Mixtures of prion substrains in natural scrapie cases revealed by ovinised murine models

Abstract

Phenotypic variability in prion diseases, such as scrapie, is associated to the existence of prion strains, which are different pathogenic prion protein (PrP^Sc) conformations with distinct pathobiological properties. To faithfully study scrapie strain variability in natural sheep isolates, transgenic mice expressing sheep cellular prion protein (PrP^C) are used. In this study, we used two of such models to bioassay 20 scrapie isolates from the Spain-France-Andorra transboundary territory. Animals were intracerebrally inoculated and survival periods, proteinase K-resistant PrP (PrP^res) banding patterns, lesion profiles and PrP^Sc distribution were studied. Inocula showed a remarkable homogeneity on banding patterns, all of them but one showing 19-kDa PrP^res. However, a number of isolates caused accumulation of 21-kDa PrP^res in TgShp XI. A different subgroup of isolates caused long survival periods and presence of 21-kDa PrP^res in Tg338 mice. It seemed that one major 19-kDa prion isoform and two distinct 21-kDa variants coexisted in source inocula, and that they could be separated by bioassay in each transgenic model. The reason why each model favours a specific component of the mixture is unknown, although PrP^C expression level may play a role. Our results indicate that coinfection with more than one substrain is more frequent than infection with a single component.

Figure 1

Western blot of inocula sourced from sheep tissues. (A) Inocula from medulla oblongata; (B) inocula from mesenteric lymph node. Note that all but one samples had banding patterns with the non-glycosylated (NG) band at 19 kDa. M: molecular weight marker; kDa: kilodaltons. Notice that cropped images are presented here; for the full image, see Supplementary Figure S3.

Figure 2

Western blot of spinal cord homogenates from second-passage TgShp XI mice experimentally challenged with (A) sheep brain-derived isolates and (B) sheep lymph node-derived isolates. Note that most samples had banding patterns with the non-glycosylated (NG) band at 19 kDa, while those of inocula 3 N, 4 N, 5 N and 7 L were at 21 kDa. M: molecular weight marker; kDa: kilodaltons. Notice that cropped images are presented here; for the full image, see Supplementary Figure S3.

Figure 3

Western blot of spinal cord homogenates from second-passage Tg338 mice experimentally challenged with (A) sheep brain-derived isolates and (B) sheep lymph node-derived isolates. Note that 13/18 samples were clearly positive with non-glycosylated bands at 19 kDa. 5/18 samples differed from this, either being negative (4 L and 5 L) or having NG bands at 21 kDa (2 L, 6 L and 8 L). M: molecular weight marker; kDa: kilodaltons. Notice that cropped images are presented here; for the full image, see Supplementary Figure S3.

Figure 4

Lesion profiles of second-passage TgShp XI mice. Both inocula associated to the 19K phenotype (A) and inocula associated to the 21K-TgShp XI phenotype (B) provoked similarly shaped lesion profiles in TgShp XI mice (C), most of them characterized by peak scores at mesencephalon and thalamus. The brains of all mice dying after the onset of clinical signs in each challenged group were analyzed to plot lesions profiles (usually 6 and never less than 3 animals per group). Mobl: medulla oblongata, Cb: cerebellar cortex, Mes: mesencephalon, Hy: hypothalamus, Th: thalamus, Hp: hippocampus, Sn: septal nuclei, Tc: cortex at the level of thalamus, Fc: frontal cortex. Error bars represent SEM.

Figure 5

Lesion profiles of second-passage Tg338 mice. Inocula associated to the 19K phenotype provoked similar lesion profiles in Tg338 mice, mostly characterized by peaks at medulla oblongata and thalamus (A). However, isolates linked to the 21K-Tg338 phenotype triggered lower and flatter profiles (B,C). The brains of all mice dying after the onset of clinical signs in each challenged group were analyzed to plot lesions profiles (usually 6 and never less than 3 animals per group). Mobl: medulla oblongata, Cb: cerebellar cortex, Mes: mesencephalon, Hy: hypothalamus, Th: thalamus, Hp: hippocampus, Sn: septal nuclei, Tc: cortex at the level of thalamus, Fc: frontal cortex. Error bars represent SEM.

Figure 6

PrP^Sc distribution in representative brain sections from TgShp XI mice infected with inocula 2 N (A), 3 N (B), 7 L (C) and 8 L (D). Two different trends of PrP^Sc accumulation were observed on PET-blots: a generalized low-intensity immunostaining (A,B,C), and a localized high-intensity pattern, usually associated with the alveus of hippocampus (B, D, arrowheads), and subependymal areas (D, arrowheads), and which correlated with coarse/coalescing plaque-like deposits on immunostochemistry (E).

Figure 7

PrP^Sc distribution in representative brain sections from Tg338 mice infected with inocula 1 L (A), 3 L (B), 7 L (C) and 8 L (D,E). A general pattern arose in all examined brain samples consisting of intense staining of brainstem and subcortical areas. Additionally, conspicuous immunostaining was observed in specific structures, including the cingulate gyrus (A,B,E; thick orange arrows), habenular nuclei (A,C; red arrowheads), and the stratum lacunosum-moleculare and the alveus of hippocampus (B,E; thin green arrows). High-intensity staining of the medial amygdaloid nuclei (D,E; blue triangles) was observed in a subset of samples obtained from animals challenged with isolates causing prolonged survival periods and low vacuolization scores.

Figure 8

Unicentric (A to C), florid (D to F) and amorphous perivascular plaques (G to L) in brain of second-passage Tg338 mice infected with inocula 6 L and 8 L. Haematoxilin-eosin staining (A,D,G,J), immunohistochemistry with anti-PrP mAb SAF84 (B,E,H,K) and Congo red staining that proves the amyloid nature of the plaques (C,F,I,L).