How prions damage neurons

Is PrPSc itself the toxic species?

Hughes and Halliday (Pathogens 2017) state that aggregated PrPSc appears to be minimally toxic and that smaller soluble oligomers are the more likely drivers. They report that oligomers are more toxic than fibrils in vitro and in vivo, and that the most infectious particles are non-fibrillar assemblies of roughly 14 to 28 PrP molecules. A transient intermediate between PrPC and PrPSc is also considered a possible toxic species.

Solomon, Schepker and Harris (2010) reach a related view from non-infectious PrP mutants. They write that aggregation, toxicity and infectivity are distinct properties of PrP. They note that pathology can appear with little PrPSc, that some familial prion diseases lack protease-resistant PrP, and that some animals carry abundant PrPSc with few symptoms. They hypothesise a separate toxic form, PrPtoxic. In mice carrying an octapeptide insertion (PG14), they describe small beta-rich oligomers of roughly 20 to 30 PrP molecules as toxic and large, tightly packed aggregates as linked to infectivity, and compare this to small amyloid-beta oligomers in Alzheimer’s disease.

Hughes and Halliday list the exact structure and identity of the toxic species as unresolved, along with whether PrPSc is mainly a marker of infection and not the cause of neurodegeneration.

What role does normal PrPC play?

Both reviews note that PrP-null animals are largely normal, so loss of PrPC function does not appear to be the main cause of disease. Hughes and Halliday describe PrPC as required for PrPSc toxicity. PrP-null tissue grafted into infected mice was not damaged, depleting PrPC during infection reversed early spongiform change, and anchorless PrP that replicates PrPSc did not cause disease, which suggests that membrane anchoring is needed. The normal functions of PrPC are covered on PrP function.

Solomon and colleagues propose a subversion model. PrPC acts as a signal mediator, and toxic forms deliver an abnormal signal through a putative transducer whose identity is unknown. They state that both gain of toxic function and loss of protective function may contribute.

Which downstream pathways are implicated?

Hughes and Halliday describe these findings, and say it is unclear for several whether they cause or follow neuronal injury:

• Synapse loss precedes neuronal death and correlates with behavioural deficits.
• Excess NMDA receptor activity contributes to neuronal death, and PrPC can attenuate it.
• Soluble oligomers may insert into membranes and form ion-channel-like pores.
• PrP oligomers may inhibit the 26S proteasome.
• The unfolded protein response, through PERK and phosphorylation of eIF2alpha, reduces protein synthesis and precedes neuronal loss. Several interventions on this pathway were neuroprotective in animals while PrPSc levels were unchanged. These are reported as laboratory findings, not as treatments.
• Autophagy is altered, but it is unclear whether this is toxic or protective.
• Apoptosis appears downstream, because deleting Bax or caspase-12 did not stop disease.
• Oxidative stress and mitochondrial dysfunction are reported without clear causal order.
• Microglia have both harmful and protective effects in the evidence they review.

What is unresolved?

Hughes and Halliday name why PrPC is required for toxicity, how misfolding begins in sporadic cases, whether downstream pathways are direct effects or secondary responses, and how strain differences alter toxicity. Solomon and colleagues add the transducer and downstream signalling components, and why neurotoxicity is hard to reproduce in cultured cells. Whether mutant PrP findings in mice apply to natural human and animal disease is also open. Strain differences are covered on strains. The class as a whole is on the mechanisms overview and the mechanisms hub.