Potential Role of Soil in the Transmission of Prion Disease

Transmissible spongiform encephalopathies (TSE), or prion diseases, are a family of inevitably fatal neurodegenerative disorders affecting a variety of mammalian species. These diseases include scrapie in sheep and goats; bovine spongiform encephalopathy (BSE, “mad cow” disease) in cattle; chronic wasting disease (CWD) in North American deer, elk and moose; transmissible mink encephalopathy; and Creutzfeldt-Jakob disease (CJD, sporadic, familial and variant forms) and kuru in humans. These diseases are characterized by long incubation periods, spongiform degeneration of the brain and accumulation of an abnormally folded isoform of the prion protein, designated PrP^Sc, in brain tissue (Prusiner 1998).

CWD and scrapie are the only TSEs that appear to be environmentally transmitted. Therefore, this chapter focuses primarily on these two TSEs. Scrapie has been known in sheep for at least 250 years (McGowan 1922). Most clinically infected sheep exhibit the obvious feature of excessive rubbing and scratching of the skin; the term “scrapie” derives from this symptom. The peak incidence of scrapie occurs in sheep three to four years of age although the earliest cases are seen at 18 months and the latest in animals older than 10 years (Dickinson 1976). The origin of the scrapie agent is unknown, but a familial pattern exists in natural sheep scrapie suggesting that genetics and, possibly, vertical transmission are important. Scrapie has a world-wide distribution and has been documented wherever sheep are raised, with the exception of Australia and New Zealand.

CWD was first identified at a Colorado research facility in 1967 (Williams and Young 1980) and has since been identified in captive cervid populations in Wyoming, Wisconsin, Saskatchewan, South Dakota, Oklahoma, New York, Nebraska, Montana, Minnesota, Kansas and Alberta (Fig. 1¹). In the free-ranging cervid population, CWD has been found in Wyoming, Wisconsin, Utah, South Dakota, Saskatchewan, New Mexico, Nebraska, Illinois, New York, West Virginia, Alberta and Colorado (Fig. 1¹). The increase in the known range of CWD in free-ranging and captive cervids may be due in part to increased surveillance. Epidemiological data from expanding affected regions suggest that lateral transmission is the primary mode of disease spread (Miller et al. 2000).

Oral transmission of prion diseases is well established raising concern over the potential interspecies transmission of animal TSEs to humans. Interspecies transmission of BSE to sheep, felines and ungulates has occurred (Prusiner 1997; Horiuchi et al. 2000) and is responsible for the emergence of variant CJD in humans. On the other hand, sheep scrapie does not appear to be orally transmissible to humans since the incidence of TSEs in populations consuming scrapie-infected sheep is not significantly higher than in other populations (Brown et al. 1987). Although CWD has been experimentally transmitted to squirrel monkeys (Marsh et al. 2005), no evidence exists for transmission of CWD to humans, and cell-free conversion assays suggest that the likelihood is low (Raymond et al. 2000). However, unlike BSE in cattle, infectivity is present in the muscle tissue of CWD-infected deer (Angers et al. 2006); thus, concerns remain about the possibility of CWD transmission into humans.

Prion diseases were originally designated as “unconventional” or “slow” viruses based on the inability to identify a conventional virus and the long incubation periods associated with these infections (Sigurdsson 1954). The extreme resistance of these agents to ionizing and UV irradiation (Alper et al. 1978; Bellinger-Kawahara et al. 1987) combined with the inability to isolate a virus or scrapie-specific nucleic acid suggested that TSEs lacked a nucleic acid genome. These data support the hypothesis that a protein with self-replicating properties could be the TSE agent (Griffith 1967). The self-replicating protein hypothesis was refined and renamed as the prion hypothesis, which stated “prions are small proteinaceous particles which are resistant to inactivation by most procedures that modify nucleic acids” (Prusiner 1982). The discovery and characterization of the disease-associated prion protein (PrP^Sc) suggested that PrP^Sc may be a major component of the infections agent, if not the agent itself (Bolton et al. 1982). Subsequent studies demonstrated that PrP^Sc, rather than being a novel protein, was a conformational variant of a normal brain protein, PrP^C. Figure 2² shows electron micrographs of PrP^C and PrP^Sc. Circular dichroism and infrared spectroscopy indicate that, relative to PrP^C, the disease-specific isoform has a higher β-sheet and lower α-helix content (Caughey et al. 1991). Although the three-dimensional structure of refolded, recombinant PrP^C has been elucidated by nuclear magnetic resonance (NMR) spectroscopy (Wüthrich and Riek 2001), difficulties in isolating pure PrP^Sc have thwarted attempts to determine its structure by NMR spectroscopy or X-ray crystallography. Recent electron crystallography data (Govaerts et al. 2004) suggest that PrP^Sc forms structured trimers with diameters of ~10 nm that aggregate into fibrils (typically 50 – 300 nm long) (Fig. 3³).

In vitro cell culture studies have demonstrated that PrP^C is the precursor to PrP^Sc (Caughey et al. 1989; Borchelt et al. 1990). Inhibition of the migration of PrP^C to the cell surface blocks the formation of PrP^Sc, confirming that PrP^Sc arises via alternate processing or misprocessing of PrP^C (Caughey et al. 1989). Two different models for PrP^C conversion to PrP^Sc are currently being debated (Fig. 4⁴). Both involve the direct interaction of PrP^C with PrP^Sc and consider PrP^Sc capable of conferring its abnormal conformation to the normal PrP^C molecule. The major difference between the two models is the nature of the PrP^Sc. In the template-assisted conversion model (Prusiner 1991), monomeric PrP^Sc is envisioned to interact with PrP^C, forming a heterodimer. PrP^C is then converted to PrP^Sc resulting in the formation of a homodimer, which then dissociates to form two PrP^Sc molecules (Fig. 4A⁴, “Refolding” model). This cycle is then repeated resulting in the increase in both infectious titer and PrP^Sc accumulation. In the nucleation-dependent polymerization model (Jarrett and Lansbury 1993), however, the infectious PrP^Sc is present as a nucleant or seed, likely comprised of oligomers of PrP^Sc (Fig. 4B⁴, “Seeding” model). Neither model, however, predicts how the conformation of PrP^Sc is conferred to the PrP^C molecule.

The function of the normal protein, PrP^C, has not been firmly established. PrP^C is normally expressed in mammalian neural tissue and may play a role in cellular resistance to oxidative stress and metal homeostasis in the brain (Brown 1999; Thackray et al. 2002). This host-encoded, 33–35 kDa sialoglycoprotein (Chesebro et al. 1985; Oesch et al. 1985) resides on the cell surface in lipid rafts (Taraboulos et al. 1995) and is tethered by a glycophosphotidylinositol anchor (Stahl et al. 1987). PrP^C is concentrated at neuronal synapses (Sales et al. 1998; Herms et al. 1999) and axonally transported to nerve terminals (Borchelt et al. 1994), suggesting that PrP^C is important for neuronal activity. PrP^C also appears to be a metalloprotein binding copper in vivo (Brown et al. 1997), and perhaps an endocytic receptor for the uptake of extracellular copper (Pauly and Harris 1998).

The disease-associated form of the prion protein exists in vivo as an aggregate of PrP^Sc resulting in a build up of amyloid deposits, often surrounding regions of dead tissue or spongiosis (Fig. 5⁵). The relationship of cell death to protein conversion is not clear, but loss of PrP^C or accumulation of PrP^Sc may be involved. Different strains of TSE agents exhibit different lesion patterns in the brain (i.e., distribution of vacuoles), physical properties, incubation times and host ranges (Bruce et al. 1976; Nonno et al. 2006). Protein aggregation may be responsible for toxic effects, but some evidence suggests that misfolded monomers or small oligomers may be responsible for toxicity (Stefani and Dobson 2003).

Prions exhibit extraordinary resistance to conditions and treatments that inactivate conventional pathogens including exposure to ultraviolet, microwave and ionizing radiation, treatment with proteases and contact with most chemical disinfectants (Millison et al. 1976; Ernst and Race 1993; Taylor et al. 1995; Taylor 2000). Boiling does not to eliminate prion infectivity; high temperatures are more effective when combined with steam and pressure (autoclaving). A small fraction of infectivity in wet brain tissue can, however, withstand autoclaving at 134 °C for ≤60 min (Taylor et al. 1994). Dry heat sterilizes only at the extremely high temperatures (in excess of 600 °C) that can be achieved during incineration (Brown et al. 2000). Incineration at lower temperatures does not reliably eliminate prion infectivity. TSE agents withstand chemical decomposition retaining the ability to initiate infection after exposure to temperatures that should decompose or volatilize organic molecules (Brown et al. 2000, 2004). The effectiveness of incineration has been demonstrated only at the laboratory scale.

An environmental reservoir of prion infectivity appears to contribute to the maintenance of epizootics of CWD (Miller et al. 1998, 2004) and possibly scrapie (Pálsson 1979; Andréoletti et al. 2000). Evidence for environmental transmission of scrapie has long been noted but is largely anecdotal. Healthy sheep in Iceland contracted scrapie after grazing on fields previously occupied by infected animals, but not on fields where the disease had not been present (Greig 1940; Sigurdarson 1954; Pálsson 1979). These studies are difficult to interpret because they predate our understanding of the infiuence of sheep genetics on susceptibility to scrapie. More convincing are the observations of healthy elk (Cervus elaphus) contracting CWD after introduction into pens that previously housed infected animals (Miller et al. 1998; Williams et al. 2002).

In a controlled field experiment (Fig. 6⁶), Miller et al. (2004) demonstrated that the presence of decomposed infected carcasses as well as residual excreta from infected animals on the landscape were sufficient to transmit CWD to healthy mule deer (Odocoileus hemionus). This study employed three treatments (viz., CWD-infected carcass decomposed in situ, residual excreta from CWD-infected animals and live CWD-infected mule deer) to determine whether these exposures could transmit CWD to healthy deer. Healthy deer contracted CWD in all three types of paddocks.

The epidemiological and experimental studies cited above indicate that prion infectivity can be maintained in the environment for several years. Several hypotheses have been advanced on the nature of the putative environmental reservoir including soil, arthropod vectors and nematodes (Fitzsimmons and Pattison 1968; Brown and Gajdusek 1991; Post et al. 1999; Carp et al. 2000; Miller et al. 2004; Johnson et al. 2006d). As noted in a number of studies (Miller et al. 1998, 2004; Williams et al. 2002), arthropod vectors (e.g., hay mites, flesh flies) seem unlikely to account for the observed persistence of prion infectivity in the environment. Soils appear to represent a plausible environmental reservoir for prion infectivity. The remainder of this chapter focuses on the potential role of soil in the transmission of prion diseases.

For soil to serve as environmental reservoir, the following must hold: (1) prions must enter the environment; (2) prions must persist in soil; (3) prions in soil must retain infectivity; (4) prions must remain near the soil surface where they can be accessed by animals; (5) naïve animals must be exposed; and (6) the dose must be sufficient to cause infection. We discuss each of these factors below.

Several plausible routes of prion introduction into soil environments can be envisioned (Fig. 7⁷). Prions clearly enter soil environments when carcasses of infected animals decompose. Free-ranging CWD-infected animals die and decompose in the field. The remains of infected deer and elk dressed in the field by hunters (i.e., “gut piles”) also represent a route of introduction of CWD agent into soils. Placentas of infected sheep carry high levels of infectivity (Race et al. 1998). Deposition of placenta to the soil surface may represent an important route of prion introduction.

The presence of PrP^Sc in gut-associated lymphatic tissue (e.g., tonsils, Peyer’s patches, mesenteric lymph nodes) early in the disease course suggests that the agent could be shed through the alimentary system (Sigurdson et al. 1999; Miller and Williams 2002). Due to the lengthy incubation periods associated with TSE infection (long pre-clinical phase), infected animals may shed prions for extended periods. Prions have recently been detected in the saliva of CWD-infected mule deer by oral exposure of naïve white-tailed deer (O. virginianus), but in the feces or urine of the same animals (Mathiason et al. 2006). This latter finding may be due to the small sample size, the genotype of the recipient animals (Johnson et al. 2006a), and/or lower prion concentrations in urine and feces.

Urinary shedding of prions has been recently demonstrated in presymptomatic scrapie-infected animals having concurrent chronic kidney infection (Seeger et al. 2005). Earlier reports of urinary shedding in TSE-infected hamsters, cattle and humans (Shaked et al. 2001) have not been reproduced by others. The observations of Shaked et al. (2001) may have been due to a proteinase K-resistant enterobacterial outer membrane protein, of similar molecular mass as PrP, that exhibited nonspecific binding to a variety of anti-PrP antibodies (Furukawa et al. 2004).

As discussed above, prions display remarkable resistance to a variety of conditions that inactivate conventional pathogens including heat, chemical disinfectants, proteases and UV irradiation (Taylor 2000). Persistence of prions in soil environments requires that they withstand a variety of assaults including UV irradiation, freeze-thaw cycles, extracellular enzymes from bacteria and fungi, digestion by soil macrofauna, and abiotic transformation by reactive mineral phases. Furthermore, prions have the potential to interact with organic and inorganic soil components. Such interactions may influence the resistance of prions to various assaults and may alter their bioactivity.

The first study to examine the persistence of prions in soil involved interring scrapie-infected hamster brain material in garden soil for 3 years (Brown and Gajdusek 1991). Compared to a frozen control, infectivity in water washes of the soil was 1.8 to 2.6 log units less, but still represented 10^5.6 to 10^6.4 IU50. The authors did not determine whether the apparent reduction in infectivity was attributable to attachment of PrP^Sc to soil particles or to degradation. This study was limited to a single, uncharacterized soil and did not include replicates for treatments or controls. Since soil represents a complex matrix of inorganic minerals and organic matter and soil properties vary considerably from location to location, these findings cannot be easily extrapolated to other settings.

The pathological form of the prion protein is extremely hydrophobic (Prusiner 1998) and is therefore expected to associate with soil components to a significant extent. Johnson et al. (2006b) examined the association of PrP^Sc with three common soil minerals (viz., montmorillonite, kaolinite and quartz) and four soils. In this study, PrP^Sc enriched from clinically infected hamster brains (Bolton et al. 1987; McKenzie et al. 1998) was exposed to pure minerals in aqueous suspension. Mineral-bound PrP^Sc was separated from unbound prion protein by density centrifugation through a sucrose cushion. Prion protein associated with mineral particles was recovered by 10-min extraction in a solution containing 10% sodium dodecyl sulfate (SDS; a denaturing anionic detergent) at 100 °C. The amounts of bound and unbound PrP^Sc were examined by SDS-polyacrylamide electrophoresis (SDS-PAGE) followed by immunoblotting (Western blotting). Of the soil minerals examined, montmorillonite (effective hydrodynamic diameter, d_h = 0.5–2 μm) exhibited the largest PrP^Sc-binding capacity followed by quartz microparticles (d_h = 1–5 μm) and kaolinite (d_h = 0.5–2) (Table 1¹). These results suggest that mineral surface properties contributed to the differences in PrP^Sc binding.

The binding of PrP^Sc to montmorillonite was avid; only extremely harsh conditions (viz., 10-min exposure to 10% SDS at 100 °C) were effective in desorbing the protein from the clay surface (Johnson et al. 2006b). Exposure to 10% SDS at lower temperatures was less effective in removing PrP^Sc from the montmorillonite. Conditions employed by previous investigators (e.g., Morgan and Corke 1976; Docoslis et al. 2001) to desorb proteins from clay surfaces were ineffective in removing detectable PrP^Sc from montmorillonite (Johnson et al. 2006b). These conditions included increases in ionic strength, strong chaotropic agents (viz., 8 M guanidine HCl and 8 M urea) and pH extremes (pH 2.5 and 11.5). The attachment of proteins to clay surfaces is often maximal at the isoelectric point (pI) of the protein and decreases at pH > pI because of repulsive electrostatic interactions (Quiquampoix et al. 2002). Increasing suspension pH to above the protein pI often results in the detachment of proteins from clay surfaces (Armstrong and Chesters 1965; Quiquampoix et al. 1993). PrP^Sc aggregates exhibit an apparent pI of 4.6 (unpublished data), but were not desorbed in detectable amounts from the montmorillonite surface when pH was elevated to 11.5 (Johnson et al. 2006b).

Interestingly, PrP^Sc desorbed from montmorillonite, but not kaolinite or quartz, exhibited a reduction in molecular mass (Johnson et al. 2006b). Using antibodies directed against the N-terminus of the protein, Johnson et al. (2006b) determined that PrP^Sc desorbed from montmorillonite was cleaved at a site on the N-terminal flexible domain not required for infectivity (Supattapone et al. 1999). Cleavage of other proteins desorbed from Mte has not been reported.

The sorption of PrP^Sc to four soils differing in texture, organic carbon content and mineralogy was also examined (Johnson et al. 2006b). Although the sorption capacities of the soils were not quantified, the amount of sorbed PrP^Sc recovered varied among the soils suggesting differences in the strength of interaction. Clay content appeared to be an important determinant for the amount recovered. Leita et al. (2006) reported that PrP^Sc strongly interacted with two sandy loams and one clay loam. These investigators were unable to recover sorbed PrP^Sc from the soils, although the extractant used differed from that employed by Johnson et al. (2006b). Sorption of PrP^Sc to humic substances has not been investigated.

Several reports have appeared investigating the association of the non-pathogenic, nonglycosylated, recombinant PrP to clay minerals and soils (Revault et al. 2005; Vasina et al. 2005; Rigou et al. 2006) but relevance of these studies to the transmission of TSEs is limited because they did not employ the infectious agent in their experiments. As mentioned above, the tertiary structure (folding) of PrP^C differs dramatically from that of PrP^Sc resulting in the two isoforms exhibiting very different biophysical properties (Prusiner 1998; Riesner 2004).

PrP^Sc is more resistant to proteases than most other proteins. Association of prions with mineral surfaces and humic substances may further protect the agent from degradation by microorganisms and extracellular proteases. This phenomenon has been observed for a variety of enzymes and Bt toxin, the insecticidal protein of Bacillus thuringiensis (Naidja et al. 2000; Stotzky 2000). The extent of protection depends on both the nature of the mineral surface and the protein in question.

For soil to contribute to lateral transmission of TSEs, prions attached to soil particles must retain infectivity. Adsorption of proteins to mineral surfaces can be accompanied by conformational changes that cause loss or diminution of function (e.g., Morgan and Corke 1976; Vettori et al. 1999; Naidja et al. 2000; Lecomte et al. 2001). Johnson et al. (2006b) examined whether attachment to clay surfaces would decrease infectivity and estimated that adsorption of prions enhanced infectivity by a factor of ~10 relative to unbound agent when inoculated intracerebrally (Table 2²). This result was surprising, given the avidity of PrP^Sc attachment to the clay mineral, and suggested that association with montmorillonite reduced clearance from the brain (i.e., translocation out of the brain and degradation by proteases).

Bioavailability via the oral route of exposure could be diminished if detachment of PrP^Sc from mineral surfaces is necessary for uptake or enhanced if PrP^Sc-mineral complexes are directly taken up. The digestive tract secretes numerous surfactants (e.g., bile acids, lipid and protein surfactants) that may allow extraction of PrP^Sc from soil or soil minerals. Association with mineral particles could, however, protect prions from proteolysis in the gastrointestinal tract, thereby enhancing infectivity. Detachment from mineral particles may not be required for uptake. Intestinal Peyer’s patches sample microparticles similar in size to clay particles; in some cases such particles are translocated to lymph tissue (Desai et al. 1996; Hazzard et al. 1996). Both Peyer’s patches and lymphatic tissues are peripheral sites of PrP^C-to-PrP^Sc conversion, involved in the early stages of CWD and scrapie infection (Sigurdson et al. 1999; Andréoletti et al. 2000; Miller and Williams 2002), and routing particle-associated prions to either of these sites could dramatically increase the chance of initiating TSE infection. Oral infectivity assays with soil- and soil-mineral-bound prions will be required to assess bioavailability via this exposure route.

For prions in soils to contribute to the lateral transmission of TSEs, they must remain near the soil surface where animals can come in contact with them. Little published literature exists on the mobility of prions in soils. The strong attachment of PrP^Sc to clay minerals (Johnson et al. 2006b) suggests that prions may be retained near the surface of fine textured soils. Attachment to clay particles may, however, facilitate transport through the soil column in some situations, as has been demonstrated for virus particles (Jin et al. 2000). Brown and Gajdusek (1991) observed a limited amount of prion leaching from a Petri dish with holes drilled through the bottom that contained infected hamster brain material and garden soil; no infectivity was observed more than 4 cm below the dish.

Because little information is available on the environmental behavior of prions, a brief consideration of research on virus transport may serve to frame our expectations of prion mobility in soils. Viruses are biological nanoparticles; like prions, non-enveloped viruses expose a protein surface to solution. While viruses are generally larger than monomers and small oligomers of PrP^Sc, the size range of larger PrP^Sc aggregates overlaps that of small animal viruses (Silveira et al. 2005). The aggregation state of PrP^Sc excreted from infected animals or released from decomposing carcasses is unknown but is likely smaller than those in PrP^Sc-enriched fractions.

Factors influencing the transport of viruses include soil mineralogy and organic matter content (Lukasik et al. 1999; Ryan et al. 2002; Meschke and Sobsey 2003; Zhuang and Jin 2003). Iron and aluminum oxides appear to decrease virus movement (Lukasik et al. 1999; Zhuang and Jin 2003). Organic matter associated with soil particles increases retention of viruses in porous media, presumably by providing hydrophobic sites for attachment. In contrast, dissolved organic matter decreases virus attachment due to competition for surface sites (Schijven and Hassanizadeh 2000). Both electrostatic and hydrophobic interactions appear to be important in virus transport through porous media (Bales et al. 1991; Penrod et al. 1996; Redman et al. 1997; Dowd et al. 1998; Chattopadhyay and Puls 2000; Zhuang and Jin 2003). Not surprisingly, hydrophobic interactions were especially important for more hydrophobic viruses (Bales et al. 1991; Kinoshita et al. 1993; Chattopadhyay and Puls 1999, 2000). Both reversible and irreversible virus attachment have been reported in batch sorption and column leaching experiments (Bales et al. 1993; Loveland et al. 1996), depending upon the characteristics of the mineral surface. Loveland et al. (1996) demonstrated that viruses can be desorbed from mineral surfaces by raising the pH, decreasing ionic strength or adding a high ionic strength protein solution such as beef extract that can compete with virus for sorption sites (Bales et al. 1993). Surfactants reduce virus attachment to soil particles by competing for attachment sites, displacing attached viruses, or increasing virus solubility (Chattopadhyay et al. 2002). Although attachment to clay particles may limit virus (and PrP^Sc) mobility, association with colloidal clay minerals can also facilitate transport (Jin et al. 2000).

The analogy between prions and viruses is only partial. Prions exhibit surface charge heterogeneity, and aggregates are smaller and more hydrophobic than most viruses. Prion proteins exhibit a range of isoelectric points with at least eight charge isomers (Bolton et al. 1985). This charge heterogeneity is imparted by the number (0–2) and nature of N-linked glycans. Gel electrophoresis with immunoblotting yields three bands corresponding to the di-, mono- and unglycosylated protein. Mass spectrometry studies indicate that >160 PrP glycoforms exist in a single prion preparation (Baldwin 2001).

The early involvement of gut-associated lymphatic tissue in TSE infection argues for an oral route of exposure. Herbivores ingest soil both deliberately and incidentally during grazing and grooming (Beyer et al. 1994). Sites where infected carcasses, gut piles or placentas have decomposed may represent foci of TSE transmission via ingestion of prions associated with soils. The frequency that cervids [mule deer, white-tailed deer, elk, and moose (Alces alces)] and sheep visit such sites and the amount of soil sampled during visits has not been determined. The following discussion on deliberate soil ingestion focuses on mineral licks and scrapes because more is known about deer behavior at these sites than at those where carcasses, gut piles or placentas have decomposed. Given the much higher levels of infectivity in carcasses and gut piles, sites where these materials are deposited and decompose may harbor more infectivity and contribute more to TSE transmission than licks and scrapes.

In cervids, deliberate ingestion of soil occurs at mineral licks, artificial salt licks and scrapes (Atwood and Weeks 2003). Throughout the year, all members of the Cervidae family supplement their mineral intake by ingesting soil at locations that have higher levels of sodium and other cations (Atwood and Weeks 2002). Licks may be artificial (e.g., from mineral blocks used for livestock grazing) or natural, and the salts persist in clay soils for over 20 years, after diffusing into the soil (Thackston and Holbrook 1992). All natural licks used by white-tailed deer in one study (Weeks 1978) occurred in Stendal silt loams with subsoils having high clay contents. Licks formed in depressions, where runoff collected and evaporated, leaving dissolved minerals above the mostly impermeable subsoils (Weeks and Kirkpatrick 1976). Since minerals from licks are persistent and draw deer to their location on a regular basis for many years, they may serve as points of prion accumulation from deposited saliva, urine and feces.

Deliberate soil ingestion by deer also occurs at scrapes. Scrapes are created by heavy pawing of the soil to remove surface detritus, and then marked with deposits of glandular secretions and urine or feces (Hirth 1977; Kile and Marchinton 1977; Miller 1987). Primarily during the rut, male deer create these chemical signposts to communicate with other deer (Moore and Marchinton 1974; Hirth 1977). Both males and females are known to “scent mark” at established scrapes by urination (Moore and Marchinton 1971). Several studies suggest that females visit scrapes more frequently than males (Alexy et al. 2001). Multiple males visit each scrape with little indication of revisitation (Alexy et al. 2001). Soil sampling by deliberate licking the urine-saturated soil occurs by males as a means to determine the estrous condition of females and presence of other males (Moore and Marchington 1971). It is not known whether females sample soil at scrapes. Both males and females also mouth small branches above the scrapes (Kile and Marchinton 1977).

The most likely route of exposure to soil-associated CWD agent by deer appears to be via either of the above recurrent, deliberate behaviors (i.e., sampling soil at mineral licks and scrapes). Incidental ingestion of soil can also contribute to soil exposure. The annual average soil intake of sheep pastured all year was 4.5% of the dry matter intake (Fries 1996). Soil is estimated to comprise a minimum of 2% of a deer’s diet (dry matter basis), annually, and may exceed 50% during the late spring and early summer at licks (Weeks and Kirkpatrick 1976).

For ingestion of soil to lead to TSE infection, sufficient quantities of prions (i.e., ≥ 1 oral IU50) must be ingested. With intracerebral inoculation, 1 IU50 has been estimated to comprise approximately 10⁵ PrP^Sc molecules in rodent models (Bolton et al. 1991). Oral exposure is approximately ~10⁵-fold less efficient in mice and hamsters (Diringer et al. 1994, 1998). The number of molecules per IU50 has not been determined for scrapie in sheep or for CWD, nor has the efficiency of oral versus intracerebral dosing been determined for these TSEs. The infectious dose need not be acquired from a single exposure; repeated dosing can enhance transmission (Diringer et al. 1998). Frequent visitation of mineral licks and, perhaps, locations where carcasses or gut piles have decomposed, by cervids may enhance transmission. Since Johnson et al. (2006b) demonstrated that adsorption of PrP^Sc to the clay mineral montmorillonite enhanced infectivity by a factor of ~10 via the intracerebral route, sorption to particle surfaces may also enhance prion infectivity via the oral route by protecting the agent from degradation in the gastrointestinal tract (Martinsen et al. 2002). Such an effect has been demonstrated for bovine rotavirus and coronavirus (Clark et al 1998).

Levels of infectivity present in naturally prion–contaminated soils have not been measured. Until recently, methods to recover PrP^Sc from soils were unavailable. Johnson et al. (2006b) were able to recover PrP^Sc from soils and pure minerals experimentally spiked with the protein using SDS-PAGE sample buffer [10% SDS, 100 mM Tris pH 8.0, 7.5 mM EDTA, 100 mM dithiothreitol (DTT), 30% glycerol] at 100 °C. Rigou et al. (2006) described an electroelution method to extract recombinant PrP.

An environmental reservoir of prion infectivity appears to contribute to the transmission of cervid chronic wasting disease and, probably, sheep scrapie. Soil represents a likely candidate for an environmental reservoir because (1) oral exposure appears important in the natural transmission of scrapie and CWD; (2) accumulation of the infectious agent in gut-associated lymphatic tissue at early stages of infection argues for alimentary shedding [recently demonstrated for saliva (Mathiason et al. 2006)]; (3) sheep and deer ingest soil both deliberately and incidentally; and (4) prions persist in soil environments for ≥ 3 years. Research conducted in our laboratories supports a role for soil in the transmission of prion diseases of sheep, deer and elk. Prions adsorb strongly to clay minerals and may serve to maintain infectivity close to the soil surface where they can be ingested by herbivores. Known behaviors of all members of the Cervidae family (deer, elk, moose) lead to deliberate ingestion of soils that have a high probability of contact with potentially infectious bodily secretions. Attachment to clay minerals enhances prion infectivity via the intracerebral route and may also do so via the oral route of exposure.

The authors gratefully acknowledge financial support from the Department of Defense National Prion Research Program through grant DAMD17-03-1-0369 (JMA, DM and JAP) and the U.S. Environmental Protection Agency through grant 4C-R070-NAEX (JAP). PTS was supported by NIEHS training grant T32 ES07015–28.