They say bad news comes in threes. According to a study published April 17 in the Journal of Biological Chemistry, tau trimers are the smallest oligomers able to sneak into neurons and seed new fibrils—a process thought responsible for spreading tau pathology in the brain. Led by Marc Diamond of the University of Texas Southwestern Medical Center in Dallas, the study’s other central finding was that oligomers of any size upward of a trimer were taken up in near-equal measure by neurons.

This broad range of oligomer sizes capable of triggering uptake surprised Diamond. “Going into the study, I thought there would be a size preference, perhaps something like a 20-mer,” Diamond said. “But instead, we found a whole range of sizes can work.”

Mounting evidence suggests that aggregated forms of tau protein can spread from one cell to another in the brain. The toxic aggregates somehow wind up outside of the cell and are taken up by neighboring neurons. Once inside, these aggregates may coax more tau protein to fibrillize. The molecular mechanisms of propagation are still unclear (see April 2015 conference news).

Diamond and colleagues had previously reported that tau oligomers gained entry into cells by binding to cell surface heparin sulfate proteoglycans. HSPGs contain motifs that latch onto regions in the tau protein, and Diamond believes that when tau engages the receptor on neurons it triggers uptake by a process called macropinocytosis. Tau that gets into the cell eventually seeds more aggregates (see Holmes et al., 2013). In this study, the group assessed which size aggregate held the key to HSPG portal.

First author Hilda Mirbaha and colleagues started off by separating tau fibrils into oligomers of various sizes. In a test tube, they fibrillized repeat domain tau, fluorescently labeled the mixture, then sonicated it for various times until a range of tau oligomers could be isolated by size-exclusion chromatography. The researchers confirmed the size and identity of each fraction with mass spectroscopy; they ranged from monomers to larger than 40-mers. The researchers employed a second technique, fluorescence-correlation spectroscopy, to confirm the purity of each oligomer fraction.

The researchers found that the oligomers in each fraction were surprisingly stable. Monomers, dimers, trimers, and larger species remained intact even after a freeze-thaw cycle. Different species did not readily coalesce. For example, when the researchers mixed trimers with 40-mers, the trimers stayed trimers.

The researchers next tested the ability of a spectrum of fluorescently labeled oligomers to bind to cells. They mixed different-sized tau species with HEK293 cultures at 4°C, a temperature at which protein can bind the cell surface but macropinocytosis does not happen. Flow cytometry indicated that tau species of all sizes—from monomer to 100-mer—attached to the cell surface equally well. Heparin, which competes for HSPG, prevented tau from attaching to the cells. So did chlorate, which prevents the sulfation of HSPG that is necessary for binding proteins.

To see which of the tau species would trigger engulfment, the researchers mixed each fluorescently labeled tau fraction with cells at 37°C to allow uptake, then treated the surface of cells with a protease to remove any fluorescent signal from tau species that were stuck outside. As assessed by flow cytometry, monomers and dimers did not enter the cell, but trimers and larger species did. Trimers were less efficient at triggering engulfment than 10-mers—about 20 percent of cells engulfed trimers, whereas 70 percent consumed 10-mers. However, the percent uptake remained constant from 10-mers on up to the realm of 100-mers.

Once inside the cell, could any oligomer corrupt endogenous tau? To find out, the researchers used the tau biosensor cell line they had previously developed. These cells express tau protein fused to one or the other half of the click beetle luciferase. When toxic tau seeds trigger aggregation, the two parts join, reconstitute the luciferase, and produce light (see Oct 2014 news story). Adding a range of tau species to the biosensor cell lines, the researchers found that trimers were the smallest capable of seeding. After that, seeding efficiency went up slightly with increasing oligomer size, leveling off from 20-mer to 100-mers.  

All told, the work suggests that tau species from trimers on up penetrate cells and seed aggregation. Would this hold true for tau isolated from human brain? The researchers separated tau protein from one AD and one age-matched control brain using size-exclusion chromatography and performed the same cell-binding, internalization, and seeding assays. While control brains primarily contained tau monomer, AD brains contained a range of oligomers, from dimers to larger than 20-mers. Human tau behaved strikingly like that made in vitro, Diamond said. All sizes latched onto cells in equal measure, but only trimers and above were capable of internalization and aggregation. These interactions were dependent upon HSPG binding and worked in HEK293 cells and primary cortical neurons.

Diamond said he was initially surprised that oligomers of all sizes bound the surface of cells with the same strength. This indicated no enhancement by binding multiple HSPG molecules at once. That three units of tau were necessary to trigger internalization could indicate that several HSPGs must be engaged, and perhaps cluster together, to initiate macropinocytosis.

Diamond does not know why oligomers of many sizes enter cells and trigger fibrillization equally well. His lab has identified different “strains” of tau species with varying levels of toxicity (see May 2014 news). Tau isolated from different brains can maintain unique characteristics, such as the form of aggregates they create, when it is transferred into another animal. Is a tau trimer big enough to form a strain? Diamond said his lab is currently investigating this question. He hypothesized that although trimers can propagate, only larger tau species may take on the flavor of a bona fide strain. He said it was possible that much larger oligomers likely exist in the AD brain, perhaps with different propagation characteristics. In this study the researchers did not test the upper limits of tau oligomer size due to restrictions of the assays.

Diamond said his findings indicate that therapies aimed at tau may need to broadly target a variety of species, from trimers upward. “That there isn’t a single assembly capable of seeding means we have to keep our minds open to the idea that an antibody or other therapy has to work against a range of sizes,” he said.

Patrik Brundin of the Van Andel Institute in Grand Rapids, Michigan, said the study elegantly demonstrated the minimal units of tau necessary for seeding to occur, and he hopes the α-synuclein field will follow suit. “We simply do not yet know what the minimal requirements are on α-synuclein assemblies for propagation of pathology to occur from the outside to the inside of a cell,” Brundin wrote. “Taking a similar careful approach to that of Mirbaha and coworkers would be very valuable.”—Jessica Shugart


  1. The work by Mirbaha et al. provides strong evidence that tau trimer is the minimal unit that is required for both entry to cells and induction of tau aggregation inside cells, hence for cell-to-cell transmission of the aggregates. The authors elegantly characterized the oligomers with various sizes that are derived from in vitro-generated repeat domain fibrils and brain-derived aggregates prepared from the brain tissues of Alzheimer patients, and both preparations exhibited essentially the same results. This is the most advanced work so far in terms of characterization of size-species with propagation ability and makes a strong case for the trimer being the smallest unit for uptake and aggregate induction.

    I would like to point out a few things that might be important in interpreting this interesting work. First, it relies on the premise that the fibril breakdown products are identical to de novo oligomers that are en route to fibrils. The forward/assembly process may involve a different set of oligomeric intermediates from the reverse/disassembly process. Second, when comparing the activities of oligomers and higher-order aggregates, one has to take into consideration the difference between activity per-mole of monomer equivalents and activity per-particle. The current work shows that the trimers are not more efficient in “seeding” aggregation on the monomer mole basis. That indicates relatively low per-particle activity. Finally, one cannot rule out the possibility that the trimers have to convert to larger assemblies before being able to “seed” aggregation in cells. In any case, thanks to this work, we might have the minimum unit of tau assembly that contains all the information required for cell-to-cell propagation of tau pathology.

    View all comments by Seung-Jae Lee

Make a Comment

To make a comment you must login or register.


News Citations

  1. Protein Propagation Real, but Mechanisms Hazy
  2. Cellular Biosensor Detects Tau Seeds Long Before They Sprout Pathology
  3. Like Prions, Tau Strains Are True to Form

Paper Citations

  1. . Heparan sulfate proteoglycans mediate internalization and propagation of specific proteopathic seeds. Proc Natl Acad Sci U S A. 2013 Aug 13;110(33):E3138-47. Epub 2013 Jul 29 PubMed.

Further Reading

Primary Papers

  1. . Tau Trimers Are the Minimal Propagation Unit Spontaneously Internalized to Seed Intracellular Aggregation. J Biol Chem. 2015 Jun 12;290(24):14893-903. Epub 2015 Apr 17 PubMed.