S-nitrosocysteine and glutathione depletion synergize to induce cell death in human tumor cells: insights into the redox and cytotoxic mechanisms

Abstract

Nitric oxide (NO)-dependent signaling and cytotoxic effects are mediated in part via protein S-nitrosylation. The magnitude and duration of S-nitrosylation are governed by the two main thiol reducing systems, the glutathione (GSH) and thioredoxin (Trx) antioxidant systems. In recent years, approaches have been developed to harness the cytotoxic potential of NO/nitrosylation to inhibit tumor cell growth. However, progress in this area has been hindered by insufficient understanding of the balance and interplay between cellular nitrosylation, other oxidative processes and the GSH/Trx systems. In addition, the mechanistic relationship between thiol redox imbalance and cancer cell death is not fully understood. Herein, we explored the redox and cellular effects induced by the S-nitrosylating agent, S-nitrosocysteine (CysNO), in GSH-sufficient and -deficient human tumor cells. We used L-buthionine-sulfoximine (BSO) to induce GSH deficiency, and employed redox, biochemical and cellular assays to interrogate molecular mechanisms. We found that, under GSH-sufficient conditions, a CysNO challenge (100-500 μM) results in a marked yet reversible increase in protein S-nitrosylation in the absence of appreciable S-oxidation. In contrast, under GSH-deficient conditions, CysNO induces elevated and sustained levels of both S-nitrosylation and S-oxidation. Experiments in various cancer cell lines showed that administration of CysNO or BSO alone commonly induce minimal cytotoxicity whereas BSO/CysNO combination therapy leads to extensive cell death. Studies in HeLa cancer cells revealed that treatment with BSO/CysNO results in dual inhibition of the GSH and Trx systems, thereby amplifying redox stress and causing cellular dysfunction. In particular, BSO/CysNO induced rapid oxidation and collapse of the actin cytoskeletal network, followed by loss of mitochondrial function, leading to profound and irreversible decrease in ATP levels. Further observations indicated that BSO/CysNO-induced cell death occurs via a caspase-independent mechanism that involves multiple stress-induced pathways. The present findings provide new insights into the relationship between cellular nitrosylation/oxidation, thiol antioxidant defenses and cell death. These results may aid future efforts to develop NO/redox-based anticancer approaches.

Graphical Abstract

1. Introduction

Nitric oxide (NO) and its derivatives, collectively known as reactive nitrogen species (RNS), are well-established signaling and cytotoxic molecules [1]. Cytotoxic effects of NO/RNS are well documented, such as in the response of innate immune cells towards pathogens and tumor cells [1,2]. Over recent decades, researchers have sought ways to harness the cytotoxic potential of NO/RNS for developing effective antitumor therapies. These efforts have led to the development of numerous NO/RNS donor drugs, which have shown varied ability to halt the growth of tumor cells in vitro and in vivo [3,4]. One group of RNS of special interest are S-nitrosothiols (RSNO), molecules in which a nitroso group is covalently bound to a cysteine thiol. In this context, studies have examined the effects on tumor cells of various low molecular weight RSNO, including S-nitrosocysteine (CysNO) and S-nitrosoglutathione (GSNO), among others. An important aspect of the action of RSNO donors is their ability to enhance the formation of protein RSNO in cells, which in turn influence a variety of cell biological processes, including those related to cell growth and survival [3-7]. In many cases, it is observed that excessive levels of cellular S-nitrosylation (hyper-nitrosylation) leads to cell dysfunction and damage, and indeed, RSNO donors have been shown to exert cytostatic and cytotoxic effects against various tumor cells [3-7]. However, effective NO/RSNO-based antitumor therapy remains an elusive goal. This is due in part to the challenge imposed by protective mechanisms employed by the tumor cells, which promote cell survival under nitrosative stress.

A large body of research has shown that tumor cells exhibit increased resistance to stress conditions and in particular to oxidative/nitrosative stresses [8,9]. Among the antioxidant protective mechanisms employed by tumor cells, the glutathione (GSH) and thioredoxin (Trx) redox systems are of paramount importance [10,11]. These thiol reducing systems scavenge and protect against several types of RNS, including RSNO. Indeed, GSH and Trx are considered to play a central role in RSNO catabolism [12-14]. Importantly, because GSH/Trx are frequently upregulated in cancer cells [8-11] they may hinder the efficacy of NO/RSNO-based anticancer therapies. What is more, GSH and Trx often cooperate to mitigate redox stress [15]. Of note, previous research has revealed that NO/RSNO donors are capable of disrupting cellular redox homeostasis by interfering with the activity of GSH or Trx systems [16-19]. In this regard, we have recently reported that mammalian Trx reductase (TrxR) is sensitive to inactivation by RSNO donors such as CysNO [20]. In a cellular context, RSNO-dependent inhibition of TrxR is potentiated under limiting GSH conditions [20].

In this work, we sought to characterize further the response of tumor cells to RSNO-mediated stress, in particular with regards to thiol redox alterations, impact on cellular homeostasis and mechanism of toxicity. While conducting these studies, we addressed a related question, regarding the relationship between protein S-nitrosylation and S-oxidation, where S-oxidation mainly refers to the formation of protein disulfides (RSSR) or mixed disulfide (glutathionylation, RSSG). In this regard, Eaton and colleagues have recently argued that protein RSNO are quite unstable in cells because they rapidly react with excess thiols to generate RSSR [21,22]. Yet in experiments shown here, we find that under normal, GSH-sufficient conditions, S-nitrosylation is not accompanied by significant S-oxidation. Rather, only under GSH-deficient conditions, a nitrosative intervention leads to significant thiol oxidation concomitant with loss of cellular reducing activities. This state of potentiated redox stress promotes cellular dysfunction and death. The findings reported herein provide a better understanding of the cellular interplay between thiol nitrosylation, oxidation and antioxidant defenses, and offer new insights into the mechanism of nitroso-redox stress-mediated cancer cell death.

2. Materials and Methods

2.1. Antibodies and reagents

The following antibodies were used throughout this study. Anti-actin (catalogue no. SC-10731) and anti-cofilin (SC-376476) were from Santa Cruz Biotechnology. Anti-Prx1 (ab41906) and anti-Prx3 (ab16751) were from Abcam. Anti-PARP-1 (551025) was from BD Pharmingen. Anti-caspase 3 (9662), anti-cytochrome c (11940) and fluorescently labeled secondary antibodies were purchased from Cell Signaling Technology. CysNO was synthesized by combining an equimolar concentration of L-cysteine with sodium nitrite in 0.2 N HC1, and used within 1 h. Thiopropyl sepharose resin was obtained from GE Health Care. N-ethyl maleimide (NEM), Alexa Fluor 488 phalloidin and tetramethylrhodamine methyl ester perchlorate (TMRM) were purchased from Thermo Fisher Scientific. Other materials were obtained from Sigma unless otherwise indicated.

2.2. Cell culture

HeLa, A549 and EBC-1 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 1 mM sodium pyruvate and 1% penicillin/streptomycin at 37 °C under 5% CO₂. THP-1, H460, H838, H1944 and H1437 cells were maintained in Roswell Park Memorial Institute (RPMI) medium 1640 supplemented with 10% FBS and 1% penicillin/streptomycin at 37 °C under 5% CO₂. THP-1 cells were also supplemented with 1 mM sodium pyruvate. These monocytic cells were differentiated to adherent macrophages by incubation with phorbol 12-myristate 13-acetate (100 ng/ml) for 24 h followed by a 24 h rest period. Tissue culture media and reagents were from Biological Industries (Beit Haemek, Israel).

2.3. Thiol analysis

Cellular thiol content was measured spectrophotometrically after reaction with 5,5’-dithio-bis (2-nitrobenzoic acid) (DTNB, Ellman’s reagent). After different treatments as indicated in the figure legends, HeLa cells were washed and harvested in phosphate-buffered saline (PBS) by centrifugation at 13,000 rpm for 3 min at 4 °C. Cells were then homogenized by 5 passages through a 25 G needle in hypotonic buffer (10 mM Hepes, 3 mM MgCl₂ ,10 mM KCl, 0.1 mM EDTA, pH 7.5), followed by centrifugation at 13,000 rpm for 15 min at 4 °C and the protein concentration was determined by Bradford assay. THP-1 cells were lysed in digitonin lysis buffer as described elsewhere [23]. For measurement of cellular thiol content, 40 μg protein was incubated in a final volume of 200 μl of reaction mix (1 mM DTNB, 6 M guanidine hydrochloride, 90 mM HEPES, pH 7.5) for 10 min and the absorbance readings were taken at 412 nm. Values were derived by comparison with GSH standards and were normalized to protein concentration.

2.4. Biochemical analysis of S-nitrosylation

RSNO content was determined by copper-cysteine-based reductive chemiluminescence assay or Saville–Griess assay as detailed elsewhere [24]. Briefly, cells were washed and harvested with PBS, and homogenized as detailed in section 2.3. RSNO content was determined by reductive chemiluminescence using an NO analyzer (CLD88; Eco Medics AG Switzerland). Values were derived by comparison with GSNO standards and were normalized to protein concentration in the extract. Alternatively, Saville–Griess assay was performed as follows. Lysates (0.1 mg protein each) were incubated in a final volume of 200 μl of assay buffer (1% sulfanilamide, 0.1% N-(1-naphthyl) ethylenediamine dehydrochloride, 1% HCl) in the absence or presence of 1 mM HgCl₂ for 30 min and absorbance was measured at 540 nm. Mercury-dependent absorbance was converted to RSNO concentrations using GSNO standards treated identically. In the in vitro denitrosylation experiments shown in Fig. 1C, prior to RSNO measurements, excess reductants were removed using an ultrafiltration centrifugation device as previously described [24].

Fig. 1. Effects of CysNO treatment on RSNO and RSH levels in human acute monocytic leukemia-derived macrophages.

(A) THP-1-derived macrophages were treated with CysNO (0, 100, 200 or 500 μM) for 10 min (top) or with 200 μM CysNO for different times (bottom). RSNO and RSH levels in whole cell lysates were measured using Saville Griess assay and Ellman’s assay, respectively. (B) Nitrosylated lysates, prepared from THP-1 cells treated for 10 min with 200 μM CysNO, were incubated with 50 mM ascorbate (Asc) or 25 mM dithiothreitol (DTT) followed by analysis of RSNO content as in A. Data shown in graphs represent mean ± SD (n = 3).

2.5. Ox-RAC analysis of protein nitrosylation/oxidation

Detection of oxidized proteins was performed using oxidized cysteine resin-assisted capture (Ox-RAC) method [25]. Following different treatments as indicated in the figure legends, cells were incubated with PBS containing 100 mM NEM for 10 min at room temperature and then lysed in lysis buffer (50 mM HEPES, 1% NP-40, 150 mM NaCl, 1 mM EDTA, 0.1 mM DTPA, 50 mM NEM, with protease inhibitors, pH 7.5). Cell debris was removed by centrifugation at 13,000 rpm for 15 min at 4 °C. To detect the oxidized proteins, a total of 0.5 mg protein (0.8 mg/ml concentration) was used for each experimental condition and the thiol blocking step was performed at 50 °C, in the presence of 20 mM NEM with frequent vortexing. After blocking, excess NEM was removed by acetone precipitation with 3 volumes of acetone at −20 °C for 30 min. The proteins were recovered by centrifugation at 2,000g for 5 min, and the pellets were washed three times with 70% cold acetone and resuspended in HENS buffer (250 mM HEPES, 1 mM EDTA, 0.1 mM neocuproine, 1% SDS, pH 7.7). The samples were then incubated with 25 mM dithiothreitol (DTT) in the dark for 30 min at room temperature. The excess DTT was removed by acetone precipitation with 3 volumes of acetone at −20 °C for 30 min. The proteins were recovered again by centrifugation at 2,000g for 5 min, and the pellets were washed three times with 70% cold acetone and resuspended in HENS buffer. Immediately thereafter, 80 μl thiopropyl sepharose resin was added to each sample. Following rotation in the dark for 1 h at room temperature and then overnight at 4 °C, the resin was washed with 4 x 1 ml HENS buffer, then 2 x 1 ml HENS/10 buffer (HENS diluted 1:10). Captured proteins were eluted with 30 μl HENS/10 containing 100 mM 2-mercaptoethanol for 20 min at room temperature, and analyzed by Western blotting using specific antibodies, or stained with Coomassie blue staining solution for total protein analysis. Proteins were visualized using the Odyssey infrared imaging system (LI-COR Biosciences).

2.6. Assessment of cell viability and morphological changes

Cells were plated in 96-well plates and grown overnight. Thereafter, the cells were treated with BSO and/or CysNO as indicated in the figure legends. After 24 h, cell numbers were quantified using the CyQuant cell proliferation assay kit (Invitrogen) according to the manufacturer's protocol. In addition, cells were tracked with the Essen IncuCyte ZOOM live cell imaging system. Bright field images were captured every 10 min for the duration of the experiment.

2.7. Confocal analysis of the actin cytoskeleton

Cells were washed with PBS, fixed with 4% paraformaldehyde in PBS for 15 min, washed three times with PBS, permeabilized with 0.2% Triton X-100 for 5 min and washed again three times with PBS. Actin and nuclei were stained with phalloidin and 4′, 6-diamidino-2-phenylindole (DAPI) respectively for 10 min in PBS at room temperature, and washed five times with PBS. Stained cells were visualized and photographed using the Zeiss LSM 800 inverted confocal microscope (Thornwood, NY) equipped with an 63x/1.4 NA oil objective, and solid-state lasers.

2.8. Biochemical fractionation

Cells were washed, harvested with PBS, homogenized by 5 passages through a 25 G needle in hypotonic buffer and then centrifuged at 600g for 10 min at 4 °C to pellet the nuclei and unbroken cells. The supernatant (containing cytoplasm and mitochondria fractions) was centrifuged at 15,000g for 30 min at 4 °C. The new supernatant, containing the cytoplasm fraction, was transferred to a new tube; the new pellet, containing the mitochondria fraction, was washed twice in homogenization buffer and then resuspended in lysis buffer. Equal amount of total protein from each fraction was subjected to SDS-PAGE followed by Western blotting.

2.9. Determination of mitochondrial membrane potential

After cell treatments, the cells media was replaced to new media without phenol red, and then TMRM (100 nM) was added to the cell medium for 15 min at 5% CO₂ and 37 °C in the dark. Fluorescent images were captured by the Zeiss LSM 700 upright confocal microscope (Thornwood, NY).

2.10. Measurement of cellular ATP

Cells were plated at 7,500 cells/well in 96-well plates and grown overnight. After treatments, intracellular ATP was determined by CellTiter-Glo Luminescent Kit (Promega) according to the manufacturer's protocol.

3. Results

3.1. Quantitative assessment of RSNO formation in nitrosatively-challenged cells

The balance between protein nitrosylation and oxidation may have functional consequences during cell signaling and stress responses. In this regard, Eaton and colleagues have recently argued that cell exposure to nitrosative stimuli, such as CysNO or lipopolysaccharide, results mainly in the formation of RSSR as a stable modification. It was reasoned that RSNO, once formed, quickly react with RSH to yield RSSR, a process that can be summarized as: RSNO + RSH → RSSR + HNO [22]. However, much of the evidence presented in that study was based upon assays that are only semi-quantitative, which may lead to inaccurate conclusions regarding the actual RSNO and RSSR concentrations in cells. With these issues in mind, we initiated our studies by taking a quantitative approach to determine the extent of protein nitrosylation and oxidation in human cells exposed to a nitrosative challenge. For this purpose, we first used THP-1 monocytic leukemia cells, which were differentiated into macrophages with phorbol 12-myristate 13-acetate. Exposure of THP-1-derived macrophages to CysNO (0-500 μM) lead to rapid, robust and dose-dependent formation of cellular RSNO, as determined by the Saville-Griess assay (Fig. 1A, top). The increase in RSNO was closely mirrored by a decrease in RSH content, as determined by Ellman’s assay (Fig. 1A, top). A time course experiment demonstrated the transient formation of RSNO, peaking at 10 min after CysNO exposure and returning to baseline levels at 60 min (Fig. 1B, bottom). Here too, RSNO levels were inversely mirrored by the changes in RSH content. Indeed, the sum, RSNO+RSH, was virtually constant across all the samples (Fig. 1A). Of note, this sum should decrease in a sample were RSSR species accumulate, but this was not seen in any of the samples. Thus, these quantitative measurements clearly demonstrate that a nitrosative intervention (exposure to CysNO) leads mostly to formation of RSNO and is not accompanied by significant oxidation (such as RSSR formation). The data indicate that the transient nature of the RSNO is largely due to a direct denitrosylation process, RSNO→RSH, whereas the process, RSNO→RSSR→RSH, represents a minor route.

It is important to note that in the abovementioned study [22] the investigators employed indirect, “switch assays” to assess and compare the formation of RSNO and RSSR. Specifically, an ascorbate (Asc)-dependent assay was used to detect RSNO whereas a dithiothreitol (DTT)-dependent assay to detect RSNO+RSSR. The difference between the signals obtained (ie, DTT-Asc) was used to estimate RSSR content [22]. This experimental approach assumes that Asc and DTT have similar reducing capacity. However, this assumption ignores the slow rate of RSNO reduction by Asc, as demonstrated before [26] and confirmed here. Indeed, using our THP-1 model system, we observed that Asc is poor RSNO reductant, much weaker than DTT. Specifically, incubation of nitrosylated THP-1 lysate with 50 mM ascorbate for 1 h lead to only ~ 10% decrease in RSNO content as compared to ~ 95% reduction obtained with 25 mM DTT over 30 min (Fig. 1B). These findings reinforce the idea that Asc is a week RSNO reductant, and therefore, Asc-based assays are inadequate for quantitative assessment of RSNO. In particular, the data strongly suggest that Asc- and DTT-based assays done in parallel (as in [22]) greatly underestimate RSNO content relative to RSSR, and are therefore unreliable for determining RSNO/RSSR ratios.

3.2. Glutathione depletion enhances S-nitrosocysteine-induced cellular nitrosylation, oxidation, and promotes cell death

We next employed HeLa human cervical carcinoma cells in order to explore redox processes occurring in cancer cells subjected to nitrosylation-promoting interventions. Motivated by our previous work [20], we examined cellular effects elicited by CysNO treatment under GSH-sufficient and GSH-deficient conditions. We used the γ-glutamylcysteine synthetase inhibitor L-buthionine-sulfoximine (BSO) to pharmacologically deplete GSH. We previously reported that subjecting HeLa cells to BSO treatment (100 μM, 24 h) results in ~ 50% decrease in GSH content [20]. Thus, we proceeded to analyze the effects induced by CysNO and BSO/CysNO treatments on cellular nitrosylation/oxidation. The data on cellular RSNO and RSH levels are shown in Fig. 2A. In GSH-sufficient cells, the time profiles of RSNO and RSH were largely similar to that seen before in THP-1 cells. However, in GSH-depleted cells, these profiles were quite different. Indeed, whereas in GSH-sufficient cells RSNO levels returned to baseline at 60 min post-treatment, GSH-depleted cells showed an attenuated recovery, with RSNO levels decreasing by only ~ 65% (relative to maximum levels at the 10 min time point). In addition, the RSH profiles were markedly different. Whereas in the CysNO treatment group there was a modest 14% decrease in RSH content at 10 min with recovery to baseline at 60 min, in the BSO/CysNO group, RSH content was much lower at baseline, was further decreased by 40% at 10 min, and remained essentially unchanged up to 60 min. We also employed the Ox-RAC assay, a semiquantitative method that detects protein S-nitrosylation/oxidation/thiolation (see Material and Methods). Consistent with the above data, this analysis also showed that GSH deficiency leads to enhanced and prolonged protein nitrosylation/oxidation (Supplementary Fig. 1). The sustained nitrosylation/oxidation observed in the GSH-deficient cells (documented both by quantitative RSNO analysis and Ox-RAC analysis) suggests a diminished activity of the endogenous thiol reducing systems.

Fig. 2. Effects of CysNO and BSO/CysNO treatments on HeLa cancer cells: assessment of RSNO/RSH content, redox state of peroxiredoxins, and cell viability.

(A) HeLa cells were pretreated or not with BSO (100 μM, 24 h) followed by treatment with 500 μM CysNO for different times (0, 10, 30 and 60 min). RSNO and RSH levels in whole cell lysates were determined by reductive chemiluminescence and Ellman's assays, respectively, as detailed in Materials and Methods. (B) HeLa cells were treated as in A. The redox states of peroxiredoxin-1 (Prx1) and Prx3 were analyzed by non-reducing immunoblotting. (C) HeLa cells were pretreated or not with BSO (100 μM, 24 h) followed by treatment with the indicated concentrations of CysNO. After additional 24 h, cell viability was determined using CyQuant cell proliferation assay. Data shown in graphs represent mean ± SD (n > 3).

It has been shown that nitrosative challenges may compromise the activity of cellular antioxidant systems, including that of the Trx/TrxR system [17,20]. The cytosolic and mitochondrial Trx systems recycle peroxiredoxin-1 (Prx1) and Prx3, respectively [27,28]. We sought to examine the status of these thiol redox systems in response to nitrosative interventions. To this end, we used non-reducing immunoblotting to analyze the redox states of Prx1/3. As shown in Fig. 2B, Prx1 and 3 were mostly in the reduced state in resting cells, as expected. CysNO treatment hardly affected the redox state of Prx1 while it induced a transient oxidation of Prx3. Distinct from these effects, in GSH-depleted cells, both Prx proteins were extensively oxidized at 10 min after CysNO stimulation and remained so for at least 60 min (Fig. 2B). These results indicate that both the cytosolic and mitochondrial Trx systems maintain their redox activity in CysNO-challenged, GSH-sufficient cells; however, their redox activity is significantly compromised in a setting of GSH deficiency. We then assessed how these nitrosative interventions affect cellular viability. The results revealed that CysNO treatment exerted a growth inhibitory effect yet caused little cytotoxicity (Fig. 2C). In contrast, BSO/CysNO treatment lead to a marked decrease in cell viability (Fig. 2C).

3.3. Glutathione depletion enhances S-nitrosocysteine-induced cell death of lung cancer cells

To extend these observations, we wished to examine the effects of nitrosative challenge on the viability of other types of cancer cells. To this end, we evaluated the effects of CysNO and BSO/CysNO treatments on a panel of six human non-small cell lung carcinoma (NSCLC) cells. Similar to the observation in HeLa cells, treatment with BSO or CysNO alone exerted very little cytotoxicity in all the tested NSCLC cell lines (Fig. 3). Notably, BSO/CysNO combination therapy displayed significant toxicity against four different cell lines (A549, H1437, H1944 and H838), whereas two cell lines (H460 and EBC1) were relatively refractory (Fig. 3). The results thus indicate that decreased GSH levels sensitize many, but not all, cancer cells to CysNO-induced cell death.

Fig. 3. Effects of CysNO and BSO/CysNO treatments on cell viability of NSCLC cells.

NSCLC cells were pretreated or not with BSO (100 μM, 24 h) followed by treatment with the indicated concentrations of CysNO. After additional 24 h, cell viability was determined using CyQuant cell proliferation assay. Data shown in graphs represent mean ± SD (n ≥ 3).

3.4. Nitroso-redox stress-dependent disruption of the actin cytoskeleton

In the course of these experiments we observed by light microscopy that HeLa cells subjected to BSO/CysNO combination treatment underwent notable morphological changes. IncuCyte live-cell imaging showed that these changes occurred within 1 h after exposure to the nitrosative stimulus and involved cell shrinkage and rounding, with apparent weakening of overall attachment to the matrix and clustering of the cells (Fig. 4A, Supplementary Fig. 2). We noticed that these cell clusters were rather loosely attached to the matrix and were easily detached by gentle mechanical agitation (such as pipetting). The live-cell imaging, performed over a 4 h time period, revealed additional features, such as frequent events of membrane blebbing (Supplementary Movie). These observations prompted us to assess the state of the actin cytoskeleton in resting and treated cells. To do so, we used phalloidin staining, which showed long and well-organized actin stress fibers in untreated cells (Fig. 4B), consistent with their elongated and well-spread shape. Treatment with CysNO did not induce significant changes in actin distribution although the filaments appeared less organized at 60 min. In strike contrast, BSO/CysNO treatment induced marked alterations in the number and arrangement of the actin filaments (Fig. 4B). Loss of stress fibers was apparent already at 30 min, without significant change in cell size; at 60 min, extensive breakdown of the actin filaments and aggregation in the cytoplasm was evident, accompanied by massive shrinkage of the whole cell. In parallel, the cell nuclei underwent condensation, as revealed by DAPI staining (Fig. 4B).

Fig. 4. Effects of CysNO or BSO/CysNO treatments on HeLa cell morphology and cytoskeletal organization.

(A) HeLa cells were pretreated or not with BSO (100 μM, 24 h) followed by treatment with 500 μM CysNO for different times. Cells were visualized using IncuCyte ZOOM platform. (B) Cells were pretreated or not with BSO followed by treatment with 500 μM CysNO for the indicated times. The actin cytoskeleton and the cell nuclei were visualized by Phalloidin and DAPI staining. (C) Cells were treated with BSO and/or CysNO as indicated. The nitrosylation/oxidation status of actin and cofilin was assessed by using the Ox-RAC assay.

Several cytoskeletal proteins harbor thiol groups that are susceptible to nitrosylation or other oxidative modifications, which can influence cell form and function. In particular, nitrosylation/oxidation of actin and cofilin is known to promote actin filament depolymerization [29,30]. Thus, we next used the Ox-RAC assay to assess redox changes in these proteins. This analysis revealed that CysNO treatment lead to transient nitrosylation/oxidation of actin and cofilin. In contrast, BSO/CysNO treatment promoted sustained nitrosylation/oxidation of the two proteins (Fig. 4C). Collectively, the findings suggest that under GSH-deficient conditions, a nitrosating intervention triggers extensive and continuous nitrosylation/oxidation of key components of the actin cytoskeleton, which is followed by cytoskeletal collapse and disruption of cellular organization and adhesion.

3.5. Nitroso-redox stress-dependent disruption of mitochondrial function

Oxidative and nitrosative insults can impact mitochondrial functions and thus deleteriously affect cell viability [31,32]. Using TMRM imaging, we examined the effects of nitrosative interventions on the cell ability to maintain mitochondrial transmembrane potential (Δψm). We observed that CysNO treatment lead to gradual decrease in TMRM fluorescence, indicating partial mitochondrial depolarization (Fig. 5A). Notably, the effect of BSO/CysNO treatment was much more pronounced, as evident by the rapid and near-complete collapse of Δψm at 2 h post-treatment (Fig. 5A). We next measured intracellular ATP levels over a period of 0-24 h. CysNO treatment triggered a significant decrease in ATP levels at 6 h, but a recovery phase was evident thereafter, with ATP levels reaching ~ 75% at 24 h (relative to untreated control). A significantly different ATP time profile was seen in BSO/CysNO-treated cells, whereby ATP levels did not substantially change during the first 6 h, but thereafter precipitously and irreversibly declined (Fig. 5B).

Fig. 5. Effects of CysNO and BSO/CysNO treatments on mitochondrial membrane potential and ATP levels.

(A) HeLa cells were pretreated or not with BSO (100 μM, 24 h) followed by treatment with 500 μM CysNO for different times. Mitochondrial membrane potential was analyzed by TMRM fluorescence. Nuclear staining with Hoechst is also shown. (B) HeLa cells were pretreated or not with BSO (100 μM, 24 h) followed by treatment with 500 μM CysNO for different times. Total cellular ATP levels were monitored using a luciferase-based bioluminescence assay. Open circles and squares indicate CysNO treatment and BSO/CysNO treatment, respectively. Data shown in graphs represent mean ± SD (n = 3).

Previous research revealed that thiol oxidation of cofilin is involved in oxidant-induced cell death [33]. Specifically, it was shown that oxidation of cofilin causes it to lose its affinity for actin and to translocate to the mitochondria, where it induces swelling and cytochrome c (Cyt c) release via opening of the permeability transition pore (PTP) [33]. To examine the involvement of this mechanism in our experimental system we performed subcellular fractionation experiments. We found that cell treatment with CysNO resulted in a modest accumulation of cofilin in the mitochondria, an effect that was markedly enhanced in BSO/CysNO-treated cells (Fig. 6A). Further analyses indicated the accumulation of Cyt c in the cytosol of nitrosatively-challenged cells. Similar to the cofilin data, the observed relocalization of Cyt c was more pronounced in cells treated with BSO/CysNO compared to cells treated with CysNO alone (Fig. 6A).

Fig. 6. BSO/CysNO-induced cell death occur via a caspase-independent mechanism.

(A) HeLa cells were pretreated or not with BSO (100 μM, 24 h) followed by treatment with CysNO (500 μM, 1 h). Mitochondrial localization of cofilin and cytosolic localization of cytochrome c were determined by biochemical fractionation followed by immunoblot analysis. (B) HeLa cells were pretreated or not with BSO followed by treatment for various times with 500 μM CysNO. For comparison, cells were also treated for 4 h with the apoptosis inducer staurosporine (STS, 1 μM). Cleavage of caspase-3 and PARP-1 was determined by immunoblot analysis. (C) HeLa cells were pretreated or not with BSO followed by treatment for various times with 500 μM CysNO. For comparison, cells were also treated for 5 min with the alkylating agent N-methyl-N′-nitro-N′-nitrosoguanidine (MNNG, 100 μM). PARylation was assessed by immunoblotting using the anti-PAR antibody. (D) Cells were pretreated with BSO (100 μM, 24 h) and then treated or not with BAPTA (10 μM, 1 h) before exposure to CysNO. Cell viability after 24 h was determined using CyQuant assay. Data presented as mean±SD (n=4), with statistical comparison performed by one-way repeated measure ANOVA, *p < 0.05 compared to vehicle controls.

It is known that mitochondrial membrane permeabilization and Δψm dissipation can promote cell death via apoptotic or non-apoptotic mechanisms [34,35]. In general, mitochondria-mediated apoptosis (also known as intrinsic apoptosis) is mediated by the caspase-9/3 proteolytic cascade [34]. To assess apoptosis we monitored the cleavage of caspase-3. In these experiments, we used the apoptogenic agent staurosporine (STS) as a positive control. We observed that CysNO treatment lead to some activation of caspase-3, as evident by the appearance of cleaved caspase-3 on Western blots (Fig. 6B); however, this cleavage was very minor compared to that induced by STS. Notably, cleaved caspase-3 was not detected in BSO/CysNO-treated cells (Fig. 6B). We also analyzed the cleavage of the caspase-3 substrate PARP-1. Consistent with the above observations, partial cleavage of PARP-1 was seen in CysNO-treated cells (Fig. 6B). Strikingly, PARP-1 protein was not detected at all in BSO/CysNO-treated cells (Fig. 6B). These results prompted us to consider the possibility that a modification of PARP-1 could be involved in this effect. Previous studies have shown that oxidative/nitrosative stress can trigger PARP-1 hyperactivation-mediated cell death [36,37]. As activation of PARP-1 results in increased production of poly (ADP-ribose) moieties and PARylation of target proteins (including auto-PARylation of PARP-1), we next assessed protein PARylation. In these experiments we also employed N-methyl-N′-nitro-N-nitrosoguanidine (MNNG), an alkylating agent known to kill cells via PARP-1 activation [36,37]. This analysis revealed that cell treatment with BSO/CysNO, but not CysNO alone, substantially elevated levels of protein PARylation (Fig. 6C). Finally, because disruption of Ca²⁺ homeostasis has been linked with redox stress and PARP-1-mediated cell death [36] we examined the effect on viability of the Ca²⁺ chelator BAPTA. The results showed that cell treatment with BAPTA partially alleviated BSO/CysNO-induced cytotoxicity (Fig. 6D). Taken together, these data support the involvement of PARP-1 activation and Ca²⁺ overload in BSO/CysNO-mediated cell death.

4. Discussion

Because S-nitrosylation affects many cellular processes, multiple studies have focused on understanding the role of this protein redox modification in tumor cell biology [38-40]. In this regard, there has been considerable interest in the application of RSNO donors to inhibit or kill tumor cells [3,4]. Despite progress in this area, there are still significant challenges to overcome. Our understanding of antioxidant systems in driving tumor growth and chemoresistance is largely based on Nrf2 due to common occurring mutations in its inhibitor Keap1 in lung and other cancers [41,42]. There is a need to better elucidate how the cancer cell reducing systems handle elevated RSNO levels, and how NO/redox imbalance is mechanistically linked to cell death. The present study offers new insights into these questions by providing quantitative data on RSNO-induced redox changes under GSH-sufficient and GSH-deficient conditions and by highlighting some of the key events involved in RSNO-triggered cancer cell death.

Our first set of experiments demonstrated that exposure of GSH-sufficient cells to CysNO leads to a substantial increase in protein S-nitrosylation but not in S-oxidation (such as RSSR or other S-thiolations). These findings are at odds with the recent data and model proposed by Eaton and colleagues [22]. In that study, evidence was presented to support the contention that protein RSNO are unstable in cells because they rapidly react with thiols (which are found in excess) to generate disulfides. However, much of the evidence presented was derived from Asc-dependent assays. Yet, as documented here and elsewhere [26], Asc is a weak RSNO reducing agent, and thus inadequate for quantitative determination of RSNO. In particular, as our data clearly show, Asc is a much weaker RSNO reductant compared to DTT. Therefore, Asc/DTT-based switch assays (as done in [22]) lead to significant underestimation of RSNO relative to RSSR. As for the issue of ‘RSNO stability’, it is recognized that the lifetimes of protein RSNOs vary considerably in cells, and are fundamentally determined by the (site-specific) rates of S-nitrosylation and denitrosylation [14]. In this regard, our findings strongly suggest that in most cases nitrosylated thiols are directly denitrosylated (RSNO→RSH), that is, without the intermediacy of RSSR. The transition RSNO→RSSR constitutes a minor route, which might be favored in certain cases, eg, a nitrosylated thiol that is situated in close proximity to another thiol (as in vicinal thiols). It is also worth emphasizing that the abovementioned model [22] completely ignores the alternative, often preferred chemical reaction of RSNO with RSH, that is, trans-nitrosylation (RSNO+R’SH→RSH+R’SNO). The view that nitrosylation is generally not followed by thiolation is further supported by recent redox proteomic analyses demonstrating (i) the low overlap between RSNO and RSSG sites in vivo, and (ii) that, unlike RSNO, RSSG is largely-independent of NO synthase activity [43].

In stark contrast to GSH-sufficient conditions, under GSH-deficient conditions, cell exposure to CysNO promoted a sustained increase in both S-nitrosylation and S-oxidation. We have previously shown that when GSH is limiting, CysNO treatment leads to functional inactivation of the Trx/TrxR system [20]. In the present study, we observed elevated and sustained oxidation of Prx1 and Prx3 in cells treated with BSO/CysNO but not CysNO alone. These data indicate the inhibition of both the cytosolic and mitochondrial Trx systems upon BSO/CysNO challenge. These findings thus support the idea that BSO/CysNO combination therapy triggers the dual inactivation of the GSH/Trx systems across different cellular compartments. Disabling of the two main thiol reducing systems propagates and amplifies the redox stress (an auto-amplification loop), leading to cellular dysfunction and death [15].

What are the key cellular effects elicited by nitroso-redox imbalance? Our studies suggest that one of the earliest effects induced by BSO/CysNO is oxidation of cytoskeletal proteins (actin, cofilin) leading to cytoskeletal collapse within one hour. These findings are consistent with and expand on previously published results, reporting on the high susceptibility of cytoskeletal proteins to undergo nitrosylation/oxidation, which can shift the dynamic equilibrium between monomeric G-actin and polymeric F-actin towards depolymerization [29,30]. It is recognized that the actin cytoskeleton plays an important role in maintaining cell viability, in part by supporting mitochondrial function and dynamics [44,45]; accordingly, loss of cytoskeletal homeostasis may actively contribute to mitochondrial dysfunction and induction of cell death. In addition, oxidation of cofilin was shown to lead to its translocation to the mitochondria, promoting the opening of the PTP, culminating in cell death [33]. Our data on the localization of cofilin, mitochondrial depolarization and release of Cyt c are in line with this mechanism. Loss of mitochondrial membrane potential, due to PTP opening or mitochondrial outer membrane permeabilization (MOMP), and resultant bioenergetic and metabolic catastrophe are critical events in several forms of necrotic and apoptotic cell death [34,35]. MOMP-induced release of Cyt c often promotes activation of the apoptotic caspase-9/3 cascade. As shown here, BSO/CysNO-induced cell death displays some features that resemble apoptosis such as membrane blebbing and nuclear condensation; nonetheless, our data strongly suggest that cell death is caspase-independent. Indeed, this may be related to the fact that thiol oxidation may inactivate caspases upon cellular redox stress [32]. It is not yet clear if BSO/CysNO-mediated cell death occurs by one of several described mechanisms of caspase-independent cell death, which frequently involve mitochondrial release of toxic proteins such as apoptosis-inducing factor (AIF) and endonuclease G [46]. Accordingly, additional studies are required to establish the exact molecular mechanism of nitroso-redox stress-induced death. It should be noted however that although researchers often attempt to identify a single mechanism that mediates cell death, in some settings, such as oxidant-induced cellular injury, early activation of multiple stress pathways could be important in getting the cancer cell to commit to cell death. Altogether, our data suggest that BSO/CysNO-induced cell death entails mitochondrial dysfunction and energetic collapse, as well as calcium overload and overactivation of PARP- 1. The latter may imply the involvement of DNA damage upstream of PARP-1 activation as well as the subsequent mitochondrial release of AIF [36]. Overall, further research is needed to fully dissect and delineate the contribution of each of these maladaptive responses in the context of BSO/CysNO-dependent cell death.

Despite advances in the development of NO/RSNO-donors as anticancer agents [3,4], significant challenges still exist. In this context, the endogenous thiol antioxidant systems, which are often upregulated in cancer cells, represent an important barrier for translational application of RSNO-based therapy. Our results reinforce the importance of dual inhibition the GSH/Trx systems for sensitizing cancer cells to RSNO-induced cytotoxicity. It is proposed that further elucidation of the interplay between RSNO, the GSH/Trx systems and cell death pathways, along with development of strategies for tumor-targeted delivery of RSNO compounds, will be important for future progress in devising NO/redox-based anticancer approaches.

Highlights.

• Effects of S-nitrosocysteine and glutathione depletion on cancer cells were investigated

• S-nitrosocysteine promoted protein S-nitrosylation, rather than S-oxidation

• Under glutathione-deficient conditions, S-nitrosocysteine induced nitroso-redox stress

• Nitroso-redox stress involved elevated thiol oxidation and decreased reducing capacity

• Nitroso-redox stress induced cancer cell death via a caspase-independent mechanism