Multiple lines of evidence have converged on the idea that toxic variants of Aβ, α-synuclein, and tau worm their way from one cell to the next, seeding the misfolding and aggregation of normal proteins as they go. In this manner, a growing number of scientists argue, Alzheimer’s, Parkinson’s, and frontotemporal dementias are similar to prion diseases such as Creutzfeldt-Jakob and bovine spongiform encephalopathy. But is this really so? In a Perspective in the April Nature Reviews Neuroscience, Dominic Walsh and Dennis Selkoe challenged the hypothesis and cautioned about drawing analogies to prion spread. They noted that much of the evidence invoking physical spread of pathogenic proteins can also be explained by the concept of selective vulnerability, whereby toxic proteins put certain cells under so much stress that they begin to form protein aggregates all by themselves. Even if fibrillar proteins from one cell do manage to seed aggregation in another, there is little to suggest those aggregates are toxic, Walsh and Selkoe claimed, because so far no evidence links the progressive pathologic lesion to the functional defects seen in disease. Identifying the neurotoxic agents should be the field’s top priority, they said. Why? Therapeutics that prevent protein spread may not curb toxicity, while toxicity might be prevented without the need to curtail spread.

Read the perspective, made freely available to Alzforum readers by Nature. We thank Nature Reviews Neuroscience for generously opening access to this article.

We apologize for the audio quality of the webinar recording below.




By Tom Fagan

Most neurodegenerative diseases are marked by the accumulation of specific protein aggregates such as amyloid plaques or neurofibrillary tangles. These aggregates sprout in one particular locale, such as the frontal cortex in the case of plaques, or the medial temporal lobe for tau tangles, before slowly spreading to other brain regions. The aggregates form when proteins misfold, assuming shapes that form multimeric assemblies, which eventually grow into larger fibrils. The aggregates spread, so the hypothesis goes, when minute amounts of misfolded proteins physically travel to unaffected neurons, where they coax normally folded proteins to misfold, seeding the process anew. The pathogenic spread hypothesis also suggests that Aβ, tau, α-synuclein, and other aggregation-prone proteins act similarly to prions, which interconvert between innocuous and toxic conformations.

Over the last two decades, evidence has grown to support this hypothesis. Taken from affected tissue, teeny bits of Aβ, α-synuclein, and tau can seed aggregation when injected into aggregation-prone brains of recipient animals. There are even indications that Aβ seeds can be passed from one person to the next via tissue grafts, raising the specter of iatrogenic Alzheimer’s disease, akin to iatrogenic Creutzfeld-Jacob disease, a prion disorder.

But has pathogenic spread been proven? In their Nature Reviews Neuroscience Perspective, Dominic Walsh and Dennis Selkoe, from Brigham and Women’s Hospital, Boston, said the answer is no, and asked the field to step back and critically reappraise the theory. They suggested more parsimonious explanations for what goes on in the brains of affected people and cautioned about drawing analogies to prion diseases, because it has not formally been proven that pathogenic prion conformations physically move from neuron to neuron. Walsh and Selkoe noted that the observed expansion of protein aggregates over time can as easily be explained by a selective vulnerability hypothesis. It posits that certain neurons poorly cope with toxic stress, creating a cellular environment ripe for protein aggregation. Aggregation could then spread along neural networks in the brain because one sick neuron could release factors that damage its neighbors, or one sick neuron could stress connected neighbors by insufficient or dysfunctional firing. In cases of familial Parkinson’s and frontotemporal dementias caused by mutations in synuclein and tau, the selective vulnerability hypothesis makes even more sense because neurons are preloaded with mutant proteins prone to aggregate.

The authors acknowledged that physical spread and selective vulnerability are not mutually exclusive, but they claimed that the former is harder to reconcile mechanistically. For it to happen, intracellular proteins, such as α-synuclein and tau, must be released from cells, travel to other neurons, be selectively internalized in such a way that they escape lysosomal destruction, and then template misfolding of normal cytosolic protein. Details for all of these steps are lacking, the authors noted. Furthermore, they emphasized that spread does not equal toxicity. While the evidence for seeding following injection in animal models may be conclusive, there is no evidence that this causes the animal any harm. For the hypothesis to hold, spreading lesions must be linked to functional deficits. In this respect, injection experiments showing spread of Aβ or α-synuclein aggregation differ from experimental prion infectivity, which does cause death or neurological disease in animals. In fact, Walsh and Selkoe avoided analogies with prions, since little is known about their neuron-to-neuron spread. “Likening AD and PD to disorders whose mechanisms of clinical dysfunction remain to be delineated does not provide mechanistic clarity,” they wrote.

Walsh and Selkoe suggested four steps to prove or disprove the pathogenic spread hypothesis:

  • more rigorous experiments to determine if spread and toxicity occur when seeds are injected into primate models;
  • better identification and tracking of toxic forms of proteins in rodent models;
  • development of PET tracers for prefibrillary aggregates to track in vivo spread of toxic seeds;
  • in the case of PD, more rigorous methods to reconcile aggregation of α-synuclein in enteric or other peripheral sites before emergence of synuclein toxicity in the brain. 


Q: Can we say toxicity and spread reciprocate each other?

Walsh: As both David Harris and Byron Caughey mentioned, the relationship between infectivity and toxicity in prion diseases is not well understood. Certain rare prion diseases appear to have little or no infectivity, and the identification of “subclinical” prion disease states suggest that infectivity does not always lead to disease. By analogy, if infectivity and toxicity do not exhibit a simple relationship, one might expect that “spread” and toxicity might not exhibit a simple relationship.

This gets to the very crux of our understanding of diseases linked to protein aggregation—do the aggregates cause disease, or does their presence simply reflect a change in proteostasis? Clearly there are certain conditions in which aggregates cause disease simply by physical disruption, but there are other conditions where aggregates per se do not appear to be driving pathogenesis. Aggregates of Aβ, tau, and α-synuclein allow “seeded aggregation”/templating, but the aggregates that propagate may not always cause dysfunction (e.g., Liu et al., 2015). Rather, dysfunction may arise from intermediates or off-pathways products.

Q: Could extracellular aggregates or oligomers not also be the stressful secreted factor, while not seeding per se?

Walsh: This is an interesting and reasonable idea. A scenario could arise whereby a neuron that contains aggregated protein releases some of these aggregates/oligomers and that the aggregates/oligomers induce an adverse effect on “connected” neurons without the aggregates/oligomers entering the target neuron. It is also possible that released aggregates/oligomers could induce unfavorable responses from microglia, which in turn could harm connected neurons. In a complex system, direct toxicity, secondary toxicity, and the transfer of aggregates from one neuron to another may operate simultaneously.

Q: What can be done to promote more epidemiological studies to test if infectivity really happens in humans?

Walsh: Further epidemiological studies are necessary and important and one would hope that recent publications that suggest iatrogenic acceleration of amyloid deposition would encourage such studies. However, because of the prevalence of AD and PD, and the likelihood that “transmission” in humans may accelerate rather than cause disease, such studies will require sophisticated approaches that were not necessary to identify rare causative iatrogenic transmission of prion disease.

Q: For therapeutic approaches, might genetics point us in a direction that helps us to understand spreading, e.g., release or uptake?

Walsh: Yes, one would hope that genetics will identify factors that will help explain the selective loss of certain populations of neurons in specific diseases.

Q: If aggregated proteins may be secreted into extracellular space in the forms of exosomes or other membrane structures, how can antibodies capture the encapsulated target protein in brain parenchyma?

Walsh: This is a reasonable question and one that we drew attention to in our review: “If these processes occur via exosomes or tunneling nanotubes (Fig. 2), they may not be accessible to extracellular agents, such as antibodies.” While there is evidence that small amounts of tau are present in exosomes, several groups have shown that “spread” of tau aggregates can be attenuated by certain antibodies. Collectively these results indicate that the “spread” observed in tau transgenics does not rely on exosomes.


  1. The possibility is raised that the spread of protein aggregates can easily be explained by the idea that cells under stress begin to form protein aggregates and thus protein aggregation can be initiated along "stressed" neural networks rather than by the spread of the protein seeds themselves.

    This possibility is very well taken and thus it is important to control for it!

    However, it is important to note that, in the seeding paradigm, prions beget prions in recipient cells, tau seeds beget tauopathy, and synuclein seeds beget synucleinopathy. Furthermore, in many studies two different seeds (strains) of the same protein have been used to test the two alternatives.  The observation that a seed with conformation A leads to the induction and spreading of conformation A while a seed with conformation B leads to the induction and spreading of conformation B is somewhat difficult to reconcile with the idea that the spreading of aggregation is the result of neural stress, which would be expected to have relatively nonspecific effects on downstream targets.

  2. This is a well-reasoned analysis of the protein-spread hypothesis that should be studied by every AD investigator. Some important take-away messages:

    • “The molecular mechanisms of PrP proliferation and neurotoxicity in ‘classical’ prion diseases are not fully understood and therefore referring to a pathogenic process as prion-like does not provide mechanistic precision.”
    • “The Braak staging system is neither proof of, nor an argument against, the spreading hypothesis.”
    • “Need to link the progressive development of histological lesions to actual functional effects. Longitudinal studies in transgenic mice suggest that the existence of neurofibrillary tangles per se may not necessarily disrupt neural function.”
    • “It should be emphasized that, even after more than 30 years of intensive investigation, the sites of conversion of cellular PrP into infectious and neurotoxic forms of PrP are unknown.”
    • “Moreover, there remains an urgent need to identify the actual neurotoxic agents in both prion diseases and non-PrP neurodegenerative diseases. This point cannot be overstated.”
  3. This certainly is an important and thoughtful commentary that keeps us focused on unanswered questions. However, I would take issue with the statement that “the molecular mechanisms of PrP proliferation and neurotoxicity in ‘classical’ prion diseases are not fully understood and therefore referring to a pathogenic process as prion-like does not provide mechanistic precision.” Certainly there are gaps in our knowledge, particularly with regard to cell biological mechanisms (when are there not?). But the literature on PrP-prion diseases is long and deep, and the basic molecular features that define prions – even in their variability - are sufficiently well established to give heuristic power to the prion paradigm as it applies to other neurodegenerative diseases.

Make a Comment

To make a comment you must login or register.


Paper Citations

  1. . Quaternary Structure Defines a Large Class of Amyloid-β Oligomers Neutralized by Sequestration. Cell Rep. 2015 Jun 23;11(11):1760-71. Epub 2015 Jun 4 PubMed.

External Citations

  1. this article

Further Reading

No Available Further Reading

Primary Papers

  1. . A critical appraisal of the pathogenic protein spread hypothesis of neurodegeneration. Nat Rev Neurosci. 2016 Apr;17(4):251-60. PubMed.