Classical BSE was first described as a novel neurological disease in cattle, which displayed overt clinical signs characterised by changes in behaviour, sensation and locomotion , which led to reporting of many clinical cases in the UK . Atypical BSE, by contrast, has so far not been confirmed in clinical suspects: it has been found in either apparently healthy cattle at abattoirs, cattle that were subject to emergency slaughter or in fallen stock and may suggest that the clinical presentation is different to C-type BSE. Indeed, previously published experimental transmission studies of atypical BSE to cattle conducted elsewhere describe H-type and L-type BSE as diseases characterised by dullness, low head carriage, inactivity or lack of nervousness whereas cattle with classical BSE presented with nervousness or anxiety and hypersensitivity to various external stimuli [19, 20, 26, 27]. However, hypersensitivity to facial tactile or other external stimuli was also observed in H- or L-type BSE although the study did not use unchallenged controls for comparison with clinically 'normal' cattle [19, 27]. The clinical signs in the study reported here resembled those observed in the German study , with no difference in the clinical presentation between H-type and L-type BSE, but separation from other cattle and low head carriage suggestive of dullness were not a consistent finding. In fact, the majority of cattle in the current study showed no signs of dullness, even on closed circuit television observations, but were over-reactive, characterised by over-reactivity to tactile facial and visual stimuli (menace response testing) and displayed unexpected startle responses throughout the clinical phase. Cattle intracerebrally inoculated with classical BSE were not included but previous transmission studies of classical BSE in Holstein-Friesian cattle conducted at the Veterinary Laboratories Agency confirmed that over-reactivity and nervousness were a feature of classical BSE [19, 20, 30]. There is, however, variation in the severity of behavioural and sensory changes in cattle intracerebrally infected with classical BSE, particularly influenced by the frequent handling of cattle in experimental studies, which may lead to habituation to clinical tests. In the authors' opinion, the 'nervous syndrome' of cattle inoculated with atypical BSE has some resemblance to the clinical presentation of classical BSE in some cattle . Notably different to classical BSE were, however, the absence of tremor in all atypical BSE cases, both H- and L-type, and the presence of difficulty in rising, accompanied by dysmetria, early in the disease. In the experimental cases produced in the Italian study (L-type BSE), difficulty in rising and ataxia were surprisingly rare and not observed, respectively , even though the inoculum in the present study was also derived from an Italian BASE case.
We previously employed identical clinical methods to describe ovine scrapie transmitted to cattle, which also presented with a dull (low head carriage and head resting against objects) and nervous form (nervousness and over-reactivity), and all cattle developed difficulty in rising . Different prion strains in cattle appear to produce two main clinical phenotypes (dull and nervous form), with classical BSE being the most uniform and - with respect to over-reactivity - clinically the most extreme phenotype.
None of the cattle inoculated with atypical BSE in the current study progressed to permanent recumbency, which has been reported to occur within few days after the display of evident ataxia , even though they all had difficulty in rising. If this is a consistent feature of atypical BSE whereas over-reactivity is present in a milder form than classical BSE, it would explain why case reports of naturally occurring atypical BSE describe this disease in downer cows [14, 17, 31].
Our previous recommendation to detect classical BSE cases in recumbent casualty slaughter cattle on clinical grounds using the display of repeated startle responses to tests of over-reactivity and abnormal limb position in combination with low heart rate or poor body condition as criteria  may not have detected the atypical BSE cases in the present study but they were culled prior to permanent recumbency. None of the affected cattle were bradycardic at clinical end-stage (heart rate not determined for case H-2) as reported separately  but unexpected startle was observed. Hence, atypical BSE should be considered as differential diagnosis in any recumbent animal, which - based on naturally occurring cases  - is 8 years of age or older and has a history of or presents with over-reactivity to various external stimuli.
The clinical duration of atypical BSE in the current study was 4-10 months for L-type BSE and 2-7 months for H-type BSE, which is considerably longer than reported by others [19, 20, 27] except for an experimental H-type BSE study in Japan . As the survival times or incubation periods (time from inoculation to cull at clinical end-stage) and pathology were similar, the discrepancy with other studies may be explained by the difficulty in defining the exact clinical onset, which is made retrospectively and based on the presence of various clinical parameters, which may overlap with those seen in 'normal' cattle. For example, the presence of marked over-reactivity to touch, which diminished on subsequent observations, was also seen in control cattle. However, startle responses to unexpected (familiar) stimuli were not seen in controls and hence - when displayed in combination with other signs, e.g. head shyness - were interpreted as sign indicating clinical onset. Similarly, difficulty in rising was only seen in inoculated cattle and thus also used as a marker for clinical onset. On the other hand, it is possible that the differences in the clinical duration were attributable to the particular cattle breeds used in each study. Indeed, a significantly shorter clinical duration (and incubation period) was reported for Holstein-Friesian cattle compared with Alpine brown cattle intracerebrally inoculated with BASE brain . Analysis of the clinical picture of classical BSE using surveillance data has shown that some statistically significant differences in the clinical sign frequency may exist between different herd types (beef and dairy cattle) and different breeds of dairy cattle , which may be due to different methods of management or due to the different temperaments of cattle. However, it is difficult to conclude that there are indeed breed-specific differences in the clinical presentation since the display of individual signs can generally be variable, even if a detailed clinical examination is conducted , and depend on other factors, such as duration of illness and age at onset .
PrP genotyping revealed that all cattle had six octapeptide repeats and no novel polymorphisms were detected. The only mutations detected were 'silent' in that they do not affect the PrP protein sequence. The lack of wild type sequences and the comparatively high number of cattle with the P113 silent mutation (found in only three of 118 Holstein-Friesians examined in the UK ) was unusual but may be due to the breed studied; its occurrence was higher in a study of predominantly beef breeds in the United States .
The development of the dull syndrome in only those cattle that had the longest clinical duration but previously exhibited the nervous syndrome may suggest that all infected cattle might have eventually developed the dull syndrome. Based on the range of clinical signs displayed by all cattle prior to cull, which was considered to be the clinical end-point, there was however no evidence that the cattle exhibiting the dull syndrome were in a more advanced stage of disease than the cattle exhibiting the nervous syndrome at cull. The display of either the nervous or dull syndrome at end-stage in both inoculation groups did not appear to be influenced by the prion protein genotype: both syndromes were displayed by cattle with the same genotype. Equally, the observed differences did not appear to have been attributable to gender. As the majority of the brain from both cattle with the dull syndrome was frozen, it was not possible to determine whether the pathological phenotype was different; accumulation of the disease-associated prion protein (PrPd) at the level of the obex was similar for all cases. The lack of association between clinical presentation and PrPd accumulation in the brain has been previously described for scrapie in cattle  and goats .
Contrary to the findings reported for BASE/L-type BSE in the Italian study, we did not observe muscle atrophy preceded by fasciculations, which were attributed to a lower motor neuron disease . Serum CK and AST, which are useful in the diagnosis of diseases of the muscle, even though they may not provide information about the origin (myopathy or neuropathy) , were only altered in two cattle that had muscular damage due to trauma. The finding of behavioural, sensory and locomotor signs in cattle in the present study is suggestive of a brain disease, with dysmetria likely to be caused by cerebellar dysfunction. The 'panic attack' of one L-type BSE-inoculated steer, which thrashed its legs about whilst in lateral recumbency, could even be interpreted as cerebellar convulsions triggered by sudden disturbance . However, other cerebellar signs, such as an abnormal menace response or intention tremor, were not observed. The dysmetric limb movements were also not considered to be so severe that they would have caused difficulty in rising - cerebellar diseases do not cause loss of muscle strength  - although rising with the forelimbs first, which was seen occasionally, is suggestive of some inability to coordinate the movements to stand. The hind limb joints were inspected in five atypical BSE cases and both controls and - with the exception of one H-type BSE case (H1) that presented with mild joint degeneration of both stifles - none presented with abnormalities that would have explained the observed gait abnormalities. PrPd immunolabelling was observed throughout the spinal cord of both L-type and H-type BSE cases, although this also occurs in experimentally infected, classical BSE cases without causing difficulty rising . H-type BSE cases presented with marked PrPd accumulation in glia within the white matter at every level examined but there was no apparent difference in the observed gait changes between L-type and H-type BSE cases.
Elevated blood manganese has been reported for cattle experimentally infected with classical BSE . A similar phenomenon was not observed in the present study where two L-type BSE cases had abnormally low levels of blood manganese, similar to one control.
For the purposes of active screening and confirmatory diagnosis it is reassuring that the obex is involved in both H- and L-type BSE, even after intracerebral inoculation of cattle, but it is concerning that discriminating phenotypically between the different forms is not necessarily straightforward if only the brainstem (obex) is available, despite significant qualitative differences in other areas of the brain. This may in part be due to the large quantities of PrPd which have accumulated as a consequence of this experimental challenge being intracerebral, since immunolabelling in the brainstem of field cases has been reported to be much less, for the L-type in particular [11, 42], although it may also be a result of these animals being allowed to progress to clinical end-stage, rather than being killed for other reasons, and possibly detected at an earlier point in disease development.
However, in this study the immunohistochemical labelling of H- and L-type BSE differed substantially from each other and from C-type BSE in other areas of the central nervous system. Greater qualitative discrimination was possible examining more rostral areas of the brain, and particularly the cerebellum, despite this area not necessarily displaying the greatest amount of PrP . This offers the possibility of amending routine surveillance sampling to include cerebellum taken through the foramen magnum, as is currently required for sheep, thereby enhancing the discriminatory potential of the confirmatory testing. The presence of immunolabelling in muscle spindles and the trigeminal ganglion is consistent with C-type BSE. Current study design does not allow ascertainment of whether this happens at an early stage in the disease, or as a centrifugal spread in end-stage disease as is speculated for C-type BSE [28, 44, 45].
The absence of detectable involvement of the lymphoreticular and enteric nervous system is consistent with many observations for C-type BSE, and other reports of experimental H-BSE  although for C-type BSE it has also been demonstrated that infectivity can be detected in the absence of identifiable PrPd in in vitro tests [46, 47]. Besides, all cattle were challenged intracerebrally, thus by-passing the gastro-intestinal tract and possible uptake of the agent through the gut.
Interestingly, the H-type BSE challenged animals in this study did not develop plaques like the animals in another experimental challenge study conducted in Japan , which also displayed shorter incubation periods and a more restricted range of clinical signs. Unlike the Japanese study, we only used one antibody - the rat monoclonal R145 which recognises the C-terminal of PrP. However, the range of antibodies used in the Japanese study all demonstrated strongly immunolabelled plaque formations, despite more variation in intracellular labelling, so it would appear to be a specific morphological difference between the animals in these studies.
The very small number of animals represented in these studies means that it can only be speculated whether such changes reflect a degree of natural variation in the donor animals, or whether there are as yet unrecognised influences from recipient host factors.
When Western immunoblotting is used for diagnostic or confirmatory purposes, H-type BSE can readily be identified by its characteristic profile even when the signal is strong. In contrast, the characteristics of L-type BSE appear to be more subtle and the results here suggest that when the signal is too intense it is not readily differentiated by visual examination from a classical BSE profile. This observation has led to the routine dilution of samples from any positive BSE case that is subjected to Western immunoblotting under UK surveillance in order to maximise the opportunity for initial identification of this form of atypical BSE and to allow such samples to be subjected to further investigation for full characterisation.