This is Part 1 of a two-part series. See also Part 2.
10 December 2008. In discussions of what goes awry during neurodegeneration, the words genome stability, cell cycle, and DNA repair often come out in the same breath. Now, two Boston-based research teams have linked these activities to the same chromatin-modifying proteins—histone deacetylases (HDACs). Writing in tomorrow’s issue of Neuron, Li-Huei Tsai, Massachusetts Institute of Technology, and colleagues have identified HDAC1 deregulation as a common mechanism behind cell cycle re-entry and double-strand DNA breaks in an Alzheimer disease mouse model. Studying SIRT1, the mammalian homolog of the yeast HDAC Sir2, researchers led by David Sinclair at Harvard Medical School report in the 28 November issue of Cell that SIRT1 normally represses cell cycle-related and other mouse genes but redistributes to DNA breaks to promote repair. This, in turn, unleashes age-related transcriptional changes. The two studies highlight proper gene regulation as a key aspect of neuronal health. They underscore the potential for HDAC-targeted therapeutics, several of which showed preclinical promise at the Society for Neuroscience annual meeting held 15-19 November in Washington, DC (see ARF companion story).
When all is well in the brain, neurons are terminally differentiated cells whose replication machinery has remained dormant for decades. However, researchers have long observed that some neurons enter a mitotic frenzy that presumably leads to their early demise in AD (Vincent et al., 1996; Yang et al., 2003; and
ARF related news story). Even so, considerable uncertainty has lingered as to whether the errant cell cycle activity causes or results from neuronal degeneration. The Neuron paper lays to rest this chicken-and-egg controversy, Tsai suggested in an interview with ARF. Cell cycle dysfunction “actually happens before you see any signs of neurodegeneration,” she said. “Our study really suggests that this is a very early event and likely contributes to neurodegeneration as opposed to just being a terminal event.”
To start off, first author Dohoon Kim and colleagues used microarrays to probe transcriptional changes that precede neurodegeneration in Tsai’s AD mice, which also develop tau pathology and learning defects upon induction of forebrain-specific p25 (Cruz et al., 2003). Among the 225 genes that were differentially expressed in induced transgenics versus uninduced controls, an unusually high number (65) encoded proteins involved in cell cycle (e.g. PCNA, Ki-67, cyclins A, B, and E) or DNA repair (e.g., Rad51, BRCA1, Chk1). The vast majority (63 of 65) were upregulated upon p25 induction. In subsequent RT-PCR and immunofluorescence studies of p25 mouse brain cells, the researchers validated the microarray findings by showing that p25 induction in fact activates abnormal cell cycle activity and produces DNA breaks.
More hints that these ominous features might share a molecular trigger came when the team found markers for DNA damage (phosphoserine 129 histone H2AX [γH2AX]) and cell cycle progression (Ki-67) co-expressed in more than 90 percent of examined neurons. HDAC1, already implicated in transcriptional repression of various cell cycle genes, came to the fore as a possible common culprit for these problems. Confirming their hunch, the researchers demonstrated a p25/HDAC1 interaction in induced p25 mouse forebrain tissue and in 293T cells transfected with either p25 or p35 (the physiological protein from which pathological p25 is cleaved). They found that HDAC1 prefers p25 to p35, interacting with the former to a 12-fold higher extent than the latter in 293T cells.
This curious observation, combined with mapping studies showing that p25 binding occurs in an N-terminal region of HDAC1’s catalytic domain, suggested that interaction with p25 could interfere with HDAC1’s normal activities. Indeed, both p25-transfected 293T cells and induced p25 hippocampal cells had lower HDAC1 activity than their untransfected or uninduced controls. The researchers also found that inhibiting HDAC1 increased double-strand DNA breaks and cell death, and that overexpression of HDAC1 relieved these problems in primary cortical neurons, compared with untreated cells. Trying to make a case for in vivo significance, they showed that striatal HDAC1 injections given to rats one day before an ischemic insult reduced DNA damage and numbers of dying cells by more than a third relative to animals treated with catalytically dead HDAC1.
Interestingly, in the immunofluorescence studies of p25 transgenic neurons, the DNA damage signal that showed up after two weeks of p25 induction disappeared with four subsequent weeks of p25 suppression. This suggests that “the early stage when we start to see DNA lesions is still a reversible window of time,” Tsai told ARF. “If you can do something during this window, then probably neurodegeneration can be prevented.”
SIRT1 Resembles Yeast Cousin Sir2 as Genome Stabilizer
For its part, Sinclair’s work paints a similar picture of HDACs as factors that relocalize to DNA lesions to initiate repair and, in doing so, stray from their typical gene regulatory functions. However, his team arrived at these conclusions by entirely different means. Their study’s point of departure was the longstanding observation that Sir2 proteins in yeast function primarily by stabilizing the genome. Interested in epigenetic effects during aging, first author Philipp Oberdoerffer wondered, for the mammalian Sir2 homolog SIRT1 (a class III HDAC), “Is this function conserved, and does it impact proper organismal function?”
Initial clues that aging-related stressors might disrupt the transcriptional silencing function of SIRT1 came from experiments in mouse embryonic stem (ES) cells, where SIRT1 associates with and represses highly repetitive DNA. When the researchers subjected the cells to oxidative stress (H2O2), the amount of SIRT1 bound to satellite repeats dropped, and transcription at these loci went up—an effect reproduced by the pan-sirtuin inhibitor nicotinamide and counteracted by expression of a SIRT1 transgene.
To extend these findings to protein-coding genes, Oberdoerffer and colleagues performed “ChIP on chip” (chromatin immunoprecipitation combined with a genome-wide promoter array) to identify SIRT1 target genes. In untreated ES cells, SIRT1 associated with promoters of a variety of genes, most prominently those involved in chromatin assembly, transcriptional repression, ubiquitin-regulated protein degradation, and cell cycle regulation. However, in H2O2-treated cells, SIRT1 associated with less than 10 percent of these genes, and its binding pattern no longer fit the observed functional groups, suggesting that oxidative stress shifted SIRT1 to random sites across the genome.
Taking after its yeast homolog, SIRT1 relocalizes from silent genes to DNA lesions in response to stressors, the researchers found. In fact, SIRT1 physically associates with double-strand break sites and helps bring the double-strand break repair crew where it needs to go, as recruitment of Rad51 (a key double-strand break repair protein) to DNA lesions was down in SIRT1-deficient ES cells. Knockdown of SIRT1 or pharmacological inhibition of its activity reduced both pathways of double-strand break repair—homologous recombination-mediated repair and, to a lesser extent, non-homologous end-joining. Consistent with its key role in DNA repair, SIRT1 seems to be important for maintaining overall genomic stability. When the researchers subjected wild-type and SIRT1-deficient ES cells to H2O2, and analyzed them for chromatid breaks, chromosomal fusions, translocations, and other chromosomal abnormalities, they found many more aberrations in the SIRT1-deficient cells.
To put these findings into an in-vivo context, the scientists turned to p53+/- mice. Missing one copy of this tumor suppressor gene, these animals frequently develop cancer when exposed to ionizing radiation because the irradiation-induced DNA damage disrupts their remaining p53 allele. Hence, survival and cancer rates provided a nice readout for testing SIRT1 manipulations. The researchers found that, as expected, boosting SIRT1 expression and/or activity increased survival rates and reduced cancer frequency in irradiated p53+/- animals. In normal mice, two-thirds of the SIRT1-bound genes whose repression lifted with aging (comparing five-month-old and 30-month-old mice) also were re-expressed in H2O2-treated ES cells. Furthermore, in animals engineered to overexpress SIRT1 in the brain in an inducible fashion, switching on SIRT1 delayed these age-related transcriptional changes.
Together, the in-vivo findings encouraged the scientists because they indicated that having more SIRT1 around doesn’t detract from either of its functions. “You maintain your genomic integrity and at the same time you have extra SIRT1 to take care of the DNA damage,” Oberdoerffer told ARF. “Based on our paper, you would say that in general, in a DNA damage-driven aging environment, it might be a good thing to have more SIRT1.” SIRT1’s benefits extend to neurodegenerative disease, too. Last year, Tsai, Sinclair, and colleagues published a study showing that SIRT1 wards off neurodegeneration in the inducible p25 transgenic mouse and in cell-based models for AD and amyotrophic lateral sclerosis (Kim et al., 2007).
Oberdoerffer is keen on seeing whether other chromatin modifiers such as methylases and other HDACs exhibit similar protective effects as SIRT1. This question could be hard to tackle, as isoform-specific differences likely abound. Whereas Oberdoerffer’s paper suggests that a specific sirtuin guards against aging-driven transcriptional changes, a recent report from Frank LaFerla’s group at the University of California, Irvine, showed that a pan-sirtuin inhibitor (nicotinamide) unexpectedly boosted cognition and reduced tau pathology in AD mice (see ARF related news story).
Still, hopes remain high that deacetylases and other agents of epigenetic change will prove their worth as therapeutic targets for aging and neurodegenerative disease. “If you silence a gene, you can at least in theory always reverse that,” Oberdoerffer told ARF. “If you have a good understanding of the epigenomic picture, then you might understand a little better the aging process, and where and how to interfere with it.” For a glimpse at several promising HDAC-targeted drug candidates, read Part 2 of this series.—Esther Landhuis.
This is Part 1 of a two-part series. See also Part 2.
Oberdoerffer P, Michan S, McVay M, Mostoslavsky R, Vann J, Park SK, Hartlerode A, Stegmuller J, Hafner A, Loerch P, Wright SM, Mills KD, Bonni A, Yankner BA, Scully R, Prolla TA, Alt FW, Sinclair DA. SIRT1 Redistribution on Chromatin Promotes Genomic Stability but Alters Gene Expression during Aging. Cell. 2008 Nov 28;135(5):907-18. Abstract
Kim D, Frank CL, Dobbin MM, Tsunemoto RK, Tu W, Peng PL, Guan J-S, Lee B-H, Moy LY, Giusti P, Broodie N, Mazitschek R, Delalle I, Haggarty SJ, Neve RL, Lu Y, Tsai LH. Deregulation of HDAC1 by p25/Cdk5 in Neurotoxicity. Neuron. 2008 Dec 11;60:1-15. Abstract