Prion hypothesis proven?

In vitro infectivity study in Cell stirs tempest in a test tube

Apr 21, 2005
Brendan Maher(

Protein aggregates generated in a test tube infected wildtype hamsters with a disease much like scrapie, according to an article appearing this week in Cell. Such a demonstration has, in the past, been called the gold standard of proof for the prion hypothesis, Stanley Prusiner's Nobel-winning assertion that infectious, self-replicating protein isoforms are the culprit in transmissible spongiform encephalopathies (TSEs) like scrapie, Creutzfeldt-Jakob disease, and mad cow disease.

Study coauthor Claudio Soto, said that this demonstration, together with a paper published by Prusiner's group last summer, should allay most doubts. "There is really little room for skepticism," he told The Scientist.

But the study has done little to quiet prion hypothesis skeptics. "I'm not going to abandon alternative hypotheses for the time being," said Robert A. Somerville of the Institute for Animal Health, Edinburgh.

While Prusiner's group had successfully infected a mouse with a recombinant protein derived from bacteria, some argued that their use of transgenic mice susceptible to the disease undercut the power of the demonstration. In the new study, researchers at the University of Texas Medical Branch, Galveston, Universidad Autonoma, Madrid, and the University of Chile in Santiago fine-tuned a cyclical process for amplifying aggregated protein from an infected hamster brain. Through serial dilutions, they were able to infect a wildtype hamster with in vitro–produced aggregates without any traces of the original infectious brain. But skeptics, including a member of Prusiner's group, argue that using material from a diseased hamster brain could have resulted in residual contamination.

Soto's group has been using a process that they call protein misfolding cyclic amplification (PMCA), which aids the aggregation of the normal cellular protein PrPc into the misfolded, polymer-forming PrPres that is associated with TSE pathology. The process works in a fashion similar to polymerase chain reaction (PCR) amplification of oligonucleotides. After seeding PrPc with PrPres, the solution is incubated and sonicated. "Once the aggregates become long enough, we split them into smaller pieces so that in a new conversion, a new incubation, they are able to convert more and more of the normal protein," Soto explained.

Crucially, however, the PrPres "seed" comes from infected hamster brain homogenate, while the normal PrPc comes from healthy hamster brain homogenate. "They actually started from infectious material, and we didn't," said Giuseppe Legname, of the University of California, San Francisco, and co-author on the Prusiner paper. "It's an alternative approach to demonstrate that you might make prions, but to say that these are synthetic prions, it's very difficult."

Soto insisted that serial dilutions between rounds of PMCA reduce scrapie brain homogenate to an amount equivalent to a 10 to the minus 10th and a 10 to the minus 20th–fold dilution. Infectivity generally drops off after 10 to the minus 9th, according to the paper. "We've completely ruled out the possibility that the infectivity is still remaining from… the original brain," Soto said.

Laura Manuelidis, Yale professor and vociferous critic of the prion hypothesis, pointed to differences in survival between animals inoculated with PMCA-amplified PrPres and control animals inoculated with equivalent amounts of scrapie-derived PrPres. Control animals died from nearly 60 to more than 70 days earlier, indicating, she said, a vast reduction in infectivity for PMCA-derived PrPres. "If PrPres was infectious, you'd expect to make a lot of infectivity… I would argue that it's 100,000 to 1 million times less infectious than it should be." In the paper's results, the authors note a "significant difference" in infectivity for in vitro–generated PrPres, "between 10 and 100 times less infectious."

Somerville said, "I think it's probably at least on the order of 1000-fold. I think they've been quite generous to themselves, there."

Soto disagreed with these interpretations but noted that his group is "baffled" by the lower infectivity derived from in vitro–produced PrPres. "We think it's because the distribution of these aggregates is different in our preparation from the in vivo–produced material," he said. "We may have the same number of protein molecules, but organized in a different number of units." Alternative explanations include the possibility that they created a different prion strain, Soto said.

Legname noted that similar problems have made it difficult to provide definitive evidence. "I agree the amount of infectivity we created is very little," he said. "This isn't because we didn't create infectivity, but because we don't know how to create a high level of infectivity." This is why his group had used transgenic mice producing 16 times the normal amount of normal PrP, he said. The bacterially produced prions they created only worked in sensitive animals.

Still, researchers agreed that the PCR-like process Soto's team has refined may make study easier. "It will be very interesting to see whether these experiments can… work with other TSE strains," Somerville said. Regardless of the validity of the prion hypothesis, PMCA might aid in diagnostics. "If their work is robust and reproducible, then it should be applied sooner rather than later for the detection of prions in tissue where still it's very difficult to detect prions, like blood," said Legname.