The human neuroblastoma cell line SH-SY5Y was grown in RPMI-1640 medium supplemented with 2 mM glutamine, standard antibiotics (100 U/ml Penicillin, 100 μg/ml Streptomycin) and 2% fetal calf serum (FCS). For toxicity studies cells were seeded at 3 × 10⁴ cells per well in 48 well plates, treated with cytokines and allowed to adhere overnight before use. After 24 hours, different concentrations of peptides, staurosporine or hydrogen peroxide were added. Cell viability and/or prostaglandin E₂ content were determined after a further 24 hours.

Primary cortical neurons were prepared from embryonic day 15.5 mice as previously described 9. Neuronal progenitors were seeded at 500,000 cells per well in 48 well plates in RPMI-1640 supplemented with 2 mM glutamine, standard antibiotics and 10% FCS. After 2 hours, cultures were washed and subsequently grown in neurobasal medium containing 2 mM glutamine and B27 components (Invitrogen, Paisley, UK). Primary cerebellar neurons were prepared from the brains from newborn mice pups following dissection of the cerebellum, removal of the meninges and cell dissociation as previously described 9. Neuronal progenitors were plated in 10% FCS for 2 hours, and then grown in neurobasal medium containing glutamine and B27. In both these neuronal cultures, medium was supplemented with 5 mM L-leucine methyl ester to reduce the numbers of contaminating microglial cells. After 7 days, cultures were treated with cytokines for 24 hours before the addition of neurotoxins/peptides. Caspase-3 activity was measured 24 hours after the addition of neurotoxins using a flourometric immunosorbent enzyme assay kit as per the manufacturer's instructions (Roche Diagnostics, Lewes, UK). Results are expressed as fluorescent units which are proportional to caspase-3 activity. For toxicity assays medium was replaced 48 hours after the addition of neurotoxins/peptides and cell viability was determined after another 48 hours (4 days after the addition of neurotoxins/peptides).

A peptide corresponding to amino acids 1 to 42 of the amyloid-β protein (amyloid-β_1-42) and a control peptide (amyloid-β_42-1) were obtained from Bachem (St Helens, UK). Peptides containing amino acid residues 82 to 146 of the human PrP protein (HuPrP82-146) corresponding to a PrP fragment found in certain prion-infected human brains 10, a control peptide containing the same amino acids in a scrambled order (HuPrP82-146scrambled) were a gift from Professor Mario Salmona (Mario Negri Institute, Milan).

To determine cell survival, cultures were treated with WST-1 (Roche Diagnostics Ltd, Lewes, UK) for 3 hours and optical density was read on a spectrophotometer at a wavelength of 450 nm. WST-1 is cleaved to formazan by mitochondrial dehydrogenases and the amount of dye formed correlates to the number of metabolically active cells. Percentage cell survival in cultures was calculated by reference to untreated cells incubated with WST-1 (100%).

SH-SY5Y neuroblastoma cells were lysed in an extraction buffer containing 10 mM Tris-HCl, pH 7.8, 100 mM sodium chloride, 10 mM EDTA, 0.5% Nonidet P-40, 0.5% sodium deoxycholate and 2 mM phenylmethylsulphonylflouride at 1 × 10⁶ cells per ml. Protein content was determined using a BCA kit (Pierce, Cramlington UK) and protein concentrations standardised. 20 μl samples were analysed via PAGE or blotted onto a PVDF membrane. Where appropriate, dilutions of lysates were made prior to blotting. Blots were probed with monoclonal antibodies (mabs) to cPLA₂ or phospholipase C (PLC)γ-1 (Upstate, Milton Keynes, UK) and developed with an anti-mouse IgG-alkaline phosphatase conjugate followed by BCIP/NBT (Sigma).

Analysis of total prostaglandin E₂ levels was performed using an enzyme-immunoassay kit Amersham Biotech (Amersham, UK).

Recombinant murine TNFα, IL-6, IL-1β, IFN-γ were supplied from (R&D systems, Abingdon, UK). Human IFN- was obtained from (Sigma, Poole, UK).

Comparison of treatment effects were carried out using one and two way analysis of variance techniques as appropriate. Post hoc comparisons of means were performed as necessary.

Preliminary studies examined the effects of varying concentrations of murine cytokines (0.01 to 10 ng/ml) on the survival of primary murine cortical neurons. We were unable to detect any significant reduction in the survival of neurons following culture with any of the following recombinant murine cytokines; TNF-α, IL-1β, IL-6, or IFN-γ. Similarly, none of the recombinant cytokines affected the survival of cerebellar neurons, or the survival of the SH-SY5Y neuroblastoma cells. Amyloid-β_1-42 caused a dose-dependent reduction in the survival of neurons that was not observed after the addition of a control peptide (amyloid-β_42-1). To determine if cytokines could modify the effects of amyloid-β_1-42, primary cortical neurons were pre-treated with 1 ng/ml individual cytokines, before the addition of 10 μM amyloid-β_1-42. There was no significant difference between the survival of neurons pre-treated in control medium and those pre-treated in medium containing TNF-α, IL-1β or IL-6 prior to the addition of amyloid-β_1-42. In contrast, the survival of neurons pre-treated with IFN-γ and amyloid-β_1-42 was significantly less than neurons treated with amyloid-β_1-42 alone (Figure 1). Further studies demonstrated that this effect of IFN-γ was dose-dependent; and a significant reduction in neuronal survival was still observed when cells were treated with 40 pg per ml of IFN-γ (Figure 2).

The effects of IFN-γ were tested on both primary cortical and cerebellar neuronal cultures. Pre-treatment with IFN-γ (100 pg/ml) resulted in reduced survival of both primary cerebellar and cortical neurons following the addition of 10 μM amyloid-β_1-42. Since it is possible that the effects of IFN-γ in these neuronal cultures were via effects on contaminating astroglial cells, we also tested the effects of IFN-γ on the SH-SY5Y neuroblastoma cell line. Pre-treatment with IFN-γ reduced the survival of SH-SY5Y neuroblastoma cells following the addition of 10 μM amyloid-β_1-42 indicating that IFN-γ had a direct effect on neuroblastoma cells (Table 1).

To determine if IFN-γ treated neurons show increased sensitivity to other neurotoxins, cortical neurons were treated with 100 pg/ml of IFN-γ prior to exposure to HuPrP82-146, a synthetic correlate of a neurotoxic peptide found in the brains of patients with prion disease 10, staurosporine or hydrogen peroxide. The survival of neurons pre-treated with IFN-γ was significantly less than that of untreated neurons, when incubated with HuPrP82-146. However, there were no significant differences between the survival of neurons treated with IFN-γ and untreated neurons that were exposed to hydrogen peroxide, or to staurosporine, a drug that caused programmed cell death in neurons via activation of the ceramide pathway 11 (Table 2).

Caspase-3 is an enzyme that is increased during apoptosis 12 and was measured as an alternative indicator of neuronal injury. Caspase-3 activity was increased in primary cortical neurons treated with amyloid-β_1-42 or HuPrP82-146, but not in primary cortical neurones treated with control peptides (amyloid-β_42-1 or HuPrP82-146scrambled) or with IFN-γ (100 pg/ml) alone. Following pre-treatment with 100 pg/ml IFN-γ caspase-3 activity in cortical neurons treated with either amyloid-β_1-42 or HuPrP82-146 was significantly higher than in untreated cells incubated with amyloid-β_1-42 or HuPrP82-146 (Figures 3 &4).

Since recent studies demonstrated that cPLA₂ is involved in amyloid-β_1-42 induced neuronal injury 13 we compared levels of cPLA₂ and another enzyme involved in cell signalling (PLCγ-1) in IFN-γ-treated and untreated SH-SY5Y cells. Treatment with IFN-γ did not significantly alter the total protein content of cells (data not shown). When lysed cells were diluted and analysed in a dot blot we found that SH-SY5Y cells treated with IFN-γ (100 pg/ml) had higher levels of cPLA₂ than did untreated cells (Figure 5). However, pre-treatment with IFN-γ did not affect the levels of PLCγ-1 indicating that IFN-γ up-regulates specific pathways in these neurons. We next determined if pre-treatment with cytokines affected prostaglandin production. Levels of prostaglandin E₂ were not altered by any of the cytokines tested. Prostaglandin E₂ levels were significantly raised after the addition of either amyloid-β_1-42 or HuPrP82-146, but not after the addition of control peptides (amyloid-β_42-1 or HuPrP82-146scrambled). Pre-treatment of neurons of 100 pg/ml IFN-γ resulted in increased prostaglandin E₂ production following the addition of 10 μM amyloid-β_1-42 or 10 μM HuPrP82-146. Prostaglandin E₂ levels in neurons incubated with 10 μM amyloid-β_1-42 or 10 μM HuPrP82-146 was not affected by pre-treatment with TNF-α, IL-1β, IL-6 (Table 3).