Advertisement

An Unexpected Functional Link between Lysosomal Thiol Reductase and Mitochondrial Manganese Superoxide Dismutase*

  • Branka Bogunovic
    Affiliations
    Department of Microbiology and Immunology, Georgetown University, Washington, D. C. 20057
    Search for articles by this author
  • Milica Stojakovic
    Affiliations
    Center for Cancer and Immunology Research, Children's Research Institute, Children's National Medical Center, Washington, D. C. 20010-2970
    Search for articles by this author
  • Leonard Chen
    Affiliations
    Department of Microbiology and Immunology, Georgetown University, Washington, D. C. 20057
    Search for articles by this author
  • Maja Maric
    Correspondence
    To whom correspondence should be addressed: Dept. of Microbiology and Immunology, Georgetown University Medical Center, 3900 Reservoir Rd. NW, Med/Dent C301, Washington, D. C. 20057. Tel.: 202-687-3749; Fax: 202-687-1800
    Affiliations
    Department of Microbiology and Immunology, Georgetown University, Washington, D. C. 20057
    Search for articles by this author
  • Author Footnotes
    * This work was supported by American Heart Association Scientist Development Grant 0535032N. 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.
    The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1 and S2.
Open AccessPublished:January 24, 2008DOI:https://doi.org/10.1074/jbc.M708998200
      Gamma interferon-inducible thiol reductase (GILT) is an enzyme involved in the initial steps of antigen processing and presentation. Recently we have shown that GILT is also expressed in mouse T cells, where it exerts an inhibitory role on T cell activation. In this study, we identified mitochondrial manganese superoxide dismutase (SOD2) as one of the key intermediaries affected by GILT expression in fibroblasts. Expression and activity of SOD2 is reduced in the absence of GILT because of reduced SOD2 protein stability. The forced increase in SOD2 expression in the absence of GILT restores fibroblast proliferation to wild-type levels. Thus, GILT appears to have a fundamental role in cellular proliferation mediated through its influence on SOD2 protein activity and expression.
      Enzymes of the thiol reductase family carry out reduction, oxidation, and isomerization of protein disulfide bonds in cytosol (for example, thioredoxin) (
      • Arnér E.
      • Holmgren A.
      ,

      Lundstrom-Ljung, J., and Holmgren, A. (1998) in Prolyl Hydroxylase, Protein-disulfide Isomerase and Other Structurally Related Proteins (Guzman, N. A., ed), pp. 297–314, New York, Dekker

      ), mitochondria (
      • Patenaude A.
      • Ven Murthy M.R.
      • Mirault M.E.
      ), endoplasmic reticulum (protein-disulfide isomerase) (

      Lundstrom-Ljung, J., and Holmgren, A. (1998) in Prolyl Hydroxylase, Protein-disulfide Isomerase and Other Structurally Related Proteins (Guzman, N. A., ed), pp. 297–314, New York, Dekker

      ), and lysosomes (gamma interferon-inducible thiol reductase, GILT).
      The abbreviations used are: GILT
      gamma interferon-inducible lysosomal thiol reductase
      WT
      wild type
      SOD2
      superoxide dismutase 2
      ROS
      reactive oxygen species
      EGF
      epidermal growth factor
      FBS
      fetal bovine serum
      MF
      mouse fibroblast cell line
      PMF
      primary mouse fibroblast
      DHE
      dihydroethidium
      PBS
      phosphate-buffered saline
      DCF-DA
      2′,7′-dichlorodihydrofluorescein diacetate
      NRF
      nuclear respiratory factor.
      2The abbreviations used are: GILT
      gamma interferon-inducible lysosomal thiol reductase
      WT
      wild type
      SOD2
      superoxide dismutase 2
      ROS
      reactive oxygen species
      EGF
      epidermal growth factor
      FBS
      fetal bovine serum
      MF
      mouse fibroblast cell line
      PMF
      primary mouse fibroblast
      DHE
      dihydroethidium
      PBS
      phosphate-buffered saline
      DCF-DA
      2′,7′-dichlorodihydrofluorescein diacetate
      NRF
      nuclear respiratory factor.
      The majority of these enzymes are functional at neutral or slightly alkaline conditions (
      • Raina S.
      • Missiakas D.
      ), they have similar three-dimensional structures, and all feature a conservative active site loop containing two cysteines in the sequence -CGPC- (
      • Roos G.
      • Garcia-Pino A.
      • Van Belle K.
      • Brosens E.
      • Wahni K.
      • Vandenbusche G.
      • Wyns L.
      • Loris R.
      • Messens J.
      ). GILT is a unique and unusual member of the thiol reductase family because its optimal enzymatic activity is at a low pH (4.5–5.5) (
      • Luster A.D.
      • Weinshank R.L.
      • Feinman R.
      • Ravetch J.V.
      ,
      • Arunachalam B.
      • Pan M.
      • Cresswell P.
      ,
      • Arunachalam B.
      • Phan U.T.
      • Geuze H.J.
      • Cresswell P.
      ) and has an atypical active site (-CGAC-).
      GILT is synthesized as a 35-kDa soluble glycoprotein precursor and is transported to the endosomal compartment via the mannose-6-P receptor pathway (
      • Phan U.T.
      • Arunachalam B.
      • Cresswell P.
      ). It is processed into the mature form (30 kDa) by proteolytic removal of N- and C-terminal peptides. The protein has an approximate molecular mass of 30 kDa and was therefore initially named IP-30 (
      • Luster A.D.
      • Weinshank R.L.
      • Feinman R.
      • Ravetch J.V.
      ). In addition to endosomal/lysosomal localization, GILT is secreted in the tissue culture medium of the GILT-expressing cell lines (
      • Arunachalam B.
      • Pan M.
      • Cresswell P.
      ,
      • Maric M.
      • Arunachalam B.
      • Uyen P.
      • Dong C.
      • Garrett W.S.
      • Cannon K.S.
      • Alfonso C.
      • Karlsson L.
      • Flavell R.
      • Cresswell P.
      ), and is present in mouse sera.
      M. Maric, unpublished observations.
      3M. Maric, unpublished observations.
      GILT is constitutively expressed in professional antigen-presenting cells (APCs), but it is also inducible by pro-inflammatory cytokines such as inter-feron γ, tumor necrosis factor α, and interleukin 1β (
      • Phan U.T.
      • Arunachalam B.
      • Cresswell P.
      ).
      Using GILT–/– mice as a model, we have shown that GILT catalyzes initial unfolding of antigenic protein (protein becomes more accessible for further processing by cathepsins) and therefore facilitates protein/peptide binding to MHC class II molecules (
      • Maric M.
      • Arunachalam B.
      • Uyen P.
      • Dong C.
      • Garrett W.S.
      • Cannon K.S.
      • Alfonso C.
      • Karlsson L.
      • Flavell R.
      • Cresswell P.
      ). By changing the redox state of exogenous antigenic proteins with disulfide bonds, GILT initiates the adaptive immune response. However, we have shown that GILT is constitutively expressed in T cells and has a role in the regulation of T cell activation. This is so far the only known GILT function not related to MHC class II processing (
      • Barjaktarevic I.
      • Rahman A.
      • Radoja S.
      • Bogunovic B.
      • Vollmer A.
      • Vukmanovic S.
      • Maric M.
      ). GILT–/– T cells show increased proliferation and cytotoxic T cell activity in response to anti-CD3 stimulation. This observation suggests that GILT has a more fundamental role in cellular processes than just reduction of antigens in the antigen-processing pathway.
      We hypothesized that the effect of GILT observed in T cells, is not unique and significant only to T cells, but that fundamentally affects cellular proliferation in other cell types. In support of this hypothesis, we show that GILT–/– mouse fibroblasts also have increased levels of proliferation.
      Furthermore, we show in this study that a possible mechanism for the regulatory role of GILT in cellular proliferation involves affecting steady-state levels of a mitochondrial enzyme SOD2, involved in scavenging of reactive oxygen species (ROS) (
      • Hernandez-Saavedra D.
      • McCord J.M.
      ,
      • Cadenas E.
      • Davies K.J.A.
      ). Our data indicate that the expression, stability, and function of mitochondrial manganese SOD (SOD2) are significantly decreased in GILT–/– mouse fibroblasts. These data suggest an unexpected functional link between lysosomal and mitochondrial enzymes involved in oxido-reductive processes.

      EXPERIMENTAL PROCEDURES

      Mice, Primary Cells, and Cell Lines—C57BL/6 mice were purchased from Jackson Laboratories (Bar Harbor, ME). GILT–/– mice (
      • Maric M.
      • Arunachalam B.
      • Uyen P.
      • Dong C.
      • Garrett W.S.
      • Cannon K.S.
      • Alfonso C.
      • Karlsson L.
      • Flavell R.
      • Cresswell P.
      ) were bred and maintained in the Georgetown University pathogen-free animal facility. All mice were used at 6–12 weeks of age.
      Primary fibroblasts were isolated from GILT–/– and wild-type (WT) mouse spleens and/or ears as previously described (
      • Rahimi F.
      • Hsu K.
      • Endoh Y.
      • Geczy C.L.
      ). GILT–/– and WT SV40 large T antigen immortalized mouse fibroblast cell line was generated from GILT–/– and WT mice in Dr. Peter Creswell's laboratory at Yale University. Stable GILT–/– transfectants with mouse GILT and/or hSOD2 (kind gift from Dr. M. Williams, University of Maryland) were made using a Lipofectamine 2000 (Invitrogen) standard protocol. Both constructs were subcloned into pcDNA3.1 vector with zeocin resistance.
      Cell Proliferation—GILT–/– and WT mouse fibroblast cell lines were plated in triplicates at density of 1 × 104 cells/well (unless indicated otherwise) in 96-well flat-bottom plates and incubated at 37 °C, 5% CO2 for 30 min. Cells were pulsed with 1 μCi of [3H]dT (Amersham Biosciences/GE Healthcare) overnight, and radioactive thymidine incorporation was subsequently measured on a β scintillation counter 1450 Micro-Beta™ (Wallac, Turku, Finland).
      Western Blotting and Antibodies—Typically 5 μg and/or 2.5 μg of cell lysates in Tris-saline pH 7.5, 1% Triton X-100 containing protease inhibitor mixture were separated by SDS-PAGE. Proteins from the gel were transferred to nylon membrane (Immobilon P Millipore). Membranes were incubated overnight at 4 °C, with mouse anti-SOD2 antibody (Abcam), rabbit anti-actin antibody (Sigma), anti-Lamp-1 (Iowa hybridoma bank), and 1:5000 dilution of secondary antibodies horse-radish peroxidase-conjugated goat anti-mouse IgG and goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories). Immunoreactive bands were visualized using enhanced chemiluminescence Western Lightning™ (PerkinElmer Life Sciences). In some cases, membranes were stripped using Restore™ Western blot stripping buffer (Pierce) and incubated at room temperature for 120 min, followed by incubation with anti-actin antibody as a loading control and appropriate secondary horseradish peroxidase-conjugated antibody.
      Real-time PCR—Total RNA was isolated using TRIzol followed by RNase clean-up and treated with DNAse I. Total RNA (5 μg) was reverse-transcribed using the Superscript II RT kit and random hexamers as primers (Invitrogen). All PCR reactions were done in triplicate using ABI Prism 7700 Sequence Detector (Applied Biosystems). SOD2 and 18 S rRNA were amplified using TaqMan Universal PCR master mix (Applied Biosystems), and the average threshold cycles (Ct) of the triplicates were used to compare the relative abundance of the mRNA. Ct of 18 S rRNA was used to normalize all samples. Primers and TaqMan probe for SOD2 were designed using Primer Express Software as follows: forward: 5′-cctgctctaatcaggacccatt-3′, reverse: 5′-gtgctcccacacgtcaatcc-3′, TaqMan probe, 5′-FAM-aacaacaggccttattc-MGB-3′. Primers and probe for 18 S rRNA were previously reported (
      • Bickerstaff A.A.
      • Xia D.
      • Pelletier R.P.
      • Orosz C.G.
      ). TaqMan probe for SOD2 was designed to span the junction between two exons to avoid amplification of the genomic DNA. This was not possible for 18 S rRNA, which has no introns. The DNA contamination in these samples was excluded by amplifying control samples treated identically with the exception of the reverse transcriptase step.
      Cycloheximide Chase Experiment—5 × 106 primary mouse fibroblasts or mouse fibroblast cell lines were incubated in complete RPMI 1640 medium with cycloheximide (0.2 mg/ml). Cell lysates were processed for immunoblotting with mouse anti-SOD2 antibody at 0, 1, 6, and 24 h. Equivalent amounts of cell lysates were analyzed at each time point by Western blot.
      Superoxide Dismutase Activity Assay—2 × 107 GILT–/– and WT mouse fibroblast cells were homogenized in 1 ml of cold buffer (20 mm HEPES, 1 mm EGTA, 210 mm mannitol, and 70 mm sucrose), pH 7.2, using a Dounce homogenizer. The protein concentration of the cytoplasmic lysate was measured using the BCA™ protein assay kit (Pierce). The SOD activity assay was performed as suggested by the supplier (Cayman Chemical). This protocol is based on the tetrazolium salt for the detection of superoxide radicals generated by xanthine oxidase and hypoxanthine. One unit of SOD is defined as the amount of enzyme needed to exhibit 50% dismutation of the superoxide radical. When measuring SOD2 activity to separate cytosolic and mitochondrial SOD activity, 3 mm potassium cyanide was added to some sample wells. When measuring SOD1 activity, fractions containing SOD1 (cytosolic) were separated from fractions containing SOD2 (mitochondrial) by centrifugation at 10,000 × g as recommended by the manufacturer's protocol. The absorbance was read at 450 nm using a plate reader SpectraMax 190 (Molecular Devices), and data were analyzed using SoftMaxPro software (Molecular Devices).
      ROS Detection—Changes in intracellular ROS concentration were assessed with a method adapted from Bass et al. (
      • Bass D.A.
      • Parce J.W.
      • Dechatelet L.R.
      • Szejda P.
      • Seeds M.C.
      • Thomas M.
      ). GILT–/– and WT mouse fibroblast cell lines were starved overnight in phenol red-free medium with 1% FBS. Fresh medium and 10 μg/ml of dihydroethidium (DHE) (Molecular Probes) were added and incubated for 5 min at room temperature (preliminary kinetics experiment suggested 5 min to be the optimal incubation time). Cells were washed again in PBS, resuspended in cold PBS, 0.5% bovine serum albumin buffer, and immediately analyzed by flow cytometry.

      RESULTS

      GILT-deficient Mouse Fibroblasts Show Increased Proliferation—GILT–/– T cells display stronger proliferative responses upon stimulation with anti-CD3 (
      • Barjaktarevic I.
      • Rahman A.
      • Radoja S.
      • Bogunovic B.
      • Vollmer A.
      • Vukmanovic S.
      • Maric M.
      ). Hence, we examined whether GILT might affect cellular proliferation in general. To address this question, we tested the proliferation capabilities of fibroblast cell lines (Fig. 1A) and primary fibro-blasts (Fig. 1B) isolated from GILT–/– and GILT WT mice. Because a high density of fibroblasts may induce contact inhibition of growth (
      • Meisler A.I.
      ), we titrated the numbers of fibroblasts used in our proliferation assay. Our data indicate that irrespective of the concentration of cells (1 × 105, 1 × 104, 1 × 103) in 96-well plates incubated overnight, proliferation of GILT–/– and GILT WT fibroblasts was significantly different. GILT–/– fibroblasts persistently showed at least 2-fold increased proliferation. Both GILT–/– and GILT WT fibroblasts were generated by transfection with the large T antigen of SV40 and the increased proliferation in GILT–/– could be a consequence of random insertion of the gene coding for the SV40 large T antigen, an oncogene known to deregulate the cell cycle and cause cellular proliferation. To exclude such a possibility, we tested primary mouse fibroblasts from GILT–/– and GILT WT mouse spleens (Fig. 1B). GILT–/– primary fibroblasts also showed increased proliferation when compared with GILT WT primary fibroblasts. Identical findings were observed if fibro-blasts were treated with EGF, which is known to induce cellular proliferation (Fig. 1C). Thus, GILT–/– mouse fibroblast (MF) cell lines display increased levels of both basal and EGF-induced proliferation. This effect could be ascribed directly to the presence or absence of GILT, because the transfection of the GILT–/– fibroblast cell line with either hGILT (not shown) or mGILT reverses the phenotype (Fig. 1D). GILT–/– fibroblasts transfected with GILT have decreased proliferation levels similar to WT fibroblasts.
      Figure thumbnail gr1
      FIGURE 1GILT–/– mouse fibroblasts have increased proliferation. a, GILT WT and GILT–/– mouse ear fibroblasts (MFs) were incubated at different concentrations (103, 104, 105), in 96-well plates in serum-free RPMI medium with 1 μCi/well of [3H]dT overnight. Cells were harvested, and [3H]dT incorporation was subsequently measured on a β scintillation counter. Proliferation intensity is presented as counts per minute (cpm). *, p < 0.05 probability associated with a Student's paired t test. One representative experiment of three is shown. Results are stated as mean ± S.E. b, primary fibroblasts isolated from the spleens of GILT WT and GILT–/– mice were grown for 10 days in RPMI, 10% FBS. 24 h before the proliferation assay, they were placed in serum-free RPMI medium. The following day [3H]dT was added as in a. GILT–/– primary fibroblasts show approximately one-third higher proliferation when compared with GILT WT fibroblasts. c, GILT WT and GILT–/– MFs ± 100 ng/ml of mouse rEGF were handled the same as in a. GILT–/– MFs retain a higher proliferative ability either in the presence or absence of rEGF. d, GILT–/– MFs transfected with the empty vector pCDNA3.1 retain higher proliferation when compared with GILT WT MFs transfected with the same vector. When GILT is reconstituted into GILT–/– MFs, proliferation is decreased to the levels of GILT WT MFs. This is a representative experiment with one of three different transfectants.
      SOD2 Is Down-regulated in GILT/ Mouse Fibroblasts—Most of the GILT is sequestered in the endosomal/lysosomal compartment; therefore, its effect on fibroblast proliferation is likely to be exerted by affecting the stability and/or activity of other endosomal/lysosomal proteins. To this end, we first analyzed contents of Lamp-1/cathepsin D+ Percoll fractions in GILT–/– and GILT WT fibroblast cell lines by two-dimensional gel electrophoresis followed by qualitative and quantitative assessment of protein spots. Of the several protein spots that were differentially displayed in two samples, the most intriguing was the identification of mitochondrial SOD2 (gi:17390379). SOD2 was almost absent from heavy fractions of GILT–/– mouse fibroblasts in comparison to heavy fractions of GILT WT mouse fibroblasts (data not shown). The finding of SOD2 in these fractions is not surprising because endosomal/lysosomal and mitochondrial fractions isolated using a Percoll gradient have overlapping density. However, different levels of SOD2 in GILT WT and GILT–/– cell lines were unexpected, given their distinct subcellular localization. To exclude the possibility of altered intracellular localization of SOD2 or the different density of GILT–/– and GILT WT lysosomes, SOD2 protein levels were determined in whole cell lysates using semiquantitative Western blot. As shown in Fig. 2a, SOD2 expression was significantly lower in both the GILT–/– fibro-blast cell line and GILT–/– primary mouse fibroblasts. The lack of GILT in fibroblasts does not correlate with alteration of the expression of other mitochondrial protein such as Hsp60 (data not shown).
      Figure thumbnail gr2
      FIGURE 2SOD2, but not SOD1, protein expression and activity is decreased in GILT–/– MFs and primary fibroblasts. a, protein concentrations of cell lysates in TS, 1% Triton X with protease inhibitor mixture were determined by the BCA protein assay, and either 5 or 2.5 μg/well were loaded onto 12% SDS-PAGE gels. Proteins were transferred onto a nylon membrane, and SOD2 was detected by anti-SOD2 antibody by the chemiluminescent method. Following the detection of SOD2, the membranes were stripped and reincubated with anti-actin antibody as a loading control. b and c, MFs and PMFs, respectively, were lysed, and equal amounts of total proteins were assayed for SOD2 activity as described under “Experimental Procedures.” d, cells were prepared the same as in a except that samples were run on a 15% SDS-PAGE gel and SOD1 was detected by anti-SOD1 antibody by the chemiluminescent method.
      To test whether the activity of SOD2 is decreased in GILT–/– cells, mouse fibroblasts were tested by an in vitro SOD assay based on the tetrazolium salt for detection of super-oxide radicals generated by xanthine oxidase and hypoxanthine. The lysates of cells contain both cytosolic SOD1 and mitochondrial SOD2; thus, to distinguish between these two enzymes, we added potassium cyanide that inhibits the activity of SOD1 but not of SOD2. Our data clearly show that the GILT–/– mouse fibroblast cell line (Fig. 2b) as well as primary mouse fibroblasts (Fig. 2c) have decreased activity of SOD2. Although the role of SOD2 appears to be intrinsic for cellular survival, cytoplasmic copper/zinc SOD (SOD1) is far more abundant in the cell, accounting for 90% of the total SOD activity (
      • Aquilano K.
      • Vigilanza P.
      • Rotilio G.
      • Ciriolo M.R.
      ,
      • Chance B.
      • Sies H.
      • Boveris A.
      ). Thus, we tested the protein expression of SOD1 in GILT–/– and GILT WT fibroblast cell line and primary fibro-blasts (Fig. 2d). Whereas SOD2 levels were significantly decreased in the GILT–/– fibroblast cell line and primary fibro-blasts, SOD1 levels remain similar in GILT–/– and GILTWT cells. Furthermore, the activity of SOD1 in the GILT–/– fibro-blast cell line and primary fibroblasts also remains similar (supplemental Fig. S1). Thus, SOD2 levels and activity are specifically affected by GILT.
      Reconstitution of GILT Recovers SOD2 Activity—The difference in SOD2 protein expression and activity between GILT–/– and GILT WT cells could be a consequence of GILT presence or a random event. To distinguish between these possibilities, we assessed SOD2 activity in GILT–/– fibroblasts reconstituted with mouse GILT. As shown in Fig. 3, the SOD2 activity of transfectants is increased to the level of WT fibro-blasts. These data suggest that the presence of GILT correlates with the decreased expression and activity of mitochondrial enzyme SOD2.
      Figure thumbnail gr3
      FIGURE 3SOD2 activity is reconstituted in GILT–/– MFs transfected with mouse GILT. GILT–/– MFs were transfected by the Lipofectamine method with the pCDNA3.1Zeo.mGILT construct. GILT-expressing clones were selected in selective medium containing Zeocin. SOD2 activity was tested in untransfected GILT–/–, GILT WT, and GILT–/– MFs transfected with mGILT as described under “Experimental Procedures.” GILT–/– MFs transfected with vector-only have low SOD2 activity as GILT–/– MFs (data not shown).
      SOD2 Protein Stability Is Decreased in GILT/ Mouse Fibroblasts—Our data indicate that the lack of GILT correlates with lower activity and the amount of SOD2 in the fibroblasts. The decreased expression of SOD2 in GILT–/– fibroblasts may be due to down-regulation of Sod2 gene expression and/or due to instability of SOD2 protein in the absence of GILT. Thus, we first measured relative levels of SOD2 mRNA in GILT–/– and GILT WT mouse fibroblast cell line by quantitative PCR. As shown in Fig. 4a, mRNA SOD2 levels were only moderately decreased in GILT–/– mouse fibro-blasts when compared with GILT WT cells.
      Figure thumbnail gr4
      FIGURE 4SOD 2 mRNA is less abundant, and protein stability is decreased in GILT–/– MFs. a, SOD2 gene expression is ∼35% decreased in GILT–/– MF SOD2 when compared with GILT WT MFs. The real time/quantitative PCR reaction was performed using Taqman Universal PCR (Applied Biosystems). This is a representative one of three experiments. b, MF lysates were processed for immunoblotting with mouse anti-SOD2 antibody after 0, 1, 6, and 24 h of incubation of cells with 200 μg/ml of cycloheximide. Membranes were stripped and incubated with anti-actin antibody as a loading control. The film was scanned by phosphorimager and bands quantified. c, PMF lysates were treated the same as MFs were in b.
      However, this very moderate decrease of SOD2 at the level of gene expression does not preclude the possibility that SOD2 protein is less stable in GILT–/– fibroblasts. To test the stability of SOD2 protein in GILT–/– and WT mouse fibro-blasts, we treated both GILT–/– and GILT WT fibroblasts with the protein synthesis inhibitor cycloheximide. Cell aliquots were collected at different time points up to 48 h of treatment with cycloheximide. The protein concentration of each sample was determined, and equal amounts of protein were loaded per well. Data shown in Fig. 4b indicate that SOD2 stability is reduced in the GILT–/– mouse fibroblast cell line as well as in GILT–/– primary mouse fibroblasts (Fig. 4c), although the amounts of SOD2 between the GILT–/– cell line and primary cells differ. The observed difference is likely to originate from different lengths that cells were kept under cell culture conditions. It is well known that cell culture and hyperoxic environment impose oxidative stress (
      • Halliwell B.
      • Gutteridge J.M.C.
      ). The cell line has been kept in tissue culture much longer than primary fibroblasts and therefore has been exposed to additional extracellular oxidative stress that may have destabilized SOD2 further than in primary fibro-blasts. Our data suggest a possible dual regulation of GILT on SOD2 expression: at the level of gene expression and also through its effect on the stability of SOD2 protein.
      GILT-deficient Mouse Fibroblast Cell Line Has Increased Superoxide Anion Levels—As the stability and function of an important antioxidant SOD2 is decreased in GILT–/– mouse fibroblasts, it is possible that the levels of endogenous ROS, particularly superoxide anion, are increased. It has been shown that a moderate increase of endogenous ROS can trigger signaling that induces cell proliferation (
      • Sun Y.
      • Oberley L.W.
      ,
      • Jackson S.H.
      • Devadas S.
      • Kwon J.
      • Pinto L.A.
      • Williams M.S.
      ,
      • Laurent A.
      • Nicco C.
      • Chereau C.
      • Goulvestre C.
      • Alexandre J.
      • Alves A.
      • Levy E.
      • Goldwasser F.
      • Panis Y.
      • Soubrane O.
      • Weill B.
      • Batteux F.
      ). Therefore, we tested levels of superoxide anion and ROS in GILT–/– and GILT WT fibroblast cell lines. We used DHE (Fig. 5) and 2′,7′-dichlorodihydrofluorescein diacetate (DCF-DA) (data not shown) oxidation-sensitive dyes to detect intracellular ROS. DHE is recognized and widely used as a probe relatively specific for O2·¯ and shows little oxidation by H2O2, ONOO, or HOCl, while DCF-DA is less specific and is used to assay cellular peroxides, superoxide anions. DCF-DA reacts even faster with various cellular radicals (RO2·¯, RO, NO2·, CO3·, OH, and ONOO); thus, it is not a very specific probe and easily oxidizes (
      • Halliwell B.
      • Gutteridge J.M.C.
      ,
      • Tarpey M.M.
      • Fridovich I.
      ). Fibro-blasts were incubated with DHE in RPMI for 5 min and immediately analyzed by flow cytometry. Our data indicate that GILT–/– mouse fibroblasts have higher levels of intracellular O2·¯ (Fig. 5a). However, transfection with either hSOD2 (Fig. 5b) or mGILT (Fig. 5c) decreases levels of O2·¯ to the level seen in GILT WT fibroblasts, suggesting that this is the component of the mechanism of increased cellular proliferation.
      Figure thumbnail gr5
      FIGURE 5Reconstitution of mGILT and hSOD2 in GILT-deficient mouse fibroblasts decreases endogenous ROS. Cells were starved overnight in phenol red-free medium with 1% FBS. 5 μg/ml of HE was added to cells and incubated for 5 min. Cells were washed again in PBS, resuspended in PBS, 0.5% bovine serum albumin buffer, and immediately analyzed by flow cytometry. Y-axis, cell count; X-axis, FL-2 channel. GILT WT MFs incubated with DHE (gray-filled dotted line), GILT–/– fibroblasts incubated with DHE (thick black line), transfectants GILT/tf, pCDNA3.1 (a), GILT/tf, mGILT (b), and GILT/tf, hSOD2 (c), thin dotted line.
      Restoring SOD2 Levels in GILT/ Fibroblasts Restores Proliferation—We have shown that GILT affects fibroblast proliferation and the levels of SOD2 expression. The key question is whether these two phenotypes are related. If SOD2 expression is the mechanism of GILT-regulated fibroblast proliferation, then reconstitution of SOD2 levels in GILT–/– cells (Fig. 6a) should reverse proliferation to GILT WT levels. Indeed, the proliferation of GILT–/– fibroblasts, reconstituted with human SOD2 (two independently made transfectants show the same phenotype) was markedly lower than in GILT–/– and similar to GILT WT fibroblasts. In addition, our data indicate that SOD2 activity is reconstituted in transfectant cells (Fig. 6b). Thus, regulation of SOD2 expression and activity appears to be the major mechanism of GILT-induced inhibition of fibroblast proliferation.
      Figure thumbnail gr6
      FIGURE 6Proliferation is down-modulated in GILT–/– MFs reconstituted with hSOD2. GILT–/– MFs were reconstituted with the human SOD2 gene in the pCDNA3.1 vector. a, GILT WT and GILT–/– and GILT–/– hSOD2 transfectants were incubated overnight with 1 μCi of [3H]dT, and [3H]dT incorporation was subsequently measured on a β scintillation counter. The proliferation intensity is presented as counts per minute (cpm). b, equal amounts of total protein of lysates from GILT–/– MFs reconstituted with hSOD2 were assayed for SOD2 activity as described under “Experimental Procedures.”

      DISCUSSION

      In this study we demonstrate that GILT-deficient mouse fibroblasts have increased levels of proliferation. Surprisingly, our data indicate that GILT deficiency affects the stability and activity of a major antioxidant mitochondrial enzyme SOD2. GILT–/– fibroblasts have 2–4-fold lower SOD2 activity because of a decreased half-life of SOD2. The partial loss of SOD2 activity leads to the rise in superoxide anion as measured by DHE dye and possibly other ROS produced by mitochondrial oxidative phosphorylation. Reconstitution of GILT–/– fibroblasts with GILT reverts the SOD2 activity and the proliferation to levels closely resembling that of GILT WT fibroblasts. These data suggest that the presence of GILT is necessary to maintain levels of mitochondrial enzyme SOD2 and cellular proliferation. The transfection of human SOD2 into GILT–/– fibroblasts restored proliferation to WT levels, suggesting that SOD2 is a key step linking GILT and its effects on cellular proliferation.
      SOD2 is an antioxidant enzyme responsible for the dismutation of the superoxide radical into hydrogen peroxide (
      • Halliwell B.
      • Gutteridge J.M.C.
      ,
      • Fridovich I.
      ). Although superoxide radicals have short life and also spontaneously dismutate into hydrogen peroxide, deletion of the Sod2 gene in mice is perinatal lethal (
      • Lebovitz R.M.
      • Zhang H.
      • Vogel H.
      • Cartwright Jr., J.
      • Dionne L.
      • Lu N.
      • Huang S.
      • Matzuk M.M.
      ,
      • Li Y.
      • Huang T.T.
      • Carlson E.J.
      • Melov S.
      • Ursell P.C.
      • Olson J.L.
      • Noble L.J.
      • Yoshimura M.P.
      • Berger C.
      • Chan P.H.
      • Wallace D.C.
      • Epstein C.J.
      ), severely reduces life span in Drosophila (
      • Duttaroy A.
      • Paul A.
      • Kundu M.
      • Belton A.
      ), and mice genetically modified to lack SOD2 in specific tissues show various pathologies (
      • Friedman J.S.
      • Rebel V.I.
      • Derby R.
      • Bell K.
      • Huang T.T.
      • Kuypers F.A.
      • Epstein C.J.
      • Burakoff S.J.
      ,
      • Huang T.T.
      • Carlson E.J.
      • Kozy H.M.
      • Mantha S.
      • Goodman S.I.
      • Ursell P.C.
      • Epstein C.J.
      ,
      • Samper E.
      • Nicholls D.G.
      • Melov S.
      ). In addition to SOD2 located in the mitochondria, there are two additional forms of SOD in cells of aerobic organisms, namely SOD1 and SOD3. SOD1 is primarily cytosolic (although a fraction is also found in the intermembrane space of mitochondria) (
      • Sturtz L.A.
      • Diekert K.
      • Jensen L.T.
      • Lill R.
      • Culotta V.C.
      ), and SOD3 is an extracellular enzyme. SOD1 constitutes up to 90% of the total activity in most cells and tissues and is considered a principal ROS scavenger in the cell. Despite that, genetic inactivation of SOD1 results in a relatively mild pheno-type (
      • Reaume A.G.
      • Elliott J.L.
      • Hoffman E.K.
      • Kowall N.W.
      • Ferrante R.J.
      • Siwek D.F.
      • Wilcox H.M.
      • Flood D.G.
      • Beal M.F.
      • Brown Jr., R.H.
      • Scott R.W.
      • Snider W.D.
      ). Therefore, we tested the levels and the activity of SOD1 in our model system. The levels of SOD1 were unaffected by the absence of GILT. This finding is in agreement with previous findings that cytosolic SOD1 cannot compensate for the loss of mitochondrial SOD2 and vice versa (
      • Van Remmen H.
      • Williams M.D.
      • Guo Z.
      • Estlack L.
      • Yang H.
      • Carlson E.J.
      • Epstein C.J.
      • Huang T.T.
      • Richardson A.
      ,
      • Huang T.-T.
      • Yasunami M.
      • Carlson E.J.
      • Gillespie A.M.
      • Reaume A.G.
      • Hoffman E.K.
      • Chan P.H.
      • Scott R.W.
      • Epstein C.J.
      ). Thus, despite additional systems for the removal of ROS (SOD1, catalase, glutathione peroxidase) (
      • Gilbert D.L.
      • Colton C.A.
      ), SOD2 plays a distinct and crucial role in the removal of ROS generated by the respiratory chain in mitochondria. Therefore, SOD2 is an important regulator of endogenous ROS generation (
      • Van Remmen H.
      • Williams M.D.
      • Guo Z.
      • Estlack L.
      • Yang H.
      • Carlson E.J.
      • Epstein C.J.
      • Huang T.T.
      • Richardson A.
      ).
      Several studies indicated that the balance between ROS production and antioxidant defenses is a factor influencing cell growth and differentiation (
      • Duttaroy A.
      • Paul A.
      • Kundu M.
      • Belton A.
      ). ROS were shown to be involved in processes of cell growth and proliferation as well as apoptosis in various cell types (
      • Bravard A.
      • Petridis F.
      • Luccioni C.
      ,
      • Burdon R.H.
      ). Because superoxide anion is the major target for SOD2 action, we used the fairly specific fluorescent probe DHE to compare the overall levels of O2·¯ in GILT–/– versus GILT WT fibroblasts. Our data indicate that O2·¯ levels are increased in GILT–/– cells. However, one has to be aware of caveats of this approach. DHE is a non-fluorescent cell permeant that undergoes oxidation to a fluorescent product 2-hydroxyethidium, that intercalates into nuclear DNA and shows strong fluorescence upon interaction with O2·¯ (
      • Zhao H.
      • Kalivendi S.
      • Zhang H.
      • Joseph J.
      • Nithipatikom K.
      • Vasquez-Vivar J.
      • Kalyanaraman B.
      ). It shows little oxidation by H2O2, ONOO, or HOCl. DHE can spontaneously oxidize or be oxidized by singlet O2 and if cytochrome c is released into the cytosol by mitochondria (
      • Green D.R.
      • Reed J.C.
      ), it can also oxidize DHE (
      • Halliwell B.
      • Gutteridge J.M.C.
      ,
      • Tarpey M.M.
      • Fridovich I.
      ). Although we think the O2·¯ is most likely the source of DHE fluorescence in our model system, we cannot exclude other above-mentioned factors. A more detailed study is underway to dissect further details of moderately increased oxidative stress in GILT–/– fibroblasts. SOD2 activity is directly linked to the degree of cell differentiation and inversely related to proliferation in several different systems (
      • Bize I.B.
      • Oberley L.W.
      • Morris H.P.
      ,
      • Oberley L.W.
      • Oberley T.D.
      ,
      • Allen R.G.
      ,
      • Church S.L.
      • Grant J.W.
      • Ridnour L.A.
      • Oberley L.W.
      • Swanson P.E.
      • Meltzer P.S.
      • Trent J.M.
      ). Most types of tumor cell lines have reduced levels of SOD2 in comparison to their normal cell counterparts. Several studies have shown that tumorigenicity of various tumor cell lines is decreased post-transfection with SOD2 cDNA (
      • Li S.
      • Yan T.
      • Yang J.Q.
      • Oberley T.D.
      • Oberley L.W.
      ,
      • Bravard A.
      • Ageron-Blanc A.
      • Alvarez S.
      • Drane P.
      • le Rhun Y.
      • Paris F.
      • Luccioni C.
      • May E.
      ,
      • Mukhopaday S.
      • Das S.K.
      • Mukherjee S.
      ). We hypothesize that in our system, moderately increased endogenous ROS in GILT–/– fibroblasts caused by lower SOD2 activity stimulate signaling pathways involved in the regulation of cellular proliferation.
      The precise mechanism of GILT-SOD2 interaction is currently under study. While GILT resides in the endosomal compartment, SOD2 is a resident protein of mitochondria. A remote possibility exists that these two proteins may interact directly either during their transport through the endoplasmic reticulum and Golgi or perhaps transiently co-localize in yet another vesicular compartment. Intracellular immunofluorescence confocal microscopy (supplemental Fig. S2) showed that the majority of GILT and SOD2 reside in their respective compartments (lysosomes and mitochondria). However, the resolution of this method does not completely exclude the possibility that small fractions of these molecules co-localize.
      The alternative possibility is that GILT and SOD2 interact indirectly. GILT may limit the source of cysteine necessary for glutathione synthesis and that may cause an overall redox balance alteration that may affect the activity of SOD2. Currently we do not know the exact mechanism of regulation of SOD2 expression by GILT. However, it is possible that increased amounts of O2·¯ act as scavengers for nitric oxide (NO) and turn it into a shorter-lived and less reactive peroxynitrite (ONOO.) NO has been indicated to stimulate SOD2 expression and activity (
      • Keller T.
      • Pleskova M.
      • McDonald M.C.
      • Thiemermann C.
      • Pfeilschifter J.
      • Beck K.-F.
      ,
      • Li W.
      • Jue T.
      • Edwards J.
      • Wang X.
      • Hintze T.H.
      ) via the Ras/Erk1/2 pathway (
      • Scorziello A.
      • Santillo M.
      • Adornetto A.
      • Dell'Aversano C.
      • Sirabella R.
      • Damiano S.
      • Canzoniero L.M.T.
      • Di Renzo G.F.
      • Annunziato L.
      ). The lower availability of NO may contribute to decreased SOD2 expression. Either alternatively or perhaps in addition, through interaction of an as yet unidentified signaling pathway, or through the altered overall redox status of the cell, GILT may affect the activity of transcription factors such as nuclear respiratory factors 1 and 2 (NRF-1 and NRF-2) that are involved in expression of mitochondrial proteins. The binding of NRF-2 to DNA is inhibited by ROS because of oxidation of essential thiol groups; thus, redox changes within the cell may affect the supply of SOD2 (
      • Scarpulla R.C.
      ). NRF transcription factors are not the sole potential regulators of SOD2 expression; increased oxidative stress has been implicated in cellular signaling and activation of NFκB, AP-1, and AP-2 (
      • Das K.C.
      • Lewis-Molock Y.
      • White C.W.
      ,
      • Das K.C.
      • Lewis-Molock Y.
      • White C.W.
      ,
      • Dhar S.K.
      • Lynn B.C.
      • Daosukho C.
      • St. Clair D.K.
      ,
      • Cho H.-Y.
      • Reddy S.P.
      • Kleeberger S.R.
      ). All these transcription factors have been found to be redox-sensitive and are involved in SOD2 transcription/expression. Therefore GILT might at least in part affect the expression of SOD2 through this mechanism, because we have seen small decreases of SOD2 mRNA. In addition, GILT may affect expression of other proteins that influence the stability of the SOD2 protein.
      The overall picture is that at the steady state in normal fibro-blasts, both GILT and SOD2 are expressed. The role of SOD2 is to scavenge superoxide anions (generated in the respiratory chain reaction within mitochondria) and dismutate them into peroxide that will eventually be turned into water through catalase or peroxidase action. In GILT-deficient cells, less SOD2 protein is expressed, and its activity is decreased as well. As a result, increased levels of ROS (superoxide anions and hydrogen peroxide) exist in these cells. Consequently, certain signaling pathways may have been activated in the GILT-deficient cells that are negatively regulated by GILT in the GILT WT cells.
      Our data reveal an unexpected correlation between lysosomal thiol reductase and mitochondrial reactive oxygen species scavenger. Both GILT and SOD2 are involved in regulation of cellular proliferation of fibroblasts. Overall, the importance of this study lies in the fact that GILT–/– cells can be used as a model to study novel intracellular communication pathways between lysosomal and mitochondrial compartments. In addition, our results support the role of reactive oxygen species, not solely as factors with negative consequence for cell survival but also as mediators of cellular signaling. In many diseases, redox processes gone “awry” are cause for further damage of cellular structures and tissues. Learning how to control and modify these events may lead to prevention of damage or invention of drugs that control negative consequences of our aerobic existence.

      Acknowledgments

      We thank Dr. Karen Creswell and Michelle Lombard for technical assistance with flow cytometry and Drs. R. Pad Padmanabhan (Georgetown University), M. Williams (University of Maryland), and S. Vukmanovic (Children's National Medical Center, Washington, D. C.) for helpful discussions and critical reading of the manuscript.

      References

        • Arnér E.
        • Holmgren A.
        Eur. J. Biochem. 2000; 267: 6102-6109
      1. Lundstrom-Ljung, J., and Holmgren, A. (1998) in Prolyl Hydroxylase, Protein-disulfide Isomerase and Other Structurally Related Proteins (Guzman, N. A., ed), pp. 297–314, New York, Dekker

        • Patenaude A.
        • Ven Murthy M.R.
        • Mirault M.E.
        J. Biol. Chem. 2004; 279: 27302-27314
        • Raina S.
        • Missiakas D.
        Annu. Rev. Microbiol. 1997; 51: 179-202
        • Roos G.
        • Garcia-Pino A.
        • Van Belle K.
        • Brosens E.
        • Wahni K.
        • Vandenbusche G.
        • Wyns L.
        • Loris R.
        • Messens J.
        J. Mol. Biol. 2007; 368: 800-811
        • Luster A.D.
        • Weinshank R.L.
        • Feinman R.
        • Ravetch J.V.
        J. Biol. Chem. 1988; 263: 12036-12043
        • Arunachalam B.
        • Pan M.
        • Cresswell P.
        J. Immunol. 1998; 160: 5797-5806
        • Arunachalam B.
        • Phan U.T.
        • Geuze H.J.
        • Cresswell P.
        Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 745-750
        • Phan U.T.
        • Arunachalam B.
        • Cresswell P.
        J. Biol. Chem. 2000; 275: 25907-25914
        • Maric M.
        • Arunachalam B.
        • Uyen P.
        • Dong C.
        • Garrett W.S.
        • Cannon K.S.
        • Alfonso C.
        • Karlsson L.
        • Flavell R.
        • Cresswell P.
        Science. 2001; 294: 1361-1365
        • Barjaktarevic I.
        • Rahman A.
        • Radoja S.
        • Bogunovic B.
        • Vollmer A.
        • Vukmanovic S.
        • Maric M.
        J. Immunol. 2006; 177: 4369-4375
        • Hernandez-Saavedra D.
        • McCord J.M.
        Cancer Res. 2003; 63: 159-163
        • Cadenas E.
        • Davies K.J.A.
        Free Radic. Biol. Med. 2000; 29: 222-230
        • Rahimi F.
        • Hsu K.
        • Endoh Y.
        • Geczy C.L.
        Febs J. 2005; 272: 2811-2827
        • Bickerstaff A.A.
        • Xia D.
        • Pelletier R.P.
        • Orosz C.G.
        J. Immunol. 2000; 164: 5132-5139
        • Bass D.A.
        • Parce J.W.
        • Dechatelet L.R.
        • Szejda P.
        • Seeds M.C.
        • Thomas M.
        J. Immunol. 1983; 130: 1910-1917
        • Meisler A.I.
        J. Cell Sci. 1973; 12: 847-859
        • Aquilano K.
        • Vigilanza P.
        • Rotilio G.
        • Ciriolo M.R.
        Faseb J. 2006; 20: 1683-1685
        • Chance B.
        • Sies H.
        • Boveris A.
        Physiol. Rev. 1979; 59: 527-605
        • Halliwell B.
        • Gutteridge J.M.C.
        Free Radicals in Biology and Medicine. Oxford University Press, Oxford, UK2007: 851
        • Sun Y.
        • Oberley L.W.
        Free Radic. Biol. Med. 1996; 21: 335-348
        • Jackson S.H.
        • Devadas S.
        • Kwon J.
        • Pinto L.A.
        • Williams M.S.
        Nat. Immunol. 2004; 5: 818-827
        • Laurent A.
        • Nicco C.
        • Chereau C.
        • Goulvestre C.
        • Alexandre J.
        • Alves A.
        • Levy E.
        • Goldwasser F.
        • Panis Y.
        • Soubrane O.
        • Weill B.
        • Batteux F.
        Cancer Res. 2005; 65: 948-956
        • Tarpey M.M.
        • Fridovich I.
        Circ. Res. 2001; 89: 224-236
        • Fridovich I.
        Annu. Rev. Biochem. 1995; 64: 97-112
        • Lebovitz R.M.
        • Zhang H.
        • Vogel H.
        • Cartwright Jr., J.
        • Dionne L.
        • Lu N.
        • Huang S.
        • Matzuk M.M.
        Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 9782-9787
        • Li Y.
        • Huang T.T.
        • Carlson E.J.
        • Melov S.
        • Ursell P.C.
        • Olson J.L.
        • Noble L.J.
        • Yoshimura M.P.
        • Berger C.
        • Chan P.H.
        • Wallace D.C.
        • Epstein C.J.
        Nat. Genet. 1995; 11: 376-381
        • Duttaroy A.
        • Paul A.
        • Kundu M.
        • Belton A.
        Genetics. 2003; 165: 2295-2299
        • Friedman J.S.
        • Rebel V.I.
        • Derby R.
        • Bell K.
        • Huang T.T.
        • Kuypers F.A.
        • Epstein C.J.
        • Burakoff S.J.
        J. Exp. Med. 2001; 193: 925-934
        • Huang T.T.
        • Carlson E.J.
        • Kozy H.M.
        • Mantha S.
        • Goodman S.I.
        • Ursell P.C.
        • Epstein C.J.
        Free Radic. Biol. Med. 2001; 31: 1101-1110
        • Samper E.
        • Nicholls D.G.
        • Melov S.
        Aging Cell. 2003; 2: 277-285
        • Sturtz L.A.
        • Diekert K.
        • Jensen L.T.
        • Lill R.
        • Culotta V.C.
        J. Biol. Chem. 2001; 276: 38084-38089
        • Reaume A.G.
        • Elliott J.L.
        • Hoffman E.K.
        • Kowall N.W.
        • Ferrante R.J.
        • Siwek D.F.
        • Wilcox H.M.
        • Flood D.G.
        • Beal M.F.
        • Brown Jr., R.H.
        • Scott R.W.
        • Snider W.D.
        Nat. Genet. 1996; 13: 43-47
        • Van Remmen H.
        • Williams M.D.
        • Guo Z.
        • Estlack L.
        • Yang H.
        • Carlson E.J.
        • Epstein C.J.
        • Huang T.T.
        • Richardson A.
        Am. J. Physiol. Heart Circ. Physiol. 2001; 281: H1422-H1432
        • Huang T.-T.
        • Yasunami M.
        • Carlson E.J.
        • Gillespie A.M.
        • Reaume A.G.
        • Hoffman E.K.
        • Chan P.H.
        • Scott R.W.
        • Epstein C.J.
        Arch. Biochem. Biophys. 1997; 344: 424-432
        • Gilbert D.L.
        • Colton C.A.
        Reactive Oxygen Species in Biological Systems: An Interdisciplinary Approach. Kluwer Academic/Plenum Publishers, NY1999
        • Bravard A.
        • Petridis F.
        • Luccioni C.
        Free Radic Biol. Med. 1999; 26: 1027-1033
        • Burdon R.H.
        Free Radic Biol. Med. 1995; 18: 775-794
        • Green D.R.
        • Reed J.C.
        Science. 1998; 281: 1309-1312
        • Bize I.B.
        • Oberley L.W.
        • Morris H.P.
        Cancer Res. 1980; 40: 3686-3693
        • Oberley L.W.
        • Oberley T.D.
        Mol. Cell Biochem. 1988; 84: 147-153
        • Allen R.G.
        Proc. Soc. Exp. Biol. Med. 1991; 196: 117-129
        • Church S.L.
        • Grant J.W.
        • Ridnour L.A.
        • Oberley L.W.
        • Swanson P.E.
        • Meltzer P.S.
        • Trent J.M.
        Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 3113-3117
        • Li S.
        • Yan T.
        • Yang J.Q.
        • Oberley T.D.
        • Oberley L.W.
        Cancer Res. 2000; 60: 3927-3939
        • Bravard A.
        • Ageron-Blanc A.
        • Alvarez S.
        • Drane P.
        • le Rhun Y.
        • Paris F.
        • Luccioni C.
        • May E.
        Carcinogenesis. 2002; 23: 705-711
        • Mukhopaday S.
        • Das S.K.
        • Mukherjee S.
        J. Biomed. Biotechnol. 2004; 4: 195-202
        • Keller T.
        • Pleskova M.
        • McDonald M.C.
        • Thiemermann C.
        • Pfeilschifter J.
        • Beck K.-F.
        Nitric Oxide. 2003; 9: 183-193
        • Li W.
        • Jue T.
        • Edwards J.
        • Wang X.
        • Hintze T.H.
        Am. J. Physiol. Heart Circ. Physiol. 2003; 286: H47-H54
        • Scorziello A.
        • Santillo M.
        • Adornetto A.
        • Dell'Aversano C.
        • Sirabella R.
        • Damiano S.
        • Canzoniero L.M.T.
        • Di Renzo G.F.
        • Annunziato L.
        J. Neurochem. 2007; 103: 1472-1480
        • Scarpulla R.C.
        Biochim. Biophys. Acta. 2002; 1576: 1
        • Das K.C.
        • Lewis-Molock Y.
        • White C.W.
        Am. J. Physiol. 1995; 269: L588-L602
        • Das K.C.
        • Lewis-Molock Y.
        • White C.W.
        Mol. Cell Biochem. 1995; 148: 45-57
        • Dhar S.K.
        • Lynn B.C.
        • Daosukho C.
        • St. Clair D.K.
        J. Biol. Chem. 2004; 279: 28209-28219
        • Cho H.-Y.
        • Reddy S.P.
        • Kleeberger S.R.
        Antiox. Redox Signal. 2006; 8: 76-87
        • Zhao H.
        • Kalivendi S.
        • Zhang H.
        • Joseph J.
        • Nithipatikom K.
        • Vasquez-Vivar J.
        • Kalyanaraman B.
        Free Radic Biol. Med. 2003; 34: 1359-1368