The oncology drug elesclomol selectively transports copper to the mitochondria to induce oxidative stress in cancer cells

Free Radical Biology and Medicine 52 (2012) 2142–2150

Contents lists available at SciVerse ScienceDirect

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

Original Contribution

The oncology drug elesclomol selectively transports copper to the mitochondria to induce oxidative stress in cancer cells Masazumi Nagai, Nha H. Vo, Luisa Shin Ogawa, Dinesh Chimmanamada, Takayo Inoue, John Chu, Britte C. Beaudette-Zlatanova, Rongzhen Lu, Ronald K. Blackman, James Barsoum, Keizo Koya, Yumiko Wada n Synta Pharmaceuticals Corp., Lexington, MA 02421, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 December 2011 Received in revised form 16 February 2012 Accepted 20 March 2012 Available online 17 April 2012

Elesclomol is an investigational drug that exerts potent anticancer activity through the elevation of reactive oxygen species (ROS) levels and is currently under clinical evaluation as a novel anticancer therapeutic. Here we report the first description of selective mitochondrial ROS induction by elesclomol in cancer cells based on the unique physicochemical properties of the compound. Elesclomol preferentially chelates copper (Cu) outside of cells and enters as elesclomol–Cu(II). The elesclomol– Cu(II) complex then rapidly and selectively transports the copper to mitochondria. In this organelle Cu(II) is reduced to Cu(I), followed by subsequent ROS generation. Upon dissociation from the complex, elesclomol is effluxed from cells and repeats shuttling elesclomol–Cu complexes from the extracellular to the intracellular compartments, leading to continued copper accumulation within mitochondria. An optimal range of redox potentials exhibited by copper chelates of elesclomol and its analogs correlated with the elevation of mitochondrial Cu(I) levels and cytotoxic activity, suggesting that redox reduction of the copper triggers mitochondrial ROS induction. Importantly the mitochondrial selectivity exhibited by elesclomol is a distinct characteristic of the compound that is not shared by other chelators, including disulfiram. Together these findings highlight a unique mechanism of action with important implications for cancer therapy. & 2012 Elsevier Inc. All rights reserved.

Keywords: Elesclomol Copper chelation Mitochondria Reactive oxygen species Cancer therapy Free radicals

Introduction Mitochondria play a dominant role in cellular energy production and are also crucial regulators of apoptosis [1,2]. To control the activation of apoptotic effector mechanisms, cancer cell mitochondria are structurally and functionally altered and differ from those of normal cells [3–5]. Altered redox status and increased reactive oxygen species (ROS) generation are also commonly observed in cancer cell mitochondria [6,7]. Disrupting redox homeostasis and exceeding the threshold compatible with cellular survival is a promising approach for cancer therapy [8], particularly because normal cells are less sensitive to agents that induce oxidative stress because of a lower level of basal ROS production and higher antioxidant capacity. Indeed, accumulating evidence suggests that this biochemical property of tumor cells

Abbreviations: ROS, reactive oxygen species; LDH, lactate dehydrogenase; BCS, bathocuproinedisulfonic acid; DLC, delocalized lipophilic cations; DSF, disulfiram; ATTM, ammonium tetrathiomolybdate; Dp44mT, di-2-pyridylketone 4,4,dimethyl-3-thiosemicarbazone; ETC, electron transport chain; PBMC, peripheral blood mononuclear cell n Corresponding author. Fax: þ1 781 274 8228. E-mail address: [email protected] (Y. Wada). 0891-5849/$ - see front matter & 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.freeradbiomed.2012.03.017

can be exploited for therapeutic benefit by pharmacologic ROS insults [9–11]. Elesclomol is a novel drug candidate that exerts anticancer activity through the elevation of ROS levels leading to subsequent activation of apoptosis [12]. Elesclomol has been evaluated in a number of clinical trials [13,14], and combined data from three randomized phase 2 and 3 trials have shown that therapeutic activity, including prolonged progression-free survival time, is demonstrated in patients with low levels of lactate dehydrogenase (LDH) at baseline [15]. Despite this efficacy, however, the molecular mechanisms by which elesclomol generates ROS and how its clinical benefit is related to the baseline LDH plasma level are currently not defined. Clearly an increased molecular understanding of the mechanism of action of elesclomol will contribute to its continued development and clinical applicability. In this study we show that elesclomol-induced ROS generation is dependent on the chelation and redox cycling of copper within mitochondria. Elesclomol chelates extracellular copper, which in turn facilitates its uptake into cells as elesclomol–Cu(II). Upon entering the cell, elesclomol–Cu(II) rapidly transports copper to mitochondria. Our results support the hypothesis that elesclomol generates mitochondrial ROS by redox reduction of Cu(II) to Cu(I) and that this process is necessary for its anticancer activity.

M. Nagai et al. / Free Radical Biology and Medicine 52 (2012) 2142–2150

After dissociation from copper, elesclomol is rapidly effluxed from cells and continues shuttling elesclomol–copper complexes into the cell. These findings provide a biological framework for understanding the unique mode of action of elesclomol. This novel mechanism, which involves selective induction of oxidative stress in cancer cell mitochondria, represents an approach distinct from that of chemotherapy or kinase inhibition for therapeutic intervention in human malignancies.

Materials and methods Cell lines and reagents Cell lines were obtained from the American Type Culture Collection (Rockville, MD, USA). MDA-MB435 melanoma and HL-60 leukemia cell lines were maintained in 10% fetal bovine serum (FBS)-supplemented Dulbecco’s modified Eagle’s medium (DMEM) and RPMI 1640, respectively, at 37 1C in 5% (v/v) CO2. PBMCs were isolated from whole blood collected from volunteers at Synta Pharmaceuticals in compliance with the Declaration of Helsinki Protocols. All reagents, including the copper chelators disulfarim (DSF) and ammonium tetrathiomolybdate (ATTM), were purchased from Sigma–Aldrich (St. Louis, MO, USA) unless otherwise noted. Elesclomol, elesclomol–Cu, di-2-pyridylketone 4,4,-dimethyl-3-thiosemicarbazone (Dp44mT), and the bis(thiosemicarbazonato) compounds ATSM and GTSM were synthesized by Synta Pharmaceuticals Corp.

2143

Copper assays Cell pellets were prepared and weighed after three washes in phosphate-buffered saline (PBS). Inductively coupled plasma mass spectometry (ICP-MS) analysis of cellular copper levels was performed by Maxxam Analytics (Burnaby, BC, Canada). In the tracer experiments, HL-60 cells were enriched in medium with 4 mM 63CuCl2 for 120 h. 63Cu-enriched HL-60 cells (2  107) were then incubated with 100 nM 65Cu–elesclomol in medium with no supplemental Cu or in the presence of 65CuCl2 (as a control). Copper levels were also determined by a photometric assay using the Cu(I)-selective indicator bicinchoninic acid (BCA) [16] with a minor modification. Briefly, copper was extracted in 333 mM hydrochloric acid before deproteinization with trichloroacetic acid. This method was validated by the demonstration of Cu levels equivalent to those detected by ICP–MS (Supplementary Fig. 1B). Total copper content was measured in the presence of reducing agent (ascorbic acid), whereas spontaneously released Cu(I) in isolated mitochondria was determined in the absence of the reducing reagent.

Electrophoresis assay Elesclomol–Cu, phenol red, and rhodamine-123 were dissolved at 10 mM in dimethyl sulfoxide (DMSO), and 5 ml of each compound was spotted onto a cellulose chromatography paper (Whatman 3017-915) and electrophoresed horizontally at 770 mV/cm in a Tris-acetate buffer (40 mM Trizma base, 20 mM sodium acetate, pH 8.2).

Cell viability assays Measurement of ROS induction To investigate the effects of metals on elesclomol activity, MDA-MB435 cells were seeded overnight in 10% FBS-containing DMEM at 2  103 cells/well in 96-well plates. Cells were then treated with graded concentrations of elesclomol for 1 h in DMEM plus 10% FBS or serum-free DMEM containing 1% bovine serum albumin(BSA) supplemented with 2 mM Cu2 þ , Fe2 þ , Mn2 þ , or Zn2 þ . Cell viability was assessed after 48 h using the CellTiter-Glo ATP assay (Promega, Madison, WI, USA). MDA-MB435 cells were also continuously treated with increasing concentrations of elesclomol in the presence or absence of 200 mM bathocuproinedisulfonic acid (BCS) and viability was assessed at 48 h. For PBMCs and HL-60 cells, 6  104 cells were seeded into 96-well plates and treated with graded concentrations of elesclomol or elesclomol– Cu for 24 h. For comparative kinetics, HL-60 cells were exposed to 200 nM elesclomol–Cu or DSF–Cu over a 48-h time course.

For superoxide measurement in intact cells, PBMCs obtained from two independent donors and HL-60 cells were incubated in the presence 100 nM elesclomol or control (DMSO) for 3 h. Cells were washed and stained with 5 mM MitoSOX red superoxide indicator (Invitrogen, Eugene, OR, USA) in Hanks’ buffer at 37 1C for 10 min. Oxidized ethidium fluorescence was then analyzed by FACS analysis. Isolated mitochondria were washed once in buffer (210 mM mannitol, 70 mM sucrose, 2.5 mM KH2PO4, 2 mM MgCl2, 1 mM BAPTA, 2 mM Hepes, pH 7.4, and 0.1% fatty acid-free BSA) and resuspended at 1 mg wet wt/ml. The effects of Cu complexes on mitochondrial ROS production were examined in the presence of 12.5 mM each glutamate, malate, and pyruvate; 25 mM succinate; 1.65 mM ADP, and 1 mM MitoSOX red. ROS-dependent fluorescence was excited at 535 nm and read at 585 nm.

Subcellular fractionation and mitochondrial isolation Redox potential analysis Cells were lysed in a hypotonic buffer (10 mM Hepes, pH 7.0, 0.05% digitonin). The lysate was divided into two halves and tonicity was adjusted with 0.5 M sucrose plus or minus 90% Percoll (GE Healthcare, Piscataway, NJ, USA). To obtain the cytosolic fraction, supernatant was collected after centrifugation of the isotonic lysate (without Percoll) at 12,000g for 10 min. Percoll density centrifugation was used to separate the nuclear and mitochondrial fractions in the remaining lysate using a 5-mmpore centrifugal filter (Ultrafree-MC; Millipore, Billerica, MA, USA). After centrifugation at 700g for 2 min, nuclei and mitochondria were collected from the upper filter cup and flowthrough, respectively. Purity of the isolated mitochondrial fraction was confirmed using rhodamine-123 fluorescence (Supplementary Fig. 1A).

The redox potentials of copper complexes with elesclomol, its analogs, or disulfiram were all determined by cyclic voltammetry. Cyclic voltammograms were recorded using an EC Epsilon electrochemical workstation from Bioanalytical Systems (West Lafayette, IN, USA) with platinum working electrode, Ag/Ag þ reference electrode, and platinum wire auxiliary electrode. The reference electrode was an Ag/Ag þ reference with a fill solution of DMSO containing 0.01 M AgNO3 and 0.1 M tetraethylammonium perchlorate. A copper complex sample (5.0 mg) was dissolved in a solution of 0.1 M tetra(n-butyl)ammonium hexafluorophosphate in DMSO (10 ml). The solution was added to the cell vial and purged with nitrogen for 10 min. The cyclic voltammetry was performed at 50 mV/s scan rate.

2144

M. Nagai et al. / Free Radical Biology and Medicine 52 (2012) 2142–2150

Fig. 1. Copper-dependent cytotoxicity and cellular uptake by elesclomol. (A) Chemical structure of elesclomol. The native nonplanar conformation of elesclomol (top) and the flat-planar conformation of the elesclomol–Cu complex (bottom) are shown. (B) MDA-MB435 melanoma cells were washed with PBS to remove metal ions in serum and the medium was replaced with serum-free DMEM supplemented with 1% BSA. The cells were then treated with graded concentrations of elesclomol for 1 h with or without supplementation of metal ions as indicated. After treatment, the compounds were removed and cell viability was determined by measuring cellular ATP at 48 h. Data are presented as means7 SD (N ¼4). (C) MDA-MB435 cells were treated with increasing concentrations of elesclomol in 10% FBS-supplemented DMEM in the absence or presence of 200 mM BCS. Viability was determined by measuring cellular ATP at 48 h. Data are presented as means7 SD (N ¼ 4). (D) MDA-MB435 cells were incubated with 100 nM elesclomol for 3 h in the absence or presence of 200 mM BCS. Cellular uptake of copper was determined by ICP–MS. (E) HL-60 cells were treated with graded concentrations of elesclomol or elesclomol–Cu complex. Cell viability was determined by measuring cellular ATP at 24 h. Data are presented as means 7SD (N¼ 4). (F) HL-60 cells were treated with graded concentrations of elesclomol or elesclomol–Cu complex and cellular copper levels determined by ICP–MS. Data are presented as means7 SD (N ¼3).

Liquid chromatography/tandem mass spectrometry (LC/MS/MS) analysis Samples were analyzed with an Agilent 1100 HPLC (Santa Clara, CA, USA) interfaced to an API 4000 tandem mass spectrometer (Applied Biosystems, Foster City, CA, USA) using a SymmetryShield RP18 column (5 mm, 2.1  100 mm; Waters, Milford,

MA, USA) at a flow rate of 0.5 ml/min. Mobile phase consisted of 0.1% formic acid in water (A) and 0.1% formic acid in methanol (B). Total run time was 9 min with the gradient elution. Detection was achieved with turbo ion spray ionization under the positiveion mode by multiple reaction monitoring. All standard curves were fit to a linear equation with a weighing factor of 1/(concentration)2 to calculate elesclomol concentrations.

M. Nagai et al. / Free Radical Biology and Medicine 52 (2012) 2142–2150

2145

Results Copper is essential for the cytotoxic activity of elesclomol It was recently reported that elesclomol (Fig. 1A) readily forms a copper chelate at a 1:1 M ratio in competition assays [17] and preferentially over the other major transition metal ions iron, manganese, and zinc. Elesclomol is potently cytotoxic to cancer cells even with brief exposure in serum-containing culture medium [12]. To examine whether copper was required for the activity, MDA-MB435 melanoma cells were treated with elesclomol for 1 h in 10% FBS-containing medium or 1% BSA–Cufree culture medium in the presence or absence of various metal ions, and cytotoxicity was evaluated after 2 days incubation without elesclomol in 10% FBS medium (Fig. 1B). Under Cudepleted conditions elesclomol did not demonstrate any significant cytotoxic activity. Supplementing the 1% BSA medium with Cu ions, but not those of iron, manganese, or zinc, restored the cell killing activity of elesclomol. Indeed cell viability was decreased to levels comparable to that seen with elesclomol treatment in 10% FBS medium. Thus copper is essential for the activity of elesclomol. To examine the effect of decreasing accessible copper in the medium, elesclomol activity was evaluated using MDA-MB435 cells in 10% FBS medium in the presence and absence of the membrane-impermeative copper-specific chelator BCS. Addition of BCS significantly inhibited cell killing, indicating that formation of Cu chelate was a rate-limiting step for the activity of elesclomol (Fig. 1C). Copper uptake studies, analyzed by ICP–MS, further showed that elesclomol promoted significant transport of copper into the cell, and this effect was abrogated in the presence of BCS (Fig. 1D). We next compared the activity of elesclomol with a preformed elesclomol–Cu complex. The preformed elesclomol–Cu complex and elesclomol were similarly active in killing HL-60 promyelocytic leukemic cells (Fig. 1E) and MDA-MB435 cells (Supplementary Fig. 2), with IC50 values below 100 nM. As shown for HL-60 cells the cytotoxic activity of the preformed elesclomol–Cu complex increased in a dose-dependent manner up to 1 mM, whereas the activity of free elesclomol saturated at Z100 nM. When the respective intracellular copper levels were compared, it was found that copper levels resulting from elesclomol–Cu complex treatment increased in a linear fashion (Fig. 1F). In contrast, and consistent with the cell killing activity, copper levels induced by addition of free elesclomol reached a plateau at 100 nM. This saturation of intracellular copper levels by free elesclomol is probably due to insufficient copper availability in 10% FBS for the formation of chelate at higher concentrations. Taken together, these results indicate that the potent anticancer activity of elesclomol is dependent on obtaining copper from the culture medium to form an active Cu chelate for entry into cells. Copper accumulation is facilitated by elesclomol–copper complex shuttling The comparative cellular concentrations of elesclomol and copper were analyzed in HL-60 cells. Copper levels increased after elesclomol treatment as a function of time (Fig. 2A), whereas intracellular levels of elesclomol remained consistently low ( r0.6 pmol/mg). Elesclomol, but not elesclomol–Cu, is a p-glycoprotein pump substrate, suggesting that elesclomol may be readily effluxed from cells upon dissociation from copper. To explore this hypothesis, HL-60 cells were washed after a 3-h elesclomol treatment and the retention of copper and elesclomol was compared (Fig. 2B). Copper levels remained high, in direct contrast to the rapid elimination of elesclomol within 1 h. This

Fig. 2. Copper accumulation results from elesclomol–copper complex shuttling over time. (A) HL-60 cells were treated with 100 nM elesclomol and cellular concentrations of copper and elesclomol were measured at the indicated time points. Copper and elesclomol levels were determined by photometry using BCA and LC/MS/MS, respectively, as described under Materials and methods. (B) HL-60 cells were treated with 100 nM elesclomol for 3 h before washout. Intracellular copper and elesclomol concentrations were determined as for (A) at the time of 3 h and at 1 h post-washout. (C) 65Cu was complexed with elesclomol as a tracer. 63 Cu-enriched HL-60 cells were treated with 100 nM elesclomol–65Cu for 0, 0.5, 3, and 9 h in 10% FBS-containing medium. Cellular 65Cu and 63Cu levels were measured by ICP–MS.

result suggests that, after cellular uptake, copper is dissociated from the elesclomol–Cu complex and retained within the cell and unbound elesclomol is rapidly effluxed.

2146

M. Nagai et al. / Free Radical Biology and Medicine 52 (2012) 2142–2150

By extension, we investigated whether effluxed elesclomol could chelate and transport additional extracellular copper into cells, thereby establishing a shuttling mechanism to account for the observed copper accumulation. To do this elesclomol was preformed into a complex with 65Cu. HL-60 cells previously enriched with 63Cu were treated with 100 nM elesclomol–65Cu or free 65Cu (not complexed to elesclomol) for 0, 0.5, 3, and 9 h in 10% FBS medium, and cellular 63Cu and 65Cu levels were measured by ICP–MS. After a 0.5-h exposure to elesclomol–65Cu, 65Cu, but not 63Cu, increased in cells, indicating that this copper was derived from the preformed elesclomol–65Cu complex (Fig. 2C). However, at 3 and 9 h, cellular 63Cu levels also increased. This strongly suggests that elesclomol exits cells after releasing 65Cu, chelates 63Cu in the culture medium, and then continues shuttling elesclomol–copper complexes into the cell. In total, elesclomol treatment increased cellular copper levels over 60-fold from baseline to 9 h; this increase represents over 300% of the total elesclomol (at molar ratio) added to the culture. Therefore copper accumulation seems reliant on the continuous shuttling of elesclomol–copper complexes between the extracellular and the cellular compartments. Control treatment with 100 nM free 65Cu did not alter cellular copper levels at any time point (data not shown).

marked increase in cellular copper levels in elesclomol-treated cells, no significant increases were seen with the other agents at this concentration. Thus rapid copper accumulation by elesclomol is a distinctive characteristic of this compound not shared by these other chelators. It was recently reported that DSF induced ROS and apoptosis in tumor cells in a copper-dependent manner [20]. Indeed preformed DSF–Cu chelate increased intracellular copper after a 9-h incubation (Supplementary Fig. 3). However, subcellular fractionation revealed a distinct copper distribution between elesclomol–Cu and DSF–Cu (Fig. 3B). In elesclomol–Cu-treated HL-60 cells, marked copper accumulation was observed primarily within the mitochondrial fraction. In contrast, the majority of copper was retained in the cytosolic fraction in DSF–Cu-treated cells. To examine the effects of these divergent copper distributions on cell viability, we then compared the kinetics of cell killing with each compound. As shown in Fig. 3C, DSF–Cu treatment resulted in a significantly delayed induction of HL-60 cell death and with weaker potency than elesclomol–Cu. Together these data highlight the novel selectivity of elesclomol–Cu for mitochondria and support a putative role for mitochondrial apoptosis to account for the rapid cytotoxic activity of elesclomol. Elesclomol–Cu is electrophoretically neutral

Unique mitochondrial selectivity of elesclomol–Cu A number of metal chelators are known to induce growth inhibition and apoptosis in cancer cells [18,19]. HL-60 cells were treated with a panel of copper chelators (DSF, Dp44mT, ATTM, ATSM, and GTSM) at 100 nM for 3 h and their capacity to chelate extracellular copper with consequent intracellular accumulation was compared to that of elesclomol (Fig. 3A). In contrast to the

Delocalized lipophilic cations (DLCs) are a class of compounds that can selectively target the mitochondria by virtue of their net positive charge. X-ray crystallography indicated that elesclomol released two protons from the N3 and N7 positions upon chelate formation with the Cu2 þ ion (data not shown), suggesting that the complex is neutral. Electrophoresis showed that elesclomol– Cu remained at its spot of origin, confirming the net neutral

Fig. 3. Elesclomol–Cu selectively accumulates in mitochondria and rapidly induces cell death, yet is electrophoretically neutral. (A) HL-60 cells were treated with 100 nM concentrations of the indicated copper chelators for 3 h and cellular copper levels determined by BCA assay. (B) HL-60 cells were treated with 100 nM elesclomol–Cu or DSF–Cu for 9 h. Cells were fractionated into cytosolic, nuclear, and mitochondrial fractions and the subcellular copper content was determined by BCA assay. Data are presented as means 7SD (N ¼3). (C) HL-60 cells were treated with 200 nM elesclomol–Cu or DSF–Cu over a 48-h time course and cell viability was assessed by measuring cellular ATP levels. (D) Elesclomol–Cu was electrophoresed as described under Materials and methods. Anionic phenol red and cationic rhodamine-123 were included as reference compounds.

M. Nagai et al. / Free Radical Biology and Medicine 52 (2012) 2142–2150

2147

Fig. 4. Elesclomol–Cu, but not DSF–Cu or CuCl2, induces mitochondrial Cu(I) release and ROS generation. (A) Isolated mitochondria from HL-60 cells were treated with the indicated concentrations of elesclomol–Cu, DSF–Cu, CuCl2, or DMSO (control) for 30 min and Cu(I) levels measured using the BCA assay in the absence of reducing agent. Data are presented as means7 SD (N ¼ 3). (B) Isolated HL-60 mitochondria were treated as for (A) at the indicated concentrations and ROS production was measured with MitoSOX red. Data are presented as means 7SD (N ¼ 3).

charge of elesclomol–Cu (Fig. 3D). Therefore the mechanism of targeted mitochondrial specificity by elesclomol–Cu is distinct from that used by DLCs. Brief exposure to elesclomol results in redox reduction of copper and the generation of mitochondrial ROS Copper is capable of redox cycling and may play a direct role in the induction of oxidative stress as a consequence of electron exchange inside mitochondria. To examine the hypothesis that elesclomol generates ROS by redox cycling of Cu(II) to Cu(I), we first measured the appearance of Cu(I) in isolated mitochondria from HL-60 cells after a brief exposure (30 min) to elesclomol–Cu, DSF–Cu, and CuCl2 in 1% BSA buffer. Cu(I) levels in isolated mitochondria were selectively measured using BCA without reducing agent. Elesclomol–Cu increased mitochondrial Cu(I) in a concentration-dependent manner with significant increases at concentrations of 0.2 mM and higher (p o0.05; Fig. 4A). No changes in Cu(I) levels relative to control were observed with DSF–Cu or free Cu(II) (from CuCl2). We next measured mitochondrial ROS levels in the same cell-free assay system. As shown in Fig. 4B, efficient generation of ROS was observed after treatment with elesclomol–Cu but not DSF–Cu or free Cu(II), in accordance with both the copper redox reduction and the cytotoxicity data. Together these findings suggest that the increased Cu(I) in mitochondria induces mitochondrial ROS, which ultimately results in cell death. Optimum redox potential is required for elesclomol activity The increase in mitochondrial Cu(I) levels by elesclomol–Cu suggests that redox cycling is occurring and that it is responsible for the ROS generation. To investigate the correlation between redox cycling ability and anticancer activity, we assessed the redox potential of elesclomol–Cu and Cu chelates of elesclomol analogs. The redox potential of elesclomol–Cu is 333 mV. Importantly, only compounds with identified redox potentials between 50 and 400 mV increased mitochondrial Cu(I) levels (Fig. 5A, Supplementary Table 1), indicating an optimum redox potential associated with this class of compounds. In contrast, the redox potentials of inactive analogs of elesclomol were either above  50 or below 400 mV, despite the fact that they all stably formed Cu chelates. Moreover, cytotoxic activity also appeared dependent on redox cycling, as only compounds within the 50 to  400 mV range of potential elicited significant cell killing in HL-60 cells (Supplementary Table 1). This result was

confirmed in the MDA-MB435 line, in which all analogs of elesclomol with IC50 values below 300 nM also possessed redox potentials ranging from  50 to  400 mV (Fig. 5B). Collectively these data show that the unique physicochemical properties of elesclomol facilitate redox reduction and the generation of Cu(I) in mitochondria and that this process is associated with potent anticancer activity. Functional selectivity of elesclomol in cancer, but not normal, cells To date over 800 patients have received elesclomol in clinical trials, and no elesclomol-related organ toxicity has been observed [14]. In addition, although elesclomol potently kills almost all types of cancer cells in vitro, with IC50 values below 100 nM, no cytotoxicity is seen in normal PBMCs (IC50 410 mM; data not shown). To investigate the basis of this selectivity, we analyzed cellular elesclomol and copper levels in human PBMCs treated with elesclomol and compared them to the HL-60 leukemic cell line. Despite efficient cellular uptake of elesclomol in both cell types, copper accumulation was observed only in the HL-60 cells (Fig. 6A). We next examined the functional consequences of this difference by examining oxidative stress induction in each of the cell types. First, PBMCs (from two independent donors) and HL-60 cells were treated with 100 nM elesclomol for 3 h and superoxide generation was determined using a fluorescent probe for ROS induction. As shown in Fig. 6B, ROS were readily observed in the leukemic cell line but not in the PBMCs. Second, we compared ROS generation in isolated mitochondria, the functional integrity of which was confirmed by measuring ATP generation (Supplementary Fig. 4). As expected, the basal level of ROS was significantly lower in isolated mitochondria from all donor PBMCs compared to that found for mitochondria from HL-60 (Fig. 6C). In contrast to the increased ROS levels seen in the mitochondrial fraction of HL-60 cells after a 30-min treatment, elesclomol–Cu did not induce ROS in the PBMC-derived mitochondria. These results support the model that elesclomol–Cu selectively delivers copper to cancer cell mitochondria, which ultimately produces critical elevations in oxidative stress.

Discussion Elesclomol offers significant potential as a novel therapeutic for cancer treatment. It has been shown that its potent anticancer activity results from elevation of ROS levels leading to the initiation of apoptosis [12]. We previously carried out experiments using

2148

M. Nagai et al. / Free Radical Biology and Medicine 52 (2012) 2142–2150

Fig. 5. Optimum redox potential is required for mitochondrial Cu(I) release and cytotoxic activity of elesclomol and its analogs. (A) Elesclomol–Cu and related analogs are active when their redox potentials are between  400 and  50 mV. Mitochondria were isolated from HL-60 cells and treated with Cu complexes of elesclomol and its analogs for 30 min. Cu(I) levels were measured using the BCA assay in the absence of reducing agent. (B) Plot showing relationship between IC50 cytotoxicity values of Cu complexes of elesclomol and its analogs in the MDA-MB435 melanoma cell line with redox potential.

isolated mitochondria in the absence of copper, thus hindering identification of the essential effects of this ion [12]. Here we have deciphered the mechanism by which elesclomol induces ROS in cancer cells, a unique mode of action that is dependent on its copper-chelating ability. Elesclomol readily chelates copper in serum to form a highly cell permeative and active elesclomol–Cu entity. After cellular uptake, the elesclomol–Cu complex rapidly and selectively targets the mitochondria. Within this organelle, redox reduction of Cu(II) to Cu(I) occurs, resulting in the generation of oxidative stress. Importantly, efflux of elesclomol after dissociation from the complex facilitates the repeated shuttling of elesclomol–Cu complexes to promote copper accumulation and ROS generation sufficient to cause cell death. Many metal chelators and metal ions themselves have been shown to generate ROS in cells and demonstrate cytotoxicity [18,21]. For example, the copper chelator DSF can induce ROS and activate apoptosis in melanoma cells in a copper-dependent manner [20] and is currently undergoing clinical evaluation in a number of malignancies [21]. Although elesclomol and elesclomol– Cu each promote similar levels of cellular copper accumulation, increased copper levels were observed only with preformed DSF–Cu,

and not free DSF, indicating that the capacity of DSF to chelate copper in serum is much lower than that of elesclomol. Further, the majority of this copper in DSF–Cu-treated cells remained in the cytosol. Selective mitochondrial copper uptake therefore seems to be an exclusive characteristic of elesclomol and is sufficient to explain its higher relative potency. The use of mitochondrial-targeted compounds has provided new experimental strategies for cancer therapy [22]. DLCs have been developed to selectively target this organelle because of their positive charge [23]. The Isatin–Schiff base copper(II) complex Cu(isaepy)2 can also act in a manner similar to that of DLCs to induce mitochondrial oxidative damage and consequent apoptosis [24]. However, the proapoptotic activity of Cu(isaepy)2 is weak [25]. In contrast, elesclomol–Cu is electrophoretically neutral and demonstrates strong anticancer activity and mitochondrial selectivity, suggesting that the mechanism of the targeted specificity for mitochondria by elesclomol–Cu is distinct from that of DLCs. In isolated mitochondria, ROS induction accompanied the reduction of Cu(II) to Cu(I), suggesting that transport by elesclomol–Cu allows for the transition between copper oxidation states. In addition,

M. Nagai et al. / Free Radical Biology and Medicine 52 (2012) 2142–2150

Fig. 6. Elesclomol selectively induces copper accumulation, oxidative stress, and mitochondrial ROS generation in HL-60 leukemic cells but not in PBMCs. (A) PBMCs and HL-60 cells were incubated with 100 nM elesclomol for 3 h before measurement of cellular elesclomol and copper levels. The y-axis units are in femtomoles and picomoles. Control (vehicle-treated) PBMC and HL-60 cellular copper content values were 14.8 and 34.6 pmol/106 cells, respectively. (B) PBMCs and HL-60 cells were treated for 3 h with 100 nM elesclomol or DMSO (control). At the end of the incubation, the cells were stained with MitoSOX red and ROS levels assessed by measuring hydroxyethidium by flow cytometry. PBMC data are presented as a representative result obtained from two independent donors. (C) Isolated mitochondria from human PBMCs and HL-60 cells were treated with the indicated concentrations of elesclomol for 30 min in a MitoSOX red buffer. Data are presented as means7SD from three independent donors (PBMCs) or three individual experiments (HL-60).

2149

the anticancer activity of elesclomol and its analogs is highly dependent on the redox potential of their copper complexes, and only active compounds with redox potentials between  50 and  400 mV lead to reduction of Cu(II) to Cu(I) in mitochondria. Interestingly, this range of redox potential closely matches the potential drops along the ETC [26], suggestive of an association between electrons from an actively working ETC and the redox cycling of copper. This finding is in complete agreement with our recent identification of mitochondrial electron transport as a putative cellular target of elesclomol using a powerful yeast genomics approach [27]. For any cancer therapeutic the ability to selectively target tumors by exploiting fundamental differences between normal and cancer cells represents a promising approach. No elesclomolrelated organ toxicity has been observed in clinical trials, unlike most copper-based antitumor agents that show relatively high toxicity [28]. In the comparative experiments using the myeloid tumor line and PBMCs copper accumulation, mitochondrial ROS generation, and subsequent cellular oxidative stress induction did not occur in the nontransformed cells despite similar cellular uptake of elesclomol in both cell types. Because copper accumulation in cancer cells results from elesclomol–Cu complex shuttling, which requires the reduction of Cu(II) to Cu(I), it is reasonable to suggest that the lack of copper accumulation in PBMCs may be linked to a lower redox cycling capacity. Also, because cancer cells typically exhibit higher ROS and lower antioxidant reservoirs in comparison to normal cells [6,29] even a minor augmentation of ROS induction in tumor mitochondria might be sufficient to push total ROS levels beyond a critical threshold that results in apoptosis. Although the underlying molecular basis for the cancer cell-selective cytotoxicity of elesclomol remains to be defined, the findings we are presenting here suggest that redox cycling-dependent mitochondrial ROS generation may be an effective mechanism to provide cancer cell selectivity. The new mechanistic insights into elesclomol activity presented here have important implications for clinical observations seen with elesclomol by providing a rationale for a potential molecular predictor of response identified through clinical trials. In a phase 3 trial of elesclomol in combination with paclitaxel in metastatic melanoma, the primary endpoint of progression-free survival was achieved in 68% of patients exhibiting normal LDH levels with a significant improvement in median PFS, whereas there was an increase in mortality in the elevated LDH population [15]. There are two genes that encode LDH (LDHA and LDHB) [30]. LDHA, which favors the conversion of pyruvate to lactate, is increased in tumors reliant on glycolysis for energy production instead of mitochondrial respiration through the ETC. This high LDHA in tumors, in turn, results in high LDH levels detected in the bloodstream. The clinical benefit selectively observed with normal-LDH patients implies that tumors with high mitochondrial respiration respond to elesclomol treatment. Here we show a relationship between redox potential of elesclomol–Cu chelates and Cu(I) release/ROS generation in mitochondria. Together with the proposed mechanism targeting mitochondrial electron transport in yeast [27], the LDH dependency observed in clinical studies may be explained by the requirement of active ETC for optimal activity of elesclomol–Cu to induce mitochondrial apoptosis in human tumors. In summary, this study describes the unique mode of action of elesclomol in directly targeting cancer cell mitochondria and elevating reactive oxygen species. Elesclomol induces ROS by redox reduction of Cu(II) to Cu(I) within the mitochondria and this process is required for its anticancer activity. Generating mitochondrial oxidative stress by this mechanism provides a compelling new strategy to selectively target cancer cells for therapeutic intervention in a variety of human malignancies.

2150

M. Nagai et al. / Free Radical Biology and Medicine 52 (2012) 2142–2150

Acknowledgments The authors would like to thank Drs. Richard Bates, Pat Rao and David Proia for their valuable contributions to this manuscript. All work was funded by Synta Pharmaceuticals Corp. All authors are current or former employees of Synta Pharmaceuticals Corp.

Appendix A. Supplementary Information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.free radbiomed.2012.03.017.

References [1] Huttemann, M.; Pecina, P.; Rainbolt, M.; Sanderson, T. H.; Kagan, V. E.; Samavati, L.; Doan, J. W.; Lee, I. The multiple functions of cytochrome c and their regulation in life and death decisions of the mammalian cell: from respiration to apoptosis. Mitochondrion 11:369–381; 2011. [2] Ulivieri, C. Cell death: insights into the ultrastructure of mitochondria. Tissue Cell 42:339–347; 2010. [3] Solaini, G.; Sgarbi, G.; Baracca, A. Oxidative phosphorylation in cancer cells. Biochim. Biophys. Acta 1807:534–542; 2011. [4] Fulda, S.; Galluzzi, L.; Kroemer, G. Targeting mitochondria for cancer therapy. Nat. Rev. Drug Discovery 9:447–464; 2010. [5] Gogvadze, V.; Orrenius, S.; Zhivotovsky, B. Mitochondria in cancer cells: what is so special about them? Trends Cell Biol. 18:165–173; 2008. [6] Trachootham, D.; Alexandre, J.; Huang, P. Targeting cancer cells by ROSmediated mechanisms: a radical therapeutic approach? Nat. Rev. Drug Discovery 8:579–591; 2009. [7] Hervouet, E.; Simonnet, H.; Godinot, C. Mitochondria and reactive oxygen species in renal cancer. Biochimie 89:1080–1088; 2007. [8] Fruehauf, J. P.; Meyskens Jr. F. L. Reactive oxygen species: a breath of life or death? Clin. Cancer Res. 13:789–794; 2007. [9] Cabello, C. M.; Bair 3rd W. B.; Wondrak, G. T. Experimental therapeutics: targeting the redox Achilles heel of cancer. Curr. Opinion Invest. Drugs 8:1022–1037; 2007. [10] Raj, L.; Ide, T.; Gurkar, A. U.; Foley, M.; Schenone, M.; Li, X.; Tolliday, N. J.; Golub, T. R.; Carr, S. A.; Shamji, A. F.; Stern, A. M.; Mandinova, A.; Schreiber, S. L.; Lee, S. W. Selective killing of cancer cells by a small molecule targeting the stress response to ROS. Nature 475:231–234; 2011. [11] Schumacker, P. T. Reactive oxygen species in cancer cells: live by the sword, die by the sword. Cancer Cell 10:175–176; 2006. [12] Kirshner, J. R.; He, S.; Balasubramanyam, V.; Kepros, J.; Yang, C. Y.; Zhang, M.; Du, Z.; Barsoum, J.; Bertin, J. Elesclomol induces cancer cell apoptosis through oxidative stress. Mol. Cancer Ther. 7:2319–2327; 2008. [13] Berkenblit, A.; Eder Jr. J. P.; Ryan, D. P.; Seiden, M. V.; Tatsuta, N.; Sherman, M. L.; Dahl, T. A.; Dezube, B. J.; Supko, J. G. Phase I clinical trial of STA-4783 in combination with paclitaxel in patients with refractory solid tumors. Clin. Cancer Res. 13:584–590; 2007.

[14] O’Day, S.; Gonzalez, R.; Lawson, D.; Weber, R.; Hutchins, L.; Anderson, C.; Haddad, J.; Kong, S.; Williams, A.; Jacobson, E. Phase II, randomized, controlled, double-blinded trial of weekly elesclomol plus paclitaxel versus paclitaxel alone for stage IV metastatic melanoma. J. Clin. Oncol. 27: 5452–5458; 2009. [15] Vukovic, V. M.; Hauschild, A.; Eggermont, A. M.; O’Day, S. Phase III, randomized, double-blind study of elesclomol and paclitaxel versus paclitaxel alone in stage IV metastatic melanoma (MM): 1-year OS update. J. Clin. Oncol. 28(15s):8550; 2010. [Abstract]. [16] Brenner, A. J.; Harris, E. D. A quantitative test for copper using bicinchoninic acid. Anal. Biochem. 226:80–84; 1995. [17] Wu, L.; Zhou, L.; Liu, D. Q.; Vogt, F. G.; Kord, A. S. LC-MS/MS and density functional theory study of copper(II) and nickel(II) chelating complexes of elesclomol (a novel anticancer agent). J. Pharm. Biomed. Anal. 54:331–336; 2011. [18] Marzano, C.; Pellei, M.; Tisato, F.; Santini, C. Copper complexes as anticancer agents. Anticancer Agents Med. Chem. 9:185–211; 2009. [19] Easmon, J. Copper and iron complexes with antitumor activity. Expert Opinion Ther. Pat. 12:798–818; 2002. [20] Morrison, B. W.; Doudican, N. A.; Patel, K. R.; Orlow, S. J. Disulfiram induces copper-dependent stimulation of reactive oxygen species and activation of the extrinsic apoptotic pathway in melanoma. Melanoma Res. 20:11–20; 2010. [21] Cvek, B. Targeting malignancies with disulfiram (Antabuse): multidrug resistance, angiogenesis, and proteasome. Curr. Cancer Drug Targets 11: 332–337; 2011. [22] Pathania, D.; Millard, M.; Neamati, N. Opportunities in discovery and delivery of anticancer drugs targeting mitochondria and cancer cell metabolism. Adv. Drug Delivery Rev. 61:1250–1275; 2009. [23] Kurtoglu, M.; Lampidis, T. J. From delocalized lipophilic cations to hypoxia: blocking tumor cell mitochondrial function leads to therapeutic gain with glycolytic inhibitors. Mol. Nutr. Food Res. 53:68–75; 2009. [24] Filomeni, G.; Piccirillo, S.; Graziani, I.; Cardaci, S.; Da Costa Ferreira, A. M.; Rotilio, G.; Ciriolo, M. R. The Isatin–Schiff base copper(II) complex Cu(isaepy)2 acts as delocalized lipophilic cation, yields widespread mitochondrial oxidative damage and induces AMP-activated protein kinasedependent apoptosis. Carcinogenesis 30:1115–1124; 2009. [25] Filomeni, G.; Cerchiaro, G.; Da Costa Ferreira, A. M.; De Martino, A.; Pedersen, J. Z.; Rotilio, G.; Ciriolo, M. R. Pro-apoptotic activity of novel Isatin–Schiff base copper(II) complexes depends on oxidative stress induction and organelleselective damage. J. Biol. Chem. 282:12010–12021; 2007. [26] Kussmaul, L.; Hirst, J. The mechanism of superoxide production by NADH:ubiquinone oxidoreductase (complex I) from bovine heart mitochondria. Proc. Natl. Acad. Sci. USA 103:7607–7612; 2006. [27] Blackman, R. K.; Cheung-Ong, K.; Gebbia, M.; Proia, D. A.; He, S.; Kepros, J.; Jonneaux, A.; Marchetti, P.; Kluza, J.; Rao, P. E.; Wada, Y.; Giaever, G.; Nislow, C. Mitochondrial electron transport is the cellular target of the oncology drug elesclomol. PLoS One 7:e29798; 2012. [28] Wang, T.; Guo, Z. Copper in medicine: homeostasis, chelation therapy and antitumor drug design. Curr. Med. Chem. 13:525–537; 2006. [29] Gupte, A.; Mumper, R. J. Elevated copper and oxidative stress in cancer cells as a target for cancer treatment. Cancer Treat. Rev. 35:32–46; 2009. [30] Adams, M. J.; Buehner, M.; Chandrasekhar, K.; Ford, G. C.; Hackert, M. L.; Liljas, A.; Rossmann, M. G.; Smiley, I. E.; Allison, W. S.; Everse, J.; Kaplan, N. O.; Taylor, S. S. Structure–function relationships in lactate dehydrogenase. Proc. Natl. Acad. Sci. USA 70:1968–1972; 1973.