Brain samples were obtained from cattle experimentally inoculated (IC) with various isolates of classical and atypical BSE. Animals were euthanized upon development of clinical signs (see Table 1 for a comparison of incubation times). To confirm that strain-specific properties of each inoculum were transmitted to the experimental animals, obex tissue from selected animals was PK-digested, and PrPSc was detected on a western blot with the 6H4 antibody (Figure 1A). H-type and L-type samples are positively identified by differences in electrophoretic mobility on the blot as compared to classical BSE, and the ratios of the glycosylated forms of PrPSc for each sample are also consistent with the expected patterns for each strain. The appearance of PrPSc from the experimentally-infected animals was consistent with that of inoculated material (data not shown).
In order to rapidly and directly measure strain stability in BSE-infected brain, we utilized a method in which the fraction of PrPSc remaining at increasing GdnHCl concentration is measured with a commercial ELISA that specifically recognizes PrPSc (but not PrPC or denatured PrPSc). This methodology has previously been used to distinguish between different isolates of scrapie in sheep brainstem , but until now has not yet been applied to cattle. Here, cattle brainstem homogenate was incubated with 0.25-4 M GdnHCl, and the proportion of PrPSc remaining at each concentration was detected by the ELISA after dilution of the samples to low final GdnHCl concentration.
Comparison of classical, H-type, and L-type BSE PrPSc by the stability assay
Initially, classical (or C-type) BSE was compared with other TSEs in their natural animal hosts by our standard stability assay. The stability curve for classical BSE samples was found to be approximately sigmoidal, consistent with curves obtained for other TSEs (Figure 1B) [12, 14, 33]. The stability of an elk CWD isolate  was similar to the stability of classical BSE, whereas a United States isolate of sheep scrapie (136-VDEP; ) had a lower stability, demonstrating the ability of this methodology to distinguish between isolates with different stability profiles (Figure 1B). Similarly, the stability of the 136-VDEP isolate (Figure 1B inset, closed circles) is distinct from the stability of a different U.S. scrapie isolate, 13–7 (Figure 1B inset, open circles), in this assay [14, 35]. These results are consistent with previously reported results in mice; previous studies conducted by others determined that the [GdnHCl]1/2 values of (protease-resistant) PrPSc from transgenic mice inoculated with BSE or sheep scrapie were 2.8 and 2.2 M, respectively . Results for scrapie might be expected to vary depending on the specific isolate inoculated.
As compared to classical BSE PrPSc, PrPSc from cattle infected with L-type BSE from French and Italian (BASE) sources exhibited similar behavior in the steep region of the stability curve; however, some samples (derived from the French L-type inoculated cattle) exhibited a striking increase in ELISA signal at low GdnHCl concentrations (Figure 1C). While the specific cause of the increased ELISA signal at low GdnHCl concentration is not understood, the simplest explanation for this observation is either increased binding to the capture surface and/or enhanced exposure of the epitope for the detection antibody. With the exception of the effects noted in the upper baseline of the curve, the stability curve profile of classical BSE (Figure 1C, closed circles) was indistinguishable from that of L-type BSE (averaged over all L-type isolates; Figure 1C, closed squares). Due to the curve shape and the deviation from a perfect logistic fit, [GdnHCl]1/2 values (or the [GdnHCl] required to reduce the level of the PrPSc-specific signal by half) were determined for isolates based on a smooth curve fit (Table 2). [GdnHCl]1/2 values for classical (2.9 ± 0.04 M) and L-type (2.9 ± 0.05 M) were also indistinguishable. This result is in agreement with that from a single isolate of Japanese L-type BSE in a different, proteinase K-dependent assay; when the stability of Japanese L-type BSE was compared to classical BSE, [GdnHCl]1/2 values were not significantly different (3.1 ± 0.1 M and 2.9 ± 0.3 M, respectively) .
Since many of the samples of H-type brainstem, as well as of H-type colliculus, from our experimental inoculation exhibited lower absorbances in the ELISA-based assay, the final concentration of BH in the assay was increased to 10% for a comparison with classical BSE. Some modifications in assay conditions (Buffered Method) were used to control potential variables across the experiment (including effects of brain pH and other brain components on the reaction). Using this methodology, an increase in the stability of H-type BSE (Figure 1D, open circles) was observed as compared to the stability of classical BSE (Figure 1D, closed circles). While this observed increase was small, the average [GdnHCl]1/2 value for three independent classical BSE samples (3.18 ± 0.02) was significantly different than the average [GdnHCl]1/2 of 3.40 ± 0.03 for the three independent H-type BSE samples (student’s unpaired t-test, p = 0.003; Table 2). Therefore, the increased size of the PK-resistant core in H-type PrPSc is associated with a somewhat higher stability as determined by GdnHCl unfolding. We note that this is merely an association, and the increased size is not necessarily causative with regard to the increased stability.
Stability properties of E211K PrPSc
Next, we examined the properties of BSE PrPSc obtained from passage of an H-type isolate of BSE from a cow with a polymorphism at codon 211 of PRNP (U.S. 2006 BSE case) to a calf also heterozygous for the E211K polymorphism . (As this PrPSc may contain a mixture of E211 and K211 protein, we refer to this isolate as E211K and cattle carrying the mutation as EK211). This isolate serves as a potential example of a genetic TSE in a non-human species, as well as an additional example of an H-type BSE transmission. PrPSc material from this passage has a western blot profile consistent with that of H-type BSE . Using our standard assay conditions, a higher stability of the passaged E211K PrPSc than for classical PrPSc from the brain of a wild-type calf was observed (data not shown). A test of 20% classical BH (Animal #3) and EK211 BH (Animal #25) on pH paper indicated that the pH of each was slightly lower than the pH of our 1X PBS (originally pH 7.4); while this suggests that acidification was present in the brain tissue, we note that no differences in pH between the homogenates were evident in this crude assay.
To further confirm this result while controlling for the potential pH variable, EK211 brain sample (Animal #25) was homogenized in a version of PBS buffer with 10X buffering capacity and compared to a PrP wild-type calf that had been inoculated with classical BSE (Animal #26) in parallel; an increase in stability of the EK211 PrPSc was observed in this version of the assay as well (Figure 2A). We note that we are limited in our ability to assess the statistical significance of this result on its own, as only one biological sample of passaged E211K is available worldwide. However, this is the first assessment of the PrPSc stability of a purported hereditary TSE in a non-human species.
Previous Proteinase K-western blot analysis of regions of the EK211 (Animal #25) brain identified a difference between the pattern of PrPSc from the obex and from the cerebellum, with an additional 23 kDa band apparent in the EK211 cerebellum only (which is not observed in other H-type samples) . We compared EK211 obex (Fig. and cerebellum PrPSc in the stability assay, but did not observe a difference between these tissues (Figure 2A, inset).
Additional evidence for the increased stability of H-type BSE PrPSc
Since the pattern of increased PrPSc stability for H-type BSE samples was similarly observed in our H-type E211K isolate, we wanted to provide additional evidence to support our conclusion that this higher stability was statistically significant. Therefore, we utilized a slightly different experimental condition (Intermediate Method) paired with the use of more data points (from 0.25-5 M GdnHCl) with the aims of better establishing the difference between the [GdnHCl]1/2 values and better defining the stability curve for our samples. Since this method necessarily involved the use of a lower final concentration of BH, we identified two H-type samples with strong enough signal in the ELISA for analysis (Animals #6 and #7, representing U.S. H-type BSE) and compared their stability curves to those of three classical BSE isolates. We combined this data with data collected from animal #25 (E211K BSE) as a third example of H-type BSE tested under the intermediate experimental conditions, as well as with animal #26 (as another example of classical BSE) to generate the curves in Figure 2B. Using our previous methodology (Table 2), the average [GdnHCl]1/2 value from the classical BSE group (2.9 ± 0.1 M) was significantly different from the average [GdnHCl]1/2 value from the H-type BSE group (3.5 ± 0.1 M) (student’s unpaired t-test, p = 0.02).
Since a logistic fit has been used to analyze GdnHCl stability curves in previous studies (i.e. [9, 12]), the average H-type curve (Figure 2B, open circles) and the average classical type curve (Figure 2B, closed circles) are also displayed here with 4-parameter logistic fits to further demonstrate the difference between the isolate types. Finally, since much of the deviation from a logistic fit occurs between 0.25 and 2 M Gdn for our samples, we also normalized these curves to the ELISA value at 2 M GdnHCl and performed logistic fits (Figure 2B, inset; average H-type curve in open circles and average classical curve in closed circles). When we performed the same normalization on the curves from the independent biological replicates, the calculated [GdnHCl]1/2 values from these fits were 3.49 and 3.42 for the H-type BSE isolates and 3.06, 2.96, and 2.95 for the classical BSE isolates. In combination with the data and analysis of Figure 1D, we believe that this provides additional confirmation of our reported higher stability of H-type BSE.
Stability comparison of BSE to other cattle-passaged TSEs
In addition to comparing the different BSE strains, we also used the stability assay to characterize the biochemical properties of other TSEs passaged into cattle. Scrapie and CWD are both transmissible into cattle by IC inoculation, leading to PrPSc accumulation--but not significant spongiform changes--in the brain [29, 30]. Transmissible mink encephalopathy has been hypothesized to have originated from the feeding of downer cattle, possibly carrying atypical, L-type BSE, to farm-raised mink . We wanted to determine if the profiles of PrPSc from these TSEs passaged in cattle brain were distinguishable from each other or from other BSE strains, with potential implications for understanding strain origins and/or improving (non-BSE) TSE diagnosis in cattle.
Stability curves for white-tailed deer-derived CWD on the first passage into cattle (Figure 2C, open triangles) matched the profile of classical BSE PrPSc (Figure 2C, closed squares), which is not surprising due to the similarity of the stability of classical BSE and CWD in their natural hosts (Figure 1B). For TME, samples from a 4th passage of a TME isolate into cattle (Bovine TME; performed in parallel to the BSE strain inoculation experiment) were analyzed. Of the four biological replicates analyzed, one sample exhibited a significant deviation from the baseline (Figure 2C, inset). Possibly, this reflects differences in higher-order PrPSc structure, or another biochemical difference in this sample. The other three TME samples were averaged to create the stability curve in Figure 2C (open diamonds), which matches the average curve for classical BSE (Figure 2C, closed squares). [GdnHCl]1/2 values for CWD in cattle (3.0 ± 0.04) and for Bovine TME in cattle (2.9 ± 0.03) were indistinguishable from that of classical BSE. To determine if the stability of TME had changed upon repeated passage into cattle, we compared this 4th passage stability curve to that of cattle infected directly with mink TME. The 1st passage of TME from mink into cattle generated PrPSc with a stability profile consistent with both 4th passage TME and classical BSE (Figure 2C). We note that the previous studies of passage of TME into cattle did not note a decrease in disease incubation time between the original and second passages into cattle, suggesting little or no species barrier between the mink TME and the cattle prion protein .
In the case of sheep scrapie inoculated into cattle, the average curves over three available animals (Figure 2C, closed circles) showed a small increase in stability ([GdnHCl]1/2 = 3.3 ± 0.03 M) as compared to the stability of classical BSE (Figure 2C, closed squares; [GdnHCl]1/2 = 2.9 ± 0.04). The [GdnHCl]1/2 values from this experiment were significantly different using the unpaired student’s t-test at p < 0.05. However, due to the comparatively low absolute ELISA values for two of the bovine scrapie samples, we elected to conduct additional analysis using the intermediate method (see Methods) to better compare the stability of bovine scrapie to that of classical BSE with curves conducted in parallel on the same plate. Again, a small increase in stability was observed for the scrapie isolate in cattle brain (Figure 2D, open circles; [GdnHCl]1/2 = 3.6 ± 0.2) as compared to the stability of classical BSE in cattle brain (Figure 2D, closed circles; [GdnHCl]1/2 = 3.2 ± 0.1). However, a higher than average standard error of the mean was observed in this experiment for the classical BSE data set, and the difference in stability was significant at the p = 0.10 but not the p = 0.05 level. One major challenge of the bovine scrapie analysis was the issue with low signals in some of the other available samples attempted, limiting our sample population for analysis. Therefore, we suggest that it is likely that the bovine scrapie PrPSc has a slightly higher stability than classical BSE, and that further analysis of these isolates upon future passage would be instructive for continued investigation of these relationships.
Previous work by others demonstrated that the stability of BSE PrPSc ([GdnHCl]1/2 = 2.78 M) when passaged into transgenic mice expressing bovine PrP was significantly higher than the stability of scrapie PrPSc passaged into these mice ([GdnHCl]1/2 = 2.25 M)  even though the stability of PrPSc from the original scrapie-infected sheep brain material was similar to the stability of PrPSc from the BSE-infected cattle brain material (resistant to > 3.0 M GdnHCl). The original conclusion was that bovine scrapie PrPSc has a lower stability than bovine BSE PrPSc; however, our results lead us to propose that different isolates of scrapie induce different stability profiles following passage in cattle. Alternatively, this result could be due to differences in the molecular properties probed by the two different versions of the GdnHCl stability assay. We have observed changes in the stability profile of some scrapie isolates when compared in the presence and absence of proteinase K digestion (C.E. Vrentas, unpublished), and the same phenomenon has been observed in human CJD strains . The original scrapie inoculum for the cattle in this study, which consisted of a pool of nine U.S. field cases from 1992, was not available for testing. However, as the stability of the U.S. 13–7 scrapie isolate in sheep brain was similar to the stability of BSE but not of the U.S. 136-VDEP isolate (compared under parallel conditions, data not shown), we propose that the scrapie isolate(s) inoculated into to the infected cattle used here could have similar properties to the 13–7 strain.