JBIC Journal of Biological Inorganic Chemistry

, Volume 17, Issue 4, pp 589–598

Synchrotron radiation induced X-ray emission studies of the antioxidant mechanism of the organoselenium drug ebselen


  • Jade B. Aitken
    • School of ChemistryThe University of Sydney
    • Australian Synchrotron
    • Institute of Materials Structure Science, KEK
    • School of ChemistryThe University of Sydney
  • T. T. Hong Duong
    • Discipline of Neurosurgery, Australian School of Advanced MedicineMacquarie University
  • Roshanak Aran
    • Discipline of Pathology, Bosch Research Institute, Sydney Medical SchoolThe University of Sydney
  • Paul K. Witting
    • Discipline of Pathology, Bosch Research Institute, Sydney Medical SchoolThe University of Sydney
  • Hugh H. Harris
    • School of Chemistry and PhysicsUniversity of Adelaide
  • Barry Lai
    • X-ray Science DivisionArgonne National Laboratory
  • Stefan Vogt
    • X-ray Science DivisionArgonne National Laboratory
    • Department of Pharmacology and ToxicologyUniversity of Otago
Original Paper

DOI: 10.1007/s00775-012-0879-y

Cite this article as:
Aitken, J.B., Lay, P.A., Duong, T.T.H. et al. J Biol Inorg Chem (2012) 17: 589. doi:10.1007/s00775-012-0879-y


Synchrotron radiation induced X-ray emission (SRIXE) spectroscopy was used to map the cellular uptake of the organoselenium-based antioxidant drug ebselen using differentiated ND15 cells as a neuronal model. The cellular SRIXE spectra, acquired using a hard X-ray microprobe beam (12.8-keV), showed a large enhancement of fluorescence at the Kα line for Se (11.2-keV) following treatment with ebselen (10 μM) at time periods from 60 to 240 min. Drug uptake was quantified and ebselen was shown to induce time-dependent changes in cellular elemental content that were characteristic of oxidative stress with the efflux of K, Cl, and Ca species. The SRIXE cellular Se distribution map revealed that ebselen was predominantly localized to a discreet region of the cell which, by comparison with the K and P elemental maps, is postulated to correspond to the endoplasmic reticulum. On the basis of these findings, it is hypothesized that a major outcome of ebselen redox catalysis is the induction of cellular stress. A mechanism of action of ebselen is proposed that involves the cell responding to drug-induced stress by increasing the expression of antioxidant genes. This hypothesis is supported by the observation that ebselen also regulated the homeostasis of the transition metals Mn, Cu, Fe, and Zn, with increases in transition metal uptake paralleling known induction times for the expression of antioxidant metalloenzymes.


AntioxidantDrugEbselenOrganoseleniumSynchrotron-radiation-induced X-ray emission



Antioxidant response element


Endoplasmic reticulum


Glutathione peroxidase


Reduced glutathione


Oxidized glutathione


Reactive oxygen species


Superoxide dismutase


Synchrotron-radiation-induced X-ray emission


The organoselenium agent ebselen [2-phenyl-1,2-benzisoselenazol-3(2H)-one, originally known as PZ51] shows considerable promise as an antioxidant drug. In humans, ebselen (Fig. 1) has demonstrated neuroprotective activity in phase II clinical trials [1] and has entered phase III trials. Concurrently, a range of ebselen analogues are being evaluated as therapies for oxidative-stress-related diseases [211]. However, despite the growing evidence of the widespread utility of this class of drug, the mechanism(s) of action that determines the in vivo efficacy remains to be established.
Fig. 1

Structure of the organoselenium drug ebselen

Ebselen has long been known to demonstrate an in vitro protective effect by acting as a molecular mimic of the antioxidant enzyme glutathione peroxidase (GPx) [12]. Intracellular GPx protects the cell by catalyzing the reaction between the antioxidant thiol reduced glutathione (GSH) and peroxide-based reactive oxygen species (ROS). ROS generation is a major cause of cellular damage in disease, and GPx therefore preserves organ and tissue function under conditions of oxidative stress [13]. In a manner analogous to that of GPx, in vitro ebselen has been shown to redox-cycle and detoxify peroxides (Fig. 2). Although the drug is several orders of magnitude less active than the native enzyme, in vivo ebselen’s GPx activity may be enhanced through interactions with the thioredoxin system [14, 15]. However, as its redox cycle does not display pronounced substrate specificity for GSH, ebselen is not a high-fidelity mimic of GPx and can alternatively utilize a range of structurally diverse thiol-containing molecules [16]. One implication of the generic thiol recognition of ebselen is that, under conditions of oxidative stress, ebselen can also catalyze the oxidation of protein cysteine residues [4]. Reports have indicated that ebselen-induced protein oxidation could have multiple cellular effects, as the redox status of cysteine acts as a posttranslational regulator of a range of key cell signaling proteins, including the zinc storage protein metallothionein (modulating Zn2+ homeostasis) [17], the ryanodine receptor (modulating Ca2+ flux) [18], and a family of transcription factors (modulating protein expression) that contain zinc finger DNA binding domains [3, 19, 20].
Fig. 2

Molecular mechanism of ebselen catalysis. The glutathione peroxidase (GPx)-mimicking catalytic activity of ebselen 1 has several potential mechanisms, with the predominant pathway dependent upon cellular conditions. Ebselen is considered to act as an antioxidant by reducing peroxides to alcohols (overall R′OOH + 2R–SH → R′–OH + R–SS–R + H2O) [21]. However, in the absence of reactive oxygen species, ebselen can also act as thiol-specific oxidant to form the mixed selenylsulfide 2. Unlike the enzyme GPx, ebselen does not display pronounced substrate specificity and the thiol species (R) can be either reduced glutathione or other thiol-containing products of cellular metabolism, such as proteins containing a cysteine residue [22]. In the oxidative cycle, ebselen 1 can be directly oxidized by atom transfer from peroxide species (R′–OOH) to form selenoxide 6. This catalytic intermediate can then undergo reduction using a thiol electron donor (R–SH) to form the highly reactive selenenic acid 4 via the mixed selenylsulfide 5. Species 4 can then undergo a ring-closing step to regenerate ebselen 1. In the thiol-addition cycle, ebselen 1 can form the selenylsulfide intermediate 2 via addition of a thiol (R–SH). An excess of the thiol can then drive the equilibrium to form the catalytic selenol 3. Species 3 can then react with a peroxide species (R′–OOH) to form selenenic acid 4 and regenerate ebselen 1 via ring closure

In addition to redox cycling, the initial step in ebselen’s catalytic mechanism is its reaction with a thiol (forming intermediate 2, Fig. 2), which can proceed in the absence of ROS. Within the cell, the outcome of this reaction is dependent upon the competition between the thiol group of GSH and other cellular thiol species, including protein cysteine residues. The oxidation of protein cysteine residues to form cysteine–ebselen adducts allows ebselen to act as an indirect antioxidant by modifying the function of oxidative-stress-related enzymes. Analysis of this pathway has shown that ebselen can block the production of inflammatory mediators through covalent modification of lipoxygenase [23] and by downregulating the expression of cyclooxygenase 2 [24]. The propensity for ebselen to react with cysteine residues also explains the ability of the drug to interfere with the enzymatic production of nitric oxide and ROS via inhibition of endothelial nitric oxide synthase [25] and NADPH oxidase [26], respectively. Cysteine modification has also been proposed to result in a generic increase in the antioxidant capacity of the cell though activation of the Nrf2 transcription factor (a global inducer of the antioxidant response) as a consequence of the oxidation of cysteine residues on its negative regulator Keap1 [27].

Owing to the existence of multiple pathways by which the drug can interfere with cellular biochemistry, the most relevant in vivo mechanisms that determine the observed antioxidant properties of ebselen are still controversial. An emerging trend in the analysis of drug function is the realization that the activity of a drug must, to a large extent, be regulated by the extent to which it is partitioned among intracellular organelles [2830]. This has important implications for the proposed mechanism(s) of action of ebselen, as many of its known targets are highly compartmentalized to discrete regions of the cell. Consequently, it would greatly aid in developing an understanding of the predominant mechanism(s) of action of ebselen if the extent to which the drug is partitioned into cellular organelles were determined. Unfortunately, as ebselen is not fluorescent, drug localization is not amenable to analysis via visible fluorescence microscopic methods, and the attachment of tags or reporter groups would fundamentally change the physiochemical properties of the drug. However, ebselen does have one potentially useful spectroscopic handle in that it incorporates an organoselenium moiety within its structure. As selenium (Se) is a trace element, it is present in cultured cells in only extremely low concentrations [31]. Therefore, it was anticipated that the Se content of ebselen-treated cells would be much higher than that of cells with normal background Se levels. On the basis of this selective enhancement of the Se signal, it was anticipated that an analysis of cellular Se distribution would provide information on the intracellular localization and concentration of ebselen. Synchrotron radiation induced X-ray emission (SRIXE) spectroscopy, which has been an invaluable tool in understanding the effects of certain drugs on target cells [32, 33], was used to examine drug uptake and distribution in neuronal cells following exposure to ebselen.

Materials and methods

Cell culture

As ebselen has displayed neuroprotective properties, the hybrid dorsal root ganglia (rat)–neuroblastoma (mouse) cell line ND15 was selected to study cellular interactions with ebselen. This cell type can be induced to differentiate in order to develop features that are characteristic of neuronal morphologies with extended neurite outgrowth from the cell body. ND15 cells (ECACC, Salisbury, UK) were cultured in Dulbecco’s modified Eagle’s medium/Ham’s nutrient mixture F12 (JRH Biosciences, Lenexa, KA, USA) supplemented with 10% (v/v) fetal calf serum (Sigma, Sydney, Australia), 100 U/mL penicillin, and 100 μg/mL streptomycin, at 37 °C in a humidified atmosphere of 5% CO2. For SRIXE measurements, cells were seeded at a density of 105 cells per milliliter onto silicon nitride windows (500-nm thickness; Silson, Northampton, UK) and incubated in medium for 24 h. To induce the neuronal phenotype, the cells were then differentiated with retinoic acid (20 μM) and subcultured for a further 6 days [34]. Following differentiation, the cells were incubated with ebselen (10 μM) for 60, 120, or 240 min. The cells were then fixed in methanol/acetone at −20 °C for SRIXE analysis following a standard procedure [34, 35], which has been used previously with neuronal cells [36].

SRIXE imaging of the intracellular distribution of ebselen

Hard X-ray microprobe experiments were performed by raster scanning the cells on XOR beamlines 2-ID-D and 2-ID-E at the Advanced Photon Source (Argonne National Laboratory, Chicago, IL, USA), as described previously [37, 38]. Individual cells were located on the silicon nitride windows using a phase-contrast optical microscope (Leica DMRXE). All measurements were conducted using a 12.8-keV monochromatic X-ray incident beam using a single zone plate (2-ID-E) or a dual zone plate (2-ID-D) and an order-sorting aperture device under a helium atmosphere. The fluorescence signal was detected for either 0.5 or 1 s per spatial point at 90° to the incident beam using an UltraLEGe single-element solid-state detector (Canberra) (2-ID-E) or a Vortex-EM single-element silicon-drift detector (2-ID-D) [39]. Quantitative analysis (elemental area densities in micrograms per square centimeter) was performed using the MAPS software package [40] by fitting the full fluorescence spectrum at every single point to modified Gaussians [41], and the intensities of peaks were then compared with corresponding measurements on thin-film standards NBS-1832 and NBS-1833 from the National Institute of Standards and Technology (Gaithersburg, MD, USA). The statistical significance of changes in the elemental content of treated cells compared with controls was assessed using Student’s t test assuming unequal variances.


SRIXE analysis of untreated differentiated ND15 cells

When ND15 cells were fixed at −20 °C, their structure was preserved for X-ray analysis (Fig. 3), which allowed the identification of both the cell body and neurite morphologies. The potassium (K) map was routinely used to delineate the cell, as this element has been shown to display a strong signal-to-noise ratio at a range of energies [42]. Using a 12.8-keV beam focused to 0.3 μm, we obtained elemental maps of individual cells, which allowed the cell body, and in most cases the neurite regions, to be observed (a representative cell is shown in Fig. 3). Owing to the length of time required to map the full extent of the neurite outgrowth for some cells, neurite outgrowth was not always fully mapped, as this would have resulted in substantially larger scan areas that contained mainly background. In general, the untreated cells displayed distinct elemental maps for the highly abundant elements of interest (P, S, Cl, K, Ca, and Zn), and some transition metals were also observable under the experimental conditions. In untreated cells, the innate cellular Se content produced a very weak X-ray fluorescence signal, but no clearly definable areas of intracellular Se localization were observed before ebselen treatment. For all cells studied, the elemental content was predominantly located within the cell body, with the neurite growths displaying lower levels of all elements, e.g., in K maps.
Fig. 3

Synchrotron radiation induced X-ray emission (SRIXE) mapping of cultured ND15 neuronal cells. An optical image and representative SRIXE elemental maps for an untreated and differentiated ND15 cell are displayed. ND15 cells were seeded onto silicon nitride windows and then differentiated with retinoic acid (20 μM) before fixation. For each cell, a phase-contrast image was taken using a conventional inverted microscope, and the image was then compared with the SRIXE maps acquired using a 12.8-keV beam. The SRIXE elemental distribution maps show false-color images that correspond to the density and distribution of key elements (P, S, K, Ca, Zn, Se) within the cell. The scatter map (s_a) provides an indication of the thickness of the cell at each point

Uptake and intracellular distribution of ebselen

Analysis of the SRIXE spectra demonstrated that cells treated with ebselen displayed a pronounced Kα signal for Se at 11.2-keV (Fig. 4). This signal was very weak in the control samples, which indicated that the observed X-ray fluorescence was dominated by ebselen and/or its Se-containing metabolites, as there was no significant overlap with other fluorescence signals. Quantitative analysis of the pharmacokinetics of ebselen uptake (Table 1) revealed that the maximum Se levels were obtained 2 h after bolus addition of ebselen (10 μM). After 1 h, there was only a slight increase in Se levels when compared with the control, whereas by 4 h the intracellular Se content had diminished. The Se maps revealed that ebselen was highly compartmentalized within the cell (Fig. 5), with most of the Se signal intensity originating from a contiguous area. The regions of highest intensity in the Se map were also adjacent to regions of high intensity in the Zn and S maps. The positioning of these regions of high intensity was still present even when the data were normalized by the scatter map to account for cell thickness (Fig. 5).
Fig. 4

Cellular fluorescence spectra following ebselen treatment. Representative fluorescence spectra are displayed that were integrated over one cell, and were collected at an incident X-ray energy of 12.8 keV. The Se Kα (11.22 keV) X-ray fluorescence line is shown (red). A control cell (no ebselen), B cell treated with ebselen (10 μM) for 2 h

Table 1

Changes in elemental levels within ND15 cells following ebselen treatment



Time (h)





1.1 (4)

0.5 (3)

0.13 (5)*

0.11 (3)*


0.46 (8)

0.3 (2)

0.18 (6)*

0.18 (2)*


3 (1)

1.2 (6)

1.5 (4)

1.1 (2)*


0.36 (7)

0.14 (9)

0.04 (1)*

0.11 (1)*


0.0123 (5)

0.007 (4)

0.007 (2)*

0.16 (3)*


0.00051 (6)

0.0003 (2)

0.00014 (1)*

0.0042 (8)*


0.0053 (8)

0.004 (2)

0.0031 (5)*

0.017 (4)*


0.00018 (3)

0.00006 (2)

0.00010 (2)*

0.0026 (7)*


0.00112 (6)

0.0005 (2)

0.0004 (2)*

0.007 (1)*


0.015 (5)

0.009 (5)

0.0018 (1)*

0.009 (2)*


0.00023 (4)

0.0004 (1)

0.015 (7)*

0.007 (4)*

Differentiated ND15 cells were exposed to ebselen (10 μM) and samples were fixed between 1 and 4 h for synchrotron radiation induced X-ray emission analysis as described in “Materials and methods.” The mean relative intracellular elemental density (μg/cm2) was obtained by X-ray fluorescence microprobe studies averaged over the whole cell region. Data represent the mean (± standard deviation) from each drug-treated data set

* P < 0.05, Student’s t test compared with the control (n = 4 for all treatment conditions except for 1 h ebselen administration, where n = 2)

Fig. 5

Time course of elemental localization for K, P, and Se after ebselen administration. Optical images and synchrotron radiation induced X-ray emission elemental distribution maps for differentiated ND15 cells treated with ebselen (10 μM) for 1, 2, and 4 h (AC, respectively) are displayed. The false-color maps are indicative of cell morphology (K), the nuclear compartment (P), and ebselen distribution (Se)

Most of the ebselen and its metabolites were found in a region adjacent to the nucleus

The P maps are indicative of the nuclear region due to localization of P from DNA [32, 43]. Using the SRIXE P maps as markers, most of the ebselen and its metabolites were found within an area close to, but not overlapping, the nucleus. This pattern of Se distribution was obtained at the earliest time point, 1 h after ebselen administration, and remained constant throughout the 4-h time course. The overlay of the P, Zn, and Se maps revealed little overlap between the region of high Se concentration and the regions of maximum intensity on the P maps (Fig. 6). Therefore, the nuclear region itself has low levels of Se, which shows that ebselen and its metabolites either did not appreciably enter this compartment during the time course of our studies or were subjected to effective efflux mechanisms, or both.
Fig. 6

Representative optical images and elemental colocalization maps for ebselen-treated ND15 cells. Ebselen (10 μM) was administered to cells for 1 h (a) and 4 h (b). The false-color maps indicate the distribution of P (red), Zn (green), and Se (blue). The region of highest ebselen concentration is predominantly adjacent to the nuclear compartment, as delineated by the highest elemental concentration in the P map

SRIXE analysis of cellular elemental levels following ebselen administration

To determine the effect of ebselen administration upon cellular concentrations of elements other than Se, the corresponding integrated SRIXE intensity for the whole cell region of each map was evaluated (Table 1). At the initial 1-h time point after ebselen administration, which corresponded to low levels of Se uptake, there were no significant changes in the intracellular contents of other elements when compared with those of the control cells. At the 2-h time point, when maximum Se uptake was observed, significant decreases, approximately twofold to tenfold when compared with control cells, had occurred in the levels of all of the other elements studied (P < 0.05), with the exception of Cl, the level of which had also decreased 2.5-fold; however, this change was not significant (P > 0.05). By the 4-h time point, Se levels within the cell had already peaked following the bolus administration of ebselen, and had declined to approximately half the maximum cellular concentration. The intracellular P, S, and K contents remained low, at essentially the same levels as determined at 2 h, and the earlier indication that the Cl content was decreasing was confirmed, as a further decrease to 2.7-fold of that within the control cells attained significance (P < 0.05). Importantly, however, during this later time interval, the cellular Ca, Mn, Fe, Ni, Cu, and Zn contents demonstrated a dramatic reversal in the previously observed direction of metal trafficking. Between 2 and 4 h, the levels of all of these elements increased between fivefold and 30-fold. This resulted in cells at 4 h after ebselen treatment containing substantially higher levels of Ca, Mn, Fe, Ni, and Cu than were found in the untreated cells (threefold to 15-fold increases compared with control levels, P < 0.05). Zn levels followed this general trend, with a fivefold increase between 2 and 4 h; however, by 4 h, the cellular Zn content was still slightly below the control threshold.

In summary, the levels of the common cellular elements P, S, K, and Cl decreased following 10 μM ebselen treatment and remained low for the 4-h time course duration. The changes in the levels of the metals Ca, Mn, Fe, Ni, Cu, and Zn initially paralleled the decreases shown for the nonmetals. However, these variations appeared to be biphasic in nature, with decreases at the 2-h time point in sharp contrast to major increases seen at 4 h, with the influx of metals after 2 h resulting in much higher intracellular concentrations of transition metals in drug-treated cells compared with control cells. Overall, none of the changes in elemental content directly correlated with changes in cellular Se levels, which increased until 2 h and then decreased by 4 h. The fact that the decreases in P, K, and Cl levels were maintained over 4 h, whereas substantial increases in transition metal loading occurred between 2 to 4 h, suggested that different mechanisms were responsible for the trafficking of these elements.


Ebselen pharmacokinetics can be determined using SRIXE elemental mapping

The quantification of the cellular uptake and distribution of ebselen and its metabolites using an SRIXE approach over the experimental time course (0–4 h) demonstrated significant increases in Se levels and changes in Se distribution when compared with control cells (Fig. 5; Table 1). Interestingly, the pharmacokinetics indicated that there was a maximum intracellular loading of ebselen and its metabolites, with a maximum cellular Se content of 15 ng/cm2, which occurred 2 h after bolus addition of the drug. Beyond this time point, Se levels declined to approximately half the peak levels (7 ng/cm2) by 4 h. The known ebselen metabolism may provide a rationale for this decline in Se level, as ebselen is known to form intracellular thiol adducts, and ebselen–GSH conjugates (ebselen–SG) would be targets for cellular efflux [44]. These data, when combined with the intracellular distribution of Se and the effects of drug administration on other elements, provide a range of important mechanistic information regarding the likely antioxidant mechanism of the drug.

Ebselen initially induces a stress response in ND15 cells

The onset of oxidative stress is characterized by marked changes to cellular elemental content [37, 45]. Surprisingly, similar decreases were also observed in elemental content in response to ebselen administration, which indicated a cellular stress response. The peak intracellular Se concentration, reached 2 h after ebselen administration, corresponded to twofold to tenfold decreases in the levels of all other elements studied. By 4 h, as Se levels were decreasing, P, S, and Cl levels remained low, whereas the remaining elements showed substantial increases in intracellular concentrations. These findings indicate that the early outcome of ebselen administration is that the drug acts as a stress agent, rather than as an antioxidant, within the cell.

Ebselen appears to localize mainly to the endoplasmic reticulum

The SRIXE elemental maps reveal that ebselen is consistently located in a cellular region adjacent to, but not overlapping, the nucleus. Ebselen was found in the same region of the cell at all time points studied (1–4 h), which indicated that the drug was rapidly localized to this organelle. We hypothesize that this region of the cell is most likely to be the endoplasmic reticulum (ER), as the ER is continuous with the nuclear membrane and most of the ER lumen (which can occupy 10% of the cell volume) is found adjacent to the nucleus [4648]. The redox environment of the ER would also assist ebselen uptake, as the ER maintains a highly oxidized environment compared with other cellular organelles with a GSH to oxidized glutathione (GSSG) molar ratio approaching 1:1, compared with 100:1 for the rest of the cell [48]. The high abundance of GSSG and protein disulfides in this organelle could sequester ebselen though selenol–disulfide exchange reactions (Fig. 2) [49] once ebselen has entered the ER, with the formation of ebselen–protein adducts that would prevent Se efflux. An alternative possibility to the ER is that ebselen was sequestered into the mitochondria; however, drug uptake into mitochondria is primarily observed for positively charged compounds, owing to the negative potential of the mitochondria when energized [50]. As the mechanism of ebselen redox catalysis (Fig. 2) does not involve a charged intermediate, we considered that the ER was the more likely organelle to account for ebselen localization.

Implications for the antioxidant mechanism(s) of ebselen

The induction of a cellular stress response following ebselen administration, combined with the localization of ebselen to the ER, suggests that one potential mechanism of ebselen action is the induction of a stress response within the ER. This has potential implications for ebselen’s antioxidant mechanism, as an increase in ER stress would stimulate the cell to increase its antioxidant capacity via the activation of the antioxidant response element (ARE) [51, 52]. Ebselen has previously been characterized as a potent inducer of the ARE [22, 27], and the generation of ER stress would provide a mechanism to account for these phenotypic changes.

Ebselen stimulates transition metal uptake in a time frame comparable to that for ARE induction

A key distinction between cellular stress induced by agents, such as hydrogen peroxide, and the effect of ebselen administration is that, between 2 and 4 h, ebselen-treated cells displayed an influx of the d-block metals Mn, Cu, Ni, Fe, and Zn, whereas increases in oxidative stress result in transition metal efflux [37]. By 4 h, the levels of Mn, Fe, Ni, and Cu had recovered from the initial drop in elemental content, and these elements were present at much higher intracellular concentrations than before drug treatment, whereas the level of Zn had almost returned to the original level. This change in the direction of metal trafficking was most pronounced for Mn and Cu, with respective increases of 8.2- and 6.3-fold when compared with control cells.

We postulate that these changes in d-block metals are attributed to increases in the transcription of ARE-regulated genes, as these changes in metal trafficking parallel the known time course of ebselen-induced ARE induction [22]. Increases in the levels of transition metals are related to antioxidant activity, as the metalloenzymes that constitute major antioxidant defenses require the incorporation of their cognate transition metal cofactors for activity. Of the major antioxidant systems, cells possess two superoxide dismutase (SOD) genes under the control of the ARE promoter, a Cu/Zn SOD [53] and a Mn SOD [54]. Substantial increases in SOD protein levels would, therefore, deplete the cellular reserves of the trace elements Cu and Mn, requiring the cell to actively initiate metal uptake in order to fully constitute new SODs. In comparison with the levels of Cu and Mn, cells have relatively abundant levels of Zn [55] and this may explain why, unlike for Cu and Mn, Zn loads do not increase beyond the control values following ebselen administration, as presumably at these concentrations there is sufficient Zn to fully constitute Cu/Zn SOD activity. Similarly, induction of the ARE is known to upregulate the genes for both ferritin [56] and the transferrin receptor [57], with concomitant increases in the levels of cellular Fe. The increases in Ni levels are more puzzling, as Ni is not known to be utilized in mammalian biochemistry. However, prokaryotes do possess a functional Ni-based SOD [58], and it is possible that a stress-activated Ni uptake mechanism, although now redundant, has been evolutionarily conserved in eukaryotes. An alternative possibility is that, as many pumps are known to display low specificity for cations, the observed Ni import may have occurred as an indirect consequence of Ca and transition metal influx.

In summary, this is the first experimental report that has demonstrated that ebselen’s interactions with target cells can be measured using a SRIXE approach. The data reveal several key mechanistic steps following ebselen administration. Initially, there is a drug uptake phase, where uptake pathways predominate over drug efflux. Ebselen uptake appears to peak after 2 h, following which the cellular Se content declines, presumably due to a combination of metabolic conjugation and drug efflux. Analysis of the elemental maps suggests that ebselen or its metabolites immediately localized to the ER following uptake, and that they remained predominantly partitioned into this organelle for the 4-h duration of the study; the induction of ER stress is, therefore, likely to be a significant mechanism by which ebselen exerts antioxidant activity. Analysis of the cellular elemental content revealed that cells initially responded to ebselen administration in a manner analogous to that for oxidative stress, with an efflux of P, Cl, Ca, K, and transition metal species. However, unlike the stress response, at later time periods ebselen-treated cells increased their uptake of the transition metals Cu, Zn, Fe, Mn, and Ni, with the time course of metal uptake mirroring induction times for ARE-dependent protein synthesis. This transition metal influx has implications for ebselen’s protective effects, as the imported metal ions may be essential cofactors for newly synthesized Cu/Zn and Mn SOD proteins. Inherent in these findings is that structural modifications to ebselen and other GPx mimics could change drug localization and prevent ER accumulation, which would modulate the antioxidant properties of the drug. We are currently engaged in establishing the structure–activity relationships that govern ebselen’s localization, which will greatly advance our understanding of the pharmacological properties of ebselen and contribute to improved antioxidant drug design.


We are grateful for financial support provided by a University of Sydney Postdoctoral Fellowship (G.I.G), Australian Research Council (ARC) Discovery grants (P.A.L., H.H.H., and P.K.W.), including an ARC Australian Professorial Fellowship (to P.A.L), an ARC QEII Fellowship (to H.H.H.), and an ARC Australian Research Fellowship (to P.K.W.), a National Heart Foundation Grant-in-Aid (to P.K.W), and an Australian Synchrotron Research Program (ASRP) grant for access to the Advanced Photon Source facilities. The ASRP was funded by the Commonwealth of Australia under the Major National Research Facilities Program. The use of Advanced Photon Source facilities was supported by the US Department of Energy, Basic Energy Sciences, Office of Science, under contract number W-31-109-Eng-38.

Copyright information

© SBIC 2012