The Tyrosine Kinase Inhibitor STI571 Induces Cellular Clearance of PrP Sc in Prion-infected Cells*

The conversion of the cellular prion protein (PrP c ) into pathologic PrP Sc and the accumulation of aggregated PrP Sc are hallmarks of prion diseases. A variety of experimental approaches to interfere with prion conversion have been reported. Our interest was whether interference with intracellular signaling events has an impact on this conversion process. We screened (cid:1) 50 prototype inhibitors of specific signaling pathways in prion-infected cells for their capacity to affect prion conversion. The tyrosine kinase inhibitor STI571 was highly effective against PrP Sc propagation, with an IC 50 of < 1 (cid:1) M . STI571 cleared prion-infected cells in a time-and dose-dependent manner from PrP Sc without influencing biogenesis, localization, or biochemical features of PrP c . Interestingly, this compound did not interfere with the de novo formation of PrP Sc but activated the lysosomal degradation of pre-existing PrP Sc , lowering the half-life of PrP Sc from > 24 h to < 9 h. Our data indi cate that among the kinases known to be inhibited by STI571,

The conversion is thought to take place either directly at the plasma membrane or in the early compartments of the endocytic pathway, e.g. caveolae or rafts, specialized regions enriched in sphingolipids, cholesterol, and glycosyl phosphatidylinositol-anchored proteins (7,8). The conversion process comprises profound changes in the structure and the biochemical properties of PrP. The ␣-helical conformation of PrP c is converted into the mainly ␤-sheeted structure of PrP Sc . In contrast to PrP c , the pathogenic PrP Sc is highly insoluble and partially resistant to proteolytic digestion (1-3, 9, 10). The exact mechanism of the conversion is still enigmatic, although two models are currently under consideration. The first model favors a crystallization reaction, where PrP Sc acts as the crystal seed. Newly converted PrP Sc molecules are added to that seed, forming PrP Sc aggregates (11). The second model postulates a template-assisted conversion with intermediates, possibly PrP c ⅐PrP Sc heterodimer complexes (1,9). A variety of experimental approaches for interfering with prion conversion have been reported. Some of them target PrP c , as the conversion can be blocked by removing the substrate of the process. This can be achieved by preventing the expression of PrP c (1,3,12) or by inhibiting its transport to the plasma membrane (13). A different mechanism has been suggested for the archetypical anti-prion compound Congo Red. Like chemical chaperones (14), Congo Red is thought to overstabilize PrP c in its equilibrium with suggested folding intermediates (1,9,15) in the prion conversion process (16). Other substances are thought to interfere with the interaction of PrP Sc with complexes consisting of PrP c and folding intermediates. Such substances include ␤-sheet breakers, anti-PrP aptamers, and anti-PrP antibodies (16 -24). There is evidence that additional cellular components are involved in prion conversion (factor X) (1,16). A soluble PrP dimer molecule has been postulated to prevent the binding of such components to folding intermediates, thereby inhibiting PrP Sc propagation (25). Only very few compounds like branched polyamines directly target PrP Sc , increasing its intracellular clearance (26 -28).
The physiological function of PrP c is not known. PrP-null mice develop normally and show no gross behavioral abnormalities (29), albeit alterations in circadian activity and sleep were reported (30). Furthermore, it has been shown that copper binds to the octarepeat-containing N-terminal region of PrP c , resulting in a suggested antioxidant activity of the protein (31). Therefore, a functional role for PrP c in copper metabolism and cellular protection against oxidative stress has been proposed (32). Recent work points to a role for PrP c in signal transduction. Antibody-mediated cross-linking of PrP c on the surface of the murine neuronal 1C11-differentiated cell line promotes the dephosphorylation and activation of Fyn kinase (33). A further implication of PrP c in neuronal survival and differentiation is supported by the finding that different signal transduction pathways involved in neurite outgrowth and neuronal survival are elicited by PrP c (34), and PrP Ϫ/Ϫ mice expressing N-terminally truncated PrP show severe neurodegeneration soon after birth (35). We have shown previously that PrP c interacts in vitro and in cultured cells with pint, synapsin 1, and grb2, the latter two representing proteins involved in cellular signaling (36). In the present study we investigate the potential role of cell signaling interference in PrP Sc propagation. Multiple substances were screened that interfere with specific signaling pathways in prion-infected cells for their effect on PrP Sc . STI571 (Gleevec®, imatinib mesylate), an inhibitor of the tyrosine kinase c-Abl, was found to be highly effective in inducing the clearance of PrP Sc in prion-infected cells without interfering with biogenesis, localization, and biochemical features of PrP c . Our results demonstrate that, although STI571 has no major effect on the de novo formation of PrP Sc , it activates the lysosomal degradation of pre-existing PrP Sc .

EXPERIMENTAL PROCEDURES
Reagents-STI571 was purchased from Novartis Pharmaceuticals Corp. (Basel, Switzerland). Proteinase K, Pefabloc proteinase inhibitor, and pepstatin A were obtained from Roche Applied Science. The inhibitor CT52923 was kindly provided by Millenium Pharmaceuticals (Cambridge, MA). Suramin was purchased from Bayer (Leverkusen, Germany). Immunoblotting was done using the enhanced chemiluminescence blotting technique (ECL plus) from Amersham Biosciences. Monoclonal anti-PrP antibody 3F4 (Signet Pathology, Dedham, MA) recognizes amino acids 109 -112 (numbering according to 37) of human and hamster PrP. Monoclonal mouse anti-PrP antibody 4H11 was prepared using a dimeric murine PrP as immunogen. 1 The polyclonal anti-PrP antibody A7 has previously been described (24). Monoclonal mouse anti-Abl antibody was obtained from Pharmingen. Secondary antibodies for immunofluorescence or flow cytometry (FACS) 2 were fluorescein isothiocyanate-conjugated immunoglobulins obtained from Dako or Dianova (Hamburg, Germany). [ 35 S]Met/Cys (Promix; 1000 Ci/mmol) and [␥-32 P]ATP (3000 Ci/mmol) were obtained from Amersham Biosciences. Protein A-Sepharose and Protein G-Sepharose were from Amersham Biosciences. c-Abl substrate was from New England BioLabs (Beverly, MA). Cell culture media and solutions were obtained from Invitrogen (Karlsruhe, Germany). All other chemicals were from Roche Applied Science.
Cell Culture and Mode of Drug Application-The mouse neuroblastoma cell lines N2a (ATCC CCL 131), 3F4-PrP overexpressing N2a and ScN2a, the mouse hypothalamic cell line ScGT1, and the non-neuronal mouse cell line SMBs. 15 have been described (13,38,39). Cells were maintained in Opti-MEM medium containing 10% fetal calf serum, antibiotics, and glutamine, with exception of SMBs.15 cells, which were maintained in Medium 199 with Earle's salts containing 10% newborn calf serum and 5% fetal calf serum, antibiotics, and glutamine. STI571 was dissolved in Me 2 SO at a stock solution of 10 mM (storage at Ϫ20°C) and was added to the medium to a final concentration of 10 M (if not otherwise stated). Suramin was dissolved in NaCl (0.9%) at a stock solution of 200 mg/ml (light protected at 4°C) and was added to the medium to a final concentration of 200 g/ml. As published recently (40), CT52923 was dissolved in Me 2 SO at a stock solution of 3 mM (storage at Ϫ20°C) and used as indicated. In short term experiments (3 days) there was no medium change in between, and in long time studies the medium including the drug was changed on a 3-day basis (if not otherwise stated). In all experiments mock-control cells were used that were treated with solvent in an identical fashion to drug-treated cells.
Immunoblot Analysis and Detergent Solubility Assay-Immunoblot analyses were performed as previously described (13). Confluent cell cultures were lysed in cold lysis buffer (10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10 mM EDTA, 0.5% Triton X-100, 0.5% deoxycholate) for 10 min. For proteinase K (PK) treatment post-nuclear lysates were divided into two halves. One-half was incubated with PK (20 g/ml) for 30 min at 37°C, and digestion was stopped by the addition of proteinase inhibi-tors (5 mM phenylmethylsulfonyl fluoride, 0.5 mM Pefabloc, and aprotinin) and directly precipitated with ethanol. The sample without PK treatment was directly supplemented with proteinase inhibitors and precipitated with ethanol. After centrifugation for 30 min at 3500 rpm (4°C), the pellets were re-dissolved in TNE buffer, and gel-loading buffer was added. After boiling for 5 min, an aliquot was analyzed on 12.5% SDS-polyacrylamide electrophoresis gels. For solubility assay, post-nuclear cell lysates were supplemented with proteinase inhibitors and N-lauryl sarcosine to 1% and ultracentrifuged in a Beckman TL-100 table ultracentrifuge for 1 h at 40,000 rpm (100,000 ϫ g; TLA-45 rotor; 4°C). Supernatant fractions were precipitated with ethanol, and pellet fractions were resuspended in 50 l of TNE (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA) and analyzed in immunoblot or radioimmune precipitation assays.
PIPLC Treatment-For PIPLC treatment, ϳ80% confluent cells were washed twice with PBS. 200 milliunits of PIPLC (Sigma) and STI571 were added in 4 ml of serum-free Opti-MEM medium and incubated for 4 h at 37°C. The medium was collected, and the cells were washed extensively with PBS and lysed as described for immunoblotting. After a methanol precipitation of the medium and the cell lysate, both fractions were analyzed in immunoblot.
Metabolic Radiolabeling and Immunoprecipitation Assay-Metabolic radiolabeling and immunoprecipitation assays were done as described previously (13). In brief, confluent cells were washed twice with PBS and incubated for 1 h in RPMI without methionine/cysteine containing 1% fetal calf serum. Due to the long labeling period, the addition of 1% fetal calf serum is more convenient for the cells and had no negative influence on the labeling efficiency in our hands. The medium was supplemented with 800 Ci/ml L-[ 35 S]Met/Cys (Amersham Biosciences) for 16 h. For analysis of the de novo generation of PrP Sc , ScN2a cells were simultaneously treated with STI571 (10 M) or with suramin (200 g/ml) (no chase). For determination of PrP Sc half-life, STI571 (20 M) was added only during the chase period. For both experiments, cells were washed twice in ice-cold PBS and lysed in cold lysis buffer (100 mM NaCl, 10 mM Tris-HCl, pH 7.8, 10 mM EDTA, 0.5% Triton X-100, 0.5% deoxycholate). Insoluble material was removed by centrifugation at 14,000 rpm for 1 min. Then lysates with and without PK treatment were subjected to a solubility assay as described above. Pellets were re-suspended in 100 l of radioimmune precipitation assay buffer (0.5% Triton X-100, 0.5% deoxycholate in PBS) with 1% SDS and boiled for 10 min. The pellet fraction was diluted with 900 l of radioimmune precipitation assay buffer (supplemented with 1% sarcosyl), and the primary antibody (1:300) was incubated overnight at 4°C. Protein A-Sepharose beads were then added for 60 min at 4°C. The immunoadsorbed proteins were washed in cold radioimmune precipitation assay buffer supplemented with 1% SDS, subjected to a deglycosylation step with peptide N-glycosidase F at 37°C, and analyzed on 12.5% SDS-PAGE followed by autoradiography.
FACS Analysis-For surface protein analysis, STI571-treated (10 M, 3 days) or mock-treated cells were suspended with PBS containing 1 mM EDTA, centrifuged, resuspended in FACS buffer (PBS with 2.5% fetal calf serum and 0.05% sodium azide), and incubated for 5 min on ice. Primary antibody (A7) was incubated in a 1:100 dilution in FACS buffer for 45 min on ice and washed 3 times in FACS buffer, and the secondary antibody (fluorescein isothiocyanate-labeled, 1:100) was incubated for another 45 min. After the last wash cells were resuspended in FACS buffer with propidium iodide (2 g/ml). Cells incubated only with the secondary antibody served as a control. The FACS analysis was performed in a BD Biosciences FACSCalibur apparatus. In total, from each sample 10,000 living cells, as determined by the propidium iodide staining, were analyzed.
Transient Expression of a Trans-dominant Negative c-Abl Mutant-For generation of a trans-dominant negative mutant of c-Abl tyrosine kinase, the lysine residue at position 271 was mutated to a methionine. This mutation is located in the ATP binding domain of the protein, leading to a complete inactivation of the kinase. As vector we used pcDNA3.1/Zeo (Invitrogen). The resulting plasmid pcDNAVSVc-abl-K271M encodes a VSV-tagged c-Abl mutant, which when overexpressed in cells inhibits the kinase activity of endogenous c-Abl. Lipofection of ScN2a cells with this construct was done using FuGENE transfection reagent (Roche Applied Science) according to standard protocols. Four days after transfection the cells were lysed and analyzed as described for immunoblotting.
Immunoprecipitation of c-Abl from N2a and ScN2a Cells and in Vitro Kinase Assay-N2a and ScN2a cells were grown on 15-cm dishes to 90% confluency, washed twice in ice-cold PBS, and lysed in cold lysis buffer (150 mM NaCl, 10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1% Triton X-100, 1% deoxycholate, 10 mM NaF, 2 mM sodium orthovanadate, 0.1 mM aprotinin, 1 mM phenylmethylsulfonyl fluoride). Insoluble material was removed by centrifugation for 10 min at 14,000 rpm (4°C). Equal amounts of total protein of both lysates were adjusted to 500 l with lysis buffer. Protein G-Sepharose beads and 2.5 g of anti-Abl antibody (Pharmingen) were added, and the reactions were adjusted to 1 ml with cold HNTG buffer (150 mM NaCl, 50 mM Hepes, pH 7.5, 1 mM EDTA, 10% glycerin, 0.1% Triton X-100, 10 mM NaF, 2 mM sodium orthovanadate, 0.1 mM aprotinin, 1 mM phenylmethylsulfonyl fluoride) and incubated for 3 h at 4°C. The immuno-adsorbed kinase was washed twice in cold HNTG buffer and twice in basis buffer (30 mM Tris-HCl, pH 7.5, 15 mM MgCl 2 , 1 mM MnCl 2 , 15 M Na 3 PO 4 , 1.5 mM dithiothreitol) and then subjected to the in vitro kinase assay. The reaction was carried out in a total volume of 40 l of basis buffer with 2 M ATP (Sigma), 1.5 Ci [␥-32 P]ATP (Amersham Biosciences), and 100 M c-Abl substrate (New England BioLabs) at 30°C. After 20 min the reaction was stopped with 10 l of EDTA (100 mM). The samples were briefly centrifuged and spotted on a P81 filter (Upstate Biotechnology, Inc., Waltham, MA). After extensive washing with 0.15% phosphoric acid the filters were rinsed in acetone, embedded in scintillation solution, and measured in a scintillation counter.

STI571 Reduces PrP Sc in a Time-and Dose-dependent Manner in Prion-infected
Cells-Because various studies suggest a possible role of PrP c in cellular signaling, we screened about 50 prototype inhibitors known to interfere with specific intracellular signal transduction pathways in prion-infected neuronal cells for their effect on PrP Sc propagation (by proteinase K digestion and immunoblotting). The initial screening was done by adding the compounds for 3 days at a concentration of 10 M to the culture medium ( Fig. 1). Compounds that had no detectable effect on PrP Sc were excluded from further analysis. Compounds that were obviously toxic for the cells and those that had an effect on both PrP Sc and PrP c were tested again at a concentration of 2 M for 3 days. Most remaining compounds either had no effect on PrP Sc or also profoundly affected PrP c . Several substances were toxic to the cells even at low concentrations. Among the compounds tested, only one substance (STI571, signal transduction inhibitor 571, also known as imatinib mesylate, Gleevec®) that had a highly significant effect on PrP Sc propagation without cytotoxicity (as revealed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay and trypan blue staining, data not shown) was identified. Treatment of prion-infected ScN2a, ScGT1, and SMBs.15 cells with STI571 for 3 days (ScN2a and ScGT1) or 10 days (SMBs.15), respectively, reduced PrP Sc to undetectable levels ( Fig. 2A, lanes 4, 8, and 12, respectively). In contrast, mocktreated cells showed significant amounts of PrP Sc ( Fig. 2A, lanes 2, 6, and 10, respectively). The reduction of PrP Sc was found to be time-and dose-dependent (Fig. 2, B and C). ScN2a cells treated for 1 day with 10 M STI571 harbored the same amount of PrP Sc as mock-treated control cells. After the second day, the signal for PrP Sc was significantly reduced and was undetectable after 3 days of treatment (Fig. 2B, lanes 1-8). ScN2a cells treated for 6 days with different concentrations of STI571, varying from 0.25 to 10 M (procedure according to Korth et al. 41), demonstrated a slight reduction of PrP Sc with 0.25 M STI571. Treatment with 1 M STI571 led to a significant reduction of PrP Sc to less than 10% of the amount present in the mock-treated control (Fig. 2C, lanes 1-7). Under these experimental conditions the IC 50 , which is defined here as the concentration with a 50% inhibitory effect on PrP Sc propagation, was measured to be between 0.75 and 1 M. Because the optimal results for the clearance of PrP Sc in ScN2a cells were obtained with treatment for 3 days with 10 M STI571, we decided to use a concentration of 10 M for treatments longer than 1 day and 20 M for 1-day treatments. In a longer term experiment, the efficiency of PrP Sc clearance by STI571 was investigated. To test whether the observed effect on PrP Sc is reversible or whether the cells remain free of PrP Sc even after the removal of the drug (Fig. 2D), ScN2a cells were treated for 10 days with 10 M STI571. In immunoblotting assays, the signal for PrP Sc entirely disappeared in treated cells ( lanes  1-4). The cells were then cultured for further 30 days without the drug. Every 10 days, an aliquot of cells was tested for the presence of PrP Sc by immunoblotting. Even after 30 days without STI571 treatment, PrP Sc could not be detected (lanes [5][6][7][8][9][10][11][12][13][14][15][16]. Taken together, our studies show that STI571 is able to reduce PrP Sc in a time-and dose-dependent manner in various prioninfected cell lines and that this clearance is irreversible. The Biogenesis and Localization of PrP c Are Not Affected by STI571-Because PrP Sc propagation is dependent on a preceding PrP c localization at the plasma membrane, we tested whether the clearance of PrP Sc by STI571 is related to changes in the expression, biochemical features, or localization of PrP c . Previous experiments demonstrated that there is no obvious difference in the expression level of PrP c between treated and mock-treated cells (Fig. 2, A-D). To address the subcellular localization of PrP c in more detail, we first performed a diges-  5-12). Bars on the right indicate the three bands specific for non-, mono-, and diglycosylated PrP. B, the effect of STI571 on PrP Sc propagation is time-dependent. ScN2a cells were mock-treated or treated for 1, 2, or 3 days with 10 M STI571. Cell lysates were either subjected to PK digestion (ϩPK) or left untreated (ϪPK) and analyzed in immunoblotting using mAb 3F4. Bars on the right indicate PrP-specific bands. C, the IC 50 of STI571 is Յ1 M. ScN2a cells were mock-treated (lane 1) or treated for 6 days with rising concentrations of STI571 (lanes 2-7). The medium was changed every day. Cell lysates were digested with PK and analyzed in immunoblotting using mAb 4H11. D, STI571 irreversibly removes PrP Sc from prion-infected cells. ScN2a cells were mock-treated (0/0) or treated for 10 days with 10 M STI571 (10/0). After removal of STI571, the untreated and treated cells were cultured for further 10  tion with PIPLC that cleaves glycosyl phosphatidylinositol (ϩ)anchored proteins from the outer leaflet of the plasma membrane, shedding them into the medium. Upon incubation of ScN2a and N2a cells with PIPLC, PrP c was digested from the cell surface and could be detected in the medium fraction of both control cells and cells treated with STI571 (Fig. 3A, lanes  4 and 8 or 14 and 16, respectively). Accordingly, the signal for PrP c in the lysate fraction of PIPLC-digested cells was decreased, indicating that STI571 does not influence the expression of PrP c on the plasma membrane (Fig. 3A, lanes 3 and 7 or  10 and 12). As expected, the signal for PrP Sc in ScN2a cells decreased upon treatment with STI571. Confocal microscopy and surface FACS analysis of N2a cells treated with STI571 also revealed a surface expression of PrP c comparable with that of control cells, confirming the observations from the PIPLC digestion experiments (Fig. 3, B and C).
A biochemical hallmark of PrP c is its solubility in non-ionic detergents, whereas PrP Sc forms insoluble aggregates under such conditions. Alterations of the folding and aggregation behavior of PrP c is often followed by changes in its solubility. Therefore, we examined the solubility of PrP c after 3 days of treatment with STI571 in a standard solubility assay (Fig. 3D). In mock-treated ScN2a cells, soluble PrP c could be detected in the supernatant, and insoluble PrP Sc could be detected in the pellet fraction. The same behavior was observed in cells treated with STI571. PrP c remained soluble and was present in the supernatant, and insoluble PrP was not detectable in the pellet fraction. In summary, STI571 clears prion-infected cells of PrP Sc without interfering with the expression level, the localization, or biochemical properties of PrP c .

STI517 Decreases the Half-life of PrP Sc by Inducing Its
Lysosomal Clearance-Our previous studies showed that STI571 clears prion-infected cultured cells from PrP Sc . To characterize the effect of STI571 on PrP Sc propagation in more detail, we analyzed the de novo formation of PrP Sc in cells treated with STI571. Immunoprecipitation assays of radioactively labeled PrP Sc were performed. ScN2a cells were metabolically labeled overnight in the presence of STI571. As a control, ScN2a cells were treated with the compound suramin, a drug that inhibits the de novo formation of PrP Sc (13). Cells were lysed, half of each lysate was subjected to digestion with PK, and a solubility assay was performed. After immunoprecipitation of the insoluble pellet fraction with an anti-PrP antibody, the samples were deglycosylated and analyzed by autoradiography (Fig.  4A). Whereas in cells treated with suramin PrP Sc was not detectable (lane 5), there was a signal for newly synthesized PrP Sc (19 kDa) in the control cells and in cells treated with STI571 (lanes 4 and 6). In suramin-treated cells aggregated and, therefore, insoluble PrP c migrated at a higher molecular weight (lane 2). Under STI571 treatment, full-length, deglycosylated PrP c remained soluble and could not be detected in the pellet fraction (lane 3). These data show that STI571 does not significantly affect the de novo formation of PrP Sc but induces the degradation of pre-existing PrP Sc . To confirm this observation and to analyze the half-life of PrP Sc under STI571 treatment, ScN2a cells were metabolically labeled overnight. Cells were either harvested directly or cultured further in the presence or absence of 20 M STI571. At different time points, the cells were lysed, digested with proteinase K, and subjected to a solubility assay. Subsequently, the insoluble PrP was immuno-  1-4, and 9, 10, 13, and 14, respectively). Subsequently, a PIPLC digestion of all cells was performed. Cell lysates (L) and media fractions (M) without (lanes 1, 2, 5, 6, and  9, 11, 13, 15) and with (lanes 3, 4, 7, and 8  and 10, 12, 14, and 16) PIPLC treatment were subjected to an immunoblot analysis using the mAb 3F4 (ScN2a) and 4H11 (N2a). Bars on the left indicate PrP-specific bands. B, STI571 treatment does not affect surface localization of PrP. Indirect immunofluorescence and confocal microscopy were performed with 3F4-N2a cells cultivated for 3 days in the absence (left panel) or presence (right panel) of STI571 (10 M). C, amount of surface PrP c is not reduced by STI571. 3F4-N2a cells were treated for 3 days with 10 M STI571 or mock-treated. Levels of surface PrP expression were analyzed by surface FACS analysis. As a control, the primary antibody was omitted (left and middle panel). In the right panel an overlay of the signals of mock-treated and STI571-treated cells (10 M for 3 days) is shown. From each sample, 10,000 living cells were analyzed. D, no induction of insoluble PrP. Lysates of treated (10 M for 3 days) (lanes 3 and 4) or mock-treated (lanes 1 and 2) ScN2a cells were subjected to a solubility assay. Supernatant (S) and pellet (P) fractions were analyzed in immunoblot. For detection of PrP-specific signals, mAb 3F4 was used. The molecular weight marker is shown on the left. Bars on the right indicate PrP-specific bands. precipitated and deglycosylated (Fig. 4B). In mock-treated control cells, the signal for PrP Sc did not change during the chase points, corresponding to the predicted 24-h half-life of PrP Sc in these cells (lanes 1-6) (42). However, there was a rapid decrease in the signal from cells treated with STI571, indicating that PrP Sc was degraded more rapidly (lanes 7-12). Fig. 4C shows the densitometric evaluation of three independent experiments. Taken together, the treatment with STI571 induced the degradation of pre-existing PrP Sc and lowered the half-life of PrP Sc from more than 24 h to less than 9 h.
Because STI571 obviously activates the intracellular degradation of PrP Sc , we asked whether the simultaneous treatment with an inhibitor of proteolytic degradation could counteract this effect. Therefore, we analyzed the effect of STI571 on ScN2a cells in the presence of ammonium chloride (NH 4 Cl) by immunoblot analysis. NH 4 Cl is known to inhibit lysosomal degradation by raising the lysosomal pH. The cells were mocktreated or treated for 3 days with 10 M STI571. The medium was changed, and the cells were again mock-treated, treated with 10 M STI571, 10 mM NH 4 Cl, or with both substances simultaneously. As shown in Fig. 4D, NH 4 Cl was able to inhibit the degradation of PrP Sc induced by STI571 (Fig. 4D, lane 4  and lane 6, respectively). Interestingly, treatment of the cells with NH 4 Cl alone also led to a slight increase in the signal for PrP Sc , indicating that NH 4 Cl inhibits the physiologic degradation of PrP Sc .
Thus, treatment with STI571 decreases the half-life of PrP Sc in prion-infected cells by inducing its cellular clearance. The finding that this degradation can be counteracted by the lysosomal inhibitor NH 4 Cl suggests that STI571 activates the lysosomal degradation of PrP Sc .
The Effect of STI571 on PrP Sc Is Presumably Caused by Inhibition of the Tyrosine Kinase c-Abl-The studies above have shown that STI571 is able to reduce PrP Sc in prion-infected cells by inducing its cellular degradation. We were then interested in further characterizing the cellular signaling pathways that are affected by STI571, thus leading to the degradation of PrP Sc . The known targets of STI571 are the tyrosine kinases c-Abl, the  1-3; ϪPK) or with (lanes 4 -6; ϩPK) PK digestion was immunoprecipitated using polyclonal antibody A7. Samples were deglycosylated and analyzed on SDS-PAGE followed by autoradiography. Bars on the right indicate deglycosylated PrP c (upper bar) and PrP Sc (lower bar). B, half-life of PrP Sc is significantly reduced during STI571 treatment. ScN2a cells were metabolically labeled overnight and then chased in medium without STI571 (lanes 1-6) or medium containing STI571 (lanes 7-12) at a concentration of 20 M. After various time points as indicated (0, 3, 6, 9, 12, and 24 h, respectively) cells were lysed, and immunoprecipitation was performed as described in A. Only samples after PK digestion are shown. A molecular weight marker is given on the left. The arrow on the right indicates the molecular size specific for deglycosylated PrP Sc . C, schematic representation of PrP Sc half-life. The percentage ratios (y axis) of radioactively labeled PrP Sc signal after various chase periods (x axis) with and without STI571 treatment are compared. The average of three individual experiments is shown, and the S.D. is indicated. D, inhibition of lysosomal activity counteracts the effect of STI571 on PrP Sc . ScN2a cells were either mock-treated or treated for 2 days with 10 M STI571. Then the medium was changed, and the cells were kept for one further day in medium containing either 10 M STI571 or 10 mM NH 4 Cl or both substances. Mock-control cells were treated with Me 2 SO. Then the cells were harvested, and half of each lysate was PK-digested (ϩPK) or left untreated (ϪPK) and, after methanol precipitation, analyzed in immunoblot using the mAb 4H11.
PDGF receptor, and c-Kit (43,44). To specify the signal transduction pathways that are involved in the induction of the degradation of PrP Sc , the specific tyrosine kinase responsible for the observed anti-prion effect needed to be determined. Therefore, ScN2a cells were treated for 3 days with CT52923, a substance that specifically inhibits PDGF receptor and c-Kit kinase without having an effect on the activity of c-Abl (40). In a cell-based assay the autophosphorylation of PDGF receptor and c-Kit is inhibited by CT52923, with an IC 50 of about 100 and 200 nM, respectively (40). The IC 50 for CT52923 in inhibition of cellular responses to PDGF-like proliferation and migration has been reported to be 280 and 64 nM, respectively (40). Therefore, cells were treated accordingly with concentrations up to 30 M, the highest concentration that was non-toxic. However, even at this concentration of CT52923 no effect on PrP Sc could be detected, in contrast to STI571-treated cells (Fig. 5A, lane 6 and 8). A further experiment to confirm the specificity of c-Abl was to mimic the effect of STI571 in the cells by overexpressing a trans-dominant negative mutant of the kinase that causes an abrogation of the activity of endogenous c-Abl. ScN2a cells were transiently transfected with a trans-dominant negative mutant of c-Abl and after 4 days were assayed for PrP Sc by PK digestion and immunoblotting (Fig. 5B). In cells transiently expressing the c-Abl mutant there was a significant reduction of PrP Sc (lane 6) as compared with untransfected (lane 2) and mock-transfected cells (lane 4). To investigate the effects of wild type c-Abl overexpression on PrP Sc , ScN2a cells were transfected with a wild type c-Abl construct. Unfortunately, 2 days after transfection the cells stopped growing and began to die (data not shown). The observed effects are consistent with reports of cytostatic and cytotoxic effects of overexpressed c-Abl (45)(46)(47). To analyze the inhibitory effect of STI571 on endogenous c-Abl in N2a and prion-infected N2a cells, we performed in vitro kinase assays. C-Abl was immunoprecipitated from equal amounts N2a and ScN2a cell lysates and subjected to an in vitro kinase assay. Different concentrations of STI571were added to the reactions (Fig. 5C). STI571 specifically inhibited the activity of c-Abl isolated from the N2a and ScN2a cells, with no detectable differences between uninfected and prion-infected cells. The IC 50 of STI571 was 0.1 M as determined by the in vitro assay. Thus, c-Abl is the most likely tyrosine kinase responsible for the antiprion effect of STI571 in prion-infected cells. DISCUSSION This study aimed to investigate whether interfering with cellular signal transduction can interfere with PrP Sc accumulation. We performed a screen of well characterized compounds known to interfere with specific signaling pathways to examine their effect on PrP Sc propagation in prion-infected cells. The tyrosine kinase inhibitor STI571 was, thus, identified as the compound with the most pronounced effect on PrP Sc .

STI571, a Compound with Applications in Modern Cancer
Therapy-With STI571 we describe a new class of compounds with a significant effect on PrP Sc in prion-infected cultured cells. STI571 (signal transduction inhibitor 571), also known as Gleevec® or imatinib mesylate, is a derivative of 2-phenylaminopyrimidine and inhibits the tyrosine kinase c-Abl by blocking its ATP binding site (43). It has been developed against chronic myeloid leukemia. Chronic myeloid leukemia is  8). Cell lysates were left untreated (ϪPK) or digested with PK (ϩPK) and analyzed in immunoblotting using mAb 3F4. B, a trans-dominant negative c-Abl mutant can inhibit PrP Sc -propagation. ScN2a cells were either untransfected, mock-transfected, or transfected (TF) with the construct encoding the transdominant negative mutant of c-Abl or treated with 10 M STI571 for 3 days. Cell lysates were either digested with PK or left untreated (ϩ/Ϫ) and analyzed in immunoblotting using mAb 3F4. C, STI571 specifically inhibits c-Abl isolated from N2a and ScN2a cells. c-Abl tyrosine kinase was isolated from cell lysates of N2a or ScN2a cells by immunoprecipitation and subjected to an in vitro kinase assay using a specific c-Abl substrate peptide and [␥-32 P]ATP. The kinase activity of the reaction with three different concentrations of STI571 was compared with the activity in the reaction without STI571 (/). Negative controls were a reaction in which the immunoprecipitation was performed without anti-c-Abl-antibody (ϪAb) and a reaction without c-Abl substrate (Ϫsubstrate). Samples were spotted on a P81 filter, and the kinase activity was measured by scintillation counting. caused by Bcr-Abl, the product of the t(9;22) Philadelphia chromosome translocation. In this fusion protein, the kinase activity of c-Abl is constitutively active, causing an unregulated proliferation of blood cells. STI571 also inhibits the PDGF receptor and the c-Kit kinase (43,44). Deregulation of c-Kit activity in gastrointestinal stromal tumors was shown to be efficiently inhibited by the compound (48). Although STI571 inhibits the wild type kinases in healthy cells, there are only minor side effects, probably due to redundancy in signal transduction pathways. Since May 2001 STI571 has been available to patients as an oral treatment against chronic myeloid leukemia, and since February 2002, it has also been prescribed against gastrointestinal stromal tumors. In addition, STI571 may be of potential therapeutic use against the neurodegenerative disorder Alzheimer's disease. In a recent report it has been shown that STI571 inhibits the ␤-amyloid production without interfering with the cleavage of Notch-1 by the ␥-secretase complex. The authors suggested STI571 to be a useful tool for the development of novel therapies for Alzheimer's disease (49).
Compounds exhibiting anti-prion activity in prion-infected cells include Congo Red (50), dextran sulfate, pentosan polysulfate (51), suramin (13), branched polyamines like poly(propyleneimine) (PPI) dendrimers and poly(amidoamine) (PAMAM) dendrimers (26 -28), acridine derivatives like quinacrine and chlorpromazine (41), and lovastatin (52). We have demonstrated that STI571 significantly reduces PrP Sc without influencing PrP c . This makes the drug a potential candidate for prophylaxis and therapy, if some important requirements are met. (i) Results from cell culture studies usually cannot be transferred directly to the in vivo situation. For example, quinacrine has been shown to interfere with the de novo formation of PrP Sc in vitro (41). Unfortunately, when tested therapeutically it did not prolong the survival of prion-infected mice (53). Nevertheless, quinacrine might be a candidate for a post-exposure prophylaxis, as is the case for suramin (13). (ii) Similarly, the human situation may not be reflected by a murine in vivo system. The metabolism of STI571, as observed in mice, is completely different from that in humans. The biological halflife of the drug upon oral application is about 12-16 h in humans, whereas in mice it is less than 4 h (54). Therefore, it is difficult to achieve and maintain the necessary concentration of STI571 in the blood of mice. (iii) A therapeutic anti-prion drug has to reach the CNS and, therefore, must cross the blood-brain barrier (55). It has been reported that only about 2.8% of orally administered STI571 is found in the cerebrospinal fluid (56). This amount might not be sufficient to elicit a therapeutic anti-prion effect. If STI571 or derivatives could overcome the discussed problems, they could be considered as candidates for in vivo studies in animals or eventually in humans. The results with STI571 in vitro demonstrate a proof of principle that this signal transduction inhibitor is able to interfere with PrP Sc propagation. Therefore, signal transduction pathways should be considered as potential targets for prophylaxis and therapy of prion diseases.
Anti-prion Effect by Inhibition of c-Abl?-STI571 specifically inhibits the tyrosine kinases c-Abl, c-Kit, and the PDGF receptor (43,44). Because each tyrosine kinase plays a role in specific signal transduction pathways, we tried to specify the kinase pathway that is responsible for the observed anti-prion effect in prion-infected cells. Our experiments revealed that specific inhibition of the c-Kit kinase and the PDGF receptor had no effect on PrP Sc propagation. In contrast, abrogation of endogenous c-Abl activity by overexpression of a trans-dominant negative c-Abl mutant led to a decrease of PrP Sc in ScN2a cells. The reported IC 50 of STI571 for c-Abl activity ranges between 0.1 and 0.3 M for the inhibition of autophosphorylation (in vitro) and Ͻ1 M for the inhibition of cellular proliferation (as determined by cell culture experiments) (57). We observed a comparable difference between the IC 50 of STI571 in vitro (ϳ0.1 M) and that determined for the anti-prion effect in cell cultures (Յ1 M). The wide range is likely due to the diverse effects that are measured. The in vitro assays show the direct effect of STI571 on the phosphorylation of a synthetic substrate of c-Abl. The anti-prion effect, calculated in cell culture experiments, is an indirect effect lying downstream of a signaling cascade. Therefore, it is difficult to compare IC 50 values determined by the two different assays. A concentration of STI571 leading to a 50% inhibition of the phosphorylation of a synthetic substrate in an in vitro kinase assay most probably is not sufficient to activate the degradation of PrP Sc , resulting in a 50% reduction of PrP Sc in prion-infected cells. Based on our studies we conclude that among the known targets of STI571, the tyrosine kinase c-Abl is responsible for the enhanced lysosomal degradation observed in our experiments. Interestingly, c-Abl seems not to be involved in the observed reduction of ␤-amyloid production. STI571 inhibited the production of ␤-amyloid in wild type as well as in c-Abl Ϫ/Ϫ mouse fibroblasts (49). This would suggest that different mechanisms underlie the anti-prion effect and the effect of STI571 on the ␤-amyloid production.
C-Abl is a ubiquitously expressed protein that has a variety of functions, but little is known about its activators and targets. In neurons and hematopoetic cells, c-Abl is primarily located in the cytosol, where it is involved in cytoskeletal rearrangement. In most other cell lines, c-Abl is found in the nucleus and has various functions. Under normal conditions, c-Abl is inactive or is strictly regulated. Constitutively active c-Abl has transforming properties (58), and overexpression of c-Abl in cell culture induces cell cycle arrest in G 1 , with subsequent apoptosis (45)(46)(47). This is consistent with the failure of overexpressing c-Abl in our cell culture experiments. Studies to further characterize the underlying anti-prion mechanism, the affected downstream signaling cascade that leads to the activation of the degradation, and the involved proteins are in progress. The results might help define the normal cellular PrP Sc degradation mechanism and indicate potential targets for the therapy of prion diseases.
Clearance of PrP Sc without Interference within Biogenesis and Localization of PrP c -The treatment of prion-infected cells with STI571 leads to a time-and dose-dependent decrease of PrP Sc . The clearance of PrP Sc from treated cells was irreversible, and even 30 days after the removal of STI571, PrP Sc did not reappear. Bioassays in mice are in progress to confirm that STI571-treated cells are also cleared of specific prion infectivity. Cell specific differences were observed when ScGT1 and SMBs.15 cells were treated with STI571. ScGT1 and ScN2a cells were both almost equally sensitive to STI571. After a treatment for 3 days, PrP Sc was undetectable in both cell types. In the non-neuronal SMB cell line, the anti-prion effect could only be seen after treatment for 10 days. Cell type-specific effects might be due to signaling differences and could also be explained by differences in the protein equipment of the cells. Of note, both neuronal cell lines were highly sensitive to STI571, whereas the more resistant SMBs.15 cells are of nonneuronal origin (39). A compound can inhibit PrP Sc propagation in prion-infected cells by various mechanisms. For example, inhibition can be caused by overstabilizing the conformation of PrP molecules as shown for Me 2 SO and Congo Red (14,59) or by down-regulating the amount of PrP c in the cell. Suramin induced the formation of insoluble PrP c aggregates and the intracellular re-routing of PrP c (13). Thus, PrP c was removed from the cell surface and was no longer available for the conversion into PrP Sc . Therefore, it was mandatory to analyze possible effects of STI571 on PrP c . In contrast, confocal microscopy and surface FACS analyses showed no difference in the subcellular localization of PrP c between mock-treated cells and cells treated with STI571. This was biochemically confirmed by PIPLC digestion. Solubility assays demonstrated that the biochemical features of PrP c were not changed by STI571 treatment. We conclude that the anti-prion effect of STI571 is not due to an altered localization or biochemical behavior of PrP c , as has been reported for other drugs, suggesting the cellular isoform of the prion protein is not affected by STI571.
Inducing the Lysosomal Degradation of Pre-existing PrP Sc -Many substances have been identified that inhibit PrP Sc propagation in experimental systems. Interestingly, almost all of the anti-prion compounds interfere with the de novo formation of PrP Sc . For branched polyamines, however, it has been reported that they affect pre-existing PrP Sc and that they act in lysosomes. These studies also raised the possibility that lysosomal proteases normally degrade PrP Sc in cultured cells. The authors claimed that branched polyamines could render PrP Sc molecules protease-sensitive by dissociating PrP Sc aggregates or that they could facilitate the transport of PrP Sc from the membrane into secondary lysosomes (26,27). Similarly, the two anti-prion compounds quinacrine and chloroquine were reported to act in lysosomes, eventually by altering the lysosomal pH (60). STI571 has no major effect on the de novo conversion of PrP c to PrP Sc but induces the degradation of preexisting PrP Sc by decreasing its half-life from more than 24 h to less than 9 h. We were able to inhibit this effect with NH 4 Cl, which blocks the degradation in lysosomes by raising the lysosomal pH. Therefore, we suggest that STI571 activates the lysosomal degradation of pre-existing PrP Sc . Notably, treatment with NH 4 Cl alone seemed to lead to a slightly increased PrP Sc accumulation. This confirms the observation that lysosomal proteases are involved in the normal degradation of PrP Sc in prion-infected cells and that STI571 activates this process. Interestingly, the effect of branched polyamines on PrP Sc accumulation could not be inhibited by NH 4 Cl (26,27). Therefore, STI571 and these compounds employ different pathways. It is important to note that the amount of PrP c was not influenced by STI571 treatment, indicating that PrP c and PrP Sc are degraded by different proteolytic activities. This is, therefore, not contradictory to studies with proteasomal inhibitors that induce the accumulation of self-perpetuating "PrP Sclike" aggregates in the cytosol (61,62). The authors claim that defective proteasomal degradation might have an important role in the origin of PrP Sc . This might be of relevance for the spontaneous initiation in the sporadic or familial forms of prion diseases. Nevertheless, further propagation of PrP Sc and the spread from cell to cell primarily occurs on the cell surface and/or in early compartments of the endocytic pathway rather than in the cytoplasm. Therefore, the cellular clearance of PrP Sc is most probably carried out by lysosomal proteases in secondary lysosomes.
In summary, our data show that STI571 activates the cellular clearance of PrP Sc in prion-infected cells, most probably by inhibition of the tyrosine kinase c-Abl. Clearance occurs in lysosomes, as lysosomal inhibition can abolish the effect. Moreover, our finding that inhibition of lysosomes alone led to an increase in PrP Sc is consistent with the model in which lysosomes play an important role in the degradation of PrP Sc in prion-infected cells.