Modulation of Glycosaminoglycans Affects PrP^Sc Metabolism but Does Not Block PrP^Sc Uptake

ABSTRACT

Mammalian prions are unconventional infectious agents composed primarily of the misfolded aggregated host prion protein PrP, termed PrP^Sc. Prions propagate by the recruitment and conformational conversion of cellular prion protein into abnormal prion aggregates on the cell surface or along the endocytic pathway. Cellular glycosaminoglycans have been implicated as the first attachment sites for prions and cofactors for cellular prion replication. Glycosaminoglycan mimetics and obstruction of glycosaminoglycan sulfation affect prion replication, but the inhibitory effects on different strains and different stages of the cell infection have not been thoroughly addressed. We examined the effects of a glycosaminoglycan mimetic and undersulfation on cellular prion protein metabolism, prion uptake, and the establishment of productive infections in L929 cells by two mouse-adapted prion strains. Surprisingly, both treatments reduced endogenous sulfated glycosaminoglycans but had divergent effects on cellular PrP levels. Chemical or genetic manipulation of glycosaminoglycans did not prevent PrP^Sc uptake, arguing against their roles as essential prion attachment sites. However, both treatments effectively antagonized de novo prion infection independently of the prion strain and reduced PrP^Sc formation in chronically infected cells. Our results demonstrate that sulfated glycosaminoglycans are dispensable for prion internalization but play a pivotal role in persistently maintained PrP^Sc formation independent of the prion strain.

IMPORTANCE Recently, glycosaminoglycans (GAGs) became the focus of neurodegenerative disease research as general attachment sites for cell invasion by pathogenic protein aggregates. GAGs influence amyloid formation in vitro. GAGs are also found in intra- and extracellular amyloid deposits. In light of the essential role GAGs play in proteinopathies, understanding the effects of GAGs on protein aggregation and aggregate dissemination is crucial for therapeutic intervention. Here, we show that GAGs are dispensable for prion uptake but play essential roles in downstream infection processes. GAG mimetics also affect cellular GAG levels and localization and thus might affect prion propagation by depleting intracellular cofactor pools.

INTRODUCTION

Prion diseases are progressive, fatal neurodegenerative diseases of humans and other mammals. The formation of protease-resistant aggregates of the host-encoded prion protein is central to pathogenesis. Growing evidence supports the hypothesis that aggregates of the misfolded host prion protein PrP^C, termed PrP^Sc, or folding intermediates thereof are the main, if not the sole, constituent of the infectious agent (1, 2). Prion strains from different donor species have been adapted to laboratory rodents, where they manifest themselves with different disease progressions and pathologies (3). Prion strains target different brain areas in vivo (4) and exhibit restricted cell tropism in vitro (for a review, see reference 5). A growing body of evidence argues that strain information is encoded within the respective three-dimensional fold of the PrP^Sc aggregates (6).

The early steps of the prion entry process, the manifestation of a productive infection, and the exact sites of prion conversion are not fully understood (for a review, see reference 5). PrP^Sc formation occurs either on the cell surface or along the endocytic pathway upon interaction of PrP^Sc with PrP^C (7–12). It has been proposed that PrP^Sc formation requires cofactors, such as nucleic acids, phospholipids, or glycosaminoglycans (GAGs), for internalization and/or PrP^Sc formation (13, 14). GAGs, such as heparan sulfate (HS) and chondroitin sulfate (CS), are linear polysaccharides consisting of amino sugars and uronic acid that undergo extensive N- or O-sulfation and constitute ubiquitous components of the cell surface and the extracellular matrix (15). PrP^C associates with HS and CS through interaction of positively charged PrP residues with negative charges of the carbohydrates (16, 17). This interaction might modulate endocytosis of PrP^C (18, 19). Both PrP^C and PrP^Sc bind to sulfated GAG heparin in vitro (20–22). Low-molecular-weight heparin also modulates the thermodynamic stability of recombinant PrP (23). GAGs have been implicated as cofactors that catalyze the conversion of PrP^C into PrP^Sc, likely by serving as a scaffold for PrP^C-PrP^Sc interactions (13). The importance of GAGs in prion pathogenesis is supported by the findings that HS colocalizes with abnormal prion protein deposits in vivo (24, 25). Furthermore, GAG modulators exhibit antiprion activity in animal models (21, 26–29). Studies addressing the question of whether cell-associated GAGs represent attachment factors that enable prion uptake have yielded inconsistent results (21, 30, 31). Importantly, most studies were performed with detergent-extracted or proteinase K-treated prions. Those treatments, however, have drastic effects on the structure and/or amino acid sequence of PrP^Sc (32) and can alter its cellular uptake and infectivity (33–35). So far, it is unclear if cell-type- and strain-specific differences in the GAG requirements for prion entry and the establishment of chronic infections exist.

Soluble GAGs, such as HS and heparin, as well as GAG-related sulfated polysaccharides, including dextran sulfate, pentosan polysulfate, and suramin, act as GAG mimetics with potent antiprion activity in vivo and ex vivo (12, 20, 26, 29, 31, 36–40). Sulfate moieties of GAG mimetics are required for the antiprion activity (40). Sodium chlorate, a competitive inhibitor of the cellular 3′-phosphoadenosine 5′-phosphosulfate, prevents both HS and CS sulfation (41–43) and also decreases PrP^Sc accumulation in persistently infected cells (31, 44). GAG sulfation also affects PrP^Sc formation in in vitro assays and thus directly acts on PrP^Sc amplification (45). So far, a comparative analysis of the effects of GAG modulators on host cell PrP^C, on endogenous sulfated GAGs, and on the individual stages of infection by different strains has not been performed. In this study, we analyzed how the GAG mimetic DS-500 and sodium chlorate (NaClO₃) affect acute and persistent prion infections by the mouse-adapted prion strains RML and 22L. We analyzed in detail if cellular GAGs act as essential receptors for prion internalization. Our study demonstrates that both DS-500 and sodium chlorate reduce endogenous sulfated GAGs but have divergent effects on cell surface and total PrP^C levels. Neither RML nor 22L prions require endogenous GAGs to gain entry into the cell. However, although PrP^Sc is efficiently taken up by cells, DS-500 or undersulfation during exposure to prions affects the establishment of productive infections and strongly reduces PrP^Sc in chronically infected cells. Our data underscore the important role of sulfated GAGs as general cofactors for prion replication, either by directly engaging in PrP^Sc formation or by modulating the cellular levels and distribution of PrP^C.

MATERIALS AND METHODS

Cell culture and reagents.

This study was conducted under biosafety containment level 2 in accordance with the German Engineering Act of April 2008. The susceptibility of L929 cells (46) to prion strains was increased by several rounds of cloning and selection of susceptible clones. Cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS) and antibiotics. L929 cell clone 15.9 was used for the generation of cell populations persistently infected with 22L and RML prions. CHO cells (ATCC CCL-61) and mutant CHO cells that were variably deficient in GAGs were purchased from the American Type Culture Collection (ATCC) and maintained in Ham's F-12K medium supplemented with 10% FCS and antibiotics. CHO pgsA-745 (ATCC CRL-2242) cells are defective in GAG synthesis, as they lack functional xylosyltransferase activity. CHO pgsD-677 (ATCC CRL-2244) cells are deficient in N-acetylglucosaminyltransferase and glucuronyltransferase and do not produce HS (47). N2a cells were cultured in Opti-MEM supplemented with 10% FCS and antibiotics.

Unless otherwise noted, chemicals were from Sigma (Steinheim, Germany). Proteinase K was obtained from Roth (Karlsruhe, Germany) and Pefabloc from Roche (Mannheim, Germany). Cell culture media and supplements were purchased from Invitrogen (Darmstadt, Germany), and fetal bovine serum was from PAA Laboratories (Pausing, Austria). The ECL plus chemiluminescence kit was from GE Healthcare (Buckinghamshire, United Kingdom). Anti-PrP antibody 4H11 has been described previously and was generated using dimeric murine PrP as an immunogen (48). Anti-HS antibody 10E4 was purchased from Amsbio (Frankfurt/Main, Germany). The epitope in HS comprises N-sulfated acetylglucosamine residues. The antibody does not react with hyaluronan, CS, dermatan sulfate, keratan sulfate, or DNA. Fluorescein-conjugated secondary antibodies were from Dianova (Hamburg, Germany), and mouse monoclonal anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody 374 was from Millipore/Chemicon (Billerica, MA, USA).

Prion infection.

Standard brain homogenate preparations from terminally sick mice were used to circumvent potential artifacts due to detergent extraction of scrapie-associated fibrils (35). Brain homogenates are routinely used for in vivo and ex vivo infection (31, 46) and exhibit very high prion titers (34). The mouse-adapted scrapie strains RML and 22L were passaged in C57BL/6 mice. Brain homogenates (10%) of healthy control mice and terminally sick mice were prepared in Opti-MEM using a Dounce homogenizer. Cell debris was pelleted by centrifugation at 872 × g for 5 min at 4°C, and the resultant brain homogenates were stored at −80°C. A total of 2 × 10⁴ to 1 × 10⁵ cells per well were plated in 24-well plates or on ibidi dishes (ibidi, Martinsried, Germany). The medium was replaced after 24 h or 48 h by 200 μl 1% brain homogenate in fresh medium with supplements. After 5 h, the brain homogenate was replaced by 1 to 2 ml of culture medium. For studying early events of a prion infection, cells were analyzed 18 h postexposure to brain homogenate. Cells were routinely split at a ratio of 1:8 or 1:10 every 3 to 4 days. Prion infection was determined by testing lysates of cells at passage 2 or 3 after prion exposure for protease-resistant PrP^Sc by Western blotting.

Dextran sulfate and sodium chlorate treatment.

For de novo infections, cells were plated in 24-well plates; 24 h later, the cells were preincubated with 50 mM NaClO₃ for 24 h or treated with 1 μg/ml DS-500 at the time of exposure to brain homogenates. The cells were exposed to 1% brain homogenates of uninfected or RML or 22L prion-infected mice in the presence or absence of compounds. Five hours postexposure, the medium containing brain homogenate was replaced by fresh medium with or without compounds. The cells were passaged three times before PrP^Sc formation was assessed by Western blotting. To determine the effects of the compounds on PrP^Sc accumulation in persistently infected cells, subconfluent monolayers were exposed to 0.01 to 1 μg/ml DS-500- or 5 to 50 mM NaClO₃-containing medium. Seventy-two hours postexposure, the cells were lysed, and the lysates were tested for PrP^Sc content by Western blotting.

Viability assay (XTT).

The viability of a cell population upon treatment with different compounds was determined with the XTT assay (Roche, Mannheim, Germany). L929 cells were plated at a density of 1.5 × 10⁴ cells per well in 96-well plates. The following day, the cells were treated for 72 h with various concentrations of the indicated compounds. Subsequently, 50 μl of the XTT reagent was added to each well. After incubation for 4 h, absorption at 450 nm was measured with a Fluostar Omega plate reader (BMG Labtech, Offenburg, Germany). The average absorption of four control wells was set as 100% viability. The viability of treated cells was compared to the viability of the control cells.

Blyscan sulfated glycosaminoglycan assay.

The Blyscan sulfated glycosaminoglycan assay kit (Biocolor, Carrickfergus, United Kingdom) quantitatively measures sulfated proteoglycans and GAGs in biological samples. The assay was performed according to the manufacturer's instructions. Briefly, 4 × 10⁵ cells were plated in a T25 flask, and 48 h later, the cells were treated with 1 μg/ml DS-500 or 50 mM NaClO₃. Twenty-four hours after treatment, the medium was aspirated, and the cells were rinsed with phosphate-buffered saline (PBS). The cells were lysed in papain extraction reagent added to the cell monolayer for 3 h at 65°C. The total cell extract containing total GAGs was harvested, and samples were centrifuged at 10,000 × g for 10 min. A total of 100 μl of the supernatant was used for the assay.

Flow cytometry.

For cell surface PrP^C detection, cells were rinsed with PBS and detached with 1 mM EDTA in PBS. The cells were pelleted for 2 min at 4°C and 314 × g. The cells were resuspended in fluorescence-activated cell sorter (FACS) buffer (2.5% FCS, 0.05% sodium azide in PBS) and incubated with primary antibody for 45 min on ice. After rinsing the cells three times with FACS buffer, they were incubated for 45 min on ice with the appropriate fluorophore-conjugated secondary antibody diluted in FACS buffer. The cells were rinsed three times and resuspended in FACS buffer. To detect dead cells, 7-aminoactinomycin D (7-AAD) was added to all samples. 7-AAD-positive cells were excluded from the analysis. To measure the total PrP^C content in the cells, they were detached and fixed with Roti-Histofix (Roth, Karlsruhe, Germany) for 10 min at room temperature (RT). The cells were rinsed with PBS and quenched with 50 mM ammonium chloride, 20 mM glycine for 10 min at RT. After rinsing with 0.1% saponin in FACS buffer, the cells were incubated with the primary antibody diluted in saponin buffer for 45 min at RT. After three rinsing steps with saponin buffer, the fluorophore-conjugated secondary antibody was added to the cells for 45 min at RT. The cells were again rinsed three times and resuspended in saponin buffer. Control samples incubated without the first antibody were included in each experiment. Nonspecific fluorescence signals of the secondary antibody (background) in these control samples were subtracted from the measurements of all other samples. Experiments were performed using a Gallios flow cytometer (Beckman Coulter, Krefeld, Germany). A total of 15,000 events per sample were measured using Gallios software (Beckman Coulter, Krefeld, Germany).

Detection of abnormal prion protein by confocal microscopy.

Cells plated on ibidi dishes were rinsed thoroughly with PBS and fixed with 4% paraformaldehyde (PFA) for up to 20 min and permeabilized with 0.1% Triton X-100 for 10 min. For detection of PrP^Sc, cells were incubated with 6 M guanidine hydrochloride (GdnHCl) and thoroughly rinsed with PBS. Samples were blocked with 0.2% gelatin for 30 to 60 min, followed by incubation with primary antibodies diluted in 0.2% gelatin for 45 to 60 min. After three rinsing steps, the cells were incubated for 45 min with fluorophore-conjugated secondary antibodies. Nuclei were counterstained with Hoechst (1:10,000 in PBS). In some instances, the cytoplasm was stained with HCS CellMask blue stain (1:5,000 in PBS) (Life Technologies, Darmstadt, Germany) for 30 min. Confocal laser scanning microscopy was carried out using an LSM 700 laser-scanning microscope (Zeiss, Jena, Germany). Uninfected (mock) cells were included for every individual time point and used to control for background staining of normal cellular PrP. To quantify the differences in fluorophore intensities, the same settings were used for each channel during the experiment.

Western blot analysis.

Cells were lysed in lysis buffer (10 mM Tris-HCl, pH 7.4, 100 mM NaCl, 10 mM EDTA, 0.5% sodium deoxycholate, and 0.5% Triton X-100), and the lysates were cleared of cell debris by centrifugation (1 min; RT; 20,817 × g). The protein concentration in the lysates was determined with a bicinchoninic acid (BCA) protein assay kit (Thermo Scientific, Rockford, IL, USA), and the lysates were adjusted to the same protein concentration. For PrP^C detection, Pefabloc was added to the cell lysate, and proteins were precipitated with 3 volumes of methanol at −20°C overnight. To detect PrP^Sc, the cell lysate was incubated with 20 μg/ml proteinase K (PK) for 30 min at 37°C. The reaction was terminated by the addition of Pefabloc. PrP^Sc was pelleted by centrifugation. The pellets were resuspended in sample buffer (90 mM Tris-HCl, pH 6.8, 7% SDS, 0.01% bromphenol blue, 30% glycerol, 20% β-mercaptoethanol) and assayed on 12.5% SDS-PAGE gels. PrP was detected by Western blotting using the mouse monoclonal antibody (MAb) 4H11, horseradish peroxidase-conjugated secondary antibody, and enhanced chemiluminescence. The membranes were stripped with 1× Re-Blot Plus Strong Solution (Millipore, Temecula, CA, USA; 1:10) and reprobed with antibodies.

Image data analysis.

Z-stacks of confocal images were converted to two-dimensional (2D) data by maximum-intensity projection (MIP) using Fiji (http://rsbweb.nih.gov/ij). The detection of cells with PrP^Sc uptake was conducted with CellProfiler cell image analysis software (Broad Institute). All further calculations were performed with Matlab 2010b (Mathworks GmbH, Ismaning, Germany). Single cells were identified on the basis of nucleus detection. The nucleus image was smoothed with a Gaussian filter (sigma = 2.97 μm), and nucleus positions were defined by local intensity maxima. Cell regions were subsequently defined by applying a morphological watershed (Matlab function, watershed), where the maximal cell radius was restricted to 21.0 μm. Cell regions touching the image border were excluded from analysis. Subsequently, the mean intensity of the 10E4 signal per cell was determined. The same processing routine was used for determining PrP^Sc intensities in cells exposed to prions. Here, only cells detected as positive for PrP^Sc uptake, defined by an intensity threshold, were selected for image analysis. To analyze HS and PrP^Sc spatial distributions within the cell, we adapted a method for averaging of cell images, as previously described (49). In brief, the orientation of each cell was defined by the angle where the spatial spread of the fluorescence signal was maximal. On the basis of this angle and the nucleus position, each cell was transformed into a coordinate system with the nucleus located at the system's origin and the cell's orientation along the major axis. This allowed the calculation of “average cells” (see Fig. 1c, 3d, and 4c).

FIG 1.

Effects of DS-500 and NaClO₃ treatment on endogenous sulfated GAGs, PrP^C, and cell viability. (a) Total sulfated GAGs in L929 cells treated with 1 μg/ml DS-500 or 50 mM NaClO₃ for 24 h determined by the Blyscan sulfated glycosaminoglycan assay. Total sulfated GAGs of untreated control (ctrl) cells were set to 100%. The bars represent mean values and SD (n = 3). (b) L929 cells were treated with 1 μg/ml DS-500 or 50 mM NaClO₃ for 24 h. HS was detected using MAb 10E4. Nuclei were stained with Hoechst. Scale bar, 10 μm. (c) HS signals in panel b averaged at least 97 cells per treatment. (Left) Cells were aligned along the vertical axis by an automated image-processing routine. (Right) HS intensities of DS-500- and NaClO₃-treated cells were quantified relative to control cells set to 100%. The bars represent mean values and SD. (d) XTT assay of cells treated with different concentrations of DS-500 and NaClO₃ for 72 h. The viability of treated cells was compared to the viability of control cells, set to 100%. The bars represent mean values and SD (n = 4). (e) Western blot of lysates of L929 cells incubated with 1 μg/ml DS-500 or 50 mM NaClO₃ for 24 h. PrP^C was detected by antibody 4H11. Additional lanes were excised for presentation purposes. Molecular mass markers are shown. (f) Cell surface PrP^C and total PrP^C expression analyzed by flow cytometry 72 h after exposure to the compounds. The shaded graph represents untreated control cells. The dark-gray line represents DS-500 (1 μg/ml)-treated cells, and the light-gray line shows NaClO₃ (50 mM)-treated cells. PrP^C at the cell surface was detected in nonpermeabilized cells, and total PrP^C was detected in fixed and permeabilized cells. (g) Mean fluorescence intensities (MFIs) of DS-500- and NaClO₃-treated cells were compared to the MFIs of control cells, set to 100%. The bars represent mean values and SD (n = 3) (***, P ≤ 0.001; **, P ≤ 0.01; *, P < 0.05; ns, not significant).

FIG 3.

Initial uptake of RML and 22L prions is not inhibited by DS-500 and NaClO₃. (a) PrP^Sc detected 18 h after exposure to brain homogenate originates from the inoculum. Uninfected L929 cells (top) or uninfected 3F4-L929 cells (bottom) were exposed to RML or 22L prions for 5 h. Eighteen hours postexposure, the cells were lysed. Total PrP^Sc and de novo generated PrP^Sc were detected following PK treatment of cell lysates using MAb 4H11 or MAb 3F4, respectively. Experiments were performed in triplicate. Additional lanes were excised for presentation purposes. The loading control (ldg ctrl) for L929 cells was PK-treated lysate of persistently infected L929^RML and L929^22L cells. The loading control for 3F4-L929 cells was PK-treated lysate of persistently infected 3F4-L929^22L cells. Note that the glycosylation profile and size of cell-derived PrP^Sc in the loading controls differed from those of brain-derived PrP^Sc detected 18 h postexposure. (b) Confocal microscopy analysis of immunofluorescence staining of PrP^Sc in compound-treated cells 18 h after exposure to prions. Uninfected L929 cells were pretreated with 1 μg/ml DS-500 or 50 mM NaClO₃ for 24 h prior to exposure to prions in the presence of compounds for 5 h. Inoculation was performed in the presence of compounds for 18 h. Cells exposed to mock brain homogenate and untreated cells served as controls. PrP^Sc was detected by antigen retrieval using MAb 4H11. Nuclei were counterstained with Hoechst. Scale bars, 10 μm. (c) Maximum-intensity projections of the cells in panel b were analyzed using Cell Profiler software. A total of at least 360 cells from three independent experiments were analyzed. The number of treated cells with PrP^Sc uptake was normalized to the number of control cells with PrP^Sc uptake, set to 100%. The bars represent mean values and SD (n = 3). (d) PrP^Sc signal in panel b averaged over multiple PrP^Sc-positive cells aligned along the vertical axis. At least 176 cells per treatment were analyzed. (e) Quantification of PrP^Sc intensities of positive cells. The PrP^Sc intensities of DS-500- and NaClO₃-treated cells were normalized to untreated control cells, set to 100%. The box plots illustrate the median of PrP^Sc intensity, and the whiskers display the 5th to 95th percentiles (***, P ≤ 0.001; **, P ≤ 0.01; ns, not significant; Kruskal-Wallis test). (f) N2a cells were exposed to mock or 22L scrapie brain homogenate in the presence or absence of DS-500 or sodium chlorate. Five hours postexposure, the inoculum was removed, and the cells were cultured for an additional 13 h. PrP^Sc internalization was assessed by immunofluorescent staining. PrP^Sc and nuclei were detected as for panel b. Scale bars, 10 μm.

FIG 4.

Strain-independent internalization of PrP^Sc by GAG-deficient CHO cells. (a) CHO cells and mutant CHO pgsA-745 and pgsD-677 cells were exposed to mock brain homogenate or RML and 22L brain homogenates. PrP^Sc was detected following antigen retrieval by confocal microscopy analysis. Nuclei were stained with Hoechst. Scale bars, 10 μm. (b) Percentages of cells with PrP^Sc uptake shown in panel a. Maximum-intensity projections of cells shown in panel a were analyzed using Cell Profiler software. A total of at least 800 cells from three independent experiments were analyzed per treatment. The numbers of pgsA-745 and pgsD-677 cells with PrP^Sc uptake were normalized to the number of control CHO-K1 cells with PrP^Sc uptake, set to 100%. The bars represent mean values and SD (n = 3). (c) Fluorescent signal of PrP^Sc (red channel in panel a) averaged over a minimum of 620 cells per cell line/treatment, which were defined as PrP^Sc positive. The cells were aligned along the vertical axis by an automated image-processing routine. (d) PrP^Sc intensities of PrP^Sc-positive cells in panel c were calculated using an automated image-processing routine. The PrP^Sc intensities in pgsA-745 and pgsD-677 cells were normalized to CHO-K1 control cells, set to 100%. The box plots illustrate the median of PrP^Sc intensity, and the whiskers display the 5th to 95th percentiles (***, P ≤ 0.001; *, P < 0.05; ns, not significant; Kruskal-Wallis test).

Statistics.

Data sets were analyzed with one-way analysis of variance (ANOVA). To evaluate statistically significant differences between control sample data and treated-sample data, Dunnett's or Kruskal-Wallis multiple-comparison post hoc tests were used. P values of less than 0.05 were considered significant. The sample size was three or more.

RESULTS

DS-500 or NaClO₃ decreases endogenous GAG levels and divergently affects PrP^C levels.

Sulfation inhibition and the GAG mimetic DS-500 have been shown to strongly impact PrP^Sc formation in several ex vivo models of prion replication, but their precise effects on the establishment of prion infections remain unclear (20, 37, 39, 40). The effects of DS-500 and NaClO₃ treatment on endogenous levels of sulfated GAGs in L929 fibroblasts was first studied with the Blyscan sulfated glycosaminoglycan assay. The assay revealed that the total amount of sulfated GAGs was significantly decreased upon treatment with either DS-500 or NaClO₃ (Fig. 1a). To estimate the reduction of cellular HS, immunofluorescent staining of cells using the HS-specific antibody 10E4 (50) was performed (Fig. 1b). In untreated L929 control cells, HS was predominately localized to the cell membrane, with some punctate staining present in the cytoplasm. Upon NaClO₃ treatment, cell surface HS was drastically decreased and was redistributed to intracellular vesicular structures. Interestingly, the GAG mimetic DS-500 also reduced cell surface HS. Automated image analysis revealed a significant reduction of HS per cell upon DS-500 and NaClO₃ treatment (Fig. 1c). Doses up to 1 μg/ml DS-500 or 50 mM NaClO₃ for 72 h did not negatively affect cell viability (Fig. 1d). Western blot analysis of L929 cells treated with DS-500 or NaClO₃ demonstrated that PrP^C was properly glycosylated (Fig. 1e). The effects of both compounds on PrP^C levels were tested by flow cytometry. DS-500 treatment significantly reduced cell surface and total PrP^C levels (Fig. 1f and g). In contrast, NaClO₃ had an opposite effect and slightly but significantly increased both cell surface and total PrP^C (Fig. 1f and g). The finding that DS-500 reduces cell surface levels of PrP^C is in line with previous studies that argued that GAG mimetics can stimulate endocytosis of PrP^C (18). In conclusion, DS-500 and NaClO₃ both decrease endogenous GAGs but divergently affect PrP^C levels.

DS-500 or NaClO₃ reduces PrP^Sc accumulation in chronically infected L929 cells.

To test the antiprion activities of both compounds in our cellular model, L929 cells persistently infected with mouse-adapted prion strain RML or 22L were exposed to DS-500 and NaClO₃ for 3 days. Western blot analyses of PrP^Sc revealed a robust, dose-dependent decrease in PrP^Sc abundance (Fig. 2a and b). Thus, in agreement with previously published data (31), our data confirm that the GAG mimetic DS-500 and NaClO₃ exert general antiprion effects in chronically infected cells independent of the scrapie strain. The fact that both treatments reduced PrP^Sc content but had divergent effects on total and cell surface PrP^C levels argues that PrP^C and PrP^Sc levels are not necessarily correlated.

FIG 2.

DS-500 and NaClO₃ treatments reduce PrP^Sc in persistently infected L929 cells independently of the scrapie strain. Shown is reduction of PrP^Sc in L929^RML (a) or L929^22L (b) cells treated for 72 h with increasing concentrations of DS-500 or NaClO₃. (Left) Intensities of PrP^Sc signals detected by Western blotting were used for statistical analysis. (Right) PrP^Sc levels in the control cells were set to 100%. The bars represent mean values and SD (n = 3) (***, P ≤ 0.001; **, P ≤ 0.01; *, P < 0.05; ns, not significant).

Internalization of PrP^Sc does not require sulfated GAGs.

To test the effects of DS-500 and NaClO₃ on PrP^Sc uptake, uninfected L929 cells pretreated with compounds were exposed to prion strain RML or 22L or uninfected control brain homogenate for 5 h. The rinsed cells were subsequently cultured for an additional 13 h in the absence of inoculum. Western blot analyses of cells lysed 18 h after exposure to prions demonstrated that the PrP^Sc signal in inoculum-exposed cells differed in both glycosylation profile and molecular mass from that in persistently infected L929 cells loaded as a control (Fig. 3a, top). This strongly argues that PrP^Sc associated with the cells represents PrP^Sc from the inoculum and not de novo formed PrP^Sc. The result was further confirmed by exposing prion-susceptible L929 cells expressing antibody epitope-tagged mouse PrP (3F4-L929 cells) (46) to prion strains RML and 22L. No 3F4-positive PrP^Sc could be detected within 18 h (Fig. 3a, bottom), strongly arguing that PrP^Sc detected at this time point originates from the inoculum. For detection of internalized PrP^Sc, fixed cells were treated with GdnHCl prior to immunofluorescence staining, a method known to drastically increase immunostaining for PrP^Sc (Fig. 3b) (51, 52). Image analysis revealed that approximately 63.92% ± 25.71% of L929 cells had taken up exogenous RML PrP^Sc. 22L PrP^Sc was taken up by approximately 81.5% ± 10.29% of the cells. For statistical analysis between independent experiments, the numbers of cells that had taken up PrP^Sc were normalized to the untreated control cells, set to 100% (Fig. 3c). No significant difference in uptake was observed for RML PrP^Sc in the presence or absence of the compounds. Interestingly, significantly fewer cells internalized 22L PrP^Sc when exposed to NaClO₃, suggesting that 22L PrP^Sc uptake is more sensitive to sulfation perturbation (Fig. 3c). We further quantitatively assessed the amount of PrP^Sc taken up per cell in the presence or absence of either compound (Fig. 3d and e). No significant difference in RML PrP^Sc was detected upon DS-500 treatment, whereas 22L PrP^Sc was slightly but significantly decreased. Assessment of at least 176 cells per treatment revealed that undersulfation by NaClO₃ slightly but significantly increased total levels of PrP^Sc per cell, independent of the scrapie strain (Fig. 3d and e). The reason for the increased PrP^Sc signal per cell is unclear but might be related to a slight change in cell size upon NaClO₃ treatment (Fig. 3b). Overall, while slight differences in uptake efficiency exist, L929 cells treated with DS-500 or NaClO₃ are still capable of internalizing RML and 22L PrP^Sc. Efficient PrP^Sc uptake was also observed in N2a cells treated with DS-500 or sodium chlorate, arguing that the results were not dependent on the fibroblast line (Fig. 3f). Thus, the antiprion effects of DS-500 and NaClO₃ are not due to blockage of PrP^Sc internalization.

The experiments were repeated in CHO cells deficient in GAG synthesis (Fig. 4). CHO cells are not permissive to prions but can be used to monitor GAG-independent uptake of pathogens and cargo from the extracellular space (21). CHO pgsD-677 cells lack functional N-acetylglucosaminyltransferase and glucuronyltransferase and thus do not synthesize HS. CHO pgsA-745 cells are defective in xylosyltransferase and do not produce any GAGs (47). Lack of GAG or HS synthesis did not drastically affect PrP^Sc uptake or even increased intracellular PrP^Sc levels (Fig. 4a to c). CHO cells lacking HS expression appeared to be efficient in internalization of both RML and 22L PrP^Sc, as the numbers of cells that had taken up PrP^Sc were significantly increased compared to wild-type CHO cells (Fig. 4a and b). Deficiency in HS synthesis also increased PrP^Sc levels per cell (Fig. 4c and d). The fact that PrP^Sc was efficiently endocytosed by prion-permissive L929 cells with reduced sulfated-GAG expression and by CHO cells that did not express GAGs strongly argues for a GAG-independent mode of attachment and internalization of PrP^Sc.

DS-500 and NaClO₃ inhibit initiation of prion infection.

Next, we tested if internalization of PrP^Sc in the presence of compounds results in productive infections. For DS-500 exposure, cells were incubated for 5 h with infectious brain homogenate in the presence or absence (control) of the compound. Subsequently, brain homogenate was replaced with fresh medium or with medium containing DS-500 for an additional 19 to 43 h (Fig. 5a). Cells were expanded and tested for PrP^Sc production two passages postinfection by Western blot analysis (Fig. 5b). Interestingly, even short exposure of cells to DS-500 during the infection process drastically impaired PrP^Sc formation. The effect was more pronounced for strain RML, where 5-h treatment during infection decreased PrP^Sc levels in the infected cultures to approximately 8% of that in controls (Fig. 5c). PrP^Sc levels in cells exposed to 22L prions were approximately 15% of that in control cells after 24-h treatment with DS-500.

FIG 5.

DS-500 and NaClO₃ negatively affect establishment of RML and 22L prion infections. (a) L929 cells were seeded, and 48 h later, the cells were exposed to RML or 22L prions in the presence of 1 μg/ml DS-500. After 5 h, the inoculum was removed, and the cells were cultured in either medium or medium containing DS-500. Twenty-four hours postexposure, the medium was again changed, and after splitting of the cells for expansion, all the cells were cultured in medium without DS-500. The cells were lysed in passage 2 (P2) postexposure and analyzed by Western blotting. (b) Uninfected L929 cells were treated with DS-500 starting at the time of infection. Prion infection was analyzed by Western blotting for PrP^Sc (+PK) or total PrP (−PK) two passages postinfection. GAPDH protein levels were detected as a loading control on the −PK blot. (c) Quantification of PrP^Sc levels. The bars represent mean values and SD (n = 3). (d) Experimental setup. L929 cells were seeded, and 24 h later, the cells were treated with 50 mM NaClO₃. After 24 h of treatment, the cells were exposed to prions in the presence of NaClO₃. Five hours later, the inoculum was removed, and the cells were cultured in either medium or medium containing NaClO₃. Twenty-four hours postexposure, the medium was again changed, and after splitting of the cells for expansion, all the cells were cultured in medium without NaClO₃. The cells were lysed in passage 2. (e) Uninfected L929 cells were treated with NaClO₃ 24 h before exposure to RML or 22L prions. The cells were analyzed for PrP^Sc (+PK) or total PrP (−PK) two passages postinfection. GAPDH protein levels were detected on the −PK blot. Additional lanes were excised for presentation purposes. (f) Quantification of PrP^Sc levels. The bars represent mean values and SD (n = 3). Significant changes in the mean values relative to control samples (set to 100%) were calculated (***, P ≤ 0.001; **, P ≤ 0.01; ns, not significant).

Cells pretreated with NaClO₃ for 24 h to inhibit endogenous GAG sulfation were subsequently exposed to prions in the presence of the inhibitor for 5 h. Drug treatment was terminated or continued for an additional 19 to 43 h (Fig. 5d). Again, drug treatment had more pronounced effects on the establishment of RML prion infections (Fig. 5e and f). Here, pretreatment for 24 h and subsequent infection in the presence of the sulfation inhibitor reduced PrP^Sc levels to less than 30% of those of untreated controls (Fig. 5e). Interestingly, NaClO₃ treatment during the first 5 h of infection (29-h total drug treatment) did not significantly impair the establishment of 22L infections (Fig. 5f). Only after extended incubation with NaClO₃ (i.e., a total of 48 h) was a reduction in the PrP^Sc signal observed in 22L-infected cells (Fig. 5f). Thus, both DS-500 and NaClO₃ significantly impair the establishment of prion infections when administered within the first hours of infection, albeit to different degrees. In summary, DS-500 and NaClO₃ do not antagonize initial prion-host cell interactions, but rather, target downstream events following prion uptake.

DISCUSSION

Interference with endogenous GAG metabolism by GAG mimetics or impairment of cellular sulfation has been shown to drastically interfere with PrP^Sc formation in prion-infected cells (21, 30, 31, 37, 39, 40, 44), but the effects of the substances on different stages of the infection process have not been thoroughly assessed. In this study, we have examined in detail the effects of the GAG mimetic DS-500 and GAG undersulfation on endogenous GAG and PrP^C levels, as well as on PrP^Sc uptake and PrP^Sc accumulation during acute and persistent infection. We also compared the effects of the compounds on two different prion strains to assess any potential differences in prion strain infections. While slight differences in the sensitivities of strains RML and 22L to DS-500 and undersulfation exist, both treatments impaired the establishment of productive infections and significantly reduced PrP^Sc levels in chronically infected cells. The observed antiprion effects of the GAG mimetic or GAG undersulfation underline the importance of GAGs as general factors for prion biogenesis (14).

Cellular pathogens, such as viruses and microorganisms, commonly bind to GAG moieties on the cell surface to adhere to and gain entry into their target cells (15). The affinity of intracellular pathogens for certain GAGs that serve as coreceptors can be an important determinant of organ tropism and pathogenesis (53) and could also affect the internalization of different prion strains. Indeed, several studies argue that GAGs are essential coreceptors for PrP^Sc uptake (30, 54, 55). Recent studies even suggest that GAGs also mediate cellular internalization of misfolded protein aggregates in several neurodegenerative diseases (56, 57). Of note, however, quantitative assessment of PrP^Sc internalization was exclusively performed by Western blot analysis (21, 30, 31), a method that cannot accurately discriminate between cell surface-bound and internalized PrP^Sc. We therefore performed detailed analysis of immunolabeled cells, which allowed us to determine both the number of cells that had taken up PrP^Sc and the PrP^Sc intensity per cell. Our studies showed that treatment of L929 cells with DS-500 or sodium chlorate significantly reduced endogenous sulfated-GAG levels but did not inhibit PrP^Sc internalization. Similarly, the mouse neuroblastoma cell line N2a also efficiently took up PrP^Sc under the same treatment conditions, arguing that endogenous GAGs might not be required for prion invasion. In line with these findings, both RML and 22L PrP^Sc were efficiently taken up by CHO cells deficient in GAG expression. Interestingly, PrP^Sc uptake by CHO cells appeared to be significantly increased when the cells did not synthesize HS. One possible explanation for this finding is that bypassing PrP^Sc attachment to HS on the cell surface could allow direct binding to functional prion receptors, such as the laminin receptor precursor LRP/LR (54, 55, 58) or low-density lipoprotein-related protein 1 (LRP1) (59), both of which are expressed by L929 cells (data not shown). Our findings are consistent with results by Paquet and colleagues, who demonstrated efficient sheep scrapie PrP^Sc association with rabbit kidney epithelial cells in the presence of sodium chlorate or the GAG mimetic DS-500 (31). Likewise, another study demonstrated that DS-500 concentrations that inhibit PrP^Sc accumulation in RML-infected N2a cells were ineffective at inhibiting PrP^Sc binding to the cell (30). The conflicting results for PrP^Sc binding and uptake might, at least in part, be explained by prion strain differences. Some indication of this comes from in vitro studies using human enterocytes that readily take up bovine spongiform encephalopathy (BSE) PrP^Sc but do not appear to internalize mouse-adapted scrapie (55). While we did not observe such drastic strain-specific effects, slight differences in strain-dependent uptake were apparent upon undersulfation of endogenous GAGs. Thus, the presence of endogenous sulfated GAGs might still slightly influence the uptake of certain prion strains. Importantly, however, different PrP^Sc preparations could also have an impact on cellular uptake. Several studies have been performed using PK-treated prion brain homogenate or PK-treated, detergent-extracted PrP^Sc rods (21, 30). As PK also eliminates approximately 90 amino-terminal PrP residues, including GAG-binding motifs comprising amino acid residues 23 to 52 and 53 to 93 (16), binding affinities of PrP^Sc to putative cell surface receptors might be altered. Furthermore, detergent extraction has been shown to drastically change PrP^Sc infectivity and the kinetics of PrP^Sc uptake, possibly by artificially altering its aggregation state upon extraction (32–35, 55). The effect of prion preparations on PrP^Sc uptake was nicely demonstrated in human Caco-2/TC7 enterocytes that internalized BSE PrP^Sc from brain homogenate but did not take up PK-treated scrapie-associated fibrils (55). Of note, components other than PrP^Sc present in the brain homogenate could also influence prion uptake. However, even highly purified fractions of PrP^Sc still contain substantial amounts of non-PrP components (60–62). Clearly, the still undefined exact composition and structure of infectious prions add another layer of complexity to prion cell biology.

Our data suggest that GAG mimetics and prevention of endogenous GAG sulfation affect events downstream of the initial PrP^Sc attachment and internalization. Antiprion effects were apparent when drugs were administered during the early phases of prion infection and in persistently infected L929 cells, arguing that sodium chlorate and DS-500 target both initiation and maintenance of productive prion infections. Interestingly, higher concentrations of drugs were required for reducing PrP^Sc levels in L929 cells acutely or persistently infected with prion strain 22L. Whether this is related to higher levels of PrP^Sc in 22L-infected than in RML-infected L929 cells or strain-specific differences in sulfated-GAG dependence is so far unclear. However, the fact that reduced PrP^Sc levels were detected in both 22L- and RML-infected L929 cells strongly suggests that sulfated GAGs are generally required for efficient prion infection and PrP^Sc accumulation.

The effect of GAG undersulfation or GAG mimetics on PrP^Sc accumulation and prion infection could also be due to changes in substrate PrP^C and GAG metabolism (18). Both sodium chlorate and DS-500 treatments significantly reduced cellular HS levels and total sulfated GAGs. PrP can interact with endogenous GAGs through specific binding sites comprising amino acid residues 23 to 52, 53 to 93, and 110 to 128 (16), and interactions of PrP^C with GAGs is required for proper intracellular trafficking (18, 20, 44). Thus, any change in cellular GAG levels and distribution likely also affects PrP^C metabolism. Changes in endogenous sulfated-GAG levels could affect transport of PrP^C to subcellular sites of conversion and thereby modulate PrP^Sc metabolism. The conformational transition of PrP^C to PrP^Sc is believed to involve direct contact between the two PrP isoforms. As PrP^Sc formation occurs on the cell surface within sphingolipid-enriched microdomains, termed rafts, and/or along the endocytic pathway (9, 51, 63), reduced levels of PrP^C on the cell membrane could impair PrP^Sc formation. Thus, the antiprion activity of DS-500 could be related to its effect on cell surface PrP^C. However, the finding that undersulfation of endogenous GAGs had an opposite effect on membrane-associated PrP^C argues that high cell surface PrP^C levels do not generally correlate with increased PrP^Sc levels. In most cell lines, PrP^C is localized in rafts prior to its internalization via clathrin-coated pits (64, 65). Thus, NaClO₃ treatment could potentially also affect PrP^C cell membrane distribution, the endocytosis route, and the kinetics of endocytosis, all of which could influence PrP^Sc formation. Alternatively, changes in the intracellular distribution or trafficking kinetics of PrP^Sc and endogenous GAGs could contribute to the antiprion activity of GAG modulators. In vitro studies demonstrate a direct effect of GAGs and GAG analogues on PrP^Sc amplification, potentially by providing a scaffold for PrP^C-PrP^Sc interactions and conversion (13, 45, 66). The observed reduction in endogenous GAGs could thus directly affect the conversion process and thereby abrogate prion propagation. At least in vitro, GAG mimetics have also been shown to affect the association of PrP with factors stimulating conversion, such as RNA (22). If and where PrP associates with nucleic acid in the cell and how GAGs might modulate this interaction remain to be established.

We conclude that DS-500 and sodium chlorate could affect PrP^Sc biogenesis downstream of PrP^Sc uptake by (i) altering the subcellular distribution and levels of PrP^C, (ii) changing the levels and localization of endogenous sulfated glycosaminoglycans, or (iii) directly interfering with the conversion process. A better understanding of the subcellular compartments of prion propagation and the molecular process of PrP^Sc formation will help to define the exact mechanism by which GAGs affect prion biogenesis.