25 August 2008. The development of tau-based drugs for Alzheimer disease and other tauopathies depends on a deeper understanding of how tau begins to run amok in the first place. Molecular details for how this might occur in AD have emerged from mouse studies that have begun to zero in on how specific domains of the tau protein contribute to Aβ-induced neurodegeneration. Questions about which forms of tau help or harm neurons have long hampered the development of tau-based therapeutics. A recent study (Roberson et al., 2007) laid aside such concerns and showed, astonishingly, that AD transgenic mice could be protected from Aβ-induced cognitive decline and premature lethality simply by halving their levels of tau protein. By crossing tau knockout mice with human APP-overexpressing transgenics, Lennart Mucke’s group at the University of California, San Francisco, showed that lowering tau not only helped neurons resist the toxic effects of Aβ but also seemed to protect the brain from other excitotoxic insults (see ARF related news story).

Jürgen Götz and colleagues at the University of Sydney, Australia, wondered which regions of the tau protein mediated those powerful effects. To tackle this question, Lars Ittner in the group took tau’s “projection” domain (aka Δtau)—a 255-amino acid N-terminal segment that lacks the domain associated with microtubule binding and pathological aggregation—and expressed it in mice that express mutant human APP and lack full-length tau. These animals were generated by crossing APP23 mice with tau knockouts. At the International Conference on Alzheimer’s Disease in Chicago last month, Götz reported that expression of the Δtau fragment seemed to do just as well as the sweeping tau reductions at protecting mice from early death induced by APP overexpression. By 10 months of age, nearly half of APP23 animals die, but those expressing either Δtau or hemizygous levels of full-length tau enjoyed an 80-90 percent survival rate. APP23 mice with both tau manipulations got a complete rescue, with 100 percent surviving at 10 months. On a non-APP overexpressing background, Götz reported, both Δtau expression and tau reduction (tau-/-) reduced the mean severity of pentylenetetrazole-induced seizures by 30-40 percent compared to controls with normal APP and tau. In APP23 mice, each tau gene alteration increased latency for more severe seizures by about twofold. These findings extend another recent report from the Mucke lab (Palop et al., 2007) and tap into growing interest in abnormal network activity in AD.

How might the Δtau fragment protect against excitotoxicity? Götz reminded the audience that excitotoxicity is mediated by formation of a complex between an NMDA receptor (NR) and post-synaptic density protein PSD-95. This complex is stabilized by phosphorylation of the NR subunit NR2B by the kinase Fyn. The researchers examined these interactions and found that NMDA receptor surface expression dropped by about 75 percent in tau-/- mice relative to wild-type. This was associated with reduced assembly of the NR/PSD-95 complex, as neurons from both tau knockout and Δtau transgenic mice had no NMDA receptor subunit 1 (NR1) co-immunoprecipitating with PSD-95. Furthermore, both tau-/- and Δtau-expressing mice had lower levels of phosphorylated NR2B, which typically stabilizes the NR/PSD-95 interaction.

To see whether Fyn abnormalities might be responsible for the NR/PSD-95 defects, the researchers looked at Fyn-tau interactions and found less Fyn interacting with endogenous tau in Δtau transgenic mice. In both tau-/- and Δtau-expressing mice, Fyn was found at much higher levels in cell bodies than dendrites of CA1 neurons and at lower levels in synaptosomal preparations from hippocampi of these mice. Considering that Δtau is targeted to axons and excluded from dendrites in the Δtau mice, these findings suggest that Δtau traps Fyn in cell bodies, Götz said. This would prevent Fyn from phosphorylating NR2B and thereby destabilize the NR/PSD-95 interaction that mediates excitotoxicity. Consistent with this proposed model, Götz reported in Chicago that the peptide PSD-95 inhibitor Tat-NR2B9c was able to reduce Aβ toxicity in primary neuronal cultures in a dose-dependent fashion.—Esther Landhuis.

This concludes our three-part series on tau news from ICAD. See Part 1 and Part 2.