Addition of N-acetylglucosamine (O-GlcNAc) to protein hydroxyl groups is a post-translational modification made to thousands of proteins in mammals. In many cases, addition of the sugar occurs at phosphorylation sites, and the two processes are thought to compete with one another (for a review, see Hart et al., 2011). In the brains of people with AD, hyperphosphorylated tau is mirrored by a decline in O-GlcNAc-modified tau (see Liu et al., 2009). One reason for the reduced glycosylation, which depends on adequate nutrient levels, could be alterations in glucose metabolism (see Schubert, 2005). Vocadlo and colleagues decided to test whether increasing O-GlcNAc modification of proteins would protect neurons and guard against neurofibrillary tangle (NFT) formation in a mouse model of tau neurodegeneration.

Co-first authors Scott Yuzwa and Xiaoyang Shan laced the drinking water of nine- to 12-week-old hemizygous JNPL3 mice with a compound developed in Vocadlo's lab. Thiamet-G inhibits the glycoside hydrolase (O-GlcNAcase) that removes N-acetylglucosamine from proteins. The transgenic mice carry the gene for human tau with the P301L mutation and develop neurodegeneration in the brain. Twenty mice imbibed the spiked water for 36 weeks, while 20 control mice lapped up untreated water.

JNPL3 mice show dramatic motor neuron loss in their spinal cords—ending up with about half the number that wild-type mice have—so the researchers first checked there for neuron rescue. At the end of the 36 weeks, thiamet-G-treated mice retained 1.4-fold more motor neurons in their cervical spinal cords than did untreated mice. The treated mice also maintained more of their body weight, losing less skeletal muscle than is common in this model.

What led to the protection? The AT8 antibody, which recognizes phosphorylated tau, detected 23 to 62 percent less phospho-tau immunoreactivity around neurofibrillary tangle-like structures in mouse brainstem, spinal cord, hypothalamus, and cortices. At the same time, O-GlcNAc jumped four- to fivefold in the brainstems of treated mice compared to controls. It seemed the neurons benefited from fewer NFTs, presumably through reduced phosphorylation, as previous research suggests. However, again using AT8 antibody, the researchers found no change in the global tau phosphorylation levels (including soluble forms) in homogenized JNPL3 brain tissue compared to control samples. That result led the team to hypothesize that added sugars reduced tau's clumping into tangles without blocking its phosphorylation sites. "One of O-GlcNAc's major functions might be to serve as a bumper to keep proteins from bumping into each other and aggregating," said Gerald Hart, Johns Hopkins University School of Medicine, Baltimore, Maryland, told ARF. Hart was not involved in the study.

To further test whether the added sugar reduced aggregation without hindering phosphorylation, the team did in-vitro tests on a truncated, fast-aggregating tau snippet. Glycosylating the snippet slowed that peptide's aggregation even in the absence of a phosphorylation effect. "We were surprised," Vocadlo told ARF. "It shows that the sugar modification on its own can influence the aggregation of tau, at least in vitro, in the absence of other factors, including phosphorylation." Vocadlo's team showed in a previous study that acutely treating wild-type rats with thiamet-G for a single day did reduce tau phosphorylation, while boosting the addition of O-GlcNAc to tau (see ARF related news story on Yuzwa et al., 2008). However, cells may adapt to treatment with chronic exposure, said Vocadlo, which could explain why continuous treatment of the JNPL3 mice is not accompanied by a reduction in tau phosphorylation. Since in this study, O-GlcNAc also hindered the heat-induced accumulation of an unrelated protein, the sTAB1 binding protein, the authors suggested that a general role of O-GlcNAc may be to prevent protein aggregation.

If glycosylation counteracts aggregation, what promotes it? Here, acetylation might come in, bespeaking the complexity of tau biology (see ARF related news story; also Min et al., 2010). In their new paper, the Trojanowski group extended earlier work showing that acetylation led to both a loss of one of tau’s major functions, i.e., promoting microtubule assembly, and a gain of toxic function, i.e., pathological tau aggregation. First author David Irwin and colleagues examined, postmortem, brains of people who had AD, corticobasal degeneration, or progressive supranuclear palsy, though no cognitively normal controls. They report that tau acetylated at lysine 280 appeared in similar structures as hyperphosphorylated tau in all cases. While acetylated tau appeared at all stages of AD, it was most prevalent in middle and late stages of the disease. Hence, it appears to occur mostly after tau phosphorylation, which may actually open the door for the acetylase to modify the protein, suggest the authors. They point out that most phosphorylation sites in tau flank the microtubule-binding repeat where lysine 280 sits. The UPenn authors suggest that acetylation may exacerbate loss of normal tau function and foster fibrillization, offering a new therapeutic target for AD and other tauopathies. If glycosylation modified those same phosphorylation sites, as has been proposed, it might also protect tau against subsequent acetylation. "We are interested in the potential interaction between different modifications on tau, but have not looked at the effects of acetylation on tau aggregation," Vocadlo told ARF in an e-mail.

Cheng-Xin Gong, New York State Institute for Basic Research, Staten Island, is not completely convinced of Vocadlo and colleagues’ finding that O-GlcNAc has no effect on tau phosphorylation. Since the researchers did not examine all of tau’s phosphorylation sites in their tests, other sites could have been modified, he pointed out. While further study is warranted on that point, "it is possible this drug may work through different pathways to benefit people with AD," he said, referencing a study suggesting that O-GlcNAc addition steers APP processing away from the amyloidogenic pathway (see Jacobsen and Iverfeldt, 2011). Recent research also suggests that O-GlcNAc plays a direct role in learning and memory, Hart pointed out (see Rexach et al,. 2012 and Kaleem et al., 2011). The Canadian team did not characterize cognition in these mice because they don't show the cognitive defects that other models do, Vocadlo said. As for motor function, treated mice fared no better than untreated. In the cage hang test of motor skills and coordination, untreated mice declined faster than treated mice, but the treated mice hung on for shorter times overall, possibly because they weigh more, so results are inconclusive, the authors note. In the rotarod test of strength and balance, both groups performed equally. The scientists plan to search for other inhibitors of O-GlcNAc addition, and test thiamet-G in other animal models to further probe its effects, Vocadlo said.

Proteins related to DNA transcription, protein translation, trafficking and degradation, cancer, and the cytoskeleton are all glycosylated and regulated by O-GlcNAc. The sugar's far-reaching influence prompts concerns about side effects when tilting the balance of glycosylation. For example, excess O-GlcNAc protein modification may worsen insulin resistance (see Vosseller et al., 2002). The team saw no negative effects on weight, food consumption, or motor neuron counts in wild-type mice treated with thiamet-G over 22 weeks. "This is hard to believe," given O-GlcNAc's range of effects, said Gong. He was unsure why that might be, but said that perhaps glycosylation less dramatically contorts protein conformation than phosphorylation, or that the body somehow buffers the compound's effects. "What is important here is they have shown that, in vivo, thiamet-G can reduce tau pathology," said Khalid Iqbal, New York Institute for Basic Research in Developmental Disabilities, Staten Island. Before the compound can be considered for human clinical trials, researchers will need to look for cognitive benefits in other animal models of tau pathology, especially ones that more closely mimic the disease mechanisms of sporadic AD, he said. The team will also need to do further work to elucidate the mechanism of thiamet-G, especially whether it has effects on phosphorylation, researchers agreed.—Gwyneth Dickey Zakaib.

References:
Yuzwa S, Shan X, Macauley MS, Clark T, Skorobogatko Y, Vosseller K, Vocadlo DJ. Increasing O-GlcNAc slows neurodegeneration and stabilizes tau against aggregation. Nature Chem Bio 2012 February 26; Advance Online Publication. Abstract

Irwin DJ, Cohen TJ, Grossman M, Arnold SE, Xie SX, Lee VM-Y, Trojanowski JQ. Acetylated tau, a novel pathological signature in Alzheimer's disease and other tauopathies. Brain. 2012 March;135 (3): 807-818. Abstract