Overstimulation of PrPC Signaling Pathways by Prion Peptide 106-126 Causes Oxidative Injury of Bioaminergic Neuronal Cells*

Transmissible spongiform encephalopathies, also called prion diseases, are characterized by neuronal loss linked to the accumulation of PrPSc, a pathologic variant of the cellular prion protein (PrPC). Although the molecular and cellular bases of PrPSc-induced neuropathogenesis are not yet fully understood, increasing evidence supports the view that PrPSc accumulation interferes with PrPC normal function(s) in neurons. In the present work, we exploit the properties of PrP-(106-126), a synthetic peptide encompassing residues 106-126 of PrP, to investigate into the mechanisms sustaining prion-associated neuronal damage. This peptide shares many physicochemical properties with PrPSc and is neurotoxic in vitro and in vivo. We examined the impact of PrP-(106-126) exposure on 1C11 neuroepithelial cells, their neuronal progenies, and GT1-7 hypothalamic cells. This peptide triggers reactive oxygen species overflow, mitogen-activated protein kinase (ERK1/2), and SAPK (p38 and JNK1/2) sustained activation, and apoptotic signals in 1C11-derived serotonergic and noradrenergic neuronal cells, while having no effect on 1C11 precursor and GT1-7 cells. The neurotoxic action of PrP-(106-126) relies on cell surface expression of PrPC, recruitment of a PrPC-Caveolin-Fyn signaling platform, and overstimulation of NADPH-oxidase activity. Altogether, these findings provide actual evidence that PrP-(106-126)-induced neuronal injury is caused by an amplification of PrPC-associated signaling responses, which notably promotes oxidative stress conditions. Distorsion of PrPC signaling in neuronal cells could hence represent a causal event in transmissible spongiform encephalopathy pathogenesis.

Transmissible spongiform encephalopathies (TSE), 4 also known as prion diseases, are a group of neurodegenerative disorders that include Creutzfeldt-Jakob disease, Gerstmann-Straussler-Scheinker syndrome, and fatal familial insomnia in humans, as well as scrapie in sheep, bovine spongiform encephalopathy in cattle, and chronic wasting disease in deer and elk (1,2). All of these fatal affections are characterized by widespread neuronal dysfunction and loss, spongiosis, and accumulation of PrP Sc , the pathological isoform of a host-encoded prion protein (PrP C ).
PrP C is a glycosylphosphatidylinositol-anchored protein mainly expressed at the cell surface of neurons. Conformational conversion of PrP C into PrP Sc lies at the root of prion diseases (1). However, the causal link between TSE neuropathology and PrP Sc accumulation remains enigmatic. Expression of glycosylphosphatidylinositol-anchored PrP C at the cell surface of neurons appears to be an essential requirement for PrP Sc to exert its toxicity (3,4). Indeed, neuronal depletion of PrP C in infected mice hinders TSE-associated neuronal injury, whereas abundant PrP Sc accumulation still occurs in glial cells (5). In addition, neurons from transgenic mice expressing a prion protein lacking the glycosylphosphatidylinositol moiety are not vulnerable to PrP Sc (4).
It is currently suspected that prion pathology involves alterations of normal PrP C function(s) as a result of the conversion of PrP C into PrP Sc . Observations that PrP null (PrPϪ/Ϫ) mice exhibit only minor phenotypic abnormalities and that PrP C depletion in neurons of adult mice results in healthy animals, with no evidence of neuronal loss (6), argue against the idea that PrP C loss of function would be a major cause of prion-induced neurodegeneration (7). Instead, in a gain of function hypothesis, prion pathology could arise from a PrP Sc -induced overflow of PrP C -coupled signals that normally take part in neuron homeostasis.
Evidence supporting a role for PrP C in cell signaling has come from our previous studies showing that, in fully differentiated 1C11 5-HT serotonergic or 1C11 NE noradrenergic neuronal cells derived from the 1C11 neuroectodermal cell line, antibodymediated cross-linking of PrP C triggers a caveolin-1-dependent activation of Fyn (8). PrP C ligation with specific antibodies was further shown to elicit a NADPH oxidase-dependent reactive oxygen species (ROS) production in cells of either neuronal or non-neuronal origin, including 1C11 precursor cells, 1C11 5-HT and 1C11 NE neuronal cells, GT1-7 neurohypothalamic cells, and BW5147 T lymphocytes (9). ROS synthesized in response to PrP C ligation then act as "second message" signals and trigger rapid phosphorylation of extracellular signal-regulated kinases (ERK) 1/2. In 1C11-derived bioaminergic neuronal cells, these targets are under the control of the PrP C -Caveolin-Fyn platform that confers neurospecificity to PrP C signaling (9). Altogether, these observations allow consideration of PrP C as a regulator of the cellular redox equilibrium and of cell homeostasis.
Oxidative stress, which results from an imbalance of cell redox equilibrium, i.e. high levels of ROS and dramatic reduction in the activities of antioxidant systems (10,11), is widely suspected to be responsible for neuronal cell death associated with prion infection (12)(13)(14). The demonstration of a biochemical link between PrP C stimulation and intracellular ROS production (9) raises the possibility that pathogenic prions promote oxidative stress conditions through deviation of PrP Cassociated signaling function. To challenge this idea, 1C11 neuroepithelial cells and their neuronal progenies, as well as GT1-7 hypothalamic cells, were exposed to the amyloidogenic prion peptide 106 -126 (PrP-(106 -126)). This peptide represents a suitable model to study the deleterious effect of PrP Sc because it shares several physicochemical properties with entire PrP Sc and is neurotoxic in vitro as well as in vivo (15)(16)(17)(18)(19)(20)(21)(22)(23). The present study shows a selective sensitivity of 1C11 5-HT and 1C11 NE bioaminergic neurons to PrP-(106 -126). The pathogenic peptide mobilizes apoptotic pathways in neuronal cells while having no effect on cells of non-neuronal origin including 1C11 precursor and GT1-7 hypothalamic cells. We establish that the PrP-(106 -126)-associated neurotoxic effect observed depends on cell surface expression of PrP C and relates to a deviation of PrP C -associated signaling function. Exposure to the peptide promotes NADPH oxidase overstimulation, ROS accumulation, oxidative stress injury, and recruitment of apoptotic signals.
The NADPH oxidase activity was selectively inhibited using DPI (Sigma). PP2 (Calbiochem) was used to inhibit Fyn kinase. Enzyme inhibitions were performed by pretreating cells at 37°C for 30 min in their culture medium with the appropriate inhibitor prior to exposure to PrP-(106 -126) or PrP-(Scr).
For small interfering RNA (siRNA) transfection, siRNAs corresponding to the mouse Prnp gene sequence (codon 167-173) were synthesized by Proligo. Typically, siRNAs were designed such as to be constituted of 19 ribonucleotides followed by two extra thymidine bases at the 3Ј end overhang on both strands. Specific siRNA sequences used were: 5Ј-CAGUACAGCAAC-CAGAACATT-3Ј (sense) and scramble 5Ј-AACGAUGACAC-GAACACACTT-3Ј (sense). Annealing for duplex siRNA formation was performed by incubation at 90°C for 1 min in 30 mM HEPES, pH 7.4, containing 100 mM kalium acetate, and 2 mM magnesium acetate, followed by a 60-min incubation at 37°C. siRNA duplexes (10 M) were stored at Ϫ20°C until use. 1C11 precursor cells and their neuronal progenies, left to grow to 30 -40% confluence, were transfected with the Lipofectamine 2000 reagent according to the manufacturer's instructions (Invitrogen). The cells were exposed to PrP-(106 -126) 36 h after transfection.
Cell Contour Immunostaining with Anti-CD9 Antibody-PrP-(106 -126)-treated and mock treated cells for 24 h were washed with cold buffer A (PBS supplemented with 1 mM Ca 2ϩ , 1 mM Mg 2ϩ , and 0.1% azide), and incubated for 1 h at room temperature with rat anti-tetraspanin CD9 antibody (2.5 g/ml) in blocking buffer (buffer A enriched with 2% fetal calf serum) and then with fluorescein isothiocyanate-conjugated rabbit anti-rat immunoglobulins (5 g/ml) (Jackson Laboratories). The cells were fixed with 3.6% formaldehyde in PBSglycine (1.5 mg/ml). Immunolabeling was observed with a fluorescence microscope (Zeiss Leica).
Cell Surface Detection of PrP C and PrP-(106 -126) by Confocal Microscopy-1C11 5-HT and 1C11 NE neuronal cells grown on glass coverslips and exposed to PrP-(106 -126) were washed with cold buffer A and incubated simultaneously with antibodies staining either PrP-(106 -126) (3F4, 5 g/ml) or PrP C (1A8, 10 g/ml) in blocking buffer for 1 h at room temperature. Texas Red-conjugated anti-mouse (Jackson Laboratories) and fluorescein isothiocyanate-conjugated anti-rabbit (Jackson Laboratories) immunoglobulins (5 g/ml) were used as secondary antibodies. The cells were then fixed with 3.6% formaldehyde in PBS-glycine (1.5 mg/ml). Immunolabeling was observed with a Leica TCS-NT/SP inverted confocal laser scanning microscope using an Apochromat ϫ63/1.32 oil immersion objective (Zeiss Leica).
Cell Extract Preparation and Western Blot Analyses-Extracts of 1C11 precursor cells, their 1C11 5-HT and 1C11 NE neuronal progenies, and GT1-7 hypothalamic cells were prepared as described in Ref. 9. Protein concentrations in the lysates were measured using the bicinchoninic acid method (Pierce), and Western blots were performed as previously described (9).
Immunoprecipitation of Cell Surface PrP C -1C11 precursor cells and their neuronal progenies were labeled in ice-cold PBS supplemented with 1 mM Ca 2ϩ , 1 mM Mg 2ϩ (buffer B) containing 0.5 mg/ml EZ-link TM sulfo-NHS-LC-biotin (Pierce) for 1 h. The cells were lysed on ice after three washes in ice-cold buffer B as described in Ref. 9. Immunoprecipitation was carried out on 200 g of protein lysate using protein A-Sepharose beads (Amersham Biosciences) coupled to monoclonal SAF32 antibody. Immunoprecipitation was performed overnight at 4°C under gentle mixing. Biotin-labeled PrP C was eluted from protein A-Sepharose beads with gel loading buffer containing 1% ␤-mercaptoethanol. Immunoprecipitates were analyzed by Western blotting, and biotin-labeled PrP C was revealed using a streptavidin-horseradish peroxidase conjugate (Vector Laboratories, Inc.).
Immunoprecipitation of PrP-(106 -126)-PrP-(106 -126) was immunoprecipitated by using protein A-Sepharose beads (Amersham Biosciences) coupled to 3F4 antibody (30 g/ml). Aliquots corresponding to 100 g of proteins of cell extracts were incubated with 50 l of beads coupled to 3F4 antibodies overnight at 4°C under gentle mixing. The beads were then washed in high salt buffer (0.5 M NaCl) and heated to 100°C in loading buffer without reducing agent. Immunoprecipitates were analyzed by Western blot.
Fluorescence Measurement of Intracellular Reduced Glutathione-The level of GSH was determined using the GSHsensitive probe Celltracker Green CMFDA (Molecular Probes). Adherent 1C11, 1C11 5-HT , 1C11 NE , and GT1-7 cells were exposed to 25 M PrP-(106 -126) or PrP-(Scr) up to 24 h. The cells were then washed twice with Hanks' balanced salt solution buffer (Invitrogen) and further incubated for 30 min at 37°C in Hanks' balanced salt solution in the presence of 1 M fluorogenic reagent. Hanks' balanced salt solution was removed, and the cells were left to reconstitute in Dulbecco's modified Eagle's medium, 10% fetal calf serum for 30 min at 37°C before lysis in NET buffer. Fluorescence intensity of cell lysates was recorded at ϭ 517 nm (slit width ϭ 5 nm) with excitation at ϭ 492 nm (slit width ϭ 5 nm) using a Cary Eclipse fluorometer (Varian Inc.). The reference level of intracellular reduced GSH (100%) was obtained using mock treated cells.
PrP-(106 -126) Triggers Apoptotic Signals in 1C11-derived Bioaminergic Neuronal Cells-In scrapie-infected mice, dendritic atrophy and synaptic alteration of hippocampal pyramidal neurons have been reported to be associated with cell death through apoptotic mechanisms (26 -29). Accordingly, PrP-(106 -126) treatment of 1C11 5-HT and 1C11 NE cells caused a reduction in cell viability. By 6 h, nearly 40% of 1C11-derived neuronal cells died, as estimated by following MTT reduction to formazan. On another hand, PrP-(106 -126) had no effect on the viability of 1C11 undifferentiated cells or GT1-7 cells. To investigate whether the toxic effect of PrP-(106 -126) on 1C11 5-HT and 1C11 NE cells was related to an activation of apoptotic signaling pathways, mitochondrial membrane depolarization-dependent cytochrome c release and caspase-3 activation were followed. As depicted in Fig. 2A, cytochrome c release in 1C11 NE and 1C11 5-HT cells could be detected as soon as 60 and 120 min, respectively, after PrP-(106 -126) exposure (25 M). Subsequent caspase-3 activation was observed in PrP-(106 -126)-treated 1C11 NE and 1C11 5-HT cells by 240 and 360 min, respectively (Fig. 2B). Again, the latter effects were restricted to fully differentiated neuronal 1C11 5-HT and 1C11 NE cells, with no trace of cytochrome c release or caspase-3 activation in 1C11 precursor and GT1-7 hyptothalamic cells up to 24 h of exposure to PrP-(106 -126). The activation of such apoptotic markers in 1C11 5-HT serotonergic and 1C11 NE noradrenergic neuronal cells exposed to the prion peptide followed a dose-response relationship (Fig.  2C). Cytochrome c release (not shown) and caspase-3 activation (Fig. 2C) became detectable when neuronal cells were treated with 5 M PrP-(106 -126). The effect was maximal when PrP-(106 -126) concentration reached 25 M up to 50 M. As expected, PrP-(Scr), used as control, had no impact on the two above apoptotic hallmarks up to 24 h of exposure whatever the cell type examined (Fig. 2, A and B). Altogether, our results demonstrate that PrP-(106 -126) elicits apoptotic signals in 1C11-derived cells harboring a neuronal phenotype only.
PrP-(106 -126)-induced Apoptotic Reponses in 1C11 5-HT and 1C11 NE Cells Depend on Cell Surface Expression of PrP C -To investigate whether the expression of PrP C at the neuronal cell surface was required to mediate PrP-(106 -126)dependent apoptotic pathways in 1C11 5-HT and 1C11 NE cells, we designed a mouse Prnp gene-specific siRNA (PrP-siRNA) with the aim to down-regulate the expression of the PrP C protein. As shown by Western blot analysis and cell surface immunostaining, expression of PrP C was specifically reduced by ϳ95% upon 1C11 5-HT and 1C11 NE cell transfection with this PrP-siRNA (Fig. 3A). On the other hand, the PrP-siRNA had no impact on the expression of unrelated proteins such as ERK1/2. A scrambled siRNA unrelated to any known gene was used as control. It did not change the level of PrP C expression in 1C11 5-HT and 1C11 NE cells (Fig. 3A).
PrP C -depleted 1C11 5-HT and 1C11 NE cells were exposed to 25 M PrP-(106 -126). Under these conditions, neither cytochrome c release nor subsequent caspase-3 activation could ever be detected, up to 24 h of treatment (Fig. 3B). Apoptotic responses were still observed in 1C11-derived bioaminergic cells transfected with the scrambled siRNA (Fig. 3B). The observation that PrP-(106 -126) depends on PrP C expression to recruit apoptotic pathways in 1C11 5-HT and 1C11 NE neuronal cells prompted us to probe an interaction of PrP-(106 -126) with PrP C . We first carried out immunofluorescence experiments and selectively followed PrP-(106 -126) using the 3F4 monoclonal antibody, which does not cross-react with mouse PrP C . PrP C was visualized with the help of the 1A8 polyclonal antibody targetting a PrP C epitope that is C-terminal to the 106 -126 region. 1C11 5-HT and 1C11 NE cells were exposed to 25 M PrP-(106 -126) for 30 min and immunostained for both PrP C (Fig. 4A, red) and PrP-(106 -126) (Fig. 4A, green). As for Cell surface of 1C11, 1C11 5-HT , 1C11 NE , and GT1-7 cells was immunostained using CD9-specific antibodies and visualized by confocal microscopy. CD9 is a tetraspanin homogeneously distributed throughout the cell surface of all cells (panels a, d, g, and j ). Exposure to 25 M PrP-(106 -126) (panels c, f, i, and l ) triggers morphological alterations in 1C11 5-HT and 1C11 NE cells, affecting both the neurites (arrows 1-5) and cell bodies (arrows 6 and 7) while having no effect with 1C11 and GT1-7 cells. Scrambled peptide PrP-(Scr) (panels b, e, h, and k) does not alter cell morphology whatever the cell type. The scale bar, applicable to all panels, is 10 m.
PrP-(106 -126) labeling, confocal analysis revealed a punctuate staining covering the whole surface of 1C11 5-HT and 1C11 NE cells. 90% of this staining co-localized with that of PrP C (Fig. 4A,  yellow). It is noteworthy that in 1C11 5-HT and 1C11 NE cells depleted in PrP C with PrP-siRNA, PrP-(106 -126) immunostaining was cancelled (data not shown). We conclude that, under our experimental conditions, PrP-(106 -126) binds the cell surface of 1C11 5-HT and 1C11 NE cells at the vicinity of PrP C molecules. In a second set of experiments, we used 3F4 antibodies to immunoprecipitate PrP-(106 -126) from lysates of 1C11 5-HT or 1C11 NE cells exposed to the neurotoxic peptide. The presence of PrP C in the immunoprecipitates was assessed by Western blot analysis with SAF32, a monoclonal antibody specific to mouse PrP C . As shown in Fig. 4B, PrP C co-immuno-precipitates with the pathogenic peptide. As expected, when using PrP C -depleted 1C11 5-HT or 1C11 NE cells, the immunoprecipitate was devoid of PrP C (Fig. 4B). It is noteworthy that the absence of PrP-(106 -126)-induced apoptotic response in 1C11 precursor cells cannot be accounted for by a defect in cell surface PrP C expression. As shown in Fig. 4C, 1C11 precursor cells and their neuronal progenies do express similar amounts of PrP C . However, the restriction of PrP-(106 -126) effects to mature neuronal cells may relate to a lack of interaction with PrP C on undifferentiated cells, as depicted in Fig. 4B. Altogether, the above data underscore that the apoptotic effect of the neurotoxic peptide PrP-(106 -126) requires the presence of endogeneous PrP C at the surface of 1C11 5-HT and 1C11 NE neuronal cells. They also support the idea that PrP-(106 -126) has to directly or indirectly interact with PrP C to achieve the transduction of apoptotic signals.
PrP-(106 -126)-mediated Apoptotic Signals in 1C11 5-HT and 1C11 NE Cells Rely on the PrP C -Fyn Coupling-The above observations prompted us to examine whether PrP-(106 -126)-induced apoptotic response in 1C11 5-HT and 1C11 NE cells could relate to an engagement and/or a deviation of PrP C -associated signaling function. Previously, we used PrP-targeted antibodies FIGURE 2. PrP-(106 -126) induces apoptotic signals in 1C11-derived bioaminergic neurons. 1C11 precursor, GT1-7, 1C11 5-HT , and 1C11 NE cells were exposed up to 24 h with 25 M PrP-(106 -126) or PrP-(Scr). Western blots show cytochrome c (Cyt c) release from mitochondria and caspase-3 activation. A, cytochrome c release became detectable at 60 min in 1C11 NE and at 120 min in 1C11 5-HT . B, caspase-3 activation could be monitored as soon as 240 min in 1C11 NE and 360 min in 1C11 5-HT cells. No apoptotic signal was detected in 1C11 or GT1-7 cells treated up to 24 h with PrP-(106 -126). C, dose-response effect of PrP-(106 -126) (0 -50 M) on caspase-3 activation in 1C11 5-HT and 1C11 NE cells. Caspase-3 activation was quantified using the densitometric analysis tool of the QuantityOne software (Bio-Rad). The data shown are representative of a set of three independent experiments.

FIGURE 3. PrP C mediates PrP-(106 -126)-associated neurotoxic effect in 1C11 5-HT and 1C11 NE cells.
Wild type 1C11 5-HT , 1C11 NE , and PrP C -depleted neuronal (PrP-siRNA) cells were exposed up to 24 h with 25 M PrP-(106 -126). A, effect of siRNA duplexes on PrP C expression level in 1C11-derived neuronal cells. Immunoblot of a representative experiment performed with 1C11 5-HT cells in the absence (Ϫ) or in the presence (ϩ) of PrnP gene-specific siRNA or of scrambled siRNA (Ctrl ) duplexes. PrP C was first detected with the SAF32 monoclonal antibody. ERK1/2 was immunoprobed on the same blot with polyclonal antibodies against p44 ERK1 and p42 ERK2 proteins to assess equivalent protein loading and specific siRNA effect toward PrP C expression. B, Western blots showing cytochrome c release and caspase-3 activation in 1C11 NE and 1C11 5-HT cells exposed to the neurotoxic peptide (time points correspond to those indicated in Fig. 2). No apoptotic signal was detected in PrP C -depleted 1C11 5-HT and 1C11 NE cells (PrP-siRNA) treated up to 24 h with PrP-(106 -126).
to mimic an extracellular ligand for PrP C and identified a PrP C -Caveolin-Fyn coupling restricted to fully differentiated 1C11derived neuronal progenies (8). We thus investigated in the present study whether Fyn was involved in the PrP-(106 -126)induced neurotoxicity of 1C11 5-HT and 1C11 NE cells. Importantly, we first observed that antibody-mediated ligation of PrP C itself did not promote cytochrome c release or caspase 3 activation in either 1C11 5-HT or 1C11 NE cells (data not shown). On another hand, PrP-(106 -126)-mediated cytochrome c release and caspase 3 activation in 1C11 bioaminergic neurons were fully cancelled upon inhibition of the Fyn kinase with PP2 (50 pM) within a time scale ranging from 60 to 360 min (Fig. 5). Such results support the view that PrP-(106 -126)-mediated apoptotic cascade in both 1C11 5-HT serotonergic and 1C11 NE noradrenergic cells may depend, in some way, on the mobilization of the PrP C -Fyn coupling.
The fluorogenic probe CM-H 2 DCFDA (Molecular Probes) was used to monitor intracellular ROS production. First, we verified that exposure of 1C11 precursor and GT1-7 hypothalamic cells to PrP-(106 -126) did not elicit any ROS response (Fig. 6A). In 1C11 5-HT and 1C11 NE cells, PrP-(106 -126) promoted ROS accumulation by 10 -30 min (Fig. 6A). ROS level continuously increased within a 30 -300min time scale (Fig. 6A), leading to a ROS index 10-fold above the basal level. Such a continuous accumulation is in contrast with the transient ROS response obtained upon antibody-mediated PrP C ligation, which ends by 30 -45 min (Ref. 9 and Fig.  6A) and whose amplitude does not exceed one third of that reached with the neurotoxic peptide (Fig.  6B). It is noteworthy that upon the addition of 100 M DPI, a selective inhibitor of NADPH oxidase, the PrP-(106 -126)-mediated ROS accumulation was fully cancelled in 1C11 5-HT and 1C11 NE cells (Fig. 6C). Therefore, NADPH oxidase appears to be at the origin of the robust ROS production induced by the pathogenic peptide in neuronal cells.

PrP-(106 -126)-mediated ROS Overproduction Is Accompanied by a Reduction of the Level of Glutathione in Its Reduced
Form-Evidence for an elevated ROS index is not sufficient to fully account for neuronal dysfunction, because cells have evolved several defense and repair mechanisms that would normally counteract the injurious effects of ROS (11). For instance, the thiol reducing agent GSH, the cellular concentration of which is in the millimolar range in brain cells (30), may act as a major antioxidant agent. However, in some neurodegenerative diseases, a reduction in cell antioxidant systems, including GSH, has been shown to contribute to the increased oxidative stress (31)(32)(33). We therefore examined whether the PrP-(106 -126)-induced ROS overproduction in either serotonergic or noradrenergic 1C11 cells could be accompanied with an alteration in the cellular level of GSH. In both cell types treated with 25 M PrP-(106 -126), the concentration of reduced GSH significantly declined, as monitored by recording the fluorescence of Celltracker Green CMFDA (Molecular Probes) (Fig. 7). The decrease was observed by 4 h, with a maximal ϳ40% drop-off by 6 h. In control cells treated with 25 M PrP-(Scr), GSH concentration remained stable over several hours (Fig. 7). Experiments with PrP-siRNA-treated cells showed that the PrP-(106 -126)induced diminution of GSH level again depended on the presence of PrP C , as expected (Fig. 7).
The above data establish that PrP-(106 -126) causes an imbalance in the cellular GSH antioxidant system. Such disequilibrium likely relates to the long lasting production of ROS. Indeed, when a transient and low level NADPH oxidase-dependent ROS production was induced upon PrP C ligation with specific antibodies, the level of reduced GSH did not vary (Fig. 7). From these overall experiments, we conclude that overstimulation of PrP C -coupled signaling cascades by PrP-(106 -126) leads to oxidative stress conditions characterized by an enhanced ROS level and partial depletion of reduced glutathione.
PrP-(106 -126) Induces Unremitting Activation of ERK1/2 MAPKs, p38, and JNK1/2 SAPKs in 1C11 5-HT and 1C11 NE Neuronal Cells-Previous work (9) has shown that, in 1C11 5-HT and 1C11 NE cells, ROS produced upon antibody-mediated PrP C activation act as second message signals in signaling cascades linking PrP C to p44 ERK1 and p42 ERK2 MAPKs. We therefore investigated whether the neurotoxic peptide PrP-(106 -126) could have impact on MAPK activation. Exposure of 1C11 neuronal cells to 25 M PrP-(106 -126) promoted phosphorylation of both ERK1 and ERK2 (Fig. 8A). ERK phosphorylation became detectable as soon as 10 min after treatment and was maximal at  15-30 min. The ERK response persisted within a 30 min to 24 h time range and never returned to basal level (Fig. 8A). Such a time scale contrasted with the transient ERK1/2 activation observed upon PrP C ligation with specific antibodies (Fig. 8A and Ref. 9). As a control, we verified that 25 M PrP-(Scr) did not elicit any ERK1/2 phosphorylation (data not shown).
Because PrP-(106 -126) induces oxidative stress conditions in neuronal cells, we probed a possible activation of stress-sensitive JNK and p38 MAPKs/SAPKs. The phosphorylation level of these kinases was not sensitive to antibody-mediated PrP C ligation in 1C11-derived neuronal cells (data not shown and Ref. 9). By contrast, treatment of 1C11 5-HT and 1C11 NE cells with 25 M PrP-(106 -126) triggered long lasting activation of JNK1/2 and p38 SAPKs (Fig. 8B). Phosphorylation of JNK1/2 and p38 SAPKs became detectable after 30 and 60 min of treatment, respectively (Fig. 8B). JNK1/2 and p38 responses persisted over a long time range. The intensities of phosphorylation were greater in 1C11 NE noradrenergic cells (up to 24 h) than in 1C11 5-HT cells (Fig. 8B).
It is noteworthy that exposure of 1C11 5-HT or 1C11 NE cells to 25 M PrP-(106 -126) in combination with 100 M DPI fully abrogated the ERK1/2, JNK1/2, and p38 MAPK phosphorylations induced by the neurotoxic peptide (Fig. 8C). This observation supports the idea that MAPK/SAPK sustained activation triggered by PrP-(106 -126) relates to excessive ROS synthesis under the control of NADPH oxidase activity. Our overall data underscore that, in cells that have acquired a complete neuronal phenotype, the pathogenic peptide PrP-(106 -126) induces stress conditions leading to apoptotic responses through overstimulation of PrP C -associated signaling pathways.

DISCUSSION
Accumulation of abnormal or misfolded proteins, oxidative stress, and loss of neuronal homeostasis are common hallmarks of neurodegenerative disorders such as Alzheimer, Parkinson, or prion diseases (34 -36). Despite considerable effort to dissect the molecular pathways leading to neurodegeneration in these pathologies, the nature and the temporal order of events leading to neuronal cell demise still are enigmatic.
Here, by using the PrP-(106 -126) neurotoxic peptide as a paradigm for prion infection, we observe for the first time a link between overstimulation of PrP C signaling activity and prionassociated neuronal dysfunction. In 1C11-derived serotonergic and noradrenergic neuronal cells, exposure to PrP-(106 -126) promotes ROS overflow and subsequent sustained MAPK/ SAPK phosphorylation, cytochrome c release, and caspase 3 activation. At the cellular level, PrP-(106 -126) induces neuritic changes and reduces cell viability. A major observation is that 1C11 precursor cells and GT1-7 neuroendocrine cells, which do not display a mature neuronal phenotype, are not sensitive to the neurotoxic peptide.
PrP-(106 -126)-induced Neuronal Damage Is Relayed by PrP C -Saliently, our data underline that PrP C expression at the surface of neuronal cells is mandatory for prion peptide PrP-(106 -126) to elicit intracellular deleterious signals. Indeed, silencing of PrP C expression through transfection of 1C11 5-HT and 1C11 NE cells with specific siRNA abrogates all PrP-(106 -126)-associated effects. Moreover, as monitored by immunofluorescence and immunoprecipitation experiments in differentiated 1C11 5-HT serotonergic and 1C11 NE noradrenergic  cells, PrP-(106 -126) associates with PrP C . Such an interaction is not found in 1C11 precursor cells, which do not respond to PrP-(106 -126) exposure, although these cells display a similar pattern of PrP C expression to 1C11-derived neuronal progenies (Ref. 37 and Fig. 4C). To account for the selective action of PrP-(106 -126) on neuronal cells, we may hypothesize that the peptide displays affinity toward neuron-specific subsets of PrP C isoforms or glycoforms. Interestingly, a particular PrP C glycoform has been identified in hamster retinal and sciatic nerve axons (38). Another explanation could be that the association of PrP-(106 -126) with PrP C requires the participation of other partner(s) specific to mature neuronal cells.
From PrP C Signaling to PrP-(106 -126)-associated Apoptotic Response-Of note, our present findings draw a link between PrP-(106 -126)-mediated intracellular events and PrP C -coupled signaling pathways. In 1C11 precursor cells, their neuronal progenies, and GT1-7 cells, antibody-mediated ligation of PrP C triggers NADPH oxidase-dependent ROS production. ROS act as second messengers and promote a transient ERK1/2 activation (9). Some specificity of PrP C signaling in neuronal cells, however, relies on the implementation of a PrP C -Caveolin-Fyn complex at the neurites of fully differentiated 1C11 5-HT serotonergic and 1C11 NE noradrenergic cells (8,9). Regarding the action of the neurotoxic peptide, the main difference with antibody-mediated PrP C signaling deals with the lack of any intracellular response in 1C11 precursor or GT1-7 neuroendocrine cells exposed to PrP-(106 -126), whereas PrP antibodies do recruit transduction cascades. This result is to be brought together with the observation in 1C11 undifferentiated cells that PrP-(106 -126) does not appear to interact with PrP C . The behavior of the PrP-(106 -126) peptide hence differs from that of PrP C antibodies. It is possible that the antibodies do not finely discriminate between PrP C isoform(s) expressed in cells of neuronal or nonneuronal origin. On another hand, the PrP-(106 -126)-related effects in neuronal cells share similarities with PrP C signaling. Indeed, inhibition of the Fyn kinase activity abrogates 1C11 5-HT and 1C11 NE cell responses to PrP-(106 -126). The PrP C -Caveolin-Fyn complex appears therefore to be the core player mobilized by PrP-(106 -126) to convey deleterious signals.
In 1C11 5-HT and 1C11 NE bioaminergic cells, PrP-(106 -126) triggers ROS overflow (from 15 min up to 300 min), activates MAPK and SAPK (30 -240 min), and recruits apoptotic cascades (120 min to 24 h). These intracellular events are in agreement with observations in TSE-infected brains indicating hallmarks of oxidative stress (36,39) and enhanced levels of MAPK/SAPK (40). Several other in vitro studies carried out using the PrP-(106 -126) neurotoxic peptide also support our observations (25,(41)(42)(43). Here, however, the time schedule of the deleterious events observed in 1C11-derived neuronal cells contrasts with the long delay between PrP-(106 -126) exposure and cell response as recorded with other experimental systems (42)(43)(44). Such a discrepancy may relate to the conditions of preparation of PrP-(106 -126), which influence the aggregation state of the peptide (45,46). Indeed, regarding entire PrP Sc , oligomers containing only 14 -16 PrP Sc monomers were recently shown to behave as the most infectious prion material, as compared with larger aggregates or amyloids (47), consid-ered as inert dumping grounds for misfolded proteins (48 -50). Accordingly, to keep aggregation of PrP-(106 -126) at a minimum, we carried out all our experiments using a rather low concentration of freshly unfrozen PrP-(106 -126).
Corruption of PrP C Signaling Function by PrP-(106 -126) Causes Oxidative Stress-How PrP-(106 -126) causes a long lasting activation of PrP C -coupled intracellular targets is difficult to explain. Antibody-mediated PrP C signaling is likely to involve PrP C clustering (8,(51)(52)(53). Therefore, we may propose that, to trigger mobilization of signaling cascades, small oligomers of PrP-(106 -126) associate with several PrP C molecules. Such a scenario would be in agreement with the aggregation of PrP C obtained upon exposure of neuroblastoma cells to PrP-(106 -126) (54). The ROS response induced by PrP-(106 -126) is relatively slow when compared with that measured upon antibody exposure. It is possible that the recruitment of PrP C by PrP-(106 -126) operates with a kinetics slower than that promoted by PrP antibodies. In addition, PrP-(106 -126)-induced ROS production keeps raising over a 300-min period, leading to a ROS index five to ten times higher than that reached with antibodies. Our findings support the view that PrP-(106 -126) creates an imbalance in the dynamics of PrP C signaling as monitored at the proximate level of NADPH oxidase activation. We may speculate that an overflow of PrP C -coupled signals originates from a dysregulation of the PrP C -Caveolin-Fyn complex, with, as a result, overactivation of the Fyn kinase, in line with in vivo observations (55). Such alteration may reflect a capacity of PrP-(106 -126) to modify membrane fluidity and viscosity (56 -58). A loss of control of Fyn kinase activity would further result in an overproduction of ROS, which would loose their function of second messenger. Instead, high indices of ROS, combined with a drastic reduction of the antioxidant GSH level, would lead to oxidative stress conditions and switch on stress-oriented cascades. These include long lasting activation of ERK1/2 and mobilization of p38 and JNK1/2 SAPKs, which are known to promote apoptosis in neuronal cells (25,59,60).
Here, by taking advantage of the differentiation potential of the 1C11 cell line, we shed some light on the detrimental action of PrP-(106 -126) against mature neuronal cells. In fact, our overall data somewhat recapitulate in vivo findings, which have underscored the importance of PrP C expression at the surface of neuronal cells for prions to exert their pathogenic action (4,5). In addition, they give substantial credit to the hypothesis according to which prions may damage the brain by distorting signaling events normally controlled by PrP C and provide some clues as to the molecular pathways likely involved in TSE-associated neurodegeneration. By underlying the primary role of ROS overproduction in the deleterious cascade triggered by PrP-(106 -126), our study defines oxidative stress as a major cause of neuronal cell injury. Eventually, the demonstration that PrP-(106 -126) exacerbates PrP C -associated signals up to neurotoxic levels constitutes a direct piece of evidence arguing for the "gain of function" hypothesis in the TSE field. It may help to develop therapies against prion-induced neuronal injury by targetting intracellular mediators of PrP C signaling.