We used Tg110 transgenic mice in all inoculation experiments. This mouse line expresses bovine PrP^C (≈8× that of the level of PrP^C in cattle brain) under the control of the mouse prnp gene promoter in a mouse PrP^0/0 background (29).

The 5 BSE-H isolates used in this study comprised brainstem samples from naturally affected cows, diagnosed as atypical H-type BSE on the basis of the molecular analyses of PrP^res (23). All cows were healthy and killed at 8–15 years of age. Four (isolates 07-644, 03-440, 03-2095, and 02-2695) were provided by the Agence Française de Sécurité Sanitaire des Aliments (Lyon, France). Isolate 45 was obtained from the Polish National Veterinary Research Institute (Pulawy, Poland). For comparative studies, material obtained from brainstem of 1 cow naturally infected with BSE-C (RQ 225:PG1199/00), supplied by the Veterinary Laboratories Agency (New Haw, Addlestone, Surrey, UK), was used as BSE-C control. For mouse inoculations, all isolates were prepared from brain tissues as 10% (wt/vol) homogenates. For subpassages, 10% brain homogenates from Tg110 mice collected from primary passage were used as inocula.

All inocula were prepared in sterile 5% glucose as 10% homogenates. Each inoculum was prepared separately in a biosafety cabinet according to a strict protocol to avoid cross-contamination. To diminish the risk for bacterial infection, the inocula were preheated for 10 min at 70°C before inoculation.

Groups of 6–12 mice (6–7 weeks of age, weighing ≈20 g) were inoculated with 20 µL of the appropriate sample in the right parietal lobe by using 25-gauge disposable hypodermic syringes. UNO MICRO ID-8 ISO transponders (Roestvaststaal BV, Zevenaar, the Netherlands) were used for individual identification of mice. After inoculation, mice were observed daily and their neurologic status were assessed 2×/wk. When progression of the disease was evident, animals were euthanized. All animals were housed in accordance with guidelines of the Code for Methods and Welfare Considerations in Behavioral Research with Animals of the European Union directive 86/609EC). Necropsy was performed, and brain and spleen were taken. Part of the sample was frozen at −20°C for biochemical analysis, and the remaining part was fixed for histopathologic studies.

Survival times were calculated for each inoculum as the time between inoculation and euthanasia in days and expressed as the mean of the survival days postinoculation (dpi) of all the inoculated mice with its correspondent standard error of the mean. Data were processed by using SigmaPlot 2001 software (Systat Software, San Jose, CA, USA).

Frozen mouse brain samples were prepared as 10% (wt/vol) homogenates in 5% glucose in distilled water in grinding tubes (Bio-Rad, Hercules, CA, USA) by using a TeSeE Precess 48 homogenizer (Bio-Rad) following the manufacturer’s instructions. All samples were analyzed by Western blot by using the kit TeSeE Western Blot 355 1169 (Bio-Rad) but with some adjustments for the different amounts of samples used. To achieve the volume proposed in the manufacturer’s recommendations, 100 μL of the brain homogenates to be tested were supplemented with 100 μL of a 10% brain homogenate from PrP null mice (30). Processed samples were loaded on Criterion 12% polyacrylamide gels from Bio-Rad (165.6001) and electrotransferred to immobilon membranes (IPVH 000 10 [Millipore, Billerica, MA, USA]). For the immunoblotting experiments, Sha31 (31), Saf84 (Cayman Chemical, Ann Arbor, MI, USA) and 12B2 (32) monoclonal antibodies (mAbs) were used at a concentration of 1 μg/mL. Sha31 recognizes ₁₅₆YEDRYYRE₁₆₃ epitope, and Saf84 recognizes ₁₇₁QVYYRPVDQYS₁₈₁ epitope and 12B2 recognizes ₁₀₁WGQGG₁₀₅ epitope of the bovine PrP sequence. Immunocomplexes were detected by horseradish peroxidase–conjugated antimouse immunoglobulin G (Amersham Pharmacia Biotech, Piscataway, NJ, USA). Immunoreactivity was visualized by chemiluminescence (Amersham Pharmacia Biotech) and obtained after exposition with medical radiographic film (Agfa, Mortsel, Belgium).

All procedures involving mouse brains and spleens were performed as described (33). Briefly, samples were fixed in neutral-buffered 10% formalin (4% formaldehyde) before embedding in paraffin. Once deparaffinized, 2-µm–thick tissue sections were stained with hematoxylin and eosin. Lesion profiles of the brains were established according to the standard method described by Fraser and Dickinson (34). For paraffin-embedded tissue blots, the protocol described by Andréoletti et al. (35) was used.

The brains of inoculated mice were examined for PrP^res by Western blot analysis with Sha31 mAb. Consistent with the efficient transmission observed, PrP^res was readily detected from the first passage in all Tg110 mice inoculated with the different BSE-H isolates. In 3 BSE-H isolates (07-644, 03-440, and 03-2095), the totality (100%) of the inoculated Tg110 mice produced a PrP^res profile similar to that in cattle (H-type PrP^res) but clearly distinct from that produced by BSE-C agent (C-type PrP^res) in cattle (Figure 2, panel A; data not shown). Compared with C-type PrP^res, H-type PrP^res was characterized by a significantly higher apparent molecular mass of the unglycosylated band. The results obtained with these 3 isolates were comparable with those obtained by other authors with 2 BSE-H isolates inoculated in a different bovine PrP mouse line (27), where they concluded that the H-type agent essentially retained its biochemical phenotype upon serial transmission to bovine PrP transgenic mice.

We observed a different situation for the other 2 BSE-H isolates (02-2695 from France and 45 from Poland), where 3 and 2, respectively, of infected mice (Figure 1) showed a PrP^res profile clearly distinct from that of BSE-H in cattle (Figure 2, panel B). These 5 mice produced a PrP^res with lower (≈1.5 kDa) apparent molecular mass of the 3 PrP^res glycoforms, which was indistinguishable from that produced by BSE-C agent in these mice (Figure 2, panel B, and data not shown). Further characterization of this PrP^res with other antibodies showed that the PrP^res produced by these mice was not recognized by 12B2 mAb (Figure 3). However, 12B2 immunoreactivity against the H-type PrP^res produced by other mice (inoculated at the same time with the same isolates) remains essentially similar to that in cattle BSE-H (Figure 3). Furthermore, PrP^res immunolabeling with Saf84 mAb showed that these mice, contrary to mice with H-type PrP^res, did not retain the characteristic PrP^res band profile (4 bands) of cattle BSE-H but showed a PrP^res profile (3 bands) similar to that of BSE-C in Tg110 mice (Figure 4).

In addition, the proportion of the diglycosylated PrP^res increased in comparison with the mice with H-type features, as shown by using Sha31 mAb (Figure 2, Figure 3). Another difference was that the PrP^res level in brain was much higher than that in mouse brains with H-type PrP^res but similar to that in mouse brains inoculated with BSE-C, as shown by comparative Western blot analysis by using Sha31 mAb (Figure 5) and by the different equivalent brain tissue masses loaded to obtain equivalent PrP^res signals (Figure 2, Figure 3, Figure 4). A 10-fold equivalent brain tissue mass was loaded for brains from mice showing H-type PrP^res molecular profile than from those with a classical-like (C-like) PrP^res molecular profile to obtain equivalent PrP^res signals (Figure 2, Figure 3, Figure 4). These mice thus showed PrP^res molecular features indistinguishable from those in Tg110 mice infected with C-like features.

For these 2 isolates, a second passage was performed in Tg110 mice by using a brain homogenate derived from a mouse with either H-type or C-like PrP^res (Figure 1). Survival times did not differ substantially when the different inocula were compared (H-type vs. C-like PrP^res brain homogenate) or when compared with second passages of BSE-C in these mice. All the mice inoculated with the H-type brain homogenates showed H-type PrP^res features, whereas all the mice inoculated with the C-like brain homogenates exhibited C-type PrP^res molecular features indistinguishable from that of BSE-C (data not shown).

We next examined vacuolation and PrP^Sc distribution in the brain, which are known to vary by strain (10,34,38). In general, mice brains exhibiting H-type PrP^res correlated with overall more intensive vacuolation that is pronounced in areas such as the hypothalamus, medial thalamus, and mesencephalic tegmentum (Figure 6). However, a different situation was observed when we studied the brains of BSE-H–infected mice exhibiting C-like PrP^res features. All these mice showed a lesion pattern comparable with that in BSE-C–infected mice, in which slight differences are found only in the mesencephalic tegmentum (Figure 6). These differences consisted of moderate lesions, whereas in BSE-C–infected mice, no lesions were found in this area. These features were conserved on secondary transmissions where no remarkable differences were found when compared with primary transmissions (data not shown).

Moreover, PrP^Sc deposits were distinctly distributed after both primary and secondary transmissions when BSE-H–infected mice exhibiting C-type PrP^res features were compared with those with H-type PrP^res, as assessed by paraffin-embedded tissue blot on brain coronal sections (Figure 7). However, the PrP^Sc deposition patterns were clearly similar when these mice were compared with those infected with BSE-C, after both primary and secondary transmission.

Transmission experiments (27) showed that, contrary to BSE-C, BSE-H is poorly lymphotropic in mouse models. The comparative study of PrP^Sc accumulation in spleen from our Tg110 mice infected with both agents showed that BSE-H infected mice exhibiting H-type PrP^res in their brain were consistently scored as negative for PrP^Sc detection by paraffin-embedded tissue blot. In contrast, clear PrP^Sc deposits were always detected in BSE-H–infected mice exhibiting C-like features, as in mice infected with BSE-C (Figure 7). Similar results were obtained after secondary transmissions.