20 April 2007. While no one disputes that tau plays a critical role in diseases across a wide swath of the neurodegenerative spectrum, tau-based drug discovery efforts still lag behind those based around the amyloid hypothesis. This, researchers agree, is partly due to the difficulty of developing just the right mouse models, a problem that slows down mechanistic as well as therapeutic studies. Tau is catching up, though, and some of the energy was on display at the 8th International Conference AD/PD, held 14-18 March in Salzburg, Austria. This story offers some recent developments. It is not a comprehensive update, but as always, readers are welcome to improve it with their own observations from the meeting.
In reviewing mouse models for tau, Karen Duff of Columbia University, New York, noted that nearly two dozen have been made to date (see Alzforum Research Models), and these strains indeed model various aspects of tauopathy. Duff cautioned that none are complete models. Most share the drawback of expressing their human tau transgene more prominently in the spinal cord and brain stem than in the cortical, basal forebrain, and other areas that are primarily affected in their corresponding human tauopathy.
One new mouse model is different. Jürgen Götz of the University of Sydney, Australia, presented the first description of a strain that is a surprisingly precise model of early-onset human parkinsonism. Lars Ittner in Götz’s group used the cDNA for K369I, a human tau mutation known to cause Pick disease (Neumann et al., 2001. Pick’s is a form of frontotemporal dementia that often produces the movement symptoms of parkinsonism early in the course of disease. For more on Pick’s, see AFTD website). Ittner placed the K369I human tau gene under the control of the mouse Thy1.2 promoter, which drives postnatal expression and thus avoids developmental effects. He obtained mice that overexpress the mutant tau in the substantia nigra in addition to cortex, hippocampus, and amygdala. The substantia nigra is damaged in parkinsonism, but does not express the transgene in most published and widely used tau lines, such as P301L.
In the K369I mice, tau was hyperphosphorylated, and it aggregated with a histopathologic picture of Pick disease in that the lesions are positive for Bielshowsky and negative for Gallyas staining, Götz said. Between 4 and 6 weeks of age, the mice developed all four classic motor symptoms of parkinsonism: a resting tremor, rigidity, bradykinesia (i.e., slowness of voluntary movement), and postural instability. They also responded to treatment with L-dopa, increasing their performance on the beam test. By contrast, the dopamine receptor agonist haloperidol weakened them further. L-dopa worked only in young mice, not in older mice with more advanced disease, said Götz. Together, the symptoms and the drug response mimic parkinsonism in Pick disease, where patients also develop these symptoms early on and benefit from dopaminergic treatment at early but not late stages. Other researchers commented that the drug response might help drug developers validate this mouse model pharmacologically, and make it more practicable for drug screening studies than are some of the existing tau lines.
What causes the early motor symptoms? Aggregation of insoluble tau occurs later, so that can’t be it, Götz said. Already in young transgenic mice, the scientists noted an accumulation of tyrosine hydroxylase (TH), an enzyme needed for dopamine synthesis, in the substantia nigra. Neuronal culture experiments with K369I substantia nigra neurons and human wild-type tau-transfected SH-SY5Y cells indicated that tau impairs the transport of TH such that the enzyme does not reach the neuron’s terminals. In the mice, this happens in the absence of Wallerian degeneration, indicating that mutant tau makes certain neurons dysfunctional long before degeneration sets in. This work adds to a growing body of research on how tau alterations hamper axonal transport (reviewed recently in Lee and Trojanowski, 2006).
Further analysis will show whether, besides parkinsonism, the K369I also have some of the cognitive symptoms of tauopathy. A different new mouse model offers a look at this side of the coin. Luc Buee and colleagues at Inserm and the University of Lille, France, wanted to study the role of tau in Alzheimer disease, where tau pathology occurs in the hippocampus, entorhinal cortex, and other cortical areas. In making a new tau model, these researchers were motivated by the desire to avoid the limb weakness and paralysis that tau mice suffer as a consequence of transgene expression in the spinal cord, because this paralysis can make behavioral testing well nigh impossible. Curiously, by using human tau with two different mutations, they generated a model that has none of the early-onset parkinsonism seen by Ittner and Götz, but does exhibit the cognitive symptoms of tauopathy seen in AD. Joint first authors Katharina Schindowski and Alexis Bretteville placed a four-repeat tau construct bearing the G274V and the P301S FTD mutations under control of the same promoter Ittner used. This yielded several lines, of which line 22 is published (Schindowski et al., 2006). These Thy-Tau22 mice express increasing amounts of abnormal tau with age, in entorhinal cortex, CA1 subfield of hippocampus, amygdala, ventral thalamic nuclei, and other cortical areas but not in spinal cord. Abnormal tau first appeared in axonal tracts and later in cell bodies of neurons, and by 12 months the mice showed the full complement of tau pathology typical for AD. Starting at that age, the mice lost neurons in the hippocampus and showed a mild astro- and microgliosis. They had a deficit in basal synaptic transmission but not in LTP. They had no motor deficits but showed cognitive impairments in behavioral tests of anxiety, spatial learning, and spatial memory retention by 6 months of age. In toto, the Thy-Tau22 mice appear to model the tau component of AD, the authors argue.
In Salzburg, Buee presented follow-on data to suggest that these mice have changes in neurotrophic factors that mirror those seen in postmortem AD brain. The mice had a loss of BDNF mRNA and protein in hippocampus and cortex. Buee's group observed an inverse correlation between the AT100-immunoreactivity that indicates tau pathology and BDNF expression. However, keeping the mice in an enriched environment, which is known from other research to induce BDNF, restored BDNF expression even in brain regions strongly affected by neurofibrillary degeneration.
In short, these two new mouse strains hand scientists new tools to explore either the parkinsonian or the cognitive ends of the broad spectrum of clinical symptoms attributed to tau in various neurodegenerative diseases. Colleagues interested in obtaining mice, please contact the scientists directly at email@example.com or firstname.lastname@example.org.—Gabrielle Strobel.