News & Views | Published:

Prion diseases

Cannibals and garbage piles

There's a lot of disagreement among prion scientists, as a recent conference made very clear. Even the revered 'prion hypothesis' came under attack.

After more than 280,000 mad cows and two Nobel prizes for research on transmissible spongiform encephalopathies, we must know all there is to know about these 'prion' diseases. Or do we? This question was asked at a recent symposiumFootnote 1, which looked at topics such as the cell biology of the normal prion protein and of its misshapen, disease-associated form; how the brain becomes damaged; and diagnosis and treatment. It was clear that, in all of these areas and more, there's still a long way to go.

Prion scientists have a reputation for being a contentious bunch — a fact that was amply confirmed on this occasion. Diametrically divergent opinions emerged on central questions such as the physiological function of the normal prion protein (PrPC) and the role of its aberrant form (PrPSc) in disease.

Even the prion hypothesis, which nowadays is often regarded as dogma, was challenged. This hypothesis states that PrPSc is the infectious agent in transmissible spongiform encephalopathies (TSEs), and that it replicates by imparting its misshapen conformation onto PrPC. In a spirited lecture, however, new Nobel laureate Kurt Wüthrich (ETH, Zurich) pointed out the continued failures to create infectivity in vitro by modifying bacterially expressed prion protein — a crucial prediction of the prion hypothesis. Another important experiment involves abolishing the structure (and infectivity) of the disease-associated prion protein with specific salts, and then attempting to restore infectivity by reinstating the original structure. This, too, has so far failed. Wüthrich referred to PrPSc as simply a build-up of "garbage", and submitted that we must understand the function of the normal prion protein before we can understand prion diseases.

The TSEs are characterized by the death of nerve cells, and another point of controversy concerned the mechanisms by which this occurs. A strong case was presented that the accumulation of PrPSc within the cytosol of neurons is to blame (S. Lindquist, Whitehead Inst., Cambridge, Massachusetts)1,2. PrPC is usually located in the plasma membrane, and travels there by way of a network of internal membranes, the endoplasmic reticulum (ER). Lindquist proposed that a certain proportion of PrPC never reaches the plasma membrane, but instead re-enters the cytosol by 'retrotranslocation' from the ER — a standard means by which other misfolded proteins are directed to the cell's waste-disposal unit, the proteasome. Lindquist found that a form of the prion protein that was specifically targeted to the cytosol caused rapidly lethal neurodegeneration in mice. This protein did not acquire resistance to protein-digesting enzymes (proteases), which has long been thought to be a key characteristic of PrPSc. But proteasome inhibition led to the accumulation of a slightly protease-resistant prion protein in cultured cells. Lindquist speculated that the diverse mutations in the prion protein that are associated with familial Creutzfeldt–Jakob disease (CJD) — a human TSE — might all lead to enhanced retrotranslocation, which, upon impaired proteasome function, could trigger disease. The big surprise here is that the cytosol might be the place where PrP-mediated neuronal death begins.

However, this model was contested by D. Harris (Washington Univ., St Louis, Missouri), who found that cytosolic prion protein retains its 'signal peptide' — normally removed after proteins enter the ER — and does not contain the glycosyl phosphatidylinositol 'anchor' needed for attachment to membranes3. This suggests that the protein never entered the ER, and so could not have undergone retrotranslocation. Harris also pointed out that proteasome inhibitors have powerful effects on the levels of prion messenger RNA; these effects might have contributed to previous results.

Even within a single population of inbred animals, prions come in distinct varieties, which — upon transmission to further animals — produce characteristic, heritable incubation times and patterns of brain damage. This phenomenon of prion 'strains' continues to produce surprises. When hamster prions were inoculated into mice, the animals lived a long, TSE-free life, and mostly did not accumulate PrPSc in their brains (B. Chesebro, Rocky Mountain Labs, Hamilton, Montana). The injection of PrPSc-negative brain extracts from these mice into further mice again resulted in no clinical disease, over a study period of more than 650 days. But when brain extracts from the latter mice were injected into hamsters, the animals died rapidly. So the infectious agent had silently replicated for several years in mice but maintained full virulence towards hamsters. Given that mouse prions are generally harmless to hamsters, it is hard to understand why — in this instance — they retained their infectivity. It would seem that prion strain characteristics dominate over the amino-acid sequence of the prions from the infected host. It is challenging, but maybe not impossible, to reconcile these data with the protein-only prion hypothesis: maybe the strain-specific properties are really encoded in the tertiary or quaternary structure of PrPSc rather than in its amino-acid (primary) sequence.

At the genetic level, variations in the human prion gene that protect against the development of CJD have disseminated much more efficiently than non-protective variations throughout human populations worldwide (J. Collinge, Inst. Neurology, London)4. This provides a compelling case that these protective changes were 'selected for' during the course of evolution — but why would they have been necessary? Collinge suggested that they protected against cannibalism-transmitted prion diseases. He derived the disturbing conclusion that cannibalism was once commonplace among our ancestors, and that prion diseases such as kuru — once a prime cause of death in New Guinea tribes that practised cannibalism — ravaged human populations in the distant past.

On a different note, it was suggested that the concept of prions as proteins that exist in two or more conformations, and replicate by causing other such proteins to change shape, applies not just to those proteins implicated in TSEs (R. Wickner, NIH, Bethesda, Maryland)5. Thus, in yeast, heritable traits can be attributable to prions under the following conditions: if they can disappear and later reappear (in the same or a later generation); if their propagation depends on the presence of the gene encoding the protein in question; and if their spontaneous frequency increases upon overexpression of that protein. If one accepts this definition, self-activating enzyme precursors (zymogens) might fall into this category. For instance, yeast proprotease B will self-activate by limited proteolysis (T. Roberts, NIH, Bethesda, Maryland, and R. Wickner). As the activated state is stable and transmissible by transferring the cytosol from one cell to another, it could be regarded as a prion.

Moving into the clinic, we find improvements in the diagnosis of prion diseases. The conformation-dependent immunoassay allows sensitive diagnosis by exploiting the fact that parts of the prion protein become inaccessible to antibodies during conversion to PrPSc (S. Prusiner, Univ. California, San Francisco). This led to the discovery that, in certain circumstances, the disease-associated prion protein is conformationally changed but is not protease-resistant (J. Safar, Univ. California, San Francisco)6. So, conformational assays might be inherently more sensitive than the venerable protease-based tests used currently.

But the most startling diagnostic development is the advent of a quantitative, sensitive assay that measures the ability of samples to infect prion-susceptible cells (C. Weissmann, Imperial College, London). The test is inexpensive and can be performed in days, as opposed to conventional infectivity assays in which scores of animals must be observed for months or even years. For now, however, only mouse prionologists will profit, as the test has not been adapted to other species.

Little progress was reported on the therapeutic side. The antimalarial drug quinacrine can cure prion-infected cells7, but hard evidence for any antiprion effects in vivo is still lacking8. Instead, quinacrine-treated CJD patients suffer severe liver damage (N. Streichenberger, Univ. Hospital, Lyons)9. Stimulation of a specific type of signalling pathway in the innate immune system delays prion diseases, perhaps by eliciting the production of anti-PrP antibodies (H. Kretzschmar, Univ. Munich)10, or perhaps because long-lasting stimulation might lead to immune disruption (M. Heikenwalder), which has been documented to slow the progression of prion diseases. Finally, the antiprion properties of soluble PrP derivatives might merit further study (A. Aguzzi)11.

It is clear that exciting times lie ahead in this field. Prion diseases are far from understood, and there are many bones of contention. We could not escape the exhilarating feeling that, more than in other areas of biology, fundamental discoveries are yet to be made.


  1. 1.

    *Keystone Symposium on Molecular Aspects of Transmissible Spongiform Encephalopathies (Prion Diseases). Breckenridge, Colorado, 2–6 April 2003.


  1. 1

    Ma, J. & Lindquist, S. Science 298, 1785–1788 (2002).

  2. 2

    Ma, J., Wollmann, R. & Lindquist, S. Science 298, 1781–1785 (2002).

  3. 3

    Drisaldi, B. et al. J. Biol. Chem. doi:10.1074/jbc.M213247200 (26 March 2003).

  4. 4

    Mead, S. et al. Science 300, 640–643 (2003).

  5. 5

    Wickner, R. B. Science 264, 566–569 (1994).

  6. 6

    Safar, J. et al. Nature Med. 4, 1157–1165 (1998).

  7. 7

    Korth, C., May, B. C., Cohen, F. E. & Prusiner, S. B. Proc. Natl Acad. Sci. USA 98, 9836–9841 (2001).

  8. 8

    Collins, S. J. et al. Ann. Neurol. 52, 503–506 (2002).

  9. 9

    Scoazec, J. Y. et al. Ann. Neurol. 53, 546–547 (2003).

  10. 10

    Sethi, S., Lipford, G., Wagner, H. & Kretzschmar, H. Lancet 360, 229–230 (2002).

  11. 11

    Meier, P. et al. Cell 113, 49–60 (2003).

Download references

Author information

Correspondence to Adriano Aguzzi.

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.