Tsai and colleagues discovered the link between chromatin and memory by studying their p25 mouse model of neurodegeneration. The p25 protein is a highly stable fragment of the larger, unstable p35. p25 constitutively activates Cdk5, a kinase that has been implicated in various AD pathologies, including phosphorylation of tau (see ARF related news story) and production of Aβ (see ARF related news story).

Previously, Tsai’s group showed that mice expressing an inducible p25 transgene have learning and memory problems that are accompanied by both synaptic and neuronal loss (see Fischer et al., 2005 and ARF related news story). “The difference between ours and previous studies is that our mice have very profound neurodegeneration and neuronal loss, so we wanted to address whether even after this had taken place if it is still possible to improve cognitive function,” said Tsai in an interview.

To see if the effects of that neurodegeneration can be overcome, first author Andre Fischer and colleagues tested 11-month-old animals in an environmental enrichment paradigm. In environmental enrichment (EE), mice live in bigger cages with toys and more opportunity for exercise and social interaction than standard housing offers. EE has been shown to improve learning and memory, induce neurogenesis, and even attenuate the toxicity of amyloid-β in transgenic mouse models of AD (see ARF related news story).

First, the researchers induced p25 for 6 weeks. Next, they placed some of the mice in their standard cages, and some in an EE setting. Four weeks later, Fischer measured performance in various tests of learning and memory, including contextual fear conditioning and spatial learning. He found that the mice in EE performed like mice that had no induction of p25, and that both controls and EE mice performed significantly better than p25-induced mice living in standard cages. The finding suggests that during this month of EE, the animals somehow compensated for the toxicity of the p25 expression.

While this finding shows that EE can help mice overcome deficits wrought by p25, it does not suggest that EE can help mice retrieve memories that have apparently been lost. That particular finding comes from a more sophisticated experiment probing long-term memory. “Because the p25 gene is inducible, it allows us to turn on neurodegeneration at will and to train the animals when they are still normal and healthy,” said Tsai. Fischer and colleagues trained 11-month-old mice in a contextual fear-conditioning paradigm and then left the animals for 4 weeks so the memory could become consolidated as long-term memory. Then the researchers induced p25 for 6 weeks, split the mice up into normal and EE cages, and then tested their response to the fear stimulus another 4 weeks later. The animals that had been housed in normal cages completely forgot the fear context, freezing only about 5 percent of the time when tested, but animals that had been in the EE cages performed as well as controls, freezing on average about 37 percent of the time. Spatial memory in water maze testing gave similar results. “The fact that long-term memories can be recovered by EE supports the idea that the apparent ‘memory loss’ is really a reflection of inaccessible memories,” write the authors.

How might EE help restore these memories? EE can lead to increased neurogenesis (see ARF related news story). In this study, Fischer and colleagues examined other potential explanations, including reinvigoration of neural networks. Mice in EE cages proved to have normal levels of the synaptic marker synaptophysin, whereas this marker dropped by nearly half in p25-induced mice held in normal cages, indicating a potential loss of neural connectivity. “These data suggest that EE leads to the recovery of long-term memories by re-establishing the synaptic network,” write the authors.

What signaling pathways may be activated to support synaptophysin and synaptic activity? Enter chromatin. The researchers found that EE induced methylation and acetylation of histone proteins, which form the core and glue that holds chromatin together. These post-translational modifications weaken the hold of histone on DNA, allowing access to transcription factors. Fischer and colleagues found that EE can induce acetylation and methylation of histones H3 and H4 in as little as 3 hours. Histone acetylation has been linked to improved learning and memory (see ARF related news story). In support of this idea, the researchers also found that preventing deacetylation of histones mimics the effect of EE. In a repeat of the EE experiments, Fischer and colleagues found that sodium butyrate, a histone deacetylase (HDAC) inhibitor, enhanced learning in p25-induced mice and recovered freezing behavior in the long-term memory experiment. The finding suggests that HDAC inhibitors might be useful in treating memory disorders (they have been shown to slow neurodegeneration caused by polyglutamine expanded proteins; see ARF related news story). The authors caution that “…the effect of HDAC inhibitors on learning and memory could be a combination of modifications on chromatin and non-histone proteins,” (see Yuan et al., 2005).

The study by Shi and colleagues extends the idea that histone modification can alter neural networks. Joint first authors Mamta Tahiliani and Pinchao Mei and colleagues report that JARID1C/SMCX is an H3 lysine demethylase and acts as a transcriptional repressor. Mutations in SMCX have been implicated in Fragile X-linked mental retardation and epilepsy. The authors found that SMCX associates with REST, a transcriptional silencer that binds to many neuronal gene promoters, including sodium channel type 2A (SCN2A). When the authors knocked down SMCX by using RNA interference, they found that several genes, including SCN2A, were de-repressed. Taken together with Tsai’s work, this paper reinforces that chromatin regulation is complex. In one case, relieving chromatin constraints can lead to improved learning and memory, while in another it may lead to retardation.

Tsai emphasized that this study was undertaken in mice. Whether it translates to humans in general, and to Alzheimer disease in particular, remains to be seen. “That EE and histone deacetylase inhibition can facilitate recovery even after memory loss suggests that the memories are not completely erased but rather become inaccessible. This is very promising. As long as the memory is still there, there may be something we can do to bring it back,” suggested Tsai.

In AD there is initially a loss of synaptic activity followed by more extensive neuronal loss. “There is very dramatic regeneration of neuronal processes in our model. This supports previous computational models predicting that if you rewire the brain, you can recover memory,” suggested Tsai.—Tom Fagan.

References:
Fischer A, Sananbenesi F, Wang X, Dobbin M, Tsai L-H. Recovery of learning and memory is associated with chromatin remodeling. Nature. 2007, April 29. Advanced online publication. Abstract

Tahiliani M, Mei P, Fang R, Leonor T, Rutenberg M, Shimizu F, Li J, Rao A, Shi Y. The histone H3K4 demethylase SMCX links REST target genes to X-linked mental retardation. Nature. 2007, April 30. Advanced online publication. Abstract