Prions in the peripheral nerves of bovine spongiform encephalopathy-affected cattle

Bovine spongiform encephalopathy (BSE) is a fatal, neurodegenerative disorder of cattle. It belongs to a group of diseases called transmissible spongiform encephalopathies (TSEs), which also include scrapie in sheep and goats and CreutzfeldtJakob disease (CJD) in humans (Prusiner, 1991). TSEs are characterized by spongiform changes in the central nervous system (CNS) (Fraser, 1979; Masters & Richardson, 1978) and by the accumulation, principally in the CNS and lymphoreticular system, of PrP^Sc, a pathological, partially protease-resistant isoform of a normal cellular prion protein (PrP^C) (Prusiner et al., 1982). The conversion of PrP^C to PrP^Sc is thought to be central to the pathogenesis of TSEs, and PrP^Sc has been recognized as a major component of the pathogen, normally referred to as a prion. PrP^Sc is generally, but not exclusively, correlated with infectivity and is the only known disease-specific marker (Bolton et al., 1982; Prusiner, 1991).

A variant form of CJD (vCJD) has been detected in the UK and several other countries, and it is thought that this disease has resulted from the consumption of BSE-contaminated beef products (Chazot et al., 1996; Will et al., 1996, 1998; Cousens et al., 1997). As a consequence, BSE is considered a zoonosis. In most countries where BSE-control programmes have been introduced, the organs that have either been predicted by extrapolation from sheep scrapie data or demonstrated to be infectious in cattle are classified as specified risk materials (SRM). These tissues are excluded from the human diet and destroyed, irrespective of the outcome of post-mortem testing for BSE.

Whilst much has been learned about the pathogenesis of BSE in cattle by examination of tissues from experimentally infected cattle, killed at intervals throughout the disease course, infectivity has been detected consistently only in the distal ileum, the CNS and certain peripheral nervous system (PNS) ganglia: dorsal root ganglia (DRG) and trigeminal ganglion (Wells et al., 1998, 1999, 2005). Rarely, in experimentally infected cattle, infectivity has also been detected in bone marrow and tonsil (Wells et al., 1999, 2005). Until recently, in naturally infected clinical cases of BSE, infectivity had been detected only in the brain, spinal cord and retina by bioassay in wild-type mice (Fraser & Foster, 1994; MAFF, 1995). This apparently restricted tissue distribution of the BSE agent in cattle, compared with at least some other TSEs, may partially be a reflection of limitations of the assay sensitivity. Although assays have been conducted by intracerebral inoculation of cattle (Wells et al., 2005), evidence of infectivity in tissues to which humans may be exposed, such as muscle, peripheral nerves and lymph nodes, has remained elusive.

Recently, through the use of transgenic mice overexpressing bovine PrP^C (Tgbov XV mice), infectivity was detected in the brain, spinal cord, retina, optic, facial and sciatic nerves and distal ileum of a naturally infected cow at the terminal stage of BSE (Buschmann & Groschup, 2005). In addition, detection of PrP^Sc has been reported in the PNS and adrenal glands of a natural case of BSE in Japan (Iwamaru et al., 2005). PrP^Sc was also detected in the peripheral nerves of two BSE-positive cows that were detected during surveillance of slaughtered cattle (Iwata et al., 2006). Although Iwata et al. (2006) reported that the BSE-positive cattle were not clinically affected, the clinical signs reported at ante-mortem examination were inclusive of those recorded in British BSE-affected cattle. It is therefore probable that these animals were at least in the early stages of clinical disease.

These data indicate that PrP^Sc and/or infectivity can be detected outside the CNS and distal ileum, at least in the later stages of disease, and that the presence of agent in the PNS may represent a risk to consumers, as PNS structures are not specifically designated SRM, and are thus not removed from the food chain. In order to facilitate more accurate estimations of risk, in the context of control programmes including the testing of cattle entering the food chain, it was felt to be necessary to investigate whether PrP^Sc could be detected in parts of the PNS other than those implicated directly in the hypothetical pathogenetic spread of agent from the intestine to the CNS (McBride & Beekes, 1999; van Keulen et al., 2000; McBride et al., 2001; Wells & Wilesmith, 2004). In particular, it was of interest to determine whether PrP^Sc was present before, or only after, detection in the CNS, or after the onset of clinical signs. In this study, we investigated PrP^Sc accumulation in the PNS and adrenal gland of naturally infected BSE cases from the UK in order to confirm previous results from cattle in Japan. In addition, using samples from cattle in an experimental time-course study, we investigated the temporal relationship between detection of PrP^Sc in the CNS and certain PNS structures following oral exposure to BSE-infected brain material.

In addition, infectivity assays on selected tissues were conducted in transgenic mice expressing the bovine prion protein (PrP) gene (TgBoPrP).

Tissue samples from cattle.
All samples examined in this study were provided by the TSE Archive of the VLA, Addlestone, UK. For confirmation of earlier Japanese results, we examined the PNS: cervical and thoracic DRG (pooled C35 and T79), trigeminal, cranial cervical and thoracic ganglia, sciatic, vagus and splanchnic nerves and the adrenal gland from five clinically suspect, BSE-confirmed cases and five clinically suspect animals that were negative after diagnostic examinations of brain. These served as controls, as they presented with neurological signs consistent with BSE.

In the second part of the experiment, we examined the following tissues: brainstem, spinal cord (segments C12, or C23 and T910), DRG (pooled from segments C35 and T79), stellate ganglion, phrenic and radial nerves and adrenal gland harvested from cattle that were challenged orally at the VLA with either 100 or 1 g BSE-infected brainstem homogenate from clinically affected donors, and culled sequentially. The infectivity titre of the inoculum was determined previously by end-point titration in RIII mice to be 10^3.1 mouse intracerebral and intraperitoneal units ID₅₀ g¹, (M. E. Arnold and others, unpublished data), which is similar to other contemporary titrations in these mice of infectivity of brainstem from clinical cases of BSE (data not shown). Samples collected between 27 and 42 months post-exposure from the 100 g dose group, and between 36 and 51 months post-challenge in the 1 g dose group (Table 1), were selected for testing on the basis of prior knowledge of the time points in the sequential kill studies when PrP^Sc was first detected in the CNS (Wells et al., 1998; M. E. Arnold and others, unpublished data). The selection of material was judged to ensure that some samples were available from time points before PrP^Sc was detected by immunohistochemistry (IHC) in the CNS. Sample sets were not complete for all animals from the original experiments, as supply was dependent on stocks remaining after use in other studies. In total, 376 tissues obtained from 31 cattle challenged experimentally with a 100 g dose of BSE brainstem homogenate, and 14 challenged with a 1 g dose, were examined. A further seven undosed, control cattle, age-matched approximately to challenged animals killed at 27, 30, 33, 36, 42, 44 and 52 months after dosing, provided corresponding control samples.

Table 1. Numbers of inoculated cattle from a sequential kill, time-course pathogenesis study providing tissues for examination, according to time killed after inoculation

Processing of tissue samples.
PrP^Sc was extracted from peripheral tissues by a method described previously (Shimada et al., 2005). Briefly, the PNS tissues and adrenal glands were suspended in 0.8 ml of a detergent buffer containing 50 mM Tris/HCl (pH 7.5), 2 % (v/v) Triton X-100, 0.5 % (v/v) N-lauroylsarcosine (Sarkosyl; Nacalai), 100 mM NaCl, 5 mM MgCl₂, 2 mM CaCl₂, 20 mg collagenase and 40 µg DNase I, and incubated at 37 °C for 2 h with constant rotation with a metal bead (Metal corn; Yasui Kikai). The homogenate was digested with proteinase K (PK; Roche Diagnostics) (final concentration, 60 µg ml¹) at 37 °C for 1 h. PK digestion was terminated with 2 mM 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (Pefabloc; Roche Diagnostics). The homogenate was centrifuged at 68 000 g for 20 min (Optima MAX-E/TLA-100.3; Beckman) at room temperature (RT). The supernatant was discarded and the pellet was suspended in 6.25 % (w/v) Sarkosyl (Sigma) in 10 mM Tris/HCl (pH 7.5) and incubated at RT for 1 h with constant rotation; subsequently, it was centrifuged at 9000 g for 5 min (6200/AF-2730; Kubota). Sodium phosphotungstate was added to the supernatant to a final concentration of 0.3 % (v/v) and incubated at 37 °C for 30 min with constant rotation. Pellets were obtained by centrifugation at 20 000 g for 30 min (6200/AF-2730; Kubota). PrP^Sc was enriched from the brain according to a method described previously (Hayashi et al., 2005). The CNS tissues were homogenized in a buffer containing 100 mM NaCl and 50 mM Tris/HCl (pH 7.6). The homogenate was mixed with an equal volume of detergent buffer containing 4 % (w/v) Zwittergent 3-14 (Calbiochem), 1 % (w/v) Sarkosyl, 100 mM NaCl and 50 mM Tris/HCl (pH 7.6) and then incubated with 0.25 mg collagenase, followed by incubation with PK (final concentration, 40 µg ml¹) at 37 °C for 30 min. PK digestion was terminated with 2 mM Pefabloc. The sample was mixed with 2-butanol : methanol (5 : 1) and then centrifuged at 20 000 g for 10 min (6200/AF-2730; Kubota).

Western blotting (WB) analysis.
The pellets were resuspended in a gel-loading buffer containing 2 % SDS and heated at 100 °C for 6 min. The samples were separated by SDS-PAGE (12 % gel) and blotted electrically onto a PVDF membrane. The blotted membrane was incubated with anti-PrP monoclonal antibody (mAb) T2 conjugated to horseradish peroxidase at RT for 1 h. Signals were developed with a chemiluminescent substrate (SuperSignal; Pierce Biotechnology) (Hayashi et al., 2004).

Infectivity assays.
The transmissibility of infection from brain, vagus nerve and adrenal gland of natural cases of BSE, tissues that were found to contain PrP^Sc, was bioassayed in transgenic (Tg) mice expressing bovine PrP [Tg(BoPrP)4092HOZ/Prnp^0/0; Tg(BoPrP)]. These mice, kindly supplied by Dr S. B. Prusiner, are susceptible to BSE prions and exhibit an incubation period of <250 days (Scott et al., 1997). The tissues were each homogenized in 9 vols PBS by a multi-bead shocker (Yasui Kikai) and centrifuged at 1000 g for 5 min (6200/AF-2730; Kubota) at RT; the supernatant was used as the inoculum. Female Tg(BoPrP) mice (3 weeks old) were inoculated intracerebrally with 20 µl supernatant. After inoculation, the clinical status of the mice was monitored daily to assess the onset of neurological signs. Diseased mice were sacrificed and subjected to PrP^Sc examination as described previously (Yokoyama et al., 2001).

PrP^Sc detection in the PNS and adrenal gland of naturally BSE-affected cattle
Representative results of WB analysis of the vagus nerve and adrenal gland of naturally infected cattle are shown in Fig. 1(a, b). Samples from the diagnostically unconfirmed cattle were negative. The signal intensity from BSE-positive cattle differed between samples and from animal to animal, but typical triple banding was observed in positive tissues, including the adrenal gland and vagus nerve. A weak signal was detected from the sciatic nerve (Fig. 1c). The results of the WB analysis are summarized in Table 2. In addition to the trigeminal ganglion and the mid-cervical and mid-thoracic DRG, PrP^Sc was also detected in the peripheral nerves, adrenal gland and thoracic ganglia.

Table 5). (b) Lanes 110, vagus nerve (100 mg tissue eq.). Sample from animal 19591 (lane 4) was subjected to transmission study (Table 5). (c) Lanes 110, sciatic nerve (100 mg tissue eq.). (d) Lanes 19, midbrain (5 mg tissue eq.). BSE status (shown underneath each gel) was determined by IHC examination of the medullaobex at VLA, Weybridge, UK: +, BSE-positive cattle; , BSE-negative cattle. Lanes ac, mouse scrapie-infected brain was used as the positive control (a, 0.4 µg brain eq.; b, 1.6 µg brain eq.; c, 6.4 µg brain eq.). Molecular markers are shown on the left (in kDa).

Table 2. Numbers of cattle (clinically suspected to be affected with BSE) positive on WB for detection of PrPSc Values are no. cattle (clinically suspected to be affected with BSE) positive on WB for detection of PrPSc/no. animals tested, according to tissue and diagnostic status. Each sample was examined in duplicate and the sample was judged positive on a single positive test result. DRG, Dorsal root ganglia.

PrP^Sc detection in the CNS of cattle challenged orally with BSE-infected brainstem
The WB results for individual animals are given in Tables 3 and 4 according to time after inoculation. The results of WB analysis of brainstem and spinal cord were compared with the BSE status of the cattle, as determined previously by IHC examination of the brainstem. The WB results for CNS tissues were completely in accord with the previous diagnoses based on IHC.

Table 3. Western blot detection of PrPSc from cattle challenged orally with a 100 g dose of BSE-infected brainstem, according to time after inoculation

Table 4. WB detection of PrPSc from cattle challenged orally with a 1 g dose of BSE-infected brainstem, according to time after inoculation

PrP^Sc detection in the PNS and adrenal glands of cattle challenged orally with BSE-infected brainstem
PrP^Sc was detected in PNS by WB analysis, but only inconsistently in those animals diagnosed as BSE-positive on previous CNS IHC examinations (M. E. Arnold and others, unpublished data) (Tables 3 and 4). Furthermore, irrespective of dose, DRG were positive only in animals in which the corresponding level of spinal cord was positive, and PrP^Sc was not always detected at both cervical and thoracic levels of DRG. PrP^Sc was detected in the stellate ganglion in three of six animals of the 100 g dose group that were killed 36 months after inoculation, concurrent with DRG involvement. PrP^Sc was detected in the phrenic nerve in the single animal of the 100 g dose group that was analysed 35 months after inoculation, and in one of six animals of the 100 g dose group that were killed 36 months after inoculation, in which CNS and certain ganglia were also positive. PrP^Sc was not detected in the radial nerve, but was found in samples of the sciatic nerve (one in the 100 g dose group and two in the 1 g dose group), all in animals with positive CNS and ganglia. Whilst PrP^Sc was detected in DRG and stellate ganglion of occasional preclinical animals, PrP^Sc was detected in peripheral nerves only in those animals with definite clinical signs at the time of euthanasia. PrP^Sc was detected in adrenal gland from three samples from the 100 g dose-group cattle, which were also clinically affected. Detection of PrP^Sc was erratic in relation to time post-inoculation (Tables 3 and 4). All tests were negative in challenged and control cattle designated previously as BSE-negative by IHC (Fig. 1d).

PrP^Sc-positive vagus nerve and adrenal gland harbour infectivity
The attack-rate and incubation-period data of the mice inoculated with PrP^Sc-positive vagus nerve and adrenal gland are shown in Table 5. Typical PrP^Sc banding was detected in diseased Tg(BoPrP) mouse brains (Fig. 2). The glycoform pattern and molecular mass of PrP^Sc in Tg(BoPrP) mice were identical to those of the BSE-affected cattle brain.

Table 5. Data on transmission of PrPSc-positive tissues in Tg(BoPrP) mice

(59K): Fig. 2. WB analysis of the brains of Tg(BoPrP) mice. PrPSc was prepared and dissolved in a gel-loading buffer containing 2 % SDS. PrPSc was detected by using the T2 mAb. Lanes: 1, mouse-adapted scrapie; 2, BSE-affected cattle brain; 3, Tg(BoPrP) mice inoculated with BSE-affected cattle brain; 4, Tg(BoPrP) mice inoculated with PrPSc-positive vagus nerve; 5, Tg(BoPrP) mice inoculated with PrPSc-positive adrenal gland; 6, Tg(BoPrP) mice inoculated with normal cattle brain. Molecular markers are shown on the left (in kDa).

This was an opportunistic study, intended to build upon data arising from the Japanese surveillance programme, but also to expand an understanding of the pathogenesis of BSE in cattle in the late incubation period, around the time of initial CNS involvement. The data obtained in this study indicate that PrP^Sc accumulates in the PNS and adrenal glands of BSE-affected cattle and that this accumulation appears to coincide with, or follow rather than precede, accumulation in the CNS. Observations reported previously for the PNS and adrenal gland of a BSE case, and for the PNS of two cows slaughtered for human consumption in Japan (Iwamaru et al., 2005; Iwata et al., 2006), have now been repeated in five clinical cases from the UK by using the methods described here. This highly sensitive WB method has also detected PrP^Sc accumulation in the PNS and/or adrenal gland of five further natural BSE cases detected during the surveillance of fallen stock in Japan (data not shown). These data reinforce the results of Buschmann & Groschup (2005), which confirmed the presence of infectivity in the PNS of a clinical case of BSE in Germany. Based on these results, it seems likely that PrP^Sc accumulation in the PNS during the clinical stages of infection may be the rule in BSE, rather than the exception. PrP^Sc accumulation in the PNS has been reported in naturally and experimentally scrapie-affected sheep and hamsters, and in CJD-affected humans (Groschup et al., 1996, 1999; Beekes et al., 1998; McBride & Beekes, 1999; Haik et al., 2003; Favereaux et al., 2004; Head et al., 2004; Ishida et al., 2005). As shown in Fig. 1, the PrP^Sc signal intensity varied between samples and among cattle. Based on a comparison of the signal intensity in the WB analysis, the amount of PrP^Sc present in the PNS or adrenal glands was estimated to be <1/120 of that present in the brain.

By testing cattle from an experimental sequential kill study completed at the VLA, it was proposed to obtain further data on the correlation of CNS and PNS PrP^Sc detection relative to time after exposure and thereby extend previous studies (Wells et al., 1998, 2005). Limitations on the availability of certain samples, particularly vagus and splanchnic nerves and abdominal autonomic nervous system ganglia, prevented examinations that would have provided opportunities to investigate potential routes of entry of agent to the CNS. Previous studies have demonstrated that PrP^Sc deposition in the CNS of BSE-affected cattle is targeted to certain neuroanatomical areas and is not uniform. In this study, PrP^Sc detection in the brainstem and two levels of spinal cord tested by WB analysis confirmed previous results of the BSE status of experimentally infected cattle, based upon IHC examination of three levels of brainstem (M. E. Arnold and others, unpublished data). Interestingly, PrP^Sc was detected simultaneously in different parts of the CNS (Tables 3 and 4), even in animals that were only in the earlier stages of clinical disease, suggesting perhaps that, once entry has occurred, spread of agent throughout the CNS is a process involving periods that are shorter than the sequential time intervals in the experimental kill study. In this study, PrP^Sc detection in the PNS was an infrequent finding in the late preclinical and the clinical stages of BSE, and the available evidence would be consistent with spread from the CNS to the PNS structures examined (DRG, stellate ganglion, phrenic, radial and sciatic nerves) and adrenal glands. The results did not present inconsistencies that would refute the current understanding of peripheral pathogenesis in models of TSE after oral exposure (McBride & Beekes, 1999; van Keulen et al., 2000; McBride et al., 2001). The apparently restricted involvement of lymphoid tissues (Buschmann & Groschup, 2005; Wells et al., 2005) and enteric plexuses (Terry et al., 2003) in BSE pathogenesis in cattle suggests possible differences from events in experimental models of TSE pathogenesis. Such differences can only be investigated by experiments conducted in the natural host species. It is likely from epidemiological data and experimental studies (Wells et al., 2007) that the 1 g dose group in these experimental exposures of cattle approximates the majority of field-case exposures, perhaps suggesting less relevance to risk management of positive results from tissues taken from an animal that has been dosed with 100 g. However, the data here do not suggest marked differences between the dose groups and are insufficient for 1 g-dosed animals to indicate a reduced involvement of PNS tissues relative to dose or timing of positive results in other tissues. Although sparse PrP^Sc accumulation was observed in the PNS of experimentally infected cattle, PrP^Sc was detected in all PNS tissues examined from natural BSE cases (Tables 2, 3 and 4). This may indicate a progressive involvement of PNS in the clinical phase of disease and has implications for the risk assessment of tissues from such animals should they escape detection at slaughter.

In natural cases of clinically BSE-affected cattle, infectivity, detected by bioassay in wild-type mice, has been found only in the CNS (Fraser & Foster, 1994). In orally inoculated cattle, infectivity has been detected in the brain, spinal cord, distal ileum, DRG and trigeminal ganglion in the late incubation period by wild-type mouse assays (Wells et al., 1994, 1996, 1998). Infectivity was also detected by wild-type mouse assay in bone marrow, but only at a single time point in clinically affected, experimentally infected cattle (Wells et al., 1999). Intracerebral inoculation of calves with tissues collected from experimentally infected cattle indicated further that palatine tonsil, not detected previously as being infective by wild-type mouse assay, was infective at 10 months post-oral infection (Wells et al., 2005). Inoculation of calves has also led to the detection of infectivity in the lymphoid tissues of nictitating membranes (G. A. H. Wells, unpublished data; ) from a pool of nictitating membranes collected from naturally infected cows.

Recently, studies using genetically modified mice (Tgbov XV), shown to be 10 000-fold more sensitive than assay in RIII mice and 10-fold more sensitive than assays by the intracerebral route in cattle, have enabled the detection of amounts of infectivity lower than the previous threshold of detection (Buschmann & Groschup, 2005). The studies also expanded the range of tissues assayed and demonstrated BSE infectivity in brain, spinal cord, retina, optic nerve, distal ileum, peripheral nerves (facial and sciatic nerves) and the semitendinosus muscle. With respect to the latter tissue, it remains unclear whether the small amounts of infectivity present were attributable to muscle tissue per se or to the peripheral nerve or lymphoid tissue content of the muscle.

In the assay study described here, we have confirmed that the detection of PrP^Sc is indicative of the presence of infectivity in vagus nerve and adrenal gland of BSE-affected cattle in the clinical stage of disease. This supports the use of PrP^Sc detection as a surrogate for detection of infectivity in expanding our understanding of the pathogenesis of BSE. However, tissues without detectable PrP^Sc accumulation may harbour prion infectivity as, despite the highly sensitive PrP^Sc-detection procedure used in this study, there are clear precedents for the occurrence of infectivity in the absence of detectable PrP^Sc. The incubation periods of vagus nerve and adrenal gland suggest that the estimated infectivity in these tissues was 22.5 logs lower than that of the CNS (Safar et al., 2002). PrP^Sc accumulation in the adrenal gland, vagus nerve and stellate ganglion may result from extension from a primary routeing of BSE prions via sympathetic and parasympathetic pathways to the CNS, or secondary spread from the CNS. The adrenal gland, for example, has a rich sympathetic supply associated with the capsule and the medulla that could be notionally infected primarily or secondarily via the splanchnic nerves. The vagus nerve provides the parasympathetic inervation to the intestine and enters the CNS at the medulla, and the stellate ganglion is part of the paravertebral sympathetic chain of ganglia. Interestingly, there are no precedents for a primary role for sensory neural pathways in the pathogenesis of BSE and, hence, DRG infectivity is considered to have spread from the CNS.

Quite clearly, our data indicate that a consumer-protection policy that is based solely on SRM removal, as currently designated, will not eliminate potential exposure to BSE infectivity completely in the carcass of an animal that is CNS-positive, i.e. clinically affected animals or those close to the onset of clinical disease. For such animals, a positive BSE test at the level of the obex followed by destruction of the carcasses would provide greater consumer protection than the removal of SRM alone. It remains to be determined whether the additional protection is actually significant, taking into account the quantity of infectivity present in amounts of PNS likely to be consumed. Nevertheless, the data provided here can contribute to review of risk assessments in relation to bovine PNS tissues.

We thank Dr Stanley B. Prusiner for providing the Tg(BoPrP)4092HOZ/Prnp^0/0 mice and Ms Adel Dale for BSE sample dispatch. Further, we thank Dr Yuichi Tagawa for providing the T2 mAb and Ms Nahoko Tabeta, Kimi Shimada, Yuka Ookubo and Shuko Kodani for their technical assistance. We also thank Dr Morikazu Shinagawa for his encouragement, Ms Junko Yamada for her general assistance and Ms Che Jing Zh and the animal laboratory staff at the National Institute of Animal Health for maintaining the mouse colony. This study was supported in part by a Grant-in-Aid from the Bovine Spongiform Encephalopathy Control Project of the Ministry of Agriculture, Forestry and Fisheries of Japan, a grant for BSE research from the Ministry of Health, Labour and Welfare of Japan, and a grant from the Special Coordination Funds for strategic cooperation to control emerging and re-emerging infection from the Ministry of Education, Culture, Sports, Science and Technology, Japan. The studies from which materials were supplied for this work were funded by the former UK Ministry of Agriculture Fisheries and Food and by the UK Department for Environment, Food and Rural Affairs (Defra). We are also grateful for additional tissues supplied from EU Project FAIR CT98 3651, accessed with the agreement of the European Commission, who funded collection. Selection of material from the UK experimental study was based on data from research funded by Health Canada.