But besides PrPc’s interaction with PrPsc, it may also have a direct role in neurotoxicity, Williamson said. This is why he thinks uncovering PrPc’s normal function is important. Perhaps PrPc delivers a signal in vivo? To pursue this idea, the Scripps researchers developed an anti-PrPc antibody injection experiment. Results suggest that when PrP interacts with the antibody on the cell surface, it dimerizes and in this form then delivers an apoptotic signal to the neuron (see ARF related news story). In prion-infected brain, oligomeric PrP might have a similar effect as did the antibodies in this experiment, Williamson speculated. With regard to possible relevance to AD, formation of APP dimers or multimers is not heavily studied but has received some attention as a potential neurotoxic mechanism (see Scheuermann et al., 2001, Lu et al., 2003). Williamson noted that his data calls for caution in developing therapeutic antibodies for prion diseases.

In work that dovetails with Williamson’s, Corinne Lasmezas suggested that she observes in vitro the fatal consequences of dimerizing PrP in the membrane that Williamson’s group has modeled in vivo. Lasmezas works at the Neurovirology Service of the Atomic Energy Commission in Fontenay aux Roses, France. She presented new data on the possible role of oligomeric PrP species in the pathogenic mechanism. Her approach is part of a new avenue in prion research and appears to re-trace more established efforts in AD research to get an experimental handle on Aβ oligomers and their effects. Tau and α-synuclein oligomers are also under study. Put simply, the reasoning behind this work is that on their way to form large aggregates, proteins must pass through small oligomeric states. Might they be toxic? Lasmezas described different PrP peptides that lodge firmly in lipid bilayers. An inspiration to her current work came from Detlev Riesner at Heinrich-Heine Universitaet in Duesseldorf, Germany, who had described intermediate species of recombinant hamster PrP: a dimeric α-helical one and a tetra- or oligomeric β-pleated sheet structure (Jansen et al., 2001).

With distinct oligomeric species available for study (resulting either from a transient beta-sheeted monomer or from a covalently linked dimer intermediate), Lasmezas asked whether they might be toxic to neurons. To test this, she added the oligomers to cultured primary cortical neurons and assayed how the cells functioned and survived. Both types of oligomer were highly neurotoxic, Lasmezas reported, regardless of whether the cultured neurons expressed endogenous PrP. Certain PrP-binding antibodies protected against this toxicity, as did heparan-sulfate mimetics, which interfere with PrP interactions at the cell surface. Interestingly, the oligomers made of covalent dimers were not infectious, suggesting that separate mechanisms are responsible for infectivity and neuronal death. Lasmezas speculated that PrP oligomers have exposed hydrophobic side-chains that are normally covered. This would make them more prone to slide into the membrane bilayer and interact with membrane components or receptors in such as way as to signal apoptosis.

Questions from the audience echoed concerns frequently raised in the AD field: where exactly are these oligomers? Are they physiologically relevant? How can one visualize them and make them amenable to study? Researchers in AD still struggle with the technical difficulty of demonstrating the relevance of oligomers in vivo (but see ARF related conference story ), and there is no proof in humans. Even so, this line of research has won a growing number of converts over the last decade and led to a shift in the amyloid hypothesis.

In his presentation, Aguzzi described experiments aimed at coming to grips with the molecular mechanism of prion replication. His group picked up on the idea that a PrP—PrPsc heterodimer might form at some point, designed such a soluble dimer, and expressed it in transgenic mice. Crossbreeding showed that this dimer, dubbed PrP-Fc2, did not restore the vulnerability to infection of PrP knockout mice. Yet when crossed back into wild-type mice, the dimer blocked PrPsc accumulation following PrPsc infection in brain or spleen. In effect, when interacting—somehow—with normal PrPc, the soluble dimer slowed pathogenesis by interfering with deposition of the protein-resistant PrPsc, Aguzzi said. Immunoprecipitation experiments suggested the soluble dimer relocates and slides into lipid rafts only after the mouse has been infected, Aguzzi added. His group is currently using this dimer construct to try to identify any ancillary factors that some scientists still suspect are present when PrPc converts to PrPsc.

Aguzzi then reviewed current knowledge on the question of how prions get to the brain, a topic also studied by speaker Moira Bruce at the Institute for Animal Health in Edinborough, UK. In brief, prions enter the brain through the peripheral nervous system. They usually reach the spleen within days of infection but can take months or even years before entering the brain. They get in by way of the sympathetic innervation of the spleen. Cutting this organ’s innervation in mice delays neuroinvasion, whereas mice with excessive splenic innervation will show prions in their brain more quickly. Several labs focus their effort on defining the connection between the autonomic nerve endings and cells in peripheral organs such as the spleen. However, as with the nature of the prion, the exact point of this neuro-immune border crossing remains elusive, Aguzzi said.

What is the physiological function of PrP? It probably has one, otherwise kuru epidemics that may have raged through prehistoric human populations (Mead et al., 2003) would have selected for PrP-negative people. As in AD research, where the function for APP is nebulous and for Aβ even more so, prion researchers don’t have an answer to this question, either. And as in AD, researchers disagree on whether it even is a crucial question. Some speakers, including Williamson, think yes. Lasmezas said PrP might play a role in synaptic signaling and neuron survival related to its ability to regulate copper levels in the synaptic cleft. PrP interacts with heat shock proteins, other membrane receptors, and heparan sulfates from the extracellular matrix. Aguzzi, however, thinks not. “We should find out what that function is, but we really do not know that it will have anything to do with the pathogenesis. Once we know the function, it may prove irrelevant,” he said.

On this issue, a tantalizing mystery revolves around the Doppel (from “Doppelgaenger”) gene, which lies just downstream of PrP. The protein can interact with PrP and may, in fact, antagonize it (see also LeBlanc comment). Structurally and biochemically, the doppel protein resembles a form of PrP with its amino-terminal domain clipped away, and both can be neurotoxic in vivo. Normally, though, the brain expresses only minute amounts of doppel. To study this strange protein-protein relationship further, Aguzzi’s group knocked out Doppel and, when the knockout mice proved unable to breed, discovered that doppel is required for sperm to form properly (Behrens et al., 2002). While this is no answer to the question of PrP function, it suggests that perhaps a place to look for this answer is not the brain, but the testes, Aguzzi said.—Gabrielle Strobel.