Main

According to the widely accepted protein-only hypothesis2, prions consist of multi-chain forms of misfolded host-encoded cellular prion protein (PrPC), referred to as PrPSc (ref. 3) and are thought to propagate by an autocatalytic process of seeded fibrilisation4,5,6. Prion infections are invariably associated with prolonged, clinically silent incubation periods which in humans may exceed 50 years7 and yet are followed by an aggressive, often stereotypic, clinical phase which is typically measured in months1. In inbred laboratory mouse lines, incubation periods for a defined mouse prion strain, although also prolonged, are remarkably reproducible, with onsets in a group of intracerebrally inoculated mice typically spanning only a few days following an incubation period of six months or more. Incubation periods inversely correlate with PrPC expression levels: very short incubation periods can be obtained in transgenic mice with high levels of PrPC overexpression8, whereas mice with 50% of wild-type expression have greatly prolonged incubation periods9. The basis of neurotoxicity in prion neurodegeneration, its relationship to these silent incubation periods and the near synchronicity of clinical onset, remain obscure.

Recently, a general model has sought to incorporate the key phenomena of prion strain diversity and mutation, transmission barriers and the uncoupling of neurotoxicity and infectivity, while remaining within the constraint of requiring only a single polypeptide to constitute all strains of infectious and toxic species6. According to this model, these phenomena can be explained in terms of the kinetics of prion propagation, determined by interplay between prion strain type and the tissue/host environment. Neurotoxicity is mediated by a lethal PrP species, PrPL (refs 10, 11), which is distinct from PrPSc, but its formation is catalysed by it. The toxic effects occur when the PrPL concentration passes a local threshold6.

Many previous studies, largely published many years ago, have measured prion titres during the incubation period in rodents. These studies used conventional end-point dilution or incubation interval time bioassay of prion titre, with small numbers of mice or hamsters per assay, and had a significant margin of error. Such classical methods have reported contradictory findings. Some studies have reported prion titres as low or undetectable in the early incubation period, with an exponential rise before clinical onset and the highest titres in the terminal clinical stage12,13,14,15. Indeed, if prion propagation involved a high-order exponential, this might explain the rapidity and synchronicity of clinical onset as prion titre rose explosively in the late incubation period6. However, other studies, notably in hamsters, have suggested that prion titres may plateau in brain, albeit late in the incubation period16,17. The occurrence of such a plateau in brain has been unexplained.

The recently developed scrapie cell assay (SCA) now allows much more precise bioassay of Rocky Mountain Laboratory (RML) prions than conventional mouse assay18. In the course of experimental studies to test the general protein-only model, we have used the SCA to produce an accurate and detailed measurement of prion titres throughout the incubation period in mice with a range of PrPC expression levels and incubation periods and attempted to correlate these with clinical onset.

We inoculated large groups of the following mice intracerebrally with RML mouse prions (30 μl 1% brain homogenate containing 105.8 intracerebral LD50 (dose lethal to 50% of inoculated animals) units and groups of 5–6 mice were killed at multiple defined time points or at onset of clinical disease (Fig. 1a): inbred FVB/N mice (Prnp+/+; wild-type PrPC expression level); Prnp null mice19 (Prnpo/o; no PrPC expression; FVB/N background); hemizygous Prnp null mice (Prnp+/o; 50% wild-type PrPC expression; FVB/N background) and Tg20 transgenic mice20 (eightfold wild-type PrPC expression level; FVB/N background). As neuropathological controls, groups of mice of each genotype were inoculated with uninfected 1% mouse brain homogenate. Prion titre in mouse brain was determined by SCA or scrapie cell assay in end-point format (SCEPA) for low-titre samples18.

Figure 1: Survival times and levels of infectivity during prion infection.
figure 1

a, Survival curves in mice. Three different lines of mice were infected with RML prions and the survival time measured in days: Tg20 with eightfold overexpression of PrPC compared to wild type, Prnp+/+ with wild-type PrPC expression level, Prnp+/o with 50% wild-type PrPC expression. Prnpo/o mice with no PrPC expression were also inoculated but do not develop disease; data not plotted for clarity, titres were ∼2.5 logs at 1 and 3 days and fell to undetectable levels by 10 days. Mean incubation periods, indicated by arrows, were (in days ± s.d.): Tg20, 59.5 ± 2.0; Prnp+/+, 137 ± 1.5; Prnp+/o, 258 ± 23.5. b, Mice were culled at defined time points and prion titre determined. Log tissue culture infectious units are per gram brain; bars indicate s.e.m. and in some cases are smaller than symbol used to designate mean; group sizes were 3–6. Arrows indicate onset of clinical disease. c, The reciprocal of the length of the plateau time in days was plotted against the expression level of PrPC in the mouse lines. Length of phase 2 is inversely proportional to PrPC expression level. This plot effectively represents the rate at which lethal species of PrP are formed as a function of the concentration of PrPC. The start of the plateau was defined by the time taken for a linear fit to the logarithmic rise in infectious units at the beginning of each plot to reach the 108 log infectious units plateau level. The end of the plateau was defined by the onset of clinical disease. Plateau lengths were estimated as: Tg20, 8 days; Prnp+/+, 53 days; Prnp+/o, 114 days.

PowerPoint slide

Full size image

As expected, prions were only briefly detectable in Prnpo/o mice (that are unable to propagate prions21) and were undetectable after 10 days by SCEPA, demonstrating rapid clearance of the original inoculum and allowing clear interpretation of prion titres in the early incubation period in the other mouse groups (Fig. 1b). Remarkably, prion titres in wild-type FVB/N mice rose rapidly very early in the incubation period, reaching a plateau long before clinical onset. A similar rapid early rise to the same prion titre (of approximately 108 infectious units) was seen in both Tg20 and Prnp+/o mice, the principal difference being in the duration of the plateau phase. In the case of the short incubation period Tg20 mice this was highly truncated, whereas in the long incubation period Prnp+/o mice this was protracted and accounted for the large majority of the incubation time.

Therefore, we can now describe two distinct phases of prion propagation during the incubation period of RML prion infection in mice. In phase 1, prions propagate exponentially until a clearly defined limit is reached. After this point, a plateau phase (phase 2) continues until the onset of clinical disease. The length of phase 2 is inversely proportional to the PrPC expression level (Fig. 1c). Interestingly, the rate of prion propagation in phase 1 is not proportional to PrPC expression level and the prion titre attained in the brain during phases 1 and 2 is independent of PrPC expression level over the range we have studied (0.5–8-fold wild type).

These observations lead to important deductions about prion propagation and neurotoxicity. First, propagation of prion infectivity proceeds in a single phase (phase 1) of exponential autocatalytic conversion. During this process, the addition of PrPC is only rate-limiting at low expression levels. At high expression levels, that is, for wild-type and the Tg20 overexpresser, the rate of production of infectivity is closely similar. Therefore, the rate-limiting step at high concentrations of PrPC must be either a structural rearrangement following PrPC–PrPSc interaction6 or a step in the division of the extended PrPSc. In phase 2, where there is no increase in infectivity, the rate of formation of the toxic species is directly proportional to PrPC concentration. In this phase the nature of the conversion process has changed as it must be rate-limited by the addition of PrPC. That there are closely similar levels of infectivity at the end of phases 1 and 2, irrespective of PrPC expression level, indicates that there is either a key cofactor, at effectively fixed concentration, or a saturable number of replication sites22, that limits prion production.

Whereas RML prion titres assessed by SCA correlate closely with those obtained by conventional mouse bioassay, these comparisons were made using infectious isolates obtained from terminally clinically affected mice. We considered the possibility that the SCA may be more sensitive in detection of prions generated early in the incubation period, for example if they were more labile and degraded in mouse brain. We therefore performed conventional mouse end-point titration bioassay on samples obtained at 80 days post-inoculation in FVB/N mice (at the start of the plateau) and compared these with those from clinically affected mice (at the end of the plateau). The titre at 80 days was 108.6 LD50 units per ml 10% homogenate, indistinguishable from that at terminal disease.

These studies clearly demonstrate uncoupling of prion infective titre and neurotoxicity, that is, prion titres reach maximal levels in conventional mice with wild-type PrPC expression long before clinical signs occur. We have proposed previously that neurotoxic species, presumed to be oligomeric forms of PrP (PrPL), are not on-pathway for prion propagation but produced by a separate but linked pathway in which PrPSc particles act as a catalytic surface for production of PrPL (ref. 6). Our current data can be readily accommodated by this general model where production of PrPL is directly proportional to PrPC concentration and it is this that principally determines time to onset of clinical disease (Fig. 2). However, alternative models can be envisaged. For example, fission of infectious particles may cease, leading to their further growth but without an increase in their number, or alternatively a vital cellular component may be depleted.

Figure 2: Kinetics of prion propagation and toxicity.
figure 2

Prion replication (solid lines) is exponential (phase 1) until a limiting titre of prions is attained, which is independent of PrPC expression level. At this point there is a pathway switch leading to the production of toxic species (dotted lines) at a rate linearly dependent on PrPC concentration (phase 2). Toxic species do not accumulate until infectivity saturates and clinical signs occur after toxic threshold is crossed.

PowerPoint slide

Full size image

Whereas disease-related PrP, or PrPSc, was originally defined in terms of its protease-resistance and detergent insolubility, it is increasingly recognized that there are multiple disease-related forms of PrP, including protease-sensitive species23, which may indeed comprise the majority of infectivity in some isolates24. As yet, these are poorly defined in physical terms and an agreed nomenclature is lacking. It will be important to determine the relative proportion of these species at different stages of the incubation period. It is possible that a range of PrP species constitute the infectious and toxic entities and these populations may overlap; it is increasingly evident that prion strains constitute a cloud or quasispecies6,25.

It has also been proposed that PrPC acts as a receptor for PrPSc and mediates toxic signalling26. In the light of this, an alternative interpretation could be that variation in PrPC expression results in a different rate of toxic signalling when prion titres plateau and thereby determines the plateau time before death. However, subclinical prion carrier states exist, where conventional mice with normal neuronal PrPC expression do not develop clinical signs. These mice live a normal lifespan, despite having prion titres as high as those seen in end-stage clinically sick mice10,11,27,28,29; such observations are hard to accommodate within a receptor-based model.

Methods Summary

All procedures were carried out in microbiological containment level 3 facilities. Care of mice was according to institutional guidelines.

RML prion inoculum (I6200) was prepared as a 10% (w/v) brain homogenate in Dulbecco’s phosphate buffered saline lacking Ca2+ or Mg2+ ions (D-PBS) and titred by bioassay in CD1 mice (108.3 intracerebral LD50 per ml)24. I6200 was diluted to 1% (w/v) with D-PBS and 30 μl inoculated intracerebrally into large groups of mice30. Normal CD1 mouse brain homogenate (1% w/v) was inoculated similarly into control mice. Mice used were inbred FVB/N (Prnp+/+; wild-type PrPC expression level), Prnp null19 (backcrossed to FVB/N mice; Prnpo/o; no PrPC expression), hemizygous Prnp null FVB/N (Prnp+/o; 50% wild-type PrPC expression) and Tg20 transgenic mice20 (backcrossed to FVB/N mice; eightfold over expression of wild-type PrPC). Groups of 5–6 mice were killed at multiple defined time points or at onset of clinical disease. Brains were removed and divided sagittally, half-frozen at −70 °C and half-fixed in 10% v/v formal buffered saline. Brain homogenates (10% w/v) were prepared from frozen specimens in D-PBS and stored as aliquots at −70 °C.

Cell culture assays for RML prion infectivity were as described18. Briefly, PK1 cells were exposed for 3 days in 96-well plates to serial dilutions of 10% (w/v) brain homogenate. A serial dilution of 10% (w/v) RML brain homogenate of known infectivity titre (I6200) was performed in parallel. Subsequently cells were split and passaged appropriately for the scrapie cell assay (SCA) or the scrapie cell assay in end-point format (SCEPA)18. Infectivity titre of each sample was deduced from the reference preparation. Normal interval regression was used to calculate mean and s.e.m. for two groups that contained a single point below assay sensitivity cut-off. Mouse bioassay was performed by end-point titration using groups of ten FVB/N mice by intracerebral inoculation with 30 μl of serial tenfold dilutions of mouse brain to 10−8 (ref. 24); prion titres were calculated using the Reed–Muench formula.