Adapted by Jennifer Altman from her original article in the May-June 2012 issue of Alzheimer Actualités, a newsletter published in French by the Ipsen Foundation. The Alzforum editors acknowledge the Foundation’s generosity in making this summary freely available in English.

31 May 2012. In the mid-1980s, Stanley Prusiner, University of California, San Francisco, startled the scientific world by claiming that transmissible neurodegenerative diseases such as Creutzfeldt-Jakob (CJD) in humans were caused by self-replicating protein molecules, which he named "prions." Painstaking work to establish that prions could replicate without the need for genetic material won him the Nobel Prize in 1997. What at first seemed an unusual mechanism, restricted to a rather rare group of diseases, is now becoming a unifying concept for the study of neurodegenerative conditions: All the pathogenic proteins that characterize these diseases seem to behave like prions. At this year’s IPSEN colloquium on Alzheimer’s disease, held in Paris on 27 February 2012, 13 international experts concluded that the "prion-like" hypothesis provides a framework for understanding all neurodegenerative diseases and opens up new directions for research into therapy. The meeting was organized by Mathias Jucker, Hertie-Institute for Clinical Brain Research and German Center for Neurodegenerative Diseases, Tübingen, Germany, and Yves Christen, Fondation IPSEN, Paris.

Prion proteins are Janus-like: In their normal, soluble globular form, they seem to participate in cellular functions, but in certain circumstances they adopt a pleated β-sheet configuration, which forms insoluble aggregates that disrupt cell function. These prion aggregates are found in neurons in a group of neurodegenerative diseases, known as the transmissible spongiform encephalopathies, that include Kuru and CJD in humans, bovine spongiform encephalopathy (BSE; mad cow disease) in cattle, and scrapie in sheep. All of these can be transmitted by ingesting brain material from affected individuals—the cause of great concern in the late 1980s and early 1990s when some people developed a rare form of CJD, presumed to be the result of eating products from cows with BSE.

By the 1980s, a long hunt had failed to find either a bacterial or viral agent causing these diseases. Prusiner and his colleagues proposed instead that the infectious agent was the β-sheet form of the prion protein, which was able to replicate using itself as a template to convert normal prion molecules into the β-sheet type (see Colby and Prusiner, 2011). This was the first claim for protein replication without the need for nucleic acids—controversial, to say the least. Now, it is well accepted that "rogue" prion molecules act as seeds that trigger the formation of β-sheets. Once in this conformation, the molecule forms fibrils that aggregate into an amyloid-like deposit. Various functional prions have been identified, but the best known are those that cause the spongiform encephalopathies (now also known as prion diseases), which take their name from the punctate appearance of the postmortem brain caused by the amyloid deposits that disrupt the structure and function of the infected neurons. Prions released from a cell can be taken up by neighbors and trigger the same cascade of transformation and aggregation.

The parallels between the prion diseases and Alzheimer's disease (AD) were soon noted: Cellular proteins amyloid-β peptide and tau both convert to a β-sheet, fibrillar conformation and form aggregates, the amyloid plaques and neurofibrillary tangles. More recently, Prusiner reported, this prion-like pattern of β-sheet formation and aggregation has been demonstrated in all the neurodegenerative diseases, including Parkinson’s, Huntington’s, and motor neuron disease, although the form and location of the aggregates are specific to each. As a consequence, said Claudio Soto, University of Texas Houston Medical School, Houston, these degenerative conditions are now all being designated as protein misfolding disorders, and Prusiner considers that the proteins responsible behave rather like prions.

These neurodegenerative diseases are also all similar in having inherited forms, in which gene mutations promote β-sheet configurations. In familial, early onset AD, mutations in amyloid-β precursor protein and in the presenilin genes boost levels of Aβ, increasing its propensity to form amyloid plaques. Mutations in the genes coding for the prion-like proteins associated with familial forms of the other neurodegenerative diseases have also been identified. However, inherited forms of all these conditions are rare; most cases are "sporadic," having no obvious genetic—or other—cause.

The prion diseases have two other important characteristics. First, they can be transmitted from one individual to another by ingestion of prion-containing material, and second, the seeding propagates with time from area to area through the brain, followed by degeneration. If the misfolded proteins associated with other neurodegenerative diseases do behave like prions, they should be capable of initiating protein aggregates in unaffected animals and triggering the transformation of the cellular protein in unaffected cells.

The first transmission of an amyloid disease was demonstrated in mice over 40 years ago: Systemic AA-amyloidosis, a complication of chronic inflammation in which serum amyloid A accumulates in various organs, though not in the brain, was transmitted to healthy mice by injecting them with various extracts from mice with the disease, said Per Westermark, Uppsala University, Sweden. The term "amyloid" applies to any deposit of insoluble, fibrillar protein in β-sheet configuration, and at least 15 amyloid-forming proteins are known, affecting various body organs, usually resulting in death because of compromised vital systems.

Prusiner surveyed several neurodegenerative diseases that have now been experimentally transmitted to disease-free animals. Brain β amyloid deposits have been shown to develop in primates after inoculation of disease material, but the deposits take years to develop (Ridley et al., 2006). To model diseases, a more useful approach is to use mice genetically engineered to produce the human disease protein. With these animals, results can be obtained in months rather than years, and mechanisms and possible therapeutics can be more easily investigated. Nevertheless, injection models are useful to study how pathology spreads. When a brain homogenate from older animals that have developed Alzheimer's-like pathology is injected into the brains of susceptible but disease-free younger animals, Jucker and Soto both found amyloid plaques in the younger mice much earlier than in their non-injected littermates. Brain homogenates from patients dying with AD had a similar effect in these mice (see ARF related news story).

In a somewhat different experiment, Michel Goedert and colleagues, MRC Laboratory of Molecular Biology, Cambridge, UK, injected brain extracts from transgenic mice that produced filamentous deposits of the mutant human tau protein into another strain of mice that had never developed tau deposits. These "clean" mice then formed tau inclusions, suggesting that, at least for tau, susceptibility to the disease is not essential for transmission to occur (see ARF related news story).

It should be stressed that this transmission is experimental and so far has been demonstrated only in mice. There is no evidence that any of the neurodegenerative diseases, besides pure prion-diseases, are infectious. They have never been shown to pass from one individual to another by contact. Jucker pointed to one potential danger: that these diseases could be transmitted by use of contaminated instruments during surgery—prion proteins are heat-stable, and appropriate care should be taken with prion-like proteins as well.

Brain homogenates are a mixed bag: A purer inoculate is required to ensure that the misfolded protein itself is responsible for triggering the pathology, Goedert argued. A mutant, insoluble form of tau produced tau aggregates in injected brains, whereas normal, soluble tau did not. Purified fibrils of α-synuclein, which forms deposits, or Lewy bodies, in a range of conditions including Parkinson’s disease, accelerated the deposition of endogenous α-synuclein when injected into mice, according to Virginia Lee, University of Pennsylvania School of Medicine, Philadelphia. The injected animals subsequently developed disease symptoms and died earlier than untreated control mice. The purified fibrils more effectively induced disease than did brain homogenates (see ARF related news story).

Seeding and Propagation
These experiments support the hypothesis that injected protein seeds the conversion of the normal forms of the protein into an insoluble β-sheet conformation. In the brain, this conversion seems to spread in a cascade from cell to cell—in line with previous work from Heiko Braak’s laboratory indicating the tempo-spatial propagation of neurofibrillary tangles through the brain in AD (Braak and Braak, 1991), and more recently of Lewy bodies in Parkinson’s (Braak et al., 2003).

Jucker, Goedert, and Lee all demonstrated that with time, the protein deposits—amyloid-β, tau, or α-synuclein—had spread from the focal inoculation site to other areas of the brain. The spread of tau pathology through the brain in a transgenic mouse model that expresses human mutant tau in a restricted area of the entorhinal cortex was also reported in a poster presentation by Alix de Calignon and colleagues from Bradley Hyman’s lab at Massachusetts General Hospital, Boston; similar results have recently been published by Karen Duff and colleagues (see ARF related news story).

The spread of protein transformation in the human brain was reported by Patrik Brundin, Lund University, Sweden, from observations in patients with Parkinson’s disease who had received grafts of fetal neurons into the substantia nigra 10 or more years earlier. At postmortem, graft-derived neurons in the substantia nigra contained typical Lewy bodies, indicating transformation of the endogenous α-synuclein in the donor cells by β-sheet molecules from the surrounding host tissue (note that Lewy bodies have not been found in all patients receiving substantia nigra grafts, and their occurrence may depend how the graft was made; see ARF related news story). Proof-of-principle of such transmission is also coming from rat and mouse models overexpressing α-synuclein.

Seeding and propagation can also be demonstrated in cell culture, in which endogenous and introduced proteins can be distinguished by labeling with specific dyes. Lee showed how purified α-synuclein fibrils taken up into primary neuron cultures induce the endogenous protein to form aggregates that can be transported from cell body to axonal terminals and vice versa (see ARF related news story). Similar findings in culture models of amyotrophic lateral sclerosis and Huntington's disease were described by Anne Bertolotti, MRC Laboratory of Molecular Biology, Cambridge, UK, and Ron Kopito, Stanford University, California. In amyotrophic lateral sclerosis, lesions caused by the misfolded protein superoxide dismutase-1 (SOD1) propagate through the spinal cord. Mutations in the SOD1 gene are found in some families affected by this disease. Highly purified mutant SOD1 is taken up into cultured neurons, where it triggers the formation of fibrils, which propagate from cell to cell and remain even when the original seeds are no longer present (see Münch et al., 2011). In Huntington's disease, a mutant form of huntingtin is taken up by non-neuronal cells in culture and persists over generations of cell division.

Cell-to-cell spread implies that the protein fibrils are released from one cell and taken up by neighbors. Observations in culture made in Bertolotti’s and Lee’s laboratories indicate that uptake involves a form of endocytosis, which is a process of engulfment normally used by cells to take up proteins too large to cross cell membranes. According to Kopito, two pathways seem to trigger endocytosis, depending on the charge on the molecule. Once in the cells, fibrils must escape endosomes if they are to seed aggregation of cytoplasmic proteins; this has been demonstrated in both the superoxide dismutase and the huntingtin cell models.

An unanswered question is how many molecules of a misfolded protein are needed to seed aggregation. Probably misfolding happens all the time during our lives, so why are some people more vulnerable than others? Bertolotti and Peter Lansbury, Brigham and Women's Hospital, Boston, Massachusetts, both suggested that the answer may lie in differences in the efficiency of protein quality control mechanisms and the clearance of damaged proteins.

Extrinsic Uptake and Cross-Seeding
Since ingestion of prion-containing material transmits the prion disease Kuru between people and probably the variant form of CJD from cows to humans, could this happen in other neurodegenerative diseases? Could sporadic forms of disease result from amyloid-like material taken up from the environment or in food that makes its way to the brain? There is some evidence for this in Parkinson’s, because, as Brundin reported, the first pathological changes connected with the disease are now thought to be in the brainstem nucleus of the vagal nerve, and in anterior olfactory structures (see ARF related news story). On an experimental level, using a transgenic strain of mice that develops Alzheimer's-like pathology, Jucker found plaques forming in young animals when they are injected intraperitoneally with brain homogenate from older mice with plaques. He thinks this indicates that the seed protein can cross the blood-brain barrier, though no mechanism has been established.

One potential route into the brain is through the blood circulation, which is known to carry amyloid-β peptide in patients with AD. Soto showed data from a transgenic mouse model of AD in which plaques formed in the brains of mice transfused with blood from plaque-bearing animals, unless the amyloid-β peptide had been removed from the blood before transfusion. Again, this is entirely experimental evidence and there is no indication that AD in humans can be transmitted by blood transfusion.

Because these pathogenic proteins all seem to have the ability to seed fibril formation, potentially one type of molecule could seed another. Limited cross-seeding has been found in the systemic amyloidoses and between amyloids from different species, reported Westermark. This raises a concern about amyloids in food: amyloid extracted from duck foie gras resulted in increased systemic deposits of serum amyloid A in transgenic mice. Another complication is the existence of endogenous inducers—for instance, the medin peptide that is found with amyloid A aggregates accelerates its aggregation in vitro.

Neurodegenerative diseases are notorious for their variable symptomatic presentation, disease progression, and brain pathology. The form that the aggregates take also varies according to the brain area in which it first develops; consequently, the deposits differ in type and location among diseases—Goedert and Lee cited tau and α-synuclein as particular examples. With tau, this may, in part, be due to variations in the composition of the aggregating fibrils, which differ with the type of deposit, but Goedert thinks the intracellular environment of the neuron may also be an important determinant. The medin example mentioned above suggests that specific co-factors could be involved.

Classic prion proteins exist in different strains that determine features such as disease incubation time, possibly reflecting differences in the structure of the prion protein. So far, little attention has been paid to the possibility of strains playing a part in the heterogeneity of other neurodegenerative diseases. One place to look is in the various forms of amyloid fibrils: Robert Tycko, National Institutes of Health, NIDDK, Bethesda, Maryland, described three types of fibrils—rapidly twisting, slowly twisting, and striated ribbons. These forms are determined not by the amino acid sequence, but by the symmetry of and the bonds between the β-sheets that make up the fibrils. The differences could account for variations in fibril toxicity and speed of propagation, suggested Tycko. The β-sheets themselves exist in two forms, depending on the arrangement of the atoms making up each sheet: The parallel is more stable than the anti-parallel. While most amyloids consist of parallel sheets, fibrils of amyloid-β peptide with one of the familial AD mutations are predominantly anti-parallel, slowly converting to parallel over time, perhaps as the fibrils become more stable.

Functional Prions
What function do prions fulfill? To quote Nobel Prize winner Eric Kandel, Columbia University, New York, the function of the brain is not to produce neurodegenerative diseases, and prions probably did not evolve to kill cells. Some cellular functions of prion proteins are known, particularly in fungi. Roland Riek, ETH Zürich, Switzerland, described a prion-like molecule known as HET-s/S that allows compatible colonies of the filamentous fungus Podospora anserine to merge by converting the soluble protein into its prion form, which spreads through the colony. HET-S can be toxic to incompatible colonies by making membranes leaky, so the HET-s/S system acts as a primitive immune system. Riek also showed how amyloid-like properties are employed by mammalian hormone-producing cells in the formation of high-density storage granules that will fall apart once secreted from the cell.

But Kandel reported a prion-like protein with what is perhaps the most fascinating function yet discovered: a molecule essential for the perpetuation of memories, the cytoplasmic polyadenylation-element binding protein, more familiarly known as CPEB. Current understanding of the molecular mechanisms of long-term memory storage is that the synthesis of new proteins strengthens activated synaptic pathways. Because the instructions for protein synthesis are generated in the neuron’s nucleus, the puzzle is how new proteins are targeted to the correct synapses. CPEB has a key role in this targeting (see ARF related news story). When activated by neuronal stimulation, it associates with the protein-synthesizing ribosomes in the excited synaptic terminal and activates translation of the messenger RNAs delivered from the nucleus. A prion-like sequence gives CPEB the ability to aggregate, allowing this action to continue, so that improved structure and function can be maintained even when the synapse is quiet. To keep its prion-like properties in check, CPEB also requires a controlled activation pathway, now being unravelled. This type of prion-driven plasticity has not been reported in other organisms. Prion-like activity may also be necessary for the function of the RNA-binding protein TDP-43 (see ARF related news story), which aggregates in amyotrophic lateral sclerosis and some forms of frontotemporal dementia.

With the recognition that all the proteins associated with neurodegeneration are prion-like, a more unified approach to intervention becomes possible. Prusiner identified three steps in the cycle of amyloid production that are obvious targets: first, the synthesis of the precursor protein—the risk here is of blocking any normal cellular function of the protein; second, the conversion of the soluble protein to its potentially pathogenic, β-sheet form; and third, the aggregation of the newly formed β-sheet molecules into fibrils. The pathway for preventing runaway activation of CPEB in neurons may shed some light on this third step.

Several speakers remarked on the lack of progress in developing meaningful therapies for any of the neurodegenerative diseases. One well-recognized stumbling block, pointed out by both Jucker and Soto, is the lack of an easily administered diagnostic test for early identification of patients. Indications that amyloid-β peptide circulates in the blood in mouse models of AD once again bring attention to the possibility of developing a simple blood test.

The other big problem, highlighted by Lansbury, is that many pharmaceutical companies have reduced their investment in developing drugs for treating neurodegenerative diseases because of the prohibitive cost of clinical trials—a typical trial lasts 18 months and costs about $100 million. More limited, much cheaper, proof-of-principle trials are required. A trial designed to test the efficacy of a compound that promotes the clearance of misfolded α-synuclein provides some pointers: Identify the drug target; find a measure of drug binding in the brain to establish the effective dose; measure the level of the seed protein in the brain, perhaps by microdialysis or by finding a substitute peripheral measure; and identify the optimal patient population. Meanwhile, Riek alerted us to one good thing that has emerged from studying amyloids: their use as a medium for controlling the long-term release of drugs from subcutaneous implants to treat various cancers and other conditions.—Jennifer Altman

Jennifer Altman is a freelance science writer living in Todmorden, West Yorkshire, UK.


  1. This is a fascinating example of what appears to be another "functional amyloid." In this , the function is more active and interesting because it is regulated by neuronal activity, rather than serving as a scaffold for melanin, or as a storage form for pituitary peptide hormones. It seems likely that it is regulated by a conformation change in the non-prion domain of CPEB in response to 5-HT. The properties of prions may be relevant to learning and memory: they are stable and "self sustaining," at least as long as their assembly from CPEB monomer balances their disassembly by chaperones or their degradation. Another potential property that may be important is their ability to replicate or divide, which might serve to amplify their activity. Prions also are transmissible, which raises the question of whether the CPEB prions are transmitted between neurons, as has been suggested for polyQ peptides, α-synuclein, and tau amyloid oligomers. I think it is likely that most amyloids have prion-like properties such as replication and transmission, but the difference may be that CPEB has been selected by evolution to do this in a highly regulated fashion, which is why it may be functional while the same properties in unregulated amyloids, such as Aβ, may be pathological. What the functional activity consequences of CPEB oligomerization are will be an interesting question.

    View all comments by Charles Glabe

Make a Comment

To make a comment you must login or register.


News Citations

  1. Seeds of Destruction—Prion-like Transmission of Sporadic AD?
  2. Traveling Tau—A New Paradigm for Tau- and Other Proteinopathies?
  3. Synthetic Synuclein Corrupts Native Along Mouse Brain Networks
  4. Mice Tell Tale of Tau Transmission, Alzheimer’s Progression
  5. Dopaminergic Transplants—Stable But Prone to Parkinson’s?
  6. Modeling Sporadic PD in a Dish?
  7. Parkinson's: An Unlikely Proposal Gains Momentum
  8. Prion Powers Synapses by Forming Amyloid
  9. Good Prion, Bad Prion: TDP-43 Domain Plays Both Sides

External Citations

  1. Ipsen Foundation

Further Reading

Primary Papers

  1. . Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol. 1991;82(4):239-59. PubMed.
  2. . Staging of brain pathology related to sporadic Parkinson's disease. Neurobiol Aging. 2003 Mar-Apr;24(2):197-211. PubMed.
  3. . Prions. Cold Spring Harb Perspect Biol. 2011 Jan 1;3(1):a006833. PubMed.
  4. . Prion-like propagation of mutant superoxide dismutase-1 misfolding in neuronal cells. Proc Natl Acad Sci U S A. 2011 Mar 1;108(9):3548-53. PubMed.
  5. . Tau, prions and Aβ: the triad of neurodegeneration. Acta Neuropathol. 2011 Jan;121(1):5-20. PubMed.
  6. . Very long term studies of the seeding of beta-amyloidosis in primates. J Neural Transm. 2006 Sep;113(9):1243-51. PubMed.
  7. . Aplysia CPEB can form prion-like multimers in sensory neurons that contribute to long-term facilitation. Cell. 2010 Feb 5;140(3):421-35. PubMed.