Though hyperphosphorylated tau has long been implicated in Alzheimer disease (AD) and other tauopathies, such as frontotemporal dementia linked to chromosome 17 (FTDP-17) and Pick disease, just how it packs its pathological punch is not entirely clear. Are the paired helical fragments to blame, the destabilized microtubules, or is there some other downstream effect? Previously, Feany, at Brigham and Women’s Hospital in Boston, and colleagues identified an actin binding protein in a genetic screen for novel tau binding proteins (see Shulman and Feany, 2003). “Two things made us feel compelled to follow up on this line of investigation,” said Feany in an interview with Alzforum, “one being the recent flurry of papers detailing how microtubules and actin interact, and the other being the largely ignored, actin-rich Hirano bodies, which are a pathological hallmark of AD and other tauopathies.”

Bearing these observations in mind, first author Tudor Fulga set about to test if tau-actin interactions play any role in tau toxicity in Drosophila. Working together with collaborators Bradley Hyman and colleagues at Massachusetts General Hospital, Charlestown, Fulga used immunofluorescence microscopy to examine actin organization in flies expressing the R406W mutant form of tau that causes FTDP-17. The researchers found an abnormal amount of filamentous (F) actin that appears bundled up and looking like rods. These aggregates are similar to Hirano bodies in that a substantial portion of them also contain cofilin and cross-react with PHF-1 and AT-180, antibodies that recognize phosphorylated pathological forms of tau found in human Hirano bodies. The bundling of F-actin is not restricted to flies, either, because when Fulga and colleagues examined the brains of tau transgenic mice (rTg4510) they found extensive cofilin-positive, actin-rich rods in neuronal cell bodies. The data suggest that bundling of F-actin into these structures may be a common feature of tau toxicity.

One of the beauties of the Drosophila system is that it is easy to genetically manipulate the animals, and Fulga and colleagues capitalized on this to probe the tau-actin interaction in depth. They found that coexpressing an actin transgene exacerbated tau toxicity, as exemplified by a heightened rough eye phenotype seen in flies expressing the V337M tau mutant. On the other hand, coexpressing cofilin-Twinstar, a protein that destabilizes actin filaments, had the opposite effect, suppressing the rough eye feature. Similarly, decreasing total actin levels in a heterozygous actin null allele background also suppressed the toxic phenotype. These findings indicated that tau not only elicits changes in actin organization, but that those changes are related to toxicity. In fact, the researchers were able to correlate the number of actin-rich rods with neurodegeneration in the brain. In flies expressing R406W tau in an actin-rich genetic background, there were about fivefold more actin rods than in flies expressing the mutant tau alone, and the number of dying neurons increased by about fourfold. In contrast, both the numbers of rods and apoptotic cells were dramatically reduced when cofilin was coexpressed with the mutant tau.

Exactly how tau leads to formation of actin-rich rods is unclear. In-vitro experiments suggested that purified bovine tau alone can induce the formation of F-actin parallel bundles, but whether a monomer or larger form of tau induces these changes in actin needs to be elucidated. “We didn’t address aggregation of tau, but what form of tau is stabilizing the actin filaments is certainly a big question. While aggregation could be linked to the process, coaggregation with other proteins is also an interesting idea and certainly could be relevant, though there is no strong evidence for that,” said Feany.

It may come as no surprise that phosphorylation of tau, a key facet of tauopathies, seems essential for actin rearrangements in vivo. Expression of phosphorylation-incompetent tau elicited no actin bundles in flies, while pseudophosphorylated tau (all disease-related sites switched to glutamate to mimic phosphorylation) induced dramatic accumulation.

Two key observations seem particularly relevant to AD. First, mutant tau is not a prerequisite for the effects on actin—coexpressing the actin transgene with wild-type tau also leads to formation of actin-rich rods. This suggests that actin rearrangements may be an important pathology in tauopathies but are not directly caused by tau mutations. Second, Aβ exacerbates the process—when Fulga and colleagues expressed Aβ alone in flies, they failed to detect any actin rods, but in the presence of a wild-type tau transgene there was a dramatic increase in their number. Furthermore, the heterozygous null actin allele substantially reduced the combined tau/Aβ toxicity in flies, suggesting that the synergistic effects of the two proteins are mediated, at least partly, by actin.

Tau and Neurogenesis
Whether actin is related to potential neurodevelopmental properties of tau is unclear, but as Arendt, of the University of Leipzig, Germany, and colleagues suggest in the Hippocampus paper, changes in tau isoform expression might be related to migration, differentiation, and integration of new cells in the granule cell layer of the hippocampus, processes that all require cytoskeletal rearrangement.

First author Torsten Bullmann and colleagues report that a neonatal isoform of tau, which has neither of the two potential N-terminal inserts and only three of four potential microtubule binding repeats—the 0N/3R tau—occurs, contrary to expectations, in the subgranule layer of the hippocampus where adult neurogenesis takes place. Previous studies had suggested that only 4R tau is present in the adult rodent brain, but those studies may not have had the necessary resolution to identify 3R tau in such a tiny subsection of the brain.

To characterize which cells express the neonatal isoform, Bullmann and colleagues tested cells from the subgranule layer of adult rats for a variety of markers. They found that the cells expressing the 0N/3R tau also expressed doublecortin (DCX) and the highly polysialylated form of neural cell adhesion molecule, both transiently expressed in maturing neuronal precursors. The authors also found the neonatal isoform in cells that had incorporated bromodeoxyuridine, a marker of cell division. In contrast, the authors did not find DCX in any cells expressing tau isoforms containing one N-terminal insert (1N tau). “We conclude that there is a transient expression of 0N/3R-τ tau isoform during adult neurogenesis and a shift towards 1N-τ in mature granule cells,” write the authors. How this might relate to the adult human brain, where all possible isoforms of tau are reportedly expressed, is unclear.

Because 4R tau isoforms bind to microtubules with higher affinity than their 3R counterparts, this isoform switch may reflect a need for more dynamic microtubule rearrangements during adult neurogenesis, suggest the authors. “Therefore, expression of 3R-τ could serve as a marker for neuronal plasticity,” they add. In addition, studying the signaling pathways that lead to these isoform changes could shed some light on control of alternative splicing and isoform switching in adult brain.—Tom Fagan.

References:
Fulga TA, Elson-Schwab I, Khurana V, Steinhilb ML, Spires TL, Hyman BT, Feany MB. Abnormal bundling and accumulation of F-actin mediates tau-induced neuronal degeneration in vivo. Nature Cell Biology. 24 December, 2006. Advanced online publication. Abstract

Bullmann T, de Silva R, Holzer M, Mori H, Arendt T. Expression of embryonic tau protein isoforms persist during adult neurogenesis in the hippocampus. Hippocampus. 20 December, 2006. Abstract