The finding could lead to the development of a new generation of HDAC inhibitors. Researchers have put various candidates through preclinical paces as potential therapeutics for AD and other disorders (see Part 1 and Part 2 of related ARF news), and at least one company, EnVivo Pharmaceuticals in Watertown, Massachusetts, has an HDAC inhibitor in Phase 1 testing for Alzheimer’s. But most of those inhibitors are relatively non-selective, targeting multiple deacetylases. As recent experience in cancer treatment suggests, that can lead to unwelcome toxic effects (see Bruserud et al., 2007), which could be problematic when treating a memory disorder such as AD. “Cancer treatment is usually short term, which is a very different situation from chronic neurodegeneration, where the benchmark for safety is much higher,” said Tsai. Her work opens the door for the development of more specific HDAC2 inhibitors.

Tsai and colleagues used a combination of biochemistry and genetics to pinpoint the role of HDAC2 in memory. Joint first authors Ji-Song Guan, Stephen Haggarty, Emanuela Giacometti, and Jan-Hermen Dannenberg found that suberoylanilide hydroxamic acid (SAHA), a general inhibitor of class I HDACs and also of class II HDAC6, improved contextual fear memory in mice. The HDAC6 inhibitor WT-161 had no effect on memory in this paradigm, however, indicating class I HDACs mediate the SAHA effect. Using a proteome approach, the researchers found that SAHA targeted HDAC1 and HDAC2 and then chose to examine those two proteins in detail.

The researchers made both HDAC-overexpressing and HDAC-negative mice. HDAC2 overexpressors performed poorly in the contextual fear test, while HDAC1 overexpressor mice seemed normal. The same pattern surfaced from the Morris water maze test of spatial memory. HDAC2-overexpressing mice took significantly longer to find the hidden platform, and in probe trials with no platform, they showed no preference for the correct quadrant of the maze. They also performed poorly in a T-maze memory test. In contrast, HDAC2 knockout mice, which seemed perfectly viable up to one year of age, had enhanced mnemonic power. They froze almost twice as often as wild-type animals in response to a fear stimulus, and their spatial working memory was significantly better, as well. “The results were a little surprising,” said Tsai, who had earlier suggested that class II HDACs might be involved in memory, especially HDAC5, since it is involved in downregulation of brain-derived neurotrophic factor (see Tsankova et al., 2006).

How might HDAC2 specifically affect memory? To answer this, the authors first looked at the density of dendritic spines. In HDAC2-overexpressing mice, the CA1 region of the hippocampus contained significantly fewer spines than in controls, while in HDAC2-negative mice the opposite was true. Similarly, in hippocampal slices, synaptic strengthening in the form of long-term potentiation was reduced or enhanced in HDAC2-overexpressing and HDAC2-negative mice, respectively. All told, the results indicated that HDAC2 suppresses spine formation and learning and memory.

Since HDAC2 modifies the histones that package DNA, the researchers next looked to see if the deacetylase has a special appetite for gene promoters that regulate learning and memory. Using chromatin immunoprecipitation experiments, Guan and colleagues report that HDAC2 was associated with, amongst others, the promoter region of brain-derived neurotrophic factor (BDNF), Creb, and Creb-binding protein genes, all of which are implicated in memory formation. The specificity might be explained by a preference for CoREST, a co-repressor that recruits silencing machinery to segments of DNA in neurons. The researchers found that HDAC2, but not HDAC1, binds to this co-repressor.

Given that it is one of 18 deacetylases, HDAC2 might not be the only one involved in memory, emphasized Tsai, but it may be a major one. The researchers found that SAHA, which improves memory in wild-type mice and in HDAC2-overexpressing mice, had no effect on HDAC2-KO animals. “These results strongly suggest that HDAC2 is the major, if not the only, target of SAHA in eliciting memory enhancement,” write the authors. Whether that’s because the inhibitor preferentially blocks HDAC2 is not yet clear.

What regulates HDAC2 itself is also not well understood. This is something Tsai said she wants to examine in more detail. Recent work from Antonella Riccio’s lab at University College London, UK, suggests that the protein is modified by S-nitrosylation (see Nott et al., 2008) and that this leads to gene activation. The S-nitrosylation is induced by BDNF, which is elevated in the brain in response to exercise (see ARF related news story) and reduced in the CSF of people with dementia (see Li et al., 2009). It could be that the way to blocking HDAC2 is not only with a chemical inhibitor, but also by getting some more exercise—maybe by coiling more garden hoses.—Tom Fagan.

Reference:
Guan J-S, Haggarty SJ, Giacometti E, Dannenberg J-H, Joseph N, Gao J, Nieland TJF, Zhou Y, Wang X, Mazitschek R, Bradner JE, DePinho RA, Jaenisch R, Tsai L-H. Nature 2009, May 7; 459: 55-63. Abstract