Note: Prion2014 has no formal policy on social media, so I have been limiting my blogging to published work that speakers have discussed, except where speakers have given me permission to mention unpublished portions of their presentations.
Aguzzi opened by opining that not all that aggregates is a prion. In contrast to Dr. Prusiner’s talk yesterday, he argued that PrP prions are fundamentally different from Aβ, Tau, etc. because they are readily transmissible between organisms. This is why he uses the term “prionoids” instead of “prions” for everything except PrP.
He defines prion strains as maintaining their properties on serial passage. He offered that strains might differ in their quaternary structure rather than their tertiary structure – the monomers are stuck together differently. He has been using polythiophenes, aka luminescent conjugated polymers, to type prion strains, as they bind to the backbone of PrPSc fibrils and emit different wavelengths of light depending on how the backbone is bent. His and Christina Sigurdson’s work with LCPs [Sigurdson 2007] is given as evidence that prion strain is indeed encoded in quaternary structure. They next showed that if you incubate LCPs with prions, you get extremely SDS-stable higher-order PrP aggregates, suggesting that LCPs hyperstabilize the aggregates
[Margalith 2012]. This is important because the crucial moment in prion propagation is when a prion fibril breaks into two – this is the only way you go from one propagon to two. Stabilizing fibrils with the LCP LIN1001, preventing this breakage, actually reduces prion titer, as measured by the scrapie cell assay. They got similar results in the misfolded protein assay (MPA). This prompted further inquiries about LCPs, and Aguzzi then went into great detail on unpublished work which I do not have permission to share here.
He next discussed his efforts to dissect the molecular basis of prion neurotoxicity. He believes that PrPC has a native function in myelin homeostasis which requires alpha cleavage [Bremer 2010]. In prion disease, PrPSc binds the globular domain of PrPC and induces a conformational change which moves the N terminus into a different position and makes it “the executioner”. This explains why prion neurotoxicity is cell-autonomous and restricted to PrPC-expressing cells. He believes that antibodies against the globular domain of PrPC induce the same conformational change [Sonati 2013]. The antibodies are therefore a mimetic of prion toxicity. To examine this question, they developed the cerebellar organotypic cultured slices (COCS) [Falsig 2012] which are the first in vitro model to exhibit robust neurodegeneration upon infection with prions. Using this model, they launched a variety of experiments to compare the neurodegeneration induced by prions vs. by antibodies, muchy if which is in [Sonati 2013], and the rest of which is unpublished and I cannot share here. The toxicity of the antibodies depends on the epitope to which they bind, and on the PrP molecule being bound. PrPΔ32-93 does not exhibit toxicity when bound by the POM1 antibody against the globular domain, apparently because its flexible tail is shortened; antibodies against the flexible tail also abrogate toxicity [Sonati 2013]. As a final point, Aguzzi created “N-GPI” mice which express only the N terminus of PrP, linked to a GPI anchor and these mice exhibit a cerebellar toxicity like that induced by POM1.
Aguzzi concluded by presenting his evidence that ICSM-18, which MRC Prion Unit considers to be a therapeutic antibody, induces toxicity just like POM1 does. He suggested that this toxicity was not seen in MRC Prion Unit’s efforts [Klohn 2012] simply because the dose used was too low, and that at higher doses such as 4 μM, ISCM-18 is indeed toxic. He questioned MRC Prion Unit’s unwillingness to share the antibody for further research. Simon Mead and Peter Klohn each took the podium to offer an alternative perspective. Mead disputed the claim that ISCM-18 and POM1 share an identical epitope, citing evidence that ICSM-18 has a linear epitope [Khalili-Shirazi 2005, Antonyuk 2009] while POM1 is reported to have a nonlinear epitope [Polymenidou 2008], and he further argued that there exists a therapeutic window between therapeutic and toxic concentrations of ISCM-18. Klohn pointed out that three different groups (Collinge, Solforosi and Horiuchi) have found antibodies raised against PrP α-helix 1 to be innocuous, and suggested that their epitopes may differ slightly from that of POM1.
For more background on the debate, see Aguzzi’s Invited Abstract #1 and watch this video of the responses:
Mallucci opened with a general pitch for the existence of a therapeutic window in prion diseases after synapse loss begins but before frank neuronal loss, a perspective which will be familiar to those who have followed her work [White & Mallucci 2009]. She went on to discuss the temporal relationship between synapse loss, decline in synaptic transmission, onset of behavioral defecits, and multiple markers of unfolded protein response (UPR) activation [Moreno 2012]. This prompted her work demonstrating the use of PERK inhibitors as oral drugs against UPR activation as a therapeutic strategy for prion disease [Moreno 2013], which I’ve blogged about previously. She showed several videos to demonstrate the improvement in behavioral signs, at a particular timepoint, in mice treated with PERK inhibitors.
She argued that PERK is potentially a target in several other neurodegenerative diseases as well, citing [Sidrauski 2013, Stutzbach 2013, Kim 2014]. One of these studies [Sidrauski 2013] used another small molecule, ISRIB, which appears to act downstream of eIF2α phosphorylation and may be hitting eIF2β.
After the talk, Ina Vorberg asked how Mallucci was able to get such a large effect with a lentiviral injection overexpressing GADD34 [Moreno 2012] – how much area of the brain was reached with lentivirus? Mallucci said she was not sure why the hippocampus seemed to be so important, but targeting the hippocampus seemed to give a therapeutic effect out of proportion to the volume of neurons reached.
Mouillet-Richard opened by motivating a search for PrPC‘s native function: how can PrP knockout mice be so healthy despite the conservation of the Prnp gene? There is evidence for PrPC having a role in neural stem cell homeostasis [Zhang 2006, Steele 2006]. She uses anti-PrP shRNA in stem cells as a tool to examine PrP’s native function. As her title, “The cellular prion protein controls Notch signalling and neuroectodermal stem cell architecture”, indicates, she has evidence that PrP controls Notch signalling. The only reference for this that I could find online was [Martin-Lanneree, Prion2012 abstract PO-219], so I am assuming this work is unpublished and I don’t have the speaker’s blessing to share it here.
Notch is upregulated in prion disease – see this post – which raises the question of whether some of PrPSc‘s toxicity is mediated by an exaggeration of PrPC‘s native function in regulating Notch.
Lindsay E. Parrie
Parrie began by reviewing the literature suggesting that PrP is pro-survival and promotes neuronal character/differentiation. PrP is expressed in olfactory sensory neurons, where there are no glia, offering a chance to view a pure neuronal population. Also, PrPSc accumulates in olfactory neurons. They therefore considered this brain region as a good model system to look at the role of PrP in neuronal differentiation.
When I looked in the literature, there was one paper on glial expression of PrP promoting neuronal survival [Lima 2007]. Presumably Parrie’s experimental system provides a new way to disentangle the role of glial PrP from that of neuronal PrP.
After the talk, Adriano Aguzzi praised the study but warned that genetic confounders could be a factor influencing the knockout phenotypes. Six of seven independent PrP knockout mouse lines (excepting Jean Manson’s Edinburgh line [Manson 1994]) were made using 129 ES cells in Black-6 embryos, and retain genetic linkage of the 129 mouse background around the PrP locus. Aguzzi has shown genetic linkage to be the true reason for at least one reported knockout phenotype [Nuvolone 2013]. Aguzzi has recently used TALENs to create a new line of PrP knockout mice that are purely Black-6 congenic [unpublished work of Mario Hermann] and he is happy to share these with the community.
Telling completely changed his topic from the originally scheduled topic, speaking instead about structural effects of PrP polymorphisms on prion propagation, a story which is currently in revision for publication. He cited some prior evidence for polymorphism influence on propagation in a variety of species [Telling 1995, Green 2008, Saijo 2013], and particularly for the S96 polymorphism in deer influencing CWD propagation. He then introduced his work with CWD and with the SSBP/1 sheep scrapie strain, and with additional polymorphisms. I believe all of this was unpublished but should be in the literature soon.
Tanaka opened with a general introduction to yeast prions such as Swi1, Cyc8 and Mot3. His talk focused on characterization of Mod5 as a novel yeast prion [Suzuki 2012]. Sadly, I am a bit too ignorant about yeast prions to understand what was and was not novel here.
Korth’s recent work has focused on defining a specific protein misfolding pathology in DISC1-associated schizophrenia [reviewed in Korth 2012]. He hypothesizes that chronic mental illnesses such as schizophrenia may be associated with misfolded proteins which cause a more subtle neuronal impairment than that seen in neurodegenerative disease.
Korth noted that schizophrenia is diagnosed based purely on introspection – i.e. the patient’s self-reported experience. Some patients only ever have one episode, others have repeated episodes of equal severity, others have one episode followed by “residuum”, and still others have a progressive course in which episodes become more frequent and have a large amount of “residuum.” DISC1 was the first gene associated with schizophrenia through linkage in a family with apparently autosomal dominant inheritance of schizophrenia, yet common variants in DISC1 are not associated with schizophrenia [Mathieson 2012]. The familial schizophrenia mutations in DISC1 cause behavioral disturbances in multiple mouse models. The DISC1 protein is believed to be a scaffold protein that interacts with many other proteins and takes part in signaling cascades. Korth has found that DISC1 protein is insoluble in schizophrenia brains and not in neurodegenerative disease brains. He believes that these are insoluble multimers of a much smaller size than the aggregates of, say, Aβ and α-synuclein seen in neurodegenerative disease. The insoluble DISC1 species are transmissible cell-to-cell [Bader 2012. Korth has further explored DISC1 through a transgenic rat model which appears to be unpublished but about which I did find one poster citation [Korth 2014, Poster #T123, Schizophrenia Research]. He further discussed the relationship between dopamine signaling and DISC1 [Bader et al, submitted].
Castilla introduced bank vole I109 PrP as a model of prion disease, citing [Watts 2012] and plugging his group’s poster #229 from this conference. Castilla has been performing PMCA on bank vole PrP as an in vitro model of transmission barriers [for a bit of background see Cosseddu 2011, Fernandez-Borges 2009].
Puig spoke about a C-terminal deletion of PrP which leads to a fatal disease in transgenic mice. She opened with a review of deletion mutants of PrP, which I’ve also reviewed here. See Altmeppen’s poster #65 at this conference.
Chernoff began with a general introduction to yeast prions, the advantages of yeast as a model system, and the functional workings of the yeast prion Sup35 as a nonsense suppressor. I believe much of this is covered in his most recent review of yeast prions [Liebman & Chernoff 2012]. It has been shown that some yeast prions can induce prionization of other yeast prions – for instance, Rnq1 induces [PSI+].
Chernoff has recently been working on making fusion proteins of human neurodegenerative disease proteins (including PrP) with the yeast prion Sup35 and expressing these fusions in yeast. He finds that many of these proteins facilitate the spontaneous formation of the yeast prion, which is phenotypically detectable. He shows that the neurodegenerative disease relate proteins form amyloid-like oligomers in yeast, suggesting that this is bringing the fused Sup35 domains together and lowering the barrier to their oligomerization and conversion into a prion. This assay could be used for studying the amyloidogenic and prionogenic properties of mammalian proteins by applying powerful techniques of yeast genetics.
He also reviewed some evidence from the Tycko lab for existence of different Aβ strains [Petkova 2005, Lu 2013], and evidence from his own lab that these strains can be maintained on transmission [Heilbronner 2013]. He argued that if it’s true that different Alzheimer’s patients have different strains, then it is important to be able to identify these strains in live patients to allow patient stratification for more effective clinical trials.
He ended on a note that the best possible therapeutic target for Alzheimer’s would be to target Aβ seeds early in the disease process before any clinical signs, and that this is the direction we need to head in, though we don’t know how yet.
Tagliavini presented evidence for the existence of Aβ strains. As a preface, he asserted his view that mouse models of Alzheimer’s disease are indeed highly relevant to the human disease, and that he would strive to convince the audience of this. He reviewed the phenotypic heterogeneity of early-onset Alzheimer’s patients with different PSEN1 mutations. Some have phenotypes more typical of frontotemporal dementia (FTD) or dementia with Lewy bodies (DLB). Similarly, in patients with APP mutations, the Dutch (E693Q), Italian (E693K) mutations are associated with cerebral amyloid angiopathy (CAA) or combined AD/CAA phenotypes while the Arctic mutation (E693G) is associated with AD. There are different neuropathological profiles associated with different familial AD mutations. The rest of his talk was unpublished data I can’t share here.
DeArmond pointed to the large overlap in (presumably neuropathological) signatures of neurodegenerative disease in the brain between prion disease and Alzheimer’s and between Alzheimer’s and Lewy body disease and said his goal for today would be to explain why these overlaps occur. Prion disease has an incidence of 1 case per million population per year, and Alzheimer’s has an incidence of 1 per 1500 per year, so the chance of a patient having both diseases at once is .0000001 [Note: I disagree with his math here - the correct comparison would be based on the prevalence, not incidence of the two diseases, and controlling for the fact that they affect the same age group (older individuals)]. Yet patients are sometimes seen to meet neuropathological diagnostic criteria for both diseases. Alzheimer’s often induces Tau hyperphosphorylation, and so does prion disease. Prion disease can also Tau hyperphosphorylation in the absence of Aβ pathology. Aβ and PrPSc can cause PrPC to induce Tau phosphorylation through Fyn or indirectly through Calpain activation and the Cdk5 pathway.
Pirisinu presented evidence that GSS cases caused by the P102L, A117V and F198S mutations are all transmissible. She has previously shown that the Nor98 atypical sheep scrapie strain shares features with GSS and has tested its transmission to bank voles [Pirisinu 2013 and Prion2012 PO-049].
Supattapone presented the work of his PhD student Geoff Noble, who is using the Supattapone lab’s model of PMCA on minimal purified components in order to get towards understanding the structure of infectious PrPSc. Supattaapone and Noble have two new stories here.
If you’ve followed Supattapone’s work at all, you know that he has pioneered running PMCA on only the minimal components – PrPC, co-purified phospholipids, and a PrPSc seed – in order to dissect what co-factors are essential for infectious PrPSc propagation [Deleault 2007, Deleault 2012a, Deleault 2012b]. Here’s his diagram of this process:
He has previously shown that phosphatidylethanolamine (PE) is essential for propagation of infectious PrPSc [Deleault 2012a]. As long as PE is included in the mix that undergoes PMCA, infectivity is propagated at titers comparable to those found in the brain. When PE is withdrawn, you sometimes get a total absence of any PrPSc, while other times you get a different PrPSc isoform which has a different size PK-resistant fragment which propagates serially. Supattapone calls this latter isoform “protein-only PrPSc“. Once you have protein-only PrPSc, even if you add co-factor back in, you never get back to infectious PrPSc. Because these two isoforms – infectious PrPSc and “protein-only PrPSc” – came from the same original seed and differ only in whether they were propagated with co-factor, he views them as a controlled system for determining what structural features are essential for infectivity.
The Supattapone lab has therefore used three different low-resolution techniques – deuterium exchange, conformation-specific antibodies, and Raman spectroscopy – to characterize the structures of these two conformers. He found one epitope that was much more exposed in protein-only PrPSc, which might represent a co-factor binding site.
The second story revolves around PrP misfolding in fatal familial insomnia prions. Supattapone reasoned that because genetic prion diseases such as FFI are highly penetrant, PrP with the D178N mutation must be much more prone to misfold than wtPrP, and therefore he wondered whether it could misfold without the help of a co-factor. When they applied PMCA to MoPrP-D177N without a seed, it spontaneously misfolded and formed a different conformer, with a different PK-resistant fragment size than PrP seeded with other mouse prions. In fact, this could occur without any co-factor, and the conformation could template onto wild-type PrP. However, de novo FFI PrPSc did require a phospholipid co-factor in order to convert wtPrP, and in contrast to the other prions they studied, it appeared to require a different phospholipid rather than PE.
One question-asker wondered whether the “protein-only PrPSc” might not represent a non-infectious strain, but rather simply a different strain which might be infectious to different species. Supattapone said he’s looking into that. The audience member also asked whether the spontaneous FFI prions from PMCA were transmissible. Supattapone argued that this would be hard to figure out since even authentic FFI prions from post-mortem human brains have only limited transmissibility to mice.
Aguilar-Calvo has studied the impact of goat Prnp polymorphisms on scrapie susceptibility. Most or all of this appears to be recently published work [Aguilar-Calvo 2014, Lacroux 2014]. The overall conclusion was that the K222 polymorphism is the most highly protective variant and is a good candidate for selective goat breeding campaigns.
Candace K. Mathiason
Mathiason is studying the blood-borne propagation of CWD and TME prions. This is a long-standing interest of hers [Mathiason 2006] and she has more recently been using RT-QuIC to detect prions in blood [Elder 2013]. The talk revolved around unpublished updates on this work.
Sadly, I had to leave early to get ready for the poster session and I missed the talks by Emmanuel E. Comoy, Ignazio Cali, Abigail B. Diack and Kirsty A. Ireland. I believe Comoy spoke about blood-borne prion transmission in macaques and Cali spoke about transmission experiments following up on his previous work on human strain typing [Cali 2009].