The accumulation of PrPSc in scrapie‐infected neuronal cells has been prevented by three approaches: (i) transfection of ScMNB cells with an antisense laminin receptor precursor (LRP) RNA‐expression plasmid, (ii) transfection of ScN2a cells and ScGT1 cells with small interfering RNAs (siRNAs) specific for the LRP mRNA, and (iii) incubation of ScN2a cells with an anti‐LRP/LR antibody. LRP antisense RNA and LRP siRNAs reduced LRP/LR expression and inhibited the accumulation of PrPSc in these cells. The treatments also reduced PrPc levels. The anti‐LRP/LR antibody, W3, abolished PrPSc accumulation and reduced PrPc levels after seven days of incubation. Cells remained free of PrPSc after being cultured for 14 additional days without the antibody, whereas the PrPc level was restored. Our results demonstrate the necessity of the laminin receptor (LRP/LR) for PrPSc propagation in cultured cells and suggest that LRP/LR‐specific antibodies could be used as powerful therapeutic tools in the treatment of transmissible spongiform encephalopathies.
Transmissible spongiform encephalopathies (TSEs) are a group of neurodegenerative disorders that include Creutzfeldt–Jakob disease (CJD) in humans, bovine spongiform encephalopathy (BSE) in cattle and scrapie in sheep (Aguzzi & Weissmann, 1998; Prusiner et al., 1998; Weissmann, 1999; Lasmézas & Weiss, 2000). The main pathogenic event in the development of TSEs is the conversion of PrPC , the normal cellular form of the prion protein, to PrPSc . An important feature of PrPSc is its partial resistance to proteases, which makes it biochemically distinguishable from PrPC (Caughey & Raymond, 1991). Recently, we identified the laminin receptor (LRP/LR) as the cell‐surface receptor for PrPC (Gauczynski et al., 2001a). Heparan sulphate proteoglycans (HSPGs) have been shown to function as cofactors or co‐receptors for the binding of PrPC to the LRP/LR (Hundt et al., 2001). The LRP/LR has been shown to interact directly with PrPC in the yeast two‐hybrid system (Rieger et al., 1997). This interaction was confirmed by pull‐down assays in cotransfected COS‐7 cells and co‐infected insect cells (Rieger et al., 1997). Furthermore, increased levels of the LRP were found in the brain, spleen and pancreas of scrapie‐infected mice and hamsters, as well as in scrapie‐infected neuroblastoma cells, which are a well characterized in vitro model for scrapie infection (Rieger et al., 1997). These data suggest a link between the LRP/LR and prion propagation.
The non‐integrin LRP/LR is a multifunctional protein that is required for cell differentiation, movement and growth (for review see Gauczynski et al., 2001b). The LRP cDNA encodes a 37‐kDa precursor protein (LRP), also known as p40, and has been cloned from different species by several groups. This protein has been reported to be ribosome‐associated, to bind to the histones H2A, H2B and H4 and to be the precursor of the metastasis‐associated 67‐kDa mature high‐affinity laminin receptor (LR) (for a review, see Gauczynski et al., 2001b; Leucht & Weiss, 2002). The 67‐kDa LR is consistently upregulated in aggressive carcinomas, suggesting a role in cell homeostasis and cohesion. The amino‐acid sequence of the receptor is highly conserved throughout evolution, with at least 98.3% homology between the mouse, human and bovine sequences, and 99% homology between the rat and human sequences (for a review, see Gauczynski et al., 2001b; Leucht & Weiss, 2002). Published data suggest the existence of at least six LR genes in the mouse genome; one of these is localized on chromosome nine, and at least two copies are thought to be functional (Douville & Carbonetto, 1992). Using TRIBE–MCL, an algorithm for the detection of protein families (Enright et al., 2002), five LR genes were identified when the program was used to search the latest mouse draft genome sequence (Mouse Genome Sequencing Consortium, 2003, available at http://www.ensembl.org). The LRP gene on chromosome nine has seven exons and six introns but, in contrast with earlier results (Douville & Carbonetto, 1992), no LRP/LR gene on chromosome six has been identified. Interestingly, genes that affect susceptibility to prions have been identified on mouse chromosome nine (Stephenson et al., 2000).
PrP‐specific antibodies have successfully been used in preventing prion propagation in vitro and in vivo as follows: first, the accumulation of PrPSc in scrapie‐infected neuroblastoma cells was inhibited by PrP‐specific antibodies (Peretz et al., 2001); second, scrapie infection was abolished by transgenic expression of PrP‐specific antibodies in mice (Heppner et al., 2001). The epitope recognized by the antibody that has the most potent effect on PrPSc , D18, consists of amino‐acid residues 132–156 of PrP, which includes helix A (residues 144–154). Because PrP residues 144–179 have been shown to constitute a binding site for the LRP/LR (Hundt et al., 2001) we investigated whether an antibody directed against the LRP/LR, the cellular receptor for PrPC (Gauczynski et al., 2001a), can also be used to interfere with the metabolism of PrPSc . To ablate LRP/LR expression from all putative LRP/LR‐encoding genes we used an antisense RNA and a small interfering RNA (siRNA) approach. We investigated whether these strategies had an effect on prion propagation in several scrapie‐infected cell systems.
Results and Discussion
Antisense LRP RNA prevents PrPSc propagation
To produce LRP antisense messenger RNA, we cloned a region of LRP complementary DNA from nucleotide positions 65–901 into the expression plasmid pCI‐neo in the antisense orientation to produce the pCI–neo–asLRP plasmid. After transient transfection of pCI–neo–asLRP into ScMNB cells we confirmed antisense LRP RNA expression in these cells (Fig. 1A). The level of LRP mRNA was greatly reduced 48 h after transfection (Fig. 1B). Using phosphoimaging, this reduction was quantified and LRP mRNA levels were found to be 80–85% of normal LRP mRNA expression levels. A similar reduction in target mRNA levels has been shown in other studies that have used the antisense RNA method to downregulate the expression of myelin basic protein (Katsuki et al., 1988; 80% reduction) and Wnt‐1 (Erickson et al., 1993; up to 98% reduction). At the level of protein expression, no LRP protein was detected by western blotting 48 h after transfection (Fig. 1C). Analysis of cells 72 h after transfection showed an absence of PrPSc propagation (Fig. 1D) in cells with reduced LRP levels (Fig. 1C). Levels of PrPSc were unaffected in cells transfected with pCI–neo as compared with untransfected cells (Fig. 1D). In ScMNB cells we were able to detect only the diglycosylated form of PrP using the SAF70 antibody, whereas in ScN2a and ScGT1 cells we observed the classic three‐band pattern. We observed a reduction in PrPC level after antisense LRP RNA transfection (Fig. 1D), which might be caused by an altered PrPC metabolism. Previous studies have indicated that PrPC internalization is highly dependent on the presence of the LRP/LR at the cell surface (Gauczynski et al., 2001a), where the LRP/LR binds PrPC through two distinct domains: the octapeptide region and the region encompassing amino acids 144–179 of PrPC (Hundt et al., 2001). This is consistent with a recent study, in which it was found that the octapeptide region is essential for internalization of PrPC (Nunziante et al., 2003). Hence, the altered PrPC levels seen in this study are likely to be due to perturbed metabolism of the protein.
LRP‐specific siRNAs prevent PrPSc propagation
SiRNAs were used to verify the results obtained using the LRP antisense RNA construct. This method has been used succesfully in other studies to knock down target‐gene expression levels (Elbashir et al., 2001). We tested four different LRP‐specific siRNAs for their ability to repress LRP expression in ScN2a cells. All of them repressed LRP synthesis (Fig. 2A). Figure 2Bshows data from a time‐course experiment carried out to analyse the effect of siRNA‐LRP3 on PrPSc propagation in ScN2a cells. Seventy‐two hours after transfection, PrPSc propagation was completely abolished by siRNA‐LRP3, whereas siRNA‐LRP1, siRNA‐LRP4 and a control siRNA (lamin A/C, described in Elbashir et al., 2001) had a smaller effect (siRNA‐LRP1+ siRNA‐LRP4) or no effect (control) on PrPSc levels. PrPC levels were reduced in the presence of siRNA‐LRP3. The same effects were observed with LRP antisense RNA 72 h after transfection. In contrast to PrPSc , PrPC levels increased 96 h after transfection, probably due to a decrease in siRNA effectiveness with time.
We also tested the efficiency of the reduction of LRP expression using siRNAs in ScGT1 cells, which show a robust PrPSc phenotype (that is, these cells propagate PrPSc over a long period of time). The results were consistent with those obtained using ScN2a cells, with a strong reduction of PrPSc correlated with LRP downregulation (Fig. 2C).
Anti‐LRP/LR antibody W3 prevents PrPSc accumulation
LRP/LR‐specific antibodies have been used successfully to compete with recombinant prion proteins for binding to the LRP/LR in different mammalian cell types (Gauczynski et al., 2001a), showing that the LRP/LR has a crucial role in the metabolism of PrPC . Using the LRP/LR‐specific antibody, W3 (Rieger et al., 1997), in ScN2a cells we observed a reduction of PrPSc to undetectable levels (Fig. 3A, B). The antibody was used at concentrations of 6–64 μg ml−1 . At a concentration of 12 μg ml−1 a reduction in PrPSc levels was observed. At a higher concentration (64 μg ml−1 ), PrPSc accumulation was completely abolished after incubation for three days, indicating a dose‐dependent effect (Fig. 3A). In a timecourse experiment, we found a complete clearance of PrPSc after incubation for one week, using an antibody concentration of 32 μg ml−1 (Fig. 3B). These results are consistent with a previous study, in which different anti‐PrP antibodies were used to reduce PrPSc levels in cultured cells (Peretz et al., 2001; Table 1). In that study, PrP antibody concentrations of 1.2–10.0 μg ml−1 were sufficient to clear PrPSc from ScN2a cells after one week of incubation (Table 1).
We also incubated ScN2a cells in which PrPSc had been previously cleared by W3 for a further two weeks without any antibody, and showed that no PrPSc reappeared (Fig. 3B). PrPC levels in W3‐treated cells were reduced after seven days of incubation with W3, but were completely restored after a further two‐week incubation in the absence of the antibody (Fig. 3B).
Role of the LRP/LR in PrPSc propagation in cultured cells
The knock down of the LRP/LR on the cell surface by LRP antisense RNAs or by siRNAs, and the blockage of LRP/LR binding sites by the W3 anti‐LRP/LR antibody are most likely to interfere with PrP levels by blocking the PrP internalization process. However, some PrPC can still be synthesized and transported through the secretory pathway to the cell surface (Figs 1D, 2B and 3). Conversion of PrPC into PrPSc is thought to take place either at the cell membrane or in the endocytic pathway. Thus, it is possible that due to the lack of PrPC within the endocytic pathway no PrPSc can be formed, resulting in a time‐dependent reduction in PrPSc (Figs 2B and 3B). It is also possible that the LRP/LR has a function in the conversion of PrPC to PrPSc , and that the absence of the LRP/LR from the cell surface affects PrPSc formation. PrPSc propagation cannot be restored after cessation of the incubation with anti‐LRP/LR antibody (Fig. 3B) due to the absence of any PrPSc to re‐initiate the conversion process. In contrast, PrPC levels were completely restored after cessation of incubation with the anti‐LRP/LR antibody (Fig. 3B). Furthermore, depletion or blockage of the LRP/LR on the cell surface might directly prevent PrPSc binding and internalization. In summary, our results show that the LRP/LR is not only involved in PrPC metabolism, as demonstrated in previous reports (Gauczynski et al., 2001a; Hundt et al., 2001), but also has a crucial role in prion propagation. The fact that LRP/LR‐specific antibodies are able to clear PrPSc from neuroblastoma cells provides possibilities for the development of new experimental therapies for TSEs.
Construction of pCI–neo–asLRP.
Bases 65–901 of the LRP cDNA were amplified by PCR with reverse transcription (RT–PCR) from total RNA isolated from N2a cells, introducing the restriction sites Nhe I and Sma I. The LRP cassette was cloned in an antisense orientation into the plasmid pCI–neo using the Nhe I and Sma I sites to produce pCI–neo–asLRP. Cloning was confirmed by sequencing.
ScMNB and ScN2a cells (both lines are neuroblastoma cells chronically infected with scrapie) were grown in DMEM, 10% fetal bovine serum, 2 mM Glutamax (Invitrogen), 100 units ml−1 penicillin and 10 μg ml−1 streptomycin sulphate, at 37 °C with 5% CO2. ScN2a cells were produced as described previously (Bosque & Prusiner, 2000). ScGT1–7 cells (GT1 hypothalamic neuronal cells, chronically infected with the Chandler scrapie isolate) were provided by S. Lehmann, and were cultured as described previously (Mange et al., 2000), with the exception that DMEM was replaced with Opti‐MEM (Gibco Life Sciences).
Inhibition studies using the W3 antibody.
ScN2a cells (1 × 106)were incubated in normal growth medium (DMEM, 10% fetal bovine serum, 2mM Glutamax) supplemented with the purified polyclonal anti‐LRP/LR antibody, W3, at varying concentrations. After incubation with the antibody, the cells were harvested, lysed and analysed by western blotting.
Inhibition studies using LRP antisense RNA.
ScMNB cells were grown in six‐well plates to 60% confluence. Cells were transfected with pCI–neo–asLRP and pCI–neo (control plasmid) using Lipofectamine (Invitrogen) in accordance with the manufacturer's instructions. Transfection efficiencies were determined using a chloramphenicol acetyltransferase construct, and were estimated to be approximately 80% on average (data not shown). Cells were harvested 72 h after transfection, lysed and analysed by western blotting.
Inhibition studies using small interfering RNAs.
Four pairs of complementary 21‐nucleotide RNAs corresponding to regions of the LRP cDNA were made (Ambion). As a control, the lamin A/C RNA duplex was used (Elbashir et al., 2001). The single‐stranded complementary RNAs were annealed in annealing buffer (provided by the manufacturer) for 1 min at 90 °C, followed by incubation for 1 h at 37 °C. The RNA duplexes were transfected into ScN2a cells (cultured in Opti‐MEM medium, Invitrogen) using Oligofectamine (Invitrogen) in accordance wih the manufacturer's instructions. ScGT1‐7 cells were seeded in 60‐mm petri dishes (5 × 105 cells per dish) and transfected the following day with 10 μg of each of the 21‐nucleotide RNA pairs using Exgen 500 (Fermentas) in accordance with the manufacturer's instructions.
Ribonuclease protection assays.
Total RNA was purified from transfected ScMNB cells and used in a ribonuclease protection assay using the RPA III kit (Ambion). An antisense riboprobe was made by in vitro transcription from pCI–neo–asLRP, following linearization of the plasmid with Eco RI, in the presence of [α32 P]UTP. The antisense riboprobe was combined with the total RNA and the mixture was then precipitated. The precipitates were dissolved in hybridization buffer, denatured and hybridized with the total RNA. This was followed by incubation with RNAse for 30 min at 37 °C, followed by inactivation of the RNAse and ethanol precipitation of the RNA. Protected RNA fragments were separated on a 5% acrylamide/urea gel and visualized using a Storm 860 phosphorimager equipped with ImageQuant software.
Total RNA was purified from transfected ScMNB cells and cDNA synthesis was carried out using an oligo(dT) primer in an RT reaction. The resulting cDNA was then amplified by PCR using a 5′‐oligodeoxyribonucleotide corresponding to the 3′‐end of the cytomegalovirus promoter and a 3′‐oligodeoxyribonucleotide corresponding to a sequence in the 5′‐region of the simian virus 40 polyadenylation signal. PCR products were separated on 1% agarose gels and stained with ethidium bromide.
Cytoplasmic lysates were made using a buffer containing 10 mM Tris–HCl, pH 7.5, 100 mM NaCl, 10 mM EDTA, 0.5% Triton X‐100 and 0.5% sodium deoxycholate. After centrifugation, the total protein content of the lysates was measured (BCA‐Protein Assay, Pierce) and equal amounts of protein from each lysate were analysed. For PrPSc detection, cell lysates were digested with proteinase K (20 μg ml−1)for 1 h at 37 °C. The reaction was stopped by the addition of Pefabloc (1 mM) and the proteins were denatured using 6 M guanidine hydrochloride. Samples were boiled in SDS sample buffer and analysed on an SDS–polyacrylamide gel containing 12.5% acrylamide. For PrPC or PrPSc detection (from ScN2a cells), 10% Bis–Tris gels with MES running buffer (NuPAGE, Invitrogen) were used. Proteins were blotted on a polyvinylidene difluoride membrane, blocked and incubated overnight with the monoclonal antibodies SAF70, SAF32 or SAF84 (diluted 1:5,000 in blocking solution) or A7 (diluted 1:2,500 in blocking solution) for PrP detection. The polyclonal anti‐LRP/LR antibody, W3 (Rieger et al., 1997) (1:2,000), or the monoclonal antibody 43512 (1 μg ml−1)were used for LRP/LR detection and anti‐β actin antibody (Chemicon) (1:5,000) for β‐actin detection. After washing with TBS/0.05% Tween 20 the blot was incubated for 1 h with a peroxidase‐conjugated secondary antibody (Sigma) (1:2,500). Detection was carried out by enhanced chemiluminescence (Western Lightning, NEN).
We acknowledge support by grants QLRT‐2000‐02085 and FAIR‐CT‐98‐7020 from the European Union. S.W. acknowledges support by grants 01‐KO‐0106, 0313012 (both Bundesministerium für Bildung u. Forschung), LMU 3 and LMU 4 (Bavarian Prion Research Foundation). We thank J. Grassi for providing us with the SAF70, SAF32 and SAF84 antibodies, A. Bürkle for ScMNB cells, H.M. Schätzl for the A7 antibody, and K. Krüger, A. Pahlich, K. Töpolt, S. Janetzky and S. Hengge for excellent technical assistance. C.L. thanks M. Buschbeck for helpful discussions. S.W. thanks R. Grosschedl and E.‐L. Winnacker for valuable advice and continuous support.
- Copyright © 2003 European Molecular Biology Organisation