Cell Death Caused by Selenium Deficiency and Protective Effect of Antioxidants*

Selenium is an essential trace element and it is well known that selenium is necessary for cell culture. However, the mechanism underlying the role of selenium in cellular proliferation and survival is still unknown. The present study using Jurkat cells showed that selenium deficiency in a serum-free medium decreased the selenium-dependent enzyme activity (glutathione peroxidases and thioredoxin reductase) within cells and cell viability. To understand the mechanism of this effect of selenium, we examined the effect of other antioxidants, which act by different mechanisms. Vitamin E, a lipid-soluble radical-scavenging antioxidant, completely blocked selenium deficiency-induced cell death, although α-tocopherol (biologically the most active form of vitamin E) could not preserve selenium-dependent enzyme activity. Other antioxidants, such as different isoforms and derivatives of vitamin E, BO-653 and deferoxamine mesylate, also exerted an inhibitory effect. However, the water-soluble antioxidants, such as ascorbic acid, N-acetyl cysteine, and glutathione, displayed no such effect. Dichlorodihydrofluorescein (DCF) assay revealed that cellular reactive oxygen species (ROS) increased before cell death, and sodium selenite and α-tocopherol inhibited ROS increase in a dose-dependent manner. The generation of lipid hydroperoxides was observed by fluorescence probe diphenyl-1-pyrenylphosphine (DPPP) and HPLC chemiluminescence only in selenium-deficient cells. These results suggest that the ROS, especially lipid hydroperoxides, are involved in the cell death caused by selenium deficiency and that selenium and vitamin E cooperate in the defense against oxidative stress upon cells by detoxifying and inhibiting the formation of lipid hydroperoxides.

Selenium is an essential trace element for humans and many other forms of life, and a deficiency of this element induces some pathological conditions, such as cancer, coronary heart disease, and liver necrosis (1)(2)(3)(4)(5). Selenium deficiency is also accompanied by a loss of immunocompetence (6), and both cell-mediated immunity and B-cell function can be impaired (7). Supplementation with selenium has marked immunostimulant effects, including an enhancement of activated T-cell proliferation (8). Selenium is an essential component of several enzymes such as glutathione peroxidase (GPx) 1 (9), thioredoxin reductase (TR) (10), and selenoprotein P (SeP) (11), which contain selenium as selenocysteine. It is also well known that selenium is essential for cell culture when a serum-free medium is used (12). Serum-free media, especially for immune cells and neurons, contain insulin, transferrin, and sodium selenite. Without selenium, cells can neither proliferate nor survive. However, the underlying mechanism for the role of selenium in cell proliferation is still unknown.
Vitamin E, a generic term for tocopherols and tocotrienols, is one of the most potent lipid-soluble antioxidants (13). Vitamin E occurs in nature in at least eight different isoforms: ␣-, ␤-, ␥-, and ␦-tocopherols and ␣-, ␤-, ␥-, and ␦-tocotrienols (14). Tocotrienols differ from the corresponding tocopherols only in their aliphatic tail. Vitamin E deficiencies have been implicated in some pathologic conditions, such as cancer, coronary heart disease, and liver necrosis (15,16) and are also accompanied by a loss of immunocompetence (17). It is well known that selenium and vitamin E show compensative effects and that a deficiency of both elements causes massive injury in some cases (18 -20).
In the present study, we characterize the nature of cell death caused by selenium deficiency and the cell death inhibitory effect of antioxidants including vitamin E. We also demonstrate the involvement of reactive oxygen species (ROS), especially lipid hydroperoxides, on the cell death.
Cell Culture and Determination of Cell Viability-Jurkat E6 -1 cells, human T-leukemia (American Tissue Type Collection) were maintained in RPMI 1640 medium containing 100 units/ml penicillin G, 100 g/ml * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed: Human Stress Signal Research Center (HSSRC), National Institute of Advanced Industrial Science and Technology (AIST), 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, Japan. Tel.: 81-72-751-8293; Fax: 81-72-751-9964; E-mail: yoshiro-saito@aist.go.jp. streptomycin, 0.25 g/ml amphotericin B, and 10% heat-inactivated fetal calf serum at 37°C under an atmosphere of 95% air and 5% CO 2 , as described previously (22). For studies on the effects of selenium depletion, cells (3 ϫ 10 5 cells/ml) were cultured with RPMI 1640 medium containing 5 g/ml human insulin, 5 g/ml human transferrin, 92 nM FeCl 3 , and 2.5 mg/ml bovine serum albumin (ITA-RPMI). Stock solutions of sodium selenite and vitamin E were prepared in PBS and Me 2 SO, respectively. For the determination of cell viability, trypan blue assay, and MTT assay were conducted for the indicated times. In the former, cells that excluded trypan blue after incubation with PBS containing 0.04% trypan blue dye (Invitrogen) were considered viable. For the latter analysis, cells were incubated with 0.5 mg/ml MTT at 37°C for 2 h. Isopropyl alcohol containing 0.04 N HCl was added to the culture medium (3:2, by volume), and they were mixed by pipette until the formazan was completely dissolved. The optical density of formazan was measured at 570 nm using a Multiskan Ascent plate reader (Theromo Labsystems, Helsinki, Finland).
Cell Death Assay-Phosphatidylserine (PS) exposure (23) and caspase activity (24) were analyzed as described previously. PS exposure was measured by the binding of annexin V-FITC according to the protocol outlined by the manufacture in the Apoptosis Detection-kit (Sigma-Aldrich Co.). Treated cells were also stained with propidium iodide, followed by analysis with a Cytomics FC500 Flow Cytometry System (Beckman Coulter, Inc., Miami, FL) with a 488-nm argon laser. Caspase activity was measured by the cleavage of Asp-Glu-Val-Asp (DEVD) peptide-conjugated p-nitroanilide (pNA) according to the protocol outlined by the manufacture in the Caspase-3/CPP32 Colorimetric Protease Assay-kit (Medical & Biological Laboratories Co. Ltd., Nagoya, Japan). Substrate cleavage to release pNA (405 nm) was measured using a Multiskan Ascent plate reader (Theromo Labsystems, Helsinki, Finland). Absorbance units were converted to pmol of pNA using a standard curve generated with free pNA.
Cytosol Preparation-Cytosol was prepared as described previously (22). After culturing for the specified periods, cells were collected and resuspended in an appropriate volume of 50 mM Tris-HCl (pH 7.4), containing 0.25 M sucrose, 0.1 mM EDTA, 0.7 mM 2-mercaptoethanol, and 2 mM diisopropyl fluorophosphate. The cell suspension was sonicated and centrifuged at 105,000 ϫ g for 1 h at 4°C to obtain a cytosolic fraction.
Protein Assay-Protein content of cytosol and cell samples were determined using a BCA protein assay kit (Pierce) with bovine serum albumin as a standard.
Determination of Intracellular Reactive Oxygen Species-Intracellular ROS were detected using DCFH-DA as described previously (25) with a slight modification. After culturing for the specified periods, cells were collected, resuspended in PBS and incubated with DCFH-DA at a final concentration of 5 M for 15 min at 37°C. Then, cells were washed once with PBS and incubated for 1 h at 37°C. Cells were excited with a 488-nm argon ion laser in a Cytomics FC500 Flow Cytometry System, and the DCF emission was recorded at 525 nm. Data were collected from at least 10,000 events.
Determination of Intracellular Lipid Hydroperoxides-Intracellular lipid hydroperoxides were detected using fluorescence probe DPPP (25) and chemiluminescence HPLC systems (26,27) as described previously with a slight modification. For DPPP assay, cells were preincubated in PBS at 37°C at a density of 1 ϫ 10 7 cells/ml for 5 min. After addition of DPPP (in Me 2 SO) at a final concentration of 167 M, the cell suspension was incubated for 5 min in the dark. Cells were washed and resuspended in the specified medium. At the times indicated, cells were collected and resuspended in PBS. Fluorescence intensities of the cell samples were measured with the Spectrofluorophotometer RF-5300PC (Shimadzu Co., Kyoto, Japan) with excitation and emission wavelengths of 351 and 380 nm, respectively. For HPLC analysis, cell samples in PBS were mixed with chloroform/methanol (2/1) containing 0.02% butylated hydroxytoluene at twice the volume of the samples. Then, an equal volume of 0.1 M NaCl with cell samples was added to the extract and mixed. After centrifugation for 10 min at 1,500 ϫ g, the lower chloroform layer was evaporated to dryness under a stream of N 2 , redissolved in chloroform/methanol (2:1) and injected into HPLC for lipid hydroperoxide analysis. The accumulation of cholesterol hydroperoxide (FC-OOH) was followed with HPLC using a post-column chemiluminescence detector (CLD-10A, Shimadzu, Japan) and a spectrophotometric detector (SPD-10AV, Shimadzu, Japan). An ODS-2 column (5 m, 250 ϫ 4.6 mm, GL Science, Japan) was used and methanol/acetonitrile/water (45:46:9 by volume) was delivered as an eluent at 1 ml/ min. After passage through the UV detector, the eluent was mixed with a luminescent reagent in the postcolumn mixing joint in the chemiluminescence detector at 40°C. The luminescence reagent containing cytochrome c (10 mg) and luminol (2 mg) in 1 liter of alkaline borate buffer (pH 10) was loaded at the flow rate of 0.5 ml/min. Phospholipid hydroperoxides were also followed with HPLC using a chemiluminescence detector. Finepack SIL NH 2 -5 column (5 m, 250 ϫ 4.6 mm, JASCO, Japan) was used and hexane/isopropyl alcohol/methanol/water (5:7:2:1 by volume) was eluted at 1 ml/min.
Statistics-Data are reported as means Ϯ S.D. of at least three separate experiments. The statistical significance of differences between determinations was calculated by Student's t test, and values of p Ͻ 0.05 were considered significant.

Effect of Selenium Deficiency on the Viability of Jurkat
Cell-To determine the effect of selenium on viability, Jurkat cells were cultured with serum-free RPMI 1640 medium (ITA-RPMI). When cultured with a selenium-deficient medium, the viability of Jurkat cells decreased with incubation time (Fig. 1). The cell viability started to decrease after 24 h, and a higher than 95% loss was observed within 60 h. The viabilities as measured by MTT and TPB assay were in close agreement. In contrast, Jurkat cells cultured with serum-free RPMI 1640 medium containing 100 nM sodium selenite did not show any significant loss of viability.
Characterization of Cell Death Caused by Selenium Deficiency-To identify the type of cell death caused by selenium deficiency, PS exposure and caspase activity were analyzed. The selenium-deficient cells cultured with ITA-RPMI for 36 h were incubated with annexin V-FITC and propidium iodide and then subjected to flow cytometry analysis. Selenium-deficient cells showed not only signs of PS exposure but also uptake of propidium iodide (Fig. 2B). The dead cells in the selenium- deficient medium did not show any exclusion of propidium iodide for the time tested (24 -40 h, data not shown). Caspase activity in the selenium-deficient cells was also measured using DEVD peptide conjugated to the chromophore pNA. In the selenium-deficient cells cultured for 36 h, caspase activity was below the background level seen in control and selenium-sufficient cells (Fig. 2C), while activity was detected in cells treated with 50 M hydrogen peroxide for 6 h (apoptotic condition), but not in those treated with 500 M hydrogen peroxide (necrotic condition), as previously reported (28). Selenium-deficient cells did not show caspase activation for the time tested (24 -36 h, data not shown), suggesting that this cell death is necrotic rather than apoptotic.
Inhibitory Effect of Selenium-containing Protein and Compounds on Cell Death Caused by Selenium Deficiency-A dosedependent study of the inhibitory effects of selenium revealed that sodium selenite at levels higher than 10 nM protected cells almost completely (Fig. 3). The ED 50 of sodium selenite was 3.3 Ϯ 1.5 nM. SeP, which functions as a selenium supply protein (22), also demonstrated an inhibitory effect, with the ED 50 being 0.066 nM. Selenium-containing amino acids, such as seleno-DL-cystine, seleno-L-methionine, and seleno-DL-methionine, also inhibited cell death (Table I). Ebselen, which is a mimic of GPx (29), did not have an inhibitory effect. To clarify temporally the site of selenium action in selenium deficiency-induced cell death, sodium selenite was added to cells at various time points after culturing with the selenium-deficient medium. Almost complete protection of cell death was observed even when sodium selenite was added at 24 h after selenium deficiency (Fig. 4).
Effects of Selenium Deficiency on the Enzyme Activity of Selenoproteins-We next examined the effects of selenium deficiency on the enzyme activity of cellular selenoproteins in Jurkat cells. As shown in Fig. 5, cellular GPx (cGPx), phospholipid hydroperoxide GPx (PH-GPx), and thioredoxin reductase (TR) activities were reduced in selenium-deficient cells grown in a selenium-deficient medium for 24 h. cGPx, PH-GPx, and TR activities were reduced to 36, 36, and 39%, respectively, of those in control cells grown in a medium containing serum. Selenoenzyme activities of cells cultured with a selenium-deficient medium for 48 and 72 h could not be measured because of the lower cell recovery rates (Fig. 5).
Inhibitory Effect of Vitamin E and Other Antioxidants on Cell Death Caused by Selenium Deficiency-As described above, a marked decrease of selenoenzyme activities was observed in the selenium-deficient cells. It is well known that these selenoenzymes play an important role in the antioxidative defense system. To understand the underlying mechanism of the protective effect of selenium, we examined the effect of other types of antioxidants. Water-soluble antioxidants, such as ascorbic acid, N-acetyl cysteine, and glutathione, did not inhibit cell death caused by selenium deficiency even at 1 mM (Table I), whereas the lipid-soluble antioxidant ␣-tocopherol completely blocked cell death (Fig. 6), although ␣-tocopherol did not produce any decrease of selenoenzyme activies (Fig. 5). Almost complete protection of cells was also observed even when ␣-tocopherol was added at 24 h after selenium deficiency (Fig. 4). We also observed the inhibitory effect of other forms of vitamin E, such as ␤-, ␥-, and ␦-tocopherols and ␣-, ␤-, ␥-, and ␦-tocotrienols ( Fig. 6 and Table I). The ED 50 values for tocotrienols were smaller than those of the corresponding tocopherols, suggesting that tocotorienols are more potent inhibitors than the corresponding tocopherol isoforms. The cellular uptake of ␣-tocotrienol was found to be 2.2-fold higher than that of ␣-tocopherol after incubation for 72 h (data not shown), which corresponds well with the difference in their ED 50 values, the ratio being 2.8-fold (Table I). We also examined the inhibitory effect of vitamin E derivatives, such as PMC and Trolox, the former being a short-chain homolog of ␣-tocopherol and the latter a water-soluble analog of PMC. These compounds also blocked cell death (Table I), and the large difference in the ED 50 between them suggests that the lipid-soluble antioxidant retained in the membranes exerts a higher level of activity than does the hydrophilic antioxidant. BO-653, a synthetic radical-scavenging antioxidant (21), showed a similar inhibitory effect to tocopherols (Table I). Deferoxamine mesy-late, which has metal chelating properties, also completely blocked the cell death caused by selenium deficiency (Table I).
Evaluation of Intracellular Reactive Oxygen Species and the Inhibitory Effect of Selenium and Vitamin E-We determined intracellular ROS production in cells using a fluorescence probe DCFH-DA. The selenium deficiency-induced death of Jurkat cells was preceded by an increase in intracellular ROS levels (Fig. 7A). Studies on the kinetics of this change showed that DCF fluorescence was slightly higher after culture for 12 h in the selenium-deficient medium and reached a plateau at 24 h (Fig. 7B). Selenite and ␣-tocopherol prevented the accumulation of intracellular ROS as measured by DCF fluorescence (Fig. 7, A and B). The dose-dependent study of the effect  of selenite and ␣-tocopherol on intracellular ROS levels revealed that these compounds prevented the increase of DCF fluorescence in a dose-dependent manner (Fig. 7C). Watersoluble antioxidants, such as ascorbic acid, N-acetyl cysteine and glutathione, did not prevent the accumulation of intracellular ROS despite the addition of as much as 1 mM (data not shown).
Determination of Intracellular Lipid Hydroperoxides-The inhibitory effects of lipid-soluble antioxidants on cell death implicate intracellular lipid hydroperoxides in selenium deficiency-induced cell death. The levels of lipid hydroperoxides in cells were measured by fluorescence probe DPPP and HPLC chemiluminescence. DPPP has been proved to be a sensitive probe for lipid hydroperoxides (25,30,31). It reacts with lipid hydroperoxide stoichiometrically to give a fluorescent DPPP oxide. When the DPPP-labeled cells were cultured with the selenium-deficient medium, the fluorescence intensity derived from DPPP oxide increased in a time-dependent manner (Fig.  8A), but no increase was observed when cultured with sodium selenite or ␣-tocopherol. These cell death inhibitors dose-dependently suppressed the increase in DPPP fluorescence (Fig.  8B). Using HPLC chemiluminescence, cholesterol hydroperoxide (FC-OOH) was detected in cells cultured with the seleniumdeficient medium for 24 h, and the molar ratio of FC-OOH to FC was 0.41 Ϯ 0.30 pmol/nmol (n ϭ 4). Interestingly FC-OOH was detected as a major lipid hydroperoxide in this system, and phospholipid hydroperoxides were at an undetectable level (Ͻ0.0053 pmol/nmol PC-OOH/PC and Ͻ0.015 pmol/nmol PE-OOH/PE) in the selenium-deficient cells. FC-OOH was not detected in the presence of sodium selenite, ␣-tocopherol, or fetal bovine serum (Ͻ0.015 pmol/nmol FC-OOH/FC).

DISCUSSION
The essential role of selenium in nutrition has been well established. It is also well known that selenium is necessary for cell culture when using a serum-free medium. Insulin (as a growth factor), transferrin (as an iron source), and selenite are added to the serum-free media for immune and neuronal cells. Although the effects of selenium on cell viability and the cellcycle progression has been reported (12,32), the underlying mechanism of the protective effect of this element has not yet been elucidated. In the present study using Jurkat cells, a model of proliferating T lymphoma cells, the decrease in cell viability was observed when applying a serum-free medium without selenium. This cell death was completely blocked by selenium-containing materials, except for ebselen, in a dosedependent manner. It has been reported that these seleniumcontaining materials are incorporated and can be the cellular source of selenium used for synthesis of selenoprotein (22). SeP, which is a selenium-rich extracellular glycoprotein (11,33,34) that functions as selenium transport protein (22,35,36), was the most effective of the materials tested (Table I).
Although we observed a loss of cell viability after 24 h, Jurkat cells duplicated for 20 h under the serum-free culture conditions. This observation suggests that cell death occurred after a single division. We speculate that the proliferating cells became selenium-deficient, and the divided cells contained almost half of the selenoenzyme activities, such as cGPx, PH-GPx, and TR. In the presence of sodium selenite, these enzyme activies were retained or up-regulated in cells. The decrease in TR activity was lower than that of cGPx and PH-GPx activities in the Jurkat cells cultured with selenium-deficient medium for 72 h (Fig. 5). Selenoproteins have been proposed to follow a hierarchy for selenium supply in that the amounts of certain selenoproteins decrease more rapidly under selenium-deficient conditions (37). A previous study suggested that this was due in part to differences in the SECIS elements. Gasdaska et al. (38) demonstrated that the element of TR was highly active; therefore, TR levels would be better preserved when the selenium supply was limited as in a selenium-deficient medium. It has been reported that these selenoproteins play an important role in the defense against oxidative stress (1,39). In the case of lipid hydroperoxides, it is known that PH-GPx, but not cGPx, is able to reduce lipid hydroperoxides, including phospholipid hydroperoxide and cholesterol hydroperoxide, directly (40,41). It has also been proved that overexpression of PH-GPx suppresses cell death due to oxidative damage induced by radical initiator and lipid hydroperoxide (42,43). Moreover, Lewin et al. (44) reported that TR also plays a role in preventing oxidative damage induced by tert-butyl hydroperoxide and oxidized LDL, but the mechanism of the protection afforded by this selenoprotein against oxidative stress induced by lipid hydroperoxides is still unclear. At present, these selenoproteins are assumed to play an important role in the defense against oxidative stress related to lipid hydroperoxides. Under selenium-deficient conditions, the decrease in these selenoproteins is speculated to cause peroxidation in the lipid layer inside cells.
It is noteworthy that radical-scavenging antioxidants, such as ␣-tocopherol, completely blocked the cell death caused by selenium deficiency, although ␣-tocopherol did not affect the enzyme activity of selenoproteins. ␣-Tocopherol-supplemented Jurkat cells did not show any loss of cell viability despite the undetectable levels of cGPx and PH-GPx activity. Other isoforms of vitamin E, such as ␤-, ␥-, and ␦-tocopherols and ␣-, ␤-, ␥-, and ␦-tocotrienols, were also effective. Tocotrienols were more effective than the corresponding tocopherols (Table I), which may be ascribed primarily, if not solely, to the differences in the rate of cellular uptake. A higher uptake of ␣-tocotrienol than ␣-tocopherol into culture cells has been reported (45,46). Such a difference was also observed for liposomal membranes (47). Vitamin E derivatives, such as PMC and its water-soluble analogue Trolox, also showed an inhibitory effect on cell death, but their higher ED 50 values (Table I) suggest that the antioxidant incorporated into cell membranes is more effective than that localized outside the membranes. Interestingly, a synthetic radical-scavenging antioxidant, BO-653, also showed an inhibitory effect, suggesting the importance of the inhibition of lipid peroxidation in the cell membranes. The fact that selenium deficiency, which results in a decrease in the capacity to reduce lipid hydroperoxides, induced cell death and that this cell death could be inhibited by radical-scavenging antioxidants that suppress the formation of lipid hydroperoxides strongly indicates the causative role of lipid hydroperoxides in cell death.
The ED 50 value of ␣-tocopherol was as low as 36 nM (Table I). One may argue that this concentration is quite low compared with the physiological concentration; for example, 30 M in human plasma. It should be pointed out, however, that the concentration of ␣-tocopherol in the membrane is of more importance than that in the bulk phase. In the present study, the lipid concentrations were measured as follows; FC, 4.4; PC, 12; PE, 5.2 nmol/10 6 cells. Thus, the molar ratio of ␣-tocopherol to total lipids in the cell culture system (10 6 cells/ml) was 36 ϫ 10 Ϫ9 /22 ϫ 10 Ϫ6 M ϭ 1/610 mol/mol, which is similar to that in human plasma; that is, 30 ϫ 10 Ϫ6 /11 ϫ 10 Ϫ3 M ϭ 1/370 mol/ FIG. 7. Generation of intracellular reactive oxygen species and its inhibition by sodium selenite and ␣-tocopherol. Cells were cultured in a selenium-deficient medium, without and with 100 nM sodium selenite or 2 M ␣-tocopherol for indicated times, and the intracellular ROS production was measured using a fluorescence probe DCFH-DA as described under "Experimental Procedures." A, cells were cultured in each medium for 24 h. The flow cytometry data presented are representative of three independent experiments. B, cells were cultured in each medium for indicated times. *, p Ͻ 0.05 when com- mol. It may be noted that the micromolar ranges of ␣-tocopherol applied to many cell culture systems are not always physiological, but that the concentration in the membranes should be considered. Deferoxamine mesylate, a well-known iron chelator, was also found to be a potent inhibitor of cell death induced by selenium deficiency. It has been suggested that iron plays an important role in oxidative damage to cells (48,49), by reacting with hydrogen peroxide or lipid hydroperoxides to form reactive oxygen radicals.
It was found that the removal of selenium from the culture medium induced ROS production as measured by DCF fluorescence and also lipid hydroperoxides as measured by DPPP fluorescence and HPLC chemiluminescence. We are aware of the inherent drawbacks of DCFH, but it can be a useful probe for estimating semi-quantitatively the generation of ROS under specific conditions (50). Both selenium and ␣-tocopherol suppressed ROS by apparently different mechanisms, the former by enhancing the reduction of hydroperoxides, while the latter by inhibiting their formation. FC-OOH was detected as a major lipid hydroperoxide, which was unexpected since polyunsaturated lipids in PC and PE are more susceptible to oxidation than free cholesterol. One possible reason could be that the reduction of FC-OOH by PH-GPx is at least six times slower than that of phospholipid hydroperoxides (41).
Many reports have shown significant correlations between selenium deficiency and the incidence of cancer (51,52). It has been also reported that selenium supplementation prevents the generation of cancer (2,53). As shown in this study, selenium deficiency caused a significant increase in ROS and peroxidation inside cells. It is therefore considered that selenium-deficient conditions cause oxidative DNA damage that eventually leads to cancer formation.
In conclusion, the present study clearly shows that selenium deficiency decreased the activities of cGPx, PH-GPx, and TR, increased lipid peroxidation in the membranes, and eventually induced cell death. The cell death was inhibited by other types of antioxidants with different functions, such as tocopherols and deferoxamine, which inhibit lipid peroxidation in the membranes and sequester redox-active iron, respectively. These results strongly indicate that the lipid hydroperoxides play a causative role in the oxidative damage to cells induced by selenium deficiency.