The occurrence of the variant Creutzfeldt–Jakob disease (vCJD), related to bovine spongiform encephalopathy (BSE), raises the important question of the sources of human contamination. The possibility that sheep may have been fed with BSE‐contaminated foodstuff raises the serious concern that BSE may now be present in sheep without being distinguishable from scrapie. Sensitive models are urgently needed given the dramatic consequences of such a possible contamination on animal and human health. We inoculated transgenic mice expressing the ovine PrP gene with a brain homogenate from sheep experimentally infected with BSE. We found numerous typical florid plaques in their brains. Such florid plaques are a feature of vCJD in humans and experimental BSE infection in macaques. Our observation represents the first description, after a primary infection, of this hallmark in a transgenic mouse model. Moreover, these mice appear to be a promising tool in the search for BSE in sheep.
Transmissible spongiform encephalopathies (TSEs) are neurodegenerative diseases characterized by vacuolation and accumulation of the pathological form of the prion protein—PrPsc—in the brain. Ten years after the first bovine spongiform encephalopathy (BSE) cases reported in the United Kingdom, the occurrence in April 1996 of the variant form of the human TSE Creutzfeldt–Jakob disease (vCJD), affecting unusually young patients, raised the important question of the sources of human contamination (Will et al., 1996; Almond and Pattison, 1997). The histopathological hallmark in the brain of all patients with vCJD is the presence of numerous florid plaques (Fraser, 1979; Ironside and Bell, 1997) that were also found in the brain of macaques infected with cattle BSE (Lasmezas et al., 1996), thus providing the first link between BSE and vCJD. This correlation has been rapidly confirmed by disease transmission studies in inbred wild‐type mice or bovine PrP transgenic mice (Will et al., 1996; Almond and Pattison, 1997; Bruce et al., 1997; Scott et al., 1999). The similarities between incubation times and between the lesion distributions revealed that a common agent was responsible for both vCJD in human and BSE in cattle (Bruce et al., 1997; Scott et al., 1999). Moreover, studies of the quantitative ratios of the different PrPres glycoforms and evaluation of the molecular mass of the prion protein showed a typical and unusual molecular pattern named type 4 (Collinge et al., 1996; Hill et al., 1997) in all cases of vCJD compared with sporadic and iatrogenic CJD. This type 4 pattern has also been found in macaques experimentally infected with the BSE agent (Collinge et al., 1996; Lasmezas et al., 2001).
The origin of BSE remains unknown. It may have arisen from sheep TSE, scrapie, through the recycling of scrapie agent in meat and bone meal (MBM) used in cattle foodstuff. These two types of TSEs are, however, distinct in several respects. Natural sheep scrapie is rather common in sheep and has been present in sheep flocks for a long time whereas TSEs were unknown in cattle prior to the 1980s. While BSE seems to be sustained by a single major strain of agent (Bruce et al., 1997), natural scrapie is highly polymorphic with several scrapie strains that differ in terms of incubation time and lesion profile following transmission in inbred wild‐type mice (Fraser and Dickinson, 1968; Fraser and Dickinson, 1973). Since BSE can be transmitted experimentally to small ruminants, including by oral routes (Foster et al., 2001), there is a risk that BSE may have been disseminated through MBM foodstuff into the sheep population. Subsequent horizontal and vertical transmission of the BSE agent within sheep population may constitute a new source of BSE agent exposure to human. Sensitive models are urgently needed considering the dramatic consequences on animal and human health. Biochemical analysis may be discriminant (Hill et al., 1998) but there is, at least, one exception with the CH1641 experimental scrapie showing the same electrophoretic pattern as that found in a BSE‐infected sheep. In addition, the lesion profile, following transmission of the disease to wild‐type mice, is a time‐consuming method. Thus, there is unfortunately no real specific and easy manner to recognize rapidly the BSE agent in sheep. To address this question we have used ovine PrP transgenic mice [Tg(OvPrP4)] expressing, selectively in their brain, the ovine PrP gene using the neuron‐specific enolase promoter (Crozet et al., 2001). Compared with wild‐type mice, these mice are more sensitive to sheep scrapie (Crozet et al., 2001). In the present study, we have inoculated Tg(OvPrP4) mice with a brain homogenate prepared from a sheep inoculated with French cattle BSE agent (Baron et al., 2000). Here we report that the singular ‘florid plaques’ signature in the brain of infected Tg(OvPrP4) mice is identical to that of vCJD in humans. This simple criteria may thus provide the opportunity to identify possible field contamination of sheep by the BSE agent.
Results and Discussion
We used Tg(OvPrP4) mice expressing the ovine PrP gene (AA136RR154QQ171 genotype) in the brain under the control of the neuron‐specific enolase promoter (Crozet et al., 2001). Five‐week‐old mice (n = 20) were intracerebrally inoculated with a brain homogenate obtained from a sheep (AA136RR154QQ171 genotype) experimentally infected with cattle BSE agent (Baron et al., 2000). Two hundred and twenty days after inoculation, the mice started to display extensive clinical signs of encephalopathy (leanness, hunched posture, hindlimb paralysis, equilibrium trouble, plastic tail and prostration), and they died between 221 and 371 days post‐inoculation (p.i.) (mean survival time 300 ± 40 days). Biochemical analysis, performed on every other mouse brain, showed the presence of proteinase K‐resistant PrP with an electrophoretic profile comparable to that of ovine BSE (Figure 1). When comparing the electrophoretic features obtained in mice with the natural scrapie isolates, a lower apparent molecular weight of the unglycosylated PrPres was observed in Tg(OvPrP4) mice inoculated with ovine BSE. This difference is evocative of the one observed between PrPres in sheep inoculated with BSE and other natural scrapie isolates (Hill et al., 1998; Baron et al., 2000) (Figure 1) and also of those described in mice (Kuczius et al., 1998) or in humans with the so‐called type 4 pattern in vCJD (Collinge et al., 1996). This already suggests that the original biochemical BSE strain properties, characterized by a specific electrophoretic type 4‐like pattern, are maintained in infected Tg(OvPrP4) mice, even after a passage in sheep. Beside these biochemical features, histological analysis on the other mouse brains showed spongiform changes (Figure 2A). PrPsc immunostaining consisted of fine, granular deposits, always associated with vacuolation in the habenula, thalamus and hypothalamus (Figure 2B). More predominantly, plaques with a fibrillar core were observed, isolated or sometimes grouped into small clusters (Figure 2C). Remarkably, these plaques were surrounded by a ring of small vacuoles reminiscent of florid plaques (Figure 2C). In addition, on adjacent sections, Congo Red staining (Figure 2D) demonstrated the amyloid nature of these plaques, further providing evidence that they are similar to the florid plaques described for vCJD in human (Ironside and Bell, 1997). These florid plaques were numerous (Table 1) and were found mostly in the subcallosal region, hippocampus, midbrain and thalamus, and moderately in the cortex and dorsal medulla (Figure 2E and F).
The anatomical distribution of PrPsc florid plaques observed here, as early as the first passage, is similar to the amyloid plaque distribution reported in second passage experiments for bovine PrP transgenic mice infected either with vCJD or BSE (Scott et al., 1999). As a different transcription promoter was used in this last case (endogenous PrP promoter), these similarities suggest that the PrPsc plaque deposit topography depends more on the strain of agent than on the PrP gene expression pattern. Until now, florid plaques were never detected in Tg(OvPrP4) mice (n = 45) inoculated with six different isolates of scrapie (Table 1), indicating that this hallmark is peculiar to BSE agent transmitted to ovine species.
The biological meaning of florid plaques remains unclear. Originally described in wild‐type mice inoculated with three Icelandic sheep (Fraser, 1979, 1983) and in a mouse adapted‐scrapie strain (111A) (McBride et al., 1988), these florid plaques were rare in sporadic CJD (Kopp et al., 1996). No florid plaques were found in other cases of wild‐type mice inoculated from field scrapie cases. This shows that this phenomenon is probably limited to a very narrow range of scrapie strains and is not a usual feature. However, a high density of florid plaques has been reported only for vCJD and has been a decisive criterion in establishing a similar origin in these patients (Will et al., 1996; Ironside and Bell, 1997). This neuropathological marker of vCJD has also been described in macaques experimentally infected with BSE, producing a disease close to vCJD in humans and providing the first link between the BSE agent and the vCJD disease (Lasmezas et al., 1996). In addition, in control experiments still in progress over 500 days p.i., consisting of the inoculation of a BSE isolate in Tg(OvPrP4) mice, only six mice have died so far (437 ± 104 days p.i.). First, biochemical and histochemical analysis did not allow the detection of PrPsc accumulation, suggesting that cattle BSE is not easily transmissible to Tg(OvPrP4) mice. This probably reflects the species barrier effect in this transmission. Florid plaques have never been observed after transmission of BSE or vCJD to mice, even when the mice carried a human or bovine PrP transgene (Collinge et al., 1995; Bruce et al., 1997; Hill et al., 1997; Lasmezas et al., 1997; Scott et al., 1999). In previous studies, sheep experimentally infected with BSE inoculated to wild‐type mice also failed to produce florid plaques. Similarly, in our control study, still in progress, wild‐type mice inoculated with our ovine BSE isolate died from 348 days p.i. and >500 p.i. with spongiosis and PrPsc accumulation, but without florid plaques (Figure 2G and H). Thus, another factor—present in our ovine transgenic mice and common at least to some humans or macaques—could be decisive for the manifestation of the BSE agent into florid plaques. In any case, the correlation between the abundance of florid plaques and BSE contamination remains to be clarified. Since we found a high density of florid plaques in Tg(OvPrP4) mice inoculated with the brain of sheep experimentally infected with BSE, and none in any of the mice inoculated with different scrapie isolates, the presence of florid plaques could be considered in our model as a criterion to discriminate BSE from natural scrapie in sheep.
The sheep experimentally infected with BSE was of the AA136RR154QQ171 genotype. The fact that this sheep and Tg(OvPrP4) mice share the same PrP genotype could be important for the occurrence of florid plaques. Tg(OvPrP4) mice, however, were sensitive to natural scrapie isolated from AA136RR154QQ171 as well as VV136RR154QQ171 and AV136RR154QQ171 genotypes (Table 1), and the genotype did not seem to influence the nature of the PrPsc deposits. We did not attempt to investigate the effect of PrP genotype on florid plaques occurrence, for which brain homogenates of other BSE‐infected sheep would have been required.
Taken together, these results already provide an important improvement to prion strain identification in sheep. Compared with wild‐type mice, our Tg(OvPrP4) mice are more susceptible to sheep scrapie isolates (Crozet et al., 2001) and seem to be able to discriminate BSE agent in sheep. Indeed, the easy recognition of florid plaques following relatively short incubation periods may be an advantageous alternative to the lesion profile method performed in wild‐type mice. Finally, the proper calculation of sensitivity and specificity of florid plaques criterion would require testing of other sheep brain homogenates experimentally infected with BSE (Foster et al., 2001), as well as a larger number of field and experimental scrapie isolates.
The appearance of florid plaques in future inoculations of Tg(OvPrP4) mice with field scrapie could be interpreted in two ways. It might reveal BSE contamination in sheep. Alternatively, it might allow the identification of a natural scrapie strain similar to those that could have been at the origin of the BSE epidemic. In both cases this would be important, as it would allow the identification of a sheep population potentially hazardous for human health.
The Tg(OvPrP4) transgenic mice used in this study selectively expressed in neurons the ovine PrP gene (AA136RR154QQ171 genotype) but not the endogenous murine PrP. The generation of these mice and their susceptibility to scrapie are described in Crozet et al. (2001). The transgenic mice were cared for and housed in an autonomous filtered pressurized enclosure according to the guidelines of the French Ethical Committee (Decree 87‐848) and European Community Directive 86/609/EEC.
Animal inoculation and follow up.
The transgenic 5‐week‐old mice were inoculated intracerebrally with 20 μl of a 10% (wt/vol) brain homogenate in 5% glucose, from Lacaune sheep with an AA136 RR154 QQ171 that had been experimentally infected with cattle BSE (n = 20) and with six natural scrapie isolates (n = 10 per experiment) from different regions in France detected by the French Epidemiological Surveillance Network (D. Calavas, AFSSA‐Lyon, France) (Table 1). The mice were then checked weekly for the presence of clinical signs such as leanness, hunched posture, hindlimb paralysis, equilibrium trouble, plastic tail, prostration, tremors, ruffled fur, abnormal gait and clasping feet.
As soon as a mouse showed one of these clinical signs, it was isolated and monitored daily, until death. Mice were killed when the intensity of clinical symptoms appeared life threatening or in some cases were found dead. The whole brain was quickly removed, and the brain of every second mouse was fixed [paraformaldehyde, 0.1 M phosphate‐buffered saline (PBS) pH 7.4] for immunohistochemical analysis, the other frozen samples being stored at –80°C for western blot analysis as described elsewhere (Crozet et al., 2001).
Detection of abnormal PrP in scrapie‐inoculated transgenic mice.
After being fixed in paraformaldehyde in 0.1 M PBS pH 7.4 for one night at 4°C, the brains were dehydrated through graded alcohols and embedded in paraffin. Sections (5 μm) were collected onto pre‐treated glass slides (Polylysin or StarFrost, Fischer Scientific) and baked overnight at 57°C. The slides were then dewaxed and were submitted to pre‐treatments designed to destroy the normal cellular PrP. These consisted of a formic acid treatment (98%; Merck) for 10 min at room temperature, 121°C for 30 min in water, each step being followed by a 5 min rinse in water, which appeared more appropriate for detection of PrPsc by immunohistochemistry. The slides were treated with proteinase K (20 μg/ml; Roche‐Boehringer) for 10 min at 37°C. All following steps were performed at room temperature (22–24°C). Endogenous peroxidase activity was inhibited with 2% H2O2 (Merck) in 0.1 M PBS for 5 min. Non‐specific antigenic sites were blocked by a 30‐min incubation in blocking reagent (Roche‐Boehringer). SAF84 monoclonal antibody recognizing the human 142–160 PrP sequence (0.5 μg/ml) (kindly provided by Dr J. Grassi, CEA/SPI, Saclay, France) was then applied overnight. The brain sections were rinsed before detection of the primary antibody using the ABC system (Vector). These steps were followed by rinsing in 0.1 M PBS, and the peroxidase was finally revealed by incubating the sections in 0.1 M PBS containing aminoethylcarbazole (AEC; Dako) or diaminobenzidine (DAB; Clinisciences) to give red or black deposits, respectively. The slides were weakly counterstained with aqueous hematoxylin before mounting (GelMount).
Congo Red staining.
Congo Red staining was used to reveal specifically the amyloid nature of the plaques. Dewaxed slides were coloured with Harris hematoxylin for 10 min. Congo Red (1%, 30 min), lithium carbonate (15 s), 80% ethanol (30 s) were then applied successively. The slides were dehydrated and mounted.
We thank Denise Aubert for transgenesis (Ecole Normale Supérieure, Lyon), and Michèle Lavoine and Annie Kato for excellent technical assistance. This work was supported by grants from the French Programme de Recherche sur les ESST et les prions. Carole Crozet was financially supported by a grant from AFSSA (Agence Française de Sécurité Sanitaire des Aliments) and from the Fondation pour la Recherche Médicale.
- Copyright © 2001 European Molecular Biology Organization