Cytokine action and oxidative stress response in differentiated neuroblastoma SH-SY5Y cells

In the retinoic acid-differentiated neuroblastoma SH-SY5Y cells, IL-1 induced binding activity of NF (cid:1) B and up-regulated the expression and activity of MnSOD. The IL-1-elicited effects were partly reversed by IL-4 and IL-6. It is proposed that IL-4 and IL-6 may participate in the regulation of the imbalanced oxidant status induced by IL-1 in differentiated neuroblastoma cells. In the SH-SY5Y cell line, TNF (cid:2) neither activated NF (cid:1) B nor induced MnSOD expression and activity, but was capable of modu-lating the IL-1 effects. Pyrrolidine dithiocarbamate (PDTC), an inhibitor of NF (cid:1) B activation, down-regulated the expression and activity of MnSOD, which may suggest that the regulation of MnSOD by IL-1 in retinoic acid-differentiated neuroblastoma cells was mediated by the nuclear factor (cid:1) B.


2000)
. It has been well established that oxidative stress can trigger the production of cytokines which, in turn, induce the synthesis of proteins such as manganese superoxide dismutase (MnSOD) (Antras-Ferry et al., 1997;Rogers et al., 2001), which protects cells against the damage caused by reactive oxygen species. Xu et al. (1999) cloned human MnSOD gene, and found several NFkB binding sites in the 3¢ and 5¢ flanking and the intronic regions. The authors suggested that the location of NFkB elements in the MnSOD gene is critical for the IL-1-elicited induction, but in cooperation with other transcription elements. In the present study we examined the effects of IL-1b and/or IL-4, IL-6 and TNFa on the activation of NFkB and on the activity and expression of MnSOD in differentiated neuroblastoma cells. A number of neurodegenerative disorders are characterised by an increase in the level of proinflammatory cytokines (Neuroinflammatory Working Group, 2000). These cytokines can modulate several intracellular signal transduction pathways in neuronal cells, e.g. via activation of NFkB or induction of COX-2 expression (Fiebich et al., 2000), which my lead to cell destruction. Neuroblastoma cell culture appeares to be a good experimental model for examining the influence of cytokine network on neuronal metabolism. SH-SY5Y cells can be morphologically differentiated into neuronal cells, whose phenotype varies depending on the inducing factors, e.g. retinoic acid, nerve growth factor, dibutyryl cyclic AMP or TPA (Yu et al., 1988). Cell culture. SH-SY5Y neuroblastoma cells were cultured at 37°C in 75 cm 2 flasks containing DMEM supplemented with 10% FBS and antibiotics under a humidified atmosphere of 95% air and 5% CO 2 . The cells were differentiated with retinoic acid (RA) (10 mM) for 72 h (Scheibe & Wagner, 1992). For isolation of nuclear proteins and total RNA, as well as for MnSOD activity assay, the SH-SY5Y cells were cultured in 60 mm Petri dishes. The medium was changed 24 h before addition of the following cytokines: IL-1b (10 ng/ml) and/or TNFa (10 ng/ml), IL-4 (100 U/ml) and IL-6 (25 ng/ml). In some assays, the culture medium was supplemented with pyrrolidine dithiocarbamate (PDTC) (100 mM) 1 h prior to IL-1 addition.

SH
Nuclear extracts were isolated after 90 min, and total cellular RNA after 8 h of cytokine or PDTC treatment. The activity of MnSOD was estimated in cells cultured with the specific cytokines or PDTC for 24 h.
Nuclear protein extraction and EMSA. Nuclear extracts were prepared by a mini-extraction procedure (Suzuki, 1994). The retinoic acid-differentiated neuroblastoma cells were cultured for 90 min with IL-1 (10 ng/ml), or a mixture of IL-1 and TNF (10 ng/ml), or IL-4 (100 U/ml), or IL-6 (25 ng/ml); they were then washed with cold phosphatebuffered saline (PBS), collected and centrifuged for 5 min at 400´g. The cells, previously resuspended in a buffer (10 mM NaCl, 3 mM MgCl 2 , 10 mM Tris, pH 7.5, and 0.2 mM phenylmethylsulfonyl fluoride, PMSF), were incubated on ice for 15 min. Nonidet NP-40 was added, and samples were centrifuged for 60 s at 14 000 r.p.m. Pelleted nuclei were resuspended in a buffer (10 mM Hepes, 0.35 M NaCl, 5 mM EDTA, 1 mM dithiothreitol (DTT) and 0.2 mM PMSF) and centrifuged for 5 min at 14 000 r.p.m. After centrifugation at 4°C, the supernatant proteins were measured by the BCA method. The remainder of the supernatant was frozen in 10% glycerol.
For NFkB activity assay (a DNA electrophoretic mobility shift assay), nuclear protein extracts (10 mg) were incubated for 30 min at a room temperature in 25 ml of the binding buffer (0.5% Triton X-100, 2.5% glycerol, 10 mM Hepes, 4 mM DTT) containing 0.5 ng of 32 P-end-labelled NFkB-binding oligonucleotide (about 10 5 c.p.m.) and 1 mg of poly(dI-dC) (which was used as competitor). DNA-protein complexes were separated in a 5% polyacrylamide gel for 1.5 h at 140 V. The dried gels were analysed by authoradiography. The relative intensity of the bands was evaluated densitometrically using the computer imaging system Fluor S MultiImager (BioRad).

RNA extraction and Northern blots. The
Chomczynski's extraction method (Chomczynski & Sacchi, 1987) and isopropanol precipitation were used for total RNA isolation from the neuroblastoma cells treated for 8 h with the analysed cytokines or PDTC. RNA samples (10 mg) were separated by electrophoresis in a 1% agarose gel under denaturing conditions. After electrophoresis, RNA was transferred to Hybond-N membranes (Amer-sham) according to the manufacturer's instructions. The filters were prehybridized at 68°C for 3 h in 10% dextrane sulphate and 1% SDS, and were hybridized in the same solution overnight at 65°C with a 32 P-labelled human MnSOD cDNA probe and subjected to autoradiography. The relative intensity of the bands was evaluated densitometrically using the computer imaging system Fluor S Multi-Imager (BioRad). SOD activity evaluation. The cells were treated with cytokines or PDTC for 24 h, washed twice with cold PBS, harvested in 0.2 ml of PBS, frozen and thawed four times in liquid nitrogen, each time under stirring. The mixture was centrifuged for 2 min (14 000 r.p.m., 4°C) and the supernatant was used for SOD activity measurement. A BCA kit was used for cellular protein estimation. Protein samples (10 mg) were separated in a 15% polyacrylamide gel, 180 V, for 80 min in a Tris/glycine buffer. After electrophoresis, the gel was immersed in a staining buffer containing riboflavin (50 mM phosphate buffer, pH 7.8, 10 mM EDTA, 245 mM nitro blue tetrazolium, NBT, 28 mM TEMED, 30 mM riboflavin), stirred in the dark and then exposed to light until white bands appeared on a blue background.  Fig. 1 exposure of SH-SY5Y cells to 10 mM RA for 72 h led to their differentiation. We used a NFkB consensus oligonucleotide for EMSA analysis to determine whether IL-4, IL-6 or TNF cooperate with IL-1 in NFkB activation. Figure 2 demonstrates that IL-1 is the main inducer of this transcription factor. In contrast, TNFa had no effect on NFkB binding activity, and even decreased the IL-1-elicited NFkB activation. However, these results should be further analysed. As expected IL-4 and IL-6 decreased the NFkB activation induced by IL-1.

As shown in
The regulation of MnSOD expression by the tested cytokines showed a response pattern similar to that observed for NFkB activation (Fig. 3). IL-1 distinctly up-regulated MnSOD expression in the differentiated neuroblastoma cells, whereas IL-4, IL-6 and TNF partly reversed that effect. The enzyme activity was greatly enhanced by IL-1, but the modifying effect exerted by IL-4, IL-6 and TNF was less potent (Fig. 4). It was found previously that retinoic acid affected MnSOD protein stability, but not expression (Ahlemeyer et al., 2001), hence some interference of the RA-produced effects with those of IL-4, IL-6 or TNF should be considered. Neither of the cytokines examined was capable of affecting CuZnSOD activity (Fig. 4), but PDTC, an inhibitor of NFkB activation (Bowie et al., 1997), diminished the activity of this enzyme (Fig. 5B). PDTC inhibited both the IL-1-induced NFkB binding (Fig. 6) and the IL-1-elicited MnSOD expression and activity (Fig. 5A, B).

DISCUSSION
In the present report we describe the effects of proinflammatory (IL-1, TNF) and so-called antiinflammatory cytokines (IL-4, IL-6) on Several studies have reported the induction of NFkB binding activity by IL-1 (Fig. 2), whereas the effectiveness of TNF seems to be cell-type-specific. As is shown in Fig. 2, TNF is unable to activate NFkB in the RA-differentiated neuronal SH-SY5Y cells. A similar observation was made by Wong (1995) with neuroblastoma SK-N-SH cells and some other tumour cells. The nuclear factor kB is regarded as an antiapoptotic factor in neuroblastoma cells (Yabe et al., 2001;Bian et al., 2002). However, prolonged activation of NFkB may be dangerous to the cell. It seems that IL-4 and IL-6 can abrogate the IL-1-elicited effect on nuclear factor kB activation. Cellular responses to cytokines depend on receptors, signalling molecules and the stage of cell differentiation. The cytokine-activated transcription factors NFkB, C/EBP and STATs regulate individually or cooperatively the expression of target genes. Gene expression is determined by interactions between transcription factors, the promoter context of the target gene, and the presence of co-activator complexes. A cross-talk between the C/EBP, STAT and NFkB signal transduction pathways has been postulated (Luo & Yu-Lee, 2000;Kiningham et al., 2001;Cisowski et al., 2002). It is possible that the STAT pathway, which is activated by IL-4 or IL-6, may inhibit the NFkB binding induced by IL-1 in neuroblastoma cells (experiments in progress). Vol. 50 Stress response of SH-SY5Y cells 663  The brain is particularly susceptible to oxygen free radicals which are implicated in the pathology of several neurological disorders (Jenner, 2003;Pong, 2003;Klein & Ackerman, 2003). The antioxidant enzyme system of the brain may play an important role in the protection against oxidative stress. The induction of MnSOD, an enzyme converting superoxide anion to hydrogen peroxide, by IL-1 can protect cells against the damaging effects of reactive oxygen species. MnSOD overexpression, in human neuronal cells expressing mutant CuZnSOD, attenuated neuronal death (Flanagan et al., 2002). CuZnSOD -overexpressing astrocytoma cells also show increased resistance to oxidative injury (Chen et al., 2001). IL-4 and IL-6 seem to be capable of normalizing the IL-1-produced effects. Depending on the activation of astrocytes, glial and neuronal cells in the CNS, different subsets of cytokines can be generated; they can modulate neuronal cell metabolism resulting in cell protection or damage (Klegeris & McGeere, 2001).