Acute formation of protease-resistant prion protein does not always lead to persistent scrapie infection in vitro.

Transmissible spongiform encephalopathies are accompanied by the accumulation of a pathologic isoform of a host-encoded protein, termed prion protein (PrP). Despite the widespread distribution of the cellular isoform of PrP (protease-sensitive PrP; PrP-sen), the disease-associated isoform (protease-resistant PrP; PrP-res) appears to be primarily restricted to cells of the nervous and lymphoreticular systems. In order to study why scrapie infection appears to be restricted to certain cells, we followed acute and persistent PrP-res formation upon exposure of cells to different scrapie agents. We found that, independent of the cell type and scrapie strain, initial PrP-res formation occurred rapidly in cells. However, sustained generation of PrP-res and persistent infection did not necessarily follow acute PrP-res formation. Persistent PrP-res formation and scrapie infection was restricted to one cell line inoculated with the mouse scrapie strain 22L. In contrast to cells that did not become scrapie-infected, the level of PrP-res in the 22L-infected cells rapidly increased in the absence of a concomitant increase in the number of PrP-res-producing cells. Furthermore, the protein banding pattern of PrP-res in these cells changed over time as the cells became chronically infected. Thus, our results suggest that the events leading to the initial formation of PrP-res may differ from those required for sustained PrP-res formation and infection. This may, at least in part, explain the observation that not all PrP-sen-expressing cells appear to support transmissible spongiform encephalopathy agent replication.

Transmissible spongiform encephalopathies (TSEs) 1 are progressive neurodegenerative diseases that include Creutzfeldt-Jakob disease, Gerstmann-Strä ussler-Scheinker syndrome, and Kuru in humans as well as scrapie in sheep and goats, chronic wasting disease in mule deer and elk, transmissible mink encephalopathy, and bovine spongiform encephalopathy (BSE). A prerequisite for TSE infection is the expression of the protease-sensitive host prion protein (PrP-sen) (1). PrP-sen is a highly conserved mammalian sialoglycoprotein of unknown function that is anchored to the cell membrane by a phosphatidylinositol moiety (2,3). During TSE infection, PrP-sen is converted to its partially protease-resistant, pathologic isoform PrP-res. This abnormal form of PrP is closely associated with TSE infectivity and has been proposed as the protein-only agent responsible for TSE diseases (4).
PrP-sen is required for susceptibility to TSE disease (1). At the cellular level, however, it is unlikely that PrP-sen expression is the sole prerequisite for TSE infection. PrP-sen is expressed in a wide variety of tissues (5)(6)(7), yet PrP-res formation appears to be restricted primarily to cells of the nervous and lymphoreticular systems (8,9). The reason for this discrepancy in PrP-sen expression and PrP-res formation is unknown. However, the fact that different scrapie strains induce PrP-res formation in different brain areas suggests that PrP-res formation is both cell type-and scrapie strain-dependent (10 -13). Evidence for scrapie strain-dependent PrP-res formation and infection also comes from infection studies with tissue culture cells, which demonstrate that the same cell line is susceptible to some scrapie strains but not to others (14 -17).
Unfortunately, the cellular prerequisites needed for PrP-res formation and infection remain elusive. One possible explanation for the inability of TSE strains to persistently infect specific cell populations may be that these cells cannot support the formation of new PrP-res induced by these strains. However, attempts to study the very early events in cell culture leading to PrP-res formation have been difficult, since 1) only a few cell culture systems have been identified that appear to be susceptible to the scrapie agent, 2) newly formed PrP-res must be discriminated from PrP-res present in the inoculum, and 3) the amount of PrP-res produced in susceptible cell lines is often relatively low.
In this study, we have used cell culture systems that allow us to follow PrP-res formation following exposure of the cells to scrapie-infected brain homogenates (18,19). We have found that PrP-res is detectable within 24 h of exposure to mouse scrapie agents. This rapid formation of PrP-res was scrapie strain-independent and could be initiated in both neuronal and nonneuronal cells. However, acute formation of PrP-res was not always indicative of a persistent scrapie infection. Only fibroblast cells that had been exposed to the mouse scrapie strain 22L persistently produced PrP-res and infectivity. Quantitative and qualitative changes in newly formed PrP-res accompanied the establishment of a persistent infection and could be indicative of the scrapie agent localizing to a cellular compartment suitable for efficient PrP-res formation. Thus, our data suggest that whereas acute formation of PrP-res may occur in a variety of cell types, initiation of persistent PrP-res formation and scrapie infection requires additional cell-and/or scrapie strain-specific factors for the scrapie agent to adapt to and persistently infect a cell.

EXPERIMENTAL PROCEDURES
Cell Lines-The mouse neuroblastoma cell line Mo3F4-MNB and the mouse fibroblast cell cultures Mo3F4-⌿2C2 and Mo3F4-⌿2/PA317 have been described previously (16,19). Cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum. All cell cultures express high levels of mouse PrP-sen containing the epitope to the hamster PrP-specific mouse monoclonal antibody 3F4 (Mo3F4). The 3F4 epitope had previously been inserted into mouse PrP by two amino acid substitutions at positions 108 and 111 (20).
Antibodies-The mouse monoclonal antibody 3F4 was used to distinguish exogenous PrP molecules from the endogenous mouse PrP present in both mouse brain homogenate and mouse cell lines. The rabbit polyclonal antiserum R30 was used to detect mouse PrP-sen expressed in wild-type mouse neuroblastoma cells (21).
Scrapie Strains and Preparation of Inocula-The mouse-adapted scrapie strains ME7, 87V, and 22L were the kind gift of Dr. James Hope (Lasswade Veterinary Laboratory, Penicuik, Midlothian, UK) and were passaged in C57Bl/10 (ME7, 22L) or VmDk (87V) mice. Scrapie strain RML was passaged in RML mice. Brain homogenates were prepared in Dulbecco's phosphate-buffered saline (10%, w/v) using the brains of terminally sick animals. Brain tissue was homogenized and sonicated for 5 min. Cell debris was removed by low speed centrifugation (10 min, 4°C, 2000 ϫ g), and the resultant brain homogenates were stored at Ϫ70°C. For cell inoculation, brain homogenate was diluted into Opti-MEM medium (Invitrogen) and sonicated for 2 min prior to use. The level of PrP-res and infectivity present in the brain homogenates used were similar for each strain. The titers of the different strains used were based upon the amount of inoculum needed to infect 50% of animals that had been injected intracranially (IC) with an infected brain homogenate (ID 50 ). In C57Bl/10 mice, ME7 had an IC ID 50 of 2.2 ϫ 10 7.9 infectious units/g of brain tissue, whereas 22L had an IC ID 50 of 2 ϫ 10 8.8 infectious units/gram of brain tissue. Scrapie strain RML had an IC ID 50 of 3.3 ϫ 10 7.8 infectious units/gram of brain tissue when titered in RML mice. The titer of 87V in VmDk mice is ongoing.
Exposure of Cells to Scrapie Agent and Detection of Newly Formed PrP-res-Experiments were performed in 24-well microtiter plates. Cells (0.5 ϫ 10 5 /well) were overlaid with different dilutions of 10% brain homogenate in Opti-MEM. For kinetics studies, different cell numbers were plated to ensure comparable levels of cells at the time point of harvest (4 h, 4 ϫ 10 5 ; 24 h, 4 ϫ 10 5 ; 48 h, 2 ϫ 10 5 ; 72 h, 1 ϫ 10 5 ; 96 h, 0.5 ϫ 10 5 cells/ml). Kinetic studies performed with the same number of cells plated (0.5 ϫ 10 5 /well) were also done and gave similar results when compared with experiments where different numbers of cells were plated. After 4 h, 400 l of DMEM plus 10% fetal bovine serum was added. Cells were either directly lysed or incubated for up to 96 h before being lysed in lysis buffer (1 mM Tris-HCl, pH 7.4, 140 mM NaCl, 5 mM EDTA, 0.5% sodium deoxycholate, and 0.5% Triton X-100). Cell lysates were cleared of cell debris by centrifugation (5 min, room temperature, 16,000 ϫ g). For Mo3F4 PrP-sen detection, one-hundredth of the cell lysate was reserved. In order to detect PrP-res, the remaining cell lysate was incubated with 20 g/ml Proteinase K (PK) for 40 min at 37°C. The reaction was stopped by the addition of phenylmethylsulfonyl fluoride to a final concentration of 3 mM, and samples were centrifuged at 200,000 ϫ g for 1 h at 4°C. Pellets were sonicated into sample buffer (2.5% SDS, 3 mM EDTA, 2% ␤-mercaptoethanol, 5% glycerol, 0.02% bromphenol blue, and 63 mM Tris-HCl, pH 6.8) and assayed on 14% NOVEX precast gels (Invitrogen). PrP-sen and newly formed PrP-res were detected by Western blot using the mouse monoclonal antibody 3F4 and enhanced chemiluminescence or enhanced chemifluorescence according to the manufacturer's instructions (Amersham Biosciences).
Metabolic Labeling and Immunoprecipitation of PrP-sen-Confluent monolayers of cells were labeled with trans-[ 35 S]methionine/cysteine (PerkinElmer Life Sciences) as previously reported (21). Briefly, cells were rinsed in phosphate-buffered balanced salt solution and incubated in methionine/cysteine-free DMEM for 1 h. Trans-[ 35 S]methionine/cysteine was added to a final concentration of 15 Ci, and cells were incubated for 2 h at 37°C. Monolayers were lysed, and cell debris was removed by low speed centrifugation (10 min, 4°C, 2000 ϫ g). Proteins were precipitated in 4 volumes of methanol and sonicated into detergent lipid protein complex buffer (4.2 mg of L-␣-phosphatidylcholine per ml, 123 mM NaCl, 50 mM Tris-HCl, pH 7.5, 1% N-lauroylsarcosine). Samples were divided into two fractions, and PrP-sen was immunoprecipitated with either the polyclonal antiserum R30 to detect total PrP or the monoclonal antibody 3F4 to detect epitope-tagged PrP. Proteins were separated on 14% NOVEX Precast gels (Invitrogen), and radiolabeled proteins were visualized by autoradiography.
Detection of Scrapie Infectivity in Cell Cultures Exposed to Scrapie Agent-Following exposure to brain homogenate, cells were continuously cultured and harvested for inoculation at passage 12 or 13 postinfection. Cells were resuspended in DMEM plus 10% fetal bovine serum. Following five consecutive cycles of freeze/thawing, lysates were sonicated, and ϳ5 ϫ 10 5 cells/mouse were inoculated IC into C57Bl/10 weanling mice. Since the scrapie strain 87V had been propagated in VmDk mice, lysates prepared from cells exposed to 87V brain homogenate were inoculated into VmDk mice. The mouse scrapie strain 87V has an incubation time of Ͼ700 days in C57Bl/10 mice but a shorter incubation time in VmDk mice, which express a mouse PrP allele that differs from the C57Bl/10 allele at positions 108 and 189 (22). As controls, 50 l of 1% brain homogenates from uninfected mice or mice infected with scrapie strain 87V, RML, 22L, or ME7 were also inoculated IC into the appropriate mouse strain. Mice were observed at least twice a week for the onset of neurological disease, and mice showing signs of scrapie disease were sacrificed. All animal experiments were approved by the Rocky Mountain Laboratory Animal Care and Use Committee. The Rocky Mountain Laboratories are fully accredited by the American Association for Laboratory Animal Care.
Determination of the Number of PrP-res-accumulating Cells-To determine the number of cells accumulating PrP-res upon infection with scrapie strain 22L, Mo3F4-⌿2C2 cells were exposed to 1% 22L scrapie brain homogenate in 24-well plates. Cells from three wells were pooled and cloned by limiting dilution after 48, 72, or 96 h of exposure. Additionally, cells were exposed for 96 h to scrapie brain homogenate and then continuously passaged in the absence of brain homogenate. After five and 10 passages, these cells were also cloned. Single cell clones were expanded, and cell lysates of clones were tested for their PrP-res content as described above. Statistical analysis of the resulting data was done using the 2 test.

Acute PrP-res Formation following Exposure of Neuroblastoma and Fibroblast Cells to Different Scrapie Strains-In
order to study the early events that occur during infection with a TSE agent, we have used a cellular assay to detect newly formed, fully glycosylated PrP-res (19). This assay utilizes a cloned mouse neuroblastoma cell line (Mo3F4-MNB) that produces high levels of mouse 3F4-positive PrP-sen when compared with the endogenous mouse PrP-sen expressed in these cells (Fig. 1A). New PrP-res derived from mouse 3F4-positive PrP-sen can be distinguished from the PrP-res present in the infectious inoculum by simple Western blot analysis using the mouse monoclonal antibody 3F4 (18,19).
Upon exposure to different dilutions of RML scrapie brain homogenate for 4 days, Mo3F4-MNB cells produced 3F4-positive PrP-res in a brain homogenate concentration-dependent manner (Fig. 1B). By contrast, no PrP-res was present in cells exposed to normal brain homogenate (Fig. 1B, Mock). Furthermore, 3F4-positive PrP-res was also generated in Mo3F4-MNB cells upon exposure to scrapie strains 22L, ME7, and 87V ( Fig.  1C) (19). These experiments demonstrate that new (i.e. acute) formation of PrP-res in neuroblastoma cells can be driven by multiple scrapie strains.
Whereas the presence of PrP-res in the lymphoreticular tissues of infected mice can vary between scrapie strains, PrP-res is almost always found within the central nervous system, and neuronal vaculation (i.e. spongiform change) is a central hallmark of TSE diseases. As a result, it is often assumed that it is primarily cells within the central nervous system that accumulate PrP-res and are susceptible to scrapie infection. To test whether acute PrP-res formation could also be detected in nonneuronal cells following exposure to different scrapie strains, we used a mixture of ⌿2/PA317 fibroblast cells expressing Mo3F4 PrP-sen (Mo3F4-⌿2/PA317), a cell culture previously shown to be susceptible to 22L scrapie (16). PrP-res formation was detected in cells exposed to RML, 22L, ME7, and 87V scrapie, although differences in the amount of PrP-res produced were observed (Fig. 1D). The in vivo titers and PrPres content of the 22L, ME7, and RML scrapie brain homogenates used were similar (see "Experimental Procedures"), sug-gesting that significant differences in scrapie agent titer or input PrP-res cannot explain the observed differences in the amount of new PrP-res formed. Therefore, acute PrP-res is formed upon exposure to different scrapie strains and can be initiated in cells of both neuronal and nonneuronal origin.
Acute PrP-res Formation Does Not Always Lead to Sustained PrP-res Formation-To test whether acute PrP-res formation led to persistent PrP-res formation, Mo3F4-MNB and Mo3F4-⌿2/PA317 cells were exposed to mouse scrapie infectivity and continuously passaged. In subsequent passages (passages 1-13) of Mo3F4-MNB cells inoculated with any of the scrapie strains, no 3F4-positive PrP-res was detected. Multiple passages of Mo3F4-⌿2/PA317 cells exposed to the strain RML, 87V, or ME7 (Fig. 2, A and B) (data not shown) were also negative for 3F4-positive PrP-res. However, as reported previously (16), Mo3F4-⌿2/PA317 fibroblasts inoculated with 22L brain homogenate accumulated high levels of 3F4-positive PrPres upon continuous passage (Fig. 2B). Thus, whereas multiple scrapie strains could initiate PrP-res formation in both neuronal and nonneuronal cells, detectable PrP-res formation was not sustained in these cultures except for fibroblast cells exposed to the scrapie strain 22L. These data suggest that the requirements for sustained PrP-res formation are both cell type-and scrapie strain-dependent.
Acute PrP-res Formation Does Not Always Lead to Infection-Previous studies have demonstrated that cell cultures that do not accumulate detectable levels of PrP-res upon exposure to TSE agent can still propagate scrapie infectivity (23). To determine whether mouse neuroblastoma cells or mouse fibroblast cells exposed to different scrapie strains generated scrapie infectivity, cell lysates from cells passaged at least 12 times in vitro were inoculated intracranially into either C57Bl/10 (cells exposed to scrapie strain RML, 22L, or ME7) or VmDk mice (cells exposed to scrapie strain 87V) ( Table I). As controls, cells exposed to normal mouse brain homogenate ("None" in Table I) as well as brain homogenates from scrapieinfected or uninfected mice were also inoculated.
Whereas mice inoculated with 22L-infected fibroblast cells succumbed to disease within 284.6 Ϯ 39.6 days, mice inoculated with neuroblastoma cells exposed to the 22L scrapie strain did not show any signs of scrapie up to Ͼ755 days postinfection (Table I). Furthermore, mice inoculated with either neuroblastoma cells or fibroblast cells exposed to ME7, RML, or 87V mouse scrapie did not display any signs of disease. The brains of all nonclinical animals sacrificed up to 755 days postinfection (when the experiment was terminated) were an-alyzed for PrP-res. All were negative except for one mouse inoculated with fibroblast cells exposed to RML scrapie that accumulated PrP-res in the brain 624 days postinoculation (Table I). This suggests that a very low amount of RML infectivity (Ͻ1 ID 50 /5 ϫ 10 5 cells) was present in the cell culture. Overall, these experiments demonstrate that whereas mouse neuroblastoma cells and mouse fibroblasts were capable of forming detectable PrP-res upon acute exposure to scrapie brain homogenate, this did not lead to persistent infection in most cases. These data suggest that the cellular requirements for acute PrP-res formation may differ from those for sustained PrP-res formation and persistent scrapie infection.
Acute PrP-res Formation Is Cell-associated and Requires Intact Cells-The foregoing results demonstrated that acute PrPres formation could be initiated in two different cell types by multiple scrapie strains, but persistent infection did not necessarily occur. One possible explanation for these results is that de novo PrP-res formation was not cell-associated. PrP-res formation can occur in vitro without the context of a living cell (24,25), and a recent study suggested that removal of PrP-sen from the cell membrane facilitates new PrP-res formation (26). Thus, it was possible that newly formed PrP-res was being generated solely from the conversion of secreted or non-cellassociated Mo3F4 PrP-sen present in the cell medium. To confirm that the newly formed PrP-res detected was cell-associated, Mo3F4-MNB culture supernatant was harvested after  Mo3F4-MNB cells were exposed for 4 days to 1% normal mouse brain homogenate (Mock) or different dilutions of brain homogenate derived from a mouse infected with RML scrapie. Newly formed PrP-res was detected by Western blot using the mouse monoclonal antibody 3F4. C, Mo3F4-MNB cells produce PrP-res when exposed to the mouse-adapted scrapie strains RML, 22L, ME7, and 87V. Newly formed PrP-res was assayed by Western blot using the monoclonal antibody 3F4. D, PrP-res formation in a mixed culture of Mo3F4-⌿2/PA317 cells. Cells were exposed to scrapie brain homogenates, and cell lysates were checked by Western blot for newly formed PrP-res using the mouse monoclonal antibody 3F4. For all panels, the bottom arrow represents unglycosylated PrP, the top arrow indicates the low molecular mass PrP glycoform, and the bracket defines the high molecular mass PrP glycoform. Molecular mass markers in kilodaltons are indicated on the left. 96 h of incubation with brain homogenate from a mouse infected with the mouse scrapie strain ME7, and the supernatant as well as the cell lysate was checked for 3F4-positive PrP-res. Newly formed PrP-res was detectable in the cell lysate but not in the cell culture supernatant (Fig. 3A), demonstrating that the de novo PrP-res produced in this assay was cell-associated.
To study whether PrP-res formation was dependent upon intact cells, Mo3F4-MNB cells were sonicated to rupture the cell membranes and mixed with brain homogenate in cell culture medium supplemented with protease inhibitors. No 3F4positive PrP-res could be detected in samples containing either normal brain homogenate or brain homogenate from a scrapieinfected animal (Fig. 3B). These experiments are in agreement with previously published data, demonstrating that in vitro PrP-res formation driven by crude TSE brain homogenates is dependent on the presence of detergent (19,25). These results show that newly formed PrP-res is cell-associated and that its formation is dependent upon intact cells. Thus, the inability to establish a persistent infection in these cells was not due to the lack of cell-associated PrP-res formation.
Acute PrP-res Formation Is a Rapid Process-The fact that de novo PrP-res formation was easily detectable just 96 h after exposure of uninfected cells to scrapie brain homogenate suggested that PrP-res formation was occurring rapidly in these cells. To determine how rapidly PrP-res was generated, Mo3F4-MNB cells were exposed for different time periods to scrapie strains known to replicate with either long (87V) or short (RML) incubation periods in C57Bl/10 mice (27). Maximum levels of 3F4-positive PrP-res were reached after 24 h of inoculation with either RML or 87V scrapie brain homogenates (Fig. 4). Interestingly, de novo formation of PrP-res was rapid regardless of whether the PrP-res in the infected brain homogenate was derived from a short or long incubation time strain. These data are consistent with recent data using hamster scrapie demonstrating that, regardless of the strain incubation time, the same amount of PrP-res is formed after 24 h (28). Thus, our results demonstrate that the acute formation of PrP-res following exposure to a scrapie agent is rapid regardless of whether or not the PrP-res is derived from a scrapie strain with a slow or long incubation time in vivo.

Acute and Sustained PrP-res Formation Differ Kinetically-
The fact that the initiation of PrP-res formation was sustained in fibroblasts but not in neuroblastoma cells exposed to 22L brain homogenate suggested that the requirements for acute and persistent PrP-res formation may be different in these cell lines. To more closely study the kinetics of PrP-res formation, both mouse neuroblastoma and fibroblast cells were incubated with 22L scrapie brain homogenate for 4 -96 h (Fig. 5). For this particular experiment, a subclone of ⌿2 cells expressing Mo3F4 PrP-sen (Mo3F4-⌿2C2) was used that also supported sustained PrP-res formation (16). Exposure of neuroblastoma cells to 22L brain homogenate led to maximal PrP-res levels within 24 h that did not increase between 24 and 96 h (Fig. 5A). By contrast, in fibroblast cells exposed to the scrapie strain 22L, PrP-res formation drastically increased within the first 96 h postexposure (Fig. 5B). Although the amount of PrP-sen at each time point could be somewhat variable (Fig. 5), in multiple experiments, this variability did not affect the dramatic difference in the kinetics of PrP-res formation between the two different cell types. The fact that two different cell types exposed to the same infectious brain homogenate formed PrP-res at different rates suggests that specific cellular factors can contribute to persistent PrP-res formation.

Continuous Passage of 22L-infected Cells Does Not Lead to an Increase in PrP-res-accumulating Cells-One possible explanation for the time-dependent increase in the amount of
PrP-res in fibroblast cells exposed to 22L was that infectivity was spreading to more cells in the culture. Alternatively, the increase in PrP-res formation could be due to an increase in the amount of PrP-res formed per cell and not to an increase in the number of infected cells. In order to distinguish between these two possibilities, individual Mo3F4-⌿2C2 cell clones were isolated at different time points following exposure to 22L-infected mouse brain homogenates. Each clone was then assayed for 3F4-positive PrP-res by Western blot. The percentage of PrP-res positive clones did not significantly change from 48 to 96 h postinfection (Table II), although the amount of PrP-res formed did (Fig. 5B). Furthermore, the percentage of PrP-respositive cells did not significantly change even after 10 passages in vitro (ϳ30 days), making it unlikely that spontaneous loss of PrP-res-positive cells during the cloning procedure unduly influenced the results. These data are consistent with the TABLE I Inoculation of C57BL/10 mice and VmDk mice with cells exposed to different scrapie strains Cells were exposed to brain homogenates of scrapie-infected or uninfected mice and passaged 12-13 times. Lysates of approximately 5 ϫ 10 5 cells were inoculated intracranially (IC) into C57Bl/10. VmDk mice express the PrP gene necessary for long incubation time strains such as 87V and thus were used to assay the 87V strain instead of C57Bl/10 mice. Mice that showed signs of TSE disease were sacrificed at the days post-infection indicated (Incubation time). Mice were inoculated IC with 50 l of a 1% (w/v) brain homogenate in DMEM plus 10% fetal bovine serum. hypothesis that the increase in the amount of PrP-res produced over the first 96 h following infection is not solely dependent upon more cells becoming infected but is more likely related to an increase in the amount of PrP-res made per cell.
Emergence of Stable PrP-res Glycoforms in 22L-infected Cells-Since the overall level of PrP-res increased while the percentage of PrP-res positive cells remained relatively constant, the above results demonstrated that the efficiency of PrP-res formation increased over the first 96 h following infection of Mo3F4-⌿2C2 cells with 22L scrapie. One possible explanation for this observation is that the 22L scrapie agent may have "adapted" to the fibroblast cell environment. If this had occurred, then not only the quantity but also the quality of the PrP-res being formed could have changed as well.
In order to determine whether or not the quality of PrP-res in the fibroblast cells was also changing during the first 96 h following infection with 22L, we analyzed the glycoform ratio of the newly formed PrP-res. PrP-res glycoform ratios are a characteristic commonly used to define strains of TSE agent (29) and are generally determined by comparing the amount of the low molecular mass PrP-res glycoform (middle band in Fig. 5B) with the amount of the high molecular mass PrP-res glycoform (upper band in Fig. 5B). Mo3F4-⌿2C2 cells were exposed to a 22L-infected mouse brain homogenate, 3F4 epitope-tagged PrP-res was isolated at various time points postexposure, and the glycoform ratio was determined. As shown in Fig. 6A, within the first 96 h following an acute exposure to 22L scrapie, the glycoform ratio varied significantly, and no dominant glycoform ratio was apparent. By contrast, in cells that persistently produced PrP-res, there was little variability in the glycoform pattern from one passage to the next (Fig. 6B). Fur-thermore, it appeared that a single, dominant glycoform pattern had emerged that more closely resembled the PrP-res present in the 22L brain homogenate used to infect the cells.
It was possible that the dramatic difference observed in the glycoform pattern of PrP-res made within the first 96 h following infection versus that made in persistently infected cells was due to normal variations that might occur as the cells are actively dividing following the first few days after passage. In order to address this issue, we examined the glycoform ratio of PrP-res in persistently infected Mo3F4-⌿2C2 cells 24, 48, 72, and 96 h after passage. There was little variability in the PrP-res glycoform ratio at the four different time points (Fig.  6C), demonstrating that the conditions in newly passaged, rapidly dividing cells did not influence the PrP-res glycosylation pattern. The data demonstrated that, unlike cells that were in the acute stages of infection, one PrP-res glycoform pattern was dominant in cells that were persistently infected with mouse scrapie (Fig. 6, compare A with C). Thus, both   FIG. 3. Newly formed PrP-res is associated with the cell. A, Mo3F4-MNB cells were exposed to uninfected (Mock) or ME7 scrapie-infected brain homogenates. After 96 h, the cell medium and cell lysate (T), cell medium alone (SN), or the cell lysate alone (CL) were checked for newly formed, 3F4 epitope-tagged PrP-res in the presence or absence of PK. B, Mo3F4-MNB cells were ruptured by sonication and mixed with brain homogenate from either an uninfected mouse (Mock) or from a mouse infected with the scrapie strain ME7. PK-treated brain homogenates are shown on the left. This Western blot was probed with the rabbit polyclonal antiserum R30. After 96 h, samples were tested for 3F4-positive PrP-res in the presence or absence of PK (right). A, the bottom arrow represents unglycosylated PrP-res, the top arrow indicates the low molecular mass PrP-res glycoform, and the bracket defines the high molecular mass PrP-res glycoform. The asterisk indicates a cross-reactive non-PrP band that is associated with PK treatment. In both panels, molecular mass markers in kilodaltons are indicated on the left.  quantitative and qualitative changes in PrP-res production suggest that the 22L mouse scrapie agent undergoes a process of adaption that eventually leads to the persistent infection of mouse fibroblast cells.

DISCUSSION
PrP-sen is an essential susceptibility factor for productive TSE infection and is absolutely required for PrP-res formation. However, despite the nearly ubiquitous distribution of PrP-sen in mammalian tissues, PrP-res accumulation appears to be restricted almost exclusively to cells in the nervous and lymphoreticular systems. Here we demonstrate for the first time that a cell can make PrP-res whether or not it can support persistent TSE infection. These data may provide, at least in part, an explanation for the observation that the majority of PrP-sen expressing cell types do not appear to support TSE agent replication. Our data suggest that persistent TSE infection is at least a two-step process involving acute formation of PrP-res that may or may not be followed by a second phase of persistent PrP-res formation, which is indicative of a productive TSE infection.
At a cellular level, the first, acute stage of TSE infection would encompass the first few days following exposure to TSE infectivity and would involve the interaction of the PrP-sen in the target cell with the PrP-res in the infectious inoculum. During this very early stage, PrP-res would be formed rapidly, leading to detectable and, in some cases maximal, production of new PrP-res by 24 h postinfection (Fig. 4). Acute PrP-res formation does not appear to be dependent upon the strain of TSE agent or the cell type (Fig. 1), suggesting that strain-or cellspecific factors are not essential requirements for the initiation of PrP-res formation. As a result, events during this first stage of infection might occur in any cell expressing PrP-sen that also has contact with PrP-res.
Although our results clearly demonstrate that PrP-res can be formed almost immediately following exposure to an infectious scrapie inoculum, it is unclear why in most cases PrP-res formation does not persist. One possibility is that acute PrP-res formation kills the cell. In this scenario, loss of susceptible cells in the culture would account for the loss of the PrP-res signal. However, no overt cytotoxic effect was observed in cell cultures where PrP-res was lost (data not shown). Additionally, the fact that exposure of Mo3F4-⌿2/PA317 cells to 22L scrapie brain homogenate led to a sustained infection whereas exposure of Mo3F4-MNB cells to the same brain homogenate resulted in transient PrP-res formation argues against this possibility. Thus, an alternative explanation for the loss of PrP-res is that while the cells were capable of producing PrP-res upon acute exposure to scrapie agent, lack of cell-and/or strain-specific factors prevented them from becoming persistently infected.
The ability of a particular cell type to support TSE infection would therefore have to depend upon events that occurred following the initial round of PrP-res formation. In our experiments, only fibroblast cells exposed to 22L mouse scrapie persistently produced PrP-res and replicated TSE infectivity an increase in the number of PrP-res-producing cells Mouse fibroblast cell clone Mo3F4-⌿2C2 was exposed to 22L scrapie brain homogenate and cloned by limiting dilution 48, 72, and 96 h postexposure to scrapie agent. For passage 5 and 10 (P5 and P10), cells were incubated with 22L scrapie brain homogenate for 96 h and then subsequently passaged before cloning. Single cell clones were expanded and tested for the accumulation of 3F4-positive PrP-res. Shown is the number of PrP-res-producing cells to cells that did not accumulate detectable levels of PrP-res. Comparison of the data using the 2 square contingency test demonstrated that there was no significant difference in the number of positive cells at the different time points. FIG. 6. Increased variability in the PrP-res glycoform ratio in acute versus persistently infected Mo3F4-⌿2C2 cells. Cells were exposed to a 22L mouse scrapie-infected brain homogenate, and newly formed, 3F4 epitope-tagged PrP-res was analyzed by Western blot using the mouse monoclonal antibody 3F4. The glycoform ratio was determined by comparing the amount of PrP-res present in the low molecular mass PrP glycoform with that present in the high molecular mass PrP glycoform as described previously (19,29). For comparison, the glycoform data (n ϭ 11) for PrP-res isolated from a 22L-infected mouse brain homogenate is shown. These data have been reported previously (19). Each symbol represents the glycoform ratio for one independent sample, although, due to overlap in some data points, all of the symbols may not be visible. A, PrP-res glycoform ratio in Mo3F4-⌿2C2 cells during acute exposure to 22L scrapie (n ϭ 18 for each time point). B, PrP-res glycoform ratio in Mo3F4-⌿2C2 cells persistently infected with 22L scrapie. For passages 1 and 2, n ϭ 14; for passages 3 and 4, n ϭ 3. C, PrP-res glycoform ratio in Mo3F4-⌿2C2 cells persistently infected with 22L scrapie at 24, 48, 72, and 96 h after splitting for passage 11 (n ϭ 6 for each time point). (Table I) (16). Persistent scrapie infection of these cells was accompanied by both quantitative (Fig. 5) and qualitative (Fig.  6) changes in PrP-res. These data strongly suggest that it is the second stage of the infection process, where events occur that enable PrP-res formation to become persistent, that determines whether or not a particular cell type becomes infected.
The argument that cell-specific factors may influence susceptibility to TSE infection is bolstered by our observation that Mo3F4-MNB cells were refractory to infection. Mouse neuroblastoma cells expressing either wild-type mouse PrP-sen or Mo3F4 PrP-sen have long been known to be susceptible to mouse scrapie infectivity, including the mouse scrapie strains 22L and RML utilized here (23,30,31). 2 However, the Mo3F4-MNB cells used in the current experiments are cloned cells that, because of the cloning procedure, had been passaged for longer than 30 days. We have found that whereas early passage MNB cells are fully susceptible to persistent mouse scrapie infection, MNB cells that have been in passage longer than 30 days can become resistant to infection (32). 2 Thus, it is possible that the resistance of mouse neuroblastoma cells is due to the lack of specific cellular factors that are required for the second stage of PrP-res formation and sustained scrapie infection. However, other than PrP-sen expression, no cell-specific factor or factors have ever been identified that are always associated with TSE infection. The nature of these potential scrapie susceptibility factors therefore remains unclear.
Our results suggest that the scrapie strain may also influence whether or not a cell becomes persistently infected. Such strain-specific factors could include an as yet unidentified scrapie-specific nucleic acid as well as other protein and nonprotein components that may be closely associated with PrPres (33)(34)(35)(36)(37)(38). Strain specificity might also be encoded by the conformation of the PrP-res molecule (39 -41). We and others have recently provided evidence that variations in PrP-res phenotypes may be dependent upon the cellular compartment where PrP-res formation is occurring and that this compartment may vary with the strain of TSE agent (19,42). Thus, another possibility is that persistent TSE infection is dependent upon the ability of a specific scrapie agent to localize and/or adapt to a cellular compartment where sustainable formation of PrP-res can occur.
Our data demonstrating that 1) the PrP-res glycoform ratio is far more variable during the acute stage of TSE infection than during persistent infection (Fig. 6), 2) the amount of PrP-res per cell begins to significantly increase over time (Fig.  5), and 3) infectivity is not found in cells where PrP-res is only made acutely (Table I) are all consistent with this hypothesis. The PrP-res glycoform ratio would be variable during the acute stage of infection (Fig. 6) (19) because PrP-res formation is occurring at multiple sites throughout the cell. For a persistent infection to occur, PrP-res formation would become restricted to those cellular environments that could support efficient PrPres formation as well as sustain it across multiple cell passages.
It has been reported recently that TSE infectivity can spread through infected cell cultures either by direct cell-to-cell contact (43) or via the release of infectivity into the culture medium (44). Whereas a direct correlation between cell-to-cell spread of TSE infectivity and persistence of infection remains to be established, it is possible that impairment of such a mechanism on a cellular level could account for the resistance of some cells to persistent infection. Thus, spread of infectivity in cell culture could also be dependent upon both cell-mediated events and scrapie strain-dependent traits. For example, if the TSE agent is unable to reach an appropriate site, PrP-res would not be produced in a manner that is both sustainable and transferable, and infectivity would be lost. The fact that PrPres formation in most cell lines was only observed upon direct exposure to the scrapie agent but then declined after its removal and subsequent passage of the cells (Figs. 1 and 2) suggests that acute PrP-res formation was driven mainly by the PrP-res present in the inoculum. By contrast, our experiments demonstrating both quantitative and qualitative changes in acute versus persistent PrP-res formation strongly suggest that the TSE agent may have localized to a cellular environment compatible with the establishment of a productive, persistent TSE infection that is dependent upon PrP-res generated within the cell.