27 September 2010. Many efforts to find a treatment for Alzheimer disease have focused on discovering a way to reduce Aβ, with largely disappointing results so far. Now the spotlight seems to be shifting to a greater interest in tau as a therapeutic target. Two new research papers help strengthen the case for the importance of this microtubule binding protein in AD. In the September 16 PLoS Genetics, researchers led by Alison Goate at Washington University in St. Louis, Missouri, reported the results of a screen in which they looked for gene variants associated with cerebrospinal fluid (CSF) levels of phosphorylated tau (p-tau). They found an allele in the protein phosphatase B gene that not only associates with high CSF p-tau, but also with a faster rate of progression of AD. This finding supports the idea that tau increases the severity of the disease, and implies that treatments targeting tau might delay disease progression. Meanwhile, in September 23 Neuron, a research group led by Li Gan at the University of California in San Francisco described a mechanism that blocks the clearance of hyperphosphorylated tau and may contribute to the accumulation of tau in neurofibrillary tangles. Tau can be acetylated, the authors report, and acetylation prevents the protein degradation machinery from chewing up tau and eliminating it from the cell. Modifying tau acetylation, therefore, could be another approach for reducing tau levels.
The Washington University group wanted to find additional genes that contribute to AD. Traditional gene hunting methods compare the genetics of cases and controls, but misdiagnosis and overlapping pathologies can lead to a lot of noise in the data. An alternative method is to look for genes that associate with an endophenotype, or an inherited phenotype linked to the disease. Endophenotypes have been used to find genes for conditions such as osteoporosis and heart disease, but they have not yet been widely employed in AD studies. The advantage of using endophenotypes, said first author Carlos Cruchaga, is that the grouping of samples is more objective and quantitative, and the method therefore provides more statistical power than case-control studies. In the September Journal of Alzheimer’s Disease, the authors report a validation of sorts for this approach. First author John Kauwe correlated levels of CSF Aβ and tau with known AD-associated genetic polymorphisms, hoping to glean clues about how the genes contribute to the disease. Kauwe and colleagues found a marginal association of the calcium homeostasis modulator 1 (CALHM1) gene with CSF Aβ42 levels, suggesting that the endophenotype screening approach might prove useful.
In the PLoS Genetics paper, Cruchaga and colleagues took this idea a step further, selecting for analysis 355 single nucleotide polymorphisms (SNPs) in 34 candidate genes believed to have roles in tau metabolism. Cruchaga and colleagues then looked for a connection between specific SNPs and the level of CSF tau phosphorylated at threonine 181 (p-tau181). For initial screening, the authors used 353 CSF samples from the WashU Alzheimer’s Disease Research Center, then they retested SNPs with significant associations in an independent series of 493 CSF samples from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the University of Washington, Seattle. Only one SNP, located in the regulatory subunit of the protein phosphatase B gene (also known as calcineurin B), was significant in both populations. People with either one or two copies of the less common allele had higher levels of CSF p-tau181 than those homozygous for the common allele of this SNP.
Cruchaga and colleagues followed this up by looking for a connection between this SNP and disease parameters such as risk, age of onset, and rate of progression. They found that people who carried the allele associated with higher CSF p-tau181 levels had a sixfold faster cognitive decline (as measured by the change in the clinical dementia rating per year) than those homozygous for the common allele. In brains with Aβ pathology, but not in normal brains, the harmful SNP was also associated with lower levels of protein phosphatase B mRNA, and with more neurofibrillary tangles.
Moving on from the candidate gene approach, the authors are now screening the whole genome for SNPs associated with CSF levels of tau and Aβ, Cruchaga said, and they hope to publish the results before the end of the year. They will again analyze any genes they find to see if they are linked to the risk of developing AD, the age of onset, or the rate of cognitive decline. “We think we will be able to find new genetic variants that are associated with different facets of AD,” Cruchaga said.
Finding that tau levels are connected to disease progression fits well with previous research, said Henrik Zetterberg at Brigham and Women’s Hospital, Boston, Massachusetts. Not only has tau been shown to act downstream of Aβ, with several experiments showing that cognition in mouse models can be improved by intervening at the level of tau (see ARF related news story on Santacruz et al., 2005; ARF related news story on Roberson et al., 2007; and ARF related news story on Ittner et al., 2010), but other studies have seen an association between high CSF tau levels and faster cognitive decline, as well as a higher mortality rate (see Sämgård et al., 2009 and Wallin et al., 2010). “The new thing, in this paper, is that there can actually be distinct genetic traits that determine how individuals might react to amyloid pathology,” Zetterberg said. “Some people can stand amyloid pathology better than others and [this paper is] unveiling the genetics behind that.” He suggested that in the future, researchers might be able to group people with AD according to their genetic traits, and perhaps prescribe medication on the basis of genetic makeup.
Stratifying patients into subgroups will be very important for effective treatment, agreed Khalid Iqbal at the New York State Institute for Basic Research in Developmental Disabilities, Staten Island, New York. “I think it’s very clear that AD is really a whole class of diseases,” he said, which will not all be amenable to a single treatment. Iqbal also applauds a focus on tau as a therapeutic target. “I think now there’s a recognition in the field that independent of whether the amyloid cascade hypothesis is true, you have to have neurofibrillary degeneration to produce the clinical phenotype.”
Iqbal points out, however, that association studies such as this one need to be verified by experimentation. In particular, he is puzzled by the finding that protein phosphatase B associates with levels of p-tau181, as most research to date demonstrates that tau dephosphorylation, particularly at threonine 181, is instead regulated by protein phosphatase A (see, e.g., Liu et al., 2005). “The previous evidence is not in favor of protein phosphatase B to be a significant player when it comes to the hyperphosphorylation of tau,” Iqbal said. He speculated that perhaps there is some indirect effect of protein phosphatase B. Cruchaga said, however, that their group’s findings do not imply that protein phosphatase A is not important, only that they didn’t find a genetic variant in PPA that modifies tau. Cruchaga added that they are in the process of experimentally validating their PPB findings.
The finding that high p-tau levels associate with decreased calcineurin levels also appears to contradict data from Brad Hyman’s lab at Massachusetts General Hospital, Charlestown, and Chris Norris’s at the University of Kentucky in Lexington. Both showed that Aβ leads to increased activation of calcineurin, promoting downstream effects that include misshapen neurons and loss of spines (see ARF related news story and ARF news story on Wu et al., 2010). This work implicates calcineurin as one of the culprits in the pathogenesis of AD. Cruchaga says that the contradictory data might be explained in part by local effects. Aβ promotes calcium influx, which activates calcineurin in the immediate area of dendritic spines, Cruchaga said. However, his group is now examining global calcineurin protein and activity levels in human AD brains, and has preliminary data that overall calcineurin activity decreases as dementia progresses.
There are also many other sites on tau that get phosphorylated, and the second paper focused on that aspect. Aggregations of hyperphosphorylated tau are a characteristic feature of both AD and frontotemporal dementia, but it is not clear what causes their accumulation. Li Gan and colleagues at UCSF hypothesized that the problem might be faulty clearance of tau by the protein degradation system. Proteins are normally marked for degradation by the addition of ubiquitin to lysine residues, but ubiquitination can be blocked if acetyl groups are first added. First author Sang-Won Min and colleagues began by demonstrating that tau can be acetylated by the histone acetyltransferase p300 in vitro. The researchers next generated antibodies specific for acetylated tau, and used them to look for the modified protein in vivo, finding that human tau gets acetylated in transgenic mice.
Using transfected cell cultures, conditional knockout mice, and various loss of function and gain of function mutants, the authors then showed that it is the sirtuin protein SIRT1 that deacetylates tau both in vitro and in vivo. SIRT1 and tau co-immunoprecipitate, suggesting that they directly interact. In cell cultures, inhibiting the deacetylase led to slower turnover of tau, supporting the idea that acetylation prevents tau degradation. In contrast, when p300 was inhibited in cultures, acetylated tau was eliminated, and within two hours pathogenic phosphorylated tau disappeared as well, hinting at the therapeutic potential of this approach.
Finally, Min and colleagues connected these findings more directly to AD by showing that Aβ treatment of primary neurons increases the level of acetylated tau, although the mechanism is unknown. Importantly, in postmortem cortices from patients with tau pathologies, the levels of acetylated tau peaked at later stages of the disease, agreeing with the idea that tau acetylation is linked to disease in human brains. However, acetylated tau appeared before the accumulation of hyperphosphorylated tau and neurofibrillary tangles, fitting with the idea that acetylation inhibits clearance.
The results suggest that reducing tau acetylation could be an alternative way to decrease levels of phosphorylated tau, Gan said. To investigate this, the authors will cross tauopathy mouse models with SIRT1 conditional knockouts and SIRT1 overexpressors, and look for changes in disease symptoms. Gan added that they don’t think SIRT1 is the only deacetylase for tau, but it is the one they have the strongest evidence for at the moment. A possible therapeutic approach to reduce tau would be to inhibit p300, the protein that acetylates tau, but because this acetylase has numerous other roles in the body, this would likely lead to side effects, Gan said. Instead, she speculated that further research might uncover more specific, appropriate therapeutic targets in the same pathway.
Several studies have shown that SIRT1 has a role in neuronal plasticity, learning and memory (see ARF related news story on Michán et al., 2010 and ARF related news story on Gao et al., 2010), although it is not known if SIRT1’s role in memory has anything to do with its deacetylase functions. The pathways could be completely separate, Gan said, but added that altogether, the data “certainly point to the emerging role of SIRT1 in neurons.” Intriguingly, SIRT1 levels are reduced in AD brains, and this parallels tau accumulation (see Julien et al., 2009), and the sirtuin has been found to be neuroprotective in AD (see ARF related news story on Kim et al., 2007), and to suppress β amyloid production (see ARF related news story on Donmez et al., 2010).
The tau acetylation findings are novel, said Li-Huei Tsai at MIT, and the data provide “another line of evidence suggesting that SIRT1 regulates multiple components of Alzheimer disease pathology. The fact that SIRT1 deacetylase has an enzyme-substrate relationship with tau is clearly intriguing.”
Leonard Guarente at MIT concurs. Guarente co-chairs the scientific advisory board for the pharmaceutical company GlaxoSmithKline, which bought Sirtris, a biotech company that develops drugs based on sirtuins. “I think it is fascinating that SIRT1 directly affects both Aβ and tau,” he wrote to ARF. “Brain permeable SIRT1 activators may really help mitigate AD because of the multiple beneficial effects.”—Madolyn Bowman Rogers.
Cruchaga C, Kauwe JS, Mayo K, Spiegel N, Bertelsen S, Nowotny P, Shah AR, Abraham R, Hollingworth P, Harold D, Owen MM, Williams J, Lovestone S, Peskind ER, Li G, Leverenz JB, Galasko D, Alzheimer’s Disease Neuroimaging Initiative, Morris JC, Fagan AM, Holtzman DM, Goate AM. SNPs associated with cerebrospinal fluid phospho-tau levels influence rate of decline in Alzheimer’s disease. PLoS Genetics. 2010 Sep 16;6(9). Abstract
Min SW, Cho SH, Zhou Y, Schroeder S, Haroutunian V, Seeley WW, Huang EJ, Shen Y, Masliah E, Mukherjee C, Meyers D, Cole PA, Ott M, Gan L. Acetylation of tau inhibits its degradation and contributes to tauopathy. Neuron. 2010 Sep 23;67(6):953-66. Abstract
Kauwe JS, Cruchaga C, Bertelsen S, Mayo K, Latu W, Nowotny P, Hinrichs AL, Fagan AM, Holtzman DM, Alzheimer’s Disease Neuroimaging Initiative, Goate AM. Validating predicted biological effects of Alzheimer’s disease associated SNPs using CSF biomarker levels. J Alzheimers Dis. 2010 Sep;21(3):833-42. Abstract