acid; BSO, buthionine sulfoximine; carboxyH2DCFDA, 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate; DCF, dichlorofluorescein; DEM, diethyl maleate; DSB, double-strand breaks; DTNB, dithiobis-2nitrobenzoic acid; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; NAC, N-acetylcysteine; Nrf2, NF-E2-related factor 2; PI, propidium iodide; ROS, reactive oxygen species; TBHP, tert-butyl hydrogen peroxide n Corresponding author at: University of Louisville, 505 S. Hancock Street, CTRB 409, Louisville, KY 40202. Fax: þ 1 502 852 3661. E-mail address: [email protected] (P.J. Bates). 0891-5849/$ - see front matter & 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.freeradbiomed.2013.12.002

against cancer cells. Here we describe for the first time the cancerselective antiproliferative and cytotoxic effects of XB05, and propose an oxidative stress-based mechanism to explain its activity against cancer cells. Materials and methods Materials XB05 was synthesized using a previously described method [1]. Further details are provided in Supplementary Information. S-( þ)-Camptothecin, Z-VAD-fmk, L-buthionine-sulfoximine, antimycin A, diethyl maleate, maleimide, reduced L-glutathione, N-acetylcysteine, L-cysteine, and L-ascorbic acid were purchased from Sigma-Aldrich (St. Louis, MO). Antibodies against cleaved caspase-3 (No. 9661), γ-H2AX (No. 9718), phosphorylated (Ser 51) eIF2α (No. 9721), eIF2α (No. 9722), and BiP (No. 3177) were purchased from Cell Signaling Technology, Inc. (Danvers, MA).

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

Thapsigargin, anti-rabbit and anti-mouse antibodies linked to horseradish peroxidase, and antibodies against PARP-1 (sc-8007), Nrf2 (sc-365949), Ku-70 (sc-5309), and GAPDH (sc-47724) were purchased from Santa Cruz Biotech (Santa Cruz, CA).

Cell culture and treatments Cell lines were either recently purchased from the American Type Culture Collection (ATCC) or verified by short tandem repeat (STR) analysis (IDEXX Laboratories, Westbrook, ME). Cells were grown in the appropriate medium supplemented with 10% fetal bovine serum (FBS, Hyclone Laboratories, Logan, UT), 62.5 μg/mL penicillin, and 100 μg/mL streptomycin (Life Technologies, Grand Island, NY) in a humidified incubator at 37 1C with 5% CO2. The media were as follows: Dulbecco’s modified Eagle’s medium (DMEM) for A549, MDA-MB-231, DU145, and Hs27 cells; RPMI 1640 for U937 cells; Eagle’s minimal essential medium (EMEM) supplemented with Eagle’s balanced salt solution (EBSS), 2 mM L-glutamine, 1 mM sodium pyruvate, 1500 mg/L sodium bicarbonate, and nonessential amino acids (Lonza, Walkersville, MD) for IMR-90 cells; mammary epithelial growth medium (MEGM) supplemented with all of the components of the MEGM SingleQuots kit except for GA-1000 (Lonza) for MCF-10A cells. Cells were treated by direct addition of XB05 solutions into the culture medium to give the final concentrations indicated in the figures. Unless otherwise stated, cells were at approximately 40% confluence at the start of treatment. XB05 solutions were freshly prepared from stock solutions of 2 mM in 100% DMSO by dilution with cell culture medium. Final DMSO concentrations were 0.05% in both vehicle control and XB05-treated cells. As a positive control for apoptosis, cells were treated with camptothecin (6 μg/mL) for the times indicated in the figure legends. As a positive control for DNA damage, cells were irradiated with 800 μJ in a UV Stratalinker 2400 (Stratagene, Santa Clara, CA), and then allowed to grow in culture for an additional 2 h.

Cell proliferation assays The antiproliferative activity of XB05 was tested using a previously published 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay protocol [3]. Briefly, cells were seeded in quadruplicate wells in 96-well plates and allowed to adhere overnight. To account for intrinsic differences in growth rates, cells were plated at the following densities so as to achieve comparable MTT absorbance values (between 1 and 2) in untreated samples after 72 h: A549, DU145, and MDA-MB-231, 1000 cells/well; Hs27 and U937, 1500 cells/well; IMR-90 and MCF10A, 5000 cells/well. Plates were incubated with XB05 for 72 h, during which the cell culture medium was not changed. Cell viability was determined and the background corresponding to medium alone (no cells) was subtracted. Data were analyzed as described in figure legends.

111

Flow cytometric assays Analyses were performed using a FACScalibur flow cytometer (BD Biosciences, Mountain View, CA) and FlowJo software (Tree Star, Inc., Ashland, OR). Cells were treated as indicated in the figures and harvested by using TrypLE Cell Dissociation Reagent (Life Technologies). For cell death analyses, cells (adherent and detached cells) were stained with annexin-V-FITC and propidium iodide (PI) using an apoptosis detection kit (BD Biosciences, San Jose, CA), according to the manufacturer's instructions. For cellcycle distribution, cells were fixed, and stained with PI using the Cycle Test Plus kit (Becton Dickinson, Franklin Lakes, NJ). For detection of γ-H2AX, cells were washed with PBS and fixed in 4% paraformaldehyde in PBS for 10 min at 37 1C. Cells were then permeabilized in 90% methanol for 30 min on ice, blocked in 2% BSA (w/v in PBS) for 1 h, and incubated with γ-H2AX antibody or rabbit IgG isotype control (Santa Cruz, sc-2027) (0.3 μg/mL in 1% BSA) for 1 h. After washing, cells were incubated with goat antirabbit IgG antibody conjugated to Alexa Fluor-488 (Life Technologies) for 1 h at room temperature. For reactive oxygen species (ROS) detection, cells (1  106 cells/mL) were stained with 25 μM 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate (carboxyH2DCFDA, Life Technologies) in PBS for 30 min at 37 1C. After washing with PBS, cells were resuspended in PBS containing 1 μg/mL PI to allow exclusion of nonviable cells. For assays using MitoSOX Red (Life Technologies), cells were harvested, washed with PBS, and stained with 5 μM MitoSOX Red in warm PBS for 10 min at 37 1C. Cells were then washed twice with PBS, resuspended at a density of 1  106 cells/mL in PBS, and analyzed by flow cytometry. Clonogenic assays Cells were plated at low density (300 cells/well) in 6-well tissue culture plates and allowed to adhere overnight. Where indicated, the medium was then replaced with fresh medium containing antioxidants or vehicle (sterile ultrapure H2O). XB05 or vehicle control (DMSO) was then added directly to the medium at the concentrations indicated. After 10 days, cells were fixed with 4% paraformaldehyde in PBS, stained with Accustain Crystal Violet Solution (Sigma Aldrich), washed, and air-dried. Colonies were counted using the cell counter feature in the Image J software, available from the National Institute of Health (rsbweb.nih.gov/ij/ download.html). Cell extracts After treatment as indicated in figure legends, cells were washed twice with ice-cold PBS. Nuclear extracts were isolated from cells using the NE-PER nuclear and cytoplasmic extraction kit (Thermo Scientific, Rockford, IL), according to the manufacturer’s protocol. For total cell extracts, cells were lysed in RIPA buffer (Thermo Scientific) containing protease inhibitor cocktail III and phosphatase inhibitor cocktail (Calbiochem, Billerica, MA) for 5 min at 4 1C, and then cleared by centrifugation at 16,000 g for 10 min at 4 1C. All protein concentrations were determined using the DC Protein Assay kit (BioRad, Hercules, CA).

Trypan blue exclusion assay and microscopy Western blotting Cells were treated as indicated in the figures. After 72 h, cells were trypsinized (in the case of adherent cells) and stained with trypan blue solution (0.4%) (BioRad, Hercules CA), and cell number and percentage viability determined on a TC10 automated cell counter (BioRad) using triplicate measurements for each sample. Phase contrast microscopy images were obtained on an EVOSfI digital inverted microscope (Advanced Microscopy Group, Bothell, WA).

Equal amounts of protein per sample (typically, 50 μg) were resolved by SDS–Tris polyacrylamide gel electrophoresis and then electrotransferred onto polyvinylidine fluoride membranes (Millipore, Bedford, MA) in Tris–glycine buffer containing 20% methanol. Membranes were blocked with 5% milk or 5% BSA (for phosphoprotein detection) in TBS/0.01% Tween. Proteins were detected using the following primary antibody concentrations: PARP-1

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

112

(1:1000), cleaved caspase-3 (1:1000), γ-H2AX (Ser139) (1:000), phospho-eIF2α (Ser51) (1:1000), eIF2α (1:1000), Bip (1:1000), Nrf2 (1:1000), Ku-70 (1:1000), and GAPDH (1:1000). In some cases, membranes were stripped using Restore Plus Western blot stripping buffer (Thermo Scientific) and then reprobed as described in the figure legends. Where indicated, band intensities were quantified using the UN-SCAN-IT Gel digitizing software (Silk Scientific, Orem, UT) and normalized to vehicle control.

Dithiobis-2-nitrobenzoic acid (DTNB) assay for free sulfhydryls For each experiment, a 20X working solution of DTNB (2 mM DTNB, 50 mM sodium acetate in 100 mM Tris-HCl, pH 8.0) and 2X solutions of cysteine and GSH (0.5 mM in 100 mM Tris-HCl, pH 8.0) were freshly prepared. The 10 mM stock solutions of XB05, diethyl maleate, and maleimide were prepared in 100% DMSO, and subsequently serially diluted to 2X working solutions in 100 mM Tris-HCl, pH 8.0. Then 100 μl each of thiol compounds (final concentration 0.25 mM) and thiol-reactive compounds (final concentrations 0.05–5 mM) or vehicle control (DMSO) was added in duplicate to a 96-well plate and allowed to incubate at 37 1C for 30 min. Twenty microliters of 20X DTNB working solution was then added to each well (final concentration 100 μM), and allowed to incubate at room temperature for an additional 5 min. Absorbance was read on a Biotek HT Synergy plate reader (Winooski, VT) at 412 nm.

Comet assays Cells were washed once with ice-cold PBS, harvested as described, and resuspended at 1  105 cells/mL in PBS. As a positive control, cells were incubated with 100 μM tert-butyl hydrogen peroxide (TBHP) for 30 min (U937, A549) or 1 h (Hs27) at 4 1C prior to harvesting. Comet assays were performed using the Comet Assay Kit (Trevigen, Gaithersburg, MD) according to the manufacturer's instructions. Cells were visualized and digital images captured on an EVOSfI digital fluorescent microscope equipped with a fluorescein light cube.

Statistical analyses Where indicated, statistical comparisons were conducted using Student’s t test. Statistical significance was set at P r 0.05.

Reduced glutathione content analyses Results After treatment as indicated in the figures, cells were harvested as described, counted, and dispensed at a density of 10,000 cells/well in triplicate on a 96-well white, opaque bioluminescence plate (Corning). Cellular content of reduced glutathione (GSH) was analyzed using the GSH-Glo glutathione assay kit (Promega) according to the manufacturer's instructions. The chemiluminescent signal was quantified on a Biotek Synergy HT plate-reader (Biotek, Winooski, VT) using an integration time of 1 s/well.

XB05 has cancer-selective antiproliferative activity The novel molecule, XB05 (structure shown in Fig. 1A), was tested to determine its antiproliferative activity in a variety of human cell lines, which are derived from malignant tissues or from nonmalignant (immortalized but not transformed) cells, listed in Fig. 1B. As described under Materials and methods, cells

Hs27

MCF10a

IMR-90

U937

A549

MDA-MB-231

DU-145

Cell Line

Description

Hs27

Non-malignant human foreskin fibroblasts

MCF-10A

Non-malignant human mammary epithelial cells

IMR-90

Non-malignant human lung fibroblasts

U937

Human myeloid leukemia

A549

Human lung adenocarcinoma

MDA-MB-231

Human breast adenocarcinoma

DU145

Human prostate carcinoma

1.2

Normalized OD 570 nm

1 0.8

NM

0.6 0.4 0.2 0 -0.2

0

0. 5

Mal 1

1.5

2

2.5

XB05 Concentration ( M) Fig. 1. XB05 has cancer-selective antiproliferative activity. (A) The chemical structure of XB05. (B) Human nonmalignant (NM) and malignant (Mal) cell lines in which XB05 antiproliferative activity was tested. (C) Cells were treated with the indicated concentrations of XB05 for 72 h and proliferation was assessed using MTT assay. Results were normalized to untreated controls for each cell line. Data points represent the mean þ /– standard error of the mean (SEM) for three independent experiments. Statistical comparisons between nonmalignant and malignant cell lines with the same tissue of origin were conducted. Statistically significant differences were observed between A549 and IMR-90 at 0.75, 1, and 2 μM XB05 (P r 0.05) and between MDA-MB-231 and MCF-10A at all XB05 concentrations tested (P r 0.01).

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

were treated with varying concentrations of XB05 for 72 h and evaluated using the MTT cell proliferation assay. The resulting values were normalized to the proliferation of untreated cells for each cell type, which revealed that XB05 has potent antiproliferative activity against most of the malignant cell lines tested (e.g., GI50 o 500 nM for DU145, U937, and MDA-MB-231 cells), but has a much smaller inhibitory effect on the proliferation of nonmalignant cells (Fig. 1C). Clonogenic cell survival assays were also conducted for some cell lines and are shown in Supplementary Fig. S1. Based on the data shown in Fig. 1C, we chose three cell lines that had high, intermediate, or low sensitivity to XB05 for additional studies, with the goal of elucidating their differential responses to XB05. The selected cell lines were U937 myeloid leukemia cells (high sensitivity to XB05), A549 non-small-cell lung cancer cells (moderate sensitivity), and Hs27 nonmalignant skin fibroblasts (low sensitivity).

113

XB05 causes transient G2/M cell cycle arrest in U937 cells As determined by flow cytometric analysis of cells stained for DNA content, XB05-treated U937 cells displayed a pronounced increase in the proportion of cells in G2/M phase at 24 h, but this was reversible and the proportion in G2/M was reduced by 48 and 72 h (Supplementary Fig. S4). In contrast, XB05-treated A549 cells showed only a very modest increase in the G2/M fraction, suggesting that cell death does not depend upon cell cycle arrest. However, these results may explain our observation of multinucleated cells because transient arrest in G2/M followed by cell cycle progression can result in mitotic catastrophe [6], consistent with the observed morphology (Fig. 2B, white arrows, and Supplementary Fig. S2C). In agreement with the previous results, both malignant cell lines treated with XB05 exhibited a timedependent increase in the percentage of apoptotic cells with subG1 DNA content (Supplementary Fig. S4).

XB05 induces apoptotic and nonapoptotic cell death in malignant cell lines

XB05 causes an increase in reactive oxygen species

To determine whether the antiproliferative effects of XB05 are due to cytostasis or cytotoxicity, we next examined XB05-treated cells for physical, morphological, and biochemical markers of cell death. Trypan blue exclusion assays showed that XB05 induced robust cell death in U937 cells at 1 mM concentration, modest cell death in A549 cells at 2 mM, and lowest levels of cell death in the nonmalignant Hs27 cells (Fig. 2A). Microscopic examination of cells confirmed these results and revealed that XB05 induced a mixture of cell morphologies (Fig. 2B and Supplementary Fig. S2). For both U937 leukemia cells (which grow in suspension) and A549 lung cancer cells (adherent), treatment with XB05 led to some cells with classical apoptotic morphology and some with abnormal morphology characterized by cellular swelling, accumulation of perinuclear vesicles, and (in some cases) multinucleation (Fig. 2B and Supplementary Fig. S2C). In contrast, camptothecin, which is a topoisomerase I inhibitor used here as a positive control for apoptosis induction [4], triggered cell death in all three cell lines predominantly by apoptosis (Fig. 2B and Supplementary Fig. S2B). Next, we investigated the timing of cell death induction by Western blotting for biochemical markers of apoptosis. In accord with the results described above, we observed PARP-1 and caspase-3 cleavage in all three cell lines following treatment with camptothecin, but only in the malignant cells after XB05 treatment (Fig. 2C and Supplementary Fig. S3). These markers of apoptosis were apparent by 24 h after treatment with XB05 in U937 cells, but were seen later (48 h) in the less sensitive A549 cells (Fig. 2C and Supplementary Fig. S3). Finally, to evaluate the relative contribution of apoptosis to XB05-induced cell death, we performed flow cytometric analyses of annexin/PI-stained cancer cells that had been treated in the presence or absence of the pancaspase inhibitor, N-benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (Z-VAD-fmk). As expected, camptothecin treatment led to mainly apoptotic cells (seen in the annexin-positive/PInegative quadrant in Fig. 2D), whereas XB05 treatment induced two distinct populations of cells corresponding to early apoptotic cells (annexin-positive/PI-negative) and late apoptotic or necrotic cells (annexin-positive/PI-positive) (Fig. 2D). Moreover, while the presence of Z-VAD-fmk strongly inhibited cell death in response to camptothecin, it had little effect on cell death in response to XB05, as determined by the proportion of cells staining positive for either annexin or PI (Fig. 2D). These data indicate that XB05induced cell death can occur without apoptosis, which is significant because the ability to evade apoptosis (via activation of various antiapoptotic pathways) is a hallmark of aggressive cancer cells and a major contributor to chemoresistance [5].

The cancer-selective activity of XB05, as well as its structural similarities to some bioactive endogenous lipid electrophiles [7,8], led us to hypothesize that XB05 might induce oxidative stress [9]. To test this possibility, cells were treated with XB05 and analyzed for the presence of reactive oxygen species by flow cytometric analyses. The first method involved staining cells with carboxyH2DCFDA, a cell-permeable, redox-sensitive probe that fluoresces after intracellular deesterification and oxidation to dichlorofluorescein (DCF). In this assay, cells were treated with tert-butyl hydrogen peroxide as a positive control for oxidative stress. The levels of ROS, measured by the percentage of viable cells (PI-negative) that stained positive for DCF, were elevated in all cell lines after 8 h of XB05 treatment, but were significantly higher in the XB05-sensitive cell lines (A549 lung cancer and U937 leukemia) compared to the less sensitive, nonmalignant Hs27 cells (Fig. 3A). Because this DCF production may be prone to artifacts [10] (despite it being one of the most commonly used assays for ROS detection), we attempted to confirm the results using a different ROS probe that is targeted to the mitochondria, namely MitoSOX Red. Mitochondria are a major source of endogenous ROS because single electrons which escape from the electron transport chain (ETC) can reduce molecular oxygen to superoxide anion [11]. Superoxide is rapidly converted to hydrogen peroxide by manganese superoxide dismutase (SOD2); thus mitochondria are a source of both superoxide and hydrogen peroxide [12]. MitoSOX Red is a derivative of hydroethidine (HEt) conjugated to triphenylphosphonium, which preferentially accumulates within the mitochondrial matrix due to its positive charge and which generates a variety of products with red fluorescence when oxidized [13–15]. We incubated XB05-treated cells with MitoSOX Red and analyzed the staining by flow cytometry as described under Materials and methods. In these experiments, some cells were treated with antimycin A, a known inhibitor of Complex III in the ETC [16], as a positive control that increases superoxide production. We found that XB05 treatment for 8 h caused a significant increase in MitoSOX Red fluorescence, compared to the vehicle control, in U937 and A549 malignant cells (P o 0.05) but not in Hs27 nonmalignant fibroblasts (P ¼ 0.53) (Fig. 3B). Taken together, the results in Fig. 3 indicate that XB05 causes an increase in intracellular ROS selectively in malignant cells. Although apoptosis and necrosis can lead to positive staining in both of these ROS detection assays [10,15], the data are not consistent with ROS production being secondary to cell death induction. For example, maximal DCF staining was observed after 8 h with XB05, whereas maximal cell death induction occurred at much later times (at 24

114

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

Fig. 2. XB05 induces apoptotic and nonapoptotic cell death in malignant cell lines. (A) Trypan blue exclusion assay for three human cell lines following treatment with XB05 for 72 h at the concentrations indicated. Data points represent the mean þ /– SEM for three independent experiments. (B) Representative phase contrast microscopy images of cells treated with vehicle (72 h), XB05 (72 h with 1 μM for U937 and Hs27, or 2 μM for A549), or 6 mg/mL camptothecin (U937 for 6 h, A549 and Hs27 for 48 h). Cells undergoing apoptosis (black arrows) or nonapoptotic cell death (white arrows) are indicated. (C) Whole cell extracts from cells treated with vehicle, camptothecin, or XB05, as described in part B were Western-blotted for biochemical markers of apoptosis and GAPDH (loading control). Apoptosis is indicated by cleavage of full-length PARP-1 (116 kDa) to an 89-kDa fragment, and by appearance of bands (17 kDa, 19 kDa) recognized by an antibody specific to cleaved (activated) forms of caspase-3. (D) Flow cytometric analysis of annexin V-FITC/propidium Iodide (PI) staining for malignant cells treated with vehicle, camptothecin, or XB05 as described in part B, in the presence or absence of the pan-caspase inhibitor Z-VAD-fmk (100 μM).

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

115

Fig. 3. XB05 induces ROS. (A) Cells were treated with either vehicle, tert-butyl hydrogen peroxide (TBHP, 100 μM, 3 h), or XB05 (8 h with 1 μM for U937 and Hs27, or 2 μM for A549), and then stained with carboxy-H2DCFDA and PI for analysis by flow cytometry. Histograms illustrate representative data and bar graphs show the increase (over vehicle control) in the percentage of PI-negative (viable) cells with positive DCF fluorescence in each sample. Bars represent the mean þ/– SEM from three independent experiments. Statistical comparisons showed that XB05 significantly increased ROS in all cell lines (*P r 0.05) and that both A549 and U937 had a significantly greater increase compared to Hs27 (*P r 0.05). (B) Cells were treated with XB05 as described in part A. As a positive control for superoxide production, cells were treated with antimycin A (10 μM) for 1 h. Cells were harvested and subjected to MitoSOX Red staining for analysis by flow cytometry as described under Materials and methods. Histograms illustrate representative data and bar graphs show the mean þ /– SEM for three independent experiments. Statistical comparisons showed that XB05 significantly increased superoxide detection in A549 and U937 cells (*P r 0.05) but not in Hs27 nonmalignant fibroblasts (P ¼ 0.53).

or 48 h, Fig. 2C and data not shown), suggesting that increases in ROS levels precede cell death. XB05 modulates cellular stress responses Following oxidative and other stresses, cells can activate protective responses to neutralize reactive species and thereby limit damage to DNA, lipids, and proteins. These protective factors include antioxidant molecules, such as the tripeptide, glutathione [17], and transcription factors, such as NF-E2-related factor 2 (Nrf2), that mediate expression of stress response genes [18,19]. Therefore, we examined the effect of XB05 on these pathways using a luminescence-based reporter assay to monitor levels of the reduced form of glutathione (GSH) and Western blotting to detect Nrf2. The GSH-Glo assay is based on conversion of a luciferin derivative (luciferin-NT) to luciferin by glutathione-S-transferase in the presence of reduced glutathione, which is obtained from lysed cells in each treatment sample. Luciferin acts as a substrate for firefly luciferase to produce a luminescent signal that is proportional to the amount of GSH present. We found that treatment of the highly sensitive U937 cells with 1 mM XB05 resulted in a dramatic depletion of reduced GSH, which was comparable to that seen with the positive control treatment, consisting of a combination of 200 μM buthionine sulfoximine (BSO, an inhibitor of GSH synthesis) and 1 mM diethyl maleate (DEM, a GSH conjugating agent) (Fig. 4A and Supplementary Fig. S5) [20,21]. In contrast, XB05 treatment of A549 and Hs27 cells led to approximately 1.3-fold and 2-fold increases in GSH, respectively (Fig. 4A and Supplementary Fig. S5), suggesting that the decreased sensitivity of these cell lines relative to U937 could be due to their enhanced ability to induce glutathione synthesis as an adaptive response to the drug. Addition of exogenous GSH to the medium inhibited the ability of XB05 to modulate GSH levels (Fig. 4A and Supplementary Fig. S5). To confirm that GSH depletion played a significant role in the cytotoxicity of XB05, we performed similar experiments to assess whether further depletion of GSH

would increase the activity of XB05 toward malignant cells. Thus, we treated A549 and U937 cells with XB05 in the presence of sublethal concentrations (50 and 100 μM) of the glutathione synthesis inhibitor, BSO, for 72 h and then analyzed cell death by annexin/PI staining. We observed that combined treatment with BSO and XB05 led to enhanced cytotoxicity compared to XB05 treatment alone, as indicated by an increase in both apoptotic and late apoptotic/necrotic cells (Fig. 4B and Supplementary Figs. S6A and S6B). Next, we examined XB05-treated cells for activation of Nrf2, a redox-regulated transcription factor that translocates to the nucleus in response to oxidative or electrophilic stress [18,19,22]. Nrf2 can activate transcription of various ROSdetoxifying enzymes, including γ-glutamyl-cysteine synthetase (γ-GCS), the rate-limiting enzyme for glutathione synthesis [23,24]. Nuclear extracts from the three cell lines treated with XB05 for 8 or 24 h were assessed for the presence of Nrf2 and displayed differential responses (Fig. 4C and Supplementary Fig. S6C). In U937 cells, there was a marked decrease in nuclear Nrf2 after XB05 treatment, suggesting that the high sensitivity of these cells to XB05 could be related to its inhibitory effect on the Nrf2 antioxidant response pathway. This is especially significant because high Nrf2 expression has been shown to drive chemoresistance in acute myeloid leukemia [25]. In A549 cells (which have constitutive activation of this pathway [26]), we observed a slight increase in nuclear levels of Nrf2, whereas no change was apparent for Hs27 cells. The latter result indicates that the observed increase in GSH levels for XB05-treated Hs27 cells is probably Nrf2-independent. Finally, we analyzed cells for phosphorylated eukaryotic initiation factor 2 α (p-eIF2α), which is another common marker for cellular stress. Phosphorylation of eIF2α can be mediated by a variety of stress-induced kinases and results in inhibition of general translation, while allowing preferential translation of transcripts that encode for stress-relieving proteins [27]. In these experiments, we used thapsigargin, an inducer of endoplasmic reticulum (ER) stress, as a positive control. We observed strongest induction of p-eIF2α in U937 cells (highly

116

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

Fig. 4. XB05 modulates cellular stress responses. (A) A luminescence-based assay for reduced glutathione (GSH) content in cells treated 24 h with XB05 (1 μM for U937 and Hs27, or 2 μM for A549), in the absence or presence of 5 mM GSH. Some cells were treated with buthionine sulfoximine (BSO, 200 μM) plus diethyl maleate (DEM, 1 mM) for 3 h as a positive control for GSH depletion. Bar graph indicates mean þ/– SEM from three independent experiments. Raw data are provided in Supplementary Fig. S5. (B) Flow cytometric analysis of Annexin/PI staining for A549 cells treated 72 h with vehicle or 2 mM XB05 in the absence or presence of BSO as indicated. Data are representative of three independent experiments. Comparable data for U937 cells and quantitation for both cell lines are shown in Supplementary Fig. S6. (C) Western blot for Nrf2 in nuclear extracts from cells treated as in part A for either 8 or 24 h. Membranes were stripped and reprobed for Ku-70 as a loading control. Data shown are representative of three individual experiments for each cell line and quantitative data are shown in Supplementary Fig. S6. (D) Western blot for phospho-eIF2α (Ser51), total eIF2-α, BiP/Grp78, and GAPDH (loading control) in whole cell extracts from cell lines treated as indicated for 24 h with concentrations of XB05 described in part A. Thapsigargin (TG, 2 mM) was used as a positive control for induction of endoplasmic reticulum (ER) stress. Data shown are representative of three individual experiments for each cell line and quantitative data are shown in Supplementary Fig. S6.

sensitive to XB05), followed by A549 (moderately sensitive), and Hs27 cells (least sensitive) (Fig. 4D and Supplementary Fig. S6D). However, in contrast to thapsigargin, XB05 treatment did not lead to significant upregulation of BiP/Grp78 (an ER chaperone and marker of ER stress [28]) in any of the cell lines (Fig. 4D and Supplementary Fig. S6D), suggesting that XB05 does not induce ER stress or the unfolded protein response (UPR). Interestingly, phosphorylation of eIF2α in the absence of BiP/Grp78 induction is associated with activation of proapoptotic pathways rather than cytoprotective responses [29].

damage involves DSB, we conducted Western blot analysis for γH2AX (histone H2AX phosphorylated at serine 139), which is a specific marker for DSB [30]. UV irradiation was used as a positive control. We found that XB05 induces γ-H2AX within 24 h in both U937 and A549 cell lines (Fig. 5B). The result was confirmed by flow cytometry (Fig. 5C), with XB05-treated cells showing a significant increase in γ-H2AX staining (4.5-fold in U937 cells and 2.9-fold in A549 cells) compared to vehicle-treated controls (P o 0.05).

XB05 induces DNA damage in malignant cells

Exogenous thiol antioxidants inhibit the antiproliferative and cytotoxic effects of XB05

The cytotoxicity of excess ROS is due largely to DNA damage, especially double-strand breaks (DSB). To determine if XB05 induced DNA damage, we performed alkaline comet assays for cells treated with either XB05 or TBHP positive control. Using this single cell electrophoresis method, damaged DNA appears as a comet “tail.” After treatment with THBP, comet tails could be observed in all cell lines (Fig. 5A). In contrast, treatment with XB05 led to comet tails only in the malignant cell lines, U937 and A549, with no evidence of DNA damage in the nonmalignant Hs27 cells at comparable concentrations (Fig. 5A). To confirm that the DNA

We reasoned that if an increase in oxidative stress contributed to XB05-induced DNA damage and cell death, then inhibiting this process via addition of exogenous antioxidants should block the activity of XB05. Thus, we conducted MTT proliferation assays for A549 cells using increasing concentrations of XB05 in the presence of antioxidants, including N-acetylcysteine (NAC), reduced glutathione (GSH), or L-ascorbic acid (AA), used at concentrations reported in the literature to have antioxidant effects without cytotoxicity. We found that the thiol-based antioxidants, NAC and GSH, could diminish the antiproliferative effects of XB05, while AA

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

117

Fig. 5. XB05 induces DNA damage in malignant cells. (A) Alkaline comet assays for detection of DNA damage for cells treated as indicated for 72 h. TBHP was a positive control for oxidative DNA damage. Decreased DNA fluorescence in nucleoid comet heads (arrows) and increased smearing of DNA into comet tails (asterisks) are indicated. Scale bar, 400 μm. Data shown are representative of at least two individual experiments for each cell line. (B) Western blot analysis for γ-H2AX and GAPDH (loading control) in whole cell extracts from cells treated with vehicle, UV irradiation (800 μJ, positive control for DSBs), or XB05 (1 μM for U937, 2 μM for A549) after 24 h. Images shown are representative of three individual experiments in each cell line. (C) Flow cytometric detection of γ-H2AX in cells treated as described in part B. Bar graph represents mean þ /– SEM for 3 individual experiments in each cell line. Statistical comparison indicated significant differences between vehicle and XB05-treated groups in both cell lines (*P r 0.05).

had no effect (Fig. 6A). NAC and GSH also protected cells from XB05induced apoptosis and nonapoptotic cell death (Fig. 6B), and abrogated the effect of XB05 on clonogenic survival (Fig. 6C). Again, ascorbic acid treatment had no protective effect in these assays, and even slightly enhanced the proapoptotic activity of XB05 (Fig. 6B). The same experiments using U937 cells showed that NAC and GSH (but not AA) could also protect these cells from the antiproliferative and cytotoxic effects of XB05 (Supplementary Fig. S7). XB05 is weakly reactive with cysteine and glutathione in vitro We initially assumed that the protective effects of NAC and GSH were due to their antioxidant properties, but were intrigued by the inability of ascorbic acid (AA) to protect cells from XB05. We reasoned that this may be due to AA having prooxidant properties in some circumstances [31,32]. However, an alternative interpretation of our results is that XB05 can react directly with thiol groups, which are present in NAC or GSH but not AA. Reaction with NAC or GSH could potentially inhibit XB05 activity by sequestering the compound in the medium and prevent it from entering cells, as previously described for another compound [33]. To assess this possibility, we chose to examine the in vitro reactivity of XB05

with relevant thiols using the 5,5′-dithiobis-2-nitrobenzoic acid assay. This is a colorimetric assay for free sulfhydryl determination based on the reaction of DTNB (a disulfide) with thiol groups to form 2-nitro-5-thiobenzoate (NTB), which ionizes to NTB2- dianion in alkaline aqueous solution [34]. NTB2- has a yellow color which can be quantified by measuring absorbance of visible light at 412 nm [34]. We incubated increasing concentrations of XB05 (or an equivalent amount of vehicle as control) with the biological thiols, glutathione and cysteine (Cys), and monitored conversion of DTNB to NTB2-, as indicated in Fig. 7. As positive controls, we also tested diethyl maleate and maleimide, two well-known thiolreactive electrophiles. We observed that when incubated with Cys or GSH at pH 8 for 30 min at 37 1C, XB05 could lead to depletion of free thiols (presumably by reacting with Cys or GSH) only when present at very high concentrations, e.g., 5 mM XB05 led to 40-50% depletion (Fig. 7). In contrast, maleimide was very much more reactive and could completely deplete free Cys or GSH under the same conditions (Fig. 7). DEM was less reactive than maleimide, but still substantially more reactive than XB05 (Fig. 7). Although the conditions of the DTNB assay (e.g., 0.25 mM GSH, 5 mM XB05, pH 8.0) do not mimic those of the cell culture assays (e.g., 5 mM GSH, 1 mM XB05, pH 7.4), these results suggest that the

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

118

Normalized OD 570 nm

H2O 2.5 mM NAC 5 mM NAC

H2O 100 uM Ascorbic Acid 150 uM Ascorbic Acid

H2O GSH 2.5 mM GSH 5 mM

1.2

1.2

1

1

1

0.8

0.8

0.8

0.6

0.6

0.6

0.4

0.4

0.4

0.2

0.2

0.2

0

0 0

0.5

1

1.5

2

2.5

1.2

0 0

0.5

1

1.5

2

2.5

0

0.5

1

1.5

AA 100

M

2

2.5

XB05 Concentration ( M)

Negative Control

GSH 5 mM

NAC 5 mM

Vehicle

+ XB05 2 M

PI FITC-Annexin V

XB05 ( M)

Vehicle

100 AA 5 mM NAC 5 mM GSH

0.75

120

1.0 % of Control Colony Number

0.5

100 80 60 40 20 0

0

0.5 0.75 1 Vehicle

0

0.5 0.75 1 AA 100 M

0

0.5 0.75 1 NAC 5 mM

0

0.5 0.75 1

XB05 Conc. ( M)

GSH 5 mM

Fig. 6. Thiol-based antioxidants inhibit the antiproliferative and cytotoxic effects of XB05. (A) MTT proliferation assays for A549 cells treated with XB05 in combination with antioxidants N-acetylcysteine (NAC), reduced glutathione (GSH), or ascorbic acid (AA) as indicated for 72 h. (B) Flow cytometric analysis of Annexin/PI staining for A549 cells treated with vehicle or XB05 þ /– antioxidants for 72 h. (C) Clonogenic cell survival assays were conducted in A549 cells treated as indicated and grown in culture for 10 days (see Materials and methods). Bar graph indicates mean values ( þ/– SEM) for colony growth inhibition from 3 independent experiments. MTT and flow cytometric assays were also performed in U937 cells (see Supplementary Fig. S7) and gave the same results.

protective effects of GSH illustrated in Fig. 6 (and Fig. 4A) may be due to reaction of XB05 with GSH. In an attempt to further address this issue, we examined whether the protective effects of GSH in cell cultures assays depended on when it was added relative to XB05. The results of these studies indicate that adding GSH 1 h after XB05 decreases its protective activity compared to when it is added before or at the same time (Supplementary Fig. S8). These data suggest that direct reaction with thiols most likely plays a role in XB05 activity, but further research will be required to fully resolve this and other remaining mechanistic questions.

Discussion We have identified XB05 as a synthetic small molecule with cancer-selective antiproliferative and cytotoxic activity. As a fluorinated/brominated acetylenic alcohol, this compound represents a new type of potential anticancer agent, as no chemically similar bioactive compounds have been previously reported (to our knowledge). Because it was not designed to inhibit any specific protein or activity, the detailed molecular mechanism of XB05 is still unknown, but the studies reported herein indicate that

1.2

119

+ Cys

DMSO Veh

1

XB05

0.8

DEM

0.6

Maleimide

0.4 0.2 0 0

1

2

3

4

5

6

XB05/ DEM/ maleimide concentration (mM) Normalized OD 412 nm

induction of oxidative stress and modulation of cellular stress responses likely play important roles in XB05 activity. Cancer cells have much higher levels of ROS production than normal cells because of increased metabolism, mitochondrial dysfunction, activation of oncogenes, and loss of tumor suppressors [9]. Although low levels of ROS can promote tumor cell proliferation, survival, and metastasis [35–39], high levels of ROS are toxic and cancer cells must adapt by upregulating their antioxidant defenses in order to survive [40–45]. Several strategies have been pursued to try to exploit the differences in stress levels between malignant and normal cells for cancer therapy. Many of these have focused on decreasing cell stress (i.e., antioxidant therapy) in order to try to prevent cancer initiation and slow the progression of premalignant disease, but there is also an emerging interest in using prooxidative agents for treatment of advanced cancers. The rationale behind this latter approach is that induction of additional oxidative stress may push the already highly stressed cancer cells (but not normal cells) over the tolerable threshold. Similarly, agents that inhibit stress defense mechanisms are expected to cause selective toxicity because cancer cells have much higher levels of baseline stress than normal cells and are therefore much more reliant on these defenses [9,46,47]. In fact, there is growing evidence that agents which induce oxidative, electrophilic, or proteotoxic stress—especially when “antistress” responses are simultaneously inhibited—can selectively kill cancer cells by overwhelming their ability to maintain cellular stress at subtoxic levels [9,48,49]. Examples include the natural products, parthenolide, piperlongumine, withaferin A, phenylethyl isothiocyanate (PEITC), erastin, englerin A, and lanperisone [50,51], as well as endogenous reactive lipids, such as 4-hydroxynonenal (4-HNE) and 15-deoxy-Δ(12,14)-prostaglandin J2 (15d-PGJ2) [52,53]. Many of these compounds contain electrophilic centers, such as α,β-unsaturated carbonyls. Intriguingly, it was recently shown that elevation of ROS per se is not sufficient to induce cancer cell death, but that ROS inducers are often cytotoxic when combined with BSO to inhibit glutathione synthesis [50], suggesting that potent cytotoxicity requires both ROS induction and GSH depletion. Notably, in the same study, a few ROS-inducing small molecules that also contained electrophilic centers were found to be selectively cytotoxic toward oncogene-transformed cells and, like XB05, those molecules had decreased activity in the presence of thiol-based antioxidants but not in the presence of nonthiol antioxidants [50]. Thus, our data suggest that XB05 belongs to an emerging class of electrophilic small molecules with shared features including the abilities to increase ROS, to modulate glutathione levels and cellular stress response pathways, and to exert cancer-selective cytotoxicity. Moreover, XB05 may have several advantages over these other oxidative stress inducers. For instance, most of the molecules noted above are complex natural products, which must be purified from the source or made by lengthy synthetic routes, whereas XB05 is a simple molecule with a straightforward synthesis and is amenable to chemical optimization. Furthermore, XB05 has antiproliferative activity in the submicromolar range for many of the malignant cell lines examined, making it more potent than most of the other reported oxidative stress inducers [54–56] and worthy of further investigation as a potential therapeutic agent. The molecular targets of XB05 and the mechanisms by which it generates ROS and modulates stress responses remain to be elucidated. Toxic oxidative stress can be a consequence of excessive ROS production or impaired ROS scavenging capacity or a combination of both. These can arise from many causes and, based on the properties of other cancer-selective oxidative stress inducers, we can speculate on possible mechanisms that may contribute to XB05 activity. Our data hint that, like many of the other molecules noted above, XB05 may react selectively with thiols and

Normalized OD 412 nm

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

1.2

+ GSH

DMSO Veh

1

XB05

0.8

DEM

0.6

Maleimide

0.4 0.2 0 0

2

4

6

XB05/ DEM/ maleimide concentration (mM) Fig. 7. XB05 is weakly reactive with thiols in vitro: 0.25 mM cysteine (Cys, top panel) or glutathione (GSH, bottom panel) was incubated with XB05, diethyl maleate (DEM), or maleimide at the indicated concentrations, or with an equivalent amount of vehicle (DMSO), for 30 min at 37 1C. 5,5′-Dithiobis-2-nitrobenzoic acid (DTNB) assays were conducted as described under Materials and methods. Absorbance for triplicate wells was determined at 412 nm. Points represent the mean and standard deviation for triplicate wells. Data shown are representative of two independent experiments.

that could explain many features of its activity. For example, proteins involved in cellular homeostatsis and stress response are often regulated by thiol-reactive electrophiles via covalent modification of key cysteine or selenocysteine residues. These include thioredoxin, thioredoxin reductase, hsp90, hsp70, 26S proteasome, adenine nucleotide transporter, ATP synthase, NF-κB, and Keap1 [7]. Reaction with these candidate targets might also underlie the ability of XB05 and the other molecules noted to deplete glutathione, which appears to be a major determinant of their cytotoxicities (Fig. 4A and B and [50]). For example, inhibition of NF-κB could be responsible for the observed downregulation of Nrf2 and GSH in U937 cells [25], or alternatively the GSH depletion could result from inhibition of thioredoxin or thioredoxin reductase, since these enzymes are involved in converting GSSG back to reduced GSH [57,58]. Direct reaction of XB05 with GSH may also play a role; it cannot account for the depletion of GSH (1 mM XB05 could not directly deplete the millimolar concentrations of GSH in cells), but the formation of XB05-glutathione adducts could potentially inhibit GSH-utilizing enzymes, many of which regulate cellular response to stress [57,59]. Finally, the relatively low reactivity of XB05 compared to other thiol-reactive electrophiles is worthy of discussion because it may represent a substantial advantage in terms of therapeutic development. We postulate that, unlike reactive electrophiles that may react nondiscriminately with diverse biological nucleophiles and/or be deactivated before entering the cell, XB05 will react preferentially with redox-reactive cysteines and selenocysteines (Sec). Proteins containing Sec (which has a similar structure to cysteine, but with an atom of selenium taking the place of sulfur) are particularly good candidates as XB05 targets because Sec has a pKa of 5.5 (vs 8.0 for Cys), meaning it will be deprotonated to its more reactive anionic form at physiological pH [60]. Due to its high reactivity, selenocysteine is present in the active sites of many enzymes with antioxidant functions, such as thioredoxin reductases and glutathione peroxidases. In summary, XB05 is a new molecule that kills several types of malignant cells and has less effect on nonmalignant cells. Its

120

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121

mechanism of action appears to involve induction of excessive oxidative stress, which may be related to its ability to react slowly with thiol-containing biomolecules. The differential response of various cell lines to XB05 may depend on a combination of factors, including the metabolism of XB05 in those cells, baseline levels of stress, and their capacities for mitigating oxidative stress and DNA damage. Further research is required to fully elucidate the mechanism of XB05, but the ability of this compound to inhibit antioxidant defenses in leukemia cells while increasing stressprotective glutathione in nonmalignant cells suggests its potential utility as a therapeutic or chemosensitizing agent.

Acknowledgments The authors thank members of their laboratories and departmental colleagues for constructive comments. We are grateful to John Eaton, Joe Burlison, John Trent, Ned Smith, Jian Cai, Kerry Barnhart, and Alex Bridges for helpful suggestions, and to the funding agencies and programs that have supported this work, including Kentucky Lung Cancer Research Program (grants to P.J. B.); Kentucky Commercialization Fund (1324-RFP-012 to P.J.B.); Institute for Molecular Diversity and Drug Design (fellowship to F. R.S.); Spatola Endowment (fellowship to F.R.S.); Brown Cancer Center, University of Louisville.

Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.freeradbiomed. 2013.12.002.

References [1] Xu, B.; Mae, M.; Hong, J. A.; Li, Y.; Hammond, G. B. An efficient synthesis of difluoropropargyl bromides. Synthesis 2006(5):803–806; 2006. [2] Hallock, Y. F.; C.J.n.; Balaschak, M. S.; Alexander, M. R.; Prather, T. R.; Shoemaker, R. H.; Boyd, M. R. Antitumor activity and stereochemistry of acetylenic alcohols from the sponge Cribrochalina vasculum. J. Nat. Prod. 58 (12):1801–1807; 1995. [3] Girvan, A. C.; T.Y.; Casson, L. K.; Thomas, S. D.; Jüliger, S.; Ball, M. W.; Klein, J. B.; Pierce Jr W. M.; Barve, S. S.; Bates, P. J. AGRO100 inhibits activation of nuclear factor-kappaB (NF-kappaB) by forming a complex with NF-kappaB essential modulator (NEMO) and nucleolin. Mol. Cancer Ther. 5(7):1790–1799; 2006. [4] Hsiang, Y., et al. Camptothecin induces protein-linked DNA breaks via mammalian DNA topoisomerase I. J. Biol. Chem. 260:14873–14878; 1985. [5] Indran, I. R., et al. Recent advances in apoptosis, mitochondria and drug resistance in cancer cells. Biochim. Biophys. Acta Bioenerg. 1807(6):735–745; 2011. [6] Vakifahmetoglu, H.; Olsson, M.; Zhivotovsky, B. Death through a tragedy: mitotic catastrophe. Cell Death Differ. 15(7):1153–1162; 2008. [7] Higdon, A., et al. Cell Signaling by reactive lipid species: new concepts and molecular mechanisms. Biochem. J. 442:453–464; 2012. [8] Schopfer, F. J.; Cipollina, C.; Freeman, B. A. Formation and signaling actions of electrophilic lipids. Chem. Rev. 111(10):5997–6021; 2011. [9] Trachootham, D.; Alexandre, J.; Huang, P. Targeting cancer cells by ROSmediated mechanisms: a radical therapeutic approach? Nat. Rev. Drug Discov. 8(7):579–591; 2009. [10] Karlsson, M., et al. What does the commonly used DCF test for oxidative stress really show? Biochem. J. 428(2):183–190; 2010. [11] Boveris, A. Mitochondrial production of superoxide radical and hydrogen peroxide. In: Reivich, M., et al., editors. Tissue hypoxia and ischemia. US: Springer; 1977. p. 67–82. [12] Fukai, T.; Ushio-Fukai, M. Superoxide dismutases: role in redox signaling, vascular function, and diseases. Antioxid. Redox Signal. 15(6):1583–1606; 2011. [13] Robinson, K. M., et al. Selective fluorescent imaging of superoxide in vivo using ethidium-based probes. Proc. Natl. Acad. Sci. USA 103(41):15038–15043; 2006. [14] Forkink, M., et al. Detection and manipulation of mitochondrial reactive oxygen species in mammalian cells. Biochim. Biophys. Acta Bioenerg. 1797(6–7): 1034–1044; 2010.

[15] Zielonka, J.; Kalyanaraman, B. Hydroethidine- and MitoSOX-derived red fluorescence is not a reliable indicator of intracellular superoxide formation: Another inconvenient truth. Free Radic. Biol. Med 48(8):983–1001; 2010. [16] Alexandre, A.; Lehninger, A. L. Bypasses of the antimycin A block of mitochondrial electron transport in relation to ubisemiquinone function. Biochim. Biophys. Acta Bioenerg 767(1):120–129; 1984. [17] Deponte, M. Glutathione catalysis and the reaction mechanisms of glutathione-dependent enzymes. Biochim. Biophys. Acta Gen. Subj. ; 2012. (Article in press). [18] Mitsuishi, Y.; Motohashi, H.; Yamamoto, M. The Keap1-Nrf2 system in cancers: stress response and anabolic metabolism. Front. Oncol. ; 2012. [19] Taguchi, K.; Motohashi, H.; Yamamoto, M. Molecular mechanisms of the Keap1–Nrf2 pathway in stress response and cancer evolution. Genes Cells 16 (2):123–140; 2011. [20] Griffin, O.; Meister, A. Potent and specific inhibition of glutathione synthesis by buthionine sulfoximine (S-n-butyl homocysteine sulfoximine). J. Biol. Chem. 254:7558–7560; 1979. [21] Plummer, J., et al. Chemical depletion of glutathione in vivo. In: Jakoby, W., editor. Methods in enzymology. New York: Academic Press; 1981. p. 50–59. [22] Sporn, M. B.; Liby, K. T. Nrf-2 and cancer: the good, the bad and the importance of context. Nat. Rev. Cancer 12(8):564–571; 2012. [23] moinova, h. r.; mulcahy, r. t. up-regulation of the human γ-glutamylcysteine synthetase regulatory subunit gene involves binding of Nrf-2 to an electrophile responsive element. Biochem. Biophys. Res. Commun. 261(3):661–668; 1999. [24] Wild, A.; Mulcahy, R. Regulation of gamma-glutamylcysteine synthetase subunit gene expression: insights into transcriptional control of antioxidant defenses. Free Radic. Res. 32(4):281–301; 2000. [25] Rushworth, S. A., et al. The high Nrf2 expression in human acute myeloid leukemia is driven by NF-κB and underlies its chemo-resistance. Blood 120 (26):5188–5198; 2012. [26] Singha, A., et al. Dysfunctional KEAP1–NRF2 interaction in non-small-cell lung cancer. PLoS Med. 3(10):e420; 2006. [27] Donnelly, N., et al. The eIF2α kinases: their structures and functions. Cell. Mol. Life Sci. :1–19; 2013. [28] Lee, A. S. GRP78 Induction in cancer: therapeutic and prognostic implications. Cancer Res. 67(8):3496–3499; 2007. [29] Aarti, I.; Rajesh, K.; Ramajah, K. Phosphorylation of eIF2 alpha in Sf9 cells: a stress, survival and suicidal signal. Apoptosis 16(6):672–692; 2012. [30] Rogakou, E., et al. DNA double-stranded breaks induce histone H2AX phosphorylation on serine 139. J. Biol. Chem. 273(10):5858–5868; 1998. [31] Bijur, G. N.; Hitchcock, A. M.; Williams, C. L. MV Antimutagenic and promutagenic activity of ascorbic acid during oxidative stress. Environ. Mol. Mutagen. 30:339–345; 1997. [32] Bahlis, N. J., et al. Feasibility and correlates of arsenic trioxide combined with ascorbic acid-mediated depletion of intracellular glutathione for the treatment of relapsed/refractory multiple myeloma. Clin. Cancer Res. 8(12): 3658–3668; 2002. [33] Mi, L., et al. A cautionary note on using N-acetylcysteine as an antagonist to assess isothiocyanate-induced reactive oxygen species-mediated apoptosis. Anal. Biochem. 405(2):269–271; 2010. [34] Ellman, G. L. Tissue sulfhydryl groups. Arch. Biochem. Biophys. 82(1):70–77; 1959. [35] Nishikawa, M. Reactive oxygen species in tumor metastasis. Cancer Lett. 266 (1):53–59; 2008. [36] Kim, E. -Y., et al. BLT2 promotes the invasion and metastasis of aggressive bladder cancer cells through a reactive oxygen species-linked pathway. Free Radic. Biol. Med 49(6):1072–1081; 2010. [37] Shimojo, Y., et al. Attenuation of reactive oxygen species by antioxidants suppresses hypoxia-induced epithelial-mesenchymal transition and metastasis of pancreatic cancer cells. Clin. Exp. Metastasis :1–12; 2012. [38] Tochhawng, L., et al. Redox regulation of cancer cell migration and invasion. Mitochondrion ; 2012. (e-pub ahead of print). [39] Wu, W. -S. The signaling mechanism of ROS in tumor progression. Cancer Metastasis Rev. 25(4):695–705; 2006. [40] Oberley, T.; Oberley, L. Antioxidant enzyme levels in cancer. Histol. Histopathol. 12(2):525–535; 1997. [41] Kekec, Y., et al. Antioxidant enzyme levels in cases with gastrointesinal cancer. Eur. J. Intern. Med. 20(4):403–406; 2009. [42] Ramanathan, B., et al. Resistance to paclitaxel is proportional to cellular total antioxidant capacity. Cancer Res. 65(18):8455–8460; 2005. [43] Young, T. W., et al. Activation of antioxidant pathways in ras-mediated oncogenic transformation of human surface ovarian epithelial cells revealed by functional proteomics and mass spectrometry. Cancer Res 64 (13):4577–4584; 2004. [44] Hu, Y., et al. Mitochondrial manganese-superoxide dismutase expression in ovarian cancer: role in cell proliferation and response to oxidative stress. J. Biol. Chem 280(47):39485–39492; 2005. [45] Saydam, N., et al. Determination of glutathione, glutathione reductase, glutathione peroxidase and glutathione S-transferase levels in human lung cancer tissues. Cancer Lett. 119(1):13–19; 1997. [46] Schumacker, P. Reactive oxygen species in cancer cells: live by the sword, die by the sword. Cancer Cell 10(3):175–176; 2006. [47] Pelicano, H.; Carney, D.; Huang, P. ROS stress in cancer cells and therapeutic implications. Drug Resist. Updat. 7(2):97–110; 2004.

F.R. Šalipur et al. / Free Radical Biology and Medicine 68 (2014) 110–121 [48] Tew, K. D.; Townsend, D. M. Redox platforms in cancer drug discovery and development. Curr. Opin. Chem. Biol. 15(1):156–161; 2011. [49] Wondrak, G. T. Redox-directed cancer therapeutics: molecular mechanisms and opportunities. Antiox. Redox Signal. 11(12):3013–3069; 2009. [50] Adams, D. J.; Theriault, B. Z.; Wang, J. R.; Stern, A. J.; Wagner, A. M.; Shamji, B. K.; Schreiber, A. F.; High-throughput, S. L. screening identifies small-molecule enhancers of reactive oxygen species that are nontoxic or cause genotypeselective cell death. ACS Chem. Biol. ; 2013. ([Epub ahead of print]). [51] Santagata, S., et al. Using the heat-shock response to discover anticancer compounds that target protein homeostasis. ACS Chem. Biol. 7(2):340–349; 2011. [52] Semlitsch, T.; T.H.; Zarkovic, N.; Borovic, S.; Purtscher, M.; Hohenwarter, O.; Schaur, R. J. Differential influence of the lipid peroxidation product 4-hydroxynonenal on the growth of human lymphatic leukaemia cells and human periopherial blood lymphocytes. Anticancer Res. 22(3):1689–1697; 2002. [53] Hassane, D. C., et al. Discovery of agents that eradicate leukemia stem cells using an in silico screen of public gene expression data. Blood 111(12): 5654–5662; 2008.

121

[54] Raj, L., et al. Selective killing of cancer cells by a small molecule targeting the stress response to ROS. Nature 475:231–234; 2011. [55] Yang, J., et al. Selective targeting of breast cancer cells through ROS-mediated mechanisms potentiates the lethality of paclitaxel by a novel diterpene, gelomulide K. Free Radic. Biol. Med. 51:641–657; 2011. [56] Alexandre, J., et al. Improvement of the therapeutic index of anticancer drugs by the superoxide dismutase mimic mangafodipir. J. Natl. Cancer Inst. 98 (4):236–244; 2006. [57] Cai, W., et al. Small molecule inhibitors of mammalian thioredoxin reductase. Free Radic. Biol. Med. 52(2):257–265; 2012. [58] Lu, J.; Chew, E.; Holmgren, A. Targeting thioredoxin reductase is a basis for cancer therapy by arsenic trioxide. Proc. Natl. Acad. Sci. USA 104:12288–12293; 2007. [59] Powis, G.; Kirkpatrick, D. L. Thioredoxin signaling as a target for cancer therapy. Curr. Opin. Pharmacol. 7(4):392–397; 2007. [60] Byun, B. J.; Kang, Y. K. Conformational preferences and pKa value of selenocysteine residue. Biopolymers 95(5):345–353; 2011.