24 May 2004. In the waning days of March, researchers from around the world met in St. Moritz to send off the cold season with three days of science and skiing amid the glorious scenery of the Engadine Alps, in one of Switzerland’s most beautiful valleys. At 200 participants, it was the largest conference since the Swiss Society of Neuropathology began hosting this series 20 years ago. Titled “Growth and Death in the Nervous System,” the meeting was expertly organized by Adriano Aguzzi, Markus Glatzel, and their colleagues at University Hospital of Zurich. It featured a blend of science on prion diseases, Alzheimer’s, Parkinson’s, gliomas, as well as fundamental processes that cut across individual diseases, such as inflammation, proteasome function, and protein folding. The talks and discussions brought out intriguing parallels between research into prion diseases and Alzheimer’s, for example, an emerging focus on oligomers in prion research, or the struggle to come to grips with the problem of misfolding of PrP, Aβ, and α-synuclein . This summary is updated with more recent publications on some of the covered topics.
Prion Diseases: Few Answers, Many Questions
Charles Weissmann of the MRC Prion Unit, London, set the stage by reviewing current knowledge—and the most glaring knowledge gaps—of transmissible spongiform encephalopathies. In this group of diseases, neurodegeneration, ataxia, and dementia usually sets in after age 60. Once overt, disease takes a rapid course of about one year, although the incubation period may last many years. (Symptomatic Alzheimer’s progresses more slowly, though this disease, too, begins with a silent pathogenic phase of a decade or more.)
Human prion diseases include Creutzfeldt-Jakob disease, which can be sporadic or familial, the familial Gerstmann-Straussler-Scheinker syndrome and fatal familial insomnia. More notorious have been the acquired cases, either through tainted growth hormone injections, transplants, cannibalism rituals (kuru) or eating BSE-infected animal products (variant CJD). Prion diseases most obviously differ from Alzheimer’s in prevalence: they are all rare. Moreover, they have the added complexity of being infectious, with patients showing no immune reaction to the infectious agent. Alzheimer’s has never been considered a transmissible disease, and in practice still isn’t, though Mathias Jucker challenged this tenet for the sake of a fascinating argument.
Genetics has been less generous to prion researchers than to their colleagues in Alzheimer’s. Though more than 20 rare PrP mutations are known in families with familial CJD, no other causative, modifying, or risk-factor genes have since turned up. In Alzheimer’s, 4 genes are definitively known, and while the search for further ones has been a slog, linkage or association studies have at least produced many additional candidates that are awaiting independent confirmation (or invalidation) in separate patient samples or by newer methods. Genetic mosaicism is a potential factor in both prion diseases and AD, (see Beck et al., 2004). For more on neurogenetics, see John Hardy’s talk.
It took 30 years to purify the scrapie agent (scrapie is a prion disease of sheep), until Stanley Prusiner’s group at UCSF pulled it out of infected brain, Weissmann said. Its main component is PrPsc, a pathogenic and infectious form of the normal cellular glycoprotein PrPc. When years of searching for scrapie-specific nucleic acids came up empty, the field generally (though not completely) accepted Prusiner’s Nobel Prize-winning contention that the infectious unit contains only protein.
A major open question in prion disease is simply: What causes disease? Two hypotheses advanced for this question have not panned out, Weissmann said. One implicates a slow-acting virus, the other (the “virino” hypothesis) points to a yet-to-be discovered small oligonucleotide. The leading contender these days postulates that an abnormal version of the normally expressed host protein catalyzes the conversion of further PrPc into more abnormal forms. PrP knockout mice made in Weissmann’s lab by Hansruedi Bueler are healthy and normal. Surprisingly, the mice stay healthy when challenged with intracerebral inoculation of PrPsc, showing clearly that the cellular form of PrP must be present for prions to multiply, spread through the organism, and cause disease. This insight has raised important questions about the precise protein-protein interaction between PrPc and PrPsc.
Another nagging question is: What, precisely, is the infectious agent? “It’s the thing that causes prion disease,” quipped Aguzzi, adding that facetious as it may seem, this is as good an answer as the field can offer. Scientists can convert PrPc to PrPsc in a cell-free system, however, no one has been able to show that this protein is then fully infectious. “To this day, we cannot identify the infectious agent and study its structure,” said Weissmann. Similarly, perhaps, Mathias Jucker reported that endogenous but not synthetic Ab was able to seed amyloid deposition in mouse brain.
When mature PrPc converts to PrPsc, it becomes partially protease-resistant. This old finding has become a staple of prion studies, but its pathogenic significance, if any, remains unclear, Weissmann said.
Meanwhile, several investigators have cloned and expressed PrPc and PrPsc. Nobel laureate Kurt Wuethrich and his colleagues at the ETH in Zurich have solved the NMR solution structures of human, mouse, and bovine PrPc expressed in E. coli, and Wuethrich described their structural similarities and differences in St. Moritz. However, no one has yet solved the atomic structure of PrPsc. Like Aβ and α-synuclein , PrPc folds only partially, and contact with PrPsc turns the partially folded form into a conformation that is rich in β-sheet. Which other factors or environmental conditions stimulate the conversion of protease-sensitive, α-helical PrPc into insoluble, protease-resistant, β-sheeted PrPSc is the million-dollar question in prion research, many scientists agreed. Curiously, RNA reenters the picture here, as some labs have suggested that host RNAs might stimulate this crucial step (see Adler et al., 2003, Deleault et al, 2003).
Two hypothetical models exist for how PrPc might convert, Weissmann said. One is a stoichiometric notion that one PrPsc refolds one PrPc; the other is a “seeding model.” The latter posits that PrPsc would slowly form a crystallization seed while the equilibrium is strongly on the PrPc side, but then additional PrPsc would bind rapidly to this seed and form large aggregates. Peter Lansbury formulated this model for AD and scrapie (Jarrett and Lansbury, 1993; since then Lary Walker (Walker et al., 2002 and Mathias Jucker, among others, have expanded upon it but the precise process of protein misfolding remains unclear for both PrP and Aβ.
Both Weissmann and Aguzzi emphasized that another of the puzzling questions in prion research concerns the existence of strains. Scientists can characterize and propagate different strains from animals. How can this be, and what does the word strain even mean when the prion is a protein of one given sequence? Prusiner suggested that different conformations might account for this, as was later shown. New insight on this issue came out of a separate line of research dealing with prion-like elements that occur in yeast. This at first seems confusing because yeast prions are made of proteins other than PrP. Certain yeast proteins can spontaneously convert to a different conformation, and this change gets propagated onto further proteins, which then lose their function. One protein that converts in this way is the translation stopper sup35p. Weissmann quoted a recent paper from Jonathan Weissman (no relation) and colleagues at UCSF showing that different temperature conditions cause sup35p to convert into physically different structures (it forms amyloid fibers at 37 degrees) (Tanaka et al., 2004). The study provides an example for how different conformations of a protein give rise to distinct, stable prion strains driven by an environmental condition. Incidentally, prions occur not just in yeast and mammals. Prion-like properties have recently been reported in Aplysia synapses (see ARF related news story).
Finally, as in AD, no robust biomarker exists for prion diseases; the presence of PrPsc is the best surrogate marker available to date, Weissmann said. On this topic, Glatzel described Swiss and English cases of human CJD who deposit PrPSc in their skeletal muscles. His lab is investigating increasingly sensitive detection methods of intramuscular PrPsc to evaluate their potential as diagnostic tools (see also Kovacs et al., 2004).
In a related finding, French researchers yesterday reported in Nature Medicine that they detected PrPSc in muscle of scrapie-infected sheep. They spotted the first traces months before the sheep showed the first symptoms of disease (Andreoletti et al., 2004). While this work overlaps with Glatzel’s regarding the potential for developing future diagnostics, it may also stoke public fears about eating lamb and mutton. Andreoletti et al report that the infectivity of the muscle PrP was 5,000 times lower than that of brain PrPsc.
Bruce Chesebro, of NIH’s Rocky Mountain Laboratories in Hamilton, Montana, presented an extended analysis of how pathology develops after PrP infection in mouse models created in his lab. Species barriers limit this line of investigation, but it turns out that transgenic mice expressing hamster PrP develop disease when infected with hamster PrPsc. Chesebro showed images of tufts and plaques of prion deposits in mouse hippocampus. At first glance they resemble Ab amyloid deposits, but at higher magnification they clearly look extracellular as well as intracellular in neurons, astrocytes, and around blood vessels. Electron microscopy showed deposits near the plasma membrane and in lysosomes. These mice also have marked degeneration of dendrites and neurons. Chesebro compared PrP accumulation in astrocytes and neurons across different mouse strains that express transgenic PrP either in astrocytes, neurons, or both. This led him to suggest that whatever pathogenic changes occur in astrocytes lead to secondary, indirect damage of the neurons. As in AD, variability in transgene expression levels and the compressed time span of the pathogenic process confound the extrapolation of such data to humans. Chesebro’s lab is currently analyzing the role of microglia in prion pathogenesis. He also presented study of prion deposition and neurodegeneration in the mouse retina as a model organ that is more accessible than brain.—Gabrielle Strobel.