OUP user menu

Increased Proportions of C1 Truncated Prion Protein Protect Against Cellular M1000 Prion Infection

Victoria Lewis PhD, Andrew F. Hill PhD, Cathryn L. Haigh PhD, Genevieve M. Klug PGradDip(Epi.Biostats), Colin L. Masters MD, Victoria A. Lawson PhD, Steven J. Collins MD
DOI: http://dx.doi.org/10.1097/NEN.0b013e3181b96981 1125-1135 First published online: 1 October 2009

Abstract

Prion disease pathogenesis is linked to the cell-associated propagation of misfolded protease-resistant conformers (PrPres) of the normal cellular prion protein (PrPC). Ongoing PrPC expression is the only known absolute requirement for successful prion disease transmission and PrPres propagation. Further typifying prion disease is selective neuronal dysfunction and loss, although the precise mechanisms underlying this are undefined. We utilized a single prion strain (M1000) and a range of neuronal and nonneuronal, PrPC endogenously expressing and transgenically modified overexpressing cell lines, to evaluate whether PrPC glycosylation patterns or constitutive N-terminal cleavage events may be determinants of sustained PrPres propagation. Our data demonstrates that relative proportions offull-length and C1 truncated PrPC are the most important characteristics influencing susceptibility to sustained M1000 prion infection, supporting PrPC α-cleavage as a protective event, which may contribute to the selective neuronal vulnerability observed in vivo.

Key Words
  • Endoproteolytic cleavage
  • N-linked glycosylation
  • Prion infection
  • Prion protein
  • PrPC
  • PrPres

Introduction

Prion diseases are fatal neurodegenerative diseases of humans and animals that are characterized by the accumulation of misfolded protease-resistant conformers (PrPres) of the normal cellular prion protein (PrPC) (1). PrPC is a ubiquitously expressed glycoprotein (2), and like many structural and functional classes of membrane-bound proteins (3), it undergoes posttranslational proteolytic cleavage to generate distinct species. There are 2 predominant truncated PrPC species, C1 and C2. PrPC α-cleavage producing C1 occurs at Residues 111 or 112 (human PrPC sequence nomenclature) (4); significantly, this is within a domain shown to be neurotoxic and amyloidogenic in in vitro studies of a synthetic peptide fragment (5). PrPC α-cleavage is thought to be carried out at least in part by a-disintegrin-and-metalloprotease 10 (ADAM-10) and ADAM-17 (also known as tumor necrosis factor-α-converting enzyme) (6). C2 is produced from β-cleavage and is found at low levels in normal human (4, 7, 8) and cultured cells (9). β-Cleavage occurs around the C-terminus of the PrPC octapeptide repeat domain (8, 9) and is mediated by reactive oxygen species (10). The functional significance of endogenous PrPC cleavage is unresolved.

A defining feature of prion diseases is their transmissibility; this reached epidemic proportions in the United Kingdom during the outbreak of bovine spongiform encephalopathy (11). Ongoing expression of PrPC is an absolute requirement for the successful transmission and continuing pathogenesis of prion diseases (12, 13). In animal species such as rabbits and dogs, the precise molecular determinants of their apparently very high resistance to natural and experimental prion disease are not clearly defined (14, 15), although PrPC primary sequence incompatibilities have been suggested to provide protection for rabbits (16). Importantly, rabbit kidney epithelial cells that have been engineered to express exogenous PrPC can be infected with cognate prion strains derived from the same foreign species, and thereafter propagate PrPres and infectivity (17). These data highlight that the innate resistance of rabbits is linked specifically to endogenous rabbit PrPC, as opposed to the cellular machinery required for infection or propagation.

Neuropathologic features of prion diseases include vacuolation of the neuropil, astrocytic gliosis, neuronal loss, and extracellular PrPres deposits. For a single prion strain, the pattern of neuronal involvement and topographic distribution of these abnormalities can vary among host species and importantly, across different inbred host strains of the same animal species harboring the same PrP gene (Prnp) allele (18). These observations suggest nuances in PrPC biology or intrinsic cellular factors unrelated to PrPC, that differ across neuronal populations in a manner peculiar to each host, significantly impact pathogenesis. Such phenomena exemplify selective neuronal vulnerability and focus the broader concept of innate prion disease resistance to the cellular level.

Cell lines can be persistently infected with prions and retain the biologic and biochemical features of the strain (19, 20). Analogous to observations in animal models, individual cell lines have variable susceptibilities to prion infection (Table 1). In view of the absolute requirement for PrPC in prion disease transmission and pathogenesis, cellular PrPC expression profiles may influence susceptibility to sustained prion infection. Here, we used cell lines, endogenously or transgenically expressing the same murine Prnp allele, to correlate PrPC glycosylation patterns and constitutive endoproteolyic cleavage profiles with inherent susceptibility or resistance to a well-characterized mouse-adapted human prion strain (M1000). We found that expression of proportionately higher levels of full-length PrPC correlated with susceptibility to sustained M1000 prion infection, while relatively higher levels of the C1 C-terminal fragment of PrPC were associated with resistance to sustained M1000 prion infection. Overall, these results suggest that α-cleavage of PrPC is protective against stable PrPres propagation.

View this table:
TABLE 1.

Materials and Methods

Cell Lines and Culture Conditions

The following murine cell lines expressed endogenous PrPC from the Prnpa allele (data not shown): 2 cognate hypothalamic GT1-7 cell lines (21) obtained from the same laboratory approximately 2 years apart, designated GT1-7C and GT1-7H; uncloned N2a neuroblastoma cells (American Type Culture Collection [ATCC] CCL-131); NIH/3T3 fibroblast cells (ATCC CRL-1658); and OBL-21 olfactory bulb cells (22). MoRK13 cells, rabbit kidney-derived epithelial cells (RK13; ATCC CCL-37) stably transfected with the murine Prnpa coding sequence as previously described (23), were also used. Cells were maintained in either Dulbecco modified Eagle medium (OBL-21, NIH/3T3, MoRK13) or OptiMEM (N2a, GT1-7) (Invitrogen, Carlsbad, CA) containing 10% (vol/vol) heat-inactivated fetal bovine serum (Thermo Fisher Scientific, Rockford, IL) and 1% (vol/vol) penicillin-streptomycin (Invitrogen), in a humidified incubator, 37°C, 5% carbon dioxide. The medium of MoRK13 cells was supplemented with a final concentration of 2.5 μg/mL puromycin dihydrochloride (Sigma-Aldrich, St Louis, MO).

Prion Strain and Cellular Prion Infection

M1000 prions from mouse brain homogenate (10% wt/vol, prepared in sterile PBS; 140 mmol/L NaCl; 2.7 mmol/L KCl; 1.8 mmol/L KH2PO4; 10 mmol/L Na2HPO4), mean lethal dose (LD50) of approximately 109 LD50 per gram of tissue, as determined by end point titration in Tga20 mice (24), were used for all infections. To infect cells, subconfluent monolayers were overlaid with M1000 brain homogenate (or uninfected BALB/c brain homogenate as a negative control) diluted in media to between 0.1% and 2% (wt/vol) final dilution. After overnight incubation, the homogenates were removed, cells were then washed twice with sterile Dulbecco PBS (Invitrogen), and fresh medium was added. Cells were grown to confluence and subcultured routinely, with the first passage post-prion exposure designated Passage 1 (P1).

Cell Blot Detection of PrPres

A sensitive cell blot technique was used to detect PrPres in M1000-infected cell cultures, as described elsewhere (23, 25), with minor modifications. Proteinase K ([PK] Invitrogen) digestion was at a final concentration of 10 μg/mL in lysis buffer (50 mmol/L Tris-Cl pH 7.4, 150 mmol/L NaCl, 0.5% sodium deoxycholate, 0.5% Triton X-100) for 90 minutes at 37°C and was stopped by incubating for 15 minutes in 2 mmol/L phenylmethylsulfonyl fluoride (Roche Applied Science, Penzberg, Germany). Phosphate-buffered saline Tween-20 ([PBST] 140 mmol/L NaCl, 2.7 mmol/L KCl, 1.8 mmol/L KH2PO4, 10 mmol/L Na2HPO4, 0.05% Tween-20 in H2O) was used to make the block buffer (5% [wt/vol] nonfat milk in PBST) and for all washes. A 1:10,000 dilution in block buffer of both the anti-prion protein primary antibody ICSM18 (overnight at 4°C; D-Gen, London, UK) and anti-mouse-horseradish peroxidase secondary antibody (1 hour at room temperature; GE Healthcare, Buckinghamshire, UK) was used for all cell blots.

Detection of M1000 PrPres

To detect PrPres in M1000 brain homogenate, samples were digested with 50 μg/mL PK for 1 hour at 37°C. The PK was stopped by the addition of Pefabloc SC (Roche Applied Science) to a final concentration of 4 mmol/L.

Cell Lysate Preparation

Confluent cell cultures were harvested, lysed, and assessed for total protein content using the Pierce BCA protein assay (Thermo Fisher Scientific), as previously described (26). Cell lysates were used directly or treated, as described later, before sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting.

Deglycosylation of PrPC

The glycosidase PNGaseF (Roche Applied Science) was used to remove N-linked glycans from PrPC in cell lysates, as previously described (26).

Immunoprecipitation

Zymed rec-Protein G-Sepharose 4B conjugate (PGS) beads (Invitrogen) were equilibrated by washing with PBST. Cell lysates were precleared by diluting into 1 mL PBST and then incubating with 40 μL of prewashed beads for 1 hour with gentle mixing. Beads were collected by centrifugation (2 minutes, 336 × g). The supernatant was transferred to fresh tubes containing 1 μL SAF32 (Cayman Chemical, Ann Arbor, MI) and again mixed gently for 1 hour before the addition of 40 μL of fresh equilibrated beads and overnight incubation at 4°C with gentle mixing. Beads were collected as before, washed 3 times in PBST, and then used for PAGE and Western blot analysis.

PAGE and Western Immunoblotting

Untreated, PNGaseF-treated, or PK-treated samples were mixed with an appropriate volume of 4× SDS-PAGE sample buffer (170 mmol/L Tris-Cl pH 6.8, 0.5 mol/L glycine, 8% [wt/vol] SDS, 0.02% [wt/vol] bromophenol blue, 12% [vol/vol] β-mercaptoethanol), boiled (100°C) for 10 minutes, resolved by tris-glycine SDS-PAGE (percent self-made gel [26] as indicated in figure legends), and Western blotted as previously described (26). Immunoprecipitated samples on PGS beads were mixed with NuPAGE loading dye (Invitrogen), boiled (100°C) for 10 minutes, resolved on 12% Bis-tris NuPAGE gels (Invitrogen) using MES running buffer (Invitrogen), and subjected to Western blotting. The anti-PrP antibodies used were monoclonal antibodies ICSM18 (1:40,000), SAF32 (1:3000) and 8B4 (1:10,000; Alicon, Schlieren, Switzerland), and polyclonal antibody 03R19 (24) (1:5000). Anti-β-tubulin primary antibody (1:15,000; Sigma-Aldrich) and the appropriate horseradish peroxidase-conjugated secondary antibodies (i.e. anti-mouse or anti-rabbit; 1:10,000; GE Healthcare) were also used. Detection was via enhanced chemiluminescence (ECL Plus Western Blotting Detection Reagents; GE Healthcare) using Kodak Biomax film (Sigma-Aldrich) or digital capture using a Fujifilm LAS-3000 Intelligent Dark Box or Syngene GeneGnome Bioimager.

Densitometric and Statistical Analyses

Digital and scanned x-ray film images were quantitatively assessed for signal intensity using ImageJ software. Statistical analyses were carried out using GraphPad Prism v4.0a or Minitab 15 as required and as specified in the Results section.

Results

Determination of Susceptibility to M1000 Prions

The abilities of each cell line to sustain M1000 prion infection and propagate PrPres over multiple passages after exposure to M1000 brain homogenate were assessed. Control cells were either left uninfected or exposed to the equivalent dilution of normal BALB/c brain homogenate. Cells were tested a minimum of 3 passages post-prion exposure for de novo PrPres production to indicate successful establishment of M1000 infection. As previously reported (23), MoRK13 cells were highly susceptible to M1000 prions, with increasing propagation of prions over time as seen by the increase in PrPres signal with sequential passages (Fig. 1A). The GT1-7H cells and the OBL-21 cells (Figs. 1B, C, respectively) were also susceptible to M1000 prions, but they produced lower levels of PrPres compared with MoRK13 cells. In both of these cell lines after an initial slight increase in PrPres production, a loss of PrPres propagation was observed. Importantly, this is the first report PrPres propagation in an olfactory bulb cell line. PrPres was never detected in the other cell lines (Figs. 1D-F). Table 2 summarizes the cellular susceptibility or resistance to M1000.

FIGURE 1.

Representative cell blots of MoRK13 (A), GT1-7H (B), OBL-21 (C), GT1-7C (D), N2a (E), and NIH/3T3 (F) exposed to the specified inoculum (M1000, Balb/c normal brain homogenate (NBH)), or left uninfected (UN). The 2 GT1-7 cell lines were exposed to lower concentrations of inocula (that is, 1% homogenate), as higher percentage homogenates were toxic to these cells. The cells were seeded onto coverslips and grown to 90% to 100% confluence before cell blotting for PrPres as an indicator of successful M1000 prion infection. Passage numbers (post prion infection) are indicated.

View this table:
TABLE 2.

Cellular PrPC Glycosylation Profiles Do Not Correlate With Susceptibility to M1000 Prion Infection

The prion protein has 2 conserved potential N-linked glycosylation sites (27). None, one, or both of these sites can be occupied by a polysaccharide chain of variable length, which is visualized as 3 different band sizes (i.e. unglycosylated, monoglycosylated, and diglycosylated PrPC) on a Western blot. As shown in Figures 2A and B, the cell lines display markedly different PrPC glycosylation profiles; this was highlighted when these patterns were quantified (Fig. 2D). Densitometric measurements were obtained by determining the intensity of the bands at the predicted molecular weights of diglycosylated, monoglycosylated, and unglycosylated PrPC, and expressing each relative to the total PrP signal intensity. Undefined truncated PrPC species were not included for analyses. Unexpectedly, the 2 GT1-7 cell lines had noticeably different PrPC expression profiles (Fig. 2B). Validation of the authenticity of the common origin of GT1-7C and GT1-7H cells (i.e. by assessing expression of SV40-T antigen and finding no glial fibrillary acidic protein expression) showed no differences between the lines, although they had slightly different morphologies; the GT1-7H cells were always stellate in appearance and predominantly grew in clusters, whereas the GT1-7C cells were fusiform single cells but became stellate when they approached confluence (not shown). Differences observed in these originally identical cells are most likely caused by clonal drift, possibly explained by the intrinsic or extrinsic models of spontaneous generation of phenotypic heterogeneity (28).

FIGURE 2.

Prion protein (PrP) expression profiles and glycoform ratios in uninfected cells (PrPC) and M1000 brain homogenate (PrPres). Representative Western blots (WB) of PrPC expression profiles in cell lines (A) and specifically comparing the GT1-7C and GT1-7H cell lines (B); 50 μg total protein per lane, 12% Tris-glycine sodium dodecyl sulfate-polyacrylamide gel electrophoresis, WB with ICSM18, and β-tubulin primary antibodies. (C) Representative WB of M1000 brain homogenate pre (−) and post (+) proteinase K digest. (D) Quantification (mean ± SEM) of glycoform ratios of PrPC in the different cell lines and of M1000 PrPres. Numbers of replicates for each individual cell line or homogenate (n) are indicated on x axis labels. In all cases, based on apparent molecular weight, di, diglycosylated PrPC or PrPres; mono, monoglycosylated PrPC or PrPres; un, unglycosylated PrPC or PrPres; truncated, undefined endogenously cleaved PrPC species. (E) Correlative analysis of PrPC glycosylation profiles with susceptibility to M1000 prion infection. Data points represent the replicates obtained from WB quantification of diglycosylated, monoglycosylated, and unglycosylated PrPC as shown in (D). Susceptible, evidence of PrPres propagation in cell line after M1000 exposure; resistant, no evidence of de novo PrPres production after M1000 exposure; susceptible*, susceptible cell lines excluding MoRK13 from analysis. Binary logistic regression (Minitab 15) was performed; no significant differences were seen.

A previous study indicated that glycosylation of PrPC might be protective by inhibiting the propensity of PrPC to misfold and form β-sheet conformations (29). To correlate M1000 susceptibility with glycosylation pattern, the cell lines were grouped by their susceptibility or resistance (Table 2) and the proportion of each PrPC glycoform (Fig. 2D) was separately plotted against susceptibility or resistance (Fig. 2E). To assess the effect of glycosylation in an entirely endogenous expression system, the susceptible group was also separated to exclude MoRK13 PrPC overexpressing cells from the analysis. PrPC glycosylation profiles were not a significant predictor of susceptibility or resistance to M1000 infection.

There is evidence suggesting that similarity between host PrPC glycosylation pattern and prion strain PrPres glycosylation pattern can increase the likelihood of PrPC to PrPres conversion (30). Therefore, the glycosylation pattern of M1000 PrPres was quantified (Figs. 2C, D) and compared with each cell line PrPC glycosylation profile (Table 3). Only the GT1-7H and NIH/3T3 cell lines had PrPC glycosylation profiles that were not significantly different from M1000 PrPres. Although the GT1-7H cells were susceptible to M1000 infection, the NIH/3T3 cells were not; this argues against similarities of PrPC and PrPres glycoform profiles having a dominant influence on susceptibility to infection with this prion strain.

View this table:
TABLE 3.

One limitation of the glycosylation analyses is that the Western blot PrPC glycosylation profiles were obtained using the ICSM18 antibody; the epitope for this antibody is within the C-terminus of PrP, allowing detection of N-terminally truncated PrP species as well as full-length PrP. Therefore, bands seen at the presumed relative molecular weights of diglycosylated, monoglycosylated, and unglycosylated full-length PrPC may actually be a composite of the different glycosylated full-length and truncated species. To accurately visualize relative levels of glycosylated full-length PrPC and assess the influence of these levels on susceptibility to M1000 infection, N-terminal antibodies were used to immunoprecipitate and then detect by Western blot exclusively full-length PrPC species from cell lysates. The glycosylation profiles of full-length PrPC species (Fig. 3A) are generally different from the profiles in Figure 2A, thereby further indicating that the bands in Figure 2A are most likely a composite of glycosylated full-length and truncated PrPC. When comparing the different cell lines, the apparent sizes of the different glycosylated full-length PrPC species are diverse; however when the proportions of each PrPC species are quantified and graphed, they were remarkably similar (Fig. 3B). We have therefore shown that full-length PrPC glycosylation profiles are not a significant predictor of susceptibility or resistance to M1000 infection (Fig. 3C). Furthermore, when comparing the cell line PrPC glycosylation profiles in Figure 3 to the M1000 PrPres profile (Figs. 2C, D), all of the cell lines (both susceptible and resistant) had different proportions of all their PrPC glycoforms to M1000 PrPres, with the exception of GT1-7H cells for which the relative proportion of unglycosylated PrPC was not significantly different from the relative proportion of unglycosylated M1000 PrPres (not shown). These data further argue against a predominant influence of PrPC glycosylation in determining susceptibility to M1000 prion infection.

FIGURE 3.

Full-length cellular prion protein (PrPC) glycoform profiles in uninfected cells. Representative Western blot with N-terminal antibody 8B4 after immunoprecipitation with N-terminal antibody SAF32 (A) and quantification (mean ± SEM, n = 3) of glycoform ratios of PrPC (B). In all cases, based on apparent molecular weight, di, diglycosylated PrPC or PrPres; mono, monoglycosylated PrPC or PrPres; un, unglycosylated PrPC or PrPres. IgG heavy chain (IgG-H) and light chain (IgG-L) as indicated. (C) Correlative analysis of full-length PrPC glycosylation profiles with susceptibility to M1000 prion infection. Data points represent the replicates obtained from Western blot quantification of diglycosylated, monoglycosylated, and unglycosylated PrPC as shown in (B). Susceptible, resistant, and susceptible* are as defined in Figure 2. Binary logistic regression (Minitab 15) was performed; no significant differences were seen.

Expression of Lower Proportions of Full-Length PrPC Correlates With Resistance to M1000, Largely Influenced by Increased PrPC α-Cleavage

PrPC contains 2 well-characterized constitutive internal cleavage sites: around residues 111/112 (4) and at the C-terminus of the octapeptide repeat region (8, 9) (Fig. 4A). The PrPC cleavage profile of each cell line was determined by Western blot epitope mapping of deglycosylated PrPC (Fig. 4A). PrPC endoproteolysis varied dramatically among the cell lines. The C1 and C2 fragments were present in all cell lines (Figs. 4B-E), however with the exception of the GT1-7C and MoRK13 cells, the lines had proportionately low C2 levels (Fig. 4F) requiring overexposure of the Western blots for detection. The OBL-21 cell line was the only line with full-length unglycosylated PrPC (FLUG) as the clearly dominant species compared with the C1 and C2 fragments (Fig. 4F). Collectively, these data indicate that constitutive PrPC cleavage was variable among the cell lines, but was ubiquitous, suggesting that it may have functional significance.

FIGURE 4.

Cellular prion protein (PrPC) cleavage profiles of cell lines determined by Western blot (WB) epitope mapping. (A) Schematic representation of endogenous PrPC cleavage and alignment of antibodies used for epitope mapping. (B-D) Representative WB of PrPC fragments in the different uninfected cell lines; 15% Tris-glycine sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), WB with ICSM18 (B), 03R19 (C), and SAF32 primary antibodies (D). PGF, PNGaseF digested samples. (E) Comparison of PrPC fragments in the 2 cognate GT1-7 cell lines (C and H); all samples were PNGaseF treated; SDS-PAGE and WB conditions were as in (B-D); WB with 8B4 antibody is indicated. For all WBs, the amounts of total protein loaded in each well depended on relative PrPC expression levels and were adjusted to enable visual identification of PrPC truncated fragment/s. Based on apparent molecular weight and epitope mapping, diglycosylated PrPC (di), monoglycosylated PrPC (mono), full-length unglycosylated PrPC (FLUG), and the C1 and C2 truncated PrPC species are indicated. (F) Quantification (mean ± SEM) of relative levels of unglycosylated PrPC species in the different cell lines, as seen in ICSM18 WB images as this is the only antibody that detects all 3 PrPC species. The replicate numbers for each cell line (n) are indicated on x axis labels.

Similar to the PrPC glycosylation profiles previously described, there were significantly different truncated PrPC profiles in the 2 GT1-7 cell lines (Figs. 4E, F). Moreover, there were 2 anomalies in the GT1-7C cells. The first was an immunoreactive band detected with SAF32 at the electrophoretic mobility of the C2 fragment (Figs. 4D, E), suggesting that a proportion of β-cleavage in this line occurred slightly more N-terminally than in the other cell lines. The second anomaly was the apparent doublet band of “FLUG” PrPC in the GT1-7C cells (Fig. 4E). Because this faster migrating band was not detected with 8B4 (and consistent with the approximate 2- to 3-kd molecular weight difference between the 2 bands of the doublet), it likely resulted from N-terminal truncation of full-length PrPC up to at least some point within the 37-44 8B4 epitope. A PrPC doublet has been reported, but was not characterized (31). On close inspection of the blots, this fragment may also have been present in the N2a line (Figs. 4B, C), suggesting that it might occur in various cells but may be overlooked depending on PAGE resolution and Western blot conditions.

After characterization of PrPC endoproteolytic cleavage profiles in each cell line, correlative analyses comparing these profiles with the innate susceptibility to M1000 prions were performed. The cells were grouped according to their susceptibility or resistance (Table 2), again including and excluding the MoRK13 cells to allow for any influence of exogenous overexpression of PrPC. The relative proportion of each PrPC fragment (data points obtained from quantification of FLUG, C2, and C1, as shown in Fig. 4F) was then separately plotted against susceptibility or resistance (Fig. 5). The cells that were more likely to be susceptible to M1000 infection expressed a higher proportion of full-length PrPC, (analysis including MoRK13; p = 0.003; odds ratio [OR], 1.08; confidence interval [CI], 1.03-1.14; and excluding MoRK13; p = 0.003; OR, 1.09; CI, 1.03-1.15). Furthermore, the analyses also indicated that truncation specifically at the α-cleavage site was protective against M1000 infection, with the OR predicting decreased susceptibility with expression of higher proportions of the C1 fragment (analysis, including MoRK13; p = 0.027; OR, 0.94; CI, 0.89-0.99; and excluding MoRK13; p = 0.029; OR, 0.94; CI, 0.88-0.99). C2 cleavage was not protective despite any bias caused by the bimodal distribution of the data (the cluster of data points with >50% C2 was exclusively from the GT1-7C cells).

FIGURE 5.

Correlative analysis of relative susceptibility to M1000 infection and constitutive cellular prion protein (PrPC) cleavage profiles. Data points represent the replicates obtained from Western blot quantification of full-length unglycosylated PrPC (FLUG), C2, and C1 as shown in Figure 4F. Susceptible, resistant, and susceptible* are as defined in Figure 2. Binary logistic regression was performed (Minitab 15); significant results are as indicated.

Discussion

Differential susceptibility of cell lines to individual prion strains is well described (Table 1), and a number of cellular properties unrelated to PrPC may influence the intercellular variations in susceptibility and resistance to long-term propagation of PrPres. The apparent subsequent loss of prion infection, shown here for the GT1-7H and OBL-21 cells, has also previously been described (32-35). The cause of this phenomenon was not specifically investigated but might represent an adaptive cellular response to the ongoing unfavorable PrPres propagation. Such cellular changes may reflect diminished PrPC/PrPres interactions and conversion, and/or enhanced PrPres clearance or degradation, with evidence suggesting that the steady state levels of misfolded conformers are determined by the equilibrium between their production and degradation (36). Furthermore, anecdotal evidence for a finely balanced equilibrium comes from cells that have a propensity to clear prion infection where simple changes in the batches of culture media or serum used resulting in subtle differences in serum proteins or the concentration of salts, amino acids, trace metals, or other additives, can lead to a loss in prion infection. Finally, although most prion-infected cells display no obvious detrimental phenotype, changes in iron metabolism (37), copper binding (38), synaptic functions, (39) and responses to oxidative stress (40) have all been described; loss of infection may result from effects of these subtle changes, if healthier uninfected cells outcompete or outgrow the infected cells.

The primacy of PrPC in the transmissibility and pathogenesis of prion diseases (12, 13) supports the likelihood that subtle differences in PrPC biology (e.g. posttranslational processing, trafficking, or specific subcellular localization) may also constitute important susceptibility determinants in addition to the presence of expression levels of PrPC. Whilst PrPC expression levels modulate incubation period in animal models of prion disease (41, 42), it seems that in vitro, PrPC expression levels are less critical in influencing susceptibility to prion infection (36, 43). Therefore, we used the M1000 human-derived prion strain and a range of cell lines encompassing anticipated susceptibility or resistance to prion infection to explore PrPC posttranslational processing as a determinant of sustained PrPres propagation.

A previous study examined the heterogeneity of PrPC glycosylation and reported that PrPC glycoform patterns in mouse and rat brain and neuronal cell lines vary but are distinct (31). We also demonstrated diversity of PrPC glycosylation across different cell types. Because each of the cell lines investigated has different origins, the likelihood that different glycans are available for N-linked glycosylation, and varied activities of the enzymes or transferases involved in oligosaccharide processing, might explain the diversity. N-linked glycosylation sites are conserved in all mammalian prion genes (27), possibly indicating an important functional role for this posttranslational modification; however the significance of PrP glycosylation in either normal PrPC function or in prion disease pathogenesis has not yet been determined. Nonetheless, transgenic mice that express only unglycosylated PrPC can be infected with a mouse-adapted scrapie prion strain, propagate only unglycosylated PrPres, and transmit disease to wild-type mice (44). This indicates that glycosylation of PrPC or the infectious prion is not always essential for transmission. We also found that glycosylation profiles were not important determinants of inherent cellular susceptibility to M1000 prions and sustained PrPres propagation, a finding that is broadly congruent with previous studies (45-47).

There is both in vivo (44) and in vitro (30) evidence that prion strains may preferentially convert host PrPC with a more comparable glycosylation state. This phenomenon was not observed in the current study, suggesting that for the M1000 strain, there are alternative factors that override any influence of host PrPC glycosylation patterns in determining susceptibility to infection. There is also a suggestion that glycosylation of PrPC may be protective against prion infection by preventing the natural tendency of PrPC to form a β-sheet or PrPSc-like conformation (29). Our results argue against this because all of the cell lines investigated (both susceptible and resistant) expressed higher proportions of diglycosylated PrPC compared with the monoglycosylated and unglycosylated species.

The N-terminus of PrPC is not required for production of PrPres in cultured cells (48). Mice that express only N-terminally truncated PrPC (lacking residues 32-80) on a Prnp null background generated PrPres after scrapie infection and succumbed to disease after an apparently extended incubation period (41), although the significance of the longer scrapie incubation was not determined because of experimental variations in that study. In a cell-free conversion assay, progressively longer N-terminal deletion mutants (up to Residue 124) served as a template for production of PrPres, with decreasing efficiency compared with full-length PrPC, the more C-terminal the deletion (49). That purely in vitro assay, however, does not allow for possible disruption of cellular trafficking of mutant PrPC resulting in altered conversion of PrPC to PrPres. We found that when the C1 fragment constituted a higher proportion of the total expressed PrPC, there was reduced susceptibility to M1000 infection with sustained PrPres propagation; there was increased susceptibility to infection if greater relative proportions of full-length PrPC were expressed. Taken together, these results suggest that although truncated PrPC can be induced to misfold, conversion and propagation of PrPres are more efficient when there is a full-length PrPC substrate. This theory also provides the most likely explanation for why the GT1-7C cells were resistant to M1000 infection because only 10% of the total PrPC found in these cells was full-length. Furthermore, the doublet PrPC band in the GT1-7C and possibly the N2a cell lines may represent a novel truncated PrPC species that might also influence susceptibility to M1000 prion infection.

Endoproteolytic α- and β-cleavage of PrPC somewhat parallels the cleavage of the Alzheimer disease-associated amyloid precursor protein (APP) (50). Alzheimer disease pathogenesis is related to aberrant endogenous proteolytic cleavage of the transmembrane glycoprotein APP and accumulation of the resultant amyloid-β peptide. Distinct from the pathogenic switch of α-secretase APP processing to predominant β-secretase APP cleavage, our results suggest that it is not a switch from 1 PrPC processing pathway to another that is a precursor to pathogenesis, but that it is the depletion of available full-length PrPC as the optimal substrate that is protective against prion infection. Interestingly, some of the cell lines used herein that were resistant to M1000 prions have previously shown susceptibility to other prion strains (Table 1). Therefore, analogous studies using additional prion strains would provide further insight as to whether the observations made in this study are more broadly applicable or M1000 strain specific.

Overall, the results described herein highlight the complexity of PrPC biogenesis and how aspects of this, especially constitutive endoproteolytic cleavage of PrPC, are important cellular characteristics that contribute to susceptibility to nontransient M1000 prion infection, with increased α-cleavage offering protection. Thorough analysis of truncated PrPC species within different brain regions may provide insight into the selective neuronal vulnerability and topographic distribution of neuropathologic abnormalities characteristic of prion diseases.

Acknowledgments

The authors thank the following researchers for kindly providing cell lines: the 2 cognate hypothalamic GT1-7 cell lines, GT1-7C and GT1-7H, were a kind gift from Professor H.M. Schätzl, Institute for Virology, Technical University Munich, to Steven J. Collins and Andrew F. Hill, respectively. The N2a neuroblastoma cells were from Dr G. Evin, Department of Pathology, the University of Melbourne. The NIH/3T3 cells were from Dr F. Sernee, formerly of the Department of Pathology, the University of Melbourne. OBL-21 cells were a gift from Dr B. Chesebro (RML, NIAID, NIH, MT) to Victoria A. Lawson. The RK13 cells were given to Victoria A. Lawson by Victorian Infectious Diseases Reference Laboratory, and the MoRK13 cells were generated by Ms R. Sharples, Department of Biochemistry and Molecular Biology, the University of Melbourne.

Footnotes

  • This work was supported by an NH&MRC Program Grant No. 400202. Cathryn L. Haigh is supported by a University of Melbourne Early Career Researcher Grant; Steven J. Collins by an NH&MRC Practitioner Fellowship No. 400183; Steven J. Collins and Victoria A. Lawson by an NH&MRC Project Grant No. 454546; Victoria A. Lawson and Andrew F. Hill by an NH&MRC Project Grant No. 400229; and Andrew F. Hill by an NH&MRC Career Development Award No. 251745.

  • Online-only color figures are available at http://www.jneuropath.com.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
View Abstract