The Neuroprotective Functions of Selenoprotein M and its Role in Cytosolic Calcium Regulation

Abstract

Selenoproteins contain the trace element selenium incorporated as selenocysteine, the 21^st amino acid. Some members of the selenoprotein family, such as the glutathione peroxidases, have well-characterized antioxidant activity, functioning in enzymatic breakdown of hydroperoxides to protect cells against oxidative stress. However, the functions of many of the 25 human selenoproteins, including the brain-enriched selenoprotein M, are unknown. We investigated selenoprotein M function by manipulating expression in murine hippocampal HT22 cells, cerebellar astrocyte C8-D1A cells, and primary neuronal cultures. Overexpression of the protein resulted in a reduction in reactive oxygen species and apoptotic cell death in response to oxidative challenge with hydrogen peroxide. In contrast, knock-down of selenoprotein M using shRNA in primary neuronal cultures caused apoptotic cell death comparable to levels resulting from addition of hydrogen peroxide. Calcium measurements with the indicator cameleon demonstrated that overexpression of selenoprotein M decreased calcium influx in response to hydrogen peroxide. Additionally, knock-down of selenoprotein M expression in cortical cultures caused higher baseline levels of cytosolic calcium than in control cells. These results suggest that selenoprotein M may have an important role in protecting against oxidative damage in the brain and may potentially function in calcium regulation. Antioxid. Redox Signal. 12, 809–818.

Introduction

Selenoproteins are defined by incorporation of selenium (Se) into the amino acid selenocysteine. This family includes twenty-five selenoproteins in humans and twenty-four in mice (25). Selenium is an essential trace element for humans and animals, as indicated by the detrimental consequences of dietary selenium deficiency seen in regions of China and New Zealand, and further highlighted by the embryonic lethality in mice resulting from targeted distribution of the tRNA required for selenocysteine incorporation (7, 34). Symptoms of Se deficiency observed in humans include Keshan disease, a potentially fatal cardiomyopathy, Kashin–Beck disease, an osteoarthropathy occurring in regions of Tibet and China where selenium is deficient, and myxedematous endemic cretinism, a form of mental retardation occurring in selenium and iodine deficient regions of Africa (6, 14). In livestock, selenium deficiency leads to reduced weight gain, diarrhea, stillbirths, diminished fertility, and white muscle disease, a disease that affects both cardiac and skeletal muscle. Moreover, accumulating evidence implicates roles for Se in physiological and pathophysiological processes, including immune function, neurodegeneration, male reproduction, and cancer incidence (6, 23, 34).

The antioxidant defense system is associated with Se. Oxidative stress is implicated in the pathogenesis of neurodegenerative diseases, such as Alzheimer's disease and Parkinson's disease (9, 16, 32). Selenoproteins with known functions include the glutathione peroxidase family (GPx), thioredoxin reductases (TRxR), and the deiodinases (DIO) (17). GPxs reduce hydrogen peroxide and alkyl hydroperoxides at the expense of glutathione, contributing to the enzymatic antioxidant defense system in mammalian cells (8, 29). The TRxR family utilizes NADPH to reduce oxidized TRxR, which is used for regenerating cellular antioxidant systems, activating signaling molecules, reducing ribonucleotides to deoxyribonucleotides for DNA synthesis, and regulating activity of transcription factors (2, 18, 30). Deiodinases convert thyroid prohormones to active hormone, and active hormones to inactive metabolites, thus regulating hormonal activity (3, 5). In addition to these characterized families, numerous selenoproteins have been identified whose functions are not known, including a family with a common redox motif. This motif consists of CXXU, where U designates selenocysteine, and it has been identified in a subset of selenoproteins, including selenoprotein M (SelM), selenoprotein 15 (Sep15), selenoprotein W (SelW), and selenoprotein T (SelT) (13, 19).

Selenium is better retained in the brain than most other organs under conditions of dietary Se deficiency (4, 31). SelM is present at its highest levels in the brain (17). SelM has an endoplasmic reticulum localization signal and is retained there (17). Furthermore, changes in SelM levels were found to correlate with a mouse model for Alzheimer's disease that overexpresses a mutated form of human presenilin-2 (21). This mutation causes a form of early-onset Alzheimer's disease (24). Although the role of presenilin-2 is unclear, it is believed to regulate processing of the amyloid precursor protein, and the mutation is thought to cause buildup of a toxic fragment of amyloid beta. Mice overexpressing this mutant presenilin are deficient in SelM levels in the brain. This deficiency can be recovered by supplementing the mouse diet with sodium selenite (21).

We investigated a possible neuroprotective role of SelM in hippocampal and cerebellar astrocyte cell lines, as well as primary neuronal cultures. Because of the presence of the CXXU redox motif, we sought to determine the antioxidant properties and the potential role of SelM in protecting from oxidative stress-induced apoptosis (25). Additionally, because of the predicted location of SelM in the endoplasmic reticulum and recent studies indicating that another ER-localized selenoprotein, selenoprotein N, is required for ryanodine receptor calcium release channel activity in muscle (1, 12). We hypothesized that SelM could play a role in regulating calcium release from ER calcium stores (37).

Materials and Methods

Cell cultures

Murine HT22 hippocampal cells (obtained from R.R. Ratan, Department of Neurology, Harvard Medical School, The Beth Israel-Deaconess Medical Center) and C8-D1A cerebellar cells (American Type Culture Collection, Manassas, VA) were maintained in Dulbecco's modified Eagle's medium (Invitrogen Corp., Carlsbad, CA) containing 10% fetal bovine serum (FBS) (Invitrogen Corp.) at 5% CO_2, and 5.0% relative humidity. FBS lots were tested for Se content (Bodycote, Santa Fe Springs, CA), and the selenium concentration of media containing 10% FBS was 105 nM, calculated using inductively coupled plasma-mass spectrometry. For primary cultures, glass bottom tissue culture plates (World Precision Instruments, Sarasota, Fl) were coated with 0.1 mg/ml laminin in 0.1 mg/ml poly-L-lysine solution for 1 h, and then rinsed with PBS. Primary cortical cells were harvested from one day postnatal C57BL/6 mice and plated on coated dishes, and maintained in Neurobasal-A medium (Invitrogen Corp.) with 5% FBS with the addition of 100 μM glutamate (Invitrogen Corp.) to reduce glial cell growth, and enrich neuronal growth. B27 supplement (Invitrogen Corp.) was added to replace FBS after 24 h, and glutamate omitted from the media after 3 days. B27 was tested for Se content (Bodycote), and the selenium concentration of media containing 2% B27 was 93.8 nM, as determined by inductively coupled plasma-mass spectrometry.

Manipulation of SelM expression

A full-length human SelM cDNA (Invitrogen Corp.) was subcloned into pMSCVpuro retroviral vector (BD Biosciences Clontech, Mountain View, CA). RetroPack PT67 packaging cells (BD Biosciences Clontech) were transfected with SelM-MSCV retroviral vector or empty MSCV vector with Lipofectamine 2000 transfection reagent (Invitrogen Corp.). Transfected PT67 cells were selected in puromycin (Sigma Aldrich Co., St. Louis, MO) with the concentration of puromycin optimized (1–2 μg/ml) to kill the nontransfected control PT67 cells in 7–10 days. Transfected PT67 produced virus for viral infection of target HT22 and C8-D1A cells. HT22 and C8-D1A cells were selected with puromycin starting 48 h after viral transfection for a duration of ∼7–10 days. Constructs containing short hairpin RNAs (shRNA) (Thermo Scientific, Huntsville, AL) for mouse SelM and nontarget control were subcloned into the pTRIPZ vector containing the Tet-On^® inducible system with RFP. HT22 cells at 80%–90% optical confluency, or primary neuronal cultures (21 divisions) were transiently transfected with pTRIPZ vector containing nontarget shRNA control or mSelM shRNA using Lipofectamine 2000. Cells were induced with doxycycline (Clontech Laboratories, Inc.) 24 h post-transfection and incubated for 72–96 h prior to experimentation.

Reverse transcription real time PCR

RNA was harvested using an RNeasy mini kit according to the manufacture's instructions (Qiagen Inc., Valencia, CA). Synthesis of cDNA by High Capacity cDNA Reverse Transcription Kit was carried out according to manufacture's instructions (Applied Biosystems, Foster City, CA). Sequences of forward and reverse primers for real time PCR of the housekeeping gene hypoxanthine phosphoribosyl-transferase (mHPRT) (HPRT 5′: TCCTCCTCAGACCGCTTTT; HPRT 3′: CCTGGTTCATCATCGCTAATC), 18S RNA (18S 5′: CGATTGGATGGTTAGTGAGG; 18S 3′: AGTTCGACCGTCTTCTCAGC), and for SelM (SelM 5′: TGACAGTTGAATCGCCTAAAGGAG; SelM 3′: AACAGCACGAGTTCGGGGTC) are provided. Amplification of cDNA was carried out using Platinum SYBR green kits (Invitrogen Corp.) followed by detection using a 480 real time PCR machine (Roche Applied Science, Indianapolis, IN).

Radioactive ⁷⁵Se labeling of SelM

HT22 control and hSelM overexpressing cells were labeled by the addition of neutralized [⁷⁵Se] selenious acid (3 μCi/ml; specific activity, 1000 Ci/mmol), followed by incubation for 20 h. Cells were lysed with CelLytic MT (Sigma Aldrich Co., St. Louis, MO) according to the manufacturer's instructions. The proteins were separated on a 10%–20% gradient Tris-HCl Criterion Precast gel (Bio-Rad Laboratories, Inc., Hercules, CA) and transferred to polyvinylidene difluoride (PVDF). Radioactivity was detected by exposure to a phosphor screen that was subsequently imaged with a Typhoon-scanner (GE Healthcare, Piscataway, NJ).

Flow cytometry

HT22 control and hSelM overexpressing cells were plated to ∼90% confluency, and subsequently challenged with 0.5 mM, 1.0 mM, and 2.0 mM H₂O₂ for 2 h. Following peroxide challenge, cells were incubated in dihydrorhodamine 123 (DHR1,2,3) (Invitrogen Corp.) at 37°C. The cells were rinsed with PBS, dissociated from plates with trypsin/EDTA (Invitrogen Corp.), and resuspended in PBS with 2% FBS. Measurements of ROS production were carried out using flow cytometry on a FACS Caliber System (Becton Dickenson). HT22 control and shRNA SelM knock-down cells were treated in a similar fashion to overexpressing cells, except that dichlorodihydrofluorescein diacetate (H₂DCFDA) (Invitrogen Corp.), a green ROS indicator, was used due to the RFP marker in shRNA cells.

Calcium imaging

The Premo Cameleon Calcium Sensor Kit (Invitrogen Corp.) was used to transfect HT22 control and SelM overexpressing stable cell lines and primary neuronal cultures, according to manufacturer's instructions. The protein calcium indicator cameleon is a ratiometric sensor for the measurement of calcium signals with the employment of fluorescent or Förster resonance energy transfer (FRET) technology (26). The cameleon protein contains cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) modules linked with a calmodulin calcium-binding domain. The binding of calcium by the calmodulin domain causes a conformational change bringing the CFP and YFP domains in close proximity, allowing FRET to occur between the fluorophores. Excitation of CFP at 440 nm normally produces emitted light at ∼460–500 nm. After binding of calcium, this energy is transferred to YFP, shifting emission to 525–560 nm. Intracellular calcium was monitored with time lapse images of cells using an Olympus IX81 spinning-disk confocal microscope (Center Valley, CA) with an environmental chamber maintained at 37°C and ∼5% CO₂. The cameleon protein was excited at 440 nm, and emitted light was split into 485 nm (CFP) and 540 nm (YFP) with a Dual-View wavelength splitter (Optical Insights LLC, Santa Fe, NM). The ratio of 540 nm to 485 nm emission was used to determine internal calcium concentrations. In preliminary studies, cameleon-transfected cells were permeablized with ionomycin (Sigma Aldrich Co.) and imaged in the presence of varying external calcium concentrations. A baseline of calcium with a ratio of ∼4 in HT22 and ∼2 in primary cortical cells corresponds to ∼100 nM calcium.

Cell viability

HT22 control and SelM overexpressing cells were plated to ∼90% confluency and incubated with 0.5 mM, 1.0 mM, or 2.0 mM H₂O₂ for 4 h. Following peroxide challenge, propidium iodide (Invitrogen Corp.) was used to determine cell death, and was added to the cells in complete medium for 30 min. Images were acquired and quantified. HT22 nontarget shRNA control and SelM shRNA cells were plated at 10,000 cells per well in a 96-well format. CellTiter 96^® AQ_ueous Non-Radioactive Cell Proliferation Assay, (MTS) assay was used to determine cell proliferation and viability according to kit instructions (Promega Corporation, Madison, WI).

Statistical analysis

Statistical comparisons were carried out using Student's t-test and analysis of variance (ANOVA) with Bonferroni's posthoc test for multiple comparisons. Differences were judged statistically significant at p < 0.05.

Results

Previous studies have shown that SelM is highly expressed in the mouse brain, with the mRNA levels present in relatively moderate amounts (37). We investigated the expression profiles of the genes for SelM and other selenoprotein family members, along with factors implicated in selenoprotein synthesis, in mouse hippocampal neuronal HT22 and mouse cerebellar astrocyte C8-D1A cell lines, and in primary mouse cortical cultures, to determine whether these cell lines are appropriate models to study SelM. SelM mRNA was present at moderate levels in all of these cell types (Supplemental Fig. 1A-C; see www.liebertonline.com/ars).

Overexpression of SelM in HT22 and C8-D1A cells protects against hydroperoxide challenge

A retroviral gene delivery system was used to overexpress SelM RNA in neuronal cells. Stable HT22 and C8-D1A cell lines showed significant overexpression of the mRNA levels for human SelM above endogenous mouse SelM mRNA (Fig. 1A). Overexpressed SelM protein was detected by ⁷⁵Se labeling (Fig. 1B). A band of ∼15 kD, the predicted size of SelM, was clearly visible in lysates from SelM-overexpressing cells but barely detectable in control cells. The protective effects of SelM following peroxide challenge were quantified using flow cytometry. Both HT22 and C8-D1A stable overexpression cell lines were challenged using increasing concentrations of hydrogen peroxide for 2 h, and the generation of reactive oxygen species (ROS) was measured using dihydrorhodamine (DHR123) (Figs. 1C–1D). DHR123 is used as a fluorescent indicator for reactive oxygen species (ROS) generation measured such that increases in ROS production increase fluorescence (20, 33). Cells overexpressing SelM showed significantly reduced ROS production when measured by DHR123 fluorescence in the presence of varying concentrations of hydrogen peroxide (Figs. 1C–1D). Cells overexpressing SelM also exhibit increased viability compared with transfected empty vector controls following incubation with varying concentrations of hydrogen peroxide for 4 h (Fig. 1E), as shown by decreased propidium iodide (PI) fluorescence. Figure 1F indicates significantly less cell death in cells that overexpress SelM.

FIG. 1.

SelM overexpression in HT22 cells prevented H₂O₂-induced cell death. (A) Quantification of SelM mRNA overexpression, relative to HPRT, using RT-PCR in mouse hippocampal HT22 cells and cerebellar astrocyte C8-D1A cells. (B) SelM protein overexpression in HT22 control and hSelM overexpressing cells labeled with ⁷⁵Se (3 μCi/ml; specific activity, 1000 Ci/mmol). (C, D) SelM overexpression in HT22 and C8-D1A cells decreased superoxide production, as measured with flow cytometry using DHR123. Cells were exposed to increasing concentrations of H₂O₂ (0 mM, 0.5 mM, 1.0 mM, and 2.0 mM) for 2 h, as compared to MSCV empty vector controls (***p < 0.001). (E) HT22-MSCV empty vector control and HT22-SelM overexpression cells were plated equally and then challenged with H₂O₂ (0.5 mM, 1.0 mM, 2.0 mM) for 4 h, and then imaged for uptake of propidium iodide to detect cells with compromised membranes. (F) Quantification of propidium iodide positive cells from the images in (E) (**p < 0.01; ***p < 0.001). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).

Knock-down of mSelM in HT22 cells and primary cortex cultures increased oxidative damage

A construct encoding short hairpin RNA (shRNA) targeted to SelM in a lentiviral expression vector with a tet-inducible promoter (pTRIPZ) was used to knock down the levels of mouse SelM mRNA to generate stable monoclonal knock-down in HT22 cells and in primary neuronal cultures. The pTRIPZ vector also contains an RFP gene for identification of transfected cells. All pTRIPZ transfected cells were induced with doxycycline from 72 to 96 h prior to experiments to ensure adequate knock-down of SelM mRNA and promote RFP activation. As additional controls, un-induced samples were tested. Four stable HT22 cell colonies were generated, and a single colony exhibiting ∼80% knock-down was selected for subsequent studies (Fig. 2A). Primary cells exhibited a ∼60% decrease in mRNA levels compared to endogenous mRNA (Fig. 2B). ROS was also measured in knock-down cells using dichlorodihydrofluorescein diacetate (H₂DCFDA). H₂DCFDA is another indicator for ROS that is nonfluorescent until oxidation occurs within the cell to remove acetate groups (36). The stable HT22 shRNA knock-down colony was incubated with H₂DCFDA, which detected significant increases in ROS, as measured via flow cytometry (Fig. 2C). Knocking down the mRNA to ∼20% of endogenous levels increased ROS even in unchallenged cells. Additionally, cell viability was measured and the results indicate a significant decrease in viable cells when knocking down SelM (Fig. 2D). The effects observed from knocking down SelM in primary cortical cells to ∼40% of endogenous levels demonstrated the importance of this protein for neuronal survival.

FIG. 2.

Knock-down of SelM increases toxicity from oxidative stress. HT22 and primary cortical cells transfected with nontarget or SelM shRNA pTRIPZ vector and induced for 72 h. SelM mRNA levels were quantified relative to HPRT and 18S respectively using RT-PCR in (A) stably transfected HT22 cells and (B) transiently transfected primary cortical neurons. (C) Knock-down of SelM in HT22 cells increased reactive oxygen species as measured with H₂DCFDA and flow cytometry, as compared to nontargeting shRNA controls (***p < 0.0001). (D) Knock-down of SelM in HT22 cells decreased cell viability when measured with MTS assay, as a percent of the nontargeting shRNA control (**p = 0.0011).

Calcium responses induced by H₂O₂ in HT22 cells, measured with the cameleon protein calcium indicator

Having established that SelM protects against oxidative damage, and ultimately apoptosis, we hypothesized that SelM may function in regulation of calcium release from ER stores into the cytosol, due to its subcellular location within the ER, as shown in Figure 3. Cameleon was transfected into cells 48 hours prior to calcium measurements. A baseline of cytosolic calcium was initially established for each cell before oxidative challenge. Pseudocolor ratiometric images allowed visualization of calcium changes during each study (Fig. 3A). Representative panels of images from HT22 control cells transfected with the cameleon calcium indicator show the effect of 1 mM H₂O₂ on cytosolic calcium, indicated by color images (Fig. 3A). Following the addition of 1 mM H₂O₂, subsequent changes of fluorescent intensities of CFP to YFP were recorded (Fig. 3B). These fluorescent intensities are graphed as a ratiometric measurement of 530 nm (calcium-bound) to 485 nm (unbound) emission of cameleon protein over time in response to H₂O₂, indicating the changes in cytosolic calcium levels in the cell (Fig. 3C).

FIG. 3.

Calcium responses induced by H₂O₂ in HT22 cells, measured with the cameleon protein calcium indicator. Pseudocolor changes indicate cytosolic calcium changes. (A) A series of representative images corresponding to (B) intensity of CFP and YFP and (C) ratio of YFP to CFP correlating with intracellular calcium levels; the numbers on the graphs B and C correspond to the images in A. (D) Examples of calcium responses to 1 mM H₂O₂ in two cells measured for 2 h are shown in the images in (D), with YFP:CFP ratio changes for the same cells graphed in (E). The second calcium response in the right cell is followed by blebbing, indicative of apoptosis (inset, white arrows). (F and G) Cells incubated with thapsigargin overnight showed no calcium response following addition of 1 mM H₂O₂, indicating that responses depend on release of calcium from intracellular stores. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).

Calcium changes varied from cell to cell following addition of H₂O₂. Typically, peroxide caused an initial fast increase in cytosolic calcium levels, lasting many minutes before returning to basal values. In many cells, a second increase in calcium was observed, usually followed by visible signs of apoptosis, such as membrane blebbing. The second panel of images displays calcium increases following oxidative challenge, and subsequent calcium responses leading to apoptosis (Figs. 3D–3E). In order to determine whether calcium increases resulted from influx of extracellular calcium or release from endoplasmic reticulum calcium stores, cells were preincubated overnight prior to experiments with thapsigargin, to deplete internal stores (Fig. 3F). Thapsigargin is a tight-binding inhibitor of sarcoplasmic and/or endoplasmic reticulum calcium ATPase (SERCA) (35). Cells preincubated in thapsigargin did not respond to peroxide with increased cytosolic calcium levels (Fig. 3G). We also chelated external calcium using ethylene glycol tetraacetic acid (EGTA), which reduced but did not prevent responses to peroxide (data not shown). These data indicate that induction of calcium responses by peroxide is dependent upon intracellular stores and independent of extracellular calcium, although influx of extracellular calcium contributes to the magnitude of subsequent responses. Thus, the initial response to peroxide is a release of calcium from intracellular stores, and is augmented by calcium influx through membrane channels.

Overexpression of SelM in HT22 cells alters cytosolic calcium in cells under oxidative stress

Having confirmed that we could accurately measure calcium changes, we examined how changes in SelM expression could alter calcium in cells under oxidative stress. Cells overexpressing SelM had smaller calcium responses compared with control cells (Fig. 4A). Responses to oxidative stress varied from cell to cell, ranging from large brief calcium increases to smaller broad responses following the addition of 1 mM H₂O₂ (Fig. 4B). To demonstrate the reduced calcium responses in SelM-overexpressing cells, an inset with a subset of individual traces is provided. Averaged calcium measurements from Figure 4B showed that control cells had higher levels of cytosolic calcium following the addition of hydrogen peroxide than SelM-overexpressing cells (Fig. 4C).

FIG. 4.

SelM overexpression in HT22 cells alters cytosolic calcium responses to 1 mM H₂O₂. (A) Top panels are HT22 control cells, and bottom panels are HT22 cells overexpressing SelM; shown is baseline and following the addition of 1 mM H₂O₂. (B) Cytosolic calcium changes measured with cameleon control cells (blue lines, n = 10) and SelM overexpressing cells (red lines, n = 17). An inset with a subset of individual traces is shown to demonstrate the reduced calcium responses in SelM-overexpressing cells. (C) Baseline and peak response in HT22 control and SelM overexpression cells. Addition of H₂O₂ produced greater cytosolic calcium increases in HT22 control cells compared with SelM overexpressing cells. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).

Knock-down of SelM in primary cortex cultures results in higher baseline cytosolic calcium levels

Individual primary cortical cultures transfected with shRNA to knock-down SelM also exhibited variable responses to oxidative stress following addition of 100 μM H₂0₂ (Fig. 5A). Calcium responses to peroxide in cortex cultures were generally much smaller than those observed in HT22 cells. Example traces showing calcium changes for knock-down SelM and control cells are provided in an inset. Averaged calcium measurements from Figure 5A shows that baseline calcium levels were increased in cells after knock-down of SelM, but the change in calcium in response to oxidative stress relative to baseline was not significantly different (Fig. 5B). Thus, these findings with primary cortical cultures are consistent with our findings in cell lines that SelM has a neuroprotective role. A series of pseudocolor images for control and SelM shRNA indicate the changes observed in cytosolic calcium (Fig. 5C). The SelM knock-down cultures often exhibit blebbing along the processes that may be indicative of apoptosis (11). Knock-down of SelM resulted in increased blebbing following peroxide-induced calcium increases, and this blebbing was not observed in control cells (Fig. 5C). Furthermore, this blebbing can be seen along the processes following knock-down of SelM prior to oxidative challenge (Fig. 5D). A quantification of the swellings along the processes, as a measure of blebbing observed in control and SelM shRNA cultures showed a significant increase in the SelM knock-down cultures compared to controls (Fig. 5E).

FIG. 5.

Knock-down of SelM in primary cortical cells causes increased baseline calcium. (A) Cytosolic calcium measurements of primary cortical neurons transfected with cameleon protein in nontargeting shRNA control cells (blue lines, n = 20) and SelM shRNA-transfected cells (red lines, n = 21). The right panel shows two experiments demonstrating the varying calcium responses. (B) Average baseline calcium levels were higher in primary cortical cells transfected with SelM shRNA compared with nontargeting shRNA control cells (*p < 0.05), while calcium increases in response to H₂O₂ were similar in both cells (p > 0.05). (C) A series of representative pseudocolor images of nontargeting shRNA control (top panels) and SelM shRNA cells (bottom panels) corresponding to calcium responses to 100 μM H₂O₂. (D) Neurons expressing SelM shRNA show blebbing in processes (indicated by white arrows) in response to oxidative challenge. Oxidative challenge did not cause blebbing in nontargeting shRNA controls. (E) An average of the number of swellings in processes observed per cell of nontargeting shRNA control cells (blue, n = 14) and SelM shRNA-transfected cells (red, n = 17) was significantly increase in SelM shRNA cells (p = 0.03). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).

Discussion

The results reported herein show that stable overexpression of SelM in HT22 hippocampal cells and C8-D1A cerebellar cells prevented oxidative damage induced by hydrogen peroxide, indicating SelM is an antioxidant protein with neuroprotective properties. As increasing SelM levels protected against oxidative stress, we hypothesized that decreasing the endogenous levels of SelM would render cells more vulnerable to oxidative challenge. Indeed, successful knock-down of SelM in HT22 cells and primary cortical cells increased oxidative damage. Furthermore, in the absence of oxidative challenge, SelM knock-down resulted in decreased cell viability and increased ROS, further demonstrating the functional importance of SelM in preventing oxidative stress. In addition, a recent study reported that overexpression of SelM in a transgenic rat model resulted in the differential regulation of antioxidants as well as the activities of antioxidant enzymes (22).

Our results show that overexpression of SelM in HT22 and C8-D1A cell lines reduced changes in cytosolic calcium evoked by oxidative stress, without affecting basal calcium levels. This indicates that SelM influences the process through which cells release calcium from their internal stores and, ultimately, may alter apoptotic pathways (10). In primary cortical cultures, knock-down of SelM increased baseline levels of calcium. Taken together, these data suggest that calcium regulation was impaired with SelM knock-down. The inability to detect significant calcium responses to oxidative stress under conditions of SelM knock-down may have been a consequence of the increased basal calcium in these cultures.

The importance of calcium regulation for neuronal viability and responses to oxidative stress is well documented, highlighting the potential importance of SelM in this capacity (27). In addition, disruption of neuronal calcium homeostasis is implicated in the mechanism of neuron degeneration in several diseases, including Alzheimer's disease and Huntington's disease (28). A diagrammatic model of the proposed antioxidant and calcium regulation functions of SelM is shown in Figure 6. Increased oxidative stress leads to increases in calcium and ROS, triggering apoptosis. Overexpression of SelM, which is localized to the endoplasmic reticulum (15), blocks or delays apoptosis. Questions still remain as to how SelM influences calcium, and whether this action is direct or indirect, and thus further investigation of this protein will be necessary to determine the pathways through which SelM functions as a neuroprotective antioxidant.

FIG. 6.

Model for SelM function. SelM decreases calcium release from ER in response to oxidative stress and reduces apoptotic cell death.

Abbreviations Used

CFP

cyan fluorescent protein

DHR123

dihydrorhodamine

EGTA

glycol tetraacetic acid

ER

endoplasmic reticulum

FRET

fluorescent or Förster resonance energy transfer

GPx

glutathione peroxidase

GSH

glutathione

H₂DCFDA

dichlorodihydrofluorescein diacetate

H₂O₂

hydrogen peroxide

HPRT

hypoxanthine-guanine phosphoribosyltransferase

MTS

tetrazolium compound: [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt

PI

propidium iodide

ROS

reactive oxygen species

RT-PCR

reverse-transcriptase real-time PCR

Se

selenium

SelM

selenoprotein M

shRNA

short-hairpin RNA

TRxR

thioredoxin reductase

TUNEL

Terminal deoxynucleotidyl Transferase dUTP Nick End Labeling

U

selenocysteine

YFP

yellow fluorescent protein

Acknowledgments

This study was supported by NIH grants DK47320 and NS40302 to MJB; Hawaii Community Foundation grant 20061490 to FPB; and NIH RR003061 and RR016467 for the JABSOM Microscopy and Imaging core facility. We thank Dr. Robert Nichols for helpful comments and suggestions, and thank Ann Hashimoto, Andrea Takemoto, and Miyoko Bellinger for technical assistance.

Author Disclosure Statement

No competing financial interests exist.