Read a PDF of the entire series.
18 March 2013. More and more researchers believe that a neurodegenerative condition’s slow, inexorable takeover of a person’s brain might be driven by pathogenic, misfolded proteins that spread through interconnected neural networks and corrupt normal proteins along the way. It’s a wild idea that grew out of the prion field but is gaining acceptance. However, many basic questions remain, such as how these proteins enter and exit cells, and how they seed new aggregates. At the 11th International Conference on Alzheimer’s and Parkinson’s Diseases, held 6-10 March 2013 in Florence, Italy, a talk describing a means by which tau and α-synuclein can penetrate cells generated considerable buzz, not least because it points to new therapeutic targets for slowing disease progression. Other presentations focused on protein egress and toxicity. Several speakers emphasized that wild-type proteins migrate as well or even better than mutant forms, making this idea more plausible as a mechanism for sporadic disease.
Following an original line of research on Aβ, studies in recent years have shown that injecting a small amount of aggregated tau into mouse brain kicks off pathology at the injection site, which then migrates through connected brain regions (see ARF related news story; ARF news story). The same process occurs with Aβ (see ARF related news story) and α-synuclein (see ARF related news story), and it apparently involves transmission from cell to cell. Marc Diamond at Washington University in St. Louis, Missouri, previously reported that extracellular tau aggregates can enter cultured cells and seed fibril formation (see ARF related news story; ARF news story), but it is unclear how this works.
How Bad Proteins Get In
In an enthusiastically received talk at AD/PD, Diamond proposed a modus operandi for tau entry. He found that cultured primary neurons swallow tau aggregates along with extracellular fluid in a form of endocytosis known as macropinocytosis, in agreement with prior reports in the literature. In electron micrographs, Diamond saw wild-type human tau fibrils “sticking like Velcro” to cell membranes before being engulfed in vesicles. Some peptides are known to trigger endocytosis by binding to heparin sulfate proteoglycans (HSPGs) on the cell surface. Tau contains lysine-rich motifs ideal for binding HSPGs, suggesting it might enter through this route. In immunostainings, Diamond saw tau fibrils colocalizing with HSPGs. Moreover, adding heparin, or the heparin mimetic F6, to the cell media prevented uptake of the tau fibrils, presumably by competing with HSPGs for binding sites on tau. Diamond also stopped tau’s entry by knocking down a gene necessary for HSPG production, or by adding chlorate, which prevents HSPG sulfation. Altogether, the data showed that HSPGs are necessary for cells to internalize tau. To confirm the findings in vivo, Diamond injected GFP-labeled tau fibrils into mouse cortex along with the F6 heparin mimetic and saw no fibril uptake, in contrast to injections of tau alone, which were quickly ingested.
What form and size of tau propagate best? Diamond used mostly full-length tau, although in some experiments he worked with a fragment consisting of the 4R core aggregation region. In contrast to some recent reports that tau fragments are the most transmissible form (see ARF related news story; ARF news story), Diamond saw better uptake of the full-length protein. Only aggregates the size of trimers or larger were internalized, he told Alzforum; the monomer was not.
Intriguingly, α-synuclein, the villain in Parkinson’s and related diseases, also binds HSPGs, suggesting the same mechanism might apply. Diamond reported that heparin blocks α-synuclein seeding in cultures, but has no effect on the entry of huntingtin protein, which does not bind HSPGs.
The data could open up new options for tauopathy or α-synucleinopathy treatments. Ideas include using heparin mimetics, antibodies that block uptake of pathogenic proteins, or small molecules that interfere with HSPG production. Diamond told Alzforum he will target transferases that add sugars to HSPGs. It remains to be seen if pharmaceutical companies will embrace this approach. Jean-Francois Blain at EnVivo Pharmaceuticals noted that HSPGs, being large, sticky proteins, are difficult to work with. They also typically act with co-receptors, implying that another protein could be involved in tau internalization, Blain said.
Toxic Mysteries of Tau
In addition, Diamond found that cells that took up tau seemed to acquire distinct strains of misfolded protein that varied in their toxicity, solubility, ability to seed, and aggregation pattern. For example, Diamond’s “strain 9” produces a punctate pattern of staining inside neurons and consists of mostly insoluble tau. It propagates more readily and appears more toxic than the more soluble “strain 10,” which stains as a single perinuclear inclusion. Each strain, however, was faithfully passed on to all daughter cells over months of propagation in vitro, Diamond said. Adding lysate from a stable clone to a naïve cell culture re-created the exact strain. Diamond is collaborating with Stanley Prusiner at the University of California, San Francisco, to investigate whether different strains can be linked to specific clinical phenotypes. In response to audience questions, Diamond noted that mutant forms of tau give rise to distinct strains, and that he has not yet investigated what role tau phosphorylation might play.
Voicing a common opinion, Christian Haass at Ludwig-Maximilians University, Munich, Germany, called the work “fantastic” and promising. An engaged audience bombarded Diamond with questions. For example, Charlie Glabe at the University of California, Irvine, wondered how tau might get out of macropinosomes and into the cytoplasm. Diamond said he will use caged luciferin to look for leakage of endosomal contents. Another unanswered question concerns how tau deposits affect cells and whether they cause toxicity, said Karen Ashe at the University of Minnesota, Minneapolis. Ashe noted that cognition in mice improves when monomers are lowered but tangles remain (see ARF related news story). Diamond agreed that soluble oligomers, not large deposits, probably transmit toxicity.
Other talks complemented Diamond’s findings with in-vivo data. Karen Duff at Columbia University, New York City, extended her findings from transgenic mice that express mutant human tau only in the entorhinal cortex. From there, tau pathology spreads to the hippocampus and other connected regions (see ARF related news story). By the time the mice are two years old, functional MRI reveals hypometabolism in the entorhinal cortex and hippocampus, but the animals still perform normally in the Morris water maze. Duff noted this is analogous to what is seen in people at Braak stage 1 or 2, who have reduced brain metabolism on FDG-PET but no overt cognitive problems.
Modeling Sporadic Disease With Wild-Type Tau, α-Synuclein
Luc Buée at the University of Lille, France, wondered whether wild-type human tau migrates as readily as the mutant variety. Most studies of tau propagation have used mutant forms, but a majority of tauopathy cases are sporadic and involve only the wild-type species. First author Morvane Colin injected lentiviral vectors encoding human wild-type tau into rat hippocampus. Over five months, the protein spread to connected areas such as the limbic and olfactory regions, where it appeared inside neurons. By contrast, injections of mutant P301L tau stayed near the injection site, Colin reported. Buée speculated that the mutant protein may clump up more quickly, leaving fewer soluble species able to migrate through axons. The findings further implicate this mechanism of transmission in sporadic disease, Buée said.
Other groups are focusing on wild-type α-synuclein transmission as a model for sporadic PD. Virginia Lee at the University of Pennsylvania, Philadelphia, reviewed a model her group created by injecting wild-type human α-synuclein fibrils into the dorsal striatum of wild-type mice. Pathology spreads through the brain along anatomical connections, killing dopaminergic neurons by one month and impairing movement by six months, with cognition remaining normal (see ARF related news story). “The accumulation and spread of α-synuclein pathology reconstructs a neurodegenerative disease phenotype in healthy animals, kicked off by a single catastrophic event,” Lee said.
In the same vein, Elodie Angot, who works with Patrik Brundin at Lund University, Sweden, investigated the cell-to-cell transfer of wild-type human α-synuclein in vivo. She injected an adenovirus encoding the protein into the substantia nigra of wild-type rats. Three weeks later, all neurons in the substantia nigra expressed the protein. Then she grafted embryonic dopaminergic cells into the striatum of these rats. Two more weeks later, about a quarter of the grafted cells contained human α-synuclein, showing propagation from the substantia nigra to the striatum. The protein co-stained with an endocytosis marker, suggesting α-synuclein enters through endosomes, in agreement with Diamond’s tau data. Intriguingly, this wild-type α-synuclein spurred the aggregation of endogenous protein. Immunostaining showed the human protein forming the center of α-synuclein deposits. After treatment with proteinase K, which digests the human protein, the center of each aggregate disappeared (see Angot et al., 2012). An important caveat to the relevance of this finding for human disease is that the human protein was overexpressed.
Angot also spotted α-synuclein aggregates in oligodendrocytes. Since these cells do not express the protein, it must get in from outside. In the rat model, the aggregates, called glial cytoplasmic inclusions, appeared in immature, mature, and myelinating oligodendrocytes. Cultured oligodendrocytes readily took up monomers or small oligomers, but not lightly sonicated fibrils, which got stuck on the cell membrane, Angot reported. During the discussion, John Trojanowski at UPenn suggested that more thoroughly sonicated, broken-up fibrils might be internalized. The findings may be relevant to human disease. For example, α-synuclein deposits appear in oligodendrocytes in multiple system atrophy (MSA). However, questions remain about how well the findings will translate, scientists said. Another pathologist noted that in patients, oligodendrocyte precursor cells do not contain deposits, at odds with the in-vivo rat data.
How Bad Proteins Get Out
With all this attention paid to α-synuclein’s entry into cells, what about its exit? In Florence, Kostas Vekrellis at the Academy of Athens, Greece, addressed this issue. Although Vekrellis previously reported that α-synuclein can be secreted via exosomes (see Emmanouilidou et al., 2010), his recent studies suggest that this is not actually the primary route. In collaboration with Lawrence Rajendran at the University of Zurich, Switzerland, Vekrellis examined secretion in a neuronal cell line that overexpresses mutant α-synuclein. Vekrellis sonicated exosomes to break them open and measured their contents. He found that these vesicles contained but 1.5 percent of the total secreted α-synuclein. During the discussion, Michael Schlossmacher at Ottawa Hospital, Ontario, Canada, said that he sees similar values in his experiments. Vekrellis suggested that cells try to package α-synuclein in exosomes, but after the capacity of this pathway has been maxed out, they release free protein. He does not yet know whether free and exosome-associated α-synuclein represent the same species. For example, although exosome release is small, it could be that these vesicles carry the toxic species, he told Alzforum. Here, too, overexpression in cell cultures makes transfer of the results to the human brain difficult.
Vekrellis’ previous studies suggested that intracellular calcium regulates the release of exosomal α-synuclein. To look at free protein, Vekrellis infused depolarizing agents such as potassium chloride into the striatum of transgenic A53T α-synuclein mice by way of reverse microdialysis. Depolarization opens channels that flood cells with calcium, and, as expected, cells pumped out more free α-synuclein. Surprisingly, however, blocking potassium channels regulated by the sulfonylurea receptor 1 subunit (SUR1) cut α-synuclein secretion by half, even though this manipulation depolarizes cells and increases intracellular calcium. Likewise, activating SUR1 channels increased secretion. Vekrellis suspects the mechanism behind this may have to do with these channels affecting the burst firing of neurons, he told Alzforum. The SUR1 channel blocker glyburide is an approved medication for type 2 diabetes.
On the broader question of how these pathogenic proteins get out of neurons, a recent paper frequently came up in talks and hallway discussions. It suggests that neurons release tau quite physiologically in an activity-dependent manner that is influenced by intraneuronal calcium (Pooler et al., 2013).
These talks represent only some of the data on protein propagation featured at AD/PD 2013. We invite readers to highlight other pertinent posters and presentations through Alzforum commentary.—Madolyn Bowman Rogers.