Sequestration of free cholesterol in cell membranes by prions correlates with cytoplasmic phospholipase A2 activation

Background The transmissible spongiform encephalopathies (TSEs), otherwise known as the prion diseases, occur following the conversion of the normal cellular prion protein (PrPC) to an alternatively folded isoform (PrPSc). The accumulation of PrPSc within the brain leads to neurodegeneration through an unidentified mechanism. Since many neurodegenerative disorders including prion, Parkinson's and Alzheimer's diseases may be modified by cholesterol synthesis inhibitors, the effects of prion infection on the cholesterol balance within neuronal cells were examined. Results We report the novel observation that prion infection altered the membrane composition and significantly increased total cholesterol levels in two neuronal cell lines (ScGT1 and ScN2a cells). There was a significant correlation between the concentration of free cholesterol in ScGT1 cells and the amounts of PrPSc. This increase was entirely a result of increased amounts of free cholesterol, as prion infection reduced the amounts of cholesterol esters in cells. These effects were reproduced in primary cortical neurons by the addition of partially purified PrPSc, but not by PrPC. Crucially, the effects of prion infection were not a result of increased cholesterol synthesis. Stimulating cholesterol synthesis via the addition of mevalonate, or adding exogenous cholesterol, had the opposite effect to prion infection on the cholesterol balance. It did not affect the amounts of free cholesterol within neurons; rather, it significantly increased the amounts of cholesterol esters. Immunoprecipitation studies have shown that cytoplasmic phospholipase A2 (cPLA2) co-precipitated with PrPSc in ScGT1 cells. Furthermore, prion infection greatly increased both the phosphorylation of cPLA2 and prostaglandin E2 production. Conclusion Prion infection, or the addition of PrPSc, increased the free cholesterol content of cells, a process that could not be replicated by the stimulation of cholesterol synthesis. The presence of PrPSc increased solubilisation of free cholesterol in cell membranes and affected their function. It increased activation of the PLA2 pathway, previously implicated in PrPSc formation and in PrPSc-mediated neurotoxicity. These observations suggest that the neuropathogenesis of prion diseases results from PrPSc altering cholesterol-sensitive processes. Furthermore, they raise the possibility that disturbances in membrane cholesterol are major triggering events in neurodegenerative diseases.


Background
Cholesterol levels within the brain may affect the pathogenesis of some neurodegenerative diseases including Alzheimer's and Parkinson's diseases and multiple sclerosis [1,2]. Neuronal cholesterol levels are also thought to affect the progression of the transmissible spongiform encephalopathies (TSEs), otherwise known as prion diseases [3]. These diseases are associated with the conversion of the normal cellular prion protein (PrP C ) to an alternatively folded isoform (PrP Sc ) [4]. The accumulation of PrP Sc is closely associated with the main pathological features of TSEs: the spongiform degeneration of the brain, synaptic alterations, glial cell activation and extensive neuronal loss [5,6]. While a recent study reported that prion infection in vivo was associated with changes in brain cholesterol levels [7], the change in cholesterol regulation in neurons following prion infection has not been characterised extensively. Furthermore, because the brain is composed of diverse cell types, it is possible that changes in the cholesterol content of neurons may be obscured in mixed cell populations or whole brain studies. To reduce the problem of cell heterogeneity, the effects of prion infection on two neuronal cell lines were examined. We report that prion infection is associated with increased amounts of free cholesterol in the cell membrane, but also with reduced amounts of cholesterol esters suggesting that prion infection alters cholesterol regulation. The effects of prion infection on cholesterol balance were reproduced in primary cortical neurons incubated with exogenous PrP Sc preparations.
Disturbing cholesterol metabolism within cells may have profound effects on cell function. Although cholesterol is a component of normal cell membranes, the amounts of free cholesterol are increased between three-and five-fold in specialised detergent-resistant micro-domains within the plasma membrane that are commonly called lipid rafts [8]. Such lipid rafts are also highly enriched in sphingolipids and gangliosides, and contain specific proteins [9]. The raft-associated proteins include many proteins attached to membranes via a glycosylphosphatidylinosi-tol (GPI) anchor [10] including both PrP C and PrP Sc [11]. In addition, cellular receptors for folate or the p75 neurotrophin receptor are found within rafts [12,13], as are receptors for neurotransmitters including acetylcholine [14] and gamma-aminobutyric acid [15]. Such domains also contain components of signalling pathways including the Src family tyrosine kinases [16], adenylyl cyclase [17], the trimeric G-proteins [18] and cytoplasmic phospholipase A 2 (cPLA 2 ) [19]. Lipid rafts act as membrane platforms that concentrate molecules for cell signalling [20] and changes in membrane cholesterol levels may lead to abnormal cell signalling. As the neurotoxicity of PrP Sc was blocked by PLA 2 inhibitors [21] the effects of prion infection on PLA 2 activity was examined. Here we report increased activation (phosphorylation) of cPLA 2 in ScGT1 cells.

Prion infection increased free cholesterol in neuronal cell lines
The amounts of protein and cholesterol in two prioninfected neuronal cell lines (ScN2a and ScGT1 cells) were compared to that of uninfected controls (N2a and GT1 cells). There were no significant differences in the amounts of protein between infected and uninfected cells. In contrast, the amounts of total cholesterol (a mixture of free and esterified cholesterol) were significantly higher in infected ScGT1 cells than in GT1 cells (542 ng cholesterol/ mg protein ± 44 versus 453 ± 72, n = 11, P = 0.004) ( Table  1). More detailed analysis showed that the amounts of free cholesterol within ScGT1 cells were 36% higher than those in GT1 cells (500 ± 54 versus 368 ± 59, n = 11, P = 0.0003), while the amounts of esterified cholesterol were 50% less than in GT1 cells (42 ± 14 versus 85 ± 28, n = 11, P = 0.0007). Similar results were obtained when ScN2a and N2a cells were compared: amounts of free cholesterol in ScN2a cells were 23% higher than in N2a cells (473 ± 41 versus 384 ± 37, n = 11, P = 0.0001), but the amounts of esterified cholesterol were significantly lower than those of N2a cells (52 ± 14 versus 87 ± 19, n = 11, P = 0.002). Thus, in both cell lines prion infection was associ- ated with a significant decrease in the amounts of cholesterol esters and in the percentage of cholesterol that was esterified (Table 1).
Brain-derived neurotrophic factor (BDNF) increased the PrP Sc content of ScGT1 cells [22]. Here we report that treatment with BDNF, glial-derived neurotrophic factor (GDNF) or retinoic acid also increased the PrP Sc content of ScGT1 cells, while treatment with nerve-growth factor (NGF) did not ( Table 2). The increased PrP Sc content of treated ScGT1 cells was accompanied by increased amounts of free cholesterol. The free cholesterol content of ScGT1 cells was significantly higher in cells treated with BDNF (654 ng cholesterol/mg protein ± 61 versus 510 ± 48, n = 9, P = 0.003), GDNF (655 ± 59 versus 510 ± 48, n = 9, P = 0.004) or retinoic acid (705 ± 83 versus 510 ± 48, n = 9, P = 0.002) but not in cells treated with NGF (503 ± 72 versus 510 ± 48, n = 9, P = 0.64). None of the treatments increased the cholesterol content of uninfected GT1 cells showing that the increases in free cholesterol were related to the PrP Sc content of cells. To examine this relationship further, ScGT1 cells were treated with varying concentrations of GDNF and amounts of PrP Sc and free cholesterol were measured. A significant correlation coefficient between the amounts of PrP Sc and free cholesterol was observed (Pearson correlation = 0.922); see Figure 1.

PrP Sc increases the free cholesterol content of cortical neurons
As the above observations were on prion-infected neuronal cell lines, we sought to determine whether PrP Sc had the same effect on non-transformed cells. Primary cortical neurons were incubated with sub-lethal amounts of PrP Sc , or equivalent amounts of PrP C for 48 hours. The addition of PrP Sc increased the amounts of free cholesterol when compared with untreated cells or cells treated with PrP C (Figure 2). The amounts of free cholesterol were significantly higher in neurons treated with 100 pg/ml PrP Sc than in untreated cells (704 ng/mg protein ± 73 versus 504 ± 58, n = 9, P = 0.0019). Similarly, the amounts of free cholesterol were significantly higher in neurons treated with 20 pg/ml PrP Sc (632 ± 46 versus 504 ± 58, n = 9, P = 0.004). There was no significant difference in the amounts of free cholesterol in untreated neurons and in neurons incubated with 100 pg/ml PrP C (504 ± 58 versus 517 ± 46, n = 9, P = 0.51).   Correlation between the amounts of PrP Sc and free choles-terol in ScGT1 cells

Stimulating cholesterol synthesis increases cholesterol esters but not free cholesterol in neurons
The possibility that prion infection stimulated cholesterol synthesis was examined by comparing the effects of PrP Sc with those of increased cholesterol biosynthesis in cortical neurons. Mevalonate is a precursor of cholesterol synthesis that is a product of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, the rate-limiting step in cholesterol synthesis [23]. Treatment with 100 µM mevalonate significantly increased the amounts of total cholesterol in neurons (562 ng cholesterol/mg protein ± 45 versus 482 ± 54, n = 9, P = 0.001); this increase consisted primarily of cholesterol esters (96 ± 25 versus 42 ± 24, n = 9, P = 0.006) as the amounts of free cholesterol were unchanged (466 ± 50 versus 440 ± 46, n = 9, P = 0.16). Similar results were obtained in cells treated with 10 µM cholesterol, which increased the amounts of cholesterol esters (130 ± 55 versus 42 ± 24, n = 9, P = 0.001) but not free cholesterol (446 ± 36 versus 440 ± 46, n = 9, P = 0.54); see Figure 3. The percentage of cholesterol that was esterified in untreated cells (9% ± 4) was raised to 17% ± 5 in neurons incubated with mevalonate, and 22% ± 8 in neurons treated with cholesterol. The addition of mevalonate or cholesterol did not affect the protein content of cell extracts.

Prion infection increased activation of cPLA 2 in neuronal cell lines
As lipid rafts act as platforms in which signalling complexes assemble [24], the possibility that the altered composition of lipid rafts in prion-infected cells affected cell signalling was examined. More specifically, as PLA 2 was required for prion formation [25], the amounts of activated cPLA 2 in cells were examined. The amounts of activated (phosphorylated) cPLA 2 in ScGT1 cells were greater than those in GT1 cells ( Figure 4A). The relationship between activated cPLA 2 and PrP Sc was examined in ScGT1 cells treated with different neurotrophins or retinoic acid. The increased PrP Sc content of treated ScGT1 cells was accompanied by increased amounts of activated cPLA 2 . The amounts of activated cPLA 2 in ScGT1 cells was significantly higher in cells treated with BDNF (514 units/ml ± 69 versus 395 ± 33, n = 9, P = 0.01), GDNF (542 ± 65 versus 395 ± 33, n = 9, P = 0.008) or retinoic acid (553 ± 79 versus 395 ± 33, n = 9, P = 0.01) but not in cells treated with NGF (416 ± 46 versus 395 ± 33, n = 9, P = 0.41). In contrast, none of these treatments significantly increased the amounts of activated cPLA 2 in GT1 cells showing that increased activation of cPLA 2 is related to the PrP Sc content of cells (Table 3). The relationship between cPLA 2 and PrP Sc was examined further in ScGT1 cells treated with varying concentrations of GDNF. The correlation coefficient between the amounts of PrP Sc in cells and the amounts of activated cPLA 2 was significant (Pearson correlation = 0.865); see Figure 4B.

Rafts containing PrP Sc also contain cPLA 2
Immunoprecipitation was used to determine whether cPLA 2 was associated with PrP-containing lipid rafts in ScGT1 cells. Mab 4F2, which recognises both PrP C and PrP Sc , precipitated cPLA 2 out of membrane extracts from untreated ScGT1 cells. Two methods were used to show that cPLA 2 was associated with PrP Sc rather than PrP C in these ScGT1 cell extracts. First, immunoprecipitation with mab IC18, which recognises PrP C but not PrP Sc , did not precipitate cPLA 2 from ScGT1 cells. Second, immunoprecipitation with mab 4F2 precipitated cPLA 2 from ScGT1 cells from which PrP C had been removed following treatment with PI-PLC ( Figure 5A). A mab to CD55, or an IgG 2 isotype control, did not precipitate cPLA 2 from ScGT1 cells. Next we examined the distribution of activated cPLA 2 within ScGT1 cells. The amounts of activated cPLA 2 in whole cell extracts (100%) were compared with those in membranes precipitated with mab 4F2 and to the depleted membrane extract. Greater than 60% of activated   Prion infection increases the amounts of activated cPLA 2 in ScGT1 cells cPLA 2 was found in the immunoprecipitated membrane fraction ( Figure 5B).

Discussion
The major goal of this study was to investigate the impact of PrP Sc on the biochemistry of cell membranes. The amounts of total cholesterol in membranes were significantly higher in prion-infected cell lines than in their uninfected counterparts and there was a significant correlation between amounts of cholesterol and PrP Sc . More specifically, prion infection was associated with a significant increase in the amounts of free cholesterol. While much of what is known about the role of cholesterol in cell membranes is surmised from the changes in cells brought about by cholesterol depletion, either from cholesterol synthesis inhibitors or via cholesterol extraction, little is known about how neurons respond when the cholesterol content of membranes is increased.
How the presence of prions affects cholesterol levels remains to be determined. A synthetic prion-derived peptide activated HMG-CoA reductase suggesting a mechanism by which prion infection increased cholesterol production [26]. However, we were unable to replicate the effects of prion infection in non-infected cells by stimulating cholesterol biosynthesis or by adding exogenous cholesterol. The addition of mevalonate or cholesterol did not increase the amounts of free cholesterol in cell membranes; rather, they increased the amounts of cholesterol esters. This contrasts with the situation in prion-infected cells where the amounts of cholesterol esters were reduced. Cholesterol in cells is found either as free cholesterol in membranes or as cholesterol esters in cytoplasmic droplets. A dynamic equilibrium between the pools of free cholesterol and cholesterol esters is tightly controlled by acyl-coenzyme A:cholesterol acyltransferase (ACAT), an endoplasmic reticulum (ER)-resident enzyme that catalyses the formation of cholesterol esters from cholesterol and long-chain fatty acids [27]. In uninfected cells excess free cholesterol activates ACAT resulting in increased production of cholesterol esters. In these cells, increased free cholesterol levels were only seen following the addition of a combination of free cholesterol and an ACAT inhibitor (data not shown).
The situation in prion-infected cells, where the increased amounts of free cholesterol is accompanied by reduced amounts of cholesterol esters, is unusual. The increase in free cholesterol and the reduction of cholesterol esters in prion-infected cells may be a result of direct inhibition of ACAT or by sequestration of cholesterol in micro-environments that avoid ACAT. The amount of cholesterol in cell membranes is partly determined by its fatty acid composition. The high incidence of saturated fatty acids attached to sphingolipids, gangliosides and GPI-anchored proteins allows tight molecular packing and increases the solubilisation of free cholesterol [28]. Thus, the formation of PrP Sc may have a direct effect on the composition of cell membranes as the self-aggregation of PrP Sc results in the clustering of GPI anchors attached to PrP Sc . The increased density of saturated fatty acids within PrP Sc -containing micro-domains encourages the solubilisation of free cholesterol and the remodelling of those membranes.
Increasing the free cholesterol content of membranes is thought to reduce membrane fluidity and subsequently affect the endocytosis and trafficking of proteins. Therefore, the formation of PrP Sc may alter conventional lipid raft structure and the PrP C -protein interactions that occur within lipid rafts. For example, PrP C has been reported to bind to caveolin-1 [29] or N-CAM [30], proteins that reside within lipid rafts. It is unclear whether these protein-protein interactions are affected following the conversion of PrP C to PrP Sc . The sequestration of free Activated cPLA 2 is associated with PrP Sc in ScGT1 cells cholesterol into PrP Sc -containing lipid rafts may deplete free cholesterol from other cellular pools where it helps to stabilise the packing of sphingolipids, gangliosides and raft-associated proteins in the membrane. This may affect the function of such proteins. For example, free cholesterol affects the formation and function of synapses [31]. Therefore, sequestration of cholesterol by PrP Sc may affect synaptic transmission, a hypothesis supported by observations that ScGT1 cells contain altered amounts of synaptic proteins including synaptophysin [32] and that synapse damage is seen during the early stages of experimental prion diseases [6].
Cholesterol-dependent micro-domains are increasingly implicated as platforms necessary for cell signalling [24] and scrapie infections of neuronal cells are associated with increased levels of Src kinase [33]. The activation of PLA 2 that is necessary for prion formation [25] is reduced in cholesterol-depleted cells suggesting that this enzyme may reside within a lipid raft [34]. Here we show that ScGT1 cells contained four times as much activated cPLA 2 as GT1 cells and there was a significant correlation between amounts of activated cPLA 2 and PrP Sc . Immunoprecipitation studies showed that activated cPLA 2 co-localised with PrP Sc -containing lipid rafts in ScGT1 cells.
Previous studies showed that cell activation results in the translocation of cPLA 2 to endoplasmic and plasma membranes [35]. Our observations are consistent with the hypothesis that prion infection stimulates the translocation of cPLA 2 to lipid rafts containing PrP Sc . The activation of cPLA 2 is associated with the production of prostaglandins and the amounts of PGE 2 produced by ScGT1 cells were significantly higher than that of GT1 cells. Our observation that pre-treatment of ScGT1 cells with PLA 2 or cyclo-oxygenase inhibitors reduced PGE 2 production showed that PGE 2 was a valid measure of PLA 2 activity in these cells. These findings are consistent with reports of increased PGE 2 in murine scrapie [36] and raised levels of PGE 2 in the cerebrospinal fluid of patients with Creutzfeldt-Jakob disease [37].

Conclusion
We have demonstrated that the presence of PrP Sc increased the free cholesterol content of cell membranes. The increased free cholesterol could not be replicated by the stimulation of cholesterol synthesis or by the addition of exogenous free cholesterol, which increased the amounts of cholesterol esters. Our observations are consistent with the hypothesis that the clustering of saturated fatty acids, parts of the GPI anchors attached to PrP Sc , increased the amounts of free cholesterol solubilised within the plasma membrane which increased membrane rigidity. These changes in cell membranes could reduce endocytosis and the recycling of cholesterol through the ER where it is exposed to ACAT, consistent with reduced cholesterol ester production in infected cells. The increased amounts of free cholesterol in the plasma membrane were associated with increased activation of the PLA 2 pathway that is necessary for PrP Sc -mediated neurotoxicity. This is a rare example of an infective agent increasing free cholesterol levels within cell membranes and raises the possibility that disturbances in membrane cholesterol are major triggering events in neurodegenerative diseases.

Cell lines
Prion-infected ScGT1 cells, from a murine hypothalamic neuronal cell line infected by the Chandler scrapie isolate and ScN2a neuroblastoma cells, were grown in Hams F12 medium supplemented with 2 mM glutamine, 2% foetal calf serum (FCS) and standard antibiotics (100 U/ml penicillin and 100 µg/ml streptomycin; Invitrogen, Paisley, UK). Uninfected N2a or GT1 cells were used as noninfected controls and grown in the same medium. To determine the effect of neurotrophins or retinoic acid, cells were plated at 1 × 10 5 cells/well in 6 well plates. Cells were then grown with daily changes of media for 7 days.

Neuronal cultures
Primary cortical neurons were prepared from the brains of mouse embryos (day 15.5) after mechanical dissociation, cell sieving and isolation on histopaque (Sigma). Neuronal precursors were plated (1,000,000 cells/well in 24 well plates coated with 5 µg/ml poly-L-lysine) in Hams F12 containing 5% FCS for 2 hours. Cultures were shaken (600 rpm for 5 minutes) and non-adherent cells removed by two washes in phosphate buffered saline (PBS). Neurons were grown in neurobasal medium (NBM) containing B27 components (Invitrogen) for 7 days and subsequently incubated with test compounds. Immunolabelling studies showed that after 7 days cultures contained less than 5% glial cells (about 3% GFAP positive and less than 1% MAC-1 positive cells).

Cell extracts
At the end of the treatment, cells were washed twice in PBS and homogenised in an extraction buffer containing 10 mM Tris-HCl, 100 mM NaCl, 10 mM EDTA, 0.5% Nonidet P-40, 0.5% sodium deoxycholate and 0.2% sodium dodecyl sulphate (SDS) at 1 × 10 6 cells/ml. Mixed protease inhibitors (AEBSF, Aprotinin, Leupeptin, Bestain, Pepstatin A and E-46; from Sigma) were added to some cell extracts. Membranes were prepared by repeated passage with a Wheaton homogeniser; nuclei and large fragments were removed by centrifugation (300 × g for 5 minutes).
To determine the amount of PrP Sc in cells these supernatants were digested with 1 µg/ml proteinase K for 1 hour at 37°C, digestion was stopped with mixed protease inhibitors. The soluble material was heated to 95°C for 5 IgG conjugated to peroxidise and an enhanced chemiluminescence kit (Amersham Biotech).

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.
Publish with Bio Med Central and every scientist can read your work free of charge