6 March 2005. Followers of the Tao eastern philosophy advocate complete simplicity and naturalness; followers of the tau protein appreciate the latter, but can only wish for the former. At Molecular Mechanisms of Neurodegeneration, a focused meeting held in this western capital, there was a sense that tau might soon give up its secrets and enlighten us all. For though the protein is often enigmatic—is it good for microtubules or bad for microtubules?—we are closer than ever to attaining a fundamental understanding of tau, thanks in no small part to work from Eva-Maria Mandelkow's and Eckhard Mandelkow’s labs at the Max Planck Institute, Hamburg, Germany.
Of course, tau, a microtubule binding protein, is best known as the major component in the neurofibrillary tangles found in Alzheimer disease and other dementias. At the Dublin meeting, the Mandelkows (tau practitioners for many years) reviewed some recent revelations about the protein and strategies to prevent tangles.
It was apparent from Eva Maria’s presentation that to know tau is to understand time and space. In neurons, microtubules are essential for the transport of cargo between the cell body and the axons and dendrites. The Mandelkows previously demonstrated how tau, a large, natively unfolded protein that occupies a lot of hydrodynamic space, can inhibit this process—too much tau coats the microtubules, preventing access by other proteins such as kinesin, the motor that drives anterograde axonal transport (see ARF related news story). Their elegant video footage, familiar to meeting goers, showed that in N2A cells, elevated tau retards the anterograde movement of proteins, AβPP-carrying vesicles, and mitochondria away from the cell body and actually increases retrograde transport to the soma. The upshot is that neurites, starved for nourishment, shrivel and die.
Under normal conditions, cells are probably spared such ignominy by the controlled expression and phosphorylation of tau. Recently, the Mandelkows showed that microtubule affinity regulating kinase (MARK, also known as PAR1) can phosphorylate tau, making it yield its grip on microtubules (see ARF related news story). However, this process can go so far that the microtubules destabilize. For example, when CHO cells express MARK kinase (MARKK), MARK becomes highly active and the cells lose their microtubule network and rapidly apoptose, unless they are cultured in the presence of the microtubule stabilizer taxol, Eva-Maria reported. This pathway is at work endogenously in neuronal cells, she revealed, because when MARKK is knocked out in PC12 cells, they fail to differentiate when treated with nerve growth factor (NGF).
This on-again-off-again relationship between tau and microtubules does have a purpose. Eva-Maria showed how MARKK, MARK, and phosphorylated tau accumulate at the growth cone. Here there are practically no microtubules, but there is plenty of actin, and, indeed, when she stained growth cones for both phospho-tau and actin, she found that the two proteins colocalize. She proposed a scenario where phosphorylation of tau near the growth cone gives the microtubules enough temporary dynamism to advance toward the area of growth. Meanwhile, the phosphorylated tau, bound to the actin network in the cone, sits in wait for the incoming microtubule, which it then rejoins (presumably after being dephosphorylated).
How do these tau interactions relate to pathogenesis seen in various dementias? To answer this, Eva-Maria has made transgenic mice expressing human tau under control of the neural calmodulin kinase II (CaMK II) promoter and a tetracycline switch. The intention is to be able to control when and where tau is expressed and determine how age and tau expression may influence aggregation of the protein. Eva-Maria showed that when mice are reared with the promoter on, MARK phosphorylated tau (detected with the 12E8 antibody) is significantly increased in the pyramidal layer of the brain after three months. Tangle forming phospho-tau (PHF1 and AT8 antibodies) is not detectable, however, though the pathological forms of tau do appear later, at around 10 or 12 months, Mandelkow said. Because MARK phosphorylated tau is also the first to become elevated in Alzheimer brains, Mandelkow suggested that elevations in this form of the protein may be an omen that cells are trying to cope with a microtubule transport problem.
As for the influence of aging on tau aggregation, the animals that may answer that question are now only six-seven months old and Eva-Maria is waiting until they are older before she switches on human tau.
But what about the pathology? Can it be reversed? Perhaps, because Eva-Maria showed that in animals expressing human tau for nine months, a pan-tau antibody revealed dendrites packed with the protein, but when the promoter was then switched off for six weeks, the tau disappeared.
Eckhard Mandelkow also spoke to the possibility of reversing tau aggregation. First he addressed the question of why this intrinsically unstructured protein aggregates at all. Eckhard showed that tau shows very little secondary structure even when bound to microtubules, and it still has low intrinsic stability when it forms paired helical fragments (PHFs) —very low concentrations of guanidine hydrochloride, a commonly used protein denaturant, are sufficient to dissolve PHFs, he reported. This data, in fact, suggests that reversing tau aggregation might be easier than one would have thought.
So why does tau form PHFs? Last January, Eckhard showed that the aggregation of tau, like Aβ and many other fibrillogenic proteins, is driven by β-sheet formation (see von Bergen et al., 2005). In tau, the β-sheet originates from the repeat region, which is necessary for both microtubule binding and formation of paired helical fragments. Biophysical measurements, including infrared, fluorescence and circular dichroism spectroscopy, and x-ray diffraction, conducted at Eckhard’s lab indicate that two hexapeptide motifs in the second and third repeats can adopt a β-sheet structure, which also accompanies conversion of soluble to aggregated tau. Significantly, these two hexapeptides are close to the P301L and ΔK280 mutations that have been linked to familial cases of frontotemporal dementia. Currently, Eckhard is working in collaboration with Christian Griesinger at the Max-Planck-Institute, Gottingen, to study tau secondary structure using NMR. Spectra obtained from C13- and N15-labelled tau confirms that β structure is centered on the hexapeptide motifs in the repeats, Eckhard revealed. This work, still in progress, also confirms that the same amino acids are involved in both microtubule binding and PHF assembly, he reported.
Armed with the knowledge of how tau aggregates, Eckhard is now conducting high-throughput assays for chemicals that can inhibit or reverse tau aggregation. To date, over 200,000 compounds have been screened and of these, 1,266 inhibit PHF formation and 77 can break down pre-formed PHFs. These 77 were used in a homology search to find other potential PHF busters, and 241 of these are being studied in depth, he reported (for some examples, see Pickhardt et al., 2004). Eckard showed how these compounds reverse tau aggregation in N2a cells that are engineered with an inducible tau expression system. After about nine days, cells expressing a fast aggregating (ΔK280 mutation) tau have abundant PHFs and they also release lactate dehydrogenase (LDH), a sign of toxicity. The PHF busters stop aggregation and stem LDH release.
All told, data from the Mandelkows and others suggest that tau is all about yin and yang. Tau walks a fine line on microtubules. If that line is packed with tau, then axonal transport is blocked. In contrast, if there is not enough tau on the microtubules, then they fall apart. Excess or microtubule-free tau is also susceptible to aggregation, which can possibly be prevented in vivo with small molecules.—Tom Fagan.