Squalestatin cures prion-infected neurons and protects against prion neurotoxicity.

A key feature of prion diseases is the conversion of the normal, cellular prion protein (PrP(C)) into beta-sheet-rich disease-related isoforms (PrP(Sc)), the deposition of which is thought to lead to neurodegeneration. In the present study, the squalene synthase inhibitor squalestatin reduced the cholesterol content of cells and prevented the accumulation of PrP(Sc) in three prion-infected cell lines (ScN2a, SMB, and ScGT1 cells). ScN2a cells treated with squalestatin were also protected against microglia-mediated killing. Treatment of neurons with squalestatin resulted in a redistribution of PrP(C) away from Triton X-100 insoluble lipid rafts. These effects of squalestatin were dose-dependent, were evident at nanomolar concentrations, and were partially reversed by cholesterol. In addition, uninfected neurons treated with squalestatin became resistant to the otherwise toxic effect of PrP peptides, a synthetic miniprion (sPrP106) or partially purified prion preparations. The protective effect of squalestatin, which was reversed by the addition of water-soluble cholesterol, correlated with a reduction in prostaglandin E(2) production that is associated with neuronal injury in prion disease. These studies indicate a pivotal role for cholesterol-sensitive processes in controlling PrP(Sc) formation, and in the activation of signaling pathways associated with PrP-induced neuronal death.

The observation that cellular PrP C is essential for the development of prion diseases (5)(6)(7) suggests that the density and cellular location of PrP C in neurons may influence the production of PrP Sc . PrP C is found in lipid rafts or caveolae-like domains (CLDs) 1 (8), specialized membrane compartments that contain high levels of cholesterol and sphingomyelin (9). The formation of these lipid rafts is cholesterol-dependent (10), and drugs that affect cholesterol levels have been shown to influence the formation of PrP Sc . Thus, the formation of PrP Sc in prion-infected neuroblastoma cells (ScN2a cells) was reduced when cells were treated with lovastatin (11). Lovastatin is a competitive inhibitor of 3-hydroxy-3-methylglutaryl-co-enzyme A (HMG-CoA) reductase, the rate-limiting step in cholesterol production (12) and HMG-CoA reductase inhibitors are useful clinical tools used to treat hypercholesterolemia (13) or ischemic heart disease (14 -15) in man. However, although HMG-CoA reductase inhibitors reduce cholesterol levels, they exhibit many other effects, for example, they also inhibit the production of isoprenoid precursors (16). Recent studies suggest that many of the effects of HMG-CoA reductase inhibitors are independent of their effects on cholesterol metabolism (17)(18)(19). In the present study we investigated the effects of squalestatin, a specific inhibitor of squalene synthase that inhibits cholesterol production without affecting the production of nonsterol products (20) (Fig. 1), on the production of PrP res in three prion-infected cell lines, and on the levels and distribution of PrP C in non-infected cells.
Although the accumulation of aggregated PrP Sc in the CNS is a distinguishing feature of prion diseases (21), the precise mechanisms by which PrP Sc deposition leads to neuronal damage remain to be fully determined. The processes of neuronal loss have been examined by incubating neuronal cells with either prion preparations (22) or with peptides derived from the protease-resistant core of PrP Sc (23). In the present study we used two peptides derived from the human PrP protein, HuPrP(82-146), a synthetic equivalent of a PrP Sc fragment present in the brains of patients with GSS (24) and the truncated, neurotoxic version HuPrP-(106 -126) (23), and a synthetic miniprion (sPrP106) (25) to investigate if cholesterolsensitive microdomains are required for the processes by which prions kill infected neurons. We also examined the effects of cholesterol manipulation on neuronal prostaglandin (PG)E 2 production since levels of neuronal PGE 2 are closely associated with PrP-induced neuronal death (26).
Microglial activation frequently co-localizes with the accumulation of PrP Sc in the CNS (4). In vitro, microglia secrete cytokines in response to PrP peptides (27) or when incubated with prion-infected neurons (28). Moreover, the addition of microglia to neurons treated with prions or PrP peptides results in a significant further reduction in neuronal survival, which is greater than that seen on exposure of neurons to PrP peptides only (29,30). These observations suggest a significant role of microglia in neuronal loss during prion diseases and support interpretations drawn from time course studies of experimental prion disease in rodent models (1,30). Thus, we investigated the effect of squalestatin on a second model of prion-induced neurodegeneration that involves the interaction between microglia and prion-infected neuroblastoma cells or PrP-damaged neurons.

EXPERIMENTAL PROCEDURES
PrP res Production-Scrapie-infected neuroblastoma cells (ScN2a cells) were a gift of Dr. M. Rogers (University College, Dublin, Ireland). These cells produce PrP Sc and infectious agent, and were grown in Hams F12 medium containing 2 mM glutamine, standard antibiotics (100 units/ml penicillin and 100 g/ml streptomycin), 2% fetal calf serum (FCS) and 200 nM retinoic acid. SMB cells (TSE Resource Centre, Institute for Animal Health, Compton, UK), which also produce PrP Sc and infectious agent, were grown in RPMI 1640 medium containing standard antibiotics, 2 mM glutamine, 200 nM retinoic acid, and 2% FCS. ScGT1 cells are an immortalized murine hypothalamic neuronal cell line infected by the scrapie Chandler strain and persistently express PrP res and were supplied by Dr. Sylvain Lehmann (CNRS-IGH, Montpellier, France) and were grown in optimem supplemented with 2 mM glutamine, 200 nM retinoic acid, 5% FCS, and standard antibiotics. To measure the effect of squalestatin on the PrP res formation, ScN2a, SMB cells or ScGT1 cells were plated at 1 ϫ 10 5 cells/well in 6-well microtiter plates in the presence or absence of squalestatin alone (0 -100 nM) (GlaxoSmithKline, Stevenage, Herts, UK), 100 nM squalestatin plus 1-100 g/ml cholesterol (either free cholesterol or water-soluble cholesterol complexed to methyl-␤-cyclodextrin)(Sigma), or 100 nM squalestatin plus 500 M mevalonate (Sigma). Cells were then grown with daily changes of media and PrP res production was evaluated after 8 days. In some studies, ScN2a cells were grown in the serum-free neurobasal medium (NBM) (Invitrogen, Paisley, UK) supplemented with 2 mM glutamine and B27 components. Non-infected N2a cells or SMB cells that had been "cured" of infectivity by serial passages in the presence of pentosan polysulphate (PS cells) were used as controls. At the end of the treatment, cells were detached and counted to establish cell numbers. Washed cells were first lysed in distilled water and the insoluble pellet collected. The pellet was suspended in an extraction buffer containing 10 mM Tris-HCl, 100 mM NaCl, 10 mM EDTA, 0.5% Nonidet P-40, and 0.5% sodium deoxycholate and digested with 10 g/ml proteinase K for 1 h at 37°C. Digestion was stopped with 5 mM phenylmethanesulfonyl fluoride (PMSF) and the remaining supernatant was split in two. The first sample was tested in a PrP-specific enzyme-linked immunosorbent assay (ELISA). The second sample was centrifuged at 50,000 ϫ g for 4 h at 4°C. The pellet was dissolved in 50 l of Laemmli buffer (Bio-Rad). Samples were boiled for 5 min and 20 l of each sample was subjected to electrophoresis on a 15% polyacrylamide gel. Proteins were transferred onto a Hybond-P PVDF membrane (Amersham Biosciences) by semi-dry blotting. Membranes were blocked using 10% milk powder in Tris-buffered saline containing 0.2% Tween 20. PrP res was detected by incubation with mab SAF83 (a gift from J. Grassi, CEA, Saclay, France) for 1 h at room temperature, followed by a secondary anti-mouse IgG conjugated to peroxidase (1 h at room temperature). Detection of bound antibody was by the enhanced chemiluminescence kit (Amersham Biosciences). Lysates were also made from the non-infected N2a cells to evaluate PrP C content. Cells were treated as above except that proteinase K digestion was excluded.
Solubility of PrP c in Triton X-100 -To determine the effect of squalestatin on the distribution of PrP C , uninfected murine N2a neuroblastoma cells were maintained in RPMI 1640 supplemented with 2 mM glutamine, standard antibiotics, and 2% FCS in the presence or absence of 100 nM squalestatin. After 3 days cells were lysed on ice at 1 ϫ 10 7 cells/ml in either (a) an extraction buffer containing 10 mM Tris-HCl, pH 7.8, 100 mM NaCl, 10 mM EDTA, 0.5% Nonidet P-40, 0.5% sodium deoxycholate, and 5 mM PMSF or (b) a buffer containing 1% Triton X-100, 150 mM NaCl, 10 mM Tris-HCl, pH 7.8, and 5 mM PMSF. Soluble material was collected after centrifugation at 10,000 ϫ g. The insoluble pellet was dissolved in extraction buffer (a) at 1 ϫ 10 7 cells/ml, and all three supernatants were tested in a PrP ELISA.
Animals-129/Ola mice were housed at constant temperature (21 Ϯ 1°C) and relative humidity (60 Ϯ 10%) and supplied ad libitum with water and a standard diet. Procedures involving animals and their care were conducted in accordance with national and international regulations (Animals (Scientific Procedures) Act, 1986, UK, and EEC Council Directive 86609, OJ L358, 1, 12 December 1987).
Primary Neuronal Cultures-Primary cortical neurons were prepared from embryonic day 15.5 mice as previously described (28). After 2 days, medium was changed to NBM containing B27 components, 2 mM glutamine and 5 M cytosine arabinoside to prevent the proliferation of astroglia. Mature cultures were treated with 5 mM L-leucine methylester to reduce the number of contaminating microglia before further use (31). For experiments, cultures were preincubated for 3 h in media containing 100 nM squalestatin alone or a mixture of 100 nM squalestatin and 1-100 g/ml water-soluble cholesterol before the addition of peptides (HuPrP-(82-146), HuPrP-(106 -126), sPrP106, or a prion extract. After another 3 h, microglia were added to neuronal cultures in the ratio 1 microglia to 10 neurons. After 4 days, cultures were shaken and washed to remove non-adherent microglia, and neuronal viability was determined using the 3-[4,5 dimethylthiazol-2yl]-2,5 diphenyltetrazolium bromide (MTT) method.
Microglia-Microglia cultures were prepared from newborn 129/Ola mice as previously described (28). Isolated microglia was added to ScN2a cells or treated neuronal cells in the ratio of 1 microglia to 10 neuronal cells. In other studies microglia were added to 96-well flatbottomed plates at 1 ϫ 10 5 cells/well and allowed to adhere overnight. Microglia were subsequently treated with 100 nM squalestatin or a vehicle control for 3 h before the addition of 10 ng/ml lipopolysaccharide. After a further 24 h supernatants were collected and tested for IL-6.
Neurotoxicity of PrP Peptides or Prion Preparations-The human neuroblastoma SH-SY5Y cell line (European Collection of Cell Cultures) was grown in RPMI 1640 medium supplemented with 2 mM glutamine, standard antibiotics and 2% FCS. Cells were plated at 3 ϫ 10 4 cells/well into 96-well microtiter plates and allowed to adhere overnight. The following day, cells were pretreated for 3 h either with 0 -100 nM squalestatin alone or with 100 nM squalestatin plus 1-100 g/ml of water-soluble cholesterol or with 100 nM squalestatin plus 500 M of mevalonate, before the addition of peptides, sPrP106 or a diluted prion extract. Squalestatin stock solutions were made in di-methyl sulfoxide (Me 2 SO) and diluted appropriately. Cell viability was determined 24 h later using the MTT method (32). Optical density was measured at 595 nm, and results calculated by reference to untreated cells. PGE 2 levels were also determined at this time point. Microglia-mediated Toxicity-To investigate the effect of squalestatin on microglia-mediated toxicity, ScN2a cells were plated at 1 ϫ 10 5 cells/well in 24-well plates and allowed to adhere overnight. Cells were cultured in the presence or absence of 100 nM squalestatin alone, or in a mixture of 100 nM squalestatin and 1-100 g/ml water-soluble cholesterol for a further 24 h after which microglia were added in the ratio 1 microglia to 10 ScN2a cells. The survival of ScN2a cells was measured after cells had been co-cultured for 24 h using the MTT method. N2a neuroblastoma cells were used as non-infected controls.
Prion Extract Preparations-Partially purified prion extracts were obtained from ScGT1 cells to reduce the number of potential contaminants that may be present in prion preparations from infected brain material. Washed cells were suspended in an extraction buffer containing 10 mM Tris-HCl, 100 mM NaCl, 10 mM EDTA, 0.5% Nonidet P-40 and 0.5% sodium deoxycholate and digested with 10 g/ml of proteinase K for 1 h at 37°C. Digestion was stopped with 5 mM PMSF and the supernatant was loaded onto a prewashed C18 Sep-Pak column (Waters, Elstree, UK). Bound material was eluted under a gradient of n-propanol in water (5-100%), fractions containing PrP were pooled (samples tested by a PrP ELISA), washed three times in distilled water and resuspended in 0.01 M pH 7.4 phosphate-buffered saline at 100 ng/ml. Control preparations from non-infected N2a cells were produced by lysis in extraction buffer containing 5 mM PMSF. The supernatant was loaded on to a prewashed C18 Sep-Pak column and eluted under a gradient of n-propanol in water. Fractions containing PrP were pooled, washed three times in distilled water, and resuspended in phosphatebuffered saline at 100 ng/ml.
PrP ELISA-Briefly, treated lysates were added to a Nunc Maxisorb Immunoplates precoated with diluted rabbit antiserum raised to a PrP peptides (ovine PrP-  and ovine PrP-(100 -111)) conjugated to keyhole limpet hemocyanin (a gift from by Dr. J. P. M. Langeveld, Pepscan Systems, Lelystad, The Netherlands). Samples were applied and visualized using mAb SAF83 (gift from Prof. J. Grassi (CEA, Saclay, France), followed by an anti-mouse IgG-alkaline phosphatase conjugate (Sigma) and an appropriate indicator. Plates were read at 450 nM, and results were calculated by reference to a standard curve of recombinant murine PrP (Prionics, Zurich, Switzerland). PGE 2 Assay-Analysis of cellular PGE 2 levels was determined in SH-SY5Y cells by using an enzyme immunoassay kit (Amersham Biosciences) according to the manufacturer's instructions. This assay is based on competition between unlabeled PGE 2 in the sample and a fixed amount of labeled PGE 2 for a PGE 2 -specific antibody. The detection limit of this assay is 20 pg/ml. IL-6 ELISA-Levels of interleukin (IL)-6 in cultures containing microglia were assayed in a sandwich ELISA (R&D Systems, Abingdon, UK) as previously described (28). Plates were read at 450 nm, and results were calculated by reference to a standard curve of recombinant murine IL-6. The detection limit of this assay is 20 pg/ml.
Cholesterol and Protein Content-Cellular cholesterol and protein content were determined on 1 ϫ 10 7 SH-SY5Y cells, or ScN2a cells grown in either Hams F12 medium containing 2% FCS or in serum-free NBM, before and after exposure to 100 nM squalestatin for 3-24 h. At the end of incubation, cells were detached by trypsinization, washed twice with phosphate-buffered saline, and pelleted (550 ϫ g for 10 min). Total lipids were extracted according to Folch et al. (33), and total cholesterol was determined using an enzymatic assay kit (Roche Applied Science). Cellular protein concentration was determined on cell lysates using the Bio-Rad protein assay.
Statistical Analysis-Comparison of treatment effects was carried out using one and two way analysis of variance techniques as appropriate. Post hoc comparisons of means were performed as necessary. For all statistical tests significance was set at the 5% level.

Squalestatin Reduces PrP res Formation in Prion-infected
Cells-Previous studies had indicated that ScN2a cells contain protease-resistant PrP (PrP res ) and infectious agent (34). Here, we investigated the effect of squalestatin, an inhibitor of squalene synthase that inhibits the production of cholesterol (20), on ScN2a cells. As shown in Fig. 2, the accumulation of PrP res in ScN2a cells was significantly reduced by squalestatin in a dose-dependent manner. The amounts of PrP res were reduced to below detectable levels in ScN2a cells treated with 100 nM squalestatin ( Fig. 2 and Table I). This effect was observed in both lipid-depleted NBM medium and in Hams F12 medium containing 2% FCS that contained extracellular cholesterol. The effect of squalestatin on PrP res production was accompanied by a decrease in the cellular cholesterol levels. mediated by sterol derivatives, the PrP res formation was evaluated in ScN2a cells treated for 24 h with 100 nM squalestatin alone or combined with 1-100 g/ml of water-soluble cholesterol, or with 500 M of mevalonate. As reported in Table I, mevalonate did not affect the content of PrP res in squalestatintreated cells. In contrast, the addition of water-soluble cholesterol partially reversed the effect of squalestatin on PrP res production although it was not fully restored even with the higher dose of 100 g/ml of cholesterol (Table I).
To confirm the effects of squalestatin on PrP res production, two other prion-infected neuroblastoma cell lines (SMB or ScGT1 cells) was treated with 100 nM squalestatin. The PrP res content of SMB cells treated with a vehicle control (7,405 Ϯ 512 PrP res pg/ml) was significantly higher than that of SMB cells treated with 100 nM squalestatin (Ͻ 50 pg/ml, n ϭ 8, p Ͻ 0.05). Similar results were observed with ScGT1 cells, in that the PrP res content of ScGT1 cells treated with a vehicle control (10,014 Ϯ 845 PrP res pg/ml) was significantly higher than that of ScGT1 cells treated with 100 nM squalestatin (Ͻ 50 pg/ml, n ϭ 8, p Ͻ 0.05).
Effect of Squalestatin on the Solubility of PrP c in Triton X-100 -Previous studies showed that PrP C to preferentially locate in cholesterol-rich lipid rafts which are insoluble in nonionic detergents such as Triton X-100 (11). To determine the effect of squalestatin on the levels and distribution of PrP C , N2a neuroblastoma cells were grown in the presence or absence of 100 nM squalestatin. There was no significant difference in total levels of PrP C between untreated and squalestatin-treated cells. However, while the majority of PrP C in untreated cells was Triton X-100 insoluble, in squalestatintreated cells most of the PrP C was Triton X-100 soluble (Fig.  3A). PrP C in squalestatin-treated cells remained sensitive to protease digestion (Fig. 3B).
Squalestatin Protects Neurons Against PrP Peptides-The effect of squalestatin on the sensitivity of SH-SY5Y cells to PrP peptides was examined. Cells treated for 3 h with 100 nM squalestatin contained significantly less cholesterol than did untreated ones (0.394 Ϯ 0.024 g of cholesterol/mg of protein for untreated cells and 0.172 Ϯ 0.033 g of cholesterol/mg of protein for treated cells, n ϭ 6, p Ͻ 0.05). The survival of squalestatin-treated SH-SY5Y cells incubated with 40 M HuPrP-(106 -126), 10 M HuPrP-(82-146) or 10 M sPrP106 was significantly greater than that of vehicle treated SH-SY5Y cells incubated with these peptides (Fig. 4A). This protective effect of squalestatin was dose-dependent with an IC 50 ϳ1 nM when tested against 10 M HuPrP(82-146) (Fig. 4B).
Previous studies have shown that the neuronal loss is observed after the addition of partially purified prion prepara-tions (22,28). In the present study the addition of 10 ng/ml PrP extracted from three different prion preparations significantly reduced the survival of untreated SH-SY5Y cells by 50%, whereas the survival of SH-SY5Y cells was not affected by the addition of 10 ng/ml PrP extracted from uninfected cells (Table  II). Although these prion preparations were not pure, the neurotoxicity of these preparations was thought to be due to PrP res since these preparations were not toxic for neurons from PrPdeficient mice, and that the toxicity of these preparations was selectively removed following immunoprecipitation with an anti-PrP antibody (data not shown). The survival of cells pretreated with 100 nM squalestatin, before the addition of the prion extracts, was significantly higher than that of untreated SH-SY5Y cells incubated with the same preparations (Table  II). In addition, squalestatin also protected murine primary  To determine whether squalestatin-treated cells were completely resistant to PrP peptides, different amounts of HuPrP-(82-146) were added to SH-SY5Y cells treated with 100 nM squalestatin. The addition of HuPrP-(82-146) caused a dosedependent reduction in the survival of untreated cells, whereas even at high concentrations, HuPrP-(82-146) did not kill cells treated with 100 nM squalestatin (Fig. 5). Further experiments were performed to determine if the addition of cholesterol to squalestatin-treated SH-SY5Y cells would restore their sensitivity to the neurotoxic effect of PrP peptides or to prions. As shown in Fig. 6, the presence of 10 g/ml of water-soluble cholesterol reversed the protective effect of 100 nM squalestatin on the toxicity induced by 40 M HuPrP-(106 -126), 10 M HuPrP-(82-146) or by a prion extract in SH-SY5Y cells.
Reduced PGE 2 Production in Squalestatin-treated Cells-The cyclo-oxygenases (COX) are enzymes that convert arachidonic acid to PGs. The observations that the levels of neuronal PGE 2 are raised in response to PrP peptides, and that COX-1 inhibitors protect neurons against PrP peptides, suggest that activated COX is involved in neuronal toxicity (35,26). Thus, in this study, the effects of squalestatin on PGE 2 levels from SH-SY5Y cells incubated with PrP peptides or a prion extract were examined. We were unable to detect PGE 2 in untreated SH-SY5Y cells, squalestatin-treated cells, or cells incubated with scrambled peptides. Cells exposed to either 40 M HuPrP-(106 -126) to 10 M HuPrP-(82-146), 10 M sPrP106, or to a prion extract produced significant amounts of PGE 2 . SH-SY5Y cells pretreated with 100 nM squalestatin produced significantly less PGE 2 when subsequently incubated with these peptides or a prion extract (Table III). The effect of squalestatin was reversed by the addition of 10 g/ml of water-soluble cholesterol.
Squalestatin Protects ScN2a Cells Against Microglia-mediated Killing-The effect of squalestatin on interactions between ScN2a cells and microglia was also examined. As shown in Fig. 7A, the survival of ScN2a cells co-cultured with microglia (ratio of 1 microglia to 10 ScN2a cells) was significantly lower than those of non-infected N2a cells co-cultured with microglia, consistent with previous studies (28). The pretreatment of ScN2a cells with 100 nM squalestatin completely protected them from microglia-mediated toxicity, increasing cell viability from 30 to 100%. This protective effect could be fully   SO) or 100 nM squalestatin, and then incubated with 10 ng/ml PrP extracted from prion-infected cells (prion extracts), or 10 ng/ml PrP extracted from non-infected cells (control extracts). Cell viability was evaluated after 24 hours using MTT. Each value is the mean cell survival Ϯ S.D. from triplicate experiments repeated three times (9 observations  reversed by the addition of 10 g/ml of water-soluble cholesterol. The effect of the squalestatin on interactions between PrPdamaged, non-transformed neurons, and microglia was also studied. Thus, primary cortical neurons were treated with 100 nM squalestatin before the addition of 40 M HuPrP-(106 -126), 10 M HuPrP-(82-146), 10 M sPrP106 or a prion extract, and microglia then added 3 h later in the ratio of 1 microglia to 10 neurons. The survival of neurons in co-cultures containing squalestatin and PrP peptides or a prion extract, were significantly higher than the survival of untreated co-cultures exposed to these peptides or prion preparations (Fig. 7B). To determine whether squalestatin had a direct effect on microglia, isolated microglia cultures were pretreated with 100 nM squalestatin before the addition of 10 ng/ml LPS. There was no significant difference in the amounts of IL-6 produced by squalestatin-treated microglia (4205 pg/ml Ϯ 221, n ϭ 6) and vehicle-treated microglia (4286 pg/ml Ϯ 161).

DISCUSSION
In the present study, we examined the effect of squalestatin on cellular cholesterol levels, on the accumulation of PrP res in three prion-infected cell lines, and on two models of PrP-induced neurotoxicity. Squalestatin reduced the accumulation of PrP res in ScN2a, SMB, and ScGT1 cells. This effect of squalestatin on PrP res production was dose-dependent with an IC 50 ϳ5 nM, while even 50 M squalestatin was not toxic for ScN2a cells. None of the effects of squalestatin were reversed by the addi-tion of mevalonate (previously shown to reverse the effects of HMG-CoA reductase inhibitors such as lovastatin), but were reversed by cholesterol indicating that cholesterol depletion was responsible for the effects of squalestatin. Cholesterol homeostasis in mammalian cells is governed by both cholesterol synthesis and by the influx and efflux of cholesterol from the surrounding medium (36). Recent studies have shown that the statins reduce cerebral cholesterol levels (37) indicating that cholesterol production is important in maintaining neuronal cholesterol levels.
Several observations suggest that the maintenance of cholesterol levels in these neuronal cells is largely influenced by the synthesis of cholesterol rather than by the uptake of cholesterol from the surrounding medium. Firstly, the effects of squalestatin were observed in medium containing FCS (and hence cholesterol), and the addition of 100 g/ml of normal cholesterol did not affect the neuroprotective effect of squalestatin, or the effect of squalestatin on PrP res formation (data not shown). In contrast, low concentrations of water-soluble cholesterol (ϳ10 g/ml) were enough to reverse the neuroprotective effects of squalestatin, restore PrP res production and restore cellular cholesterol levels. Water-soluble cholesterol is complexed to methyl-␤-cyclodextrin that facilitates the insertion of cholesterol into cell membranes while normal cholesterol requires specific uptake mechanisms. Finally, drugs that prevented cellular uptake of cholesterol did not affect either PrP res formation or the neurotoxicity of PrP peptides. 2 Previous studies suggest that the concentration and cellular location of PrP C may be critical factors in the production of PrP Sc (6,7) since PrP Sc formation is thought to be dependent on the supply of PrP C to an intracellular environment that facilitates the conversion of PrP C to PrP Sc (38). Although cells treated with squalestatin contained significantly less cholesterol than untreated cells, the amounts of PrP C in untreated and squalestatin treated cells were not significantly different and, unlike treatment with the cholesterol-binding drug filipin (39), squalestatin did not cause the release of PrP C into culture medium (data not shown). In the present study squalestatin treatment resulted in the dispersion of PrP C into Triton X-100 soluble fractions, an observation compatible with previous studies that showed that cholesterol-depleted cells no longer contain typical caveolae (40). PrP C molecules are normally located in lipid rafts (8) and dispersion from such sites following squalestatin treatment may affect the normal cellular trafficking of PrP C . The golgi and the endosomal compartments are involved in the trafficking of a GFP-tagged PrP C (41) and in hippocampal neurons, the trafficking of cholera toxin (whose receptor is known to reside in lipid rafts) from endosomes to the golgi apparatus is blocked by cholesterol depletion (42). While the current results are suggestive, it is not known whether the trafficking of PrP C is affected in squalestatin-treated cells so as to prevent PrP C interacting with cellular components required for conversion to PrP Sc .
The loss of neurons in response to prions or PrP peptides is thought to be an apoptotic event (43) requiring receptor binding and the activation of specific metabolic pathways. The neurotrophin p75 receptor, located in CLDs (44), is thought to act as a receptor for PrP peptides and is involved in the neuronal damage of HuPrP-(106 -126) (45). Thus, the protective effect of squalestatin may be mediated by a destabilizing effect on neurotrophin p75. Since many of the molecules involved in cellular signaling are found in cholesterol-sensitive CLDs (46) the possibility that cholesterol depletion may affect the generation of second messen-  gers in response to PrP peptides, or to aggregated PrP Sc , was considered. For example, the components of the cyclic adenosine monophosphate (cAMP) signaling pathway are found in CLDs (47) and a recent report has linked cAMP with PrP C and neuroprotection (48). The observation that arachidonic acid (49) and the COX enzymes (50) are found in CLDs is also of interest. Levels of PGE 2 (a measure of COX activation) are increased in brain areas showing neuronal death in murine scrapie (51), and raised levels of PGE 2 are detected in the cerebrospinal fluid of patients with CJD (52)(53). Moreover, the addition of PrP peptides to SH-SY5Y cells increased PGE 2 production, and neurons treated with COX inhibitors were resistant to the toxicity of PrP peptides (26). Taken together these results suggest that PrP peptides stimulate the production of neurotoxic prostaglandins. The present studies are compatible with the hypothesis that COX activation by PrP peptides or prion extracts occurs within CLDs as PGE 2 production was greatly reduced in squalestatin-treated neurons. A significant role for microglia in the neuropathogenesis of prion disease is suggested by the close association between the accumulation of prion fibrils and the activation of microglia in scrapie-infected mice (1,30). Moreover, in vitro studies showed that microglia killed prion-damaged neurons (29) and selectively kill ScN2a cells, but not non-infected N2a cells (28). The presence of squalestatin in co-cultures containing microglia and prion-infected ScN2a cells resulted in a greatly increased neuronal survival. Since squalestatin had no effect on isolated microglia we concluded that the effects of squalestatin are predominantly on ScN2a cells. The addition of PrP peptides to primary cortical neurons induced a time-dependent changes that stimulate microglia to kill these cells (54). The protective effect of squalestatin against microglia-mediated killing is compatible with the hypothesis that these PrP-induced changes in neurons are sterol-dependent. The nature(s) of the changes expressed by ScN2a cells or PrP-damaged neurons that activate microglia are currently under further investigation.
Although previous studies showed that the formation of PrP Sc was reduced in lovastatin-treated ScN2a cells (11), the authors concluded that the statins were not compatible with use in man, partly due to the high concentrations of lovastatin required, and partly due to the fact that results were only seen in lipid-depleted medium. The concentration of squalestatin required to affect PrP res formation in three different prioninfected cell lines was shown to be less than 10 nM; considerably lower than that reported for lovastatin (0.3 M) (11). Moreover, the action of squalestatin was not affected by the cholesterol in the in medium and was only partially restored with high concentrations of water-soluble cholesterol. In contrast, the ability of 100 nM lovastatin to reduce PrP Sc production was reversed with 1 g/ml water-soluble cholesterol (data not shown) suggesting that statins and squalestatin affect different cholesterol pools. The different effects displayed by statins and squalestatin depended on their different pharmacological target (HMG-CoA reductase versus squalene synthase; see Fig. 1) and the effects of these on cholesterol distribution. Time course studies showed that squalestatin produced a prolonged reduction in cholesterol levels (significant reduction after 24 h) that was not observed in lovastatin-treated cells. The greater potency of squalestatin suggests that it may be a more promising candidate than the HMG-CoA reductase inhibitors for use in vivo.
In summary, an understanding of the processes by which prions accumulate in neurons, and how this process leads to neuronal loss, may aid the development of therapies to delay the progression of prion diseases. In the present study, squalestatin had a profound effect on three major aspects of the neuropathogenesis of prion disease. Squalestatin-treated ScN2a, SMB, or ScGT1 cells were cured of PrP res , ScN2a cells were protected against killing by microglia and neurons treated with squalestatin were resistant to the neurotoxicity of PrP peptides. Furthermore, the observation that squalestatin had protective effects at low concentrations, suggest that the use of this drug to treat prion diseases should be considered.