Distinct immunohistochemical localization in Kuru plaques using novel anti-prion protein antibodies

Authors


Correspondence
Takashi Onodera, Department of Molecular Immunology, School of Agriculture and Life Sciences, University of Tokyo, Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-8657, Japan. Tel: +81 3 5841 5196; fax +81 3 5841 8020; email: aonoder@mail.ecc.u-tokyo.ac.jp

ABSTRACT

By immunizing Prnp-knockout mice with synthetic polypeptides, a panel of mAbs directed to bovine PrPC was obtained. The mAb panel was characterized by the ELISA method, where synthetic polypeptides were used for epitope mapping. Different reactivity patterns were identified. The ability of these mAbs to detect abnormal PrPSc in CJD cases was studied by immunohistochemistry. All mAbs were tested for PrPSc in murine, bovine, monkey and human brain tissues. Three mAbs recognized the fragmented PrP epitope in our ELISA. Antibody 1D12 was strongly reactive to ovine and squirrel monkey tissues infected with a scrapie agent, although non-reactive to scrapie-infected mouse tissues. Antibody 2D8 was clearly reactive to type-2 but not type-1 CJD human tissues. Of particular interest was the reactivity of mAb 6C4 with the inner structure of Kuru plaques (peripheral pattern) in a type-2 CJD case and mAb T2, 1D12, 2B11, 2D8, 4B5 and 6G3-2 with the central area (central pattern). The fact that different anti-PrP mAbs possess distinct staining properties suggests that the PrPc to PrPSc conversion might involve a multiple-step process.

List of Abbreviations: 
BSE

bovine spongiform encephalopathy

CJD

Creutzfeldt-Jakob disease

ELISA

enzyme-linked immunosorbent assay

GSS

Gerstmann-Sträussler-Scheinker syndrome

mAb

monoclonal antibody

PK

proteinase K

PrP

prion protein

PrPC

cellular isoform of prion protein

PrPSc

scrapie isoform of prion protein

TSE

transmissible spongiform encephalopathy

Transmissible neurodegenerative diseases of the central nervous system (CNS), collectively termed TSE, are caused by prions. TSE include scrapie, BSE, Kuru, CJD and GSS. The pathogenetic mechanisms of these TSE have not been elucidated in detail (1), although the fundamental lesion underlying prion diseases is related with the conversion of PrPC to PrPSc (2, 3). PrPSc is known to accumulate in the scrapie-affected brain (3). Biochemical analysis of abnormal prion proteins (the only specific marker of TSE) contributes to the identification of different prion strains (4). PrPSc is partially resistant to protease digestion and yields a 20- to 30-kDa fragment after digestion with proteinase K (PK) (4). In rodent scrapie, a sequence of 62 N-terminal amino acids is digested, leaving a core of 141 amino acids (5). Using the western blot method with antibodies directed to the core globular PrP domain, three fragments (i.e. the diglycosylated, monoglycosylated and non-glycosylated forms of the protein) were separated according to their relative molecular weights. Although high levels of diglycosylated PrPSc in sheep experimentally infected with BSE or natural sheep scrapie have been documented, discrimination of the infectious agent based on the PrPSc glycoforms ratio cannot be established (6, 7).

A recently developed immunohistochemical approach for identifying BSE agent in sheep is based on the characterization of abnormal PrP epitopes detected in neurons and phagocytic cells. This concept is applicable to both the lymphoreticular system and CNS (8).

TSE diagnosis relies on the histopathological presence of spongiform lesions and immunological detection of PrPSc (9). As PrP amino acid sequences differ by 5–20% among mice, hamsters, sheep and cattle (5, 10, 11), cross-reactivities between PrP-reactive antibodies are bound to occur. As such, the present study was initiated to characterize the novel monoclonal antibodies (mAbs) raised in Prnp-knockout mice in order to recognize the different epitopes of bovine PrP. Anti-PrPC mAbs directed to conformational epitopes reacted strongly with PrPSc forms from all of the above-mentioned sources. In addition, specific differences in the reactivity of two mAbs with brain tissue from CJD patients were observed.

MATERIALS AND METHODS

Synthetic peptides corresponding to codons 153–166 of the bovine PrP gene (12) were synthesized using a tBoc method (Kurabo Industries, Tokyo, Japan). An additional cysteine residue was added to the amino- or carboxy-terminal of each peptide to facilitate conjugation of peptides to a carrier-protein, which was either bovine serum albumin or keyhole limpet hemocyanin. Prnp-knockout mice with the C57BL/6 genetic background (13, 14) have extensively been used for immunization and mAb production.

Antibody responses to peptides were assessed by an ELISA method (15). Briefly, 96-well ELISA plates were coated with appropriate amounts of bovine PrP expressed via Escherichia coli and corresponding to codons 25–242, blocked with 3% skim milk, and incubated with test serum or mAbs for 1 hr. Binding was detected in the second incubation with horseradish peroxidase-conjugated rabbit antibodies against mouse immunoglobulin (IgG; ICN/Cappel, Aurora, OH, USA) for 1 hr. Plates were monitored with a microplate reader (Molecular Devices, Sunnydale, CA, USA) at 405-nm wavelength.

Epitope mapping of the selected mAbs was established with an ELISA assay (15) using synthetic overlapping peptides (10-mer) derived from the amino acid sequence of bovine PrPC. Acetylated peptides from codons 151–160 (Ac 151–160), Ac 152–161, Ac 153–162, Ac 154–163, 155–164, Ac 156–165, Ac 157–166, Ac 158–167, Ac 159–168 and Ac 160–169 were used to detect the epitopes.

Tissue samples were fixed in 10% formalin and embedded in paraffin before sectioning. To ensure adhesion, 5-μm-thick tissue sections were harvested, placed on pretreated glass slides (Star Frost; Fisher Scientific, Pittsburg, PA, USA) and subsequently dried. After dewaxing and rehydration, sections were stained for PrPSc detection. Briefly, sections were immersed in 98% formic acid for 5 min, washed in running tap water, and immersed in 0.2% citrate buffer. Sections were then autoclaved at 121 °C for 30 min (16, 17). Endogenous peroxidase was blocked with 1% H2O2 in ethanol for 20 min. After serially rinsing in tap water and PBST buffer (0.2% Tween 20), tissue antigens were blocked with normal horse serum for 60 min. Incubation with the primary antibody was then performed overnight at 4 °C. Different PrP mAbs were used at concentrations as reported previously (15). Bound mAbs were detected by the immunoperoxidase method (Vector-Elite ABC; Vector Laboratories, Peterborough, UK). After color development, sections were immersed in 0.5% copper sulfate and counterstained with Mayer's hematoxylin to enhance the intensity of reaction products. Intraneural PrPSc labeling of target spots was scored according to the published criteria (18): no labeling (−), trace of light labeling (+/−), light labeling (+), moderate labeling (2+), and intense labeling (3+). PrP staining of neural tissues with the novel antibodies T2 and 6C4 was performed in parallel with two control antibodies (dilution rate 1:1000; mAb 8C6 (19) and mAb F99/67 from VMRD, Pullman, WA, USA (20)).

RESULTS

The results of the ELISA epitope mapping (Table 1) indicated that mAbs 4B5, and 2B11 recognized linear epitopes comprising the YEDRY sequence (aa 156–160) of bovine PrPC. As for mAbs 6C4, 1D12, and 6G3-2, they were instead strongly reactive to YEDRYYRENM (aa 156–165) and moderately to the YEDRY sequence. This suggested that mAbs 6C4, 1D12, and 6G3-2 were reactive to fragmented PrP epitope YEDRY, M (aa 156–160, 165) in our ELISA (Table 1). mAb 2D8 was weakly reactive to these peptides in the ELISA epitope-mapping procedure. Although original mice were immunized with amino acids (aa) 153–166, we could not determine the site of the epitope. Note that mAb T2 was directed to RPLIHF (aa 135–140) of mouse PrPC, although this mAb was reactive to bovine or human PrPSc.

Table 1.  Epitope mapping by ELISA
mAbIsotypeELISA epitope mapping‡
  1. †Previously reported epitopes in mouse PrP (15).

  2. ‡Epitopes were mapped by ELISA using overlapping 10-mer peptides deduced from the amino acid sequence of bovine PrP.

  3. aa, amino acid; ND, not determined due to weak reactivity in ELISA.

6C4IgG1YEDRY, M (aa 156–160, 165)
1D12IgG2aYEDRY, M (aa 156–160, 165)
2B11IgG1YEDRY (aa 156–160)
2D8IgG1ND
4B5IgG1YEDRY (aa 156–160)
6G3-2IgG1YEDRY, M (aa 156–160, 165)
T2†IgG1RPLIHF (aa 135–140)

Staining of a Japanese BSE case with mAbs 6C4, 1D12, 2B11, 4B5 and 6G3-2 yielded synaptic-type patterns that were particularly intense with mAbs 6C4 and 1D12 (Table 2). Interestingly, mAbs 6C4 and T2 displayed a typical staining pattern in the brain sections of ovine scrapie and BSE cases infected with agents from Japanese, Italian and Irish origins. Diffuse particulate PrP staining of the neuropile colocalized with 6C4 and T2 antibodies, and the staining patterns were comparable to that produced by the commercially available mAb F99/97. Interestingly, although strongly reactive to ovine or squirrel monkey tissues previously infected with a scrapie agent, antibody 1D12 was not reactive to scrapie-infected mouse tissues. Note that antibody 2D8 was clearly reactive to type-2 but not to type-1 CJD human brain tissues. We are still studying this phenomenon using other human type-1 and type-2 CJD materials. Further analysis using western blotting is warranted.

Table 2.  Reactivity of monoclonal antibodies in brain tissues: Immunohistochemistry
Brain samplemAbs
6C41D122B112D84B56G3-2T2
  1. Formalin-fixed brain tissues from TSE-affected animals were treated initially by formic acid before being autoclaving and stained (see Materials and Methods), or left undone (NA).

  2. Scrapie Japanese case: a spontaneous case reported in Hokkaido in 1984.

  3. Scrapie mouse: ICR mice inoculated with Obihiro-1 strain of scrapie and killed 120 days later.

  4. Scrapie squirrel monkey: A scrapie strain (Tsukuba-1) was inoculated intracerebrally into the brain of a squirrel monkey. Approximately 5 years later, the inoculated monkey was killed and brain tissues were tested.

  5. Type-1 CJD: Brain tissues of Japanese and Italian CJD (synaptic type) were tested.

  6. Type-2 CJD: Cerebellum tissues of Japanese and Italian CJD (Kuru-plaque type) were tested.

Scrapie (Japanese case)2+2+2+1+2+2+3+
Scrapie (mouse)2+2+1+2+2+3+
Scrapie (squirrel monkey)2+3+2+1+2+3+NA
BSE (Japanese, Italian and Irish cases)3+2+2++2+2+3+
Type-1 CJD2+3+2++2+3+2+
Type-2 CJD3+3+3+2+2+3+3+

Findings in immunohistochemistry using mAbs revealed no PrPC staining in samples previously subjected to formic acid/autoclave treatment (data not shown). Staining of human CJD cases with mAbs 6C4 and T2 using mAb 8C6 as a control produced intriguing results (Fig. 1a,b). Type-1 CJD (synaptic type) reacted with mAbs T2 and 6C4 to produce a diffuse fine margin of PrPSc staining around residual neurons in the affected cortex. Type-2 CJD (one case with the codon 129 Met/Val genotype) was recognized by the two mAbs at anatomical sites different from that of type-1 CJD. Kuru-type spherical plaques in the cerebellum were intensively stained with the mAbs; 6C4 indicated a peripheral pattern, whereas mAbs T2 (Fig. 1b), 1D12, 2B11, 2D8, 4B5, 6G3-2 and 8C6 (control; data not shown) displayed a central pattern. Positive Kuru plaques were particularly abundant in the cerebellar granular layer compared to the molecular layer.

Figure 1.

Immunohistochemical staining with monoclonal antibodies (mAbs) 6C4 and T2 in human brain tissues with Creutzfeldt-Jakob disease (CJD). Type-1 (synaptic type) CJD specimens were intensively stained with mAbs 6C4 (a) and T2 (d) in the molecular layer of cerebellum, with findings showing granular (synapsis type) staining. Type-2 (Kuru type) spherical plaques (b, c, e, f) of the cerebellum from a CJD patient with Met/Val polymorphism at codon 129 were differently stained with mAbs 6C4 and T2; the former was characterized by negative staining in the core of plaques, whereas the latter portrayed homogeneous staining of the central area (c, f).

DISCUSSION

A few commercially available mAbs react with both PrPC and PrPSc (19, 21). A library of well-characterized antibodies may contribute to an understanding of the pathogenetic events in prion diseases (19, 22). It is potentially useful that mAbs reacting with different PrP epitopes may facilitate identification of the PrPC-derived products that eventuate PrPSc accumulation in affected tissues. Apart from facilitating detection of PrPC and PrPSc, the availability of anti-PrP mAbs which can recognize epitopes of PrPC expedites monitoring of PrPC processing under normal or pathological conditions. The mAbs raised in our study are directed to distinct epitopes of PrPC fragments 156–160, and 156–160, 165 epitope. The present mAb panel is capable of recognizing PrPSc in several mammalian species, including mice, sheep, cattle, and humans. As these mAbs recognize PrPC of evolutionarily diverse species, it is likely that they will react with species hitherto untested.

These mAbs are suitable for PrPSc immunohistochemistry. With reference to type-2 CJD, mAb 6C4 recognizes the outer layer of Kuru plaques, whereas mAb T2, 1D12, 2B11, 2D8, 4B5, and 6G3-2 selectively recognize the inner Kuru plaque layer (mostly in the central area). Therefore, these antibodies will facilitate discriminative staging of the different clinical phases in CJD, besides identifying the metabolic products leading to PrPC to PrPSc conversion. Results of the present investigation indicate that the outer and inner Kuru plaque structures may contain different disease-specific isoforms of PrP. The PrP isoform accumulating in the central plaque area may exist as an intermediate that eventually produces the pathogenic PrPSc. Immunohistochemical staining of mAb T2 reproduces results that have previously been obtained with rabbit polyclonal antibodies raised against synthetic PrP peptides (23, 24). A staining pattern similar to that produced by T2 has also been reported with the mAb 6G9 (19). As such, mAbs 6G9 and T2 might recognize an intermediate isoform that is responsible for the pathogenic PrPSc isoform.

ACKNOWLEDGMENTS

This work was supported by a Grant-in-Aid from the Ministry of Health, Labor and Welfare of Japan (to TO) and a Grant-in-Aid for Scientific Research (to TO) from the Ministry of Education, Science, Culture and Technology of Japan, and an Exchange grant from the University of Tokyo (to AT). Thanks are due to Dr Shigeyoshi Itohara for providing Prnp-knockout mice, and Dr Anthony Foong for reading the manuscript.

Ancillary