Zn-superoxide dismutase

Free Radical Biology & Medicine, Vol. 31, No. 9, pp. 1084 –1089, 2001 Copyright © 2001 Elsevier Science Inc. Printed in the USA. All rights reserved 0891-5849/01/$–see front matter

PII S0891-5849(01)00691-8

Original Contribution MODULATION OF NO AND CYTOKINES IN MICROGLIAL CELLS BY CU/ZN-SUPEROXIDE DISMUTASE SU-CHIEN CHANG,*1 MING-CHING KAO,*1 MING-TSZE FU,†1

and

CHI-TSAI LIN†

†

*Department of Biochemistry, National Defense Medical Center, Taipei, Taiwan; and Institute of Marine Biotechnology, National Taiwan Ocean University, Keelung, Taiwan (Received 13 March 2001; Accepted 2 August 2001)

Abstract—The activation of microglial cells in response to neuropathological stimuli is one of the prominent features of human neurodegenerative diseases. Cytokines such as IL-1␤ and TNF-␣ and inflammation-related enzymes such as inducible nitric oxide synthase are usually induced during the activation of microglial cells. We investigated the modulation of the activation of microglial cell by transfecting a Cu/Zn-SOD cDNA into BV-2 cells. Parental and transfected BV-2 cells were then subjected to LPS stimulation. The results showed that in Cu/Zn-SOD–transfected BV-2 cells, the expression and activity of Cu/Zn-SOD increased. On the other hand, upon activation by LPS, these cells produced less NO, IL-1␤, and TNF-␣ than the parental microglial cells. This finding suggests that superoxide may be an early signal triggering the induction of cytokines and that the transfected Cu/Zn-SOD may provide a neuroprotective function via suppression of microglial activation. In addition, this approach may provide a rationale for the development of treatments for neurodegenerative diseases. © 2001 Elsevier Science Inc. Keywords—Cu/Zn-SOD, Microglia, BV-2 cell, LPS, NO, IL-1␤, TNF-␣, Free radicals

INTRODUCTION

sult in a delay of neurodegenerative processes [4,5]. It is therefore important to reduce or prevent oxidative injury to the brain. One of the antioxidative mechanisms in the cell is superoxide dismutase (SOD). SOD catalyzes the dismutation of the superoxide ion (O2⫺), which reacts easily with excessive amounts of diffusible NO and can then be converted to peroxynitrite [6]. Therefore, SOD helps prevent the destructive oxidative processes in cells. In this study, we assess the role of SOD in neuron cells by overexpressing a Cu/Zn-SOD in the BV-2 microglial cell. The production of NO and proteins (TNF-␣ and IL-1␤) generated after lipopolysaccharide (LPS) activation of the microglial cells was measured. Here we report for the first time that transfected Cu/Zn-SOD protected against the induction of NO, TNF-␣, and IL-1␤ in the activated microglial cells.

The molecular mechanisms leading to human neurodegenerative diseases such as prion diseases, Alzheimer’s, Parkinson’s, and Huntington’s diseases are not well understood. Two of the major neuropathological changes that characterize these diseases are neuronal cell apoptosis and the presence of activated microglial cells [1]. The activation of microglial cells in response to pathological stimuli is associated with the production of various cytokines, such as TNF-␣ and IL-1␤, and the induction of inflammation-related enzymes such as inducible nitric oxide synthase (iNOS) [2]. Proper activation of microglial cells may be beneficial for the neuroprotective processes, but an abnormal activation of microglia, such as caused by oxidative stresses, may become dangerous by increasing the inflammatory burden [3]. Studies using anti-inflammatory drugs and multiple antioxidants also demonstrated that reducing oxidative damage might re-

MATERIALS AND METHODS

Address correspondence to: Dr. Chi-Tsai Lin, National Taiwan Ocean University, Institute of Marine Biotechnology, 2 Pei-Ning Road, Keelung 202, Taiwan; Tel: ⫹886 (2) 2462-2192, ext. 5513; Fax: ⫹886 (2) 2462-2320; E-Mail: [email protected]. 1 Chang, S.-C.; Kao, M.-C.; and Fu, M.-T. contributed equally to this paper.

Cloning of wild-type (wt) and frame-shift (fs) Cu/ Zn-SOD gene from pineapple A full-length Cu/Zn-SOD cDNA was cloned from a pineapple (fruit) cDNA library using PCR (polymerase 1084

Cu/Zn-SOD decreases microglial activation

chain reaction) and RACE (rapid amplification of cDNA end). This Cu/Zn-SOD cDNA comprises a complete open reading frame coding for 152 amino acid residues (Accession no. AJ250667) [7]. The deduced amino acid sequence showed higher homology (78 – 86%) to the sequence of the cytosolic SOD than that of other sources. The residues required for coordinating copper and zinc are conserved as they are among all reported Cu/Zn-SOD sequences. Two conservative sense primers, that is, a wt (wild-type) primer (5⬘- GAA TTC GAT GGT GAA GGC TGT TGC TGT GC-3⬘) and a fs (frame-shiftmutant) primer (5⬘-GAA TTC GAT GAG GCA CAA TCT ACT TCA CCC AAG-3⬘), and a conservative antisense primer (5⬘-CTC GAG TTA CCC CTG CAG TCC GAT AAT TCC-3⬘) were designed and synthesized according to the reported 820 bp cDNA sequence from pineapple [7]. To a 0.5 ml microtube containing 0.1 ␮g of pineapple blunted cDNA as template, 10 pmol of 3⬘-antisense primer and 10 pmol of either wt primer or fs primer were added. Wt and fs DNA fragments (0.45 kb) were amplified by PCR, ligated with pCR2.1 (Invitrogen BV, Groningen, The Netherlands), and then transformed into E. coli TOPO 10 (Invitrogen) host. Both positive clones were selected by hybridization with 32P-labeled Cu/ZnSOD DNA as the probe, and the plasmid DNA was prepared. A suitable amount of the plasmid DNA was digested with EcoRI and XhoI and then run onto 0.8% agarose gel. A 0.45 kb EcoRI/XhoI-insert DNA (wt or fs) was recovered and subcloned into pGEX4T-1 (Pharmacia LKB Biotechnology Inc., Piscataway, NJ, USA) or pcDNA3 (Invitrogen) plasmid expression vector pretreated with EcoRI and XhoI. The recombinant DNAs, named pGEX-4T-1-Cu/Zn-SOD and pcDNA3-Cu/ZnSOD, were then transformed into E. coli XL-1-blue and E. coli TOPO 10 hosts, respectively. The wt recombinant DNA (pGEX-4T-1-Cu/Zn-SOD) was expressed in E. coli and its protein was identified by Western blot analysis. Large-scale preparation of both wt and fs recombinant DNAs was performed for use in the following transfection experiments. Cell culture and transfection of BV-2 cells with Cu/ Zn-SOD DNA BV-2 is a murine microglial cell line immortalized by infecting primary microglial cell cultures with a v-raf/vmyc oncogene-carrying retrovirus. BV-2 cells were cultured in DMEM/F-12 medium containing 10% heatinactivated fetal bovine serum (FBS, Gibco, Grand Island, NY, USA) at 37°C in a humidified chamber with 5% CO2. Prior to transfection, BV-2 cells in 60 mm dish at about 60 –70% confluence were washed twice with se-

1085

rum-free medium (without antibiotic). The washed cells were then treated with the wt- or fs-pcDNA3-Cu/ZnSOD DNA-Lipofectamine (Gibco) complex. Plasmid pCMV␤ DNA was used as internal control. The transfection reaction was carried out at 37°C for 3 h. Transfected cells were grown in fresh DMEM/F-12 medium containing 10% FBS at 37°C for at least 24 h before treatment.

Cell viability Cell viability was determined by the MTT (3-(4,5dimethy-2-thiazolyl)-2,5-diphenyl tetrazolium bromide) assay. After removal of the conditioned media from the plates, the cells were incubated with 0.5 mg/ml MTT (100 ␮l/well) for 4 h, the supernatants were aspirated, and the insoluble formazan product was dissolved in 0.5 ␮g/ml dimethyl sulfoxide (DMSO) (100 ␮l/well) for 1 h. The extent of MTT reduction reflects cell viability and was quantified by measuring the absorbance at 545 nm.

SOD activity staining on native PAGE Protein (5 ␮g) from each cell lysate was loaded onto the native PAGE gel. After electrophoresis, the gel was soaked in 2.5 mM nitroblue tetrazolium (NBT) solution for 15 min in the dark with gentle shaking, followed by an immersion with illumination in a solution containing 30 mM tetramethylenediamine and 10 ␮g/ml riboflavin [8]. SOD activity was represented by the achromatic band on the blue-colored gel.

Activation of BV-2 cells by lipopolysaccharide (LPS) About 80 –90% confluent cells in 35 mm culture dishes were challenged with LPS to a final concentration of 0.5 ␮g/ml. After being challenged, incubation continued at 37°C for either 6 or 18 h.

Glial fibrillary acidic protein (GFAP) detection After being treated with LPS for 18 h, the BV-2 cells were lysed with a lysis buffer (0.1 M Tris-8.0 containing 0.1% Triton X-100) and cell lysate was collected for the detection of GFAP by Western blotting. The primary antibody used was the goat polyclonal antihuman GFAP antibody (sc-6170: mouse, rat, and human reactive, 1:250) (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Antigoat IgG antibody conjugated with horseradish peroxidase (1:5000) (sc-2020, Santa Cruz Biotechnology) was used as the secondary antibody.

1086

S.-C. CHANG et al.

Detection of proIL-1␤ levels BV-2 cells (4 ⫻ 105) were transfected with 4 ␮g of plasmid (pCMV␤, fs Cu/Zn-SOD, or wt Cu/Zn-SOD) and cultured for 24 h. After treatment with LPS for 6 h, cell lysates were collected for the determination of proIL-1␤ levels by Western blotting. The primary antibody was the goat antimouse IL-1␤ antibody (1:1000) (R & D System, Minneapolis, MN, USA) and the antigoat IgG antibody conjugated with horseradish peroxidase (1:5000) (R & D System) was used as the secondary antibody. NO production assay BV-2 cells (2 ⫻ 10 ) were plated on 6-well plate, transfected with plasmid DNA (1 ␮g) for 24 h, and then treated with LPS (0.5 ␮g/ml) for 18 h. NO production by the BV-2 cells was measured as nitrite concentration using the Griess Reagent System (Promega, Madison, WI, USA). Briefly, after treatment with LPS, 50 ␮l of the conditioned medium was collected. To this conditioned medium, an equal volume of Griess reagent (0.5% sulfanilamide and 0.05% N-1-naphthylethylenediamine) was added and the reaction proceeded for 5 min at room temperature. The absorbance of the mixture at 530 nm was then measured. Sodium nitrite solutions (0 to 100 ␮M) were used to establish the standard reference curve. 5

Enzyme immunoassay for TNF-␣ An ELISA assay was used to measure the TNF-␣ level in conditioned medium from BV-2 cells that had been activated by LPS for 6 h. Flat-bottom 96-well microtiter plates were precoated with antimouse TNF-␣ monoclonal antibody (BD Pharmingen, San Diego, CA, USA). To these antibody-precoated wells, either recombinant mouse TNF-␣ (as a linear range standard: from 0 ng to 1000 ng per ml) or the medium sample was added. The TNF-␣ Ab-Ag mixture was then bound with biotinylated mouse TNF-␣ monoclonal antibody and avidinhorseradish peroxidase. The absorbance of the resultant colored product was measured at 450 nm.

Fig. 1. Characterization of the recombinant pineapple Cu/Zn-SOD by activity staining (A) on 15% native acrylamide gel and Coomassie blue staining (B) on 15% SDS-PAGE gel. Crude extract of the E. coli XL-1-blue host carrying pGEX-4T-1 vector only (lane 1) or the recombinant pineapple Cu/Zn-SOD clone (lane 2) was used for analysis; M ⫽ protein molecular weight marker. Arrows indicate the position of the fused protein of pineapple Cu/Zn-SOD with GST, which has a molecular mass of 42 kDa.

nant Cu/Zn-SOD. On the coomassie blue-stained gel (Fig. 1B), the distinct band at 42 kDa is the Cu/Zn-SOD fused with GST (glutathione transferase). Expression of pineapple Cu/Zn-SOD in BV-2 cells Expression of Cu/Zn-SOD protein and activity was detected in BV-2 cells transfected with recombinant pcDNA3 containing wt Cu/Zn-SOD. Cu/Zn-SOD protein expression was detected at 12 h after transfection and further increased at 24 h and 36 h (Fig. 2). However, the highest SOD activity was at 12 and 24 h after transfection. Cell viability and GFAP expression in LPS-treated BV-2 cells MTT assays of all the LPS-activated cells showed cell viabilities greater than 90%. GFAP expression increased in the LPS-stimulated cells (Fig. 3A), indicating that BV-2 cells were activated after LPS treatment. IL-1␤ production in LPS-treated BV-2 cells

RESULTS

Activity of Cu/Zn-SOD from the recombinant pineapple Cu/Zn-SOD cDNA The protein extract from E. coli transformed with the recombinant pineapple Cu/Zn-SOD cDNA was electrophoresed both in native acrylamide gel and SDS-PAGE gel. The resultant native gel was further stained for Cu/Zn-SOD activity. An achromatic zone (Fig. 1A) represents the enzyme activity expressed from the recombi-

Stimulation of BV-2 cells with LPS for 6 h led to an increase of proIL-1␤ levels (Fig. 3B). The increase in proIL-1␤ was significantly attenuated in the BV-2 cells transfected with wt Cu/Zn-SOD (lane 4) but not in the fs Cu/Zn-SOD transfected cells (lane 3). NO production in LPS-treated BV-2 cells NO production by BV-2 cells in response to LPS stimulation was indicated by the nitrite concentration in

Cu/Zn-SOD decreases microglial activation

1087

Fig. 2. Expression and activity of pineapple Cu/Zn-SOD in BV-2 cells. BV-2 cells were transfected with recombinant pcDNA3-Cu/ Zn-SOD plasmid and cultured for different periods of time. Top panel: Cu/Zn-SOD protein detected by Western blotting as described in the text. Five ␮g of cell lysate protein was loaded onto each lane and run onto a 15% SDS-PAGE. Middle panel: Cu/Zn-SOD activity staining. The achromatic band denotes the presence of active SOD. Lane 1, parental BV-2 cells (24 h); lane 2, BV-2 cells transfected with control plasmid pCMV␤ (24 h); lane 3, BV-2 cells transfected with fs Cu/Zn-SOD (24 h); lanes 4 – 6, BV-2 cells transfected with wt Cu/Zn-SOD and cultured for 12, 24, and 36 h, respectively. Bottom panel: the Western blotting of ␤-actin as control.

the medium. After 6 h of treatment with LPS, the BV-2 cells produced only insignificant amounts of nitrite (0.67 ␮M), but after 18 h, the nitrite concentration increased to about 17 ␮M (Fig. 4, bar 1). Transfection of wt Cu/ZnSOD significantly inhibited the production of nitrite in activated BV-2 cells (Fig. 4, bar 4), whereas transfection of fs Cu/Zn-SOD still resulted in an elevated nitrite concentration (Fig. 4, bar 3).

TNF-␣ production in LPS-treated BV-2 cells Elevation of TNF-␣ is another feature of activated microglial cells. After 6 h of stimulation with LPS, there was a significant increase of TNF-␣ in BV-2 cells. However, this increase of TNF-␣ was significantly inhibited in the BV-2 cells transfected with wt Cu/Zn-SOD as compared to the control or fs Cu/Zn-SOD transfected BV-2 cells (Fig. 5).

DISCUSSION

Copper/Zinc-SOD is a cytosolic enzyme of eukaryotes and certain prokaryotes [9]. The plant Cu/ZnSOD proteins have been demonstrated to be more stable and more resistant to heat and pH changes than human or

Fig. 3. (A) GFAP levels in unactivated or activated BV-2 cells. BV-2 cells were treated with LPS (0.5 ␮g/ml) for 18 h. GFAP was detected by Western blotting. (B) ProIL-1␤ levels in unactivated or LPS-activated BV-2 cells. ProIL-1␤ protein was detected by Western blotting. Ten ␮g of cell lysate protein was loaded onto each lane of 15% SDS-PAGE. Lane 1, parental BV-2 cells; lanes 2– 4, BV-2 cells transfected with pCMV␤, fs-Cu/Zn-SOD, and wt-Cu/Zn-SOD, respectively.

Fig. 4. Nitrite production by activated BV-2 cells transfected with Cu/Zn-SOD. BV-2 cells were transfected for 24 h and then treated with LPS (0.5 ␮g/ml) for 18 h. Nitrite concentration in conditioned medium was determined by Griess reagent. Each bar represents the mean and standard deviation of three independent duplicated experiments. The asterisk indicates significant difference (p ⬍ .05) of bar 4 from bars 1, 2, and 3, as determined by one-way ANOVA and Scheffe’s test.

1088

S.-C. CHANG et al.

Fig. 5. TNF-␣ levels in activated BV-2 control or Cu/Zn-SOD transfected cells. BV-2 cells were transfected for 24 h and then treated with LPS (0.5 ␮g/ml) for 6 h. TNF-␣ concentration in the conditioned medium was determined with an ELISA assay. Each bar represents the mean and standard deviation of three experiments. The asterisk indicates significant difference (p ⬍ .05) of bar 4 from the others, as determined by one-way ANOVA and Scheffe’s test.

bovine SOD [10,11]. These characteristics make them good candidates for transfection purposes. There is substantial evidence that superoxides generated from activated microglial cells contribute to progressive neuronal damage [12]. Moreover, SOD is considered part of the defense system against pathogenesis in neuronal lesions [13]. Elevated Cu/Zn-SOD and MnSOD immunoactivity was observed in rat microglial cells after cerebral ischemia [14]. An increase of MnSOD activity has been reported in BV-2 cells after activation by LPS, and it was hypothesized that this effect was mediated by NO or its metabolites [15]. However, the correlation between SOD and cell activation is not well understood. In this study we found that when microglial BV-2 cells were activated by LPS, the production of NO and cytokines was substantially increased, but in those transfected cells that exhibit high levels of Cu/Zn-SOD activity, the accumulation of nitrite was largely inhibited (Figs. 3 and 4). One mechanism that could account for this inhibition is the direct lessening of superoxide by SOD, which would in turn lead to a decreased generation of nitrite. Alternatively, this inhibition could also result from the decrease of cytokines such as IL-1␤ and TNF-␣, which have both been implicated as signals that precede the microglial generation of NO [16,17]. Our results showed that in BV-2 cells, proIL-1␤ and TNF-␣ levels were clearly elevated after 6 h of LPS treatment while nitrite levels still remained insignificant. An increase in nitrite levels was observed after 18 h of LPS activation, however, which suggests that increased NO production might occur only after cytokine accumulation. It is therefore possible that NO production in activated microglial cells is regulated through the induc-

tion of cytokines; however, it is also possible that the induction of these cytokines and NO are mediated through separate signaling pathways [18]. Studies using insulin-producing RINm5F cells have shown that overexpression of Cu/Zn-SOD provides protection against the toxic effects of NO and cytokinemediated toxicity [19]. Our present results demonstrate that an increase in Cu/Zn-SOD activity is consistent with the inhibition of cytokines and NO production in microglial cells. Thus, SOD may exert its protection effect by directly or indirectly suppressing the induction of cytokines and NO. NF-␬B is another signal involved in cytokine regulation, and it is also one of the regulatory elements of TNF-␣ DNA [20]. Overexpression of Mn-SOD in breast cancer cells was shown to block the induction of NF-␬B [21], and possibly, by this means, to downregulate the production of TNF-␣. However, the validity of this mechanism will need to be tested by further studies. In conclusion, we have shown that a Cu/Zn-SOD cDNA from pineapple could be successfully transfected and expressed in microglial BV-2 cells. The increase in active Cu/Zn-SOD led to decreased production of NO, IL-1␤, and TNF-␣ under LPS activation. These results suggest an alternative approach to the development of protection strategies for the neurodegenerative diseases in which microglial activation is a neuropathological event. Acknowledgement — This work was supported by a grant (NSC 892313-B-019-066) from the National Science Council of Taiwan, ROC, to C.-T. Lin.

REFERENCES [1] McGeer, P. L.; McGeer, E. G. Mechanisms of cell death in Alzheimer disease: immunopathology. J. Neural Transm. 54(Suppl.):159 –166; 1998. [2] Mrak, R. E.; Sheng, J. G.; Griffin, W. S. Glial cytokines in Alzheimer’s disease: review and pathogenic implications. Human Pathol. 26:816 – 823; 1995. [3] Banati, R. B.; Gehrmann, J.; Schubert, P.; Kreutzberg, G. W. Cytotoxicity of microglia. GLIA 7:111–118; 1993. [4] Sano, M.; Ernesto, C.; Thomas, R. G.; Klauber, M. R.; Schafer, K.; Grundman, M.; Woodbury, P.; Growdon, J.; Cotman, C. W.; Pfeiffer, E.; Schneider, L. S.; Thal, L. J. A controlled trial of selegiline, ␣-tocopherol, or both as treatment for Alzheimer’s disease. The Alzheimer’s disease cooperative study. N. Engl. J. Med. 336:1216 –1222; 1997. [5] Prasad, K. N.; Hovland, A. R.; Cole, W. C.; Prasad, K. C.; Nahreini, P.; Edwards-Prasad, J.; Andreatta, C. P. Multiple antioxidants in the prevention and treatment of Alzheimer’s disease: analysis of biologic rationale. Clin. Neuropharmacol. 23:2–13; 2000. [6] Beckman, J. S.; Koppenol, W. H. Nitric oxide, superoxide, and peroxynitrite: the good, the bad, and ugly. Am. J. Physiol. 271: C1424 –C1437; 1996. [7] Lin, M.-T.; Fu, M.-T.; Ken, C.-F.; Lin, C.-T. Cloning and characterization of a cDNA encoding for Cu/Zn-superoxide dismutase from pineapple. (Accession no. AJ250667) Plant Physiol. 122: 619; 2000.

Cu/Zn-SOD decreases microglial activation [8] Beauchamp, C.; Fridovich, I. Superoxide dismutase: improved assays and an assay applicable to acrylamide gels. Anal. Biochem. 44:276 –287; 1971. [9] Bannister, J. V.; Bannister, W. H.; Rotilio, G. Aspects of the structure, function, and applications of superoxide dismutase. CRC Crit. Rev. Biochem. 22:111–180; 1987. [10] Hallewell, R. A.; Imlay, K. C.; Lee, P.; Fong, N. M.; Gallegos, C.; Getzoff, E. D.; Tainer, J. A.; Cabelli, D. E.; Tekamp-Olson, P.; Mullenbach, G. T.; Cousens, L. S. Thermostabilization of recombinant human and bovine CuZn superoxide dismutases by replacement of free cysteines. Biochem. Biophys. Res. Commun. 181:474 – 480; 1991. [11] Lin, C. T.; Lin, M. T.; Chen, Y. T.; Shaw, J. F. Subunit interaction enhances enzyme activity and stability of sweet potato cytosolic Cu/Zn-superoxide dismutase purified by a His-tagged recombinant protein method. Plant Mol. Biol. 28:303–311; 1995. [12] Tanaka, M.; Sotomatsu, A.; Yoshida, T.; Hirai, S.; Nishida, A. Detection of superoxide production by activated microglia using a sensitive and specific chemiluminescence assay and microgliamediated PC12h cell death. J. Neurochem. 63:266 –270; 1994. [13] Lindenau, J.; Noack, H.; Possel, H.; Asayama, K.; Wolf, G. Cellular distribution of superoxide dismutases in the rat CNS. GLIA 29:25–34; 2000. [14] Liu, X. H.; Kato, H.; Nakata, N.; Kogure, K.; Kato, K. An immunohistochemical study of copper/zinc superoxide dismutase and manganese superoxide dismutase in rat hippocampus after transient cerebral ischemia. Brain Res. 625:29 –37; 1993. [15] Sugaya, K.; Chouinard, M. L.; McKinney, M. Induction of man-

[16]

[17]

[18]

[19]

[20] [21]

1089

ganese superoxide dismutase in BV-2 microglial cells. Neuroreport 8:3547–3551; 1997. Hartlage-Rubsamen, M.; Lemke, R.; Schliebs, R. Interleukin-1␤, inducible nitric oxide synthase, and nuclear factor ␬B are induced in morphologically distinct microglia after rat hippocampal lipopolysaccharide/interferon-gamma injection. J. Neurosci. Res. 57: 388 –398; 1999. Anderson, W. R.; Martella A.; Drake Z. M.; Hu, S.; Peterson, P. K.; Chao, C. C. Correlative transmission and scanning electron microscopy study of microglia activation by interferon-gamma and tumor necrosis factor ␣ in vitro. Pathol. Res. Pract. 191: 1016 –1022; 1995. Nakamura, Y.; Si, Q. S.; Kataoka, K. Lipopolysaccharide-induced microglial activation in culture: temporal profiles of morphological change and release of cytokines and nitric oxide. Neurosci. Res. 35:95–100; 1999. Lortz, S.; Tiedge, M.; Nachtwey, T.; Karlsen, A. E.; Nerup, J.; Lenzen, S. Protection of insulin-producing RINm5F cells against cytokine-mediated toxicity through overexpression of antioxidant enzymes. Diabetes 49:1123–1130; 2000. Takashiba, S.; Shapira, L.; Amar, S.; Van Dyke, T. E. Cloning and characterization of human TNF-␣ promoter region. Gene 131:307–308; 1993. Manna, S. K.; Zhang, H. J.; Yan, T.; Oberley, L. W.; Aggarwal, B. B. Overexpression of manganese superoxide dismutase suppresses tumor necrosis factor–induced apoptosis and activation of nuclear transcription factor ␬B and activated protein-1. J. Biol. Chem. 273:13245–13254; 1998.