Sensitive detection of aggregated prion protein via proximity ligation

Abstract

The DNA assisted solid-phase proximity ligation assay (SP-PLA) provides a unique opportunity to specifically detect prion protein (PrP) aggregates by investigating the collocation of 3 or more copies of the specific protein. We have developed an SP-PLA that can detect PrP aggregates in brain homogenates from infected hamsters even after a 10⁷-fold dilution. In contrast, brain homogenate from uninfected animals did not generate a detectable signal at 100-fold higher concentration. Using either of the 2 monoclonal anti-PrP antibodies, 3F4 and 6H4, we successfully detected low concentrations of aggregated PrP. The presented results provide a proof of concept that this method might be an interesting tool in the development of diagnostic approaches of prion diseases.

Keywords:

• diagnosis
• monoclonal antibody
• prion protein
• proximity ligation assay
• 263K

Abbreviations

BSE	=	bovine spongiform encephalopathy
CJD	=	Creutzfeldt-Jakob disease
CSF	=	cerebrospinal fluid
FIDA	=	fluorescence intensity distribution analysis
qPCR	=	quantitative real-time PCR
QuIC	=	quaking-induced conversion
PLA	=	proximity ligation assay
PMCA	=	protein misfolding cyclic amplification
PrP	=	prion protein
PrPC	=	cellular prion protein
PrPSc	=	scrapie prion protein
SP-PLA	=	solid phase proximity ligation assay

Prion diseases are fatal infectious neurodegenerative diseases of humans and animals.¹ During the asymptomatic phase, which can last for decades in humans, the disease can be transmitted between animals and humans. There is a critical need for early diagnostics in order to take measures to prevent the spread of prion diseases. The key event in prion disease pathogenesis is the conformational conversion of the normal cellular prion protein (PrP^C) into an aggregated multimeric, protease resistant form called the scrapie prion protein (PrP^Sc). There are several methods to distinguish and separate PrP^Sc from PrP^C. Classical methods to detect PrP^Sc require digestion with proteinase K, followed by immunoassays such as western blot or ELISA. A PrP^Sc detection assay with increased sensitivity has been reported using immuno-PCR,^2,3 where bound antibodies are detected by amplification of conjugated DNA oligonucleotides. However, due to difficulties of raising high quality antibodies that can distinguish PrP^Sc from PrP^C, proteinase K treatment is required to ensure specificity of the assay. Also, proteinase K-sensitive disease-related species of PrP have been found^4,5 and it is therefore important to develop new sensitive and specific methods to detect misfolded PrP without the need for treatment with proteinase K. The fact that infectious PrP can propagate itself has been the basis for development of methods that demonstrate the presence of aggregating protein via in vitro amplification of the misfolded protein. Protein misfolding cyclic amplification (PMCA)⁶ is a method where the accumulation of aggregated PrP^Sc from monomers is triggered by the presence of PrP^Sc in a sample, serving to seed the conversion of PrP^C from brain extracts into large amounts of PrP^Sc. Several assays have been developed based on PMCA for detection of PrP^Sc in blood,^7-10 but the assays are not easily adapted to clinical settings due to the slow and complex procedure, high levels of false positives, and potential health hazards. The quaking-induced conversion (QuIC) assay similarly uses recombinant PrP^C that is induced by PrP^Sc to convert into amyloid fibrils, but in a shorter amount of time compared to PMCA.^11,12 QuIC has been used for detection of PrP^Sc in human cerebrospinal fluid (CSF)¹³ to distinguish patients with Creutzfeldt-Jakob disease (CJD) from healthy controls or patients with other neurodegenerative diseases with 100% specificity and a sensitivity around 80%. The sensitivity of the assay has been further enhanced by improving the QuIC protocol and adding an immunoprecipitation step with the PrP^Sc selective antibody 15B3^14,15 prior to the assay.¹⁶ Other ways to specifically capture and separate PrP^Sc from other sample components and from PrP^C include binding to a polymeric compound (Seprion ligands),¹⁷ a method that has been the basis for the development of immunoassays sufficiently sensitive to detect PrP^Sc in whole blood from humans with variant CJD.¹⁸ Yet another method to separate PrP^Sc from PrP^C is through precipitation with sodium phosphotungstate (NaPTA).¹⁹ This procedure has been combined with a technique based on fluorescence intensity distribution analysis (surface-FIDA) to detect PrP^Sc in CSF from cattle with bovine spongiform encephalopathy (BSE)²⁰ and blood plasma from scrapie-infected sheep.²¹ The FIDA assay detects PrP^Sc based on the fact that they are aggregates of large numbers of copies of a protein. A fluorescence labeled monoclonal antibody is allowed to bind the target and only when many antibodies are bound in close proximity, i.e. to the same aggregate, will this generate a detectable signal:²² A monoclonal antibody can only bind once per monomeric PrP^C, but several antibodies can bind to aggregates of PrP^Sc, leading to a concentration of fluorophores on PrP^Sc that can be detected using fluorescence correlation spectroscopy over the randomly distributed antibodies in solution or bound to PrP^C.

Herein we describe a sensitive method to detect aggregated PrP, which is based on the principle that amplifiable reporter DNA molecules are only generated when 3 copies of a monoclonal antibody bind 3 or more identical epitopes in proximity, such as by binding an aggregate of a target protein. The specific detection of aggregated PrP is based on the solid-phase proximity ligation assay (SP-PLA).^23,24 In SP-PLA targeted proteins are first captured on a solid support via immobilized antibodies before addition of 2 PLA probes, that is antibodies with conjugated oligonucleotides, followed by washes, ligation of oligonucleotides on pairs of antibodies that have bound in proximity, and amplified detection by quantitative real-time PCR (qPCR). If all 3 affinity reagents required for detection are directed against the same epitope, then the assay can be used to detect aggregated forms of a protein recognized by the antibody. We have previously demonstrated the utility of this assay mechanism to detect Aβ-oligomers or protofibrils, thought to herald the onset of Alzheimer disease, by using a single Aβ-protofibril-specific monoclonal antibody for all 3 binding events in SP-PLA.²⁵

We established a SP-PLA protocol for detection of aggregated PrP using either monoclonal antibodies 3F4²⁶ or 6H4,¹⁴ both well known to recognize the PrP protein (Fig. 1). Briefly, a monoclonal antibody was immobilized on magnetic beads and used for enrichment of PrP from biological samples. The same monoclonal antibody was also coupled to 2 different DNA oligonucleotides in separate reactions and this pair was allowed to bind captured proteins. Only aggregates of 3 or more PrP subunits can host 2 DNA-coupled antibodies with different, ligatable sequences after capture, as required to generate a signal in the assay.

Figure 1. Schematic illustration of SP-PLA. (A) Captured antibodies are immobilized on paramagnetic beads. (B) When the sample is incubated with the beads, the targeted PrPs are captured. (C) The same clone of antibody is combined with streptavidin-DNA oligonucleotide conjugates to create PLA probes that are incubated with the captured targets. (D) Oligonucleotides attached to antibodies that have bound in close proximity to each other are joined by enzymatic ligation in the presence of a connector DNA oligonucleotide. (E) Only ligated oligonucleotide reporter molecules can be amplified, detected, and quantified by real-time quantitative PCR.

It has previously been shown that PrP^C and PrP^Sc exhibit different accessible epitopes according to their respective confirmations, and that treatment with denaturing agents such as guanidine-HCl increases the reactivity of certain PrP^Sc epitopes contrarily to PrP^C.^19,27,28 This phenomenon was observed with our SP-PLA protocol, increasing the sensitivity of this technique (Fig. 2a). As the SP-PLA requires triple recognition events our observations suggest that denaturation with high concentrations of guanidine-HCl would induce incomplete disaggregation of PrP^Sc as previously described,^29-31 with maintenance of at least trimers or small aggregates. No significant change in signal was detected upon denaturation of brain homogenate from uninfected hamster (Fig. 2b).

Figure 2. Effect of denaturation of hamster brain homogenates with guanidine-HCl (Gnd-HCl) prior to SP-PLA. (A) A homogenate from an infected hamster brain was incubated with different concentrations of Gnd-HCl and with a control without Gnd-HCl. Dilution series of the homogenates were analyzed for PrP^Sc through SP-PLA using the monoclonal antibody 3F4. (B) The effect of Gnd-HCl was also examined for a 100-fold dilution of infected and uninfected brain homogenate. Plotted are the mean cycle numbers at threshold (Ct) with standard deviations from duplicate experiments.

As a proof-of-concept experiment we used the assay to compare brain homogenates from hamsters infected with the prion strain 263K and from uninfected control animals. Using the 3F4 antibody, we could successfully distinguish between brain homogenate from hamster infected with strain 263K and brain homogenate from uninfected hamster, as the infected sample gave detection signals at 100-fold lower concentration than the sample obtained from uninfected hamster (Fig. 3). To extend our results we assayed the same brain homogenates in parallel with brain homogenate from a different infected hamster and from 2 more uninfected hamsters. The results show that the assay clearly distinguished both infected samples from all 3 uninfected controls at a 10 million-fold dilution (Fig. 4a). Using the antibody 6H4 we could similarly distinguish infected lysates from those from uninfected controls when both were assayed at a 1 million-fold dilution (Fig. 4b).

Figure 3. Distinguishing PrP^Sc aggregates in brain homogenate from an infected hamster vs. an uninfected control hamster. Dilutions of homogenates from brains of infected or uninfected hamsters in buffer were analyzed for the presence of PrP^Sc using SP-PLA with the 3F4 monoclonal antibody. A cut-off value was set at 3 standard deviations above the mean signal from a negative control (only buffer). Plotted in the graph are the ratios of reporter DNA molecules in the assayed samples relative to this cut off. The data shown are the mean values of 3 replicates with standard deviations.

Figure 4. Distinguishing PrP^Sc aggregates in brain homogenate from several infected and uninfected hamsters. (A) The 3F4 SP-PLA was used to detect PrP^Sc aggregates in brain homogenates from 2 infected hamsters and from 3 uninfected hamsters diluted 10 million times. (B) An SP-PLA based on the monoclonal antibody 6H4 was used to detect PrP^Sc aggregates in the same samples diluted 1 million times. Each sample was assayed in duplicates. Plotted are the mean cycle numbers at threshold (Ct) with standard deviations.

The results prove the validity of the concept of using SP-PLA for sensitive detection of aggregated PrP by requiring several antibody-epitope binding events per target aggregate. Here, we successfully developed assays using either the monoclonal antibodies 3F4 or 6H4, suggesting that the assays could also be adapted for detection of aggregated PrP from other species using the appropriate monoclonal antibodies. The SP-PLA also provides a novel method to study the actual aggregation state of different prion strains treated with different denaturants or other possible decontaminants.

For each reaction we only use 5 μl of sample. The capture to a solid support also allows the assay to be adapted for analyses of larger sample volumes in order to further increase the sensitivity, as has previously been used to improve the sensitivity of both immunoassays^18,32 and QuIC¹⁶ for detection of PrP^Sc, and PLA-based detection of Aβ protofibrils.²⁵ We have previously shown that SP-PLA shows similar assay performance in whole blood as in buffer or plasma.²⁴ The only instrumentation required for the assay is a qPCR machine, which represents standard laboratory instrumentation. We therefore believe that the assay given further optimization will be suitable for routine analyses of samples from body fluids such as CSF or blood, allowing ante mortem testing.

Materials and Methods

Reagents

The monoclonal antibody 3F4²⁶ was biotinylated using the ChromaLink™ biotin labeling kit (B-9007-105K, Solulink). The reaction was performed with 100 μl 3F4 antibody at a concentration of 0.9 mg/mL, and a 20-fold molar excess of ChromaLink biotin over the antibodies. The biotinylated monoclonal antibody 6H4 (01-043) was purchased from Prionics.

The oligonucleotide­streptavidin conjugates SLC1 (streptavidin­5′­CGCATCGCCCTTGGACTACGACTGACGAACCGCTTTGCCTGACTGATCGCTAAATCGTG­3′) and SLC2 (5′­TCGTGTCTAAAGTCCGTTACCTTGATTCCCCTAACCCTCTTGAAAAATTCGGCATCGGTGA-3′-streptavidin) were custom ordered from Solulink. Prior to use the streptavidin-oligonucleotide conjugates were treated with a 3-fold molar excess of free streptavidin (S4762, Sigma-Aldrich) at 65°C for 30 min as previously described.^24,33

The connector oligonucleotide (5′-TACTTAGACACGACACGATTTAGTTT-3′) serving as template for the ligation, the forward (5′­CATCGCCCTTGGACTACGA-3′) and reverse (5′-GGGAATCAAGGTAACGGACTTTAG-3′) primers were all purchased from Integrated DNA Technology. The TaqMan probe (FAM-5′­TGACGAACCGCTTTGCCTGA-3′-MGB[quencher]) was purchased from Applied Biosystems.

Preparation of hamster samples

Brains from 263K-infected hamsters were sampled at the terminal stage of the disease. Brains from the infected and uninfected animals were homogenized at 20% (weight/volume) in a 5% glucose solution.

Denaturing the samples

The brain homogenate samples were denatured by mixing the sample with guanidine-HCl solution 8 M in H₂O (G9284, Sigma-Aldrich) at a 1:1 ratio. When titrating the concentration of guanidine-HCl the samples were mixed 1:1 with 8M, 4M, and 2M guanidine-HCl in H₂O respectively. The samples were incubated at room temperature (RT) for ten minutes before they were further diluted at least 10-fold in PLA buffer (1 mM D-biotin (B-1595, Life Technologies), 0.1% purified bovine serum albumin (A7888, Sigma-Aldrich), 0.05% Tween 20 (P9416, Sigma-Aldrich), 100 nM goat IgG (I5256, Sigma-Aldrich), 0.1 μg/μl salmon sperm DNA (15632-011, Life Technologies), 5 mM EDTA (E5134, Sigma-Aldrich), 1xPBS).

Solid-phase proximity ligation assay

SP-PLA was performed as previously described.^24,34 Briefly, 1.5 μg of the biotinylated antibody – either 3F4 or 6H4 – was immobilized on 1 mg streptavidin-coated magnetic beads (Dynabeads MyOne Streptavidin T1; 65602, Life Technologies). Five μl of denatured brain homogenate sample was diluted ten times in PLA buffer and thereafter mixed with the beads to allow capture of the target molecules. After 1.5 h incubation under rotation at RT the beads were separated from the supernatant using a DynaMag-96 Side magnet (65602, Life Technologies). After washing once with 100 μl 0.02% N-lauroylsarcosine sodium salt solution (61747, Sigma-Aldrich) in 1xPBS, and twice with 100 μl washing buffer (1xPBS, 0.05% Tween 20), 250 pM of the pair of PLA probes in PLA buffer were added. These probes consisted in the same monoclonal antibody as that used for capture mixed at a 1:1 ratio with 2 streptavidin-oligonucleotide conjugates, one having oligonucleotides with free phosphorylated 5’ ends and one with free 3’ ends. The probes were allowed to find their targets during 1.5 h incubation at RT under rotation. The beads were washed twice in 100 μl washing buffer before adding 50 μl ligation and PCR mix (1xPCR buffer, 2.5 mM MgCl_{2 ,} 1.5 units Platinum Taq Polymerase (10966, Life Technologies), 0.1 μM of each primer, 0.2 μM TaqMan probe, 0.08 mM ATP (R0441, Thermo Scientific), 0.1 μM connector oligonucleotide, 0.2 mM dNTPs (with dTTP substituted by dUTP; V0195, Thermo Scientific), 0.5 units T4 DNA ligase (EL0016, Thermo Scientific), 0.1 units uracil-DNA glycocylase (UDG; EN0362, Thermo Scientific)). The ligation was carried out for 10 min at RT before starting the qPCR on an Mx3000 real-time PCR instrument (Stratagene), with an initial hot start at 95°C for 10 min, followed by 40 cycles of 15 s at 95°C and 60 s at 60°C.

Data analysis

The qPCR data was analyzed with the MxPro software (Stratagene) to extract the cycle of threshold (Ct) values. To set a cut off value for the assay we used the Ct value of a negative control sample (only PLA buffer) and subtracted 3 standard deviations from the average Ct value for triplicate assay. Based on this cut off, we calculated the ratio of ligated DNA reporter molecules in the samples relative to the amount of reporter molecules set as cut off for the assay.

Signal relative to cut off = 2^{Ct(cut off)-Ct(sample)}

Disclosure of Potential Conflicts of Interest

UL is a cofounder and shareholder of Olink Bioscience that commercializes the proximity ligation assay technology described in this paper.

Funding

This work was funded by the Knut and Alice Wallenberg Foundation, FORMAS (2006-2856, MKM), Alliance Biosecure Foundation, Åke Wiberg Foundation, IngaBritt and Arne Lundberg's Foundation, European Science Foundation, the Swedish Research Council for medicine (2007-2720, UL), and by the European Community's 6th and 7th Framework Programs.