Plasminogen-Based Capture Combined with Amplification Technology for the Detection of PrPTSE in the Pre-Clinical Phase of Infection

Background Variant Creutzfeldt-Jakob disease (vCJD) is a neurodegenerative infectious disorder, characterized by a prominent accumulation of pathological isoforms of the prion protein (PrPTSE) in the brain and lymphoid tissues. Since the publication in the United Kingdom of four apparent vCJD cases following transfusion of red blood cells and one apparent case following treatment with factor VIII, the presence of vCJD infectivity in the blood seems highly probable. For effective blood testing of vCJD individuals in the preclinical or clinical phase of infection, it is considered necessary that assays detect PrPTSE concentrations in the femtomolar range. Methodology/Principal Findings We have developed a three-step assay that firstly captures PrPTSE from infected blood using a plasminogen-coated magnetic-nanobead method prior to its serial amplification via protein misfolding cyclic amplification (PMCA) and specific PrPTSE detection by western blot. We achieved a PrPTSE capture yield of 95% from scrapie-infected material. We demonstrated the possibility of detecting PrPTSE in white blood cells, in buffy coat and in plasma isolated from the blood of scrapie-infected sheep collected at the pre-clinical stage of the disease. The test also allowed the detection of PrPTSE in human plasma spiked with a 10−8 dilution of vCJD-infected brain homogenate corresponding to the level of sensitivity (femtogram) required for the detection of the PrPTSE in asymptomatic carriers. The 100% specificity of the test was revealed using a blinded panel comprising 96 human plasma samples. Conclusion/Significance We have developed a sensitive and specific amplification assay allowing the detection of PrPTSE in the plasma and buffy coat fractions of blood collected at the pre-clinical phase of the disease. This assay represents a good candidate as a confirmatory assay for the presence of PrPTSE in blood of patients displaying positivity in large scale screening tests.


Introduction
The prion disease variant Creutzfeldt-Jakob (vCJD) was first identified in the UK in 1996 and was shown to represent the human counterpart of bovine spongiform encephalopathy (BSE), a consequence of the entry of contaminated beef products into the human food chain in the 1980s [1]. Up to August 2012, 227 clinical cases had been reported across 12 countries, 176 of which were in the UK alone, 27 in France and 24 in the other 10 countries [2]. Although the incidence of food-borne vCJD is declining, secondary transmission of vCJD through blood transfusion or tissue grafts continues to pose a genuine risk to public health. Several studies have now demonstrated the efficient transmission of Transmissible Spongiform Encephalopathy (TSE) by blood in non-human primates [3] and in sheep [4,5]. Using TSE-infected sheep models, several teams have confirmed that all blood components can transmit the infectious agent through blood transfusion into healthy sheep [5][6][7]. Moreover, O. Andreoletti et al. demonstrated that the transmission was highly efficient, as 0.2 ml of infected whole blood was sufficient to elicit the disease in sheep [8]. On the issue of transfusion risk, the United Kingdom (UK) reported five secondary cases, four of which (three clinical and one subclinical) were likely associated with the transfusion of non leukoreduced red blood cell concentrates. The fifth case concerned a patient treated with clotting factor FVIII manufactured from the plasma of a donor who developed vCJD six months after donating in 1996. However, the true size of the reservoir of asymptomatic carriers, all of whom represent potential blood donors, remains undetermined. Recently, taking into account the existence of healthy carriers of vCJD, Gharske T. et al predicted the number of vCJD cases associated with secondary transmission in the UK over the coming years will exceed the number of cases of primary transmission via contaminated food [9,10]. Interim data from a repeat appendix survey have reported 16 positives out of 32441 samples leading to a prevalence of 1 in 2000 persons in the UK population [11]. The introduction of a vCJD detection assay could provide valuable information regarding the true prevalence of pre-clinical cases and help to prevent further cases resulting from blood transfusion transmission.
The major event in this disease is a conformational change of the normal cellular prion protein (PrP C ) into an infectious misfolded isoform (PrP TSE ) which accumulates as macromolecular assemblies in the brain [12]. At variance with sporadic forms of Creutzfeldt-Jakob disease, vCJD PrP TSE accumulates not only in the brain but also in lymphoid organs [13,14], and is probably present in biological fluids.
A femtomolar level of sensitivity -i.e. 0.1 pg/mL or 10 infectious doses per mL (ID/ml) -has been estimated from the hamster model, as the minimum required to detect PrP TSE in the plasma of a donor at the preclinical phase of infection [15]. This concentration is very difficult to detect using classical PrP TSE detection methods.
Accordingly, we have been focusing our attention on an amplification technology called protein misfolding cyclic amplification (PMCA) developed by C. Soto et al. [16]. In essence, PMCA facilitates the conversion of PrP C substrate in the presence of low or undetectable amounts of PrP TSE , to achieve levels visible by conventional laboratory methods such as immunoblotting. This is achieved by subjecting samples to repeated and alternating cycles of incubation thought to enlarge the PrP TSE seed, and sonication thought to fragment it and generate new catalytic units. A serial PMCA method has been optimized for the high-efficiency amplification of PrP TSE from different animal or human TSEinfected samples [17][18][19][20][21][22]. However, direct amplification of blood PrP TSE by PMCA has proved difficult, mostly due to the difficulty in obtaining even minute quantities of prions, which generally requires large sample volumes, but also to the presence of bloodassociated conversion inhibitors that interfere with the amplification [18,23]. To circumvent this, a pre-analytical step is needed to capture PrP TSE from blood samples, in a specific manner, before amplification. Magnetic beads can be coated with plasminogen which has been reported to bind preferentially PrP TSE from multiple species including sheep and human [24][25][26]. Plasminogen also has demonstrated the ability to stimulate prion conversion in vitro [27].
We designed a specific PMCA-based confirmatory test for the detection of PrP TSE in blood, which comprises three steps: 1) plasminogen-coated nanobeads capture of PrP TSE from blood components; 2) PrP TSE amplification by serial PMCA; and 3) specific detection of PrP TSE by immunoblotting. As the number and volume of vCJD blood samples are very limited in the UK and in France, assay validation was first undertaken on sheep blood samples collected at pre-clinical or clinical stages of scrapie. The sensitivity and specificity of the test were then determined on human vCJD spiked-plasma panels provided by the National Institute of Biological Standards and Control (NIBSC). After optimization of the different steps, this test reached both the 100% specificity and the sensitivity levels required for the detection of infectivity in asymptomatic carriers.
Anonymized human whole blood samples were collected in EDTA collection tubes (Etablissement Français du Sang -Pyrénées Méditerranée). Donor written consent had been obtained for their use in research in compliance with French Law (code de la santé publique article L.1243-3) concerning Blood and Tissue Samples for Non therapeutic Use. Since our study uses plasma only as diluent, it does not belong to the field of IRB competence as defined in the French regulation, and therefore does not require the approval of the Bioethical Review Board. Plasma was then isolated and recovered after centrifugation at 1500 x g for 15 min at room temperature (RT).
Whole blood was also collected from healthy (4) and scrapieinfected sheep at preclinical (PG127 isolate at 120 days after oral challenge) (1) or terminal (natural scrapie) (4) stages of the disease. Sheep blood samples were obtained from O.Andreoletti, (Institut National de la Recherche Agronomique/Ecole Nationale Vétérinaire de Toulouse -France) as part of a research agreement. The sheep expressed the V136R154Q171 allele of ovine PrP. Sheep white blood cell (SWBC) samples were prepared from the buffy coat (BC) fraction, obtained after centrifugation at 2000 x g for 15 min at Room Temperature (RT). Residual red cells were eliminated from the buffy coat in buffer composed of 155 mM NH 4 Cl, 10 mM KHCO 3 , and 1 mM EDTA (pH 8). Aliquots of 10 7 phosphate buffer saline (PBS)-cleared SWBC were stored at -80uC.
Spiked plasma samples were prepared as follows: 500 ml of healthy human donor plasma were spiked with serial tenfold dilutions (ranging from 10 22 to 10 210 ) of either vCJD 10% IBH from affected patients or scrapie 10% 127S IBH (127S strain = mouse adapted PG127 isolate) from ovine transgenic mice (tg338 line).

Prion Protein Capture
The use of human plasminogen as an efficient ligand for the capture of prion proteins has been reported by Fischer et al. [24]. Accordingly, we used super para-magnetic nanobeads activated with carboxylic acid functionality (for bead specifications see ref 0211, Ademtech -France), which were coated with human plasminogen (Fluka Sigma-Aldrich -France) by shaking for 2 h at 37uC. The optimal ligand quantity to be used was evaluated by testing three plasminogen concentrations: 10, 20, and 30 mg/mg of beads (10 mg/mg being the lowest concentration recommended by the manufacturer). After a blocking step with 0.5 mg/ml albumin solution at 37uC, beads were stored at 4uC in suspension at 1% (w/v) in the Ademtech storage solution.
Firstly, we evaluated the optimal bead quantity to be used for each sample by testing several volumes of 1% bead suspension (2.5, 5, 10, 20, 30, 60 and 90 ml). After washing with PBS, the magnetic beads bearing PrP TSE were isolated and PrP TSE protein bound on beads was amplified by PMCA.
Different volumes of healthy and infected sheep plasma and buffy coat were mixed with the LB buffer (final volume 1 ml) before incubation with the determined quantity of plasminogencoated beads at RT for 90 min as described above.
For PrP TSE capture from SWBC samples, the cells were first incubated on ice for 30 min with LB. After centrifugation at 1000 g at 4uC for 2 min the supernatant was harvested and mixed with LB (final volume 1 ml) before incubation with plasminogencoated nanobeads as described above.

Protein Misfolding Cyclic Amplification (PMCA)
PrP C sources for PMCA were from normal brain homogenates (NBH) from either human PrP (M 129 allele, tg650 line) [29] or ovine PrP (VRQ allele, tg338 line) [30] transgenic mice. These mice overexpress the level of PrP C by 6 and 8 fold respectively, compared to wild-type mice. Brains were collected and prepared as previously described [21]. Briefly after collection, brains were rinsed in cold PBS and immediately frozen on dry ice before longterm storage at 280uC. Brains were then homogenized, using a potter-Elvehjem homogenizer on ice, at a 10% (wt/vol) concentration in the lysis/converting buffer (CB) composed of 150 mM NaCl, 1%Triton X-100 and protease inhibitor cocktail (Roche) in PBS (pH 7.2). Homogenates were centrifuged at 2000 x g for 20 seconds and frozen at 280uC in single-experiment aliquots.
Captured prion protein was first mixed with 90 ml of 10% NBH. Then, amplification by PMCA was performed using the Misonix 4000 (Misonix, N.Y., USA). Each cycle is composed of an incubation step (37uC) and a sonication step. Beforehand, the following PMCA parameters were optimized: incubation duration per cycle (30 and 60 min), sonication duration per cycle (20 and 40 s), power level (60, 70, 80%) and the number of PMCA cycles (50, 80,100) per round.
For sample analysis, after a tenfold dilution of the amplified samples with fresh NBH, a second and a third round of PMCA was performed.

Proteinase K (PK) Digestion and SDS-PAGE/ immunoblotting
Methods were performed as previously described [32]. Briefly, after bead removal, amplified products were incubated at 45uC with PK (200 mg/ml) for 60 min, before their denaturation at 100uC in SDS-PAGE sample buffer. Samples were run on 12% NUPAGE gels and electrotransferred onto PVDF membrane. Western blotting, using the SNAP system (Millipore, St-Quentinen-Yvelines, France), was performed with 3F4 or 6D11MAb as anti-PrP monoclonal antibodies (MAb) (Signet/Proteogenix, 67412 Illkirch, France) for human and sheep prion detection respectively, and an anti-mouse IgG peroxidase-linked secondary antibody for chemiluminescent reaction (ECL reagent, GE-Healthcare France).
The capture yield was calculated as follows: the amount of PrP res captured by plasminogen-beads was compared to the total amount of PrP res present in the input after acquisition of the chemiluminescent western blotting signals with the GeneGnome digital imager and quantification with the GeneTools software (Syngene, Frederick, Maryland, USA). Three independent experiments were performed.

National Institute for Biological Standards and Control (NIBSC) Panels
The NIBSC has made available a series of reference reagents prepared from autopsied human brain specimens (www.nibsc.ac. Figure 1. Ovine PMCA optimization. Tg338 127S infected brain homogenate dilutions (10 24 to 10 210 ) were tested using two rounds (one round = 80 cycles) of PMCA and detection was performed on PK-digested amplified products using western blot analysis with 6D11 as the primary antibody. Molecular weight markers are shown on the right. NC: negative control, NBH only. uk/cjd/brainsamples.html). The brain reference reagents used in this study were obtained from the CJD Resource Center in the form of 10% (wt/vol) homogenates in 0.25 M sucrose. Two panels were tested. Panel 1 was composed of ten randomly chosen normal plasma samples distributed in duplicate (a total of 20). Panel 2 was a blinded panel composed of duplicates of vCJD brain dilutions in normal plasma, and negative controls including plasma spiked with normal brain and negative plasma alone (96 tubes).

Results
Optimization of the Plasminogen-based Capture System and Serial PMCA Technology using Scrapie-infected Brain Material The PMCA parameters were optimized using normal and scrapie-infected brain homogenates from ovine PrP transgenic mice (tg338 line, VRQ allele) in the PMCA reaction. PMCA was first evaluated by comparing cycle numbers, incubation, sonication and power parameters of the Misonix. Finally, after two PMCA rounds of 80 cycles each (30 min incubation, 20 s sonication, 80% power) a PrP TSE specific signal was reproducibly detected up to 10 29 dilution (n = 3 from two different IBH) indicating a 7 log amplification compared to the signal obtained for the non-amplified 10 22 dilution (F10 22 ) (Fig. 1). No signal was detected when normal brains (about 50 samples throughout this study) were processed similarly from different NBH samples.
We evaluated the effect of PrP C overexpression on the PMCA efficacy of 10 25 and 10 26 IBH dilutions using three different dilutions of tg338 brain substrates: undiluted 10% tg338 NBH and 1:6 and 1:8 dilutions of 10% tg338 NBH in 10% PrP2/2 NBH. Only undiluted 10% tg338 NBH, demonstrated maximum amplification efficacy for both 10 25 and 10 26 IBH dilutions. Using the dilution representative of 'physiological' expression levels of PrP C found in sheep i.e. 1:8, no signal was obtained for the 10 26 IBH dilution and a very faint signal was detected for the 10 25 IBH dilution (Fig. 2).
Prior to PMCA, PrP TSE needs to be captured from blood samples and concentrated. The optimum plasminogen concentration and bead volume for PrP TSE capture were firstly calculated using a series of plasminogen-coated nanobeads. Scrapie IBH at 10 23 and 10 24 dilutions was diluted in human plasma (500 ml) and mixed with various volumes of beads with different plasminogen concentrations before PMCA amplification. Representative results are shown in Fig. 3A and B. It is worth noting that lower signals were observed with increasing bead volume (Fig. 3 B lanes 6-8) and plasminogen concentration (Fig. 3A lanes 3, 4, 5 , 7, 8, 9, 10, 11, 12). The strongest PrP TSE signal was observed using a plasminogen (Plg) concentration of 10 mg Plg/mg of beads (Fig. 3A lanes 3) and a volume of 10 ml of the 1% bead suspension per 500 ml of spiked plasma (Fig. 3B lanes 4, 6). Under these conditions, when PrP TSE was captured from 10 ml Tg338 127S IBH or 10 ml vCJD IBH diluted in 500 ml of plasma and analyzed directly by western blot without PMCA amplification, a 95% yield (percentage of recovery calculated from PrP res signal) was observed (Fig. 4). In the same way, the PrP TSE signal obtained after one PMCA round from the 10 26 plasma dilution appeared almost similar to that obtained when a control 10 26 IBH dilution (Fig. 5  C10 26 ) was used directly to seed the PMCA substrate suggesting that the association of the capture by plasminogen-coated nanobeads and PMCA reached almost 100% efficiency (Fig. 5, lanes 10 26 and C10 26 ). Both negative controls of 500 ml of human plasma mixed with plasminogen-coated nanobeads prior to PMCA remained negative (Fig. 5 lanes NC). Effect of PrP C level on PMCA efficacy. Tg338 127S infected brain homogenate dilutions (10 25 and 10 26 ) were PMCA amplified (one round) using 10% tg338 NBH (Undil.), 10% tg338NBH diluted 1/6 in 10% tg-PrP 0/0 NBH and 10% tg338NBH diluted 1/8 in 10% tg-PrP 0/0 NBH as PMCA substrates. The detection was performed on PK-digested amplified products using western blot analysis with 6D11 as the primary antibody. Undil.: 10% tg338 NBH without dilution in 10% tg-PrP 0/0 NBH. NC: negative control, NBH only. doi:10.1371/journal.pone.0069632.g002

Validation of the Plasminogen-coated Nanobeads -sPMCA Combined Assay on a Panel of Blood Samples from Healthy and Infected Sheep
We next examined whether our experimental conditions could specifically detect PrP TSE in scrapie-infected blood samples.
In four white blood cell concentrates of naturally scrapieinfected sheep (SWBC), 0/4 and 1/4 samples were detected positive respectively after one and two rounds of PMCA (results not shown). After a third round, a PrP TSE signal was obtained from all four infected SWBC amplified samples, in contrast to those from healthy sheep (0/4) (Fig. 6). These results were Figure 3. Capture optimization of ovine brain PrP TSE . 3A: Batches of nanobeads (1% suspension) coated with 10, 20, and 30 mg of plasminogen/mg of beads were used to test prion capture efficacy using 30, 60 and 90 ml of plasminogen-coupled beads per 500 ml of human plasma spiked with a 10 23 dilution of 127S IBH. After one round of PMCA, detection was performed on PK-digested and amplified products using western blot analysis with 6D11 as the primary antibody. Lane 1: NBH: normal brain homogenate without PK digestion Lane 2: F10 22 IBH dilution without PMCA (Frozen) Lane 6: negative control: plasma only 3B: Tg338 127S IBH dilution (10 24 ) in plasma was tested using different volumes of coated beads at 10 mg of plasminogen/mg of beads (1% suspension) for the prion capture. After one round of PMCA, detection was performed on PKdigested and amplified product using western blot analysis with 6D11 as the primary antibody. Lane 1: NBH: normal brain homogenate without PK digestion Lane 5: negative control: plasma only. doi:10.1371/journal.pone.0069632.g003  4) were diluted in 500 ml of plasma and incubated 2 h with plasminogen-coated beads. PrP TSE bound to the beads were PK-digested and denatured in sample buffer for western blot analysis with 6D11 and 3F4 anti-PrP MAbs. Percentage yield was quantified with Genetools software after acquisition of the chemioluminescent western blot signals with the Genegnome digital imager (Syngene, US) Lanes 1 and 2:127S IBH capture Lanes 3 and 4: vCJD IBH capture. doi:10.1371/journal.pone.0069632.g004 reproduced in three different assays. In one sheep at the preclinical phase of scrapie, a specific PrP TSE signal was detected in a 500 ml plasma sample after PrP capture and two rounds of PMCA (Fig. 7A) and in both 50 ml and 25 ml buffy coat samples (BC) after one PMCA round (Fig. 7B). These assays were reproduced twice and each sample was tested in duplicate.

Evaluation and Validation of the Plasminogen-coated nanobeads/sPMCA Assay on Human Samples
Similar PMCA optimizations were performed on vCJD human brain homogenates amplified with normal brain homogenates from human PrP transgenic mice (M 129 allele, tg650 line) as substrate. The optimal amplification parameters were found to be exactly the same for the human vCJD strain as for the ovine strain (80 cycles: 30 min incubation, 20 s sonication, 80% power).
After the PrP TSE capture from a panel of tenfold dilutions of vCJD IBH (10 24 to 10 28 ) in normal human plasma, the first PMCA round allowed the detection of PrP TSE at the 10 25 dilution. This indicated a 3 log amplification compared to the signal obtained for the non-amplified F10 22 dilution (Fig. 8). After a tenfold dilution of the amplified samples with fresh NBH, a second and a third round of 80 PMCA cycles was performed allowing detection of PrP TSE up to 10 26 and 10 28 dilutions respectively, corresponding to a 4 and 6 log amplification factor (Fig. 8). These sensitivity levels were confirmed three times with different IBH samples, one of which was the WHO reference reagent provided by NIBSC. It is worth noting that this 10 28 dilution detection after the third round was the same as that obtained when testing tenfold dilutions of vCJD IBH (10 24 -10 28 ) without the capture step (results not shown); this confirms a capture efficacy around 95% as shown in fig. 4.
The analytical performance of the test was then evaluated in terms of its sensitivity and specificity by testing blinded panels provided by the NIBSC. All samples from panel 1 (normal plasma) tested negative (20/20), demonstrating the test's 100% specificity. Results of the 96 spiked-plasma samples of blinded panel 2 (spiked with either brain or spleen homogenates) obtained after three rounds of PMCA were analyzed by the NIBSC and a sensitivity level of 10 25 vCJD brain homogenate in plasma was achieved (Table 1). For plasma spiked with spleen homogenate the 10 21 dilution was detected (4/4). The 36 control samples included in the same panel all tested negative. Results obtained on the 56 negative controls included in panel 1 and panel 2 confirmed the 100% specificity of the assay (56/56).   . Detection of PrP TSE from plasma and buffy coat of scrapie-infected sheep at the preclinical stage 7A: 500 ml of PG 127 infected and healthy plasma samples underwent the capture step in duplicate with 10 ml of coated beads at 10 mg of plasminogen/mg of beads, before PMCA amplification using two rounds of 80 cycles. Detection was performed after PK digestion of the amplified products using western blot analysis with 6D11 as the primary antibody. Pre-Cl-Plasma: plasma sample from PG127 infected sheep at the pre-clinical stage (120 days post oral challenge) Neg Plasma: plasma sample from healthy sheep 7B: 50 and 25 ml of PG 127 infected (120 days post oral challenge) and healthy BC (buffy coat) samples underwent the capture step in duplicate with 10 ml of coated beads at 10 mg of plasminogen/mg of beads, before PMCA amplification using one round (i.e. 80 cycles). Detection was performed after PK digestion of the amplified products, using western blot analysis with 6D11 as the primary antibody. Lane PC: 10 27 IBH dilution amplified by PMCA Lane NC: Buffy coat from healthy sheep. doi:10.1371/journal.pone.0069632.g007 Figure 8. Human PMCA optimization. vCJD IBH dilutions (10 24 to 10 28 ) in 500 ml of plasma were captured by 10 ml of coated beads at 10 mg of plasminogen/mg of beads. After three PMCA rounds (80 cycles/round) the detection was performed on PK-digested and amplified products using western blot analysis with 3F4 as the primary antibody. 22F: 10 22 IBH dilution without PMCA (Frozen) NC: negative control: plasma only. doi:10.1371/journal.pone.0069632.g008

Discussion and Conclusion
Results of this study demonstrate that the above-presented three-step PMCA assay can specifically detect PrP TSE at the preclinical stage of the disease in both plasma and buffy-coat samples from sheep. PrP TSE detection achieved after PMCA has previously been described using buffy coat fractions from rodents [33,34] and purified white blood cell preparations from sheep [22] at the terminal stage of the disease. However, PrP TSE amplification efficiency was compromised when the PMCA was performed directly on blood [18]. In a recent study, Tattum et al. [23] showed that when PMCA was performed directly on the whole blood of infected mice, the yield was reduced by approximately 50% because of the presence of inhibitory factors in the blood. To overcome this issue and decrease the consequential large input volumes of blood in the PMCA reaction, we developed a capture technology using plasminogen-a well-known prion protein ligand [24]-for implementation prior to PMCA amplification.
The specificity of PrP TSE capture by plasminogen is controversial. Using high resolution ultrasonography to study interactions between proteins, Negredo et al. [26] demonstrated that the interaction between plasminogen-coated beads and PrP TSE was specific. In addition these findings confirmed previous results obtained by Fisher et al. and Maissen et al. [24,25]. Conversely however, other studies have reported that recombinant prion proteins and PrP C purified from sheep or bovine brain could also interact with human plasminogen [35][36][37]. When coupling the capture step with PMCA, our results do not allow us to rule out the possible capture of PrP C isoforms from infected sheep WBCs in addition to PrP TSE by the plasminogen coated beads. However, residual PrP C in addition to PrP TSE does not seem to be a major problem for PMCA, since PrP C is the main substrate for PrP TSE amplification. One limitation could be the potential competition between PrP C and PrP TSE for capture on the covered nano beads. However, in this work, we have demonstrated no loss of PrP TSE during the capture step which allowed 95% capture of the sample PrP TSE before amplification.
We chose nanobeads over microbeads as the support for plasminogen [24] for their small size (100-140 nm) and correspondingly large surface-to-volume ratio [38]. We have previously demonstrated the need to strictly define the nanobead volume to be used for the capture in order to avoid a crowding effect due to an excess of coated nanobeads. Such molecular crowding affects the chemical reactions occurring in high solute concentrations since the volume occupied by one molecule is made unavailable to other molecules and so-called the excluded volume. Theoretical and experimental studies have demonstrated that the excluded volume effect is dependent on the size of molecules (39)(40)(41). Thus, the decrease in signal observed when the capture step is performed with an excess of coated beads could be explained by the high molecular weight of the plasminogen (88kDa) fixed on the beads.
In two independent trials, the addition of beads or plasminogen to the PMCA reaction stimulated the PMCA amplification [27,42]. By optimally combining the two enhancement strategies, we have achieved the sensitivity required for PrP TSE detection at the preclinical stage.
The plasminogen-based capture/sPMCA assay applied to scrapie-infected brains from ovine PrP transgenic mice (127S) allowed the detection at a 10 29 dilution, equivalent to 100 fg of 127S infected tg338 brain. This represents a 100-fold higher sensitivity level compared to that obtained by bioassay [43]. This finding is not unprecedented as PMCA detection levels well below infectivity levels have been reported for 263K prions in hamsters and chronic wasting disease (CWD) prions in transgenic mice for cervid PrP [20,44].
Applied to WBC samples from naturally infected sheep (n = 4), this assay showed 100% specificity and sensitivity. Moreover, we observed that a capture step is essential to detect PrP TSE in infectious blood samples. Indeed, when the same infected SWBC samples were tested by PMCA alone, the four samples gave negative results (data not shown). Recently, Lacroux et al. reported PrP TSE detection in SWBC collected from sheep experimentally infected by the PG127 scrapie isolate from 90 to 190 day post inoculation (dpi) [7]. In this study, we confirmed the detection of PrP TSE in sheep WBC samples from naturally infected sheep by using the same PMCA substrate (ovine transgenic mouse g338) as Lacroux et al. Therefore, we demonstrated that our plasminogen-based capture allows the detection of two phenotypically different sheep strains. It is worth noting that until now, PMCA detection of the pathological PrP in infected blood samples of animals was obtained using substrate and blood samples of close origins (species, strains and genotype) [21,22,45].
Finally this new assay can detect PrP TSE in a volume of buffy coat (BC) as low as 25 mL (around 0.5 ml of whole blood) and in 500 ml of plasma samples, collected from one sheep at the preclinical stage of scrapie (PG 127 isolate). This assay now needs to be repeated with additional sheep samples.
Coupling the capture step to PMCA also allowed the detection of PrP TSE in the 10 28 dilution of vCJD IBH spiked plasma (equivalent to 1 pg of initial brain sample). The assay reached a 100-fold higher level of sensitivity compared to that obtained in the PMCA study using platelets as substrate [46]. Based on our findings that the concentration of PrP C has an influence on the PMCA amplification ratio, this improvement in sensitivity may be explained by the use of tg650 NBH which overexpresses PrP by 6fold when compared to platelet homogenate.
When the NIBSC panel was tested by our PrP TSE capture followed by three rounds of PMCA, the detection limits achieved were 10 25 dilution of vCJD brain spiked in plasma (3/8) and 10 21 dilution of vCJD spleen spiked in plasma (4/4). It is worth noting that the 10 26 dilution of vCJD brain was also detected in the same experiment but only in one sample out of the four tested. Since the sensitivity of our assay ( fig. 8) was checked with vCJD brain homogenates from three different sources, this discrepancy might be explained either by differences in the procedures of material preparation or the infectivity levels of the brains at the time of collection. Furthermore, among six diagnostic assays involving the same blinded NIBSC panel, none have overcome this 10 25 detection limit with a 100% specificity [47].
Recently two blood test prototypes were published which allow the detection of endogenous PrP TSE in blood. Both are based on the combination of promising technologies for high throughput testing: i) a solid-state binding matrix to capture and concentrate the PrP TSE coupled with a direct immuno-detection allowing the detection of 71% samples of confirmed clinical vCJD patients [48]; and ii) an immunoprecipitation approach coupled to the quakinginduced conversion (eQuIC) technology which allows the detection of PrP TSE in the blood of hamsters at the preclinical stage of the disease [49].
The macromolecular structure and resistance to proteolysis of circulating PrP TSE as well as the distribution of infectivity among human vCJD blood compartments remain unknown. Based on rodent models, it has been estimated that blood PrP TSE concentrations at the preclinical phase of the disease could be around the femtomolar level, i.e. 0.1 pg/mL [50]. Recently, the PrP TSE concentration in scrapie-affected hamster brain at the clinical phase was estimated by quantitative PMCA to be around 10 25 gram/gram of brain [51]. Accordingly, the test developed here with 1pg IBH sensitivity could feasibly allow the detection of 10 217 g of PrP TSE (,180 molecules) and achieve the limit of detection (LOD) required for the detection of PrP TSE in vCJD plasma samples at the pre-symptomatic phase.
The next step will be validating our new assay on buffy-coat and plasma samples from individuals with confirmed clinical vCJD, however such samples are very limited in volume and in quantity. Fortunately, the test can overcome the problem of sample volume since it has been shown to be suitable for PrP TSE detection in a buffy-coat volume as low as 25 ml.
The last important issue will also be to determine whether or not amplified vCJD PrP TSE is infectious and if the amplification of PrP TSE correlates with the amplification of infectivity. These studies are ongoing.
In conclusion, we have developed a new highly specific and sensitive in vitro assay which may represent a good candidate for use as a confirmatory PrP TSE detection assay in plasma or buffy coat samples. This assay could ensure that blood donations detected positive by means of a large scale high-throughput screening test are indeed true positive.