Abstract
Bovine spongiform encephalopathy (BSE) is a transmissible neurodegenerative disease of cattle. Clinical diagnosis can be confirmed by investigation of both spongiform changes and abnormal prion protein (PrPsc), a marker considered specific for the disease. Tissue autolysis, often unavoidable in routine field cases, is not compatible with histological examination of the brain even though PrPsc is still detectable by immunoblotting. To determine how autolysis might affect accurate diagnosis using PrPsc immunohistochemistry, we studied 50 field samples of BSE brainstem (obex) with various degrees of autolysis. We demonstrated that the antigen-unmasking pretreatments necessary for PrPsc immunohistochemistry were compatible with the preservation of autolyzed brain sections and that PrPsc detection was unaffected by autolysis, even though anatomic markers were sometimes lost. In tissue samples in which anatomic sites were still recognizable, PrPsc accumulation was detected in specific gray matter nuclei. In samples with advanced autolysis, PrPsc deposits were still observed, at least at the cellular level, as an intraneuronal pattern. We found that the sensitivity of PrPsc immunohistochemistry as a diagnostic method for BSE was undiminished even by severe tissue autolysis.
For all these reasons, in this study we undertook to test these limits on 50 brains (of which 46 were BSE-positive cases) with varied autolysis status. Using monoclonal antibody (MAb) SAF84 (Demart et al. 1999), we demonstrate that even in brain samples with very advanced autolysis pathological PrP is still clearly detectable, at least at a cellular level, indicating that IHC remains a sensitive tool applicable to BSE diagnosis.
Materials and Methods
Brain Samples
Brains were collected from cattle aged 24 months and over, acquired through the French BSE surveillance program on cattle at risk, i.e., from cattle dead on farm or subjected to emergency slaughter because of sickness or accident. Sub-samples of each of these brains were frozen for biochemical analysis (brainstem) and were also fixed for immunohistochemical analysis (medulla oblongata at the level of the obex).
Forty-four brains that were positive using the Prionics test (Schaller et al. 1999) and two Prionics-negative brains were systematically analyzed by PrPsc IHC. In addition, four obex samples collected through the passive surveillance scheme for BSE were included in the study. These were not subjected to the Prionics test but their BSE status was confirmed by histopathology and by our in-house Western blotting procedure or, in the case of the autolyzed samples, by Western blotting alone. Two were BSE-positive and two negative on the basis of these tests.
For every brain sample, the degree of autolysis was noted and graded as shown in Table 1. Grading was based on the macroscopic appearance of the sample (clean sections, amorphous, pieces, liquid state) as well as the ability to identify the area of the obex macroscopically when the samples were cut coronally before embedding. In addition, autolysis degree was also assessed on subjective histological criteria such as survival of tissue architecture, vascular retraction, and undefined cell borders, as used elsewhere (Borras et al. 2000).
Anatomic sites recognized and autolysis status of the 46 BSE brain samples analyzed
Tissue Preparation
All brain samples were maintained in fixative (buffered 10% formalin) for a minimum of 1 week. Then they were placed in tissue-embedding cassettes (samples showing an advanced state of autolysis were filtered using a nylon bag and the filtrates placed in embedding cassettes) and, after incubation in a bath of 98–100% formic acid (Merk Eurolab; Darmstadt, Germany) for 1 hr, they were processed through graded ethanol to chloroform and embedded in paraffin. Sections from blocks were cut to a thickness of 5 μm, collected on treated glass slides (Starfrost; Medite Histotechnic, Burgdorf, Germany), and kept in an oven overnight at 55C.
Immunohistochemistry
The sections were subjected to pretreatments to enhance antigen retrieval and allow distinction of the pathological form of PrP from its normal cellular isoform (PrPc) (Bell et al. 1997; Van Everbroeck et al. 1999; Taylor 2000). They were first treated with 98–100% formic acid for 10 min at room temperature (RT) and then, after a wash in distilled water, were subjected to hydrated autoclaving for 20 min at 121C (Model E2083; Prestige Medical, AES Labs, Blackburn Lane, UK). Endogenous peroxidase activity was blocked with 3% H2O2 (Merk Eurolab) for 5 min at RT. Sections were rinsed for 5 min in distilled water and mounted in a Sequenza immunostaining system (Shandon; Cergy Pontoise, France) using PBS 0.1 M, pH 7.4, containing 0.1% Tween-20, before being digested at 37C with proteinase K (Roche Boehringer Mannheim; Mannheim, Germany) at a concentration of 20 μg/ml for 15 min.
After a wash, nonspecific antigenic sites were blocked using 1% blocking reagent (Roche Boehringer Mannheim) for 30 min at 37C. Then blocking agent was replaced by MAb SAF84 (courtesy of J. Grassi; CEA-Saclay, France) at a concentration of 0.5 μg/ml, and the sections were incubated in this for 1 hr at 37C. After washing, the secondary antibody (biotinylated goat anti-mouse; Southern Biotechnologies, Birmingham, AL) was applied at a dilution of 1:200 for 1 hr at 37C. The diluent buffer for both primary and secondary antibodies was 0.1 M PBS, pH 7.4, containing 0.001% Triton X-100. The slices were then incubated with peroxidase-labeled avidin-biotin complex (Vectastain Elite ABC; Vector Laboratories, Burlingame, CA) for 30 min at RT. Finally, peroxidase activity was revealed using a solution of diaminobenzidine intensified with nickel chloride (DAB; Zymed, San Francisco, CA). The slides were counterstained using aqueous hematoxylin solution, mounted using Eukitt, and observed under a microscope (Zeiss; Oberkochen, Germany) coupled to an image analysis workstation (Biocom; Les Ulis, France).
Omission of primary antibodies was used to check for nonspecific background staining in BSE tissues with and without autolysis and known to produce PrPsc immunolabeling.
The specificity of positive PrPsc immunolabeling was also assessed using BSE-negative bovine brain tissue, with and without autolysis.
Results
In the sections from BSE-positive tissues that were adequately preserved, pathological PrPsc was detected in gray matter only, white matter being devoid of any immunostaining. Staining was particularly marked in the nucleus of the solitary tract (NST), the dorsal nucleus of the vagus nerve (DNV), the olivary nucleus (ON), the spinal tract of the trigeminal nerve (STN), the cuneate nucleus (CN), and the reticular formation (RF) (Figures 1A and 1B). In particular, no background staining was observed. In BSE-negative control sections with adequate preservation, neither PrPsc nor background immunostaining was detected (Figure 1D).
In BSE-positive tissues with obvious autolysis, PrPsc deposits were distinguished within the obex in specific sites at a higher magnification (NST, DNV, ON, STN, CN, RF; Figure 1C), whereas in the autolyzed BSE-negative cases neither PrPsc nor background immunostaining was seen (Figure 1E).
Immunohistochemistry of abnormal prion protein (PrPsc) using SAF84 in obex region. Schematic representation of obex section from Lignereux (Lignereux 1986) (
The 46 brain samples from the French BSE surveillance programs were classified into categories A to E according to the degree of autolysis and also the extent to which it was possible to identify the anatomic site. The categories were defined as follows (see Table 1): Category A represents 13 samples which were identifiable as obex and two as spinal cord, which showed no autolysis. Obex samples (n=10) with slight autolysis were grouped into Category B. Category C contained one spinal cord and nine obex samples with distinct autolysis. In all these categories, PrPsc deposits were seen in the obex in the anatomic sites described above, with no labeling elsewhere (Figures 2A1–2C1).
In the three samples of spinal cord, PrPsc deposits were confined to the gray matter and no immunolabeling was seen in the white matter (Figure 2, Asp1-Asp2). Intact samples (n = 8) with advanced autolysis, which were not identifiable anatomically, were grouped into Category D. Category E (n=3) represents the samples received in a liquid state and impossible to identify anatomically. In these two last categories, PrPsc deposits were discernible macroscopically on the slide (Figures 2D1 and 2D2), even from samples originally received in a liquid state (Figures 2E1 and 2E2).
At the cellular level in Categories A-C, deposition of PrPsc was scattered within the neuropil in the NST, ON, or DNV, appearing as small granules (Figures 2A2 and 2B2) or as a granular delineation of neuronal processes in the RF (Figure 2C2). In addition, in this area PrPsc deposits were seen either within the cytoplasm of neurons (Figure 2, Asp3-E3) or in a perineuronal manner (Figures 2B3 and 2C3).
In Categories D and E, cells were most often very poorly preserved and PrPsc deposits were seen, either within these cells (Figures 2D3 and 2E3) or widespread within the remaining histological and cellular structures (Figures 2D2 and 2E2).
Discussion
Histopathological diagnosis of BSE, based mainly on the recognition of spongiform vacuolation in the gray matter neuropil, may not be possible if the tissue sample is poorly preserved or autolyzed. It has been shown that, because of its great resistance to degradation (Taylor 2000), PrPsc detection by immunoblotting is unaffected by tissue autolysis (Schaller et al. 1999), but less is known of the efficacy of PrPsc IHC when applied to autolyzed tissues. To clarify the limits of applicability of PrPsc IHC, we performed this method on brain samples of varying quality in terms of autolysis status, determined essentially macroscopically and during the cutting of brain specimens before embedding.
In well-preserved BSE-positive brain samples at the level of the obex, PrPsc deposits were found in this study to be restricted to the gray matter, notably NST, DNV, ON, STN, CN, and RF, and this was in accordance with previous studies (Simmons et al. 1996; Orge et al. 2000). The specificity of these PrPsc deposits was validated by the absence of any labeling in BSE-negative brain sections. In autolyzed BSE-positive brain samples that were still identifiable as being at the level of the obex, PrPsc topographic distribution was recognizable in the same anatomic structures as mentioned above, despite the poor histological preservation. In tissues autolyzed to the extent that the anatomic site was unidentifiable, disease-specific immunostaining was still evident in BSE-positive sections, in contrast to the absence of immunostaining in negative tissues similarly autolyzed. In addition, the absence of immunostaining in autolyzed negative controls showed that autolysis did not generate any nonspecific background staining.
It was possible to evaluate the “mechanical” resistance of the autolyzed brain sections to pretreatments that are necessary in the PrPsc IHC protocol, because they enhance PrPsc detection and at the same time reduce or abolish the detectability of normal PrPc (Van Everbroeck et al. 1999). Even in very advanced autolyzed samples, such as in Categories D and E, these drastic treatments, using formic acid, autoclaving, and proteinase K digestion, were compatible with the adherence and preservation of the brain sections on glass slides. Despite the sometimes very poor histological preservation, the 44 brain samples previously found positive by the Prionics test were all found positive by IHC. This result suggested that our PrPsc IHC procedure, at least for these samples, was as sensitive as the Prionics test.
Representative illustrations of PrPsc IHC within the 5 Categories A-E, corresponding to graded tissue autolysis status. (
Among the gray matter nuclei accumulating PrPsc deposits, all were not systematically found to be positively labeled. This may reflect different stages in the progression of the disease at which the different nuclei are not all affected in the same time. However, the tissues analyzed here were chosen according to the degree of autolysis only and not according to the intensity of PrPsc accumulation. Therefore, we did not attempt to investigate the effects of autolysis on the intensity of PrPsc immunolabeling, which would have required tissues in which the autolytic process was controlled.
Nevertheless, in each of those BSE cases that gave weak Prionics test results (two from Categories A and B, two others from Category D), a few cells (three to ten) still undoubtedly contained PrPsc deposits, Category A and B tissues being weaker than Category E tissues. This indicated that IHC was sensitive even in weak cases. Alternatively, it has applications as a complementary method, especially to clarify cases weakly positive by a biochemical method. In a context in which methods are needed to evaluate the true prevalence of BSE in cattle, the present study demonstrates that PrPsc immunohistochemistry is a reliable diagnostic tool regardless of tissue quality and even in cases with small amounts of PrPsc.
Footnotes
Acknowledgments
Supported by grants from the European Commission (FAIR 98–7021) and the French Ministry of Agriculture and Fisheries. Sabine Debeer was financially supported by the EU (FAIR 98–7021).
We gratefully thank Dr E. Monks (Veterinary Research Laboratories, Dublin, Ireland) for critical reading of the manuscript and valuable suggestions. We extend our grateful thanks to S. Philippe (Statistician, AFSSA-Lyon, France) for valuable comments and suggestions.