An Antibody to the Aggregated Synthetic Prion Protein Peptide (PrP106–126) Selectively Recognizes Disease‐Associated Prion Protein (PrP^Sc) from Human Brain Specimens

Abstract

Human prion diseases are characterized by the conversion of the normal host cellular prion protein (PrP^C) into an abnormal misfolded form [disease‐associated prion protein (PrP^Sc)]. Antibodies that are capable of distinguishing between PrP^C and PrP^Sc may prove to be useful, not only for the diagnosis of these diseases, but also for a better understanding of the molecular mechanisms involved in disease pathogenesis. In an attempt to produce such antibodies, we immunized mice with an aggregated peptide spanning amino acid residues 106 to 126 of human PrP (PrP106–126). We were able to isolate and single cell clone a hybridoma cell line (P1:1) which secreted an IgM isotype antibody [monoclonal antibody (mAb P1:1)] that recognized the aggregated, but not the monomeric form of the immunogen. When used in immunoprecipitation assays, the antibody did not recognize normal PrP^C from non‐prion disease brain specimens, but did selectively immunoprecipitate full‐length PrP^Sc from cases of variant and sporadic Creutzfeldt–Jakob disease and Gerstmann–Straussler–Scheinker disease. These results suggest that P1:1 recognizes an epitope formed during the structural rearrangement or aggregation of the PrP that is common to the major PrP^Sc types found in the most common forms of human prion disease.

INTRODUCTION

Human prion diseases are fatal neurodegenerative disorders that occur in idiopathic forms [sporadic Creutzfeldt–Jakob disease (sCJD)] as familial disorders associated with mutations in the prion protein gene (PRNP) [fatal familial insomnia (FFI), Gerstmann–Straussler–Scheinker Syndrome (GSS), familial CJD](26), and can also be acquired either iatrogenically through contaminated surgical instruments and bioproducts (2) or orally through the consumption of infected human tissue (kuru) (7) or meat products contaminated with the bovine spongiform encephalopathy agent [variant CJD (vCJD)]3, 36. These diseases are characterized by the conversion of the normal cellular prion protein (PrP^C) from its α‐helical rich structure to an abnormal misfolded β‐sheet rich structure, termed disease‐associated prion protein (PrP^Sc) (27). Although the details of the mechanisms involved in the initial conversion and subsequent replication of PrP^Sc are still not fully understood, this conversion process is thought by some to be the fundamental molecular event in prion disease pathogenesis.

PrP^C and PrP^Sc have the same amino acid sequence, but they differ in their physiochemical properties. Unlike PrP^C, PrP^Sc readily forms aggregates, is insoluble in non‐ionic detergents, and is partially resistant to proteinase K (PK) digestion (27). The presence of PrP^Sc is the only unambiguous prion disease marker identified to date (15), and consequently, many of the currently available diagnostic tests rely on its detection. In order to distinguish between PrP^C and PrP^Sc, many such tests typically employ limited PK digestion, which completely degrades PrP^C but leaves the PrP^Sc NH₂‐terminally truncated proteinase‐resistant core (PrP^res) intact (27). However, there is now mounting evidence to suggest that PK‐sensitive conformers of PrP^Sc exist and that these conformers may also be relevant to disease pathogenesis 23, 29, 35, 38. Reagents capable of distinguishing between PrP^C and PrP^Sc under native conditions, without the need for PK treatment, would be highly advantageous.

Monoclonal antibodies (mAbs) against PrP have been proven to be valuable in the diagnosis and in the investigation of prion disease pathogenesis (1). However, under native conditions, many of the mAbs that are currently available react only with PrP^C, for example, 3F4 (10), or react equally well with both PrP^C and PrP^Sc, for example, 6H4 (11). Although several mAbs reportedly specific for PrP^Sc under native conditions have been described 11, 17, 32, the use of these antibodies outside of the laboratories in which they were produced has, thus far, been limited.

Mice immunized with native PrP^Sc‐coated microbeads were shown to mount a predominantly IgM immune response targeting the region between PrP amino acids 101–120, suggesting that this region represented the major immunogenic region of native PrP^Sc (34). Furthermore, antibodies raised against a peptide sequence spanning amino acid residues 95–123 of human PrP have been shown to interfere with PrP^Sc propagation in prion‐infected cell cultures (16). This suggests that this region, which is highly conserved among various species, may be one of the key regions where conformational changes occur during the conversion of PrP^C to PrP^Sc. Other studies had already shown that a synthetic peptide comprising amino acid residues 106–126 of human PrP (PrP106–126) displays some of the pathogenic and physicochemical properties associated with PrP^Sc. PrP106–126 undergoes a pH‐dependent random coil to β‐sheet transformation leading to the formation of amyloid fibrils that are partially resistant to digestion with PK 5, 31. Exposure of cultured primary neurones and human neuronal cell lines to these PrP106–126 aggregates catalyses the aggregation of endogenous PrP^C to an amyloidogenic form that shares several characteristics with PrP^Sc and results in cell death 6, 33. This aggregated PrP106–126 peptide model has been used extensively to identify the intra‐ and inter‐cellular signaling events involved in prion‐associated neurotoxicity [see for example, (4)].

We reasoned that aggregated PrP106–126 and native PrP^Sc might share common conformation‐dependent epitopes that would be absent from both the monomeric peptide and normal cellular PrP^C, and that immunization and appropriate screening with aggregated PrP106–126 could be a direct route to a PrP^Sc‐specific reagent.

In this study, we report the production of one such mAb (mAb P1:1) raised against aggregated PrP106–126 which, under native conditions, can selectively immunoprecipitate full‐length PrP^Sc associated with a wide range of human prion diseases, including those characterized by the presence of type 1, type 2A, type 2B, and 8kDa PrP^res fragments. Furthermore, following PK digestion, this mAb also appears to be able to discriminate between type 1 and type 2 PrP^res, only immunoprecipitating PrP^res with a type 1 NH₂‐terminus. In addition to providing a potentially useful diagnostic and research tool, this work further validates the PrP106–126 toxic peptide model by showing that aggregated PrP106–126 and PrP^Sc do indeed share structural similarities.

MATERIALS AND METHODS

Preparation of aggregated PrP106–126

Synthetic peptides corresponding to human PrP106–126 (KTNMKHMAGAAAAGAVVGGLG) and PrP106–126‐NH₂ (KTNMKHMAGAAAAGAVVGGLG‐NH₂) were obtained from Sigma Genosys (Gillingham, Dorset, UK). Both PrP106–126 and PrP106–126‐NH₂ (2 mg/mL in 200 mM phosphate buffer, pH 7.0) were incubated at room temperature for 16 h and aggregation was monitored by turbidity measurements at 600 nm (30). Under these conditions, PrP106–126 readily formed aggregates, whereas PrP106–126‐NH₂ did not. The position of PrP106–126 in relation to full‐length PrP is illustrated in Figure 1.

Figure 1.

Schematic representation of the locations of prion protein peptide PrP106–126 (White box) and the epitopes recognized by monoclonal antibodies 12B2 (Black box A), 9A2 (Black box B), 3F4 (Black box C), and 6H4 (Black box D) within both full‐length cellular prion protein (PrP^C) and disease‐associated prion protein (PrP^Sc) and the presence of these epitopes in the proteinase‐resistant core (PrP^res) types associated with different human prion diseases.

Immunization of mice and production of mAbs

PrP^−/−null mice (13) were immunized subcutaneously with 50 μg of aggregated PrP106–126 in Complete Freunds adjuvant followed by two further booster subcutaneous immunizations of 50 µg aggregated PrP106–126 in Incomplete Freunds adjuvant at 28‐day intervals. Seven days after the final booster immunization, test bleeds were taken and the resulting serum samples were screened for antibody binding to aggregated PrP106–126‐coated microwells by enzyme‐linked immunosorbent assay (ELISA), as described below.

Wells of a 96‐well Immulon 4 HXB microtitre plate (Thermo Labsystems, Basingstoke, Hampshire, UK) were coated with 100 ng/well–aggregated PrP106–126 in phosphate‐buffered saline (PBS) (10 mM phosphate, 2.7 mM KCl, 137 mM NaCl, pH 7.4) overnight at 37°C, washed three times with PBS containing 0.05% (v/v) Tween 20 (PBST), and blot dried. The wells were then blocked [5% fetal calf serum (FCS) in PBST (200 µL/well), 60 minutes, 37°C], washed twice with PBST, and blot dried. Aliquots (100 µL/well) of test serum and pre‐immune serum from the three immunized mice (1/1000 dilution in PBST + 1% FCS) were added to triplicate wells, incubated (60 minutes, 37°C), the wells washed four times with PBST, and blot dried. Goat anti‐mouse polyvalent Ig HRP conjugate (Sigma, Gillingham, Dorset, UK) [1/2000 dilution in PBST + 1% FCS (100 µL/well)] was added to all wells. Following incubation and washing as previously described, SureBlue TMB Microwell Peroxidase Substrate (Insight Biotechnology Ltd, London, UK) (100 µL/well) was added to wells and incubated at 37°C for 30 minutes; after which, 0.18 M sulphuric acid stop solution (100 µL/well) was added and the absorbance at 450 nm was measured using a microplate reader (Dynex, Worthing, West Sussex, UK). Results were calculated as the mean absorbance 450 nm for each test sample corrected for the mean absorbance for non‐specific binding of the anti‐mouse polyvalent Ig HRP to PrP106–126 fibril‐coated wells.

The mouse chosen for hybridoma production received a final intravenous boost of 50 µg aggregated PrP106–126 in PBS and was sacrificed 4 days later. Splenocytes from the immunized mouse were fused with SP2/0 mouse myeloma cells using a conventional polyethylene glycol 1500 fusion protocol and the resulting hybridomas were selected in HAT medium. Hybridoma supernatants were routinely screened for the secretion of mAbs binding to aggregated PrP106–126 by ELISA, essentially as previously described, except that supernatant samples were screened at 1/2 dilution in PBST + 1% FCS. Hybridomas secreting mAbs binding to aggregated PrP106–126 were single cell cloned three times and frozen stocks of each cell line were laid down. The isotypes of the mAbs produced were determined using the Isostrip mouse mAb isotyping kit (Roche Diagnostics, Burgess Hill, West Sussex, UK) as per the manufacturer's instructions.

Identification of mAbs specifically binding to aggregated PrP106–126

Aliquots of each hybridoma supernatant were mixed 1:1 with PBST + 1%FCS, aggregated PrP106–126 (10 µg/mL final concentration) in PBST + 1% FCS or monomeric PrP106–126‐NH₂ (10 µg/mL final concentration) in PBST + 1% FCS, incubated on a roller mixer at room temperature for 60 minutes, and spun down at 14 000 × g for 10 minutes. Aliquots (100 µL) of each supernatant were transferred to triplicate wells of an aggregated PrP106–126‐coated microtitre plate and the ELISA was carried out as previously described. Hybridomas secreting mAbs which displayed a significant decrease in antibody binding to aggregated PrP106–126‐coated microwells following pre‐incubation with aggregated PrP106–126 but not with monomeric PrP106–126‐NH₂ were taken forward for further analysis.

Purification of IgM isotype mAbs

IgM isotype mAbs were purified from spent hybridoma culture supernatant by precipitation with 50% saturated ammonium sulphate followed by size exclusion chromatography on a Superose 6 column (Amersham Biosciences, Little Chalfont, Buckinghamshire, UK) into PBS buffer. The final concentration of the purified IgM was adjusted to 1mg/mL with PBS, 10% maltose (w/v) and 0.001% sodium azide (w/v) were added, and the purified IgM was stored in 100 µL aliquots at −40°C.

Human brain tissue

Brain tissue from patients with a range of human prion diseases and other cases initially suspected of having CJD but given an alternative final pathological diagnosis (neurological controls) were selected (see Table 1). All cases had consent for research use and the use of this material for research had local research ethics approval. The brain tissue used had been previously examined both histologically and biochemically and a definitive diagnosis had been reached by established criteria. The status of the polymorphic codon 129 [methionine homozygous (MM), valine homozygous (VV), methionine/valine heterozygous (MV)] and the presence of pathogenic mutations in the PRNP gene were determined using established methods.

Table 1.

Final diagnosis, PRNP codon 129 genotype and PrP^res type of the brain tissue sampled. Abbreviations: PRNP = prion protein gene; PrP^res = disease‐associated prion protein proteinase‐resistant core fragment; NC = neurological control; vCJD = variant Creutzfeldt–Jakob disease; MM = methionine homozygous; MV = methionine/valine heterozygous; VV = valine homozygous; FFI = fatal familial insomnia; GSS = Gerstmann–Straussler–Scheinker Syndrome; sCJD = sporadic Creutzfeldt–Jakob disease.

Sample	PRNP mutation and/or codon 129 genotype	PrPres type†	Final diagnosis
NC1*	MM	None	Vascular dementia
NC2*	MM	None	Infarction
NC3*	VV	None	Lewy body dementia
NC4*	MV	None	Vascular dementia
vCJD	MM	2B	vCJD
MM1	MM	1	sCJD MM1‡
MV1	MV	1	sCJD MV1
VV1	VV	1	sCJD VV1
MM2c	MM	2A	sCJD MM2 cortical
MM2t	MM	2A	sCJD MM2 thalamic§
MV2	MV	2A	sCJD MV2
VV2	VV	2A	sCJD VV2
FFI	D178N‐MM	None detected	FFI
GSS1	P102L‐MM	1	GSS
GSS2	P102L‐MM	8 kDa fragment	GSS

^* NC1–NC4 = non‐CJD neurological control cases.

^† Determined by western blotting following proteinase K digestion and classified according to the Parchi & Gambetti nomenclature during the original diagnosis and reconfirmed using the tissue sampled for this study.

^‡ Panencephalopathic variant of sporadic CJD.

^§ Sporadic fatal insomnia (sFI).

Immunohistochemical analysis

Immunohistochemical analysis of patterns of accumulation of disease‐associated PrP in corresponding regions of gray matter in the frontal cortex was carried out as previously described (9) on 5 µm paraffin‐embedded sections, using KG9 (TSE Resource Center, Institute for Animal Health, Compton, Berkshire, UK) and 3F4 (Dako, Ely, Cambridgeshire, UK) anti‐PrP mAbs, and a CSA amplification system (Dako) for visualization.

Immunoprecipitation experiments

In all instances, the tissue sampled for immunoprecipitation experiments was gray matter‐enriched frontal cortex. Homogenates (10% weight/volume) were prepared in extraction buffer (0.5% NP‐40, 0.5% sodium deoxycholate, Tris buffered saline, pH 7.4), clarified by centrifugation (2000 rpm, 5 minutes), and the supernatants collected. For PK digestion, PK was added to clarified homogenate at a final concentration of 50 µg/mL, incubated at 37°C for 60 minutes and digestion terminated by the addition of Pefabloc [Roche Diagnostics] (1 mM final concentration). Aliquots (10 µL) of 10% brain homogenates, both without and following PK treatment, were mixed with either 1.25 µg of mAb P1:1 or a non‐PrP‐related control IgM (Sigma) in 100 µL final volumes in 10 mM phosphate, 2.7 mM KCl, 137 mM NaCl, pH 7.4 (PBS), and incubated on a rotary mixer overnight at 4°C. Rat anti‐IgM conjugated Dynabeads (10 µL) (Invitrogen, Paisley, Scotland, UK) were added to each sample and mixed on a rotary mixer for 60 minutes at room temperature. Samples were placed in a magnetic rack to pull down the beads and the supernatants [supernatant post‐immunoprecipitation (SPI)] were collected and stored at −80°C prior to subsequent Western blot analysis [15 µL supernatant plus 5 µL 4 X NuPAGE LDS sample buffer (Invitrogen, Paisley, Scotland, UK) boiled for 10 minutes]. The beads were washed three times in PBS supplemented with 0.1% (v/v) Tween 20, resuspended in 25 µL 1 X NuPAGE LDS sample buffer and boiled for 10 minutes.

For PK digestion following immunoprecipitation, the beads were resuspended in either 10 µL extraction buffer or 10 µL extraction buffer supplemented with 50 µg/mL PK and incubated at 37°C for 60 minutes. 2 X NuPAGE LDS sample buffer (10 µL) was added and the samples were boiled for 10 minutes.

Electrophoresis and western blotting were then performed as previously described (24), except that the blots were developed using ECL Plus reagent (GE Biosciences, Little Chalfont, Buckinghamshire, UK). Unless stated otherwise, all blots were routinely probed with mAbs 3F4 (epitope amino acid residues 109–112 of human PrP) (10). Other mAbs used were 6H4 (epitope amino acid residues 144–152 of human PrP) (11), 9A2 (epitope amino acid residues 99–101 of human PrP) (12), and 12B2 (epitope amino acid residues 89–93 of human PrP) (12). All mAbs were used at the concentrations previously described (39). The relative positions of the epitopes recognized by these mAbs in both full‐length PrP^C and PrP^Sc and their presence in the PrP^res fragments associated with different human prion diseases are illustrated in Figure 1.

RESULTS

Production of mAbs against aggregated PrP106–126

PrP106–126 aggregates were formed as described and used to immunize PrP^−/− null mice. All three immunized mice mounted an immune response against aggregated PrP106–126 as determined by ELISA screening of pre‐immune and final test bleed serum samples for antibody binding to aggregated PrP106–126‐coated microwells (Figure 2A).

Figure 2.

A. Immune response of each mouse immunized with aggregated prion protein peptide (PrP106–126) as determined by serum antibody binding to aggregated PrP106–126‐coated microwells by enzyme‐linked immunosorbent assay (ELISA) for pre‐immune serum samples (white bars) and test bleed serum samples (gray bars). The increase in absorbance values obtained for the test bleed serum samples compared to pre‐immune serum samples was indicative of a positive immune response. B. Selectivity of monoclonal antibody (mAb) P1:1 for binding to aggregated PrP106–126 as determined by ELISA. mAb P1:1 was pre‐incubated in the absence of PrP106–126 peptide (None), in a 100‐fold molar excess of monomeric PrP106–126‐NH₂ (Monomer), in a 100‐fold molar excess of aggregated PrP106–126 (Aggregate), and then screened for binding to aggregated PrP106–126‐coated microwells. Inhibition of binding was detected following pre‐incubation with a 100‐fold molar excess of aggregated PrP106–126 but not with an equivalent amount of the monomeric peptide.

Out of the 360 wells seeded, post‐fusion of splenocytes from Mouse No. 3 to SP2/0 mouse myeloma cells, only one well was identified as containing actively dividing hybridoma cells secreting antibody which bound to aggregated PrP106–126 but not to monomeric PrP106–126 (Figure 2B). This hybridoma cell line (named P1:1) was single cell cloned, frozen stocks were laid down and deposited with the European Collection of Cell Cultures (ECACC, Salisbury, Wiltshire, UK) in accordance with the Budapest Treaty. The IgM isotype mAb secreted by this cell line (named mAb P1:1) was subsequently purified from spent culture medium. Both the cell line (P1:1) and mAb (mAb P1:1) are subject to an international patent application (PCT/GB2007/002205).

Neuropathological findings

In each case, the human brain tissue used in this study was selected based on an original diagnosis made according to standardized clinical, genetic, biochemical, and pathological criteria. Immunohistochemistry for PrP^Sc showed characteristic patterns of accumulation in the frontal cortex in each sCJD subtype: synaptic/granular accumulation in MM1, MV1, VV1, and MM2 thalamic; perivacuolar in MM2; cortical, perineuronal, and plaque‐like in VV2; and synaptic/granular, perineuronal, and plaque‐type in MV2. Multicentric plaques were present in the GSS cases [although less frequent in the case with type 1 PrP^res (GSS1)] and florid plaques, amorphous deposits, and cluster plaques were present in the vCJD case, but very little staining was present in the FFI case. In the non‐CJD cases, apparent upregulation of PrP^C was detectable, particularly in the cases with cerebral ischemic damage, as we have previously reported (14). The MM1 case of sCJD was a case of panencephalopathic CJD, with severe cerebral and cerebellar atrophy and white matter degeneration, while the MV1 case showed typical clinical and pathological features. The major immunohistochemical features observed conform to previous findings 20, 34, 37 and are illustrated in Figure 3.

Figure 3.

Immunohistochemistry for disease‐associated prion protein, using the KG9 prion protein (PrP) antibody with a hematoxylin counterstain, showing different patterns of accumulation in the frontal cortex (original magnification × 200 in each case). (A) Granular/synaptic positivity in sporadic Creutzfeldt–Jakob disease (sCJD) MV1 subtype; (B) Perivacuolar positivity in sCJD MM2c subtype; (C) Perineuronal/granular positivity in sCJD VV2 subtype; (D) Florid plaques, amorphous deposits and small cluster plaques in variant CJD; (E) Multicentric plaques in Gerstmann–Straussler–Scheinker Syndrome (GSS) (case GSS1, see Table 1); (F) Faint irregular staining suggestive of upregulation of cellular prion protein at the edge of a cerebral infarct in the case of vascular dementia (NC4, see Table 1).

Immunoprecipitation of PrP^Sc and PrP^res from vCJD brain tissue

mAb P1:1 was shown to immunoprecipitate PrP from the non‐PK‐treated vCJD brain homogenate (Figure 4A, Lane 7), but not from either non‐PK‐treated neurological control brain (Figure 4A, Lane 5) nor from the PK‐treated vCJD brain homogenates (Figure 4A, Lane 8). The presence of PrP^C and PrP^res in these latter two homogenates was confirmed when the respective SPIs were analyzed (Figure 4B, Lanes 5 and 8, respectively). In contrast, when a control (non‐PrP‐related) IgM isotype antibody was used instead of mAb P1:1, only trace levels of PrP were immunoprecipitated from the non‐PK‐treated vCJD brain homogenate (Figure 4A, Lane 3). Similar levels of PrP^C and PrP^res were detected in the respective non‐PK‐treated neurological control brain (Figure 4B, Lanes 1 and 5) and PK‐treated vCJD brain (Figure 4B, Lanes 4 and 8) SPIs following immunoprecipitation with the control IgM (Figure 4B, Lanes 1–4) and mAb P1:1 (Figure 4B, Lanes 5–8). However, the amount of PrP detected in the non‐PK treated vCJD brain SPI following immunoprecipitation with mAb P1:1 was significantly depleted (Figure 4B, Lane 7) when compared to the corresponding SPI following immunoprecipitation with the control IgM (Figure 4B, Lane 3). These results were consistent with a selective interaction between P1:1 and a PrP species present only in the non‐PK‐treated vCJD brain homogenate.

Figure 4.

A. Immunoprecipitation, as determined by western blotting using monoclonal antibody (mAb) 3F4, of cellular prion protein (PrP^C) from neurological control (case NC1, see Table 1) and variant Creutzfeldt–Jakob disease‐associated prion protein (PrP^Sc) and its proteinase‐resistant core (PrP^res) from non‐proteinase K treated (−PK) and proteinase K treated (+PK) brain homogenates by mAb P1:1 and a non‐PrP control IgM antibody. B. Western blot detection of PrP^C, PrP^Sc, and PrP^res remaining in the supernatants (SPIs) obtained following immunoprecipitation with mAb P1:1 and the non‐PrP related control IgM antibody. C. Determination of the PK resistance of the PrP immunoprecipitated from non‐PK‐treated vCJD brain homogenate by mAb P1:1 and the non‐PrP‐related IgM. Following immunoprecipitation, the beads were either resupended in extraction buffer (−PK) or extraction buffer supplemented with 50 µg/mL PK (+PK) and incubated at 37°C for 60 minutes. PrP^Sc and PrP^res in the samples were then determined by western blotting using mAb 3F4.

To investigate the properties of the PrP species immunoprecipitated from the non‐PK‐treated vCJD brain homogenate, immunoprecipitation was repeated using both mAb P1:1 and the control IgM. The resulting immunoprecipitates were either left undigested or digested with PK (50 µg/mL, 60 minutes, 37 C) prior to western blotting. In this experiment, little or no PrP was detected in either sample following immunoprecipitation with the control IgM (Figure 4C, Lanes 1 and 2). In contrast, mAb P1:1 efficiently immunoprecipitated PrP (Figure 4C, Lane 3) which was predominantly PK‐resistant (Figure 4C, Lane 4). Based on these results, it was apparent that under native conditions mAb P1:1 selectively bound a PrP species which was PK‐resistant; we assume this species to be full‐length PrP^Sc.

Immunoprecipitation of PrP^Sc and PrP^res associated with other human prion diseases

The above results showed that mAb P1:1 selectively immunoprecipitated full‐length PrP^Sc, but not NH₂‐terminally truncated PrP^res from a vCJD brain homogenate. We therefore decided to test the ability of mAb P1:1 to immunoprecipitate the PrP^Sc and PrP^res types that occur in different conformational isoforms and morphological deposits associated with other human prion diseases (see Table 1, Figure 3). The PrP^res types associated with these diseases were classified at diagnosis and reconfirmed by western blotting following PK digestion of the tissue homogenates used in this study (results not shown). Additionally, a series of four relevant non‐CJD neurological controls were included in this part of the study to test the selectivity of the P1.1/PrP^Sc interaction. The results obtained are shown in Figure 5, in which each panel (A–E) represents the results obtained from one set of immunoprecipitation experiments carried out in parallel. The same vCJD brain homogenate (−/+) PK digestion was included in each set of immunoprecipitation experiments, to act as a positive control and to demonstrate reproducibility. In all five immunoprecipitation experiments, mAb P1:1 immunoprecipitated full‐length PrP^Sc but not PrP^res from the vCJD brain homogenate. The band running at between 25–30 kDa observed for PK‐treated vCJD brain homogenate in Figure 5C is most likely the result of cross‐reactivity between the secondary antibody and mouse antibody light chain.

Figure 5.

A–E. Immunoprecipitation, as determined by western blotting using mAb 3F4, of cellular prion protein from non‐proteinase K treated (−PK) neurological control brain homogenates (NC1–NC4, see Table 1 for the final diagnosis associated with these individuals) and disease‐associated prion protein and proteinase‐resistant core from the corresponding non‐PK‐treated (−PK) and PK‐treated (+PK) brain homogenates from individuals with confirmed human prion diseases (see Table 1 for further details).

No PrP was immunoprecipitated from the neurological control NC1, NC2, and NC3 brain homogenates. The band running at between 25–30 kDa observed for NC3 (Figure 5C) is again consistent with mouse antibody light chain. However, a weak signal was seen in the non‐PK‐treated NC4 brain homogenate (Figure 5D), although this was not found to be reproducible in subsequent experiments using the same homogenate/tissue (data not shown).

When sCJD cases were tested, full‐length PrP^Sc was immunoprecipitated from the non‐PK‐treated brain homogenates from individuals diagnosed with sporadic CJD of all subtypes that contained type 2 PrP^res (MM2c, MM2t, MV2, and VV2 sCJD). However, no PrP^res was immunoprecipitated from any of the corresponding PK‐treated brain homogenates (Figure 5A,B). In contrast, both full‐length PrP^Sc and PrP^res were immunoprecipitated from the respective non‐PK‐ and PK‐treated brain homogenates from cases of MM1, MV1, and VV1 sCJD (Figure 5C,D). Based on the western blot staining intensities for all three type 1 sCJD cases, the amount of PrP^Sc immunoprecipitated from the non‐PK‐treated brain homogenates appeared to be greater than the amount of PrP^res immunoprecipitated from the corresponding PK‐treated brain homogenates. Therefore, a proportion of the PrP^Sc immunoprecipitated by P1:1 appears to have been PK sensitive.

When cases of inherited human prion disease were examined, no detectable PrP was immunoprecipitated from the FFI brain homogenate (Figure 5C), whereas in GSS (P102L‐MM) both full‐length PrP^Sc and PrP^res were immunoprecipitated from the respective non‐PK‐ and PK‐treated brain homogenates (Figure 5E). The PrP^res immunoprecipitated from the PK‐treated GSS1 brain homogenate displayed a similar pattern to the PrP^res immunoprecipitated from the MM1, MV1, and VV1 sCJD brain homogenates while the PrP^res immunoprecipitated from the PK‐treated GSS2 brain homogenate migrated at a molecular weight of approximately 8 kDa.

To further characterize the 8 kDa GSS fragment, we conducted western blot epitope mapping of PK digested extracts prepared from GSS2 (8 kDa fragment), MM1 sCJD (type 1 PrP^res), and vCJD (type 2 PrP^res) brain homogenates using mAbs 12B2, 9A2, 3F4, and 6H4 (Figure 6). mAbs 9A2 and 3F4 detected type 1 PrP^res, type 2 PrP^res, and the 8 kDa GSS fragment (Figure 6 Lanes 4–5 for 9A2 and Lanes 7–9 for 3F4, respectively). In contrast, mAb 6H4 detected both type 1 and type 2 PrP^res (Figure 6, Lanes 10 and 11, respectively), but not the 8 kDa GSS fragment (Figure 6, Lane 12), suggesting that COOH‐terminal cleavage had removed this epitope from the 8 kDa GSS fragment. However, mAb 12B2 detected both the 8 kDa GSS fragment and type 1 PrP^res associated with MM1 sCJD (Figure 6, Lanes 1 and 3, respectively), but not type 2 PrP^res associated with vCJD (Figure 6, Lane 2). This latter observation suggested that both the 8 kDa GSS fragment and type 1 PrP^res had similar NH₂‐terminii resulting from NH₂‐terminal cleavage at a site distinct from that found in type 2 PrP^res. A summary of the outcomes of the immunoprecipitation of PrP^Sc and PrP^res from human brain specimens using mAb P1:1 is shown in semi‐quantitative form in Table 2.

Figure 6.

Epitope mapping of type 1 proteinase‐resistant core (PrP^res) associated with MM1 sporadic Creutzfeldt–Jakob disease (Lanes 1, 4, 7, and 10), type 2 PrP^res associated with variant CJD (Lanes 2, 5, 8, and 11), and the 8 kDa fragment associated with Gerstmann–Straussler–Scheinker Syndrome (Lanes 3, 6, 9, and 12) by western blotting following limited proteinase K digestion of the respective brain homogenates. The monoclonal antibodies used for epitope mapping were 12B2 (epitope amino acid residues 89–93 of human PrP), 9A2 (epitope amino acid residues 99–101 of human PrP), 3F4 (epitope amino acid residues 109–112 of human PrP), and 6H4 (epitope amino acid residues 144–152 of human PrP).

Table 2.

Summary of the results obtained by immunoprecipitation with mAb P1:1. Abbreviations: mAb = monoclonal antibody; PrP^res = proteinase‐resistant core fragment; PrP^Sc = disease‐associated prion protein; CJD = Creutzfeldt–Jakob disease; FFI = fatal familial insomnia; MM = methionine homozygous; MV = methionine/valine heterozygous; VV = valine homozygous; GSS = Gerstmann–Straussler–Scheinker Syndrome; sCJD = sporadic Creutzfeldt–Jakob disease; sFI = sporadic fatal insomnia; vCJD = variant Creutzfeldt–Jakob disease.

Diagnosis and genotype of cases analyzed	PrPres type determined by routine western blot	Immunoprecipitation
		Full‐length native PrPSc	Proteinase K digested PrPres
Non‐CJD	None	−	−
FFI (D178N‐MM)	None	−	−
GSS (P102L–129MM)	~8 kDa	++	+
	1	++	+
sCJD (MM, MV, VV)	1	++	+
sCJD (MM, MV, VV)	2A	++	−
sCJD/sFI (MM)	2A	++	−
vCJD (MM)	2B	++	−

Key: ++ = abundant; + = detectable; − = undetectable.

DISCUSSION

We have isolated a hybridoma cell line secreting an IgM isotype mAb (mAb P1:1) which binds to both aggregated PrP106–126 and PrP^Sc associated with vCJD, but does not recognize monomeric PrP106–126 or PrP^C, even in neurological control cases in which PrP^C is upregulated in association with ischemic damage. Based on immunoprecipitation experiments, mAb P1:1 bound PrP^Sc, which was predominantly PK resistant, from a non‐PK‐treated vCJD brain homogenate; however, no PrP^res was immunoprecipitated from the corresponding PK‐treated brain homogenate. This observation is similar to that previously made for mAb 15B3 (11), an mAb raised against full‐length recombinant bovine PrP, which is reported to specifically bind PrP^Sc. mAb 15B3 appeared to immunoprecipitate intact PrP^Sc more effectively than PrP^res, possibly resulting from PK‐dependant formation of large aggregates (scrapie‐associated fibrils) that might mask the 15B3 epitope. Although this explanation might also be true for mAb P1:1, other possible explanations needed to be considered: For example, PK digestion prior to immunoprecipitation might alter the conformation of PrP^res compared to full‐length PrP^Sc, thus disrupting the conformation‐dependent epitope recognized by mAb P1:1. To investigate these possibilities, we decided to expand the mAb P1:1 immunoprecipitation experiments to include PrP^Sc and PrP^res associated with other human prion diseases and encompassing a wide range of forms of PrP^Sc as judged by immunohistochemistry and western blot typing of PrP^res.

Two major isoforms of PrP^Sc associated with human prion diseases, which probably reflect two distinct PrP^Sc conformations, have been identified 18, 19. These two types can be distinguished by the extent of NH₂‐terminal truncation and the gel mobility of their protease‐resistant core fragment following PK digestion. Type 1 PrP^res has its NH₂‐terminus at amino acid residue 82 and a gel mobility for the unglycosylated isoform of 21 kDa, whereas type 2 PrP^res has its NH₂‐terminus at amino acid residue 97 and a gel mobility for the unglycosylated isoform of 19 kDa. These two PrP^Sc types, in conjunction with the PRNP codon 129 genotype, are proposed to underlie the phenotypic variability associated with sCJD and provide a molecular basis for disease classification (MM1, MM2, VV1, etc) (21). vCJD PrP^res also has a type 2 NH₂‐terminal; however, vCJD PrP^Sc is classified as type 2B based on the differences in the glycoform ratios observed between vCJD PrP^Sc and the type 2 PrP^res associated with sCJD 8, 19.

A further conformational type is found in GSS. PrP^res associated with GSS linked to the PRNP codon 102 mutation (GSS P102L) can exist in two forms that differ in both size and glycosylation 20, 25. The first form, which is glycosylated, displays similar gel mobility to that of the type 1 PrP^res associated with sCJD and is found associated with spongiform change in the brain (25). The second form is an 8 kDa unglycosylated fragment resulting from both NH₂‐ and COOH‐terminal truncation of the PrP^Sc and is found in conjunction with multicentric amyloid plaques in the brain. The cases of GSS used in this study conform to this model to a certain extent in that GSS1 had spongiform change present in all layers of the cerebral cortex and prominent plaques, but GSS2 had some plaques in the lower layers of the cortex and a lesser degree of spongiform change. Both types of GSS PrP^res are reported to have a type 1‐like NH₂‐terminal, resulting predominantly from cleavage at amino acid residue 82, and we have confirmed this for the 8 kDa fragment detected in our experiments using mAb 12B2 (12), which binds to an epitope situated between amino acid residues 89–93 of human PrP. This epitope would be present in type 1 PrP^res but would be cleaved from type 2 PrP^res. We show here that each of these differing isoforms of PrP^Sc found in sCJD, vCJD, and GSS are all recognized in their full‐length form and can be efficiently immunoprecipitated by mAb P1:1 (see Table 2 for a summary).

Although full‐length PrP^Sc associated with the human diseases could be successfully immunoprecipitated with mAb P1:1, this was not true for all types of the corresponding PrP^res. mAb P1:1 did not immunoprecipitate PrP^res with a type 2 NH₂‐terminus. This would suggest that following PK digestion of type 2 PrP^Sc, the resulting PrP^res underwent a change in conformation such that the epitope recognized by mAb P1:1 was no longer present or available. In contrast, following PK digestion of type 1 PrP^Sc, the resulting PrP^res retained the conformation recognized by mAb P1:1. Indeed, such a change in conformation, thought to be influenced by the primary PK cleavage site, has been described, appearing to be the result of an apparent reshuffling of the residual protein structure (28). It is interesting to note that type 1 PrP^res would still contain one of the octapeptide repeat regions, whereas type 2 PrP^res would not. One could speculate that interactions between the octapeptide repeat regions of PrP molecules could stabilize PrP^Sc in a specific conformation which is recognized by mAb P1:1, thus allowing efficient immunoprecipitation of full‐length PrP^Sc. Following PK digestion, this conformation would be conserved in type 1 PrP^res as a result of the interaction between the remaining octapeptide repeat regions; however, for type 2 PrP^res, the loss of all of the octapeptide regions would lead to a change in conformation so that type 2 PrP^res was no longer recognized by mAb P1:1.

There is a clear difference between the amount of PrP^Sc and type 1 PrP^res immunoprecipitated from the MM1, MV1, and VV1 sCJD samples, with much more PrP^Sc than PrP^res detected; however, it has been reported that up to 80% of the PrP^Sc associated with sCJD is actually PK sensitive (29). P1:1 therefore presents an opportunity to capture and further biochemically define these hypothesized protease‐sensitive forms of PrP^Sc. Moreover, P1.1 immunoprecipitation may find a diagnostic application in cases of human prion disease where PrP^res levels are intrinsically low or where tissue availability is limited.

The levels of PrP^Sc in FFI are known to be lower than in other human prion diseases and the involvement of the cerebral cortex has been proposed to be a function of disease duration (22). The diagnosis of FFI in the case used in this study was based on clinical and neuropathological criteria in association with D178N mutation coupled with methionine at codon 129 of the PRPN gene (37), and intensive western blot investigation of this particular brain, including sampling from the thalamus, and the use of concentration techniques have consistently failed to demonstrate detectable PrP^res. mAb P1:1 also failed to immunoprecipitate PrP^Sc from this particular case and we judge this negative result to be significant in that it argues against the presence of substantial quantities of protease‐sensitive PrP^Sc as defined by mAb P1:1 immunoreactivity. This result is in sharp contrast to the efficient immunoprecipitation of the abundant type 2 PrP^Sc associated with the case of the thalamic variant of sCJD (sCJDMM2t), otherwise known as sporadic fatal insomnia.

PrP^Sc‐specific antibodies, while desirable for their specificity, suffer from an intrinsic limitation. Given that PrP^Sc is a conformationally altered and aggregated form of PrP^C, it follows that PrP^Sc‐specific epitopes will themselves be conformation‐dependant and/or discontinuous. Consequently, one would predict that a truly PrP^Sc‐specific mAb would fail to react with denatured materials such as those prepared for western blotting. Moreover, the pretreatments commonly used to enhance PrP immunostaining with standard PrP antibodies (guanidine treatment, antigen retrieval by hydrolytic autoclaving, and so forth) would also be predicted to disrupt PrP^Sc‐specific conformational epitopes. It is therefore not unexpected that mAb P1.1 does not recognize PrP in western blotting or when used in routine immunohistochemical assays (data not shown). Nevertheless, we can infer from the immunoprecipitation assays shown here that P1:1 may find application in assays of native materials in platforms such as ELISA, flow cytometry, or specially adapted histological techniques.

In conclusion, we have developed an mAb (mAb P1:1) that can selectively immunoprecipitate full‐length PrP^Sc of different conformations that are associated with a wide range of human prion diseases. We now intend to evaluate the use of mAb P1:1 in this and other assays for the detection of human prion disease associated PrP^Sc in a wide range of tissues and body fluids.