Prion-associated Increases in Src-family Kinases*

The prion diseases result from the generation and propagation of an abnormal conformer of the prion protein. It is unclear how this molecular event dis-rupts neuronal function and viability. Current evidence argues it is not due to loss of normal prion protein activity or direct toxic effects of the abnormal conformer. Both the normal and abnormal prion proteins are glycosylphosphatidylinositol-linked membrane proteins. Conversion to the abnormal isoform results in the formation and accumulation of prion protein aggregates. Because aggregation of glyco-sylphosphatidylinositol-linked proteins activates Src-family kinases, the activation status and levels of the Src-family kinases in prion disease were investigated. Elevations of Src-family kinases were found in a cell culture model and two separate animal models of prion disease. The elevations in Src kinases preceded the onset of symptoms and occurred concurrently with the appearance of detergent-insoluble prion protein. In addition, the total level of kinases phosphorylated at tyrosine residues associated with activation was increased. Similar alterations were not present in brain homogenates from presymptomatic animals early in the disease course, prion protein-ablated animals, or end-stage Tg2576 mice overexpressing mutant amyloid precursor protein. Identification of similar elevations in cell culture and animal model systems suggests the elevations are a specific response to the presence of the disease-associated conformer. Abnormal regulation of these signal transduction cascades may be a key element in the cellular pathology of the prion diseases. was performed using standard protocols by incubating 0.5 mg of brain homogenate with 5 (cid:2) l of pp60Src-protein A-Sepharose resin (35). The pp60Src was covalently coupled to protein A-Sepharose at a concentration of 1.0 (cid:2) g antibody/ (cid:2) l protein A resin using dimethyl pimelimidate (35). Bound antigen was eluted with SDS buffer under non-reducing conditions.

diseases are characterized by neuronal loss, astrogliosis, vacuolar degeneration, and variable degrees of amyloid plaque formation (1). Ultrastructural examination has shown vacuolar degeneration results from synaptic deterioration, cytoskeletal disruption, membrane dissolution, and membrane fusion (2). Quantitative studies suggest a progression from synaptic loss to neuritic degeneration and neuronal death (3). A wealth of data indicate these diseases are due to prions (4). Prions are self-propagating proteinaceous infectious particles composed of an abnormal conformer of a normally expressed protein known as the prion protein (PrP). 1 PrP typically exists as a normal cellular conformer (PrP C ), whereas a ␤-sheet enriched, disease-associated form is referred to as PrP Scrapie (PrP Sc ). Recently, Legname et al. (5) generated infectious prions using recombinant prion protein, adding further support to this hypothesis. The normal function of PrP C is unknown; it may participate in synaptic structure (6), neurite formation (7), copper metabolism (8), or possibly signal transduction (9). PrP Sc is produced by the conformational conversion of pre-existing PrP C to the abnormal isoform. PrP Sc uniformly aggregates, is insoluble in detergents, and tends to be resistant to proteolytic degradation (4).
Formation of PrP Sc is the initiating step in prion disease pathogenesis. Experimental inoculation of animals with prions results in the appearance of PrP Sc before the development of symptoms and the colocalization of PrP Sc with the pathologic lesions (10,11). It has been suggested that loss of normal PrP C function or, alternatively, direct toxic effects of PrP Sc are responsible for the development of disease. Loss of PrP C seems unlikely to be the principal biochemical lesion because prion protein-ablated mice (PrP o/o ) fail to develop scrapie (12)(13)(14), and furthermore, one recent publication suggests total PrP C levels may increase during the disease process (15). The toxic impact of PrP Sc also remains unclear. Conversion to PrP Sc has been reported to alter PrP metal binding and antioxidant activity (16), lysosomal function (17), and membrane fluidity (18). In addition, PrP Sc and a fragment of PrP spanning region 106 -126 have been reported to be cytotoxic to cells in culture (19,20). Nonetheless, inoculation of PrP o/o mice with PrP Sc results in neither the clinical nor pathologic changes of scrapie, whereas the neuropathologic changes of prion disease and PrP Sc formation were restricted to regions of PrP-expressing tissues in brain transplantation studies (12,21,22). Together, these findings suggest that in situ conversion of endogenous PrP C to PrP Sc per se may be essential to the pathologic process.
A number of features suggest an alternative mechanism linking PrP Sc formation to the development of disease. We have shown previously (23) that PrP C and PrP Sc are glycosylphos-phatidylinositol (GPI)-linked membrane proteins present on the external plasma membrane surface. Although the mechanism remains to be clearly defined, it is well recognized that aggregation of GPI-linked proteins, including PrP C , activates Src-family kinases (SFKs) (9,24,25). The SFKs have diverse roles ranging from control of cytoskeletal elements to regulation of the cell cycle and apoptosis (26), and they have been implicated in synaptic plasticity and neurodegeneration (27,28). GPI-linked proteins, the SFKs, and a variety of other proteins are localized to cholesterol-and sphingomyelin-enriched membrane microdomains known generically as "rafts," and specifically, PrP C and Src have been colocalized to the same raft compartments (24,26,29). As such, the inherent property of PrP Sc to form aggregates could result in the unregulated activation of these kinases. In addition, PrP Sc has a prolonged half-life (30), thus persistence of PrP Sc aggregates may affect the levels of the SFKs. The impact of PrP Sc on SFK levels and activation was therefore examined in three distinct models of prion diseases. These studies reveal elevations in multiple SFKs and demonstrate increased levels of phosphotyrosine residues associated with SFK activation. The proximate role of the SFKs in multiple diverse cellular signaling cascades suggests that dysregulation of these molecules could have profound pathologic consequences.

MATERIALS AND METHODS
Animals-Tg2866(MoPrP-P101L)PrP o/o mice were generously provided by Dr. Stanley Prusiner (University of California San Francisco) (31). These mice carry a proline to leucine mutation at position 101 of mouse PrP, the mouse equivalent of the PrP mutation causing the familial human prion disease Gerstmann-Staussler-Scheinker disease. The diagnosis of scrapie in Tg2866(MoPrP-P101L)PrP o/o mice was based on established clinical, pathologic, and biochemical criteria (31). The presence of PrP Sc in Tg2866(MoPrP-P101L)PrP o/o mice was confirmed by detergent insolubility (31) (data not shown). For the purposes of clarity, symptomatic Tg2866(MoPrP-P101L)PrP o/o transgenic animals are referred to as "scrapie-sick." PrP o/o mice were derived from back-crossings of Tg2866(MoPrP-P101L)PrP o/o mice with commercially available wild-type FVB mice. Mice were genotyped by PCR using primers specific for the mouse PrP gene (5Ј-AGGGTTGACGCCAT-GACTTTC-3Ј and 5Ј-CAGGGCCCGACACAGAGAAGCAAGAATGAG-3Ј) or the 3Ј-untranslated hamster sequence present within the mutated PrP transgene insert (5Ј-CCACCAAGGGGGAGAACTTCAC-3Ј and 5Ј-TCCCGTTGCCTTACTCAGCTAG-3Ј). PCR was performed on an Applied Biosystems GeneAmp 2700 thermocycler. Amplification of the mouse PrP gene was performed using Taq polymerase (conditions: 94°C for 5 min; 94°C for 0.5 min, 64°C for 0.5 min, and 72°C for 0.5 min ϫ 30 cycles; and 72°C for 7 min); these conditions result in a 1-kb amplicon for the wild-type PrP gene and a 1.5-kb amplicon for the ablated PrP gene containing the neomycin insert. Amplification of the 3Ј-untranslated region of the hamster insert was performed using Vent polymerase (settings: 97°C for 7 min; 96°C for 0.5 min, 65°C for 0.5 min, and 72°C for 1.0 min ϫ 30 cycles; and 72°C for 5 min), resulting in a 680-bp amplicon. Brains from wild-type FVB mice inoculated with the Rocky Mountain Laboratory (RML) prion strain or from un-inoculated controls were generously provided by Dr. Stephen DeArmond (University of California San Francisco). Ten percent brain homogenates were prepared in 50 mM Tris-HCl, pH 7.3, 5 mM EDTA, 50 mM NaF, 25 mM Na 4 P 2 O 7 , 1 mM Na 3 VO 4 , 1% Nonidet P-40, 1% deoxycholate, and Complete protease inhibitor mixture (Roche Applied Science). Brains from 24-month-old littermate control mice and end-stage transgenic Tg2576 mice carrying the K670N/M671L double mutation in the human amyloid precursor protein (APP) (32) were a kind gift from Dr. Joseph Quinn (Oregon Health & Sciences University). All animal protocols were approved by the Oregon Health & Sciences University Institutional Animal Care and Use Committee.
Cell Cultures-Mouse neuroblastoma N2a and ScN2a cell lines were a generous gift of Dr. Stanley Prusiner (University of California San Francisco). The presence of PrP Sc in ScN2a was confirmed by proteinase K digestion (data not shown). ScN2a and N2a cells were cultured as described previously (33), rinsed with phosphate-buffered saline, and then solubilized in 50 mM Tris-HCl, pH 7.3, 5 mM EDTA, 50 mM NaF, 25 mM Na 4 P 2 O 7 , 1 mM Na 3 VO 4 , 1% Nonidet P-40, 1% deoxycholate, and Complete protease inhibitor mixture (Roche Applied Science).
Western Blot Analysis-Protein concentrations were determined using the Pierce BCA assay. Fifty-microgram aliquots of brain or cell culture homogenate were separated on SDS-polyacrylamide gels and then transferred to nitrocellulose by standard techniques. Loading of equal amounts of protein was confirmed by Ponceau S staining of the nitrocellulose membrane. Nonspecific binding was blocked with 5% nonfat dried milk in 20 mM Tris, pH 7.5, 0.15 M NaCl, and 0.05% Tween 20 (TBST). All antibodies were diluted in TBST. Antibody sources, species, and dilutions were as follows: rabbit SRC-2, rabbit N-16, and mouse monoclonal antibodies against Fyn, Lck, and Yes were from Santa Cruz Biotechnology; each was used at 1:1000. Mouse monoclonal pp60Src (clone 327) (34) was a kind gift of Dr. Joan Brugge (Harvard) and used at 1:5000. Mouse monoclonal anti-human transferrin receptor (Zymed Laboratories Inc.) was used at 1:500; rabbit anti-phosphotyrosine-416 was purchased from Cell Signaling and used at 1:500. Horseradish peroxidase-labeled goat anti-mouse and horseradish peroxidaselabeled donkey anti-rabbit were purchased from Pierce and used at 1:25,000. Bound antibody was detected with SuperSignal (Pierce) chemiluminescent reagent and exposure to light-sensitive film. Western blot exposures were digitized using a flat bed scanner and analyzed using NIH Image (public domain software for Macintosh computers developed at the United States National Institutes of Health and available on the internet at rsb.info.nih.gov/nih-image/). Statistical analysis was performed using InStat (GraphPad software). Values on an individual blot were normalized against the average density from wild-type animals on that same blot. Levels of SRC-2 immunoreactive material in scrapie-sick, PrP o/o , and wild-type mice or of Src, Fyn, Lck, and Yes between mice or between cells were compared using one-way ANOVA with Bonferroni multiple comparison post-test. Comparisons between scrapie-sick animals and wild-type animals or between ScN2a cells and N2a cells were performed using the unpaired t test.
Immunoprecipitation-Immunoprecipitation was performed using standard protocols by incubating 0.5 mg of brain homogenate with 5 l of pp60Src-protein A-Sepharose resin (35). The pp60Src was covalently coupled to protein A-Sepharose at a concentration of 1.0 g antibody/l protein A resin using dimethyl pimelimidate (35). Bound antigen was eluted with SDS buffer under non-reducing conditions.

RESULTS
Src-family tyrosine kinases share a common structural motif characterized by multiple conserved functional domains and a unique N-terminal region (26). In order to maximize the probability of detecting changes in SFK levels, initial studies screened brain homogenates from scrapie-sick transgenic mice, as well as wild-type and PrP o/o controls, utilizing an antibody against the conserved C terminus (SRC-2). These studies revealed a band that was clearly increased in the scrapie-sick mice as compared with the wild-type and PrP o/o mice (Fig. 1A). The apparent molecular mass of the immunoreactive band was ϳ60 kDa, consistent with that expected for the SFKs. Wildtype and PrP o/o control mice had similar levels of SFKs. Densitometric analysis of four separate sets of scrapie-sick, wildtype, and PrP o/o mice revealed a 1.8-fold increase in SFKs in scrapie-sick mice as compared with wild-type controls. ANOVA with Bonferroni's multiple post-test comparison confirmed a statistically significant increase in the scrapie-sick Tg2866(MoPrP P101L)/PrP o/o mice ( Fig. 1B; p ϭ 0.0038), with no difference between wild-type controls and PrP o/o mice. Levels of SFKs were also examined in a scrapie-sick RML-inoculated mouse and an asymptomatic un-inoculated control (Fig. 1C); densitometric analysis (data not shown) revealed a 1.6-fold increase. Because only a single pair of brains using the RML model was available, statistical comparison could not be performed. Nonetheless, these studies revealed a remarkable similarity between the scrapie-sick RML mouse and the transgenic animals.
During the course of the prion diseases, the brain undergoes a number of reactive and degenerative changes such as astrogliosis, vacuolation, and neuronal loss. In addition, the PrP 106 -126 peptide has been shown to activate the SFKs Lyn and Syk in microglial cells (36). In order to clarify whether the observed SFK elevations in brain are due primarily to PrP Sc rather than to a secondary nonspecific reactive process, levels of SFKs were determined in N2a and ScN2a cells. Western blotting with the SRC-2 antibody revealed a marked increase in SFK immunoreactivity in ScN2a cells as compared with N2a cells (Fig. 2A). Densitometric analysis confirmed a very highly statistically significant increase of SRC-2 immunoreactivity in the ScN2a cells (  (32). Histologic examination demonstrated abundant cortical and hippocampal plaques (Fig. 3B), indicating the lack of SFK increase was not due to the absence of aggregated amyloid ␤-peptide (A␤).
In the above-mentioned studies, equal amounts of sample were compared as determined by protein determinations and Ponceau S staining of the nitrocellulose membrane before immunoblotting. However, to further ensure that the observed differences in SFK levels were not the result of irregularities in sample preparation or loading, Western blotting for the transferrin receptor was performed. The transferrin receptor is a non-raft membrane protein frequently used as a control in studies of SFK activation due to aggregation of GPI-linked proteins. As shown in Fig. 4, equivalent levels of transferrin receptor were present in scrapie-sick, wild-type, and PrP o/o mice. Similarly, immunoblotting of cell lysates revealed that the ScN2a cells had equal or slightly lower levels of transferrin receptor as compared with N2a cells (Fig. 4). Thus, the elevations in SFK immunoreactivity seen with PrP Sc infection were not due to disparities in sample loading and were not the result of a generalized increase in membrane protein content.
Confirmation that the elevated SRC-2 immunoreactive material in the scrapie-sick mice was indeed a SFK was obtained using a combination of immunoabsorption and Western blotting. Brain homogenates were first immunoabsorbed with the Src-specific monoclonal antibody pp60Src (34). Absorbed material was subsequently eluted and analyzed by Western blotting with either SRC-2 antibody as described above or N-16 (an antibody specific for the unique domain of Src). These studies revealed a solitary 60-kDa band in the scrapie-sick samples with both SRC-2 and N-16 (Fig. 5); no detectable immunoreactive material was found in the wild-type control samples. Analysis of the immunoabsorbed brain homogenate supernatants found no residual Src (data not shown). As such, the lack of immunoprecipitable Src from the control homogenate is most probably due to nonspecific adsorptive losses. These results confirm the SRC-2 immunoreactive band identified in Fig. 1A contains Src and, furthermore, specifically identify Src as one of the SFKs elevated in scrapie-sick mice.
In addition to Src, the Src-family kinases Fyn, Yes, and Lck are present in brain (26). Because the SRC-2 antibody recognizes multiple SFK members, studies were performed to determine the levels of these SFKs in scrapie infection. A composite image of typical results from Western blotting of mouse brain homogenates probed with antibodies specific for the individual SFKs is shown in Fig. 6A. Comparison of scrapie-sick and wild-type samples suggests elevations in all four SFKs, although there was considerable variability in the extent of increase. Visual inspection suggests the largest increases are in Src and Fyn, with lesser increases in Yes and Lck. Comparison of ScN2a and N2a cell lysates also shows elevations in multiple SFKs, but with a slightly different pattern (Fig. 6B). Src and Lck appear increased; however, unlike the transgenic animals, Fyn demonstrates a much more modest increase. Yes is marginally increased. Densitometric analysis of blots from multiple individual scrapie-sick and wild-type animals or homogenates of N2a and ScN2a cells is shown in Fig. 6, C and D, respectively. In scrapie-sick animals, Src is increased ϳ3-fold, whereas Yes, Fyn, and Lck are increased ϳ2-fold. Statistical analysis (ANOVA with Bonferroni post-test correction, p Ͻ 0.0001) confirms a highly significant elevation in Src (p Ͻ 0.001); the elevations in the remaining SFKs did not reach statistical significance. In ScN2a cells, Src and Lck are increased 2-fold and 3-fold, respectively, whereas Yes and Fyn are elevated by only ϳ30 -50%. Statistical analysis (ANOVA with Bonferroni post-test correction, p Ͻ 0.0001) again found a highly significant difference in SFK levels with statistically significant elevations in Src (p Ͻ 0.05) and Lck (p Ͻ 0.001). These studies demonstrate increased steady-state expression of multiple Src-family kinases in both the animal and cell culture models of prion disease.
SFK enzymatic activity is regulated by phosphorylation of key tyrosine residues. Specifically, phosphorylation of Src tyrosine 416 or its equivalent in the other SFKs is associated with increased tyrosine kinase activity (26). The phosphotyrosine status of the SFKs in PrP Sc -infected animals and cells was therefore determined using an antibody specific for this modification. A prominent band of ϳ60 kDa was identified in the scrapie-sick mice with little immunoreactive material in the wild-type or PrP o/o control animals (Fig. 7A). Densitometric analysis found a 4.7-fold increase ( Fig. 7B; ANOVA with Bonferroni post-test correction, p Ͻ 0.0001) in the scrapie-sick animals as compared with wild-type animals. Similar increases in phosphotyrosine 416 were found in the scrapie-infected cells (data not shown). These findings suggest scrapie-sick mice and ScN2a cells have increased steady-state levels of activated SFKs. Comparison of the relative increases in phosphotyrosine 416 and total levels of SFKs yields a ratio of 1.42, suggesting a relative increase in the percentage of activated kinase as compared with controls. Furthermore, comparison of wild-type mice and PrP o/o mice demonstrates a similar level of phosphotyrosine 416, indicating that ablation of PrP does not result in increased kinase activation.
The Tg2866(MoPrP P101L)/PrP o/o mice used in many of these studies offer the advantage of developing disease in a highly predictable manner. In our hands, these mice develop disease at 127.7 Ϯ 4.5 days (Fig. 8A)  spond to ϳ80% and 90% of the time through the disease time course. At each time point, the levels of total, detergent-soluble, and detergent-insoluble PrP, as well as total SFKs, were determined in brain homogenates and normalized to those from a wild-type control of 126 days of age. As shown in Fig. 8B, levels of total PrP, soluble PrP, and SFKs were the same in the 100-day-old transgenic animal and wild-type control. Trace amounts of detergent-insoluble PrP were detected at this time. By 118 days of age, levels of SFKs and detergent-insoluble PrP were clearly increased, although these animals were completely asymptomatic. Levels of soluble PrP remained equivalent to the control. With symptom onset at 126 days, the level of SFKs was increased ϳ2-fold (similar to the increase found by densitometric analysis for transgenic and inoculated animals above). Detergent-insoluble PrP continued to accumulate, whereas detergent-soluble PrP appeared to decrease. DISCUSSION Whereas the results presented in this study do not establish a direct mechanism, the following findings strongly support the hypothesis that the increase in SFKs is due to the presence of PrP Sc . First, the absence of detectable SFK abnormalities in the PrP o/o mice shows the increases are not due to the loss of PrP C . Second, adult presymptomatic Tg2866(MoPrP-P101L)PrP o/o mice had no elevation in SFKs, indicating the elevations were not due to the 8-fold overexpression of the PrP transgene (31). Third, no increase in SFK levels was detected in response to amyloid plaques produced by APP transgenic animals overexpressing mutant human amyloid precursor protein by 5-fold (32). Fourth, the increase in SFKs and detergentinsoluble PrP occurred concurrently and preceded symptom onset. Fifth, similar increases in SFKs were detected in the well-established ScN2a cell culture system as well as two discrete animal models of prion disease, Tg2866(MoPrP P101L)/ PrP o/o transgenic mice and RML-inoculated mice. In addition, the cell culture studies demonstrate that neuronal cells infected with PrP Sc are sufficient to produce the SFK elevations but do not exclude contributions by other cell types. Additional studies will be needed to delineate the contributions of the individual cellular components to the overall elevations in SFKs.
The precise mechanism linking PrP Sc formation to the activation and accumulation of the SFKs is unknown; however, a number of features suggest it is due to the raft-associated form of the abnormal protein. Specifically, we have demonstrated previously (23) that PrP Sc is associated with membrane rafts. Recently, PrP C and Src have been shown to colocalize to the same raft domain (29). In addition, because PrP Sc has a markedly increased half-life (30), it is conceivable that formation of PrP Sc aggregates leads not only to SFK activation but also to a decrease in the degradation of raft-associated proteins. In this regard, it is interesting that CD9 (37), the insulin receptor (38), and the bradykinin receptor (18) are also increased in the prion diseases, and all have been reported in rafts (reviewed in Ref. 24). Of these, messenger RNA levels have only been investigated for the insulin receptor; no increase was identified (38), consistent with an alteration in protein turnover. The observation in the current study that the transferrin receptor was not increased is also consistent with this model because it is not a raft-associated protein. In addition, it interesting to note that inoculation of mice with anti-PrP antibodies induces neuronal death (39), consistent with a model of PrP Sc aggregation-induced pathology.
The mechanism of PrP Sc aggregation is unknown. As stated earlier, peptides spanning residues 106 -126 of PrP have been shown to activate SFKs in microglial cells. In addition, Gu et al. (40) reported that this peptide could induce PrP aggregation initiating cell death in vitro. NMR studies of PrP C found that this domain existed in an unstructured form (41); however, more recent crystallographic data suggest that this region becomes structured and packed in PrP C dimers with evidence for three-dimensional domain swapping and formation of intermolecular disulfide bonds (42) in the dimers. It is therefore conceivable that this domain is involved in mediating aggregation of PrP Sc . Regardless of the mechanism of aggregation, it is likely that localization of PrP Sc to the membrane rafts is via their GPI anchor.
Rafts are enriched in cholesterol, sphingolipids, and a variety of receptors and signaling molecules. Although the existence of rafts remains controversial, these domains have been implicated in a number of biologic events, including membrane fusion (24). Perhaps PrP Sc alters normal raft function, resulting in unregulated membrane fusion as seen in the prion diseases by electron microscopy (2). Current evidence indicates that rafts are a mixed population of microdomains, each with a discrete composition; thus, not all rafts may contain PrP Sc , and not all PrP Sc -containing rafts may have the same complement of kinases. Such a distribution could explain the mixed pattern of elevations shown in Fig. 6. The impact of PrP Sc on raft composition and protein turnover is currently being investigated in our laboratory. Significant alterations in raft function and/or homeostasis may indicate a new class of disorders, i.e. "raftopathies." Src-family kinases are ubiquitously expressed tyrosine kinases that regulate and may be regulated by numerous key cellular pathways (reviewed in Ref. 26). A partial list of functions modulated by these kinases includes cell survival, apo-  (31). The percentage of disease-free animals is plotted against the age of the animals in days. As can be seen, these animals develop disease in a highly predictable manner. B, brain homogenates from Tg2866(MoPrP(P101L))/PrP o/o mice sacrificed at the indicated times were analyzed for levels of SFKs and the presence of detergent-soluble (PrP C ), detergent-insoluble (PrP Sc ), or total PrP. Each component was compared against levels from a wild-type control mouse of 126 days of age. SFK accumulation parallels the appearance of detergent-insoluble PrP Sc and precedes symptom onset. ptosis, ion channel activity, and cytoskeletal assembly. SFKs regulate these activities via direct phosphorylation of tyrosine residues and/or binding to discrete docking domains on other proteins. Whereas most studies have focused on the regulatory impact of the tyrosine kinase activity of the SFKs, a small number of articles demonstrate that regulation may also occur via kinase-independent interactions. Catalytically inactive SFKs lacking the active site have been shown to act as dominant negative regulators (43)(44)(45)(46) and to rescue SFK-dependent pathways (47,48). Thus the increased SFK levels identified in the current study may interfere with proper signal transduction by mass action effects, whereas the elevated levels of activated SFKs could result in aberrant phosphotyrosine formation. For example, abnormalities in N-methyl-D-aspartate receptors and potassium channels have been reported in the prion diseases (49,50), and both can be regulated by SFKs (51)(52)(53). Likewise, SFKs regulate cytoskeleton formation (26); therefore, aberrant SFK activity could affect cytoskeletal stability. It is conceivable that disruption of ionic homeostasis and cytoskeletal integrity may be involved in the spongiform degeneration characteristic of these disorders. Similarly, anomalous SFK activity could generate apoptotic signals leading to neuronal loss, analogous to the "synaptic apoptosis" model proposed by Mattson (54).
Although there is a growing body of literature regarding various signaling cascades in the neurodegenerative diseases (36,(55)(56)(57)(58), relatively little is known about the role of the Src-family kinases. A role in Alzheimer's disease has been suggested by the finding that A␤ can induce Src-dependent tau phosphorylation (28) and increase the stable association of Fyn with focal adhesion kinase (59). In addition, immunoperoxidase studies have suggested that increased levels of Fyn colocalize with hyperphosphorylated tau (60). Conversely, loss of Fyn may protect against A␤-induced neurotoxicity (61). Similarly, Fyn phosphorylates ␣-synuclein in vitro (55), suggesting a role in Parkinson's and Lewy body disease. If the SFKs are involved in Alzheimer's disease, why did we not see SFK increases in the APP transgenic mice? This may reflect differences in the underlying biochemical mechanisms involved in SFK activation. For example, we hypothesize that PrP Sc accumulates in rafts in essentially an irreversible manner, thus providing a persistent effect on raft-associated functions. Conversely, A␤ is likely to either interact with cells and/or membranes in a reversible manner as a soluble peptide or through limited interactions with the insoluble extracellular amyloid plaque deposits, resulting in a reduced effect that could be below our level of detection.
The steps from PrP Sc formation to development of disease clearly will be complex; nonetheless, the significance of the current study lies in the remarkable similarity of the findings from the disparate model systems. These experiments compared the impact of PrP Sc on the level and activation status of SFKs in neuroblastoma cells stably infected with RML prions and brains from symptomatic transgenic animals due to a mutation in the prion gene or wild-type animals inoculated with RML prions. Given the inherent abnormalities of signaling pathways in cultured tumor cells, in addition to the mixed cellular composition of the brain, it is not surprising that minor differences in the pattern of SFK accumulation in the in vivo and in vitro systems exist. Nonetheless, in both systems, the presence of PrP Sc is associated with a significant increase in Src. This suggests that the ScN2a cell line not only accurately converts PrP C to PrP Sc but also exhibits portions of the pathologic cascade, even though the cells fail to develop the cytopathic features of disease.
In summary, the alterations in SFKs reported in the current study, the resistance of PrP o/o animals to development of disease, and the known biophysical properties of PrP Sc are consistent with a model linking PrP Sc formation with dysregulation of the Src-family kinases. PrP and Src are localized to membrane rafts. Conversion of PrP C to PrP Sc generates a conformer prone to unregulated aggregation, and because PrP Sc has a markedly prolonged half-life, it is likely that PrP Sc aggregates exist for an extended period of time, resulting in aberrant SFK regulation and turnover. Such a disturbance could clearly have severe physiologic consequences and result in many of the pathologic lesions associated with the prion diseases.