25 November 2010. Mirroring a trend in amyloid-β research, work on the tau protein seems to be moving from tangles toward that infamous “O” word—oligomers. At this year’s annual meeting of the Society for Neuroscience (SfN), held 13-17 November in San Diego, new studies in this area sprinkled life into hallway conversations otherwise laden with funding worries (see ARF related conference story). This story distills some of the science behind the latest tau talk.
Neurofibrillary tangles containing hyperphosphorylated tau have long stood alongside amyloid-β plaques as pathological hallmarks of AD. However, in fly, fish, and mouse tauopathy models, synaptic impairment and neuron loss precede tangle formation (Wittmann et al., 2001; Paquet et al., 2009; Yoshiyama et al., 2007), suggesting that NFTs themselves may not drive disease. Stereological studies in human AD have revealed neuron loss prior to appearance of NFTs as well (see Gómez-Isla et al., 1997). Meanwhile, a growing body of research makes the case that tau oligomers could be the most dangerous tau species of all (see Marx, 2007; Berger et al., 2007; Brunden et al., 2008; Meraz-Rios et al., 2009). “It’s no longer just about plaques and tangles,” said Rakez Kayed, University of Texas Medical Branch, Galveston, said in his SfN talk on characterizing tau oligomers. “It’s time to start thinking about different Aβ and tau aggregates in the AD brain. Pre-filament forms of tau may be the most toxic and pathologically significant.”
To enable more detailed studies of this particular tau species, Kayed’s lab generated antibodies specific for tau oligomers. The researchers made polyclonal (T22) and monoclonal (TOMA) antibodies to tau oligomers they prepared in vitro by seeding soluble tau with oligomeric Aβ (Lasagna-Reeves et al., 2010). The antibodies do not recognize tau monomers in Western analysis of brain samples from AD patients or P301L (JNPL3) tau transgenic mice, nor do they pick up AT8-positive mature tangles, as judged by AD brain immunohistochemistry. But the antibodies do colocalize with the Tau 5 antibody, which recognizes both non-phosphorylated and phosphorylated tau in NTFs—i.e., early stages of tau aggregation.
Using their new antibodies, Kayed’s lab found elevated levels of tau oligomers—by ELISA, Western, and immunostaining—in postmortem AD brains, compared to control specimens. About 10 to 20 percent of total tau in AD brains is oligomeric, and the few tau structures found in control brains all appeared to be “pre-tangles,” Kayed reported at the SfN meeting. Both monoclonal and polyclonal antibodies can also detect tau oligomers in human cerebrospinal fluid by direct ELISA.
Other studies bolster the idea that tau oligomers, much more so than monomeric or fibrillar forms, are neurotoxic. As described on a poster, Cristian Lasagna-Reeves, a graduate student in Kayed’s lab, injected each of the three forms of tau into the hippocampus of wild-type mice, six animals per group. One day later, Lasagna-Reeves and colleagues analyzed the mice for cognitive deficits and killed them to extract brain tissue for immunostaining and biochemistry. Compared to the groups injected with monomeric or fibrillar tau, tau oligomer-treated mice fared poorly on parts of the novel object recognition test involving memory encoding and consolidation. The oligomer-treated mice also had more severe mitochondrial alterations and synaptic dysfunction, as revealed by immunohistochemical and biochemical experiments.
In a slide talk, James Moe of Oligomerix, Inc., in New York, described a separate set of studies demonstrating the toxicity of tau oligomers made in vitro by incubating different tau isoforms overnight at 37 degrees Celsius. Collaborating with Ottavio Arancio, Taub Institute, Columbia University Medical Center, New York, Moe and colleagues introduced tau monomers and oligomers into wild-type mouse hippocampal slices, and showed dose-dependent reduction of long-term potentiation in the oligomer-treated specimens. Confirming these findings in vivo, the researchers found that bilateral hippocampal injections of tau oligomers, but not monomeric tau, disrupted associative fear memory in mice tested the day after treatment. The findings support tau oligomers as “a good target for immunotherapeutic approaches” in AD and related tauopathies, Moe said.
Immunotherapy Targeting Pathological Tau
Tau-based immunotherapy is fairly new, but interest in the approach is growing, in part because Aβ vaccination trials (e.g., AN1792) have sustained hopes while exposing the perils of this strategy for treating AD. At the SfN meeting, Diana Castillo-Carranza, a postdoctoral fellow working with Kayed, presented a poster with initial findings of her mouse immunotherapy studies using the lab’s tau oligomer antibodies. She injected anti-tau oligomer antibody or control antibody into the hippocampus of eight-month-old JNPL3 transgenic mice, which develop tau pathology and motor impairment due to expression of the human P301L mutant tau transgene.
Transgenic mice that got a single infusion of anti-tau oligomer antibody (1 microliter of TOMA at 1 mg/ml) did far better than their mock-immunized counterparts when tested four days post-injection on the rotarod test measuring balance and coordination. At this point, it remains unclear whether the passive immunotherapy improved motor performance back to normal levels, as wild-type mice were not included in these as controls in these experiments. These animals should be old enough to analyze in early 2011, Kayed told ARF (see ARF related news story).
TOMA immunostaining of hippocampal sections removed from the JNPL3 mice after behavioral testing showed that the antibody infusion did clear tau oligomers from CA1 and dentate gyrus neurons. Western blotting of brain homogenates from vaccinated and control JNPL3 mice using the lab’s polyclonal anti-tau oligomer antibody also demonstrated removal of tau oligomers. When the researchers analyzed the same brain homogenates with Tau 5 antibody, which recognizes both oligomeric and monomeric tau, they not only saw reduction of tau oligomers on the blots, but also a significant increase in tau monomers, in the vaccinated animals. This is surprising and significant, Kayed noted, because it suggests that tau toxicity is associated with its oligomerization, not simply modification of the monomer. The findings “advise us to be careful in describing tau as soluble or insoluble,” Kayed wrote in an e-mail to ARF. “We have to be more specific.”
Scientists who saw the poster seemed intrigued by the findings. “Their work on tau oligomers is exciting. Trying to use their new antibody for passive immunotherapy is a reasonable thing to do, and that they see improvement in motor behavior is encouraging,” said Gal Bitan, who develops molecular tools for studying Alzheimer’s, Parkinson’s, and other diseases of protein misfolding at the University of California, Los Angeles.
However, the current studies do not address a number of key issues—among them whether treatment with anti-tau oligomer antibodies could improve cognition. The JNPL3 model has a predominantly motor phenotype, with only minor cognitive defects. “It will be interesting to see if their approach can be applied to models that are more relevant to AD,” Bitan noted. Kayed said his lab plans to start TOMA injections in Tg2576 mice, a widely used AD transgenic strain, in a few months.
Einar Sigurdsson and colleagues at the New York School of Medicine have data suggesting that vaccination with an AD-specific phospho-tau peptide not only clears tau pathology but prevents memory decline in a new tauopathy model more suited for cognitive assessment. The hTau/PS1 mice in this study express all six human tau isoforms (Andorfer et al., 2003) as well as mutant presenilin-1. The M146L PS1 transgene accelerates disease progression in the hTau/PS1 strain, which, unlike JNPL3 mice, develops tau pathology predominantly in the hippocampus and cortex and much less so in motor areas. The work is in press in the Journal of Neuroscience, and has been reported at previous meetings (see ARF 2008 ICAD Chicago story), though not at SfN.
In a prior study (Asuni et al., 2007 and ARF related news story), Sigurdsson and colleagues improved motor deficits of JNPL3 mice by vaccinating them with the same phospho-tau peptide used in the hTau/PS1 work. And at this year’s International Conference on Alzheimer’s Disease held in July in Honolulu, Hawaii, the lab reported that passive immunotherapy using an antibody (PHF1) targeting the same phospho-epitope led to motor benefits in treated JNPL3 mice (see ARF related news story).
How Do Tau Vaccines Work?
Conceptually, it may seem puzzling that tau immunotherapy works as well as it does in mice, given that tau pathology develops inside the cell, where it is harder for antibodies to reach. Amyloid deposits, on the other hand, are extracellular and, not surprisingly, easily cleared from amyloid mouse models by antibodies injected directly into the brain (see Wilcock et al., 2003; Wilcock et al., 2004). “My perspective is that for treating Alzheimer’s, an initial infusion of antibodies intracranially to clear deposits, followed by low-dose systemic administration to prevent future accumulation, would be an ideal strategy,” commented David Morgan of the University of South Florida in Tampa. He said Kayed’s new data support this approach for tau antibodies, and suggest further investigation of immunotherapy against tau oligomers as a consideration for treating Alzheimer’s. Morgan himself has unpublished data showing that intracranial injection of anti-tau antibody reduces histological tau deposits in the Tg4510 mouse inducible tauopathy model.
As for how antibodies might clear intraneuronal protein aggregates, Kayed proposed in a recent review (Kayed, 2010) that “depletion of the extracellular pool of [tau] aggregates by the antibody will shift the equilibrium between the intra- and extracellular pools of aggregates, hence leading to removal of the intracellular aggregates.” (See also Asuni et al., 2007, and Sigurdsson, 2008, and Sigurdsson, 2009 reviews that propose similar and alternative hypotheses.) At the 2010 International Conference on Alzheimer’s Disease in Honolulu, Pavan Krishnamurthy of Sigurdsson’s group reported results from brain slice experiments suggesting that antibody-mediated clearance of tau aggregates involves the endosomal/lysosomal system.
AD biomarker studies suggest that tau aggregates are released into the extracellular space, perhaps by dying cells (see Trojanowski et al., 2010). How tau oligomers are released from cells is unclear. In an SfN meeting poster, Kaoru Yamada of David Holtzman’s lab at Washington University School of Medicine, St. Louis, Missouri, and colleagues showed that their in-vivo microdialysis method was able to detect interstitial tau even in young wild-type mice, suggesting tau is secreted in the absence of neurodegeneration, while both Moe and Kayed reported detection of oligomeric tau in the CSF of AD patients.
Along with these data, and prior work showing that antibodies to α-synuclein boost clearance of the protein in a PD mouse model (Masliah et al., 2005 and ARF related news story) and α-synuclein oligomers secreted from cells may seed aggregation (see Danzer et al., 2009; Emmanouilidou et al., 2010; and ARF related news story), Kayed’s findings “raise the possibility that secretion of intracellular aggregates is a general phenomenon in the brain,” suggested Ben Wolozin of Boston University. “One wonders whether a combined Aβ/tau vaccine might work better on humans than simply the Aβ vaccine.”—Esther Landhuis.
Boutajangout A, Quartermain D, Sigurdsson EM. Immunotherapy targeting pathological tau prevents cognitive decline in a new tangle mouse model. J Neurosci. 2010 Dec 8;30(49):16559-66. Abstract