28 November 2005. This is the second installment of a four-part news series about the role of the microtubule-associated protein tau from the 35th Annual Conference of the Society for Neuroscience, held November 12 to 16 in Washington, D.C. See also Introduction and Part 1, Part 3, and Part 4.
Four-Repeat Tau: A Shifty Character
In one of five side-by-side posters showcasing careful expression profile analyses of single neurons from human neurodegenerative disease tissue, Stephen Ginsberg, from the Nathan Kline Institute in Orangeburg, New York, presented new research on tau. Other groups had found that the human brain creates six different isoforms of tau by alternatively splicing the freshly transcribed tau pre-mRNA. Three isoforms come with three tandem repeats, that is, three microtubule binding domains (3R tau), and three other isoforms contain four tandem repeats, that is, four microtubule binding domains (4R tau). In normal brain, 3R and 4R tau occur in about equal measure. Tau mutations in some inherited tauopathies change this ratio prior to tangle formation and cell death (for reviews, see Goedert, 2005; Lee et al., 2001).
Since cholinergic basal forebrain neurons are particularly vulnerable to neurofibrillary tangles in early AD, Ginsberg aspirated single neurons of this type from postmortem tissue of people who at death had been either cognitively normal, diagnosed with mild cognitive impairment, or with AD. Ginsberg collaborated with Elliott Mufson at Rush University Medical Center in Chicago on this study, which used tissue samples from nuns participating in the Religious Orders Study. After having isolated and amplified the mRNA in these single neurons, the scientists used custom-made cDNA arrays to analyze the levels of 3R and 4R tau from these cells. Then they compared the new data to archived data gathered from hippocampal CA1 neurons of a similar grouping of cases, and to data from schizophrenia samples for an additional control. They confirmed earlier knowledge when they found that overall levels of tau expression did not differ among the groups. They hit something new with their finding that people with AD and already with MCI, but not normal aging or schizophrenia, had a shift in the ratio of 3Rtau to 4Rtau such that they had more 4R tau in these particular neurons in basal forebrain and the hippocampal CA1 field. This indicates that not the expression level per se, but some change related to the splicing machinery or regulation may help explain the onset of tauopathies and part of the progression to AD, Ginsberg speculated. This data is in press at the Journal of Neurochemistry.
Speaking more broadly, Ginsberg emphasized the importance of single cell gene expression profiling in order to understand what goes wrong in particular neurons. The regional vulnerability of specific types of neurons remains a mystery in many neurodegenerative diseases. Profiling expression patterns in brain regions, even fairly narrowly circumcised ones, tends to cancel out important effects at the single-cell level, Ginsberg said. “These are complex disorders, where regional vulnerabilities are important. The answer will be in the details, and we need to look at local changes,” Ginsberg said.
Recent technological improvements in laser capture microdissection, RNA isolation and amplification, as well as the design of microarrays and qPCR have led to a wider acceptance of expression profiling from pure populations of cells or from minute quantities of tissue from specific subregions. The SfN conference featured a growing number of presentations in this area. At the same time, researchers are still wrestling with problems such as the minute size of the samples, low yields, and poor integrity of the RNA. Independent validation of the obtained profiles remains a challenge, as well (see Galvin and Ginsberg, 2005).
Toward an RNA-based Tau Treatment
What is it about the tau pre-mRNA that determines the 3R-to-4R shift in tau splicing? And could small-molecule drugs conceivably tweak this process therapeutically? These questions prompted Michael Wolfe’s lab, formerly known as having research interests squarely in the amyloid side of AD research, to enter the tau field. Christine Donahue, a postdoctoral fellow in Wolfe’s lab at Brigham and Women’s Hospital in Boston, picked up a clue from Mike Hutton, who had suggested that the tau pre-mRNA forms a stem loop structure at an exon-intron junction of exon 10. This stem loop, so the idea goes, controls access of a particular splicing factor to its binding site and thus influences whether exon 10 ends up included in the mRNA (this leads to 4R tau) or left out (this yields 3R tau). It was also known that FTDP-17 mutations lead to a decrease in the expression of 3R tau, possibly due to a predicted decrease in stem loop stability.
Donahue approached this question first by stabilizing the proposed tau stem loop in wild-type and FTDP-17 tau with a series of mutations in vitro. This reduced exon 10 inclusion in the splice product and argues for the existence of a stem loop in the tau pre-mRNA. A dementia-causing FTDP-17 point mutation destabilized the stem loop and increased exon 10 inclusion. Next, Donahue added a luciferase reporter to an existing wild-type tau minigene, which allowed her to detect small molecules that stabilize the stem loop. She found that Geneticin, an aminoglycoside antibiotic known to bind to RNA, indeed stabilizes the stem loop and reduces 4R tau production in vitro. Geneticin is neither sufficiently selective nor able to cross the blood-brain barrier (BBB) to become an AD drug itself. However, it proves the principle that tinkering with this RNA structure can influence tau production, and has encouraged an ongoing screen in Wolfe’s lab for more suitable compounds with the same function, Donahue said.
No tau-specific treatments are yet on the horizon, though approaches are entering the pipeline. The drug taxol has shown some promise in animal models, and a search for less toxic microtubule-stabilizing compounds that cross the BBB is underway at the University of Pennsylvania School of Medicine in Philadelphia, Lee said. Illana Gozes, of the Sackler School of Medicine in Tel Aviv, has developed a neuroprotective peptide called NAP (see ARF related news story), which appears to act in a similar way and competes with taxol for binding on microtubules. In Washington, D.C., Gozes reported that a small, double-blind, randomized phase 1a trial of an intranasal formulation of this compound has been completed. Inhibitors for certain tau kinases, for example, GSK3β, are under active study in the pharmaceutical industry. Lithium, an established drug widely used to treat bipolar disorder, also appears to stem tau pathology, and the Alzheimer Disease Cooperative Study is preparing to test it in a clinical trial of patients with mild cognitive impairment starting next year, Lee added. Compared to these approaches, stem loop stabilizers targeting RNA would represent a new generation of mechanism-based drugs; they are at the early discovery stage.—Gabrielle Strobel.
See also Introduction and Part 1, Part 3, and Part 4 of this series.