Frontal Cortex Epigenetic Dysregulation During the Progression of Alzheimer’s Disease

Abstract

Although the frontal cortex plays an important role in cognitive function and undergoes neuronal dysfunction in Alzheimer’s disease (AD), the factors driving these cellular alterations remain unknown. Recent studies suggest that alterations in epigenetic regulation play a pivotal role in this process in AD. We evaluated frontal cortex histone deacetylase (HDAC) and sirtuin (SIRT) levels in tissue obtained from subjects with a premortem diagnosis of no-cognitive impairment (NCI), mild cognitive impairment (MCI), mild to moderate AD (mAD), and severe AD (sAD) using quantitative western blotting. Immunoblots revealed significant increases in HDAC1 and HDAC3 in MCI and mAD, followed by a decrease in sAD compared to NCI. HDAC2 levels remained stable across clinical groups. HDAC4 was significantly increased in MCI and mAD, but not in sAD compared to NCI. HDAC6 significantly increased during disease progression, while SIRT1 decreased in MCI, mAD, and sAD compared to NCI. HDAC1 levels negatively correlated with perceptual speed, while SIRT1 positively correlated with perceptual speed, episodic memory, global cognitive score, and Mini-Mental State Examination. HDAC1 positively, while SIRT1 negatively correlated with cortical neurofibrillary tangle counts. These findings suggest that dysregulation of epigenetic proteins contribute to neuronal dysfunction and cognitive decline in the early stage of AD.

Keywords

Alzheimer’s disease frontal cortex histone deacetylase epigenetics mild cognitive impairment neuropathology sirtuins

INTRODUCTION

Alzheimer’s disease (AD) is an irreversible progressive neurodegenerative disorder resulting in cognitive decline, with an enormous societal cost [1]. With the rapidly expanding elderly population in the United States of America and other countries, understanding the cellular and molecular mechanism(s) driving the onset of AD are of paramount importance. AD-related cognitive decline is manifest by deficits in working memory and executive function, memory modalities associated with the prefrontal cortex [2, 3]. The frontal cortex is a major hub of the default mode network that displays extensive amyloid pathology associated with the onset of cognitive decline early in AD [4 –6]. However, the underlying cellular and molecular mechanism(s) driving frontal cortex neuronal dysfunction remain unknown. Recent studies have linked epigenetic modifications including DNA methylation and posttranslational modifications of histone proteins [7 –10] to cognitive impairment and synaptic plasticity in the medial temporal lobe memory circuit (i.e., entorhinal cortex and hippocampus) [11 –13]. By contrast, there are virtually no epigenetic studies of frontal cortex protein levels and their relation to clinical measures of cognitive performance and neuropathology criteria during AD progression.

Recently, research exploring the potential contribution of epigenetic proteins, regulators of gene expression not associated with modifications to DNA sequence [8, 14], has been of great interest in the field of AD. Histone acetylation, a posttranslational epigenetic process, is regulated by the opposite actions of histone acetyltransferases, and histone deacetylases (HDACs), which provide or block access to cellular transcriptional machinery, resulting in the tight control of gene expression [8 , 15]. HDACs, including sirtuins (SIRTs), are a family of enzymes with deacetylase activity that have been linked to learning and memory [8 , 16–20] and AD pathogenesis [21 –25]. Interestingly, monozygotic twins with identical genetic information display discordant outcomes for AD, with decrements in DNA methylation found in neurons of the temporal cortex of the AD twin [26, 27], suggesting that alterations of the epigenome may mediate molecular pathways that contribute to AD pathogenesis.

Although HDAC protein levels have been examined in the medial temporal lobe memory circuit, which displays tau pathology early in the onset of AD [28], the results have been inconsistent across studies. For example, immunohistochemical analysis of entorhinal cortex neurons containing HDAC1 and HDAC2, nuclear enzymes associated with transcriptional complexes and modulation of chromatin plasticity [8 , 14], revealed that levels of these proteins are decreased in the entorhinal cortex [29], while others report that HDAC2 but not HDAC1 or HDAC3 was increased in CA1 hippocampal and entorhinal cortex nuclei in AD compared to non-cognitively impaired aged controls [11]. A recent study demonstrated a reduction of HDAC1 expression in the AD frontal cortex compared to age-matched controls using mass spectrometry [30]. Others have examined HDACs 4 and 6, which are predominantly cytosolic deacetylases, with histone and non-histone targets in the AD brain [22 , 31]. HDAC4 is upregulated by endoplasmic reticulum stress and accumulates in the nuclei of neurons in the frontal cortex when dephosphorylated in AD [25]. HDAC6 is involved in autophagic regulation [32, 33], mitochondrial transport [34], facilitation of tau hyperphosphorylation [22, 35], and is increased in the cortex in animal models of AD and the human condition [22]. SIRT1 protein and mRNA levels were decreased in the parietal cortex but protein levels were not altered in the hippocampus or cerebellum, and parietal cortex SIRT1 levels correlated with global cognition test scores and tau accumulation [23]. Together these findings suggest that members of the HDAC family are dysregulated in cortical regions related to cognitive impairment and neurofibrillary tangle (NFT) pathology in AD. However, to our knowledge no report has examined HDAC and SIRT protein levels in the frontal cortex obtained from cognitively and neuropathologically well characterized cases during the progression of AD, and more importantly subjects classified with mild cognitive impairment (MCI), a prodromal AD stage [36].

Thus, the aim of the current study was to evaluate alterations in HDAC and SIRT1 protein levels and their association with cognitive test scores, neuropathological criteria, and demographic variables in the frontal cortex during the progression of AD. HDAC protein levels were examined in frontal cortex samples obtained from individuals who died with an antemortem clinical diagnosis of no cognitive impairment (NCI), MCI, mild/moderate AD (mAD), and severe AD (sAD). Information from this study will inform the use of HDAC inhibitors as disease modifying therapeutics in pre-symptomatic AD.

MATERIALS AND METHODS

Subjects

Frozen frontal cortex tissue was obtained from individuals who died with an ante-mortem clinical diagnosis of NCI (n = 14), MCI (n = 13), and mild/moderate AD (n = 13) from the Rush Religious Orders study (RROS), a longitudinal clinicopathological study of aging and AD in Catholic nuns, priests, and brothers [37 –39]. Additional cases with a diagnosis of severe AD (n = 8) were obtained from the Rush Alzheimer’s Disease Center (RADC). The Human Research Committees of Rush University Medical Center approved this study and written informed consent for research and autopsy was obtained from the participants or their family/guardians.

Clinical and neuropathologic evaluations

Details of the clinical evaluation of the RROS cohort have been previously reported [37 , 39–41]. Briefly, a team of investigators led by a neurologist performed annual neurological examinations and a cognitive battery in addition to reviewing each subject’s medical history. The average time from the last clinical evaluation to death was ∼8 months. Neuropsychological testing included the Mini-Mental State Examination (MMSE), global cognitive score (GCS), a composite z-score compiled from a battery of 19 cognitive tests [39], including episodic memory z-score, semantic memory z-score, working memory z-score, perceptual speed z-score, and visuospatial ability z-score [42]. Subjects with MCI had cognitive impairment insufficient to meet criteria for dementia [43]. Among the MCI cases, 3 were amnestic and 10 were non-amnestic. A final clinical diagnosis was assigned after a team of neurologists and neuropsychologists reviewed the clinical data. Neuropathological diagnosis was based on Braak staging of NFTs [28], National Institute on Aging (NIA) Reagan criteria [44], and recommendations of the Consortium to Establish a Registry for Alzheimer’s disease (CERAD) [45]. Neuritic plaque, diffuse plaque, and NFT counts in the mid-frontal cortex were obtained from modified Bielschowsky silver stained sections (6μm) [46]. Counts for amyloid density in the mid-frontal cortex, anterior cingulate, and superior frontal gyrus were obtained from sections (20μm) stained with the anti-β-amyloid (Aβ) antibody 4G8 (1:9000, Covance, WI, USA) [47]. Cases with mixed pathologies other than AD (e.g., stroke, Parkinson’s disease, Lewy body dementia, vascular dementia, and hippocampal sclerosis) were excluded.

Antibodies

Table 1 summarizes the antibodies used, respective dilution, and specificity. Briefly, rabbit polyclonal antibodies raised against HDAC1 (1:500, Abcam, Cambridge, MA) [29 , 48–52], HDAC2 (1:500, Abcam) [50 , 53–56], HDAC3 (1:500, Abcam) [57, 58], HDAC4 (1:500, Abcam) [59, 60], HDAC6 (1:500, Cell Signaling, Danvers, MA) [61 –63], SIRT1 (1:1000, Cell Signaling) [64 –66], and a mouse monoclonal antibody for β-tubulin (1:2000, Sigma, St. Louis, MO) were used.

Table 1

Summary of Antibodies

Antigen	Description of Immunogen	Source, Host Species, Cat. #	Dilution	Published antibody specificity	Reference in text
HDAC1	Synthetic peptide conjugated to KLH derived from within residues 450 to the C-terminus of Human HDAC1.	Abcam, rabbit polyclonal, ab19845	1:500	Human tissue, WB; IHC-P	[29 , 48–52]
HDAC2	Synthetic peptide conjugated to KLH derived from within residues 450 to the C-terminus of Human HDAC2.	Abcam, rabbit polyclonal, ab16032	1:500	Human tissue; WB	[50 , 53–56]
HDAC3	Synthetic peptide corresponding to Human HDAC3 amino acids 411-428 conjugated to Keyhole Limpet Haemocyanin (KLH).	Abcam, rabbit polyclonal, ab32369	1:500	Human tissue; WB, IHC-P	[57, 58]
HDAC4	Synthetic peptide to amino acids 1-19 of Human HDAC4 conjugated to KLH with C-terminal added lysine.	Abcam, rabbit polyclonal, ab12172	1:500	Human and mouse tissue; WB	[59, 60]
HDAC6	Synthetic peptide corresponding to residues surrounding Pro681 of human HDAC6 protein.	Cell Signaling, rabbit monoclonal, D21B10	1:500	Mouse tissue; WB	[61 –63]
SIRT1	Synthetic peptide corresponding to the carboxy terminus of human SIRT1.	Cell Signaling, rabbit polyclonal, D739	1:1000	Human and mouse tissue; WB	[64 –66]

WB, western blot; IHC-P, imunohistochemistry-paraffin.

Quantitative immunoblotting

Frozen tissue samples from the frontal cortex were denatured in sodium dodecyl sulfate (SDS) loading buffer to a final concentration of 5 mg/ml. Proteins (50μg/sample) were separated by 4–20% SDS-PAGE (Lonza, Rockland, ME) and transferred to polyvinylidene fluoride membranes (Immobilon P, Millipore, Billerica, MA) electrophoretically [40, 67]. Membranes were first blocked in Tris-buffered saline (TBS)/0.05% Tween-20/5% milk for 60 min at room temperature and primary antibodies against HDAC1, HDAC2, HDAC3, HDAC4, HDAC6, and SIRT1 were added to the blocking buffers. Membranes were incubated at room temperature for 30 min; and then incubated overnight at 4°C. After washes (TBS/0.05% Tween-20), membranes were incubated for 1 h at room temperature with horseradish peroxidase-conjugated goat anti-rabbit and anti-mouse IgG secondary antibodies (1:5,000 and 1:3,000, respectively). Immunoreactivity was visualized by enhanced chemiluminescence (Pierce, Rockford, IL) on a Kodak Image Station 440CF (Perkin-Elmer, Wellesley, MA) and bands quantified with Kodak 1D. Frontal cortex protein immunoreactive signals were normalized to β-tubulin signals. Samples were analyzed in three independent experiments, and only one band was visualized at the appropriate kilodalton (kDa) weight for each antibody. Controls included elimination of primary antibodies and a non-specific IgG.

Statistical analysis

The Kruskall-Wallis test was used to detect differences between clinical groups in the protein levels, age at death, education, postmortem interval (PMI), brain weight at autopsy, time between last clinical assessment and autopsy. The Conover-Inman test was used to discern group pair-wise comparisons that were statistically significant. Chi-square analysis was used to determine significance of frequency differences for gender, APOE ɛ4 allele status, Braak stage, CERAD diagnosis, and NIA-Reagan diagnosis present between clinical groups. Spearman correlations were used to assess associations between protein levels and age at death, education, PMI, brain weight at autopsy, time between last clinical assessment and autopsy, along with cognitive variables, count data for diffuse plaques, neuritic plaques, NFTs in the frontal cortex, and amyloid load in the frontal cortex, anterior cingulate, and superior frontal gyrus. Within-group differences among the HDACs and SIRT1 were tested using the Wilcoxon signed-rank test. Within each of the clinical groups, HDAC and SIRT1 differences between low and high Braak stages were assessed using the Mann-Whitney test. For correlation analyses, false discovery rate was used to correct for multiple comparisons [68, 69]. All p-values were corrected for multiple comparisons to avoid the risk of Type I error. Statistical significance was set at p < 0.05 (two-tailed) and measurements were graphically represented using Sigma Plot 12.5 software (Systat Software, San Jose, CA). A power analysis showed that a correlation of r = 0.40 could be detected with 80% power (two-sided alpha = 0.05) using a sample size of 46. Given the sample size of n = 48 for this study, there was adequate statistical power for our studies. Statistical analyses were carried out using SYSTat 13.0 (SYSTAT, Inc., San Jose, CA). G*Power 3.1 was used for the estimate of statistical power.

RESULTS

Case demographics

Clinical, demographic, and neuropathological data for RROS individuals are presented in Tables 2 and 3. No statistically significant group differences were found for age at death (p = 0.15), education (p = 0.67), PMI (p = 0.06), brain weight at autopsy (p = 0.15), and time between last clinical assessment and death (p = 0.48) (Table 2). No significant differences were found in the frequencies of gender (p = 0.98), APOE ɛ4 carrier status (p = 0.28), Braak stage (p = 0.87), CERAD diagnosis (p = 0.50), and NIA-Reagan diagnosis (p = 0.34) (Table 2). Among the cognitive variables (Table 3), the MMSE showed significant group differences as the NCI and MCI groups had significantly higher scores than the mAD group (p < 0.001). However, the NCI and MCI groups were not significantly different. A similar pattern was noted for the perceptual speed domain (Table 3) (NCI > mAD and MCI > mAD p < 0.001, NCI versus MCI p = 0.07). For GCS and episodic memory, significant group differences were noted for all group pair-wise comparisons (p < 0.001) (Table 3). Neuropathology revealed that 69% of mAD, 85% MCI, and 79% of NCI subjects were Braak stages III-VI. Using the NIA-Reagan criteria, 50% of NCI, 46% of MCI, and 69% of mAD cases were classified as intermediate to high likelihood of AD (Table 2). The CERAD diagnosis revealed that 43% of NCI, 23% of MCI, and 54% of mAD cases were probable or definite AD.

Table 2

Clinical, demographic, and neuropathological characteristics by clinical diagnosis

	NCI	MCI ^a	mAD	p-value
Number	14	13	13	–
Gender (M/F)	6/8	6/7	6/7	0.98 ^c
ApoE ɛ4 (Carrier/Non-Carrier)	1/13	4/9	2/10 ^d	0.28 ^c
Age at Death (y)	85.07±4.34	86.63±5.18	88.35±4.75	0.15 ^b
Education (y)	17.21±4.00	17.31±3.12	18.46±4.27	0.67 ^b
PMI (h)	5.77±2.13	5.97±2.27	4.42±2.27	0.06 ^b
Brain Weight at Autopsy	1,223.00±88.80	1194.77±144.71	1132.46±108.44	0.15 ^b
Duration between last clinic visit and autopsy (y)	0.70±0.83	0.92±0.98	0.77±0.54	0.48 ^b
Braak Stage
0	0	0	0	0.87 ^c
I	2	1	2
II	1	1	2
III	6	5	2
IV	4	4	4
V	1	2	3
VI	0	0	0
CERAD Diagnosis
No AD	4	5	2	0.50 ^c
Possible AD	4	5	4
Probable AD	3	1	1
Definite AD	3	2	6
NIA Reagan Diagnosis
Not AD	0	0	0	0.34 ^c
Low Likelihood	7	7	4
Intermediate Likelihood	7	4	6
High Likelihood	0	2	3

AD, Alzheimer’s disease; ApoE, apolipoprotein E; CERAD, Consortium to Establish a Registry for Alzheimer’s Disease; PMI, Postmortem Interval; MCI, mild cognitive impairment; MMSE, Mini-Mental State Examination; NCI, no cognitive impairment; NIA, National Institute on Aging. ^an = 3 MCI cases were amnestic. ^bKruskall-Wallis test, with Conover-Inman test for multiple comparisons. ^cChi-square test. ^dOne mAD case did not have ApoE genotype information.

Table 3

Summary of neuropsychological tests by clinical diagnosis category

Neuropsychological tests	NCI	MCI	mAD	p-value^a	Groupwise Comparisons
MMSE	28.14±1.35	27.23±2.68	20.08±5.57	<0.001	NCI, MCI > mAD
GCS (z-score)	0.03±0.26	–0.44±0.38	–1.19±0.60	<0.001	NCI > MCI > mAD
Episodic Memory (z-score)	0.47±0.28	–0.25±0.46	–1.49±0.89	<0.001	NCI > MCI > mAD
Semantic Memory (z-score)	–0.18±0.77	–0.40±0.60	–0.83±0.87	0.09	–
Working Memory (z-score)	–0.13±0.48	–0.34±0.68	–0.59±0.63	0.22	–
Perceptual Speed (z-score)	–0.43±0.75	–0.98±0.61	–2.21±0.80	<0.001	NCI, MCI > mAD
Visuospatial (z-score)	–0.35±0.60	–0.85±0.69	–0.99±0.71	0.07	–

Values represent mean±SD. ^aKruskal-Wallis test, with Conover-Inman test for multiple comparisons. AD, Alzheimer’s disease; GCS, Global Cognitive Score; MMSE, Mini-Mental State Examination; MCI, mild cognitive impairment; NCI, no cognitive impairment.

The RADC sAD group had an average age at death of 78.75±5.23 years (range 71–86), 63% were female, average PMI was 5.08±1.45 h (range 2–6.4) and average brain weight at autopsy was 1,105.71±98.30 grams (range 980–1220). The average MMSE score was 2.50±3.50 (range 0–9). This group was evaluated for Braak stage and MMSE scores only. Data from all eight sAD cases were included in the boxplots and summary diagram but not in the demographic tables.

Frontal cortex epigenetic protein levels

Prompted by evidence linking dysregulated HDAC levels to AD pathogenesis, we utilized quantitative western blotting to investigate whether HDACs are altered in the frontal cortex, a key component of the default mode network [70, 71], during AD progression. We chose to evaluate HDACs from all three major classes, which display both nuclear and cytoplasmic subcellular locations. HDACs 1 and 2 are nuclear proteins involved in gene transcription, and have been implicated in dysregulation of genes involved in learning and memory [8 , 16–20]. HDAC1 levels (Table 4, Fig. 1A, C) in the NCI, MCI, and mAD groups were significantly higher than the sAD (p < 0.05, p < 0.001, p < 0.001, respectively), while HDAC2 levels were not significantly different between groups (p = 0.63; Table 4, Fig. 1B, D). HDAC3 is primarily cytoplasmic but can shuttle between the nucleus and cytoplasm. HDAC3 levels (Table 4, Fig. 2A, C) were significantly lower in the NCI than the MCI group (p = 0.02), while levels in the NCI, MCI, and mAD groups were significantly higher than the sAD group (p = 0.04, p < 0.001, p < 0.001, respectively). HDAC4 is a cytoplasmic class II deacetylase, which can be translocated to the nucleus in neurodegenerative disease conditions such as AD [25]. Cortical HDAC4 levels (Table 4, Fig. 2B, D) in the MCI cases were significantly higher than the NCI (p < 0.001) and the mAD (p = 0.04) groups, whereas no significant changes were found between sAD and the other clinical groups. We also evaluated levels of a Class IIb deacetylase, HDAC6, which interacts with numerous cytoplasmic proteins. HDAC6 levels (Table 4, Fig. 3A, C) in the sAD group were significantly higher than the NCI (p < 0.001) and the mAD (p < 0.001) groups, while MCI levels were significantly higher than the NCI group (p = 0.01). SIRT1 was the only class III deacetylase evaluated, due to its linkage to tau accumulation in AD [23]. For SIRT1, the only groupwise comparison that was not statistically significant was the comparison between the mAD and sAD groups (p = 0.18; Fig. 3D). All other comparisons were statistically significant (NCI > MCI, p = 0.03; NCI > mAD, p < 0.001; NCI > sAD, p < 0.001; MCI > mAD, p = 0.02) (Table 4, Fig. 3B), indicating that deacetylases from each class are dysregulated during the progression of AD, thus contributing to the complexity of frontal cortex epigenetic changes in AD.

Table 4

Clinical group differences for HDAC and SIRT1 proteins

	NCI (n = 14)	MCI (n = 13)	mAD (n = 13)	sAD (n = 8)	p-value ^a	Groupwise Comparisons
HDAC1	0.22±0.15 (0.08–0.58)	0.44±0.43 (0.04–1.61)	0.41±0.36 (0.08–1.09)	0.09±0.05 (0.03–0.19)	0.01	MCI > sAD; mAD > sAD
HDAC2	0.53±0.18 (0.32–1.01)	0.59±0.24 (0.28–1.05)	0.56±0.18 (0.33–0.90)	0.65±0.20 (0.33–0.94)	0.63	–
HDAC3	0.38±0.39 (0.05–1.18)	0.90±0.80 (0.06–2.68)	0.55±0.48 (0.04–1.51)	0.08±0.04 (0.02–0.14)	<0.001	NCI, MCI, mAD > sAD; NCI < MCI
HDAC4	0.12±0.03 (0.08–0.17)	0.21±0.13 (0.10–0.58)	0.13±0.06 (0.06–0.27)	0.14±0.05 (0.07–0.21)	0.03	NCI < MCI; MCI > mAD
HDAC6	0.08±0.03 (0.04–0.15)	0.12±0.05 (0.05–0.25)	0.10±0.04 (0.05–0.17)	0.14±0.02 (0.12–0.18)	<0.001	NCI < MCI; NCI, mAD < sAD
SIRT1	0.73±0.33 (0.05–1.18)	0.49±0.30 (0.02–0.77)	0.20±0.14 (0.03–0.43)	0.09±0.05 (0.01–0.18)	<0.001	NCI > MCI, mAD, sAD; MCI > mAD, sAD

Values represent mean±standard deviation (range). ^aKruskal-Wallis test, with Conover-Inman test for multiple comparisons. NCI, no cognitive impairment; MCI, mild cognitive impairment; mAD, mild/moderate Alzheimer’s disease; sAD, severe Alzheimer’s disease.

Fig.1

Representative immunoblots (A and B) and box plots (C and D) of HDAC1 and HDAC2 frontal cortex levels in cases clinically diagnosed as non-cognitively impaired (NCI), mildly cognitively impaired (MCI), mild/moderate AD (mAD), and severe AD (sAD). Immunoreactive signals obtained by densitometry were normalized to levels of beta-tubulin. A, C) HDAC1 levels were significantly decreased in sAD compared with NCI (p < 0.05), MCI (p < 0.001), and mAD (p < 0.001). B, D) HDAC2 levels were stable across clinical groups. Black dots in box-plots indicate outliers. Single asterisk indicates a p < 0.05, double asterisks indicate a p < 0.001.

Fig.2

Representative immunoblots (A, B) and box plots (C, D) of HDAC3 and HDAC4 frontal cortex levels in cases clinically diagnosed as non-cognitively impaired (NCI), mildly cognitively impaired (MCI), mild/moderate AD (mAD), and severe AD (sAD). Immunoreactive signals obtained by densitometry were normalized to levels of beta-tubulin. A, C) HDAC3 levels were significantly increased in MCI compared to NCI (p < 0.05), and significantly decreased in sAD compared to NCI (p < 0.05) and MCI (p < 0.001). B, D) HDAC4 levels were significantly increased in MCI compared to NCI (p < 0.001) and mAD (p < 0.05). Black dots in box-plots indicate outliers. Single asterisk indicates a p < 0.05, double asterisks indicate a p < 0.001.

Fig.3

Representative immunoblots (A, B) and box plots (C, D) of HDAC6 and SIRT1 frontal cortex levels in cases clinically diagnosed as non-cognitively impaired (NCI), mildly cognitively impaired (MCI), mild/moderate AD (mAD), and severe AD (sAD). Immunoreactive signals obtained by densitometry were normalized to levels of beta-tubulin. A, C) HDAC6 levels were significantly increased in MCI compared to NCI, and mAD (p < 0.05), and were significantly increased in sAD compared to NCI, and mAD (p < 0.001). B, D) SIRT1 levels were significantly higher in NCI compared to MCI (p < 0.05), mAD (p < 0.001), and sAD (p < 0.001), and in MCI compared to mAD (p < 0.05). Black dots in box-plots indicate outliers. Single asterisk indicates a p < 0.05, double asterisks indicate a p < 0.001.

We examined within-group differences between HDAC proteins across the four clinical groups to determine whether select HDACs are significantly upregulated across disease progression. Significant differences in HDACs were found in all four clinical groups (p < 0.05), with SIRT1 being highest in NCI, HDAC3 and HDAC2 the highest in MCI and mAD, and HDAC2 in sAD (Supplementary Table 1).

Association between biochemical, clinical, and neuropathological measures

Having established that frontal cortex HDAC levels are altered during disease progression, we next sought to identify their association to each other and to demographic, cognitive, and neuropathologic variables. Interprotein correlations and their association with cognitive measurements are shown in Tables 5 and 6, respectively. SIRT1 showed no correlation with any of the HDAC proteins (Table 5). Whereas HDAC1 and HDAC3 were moderately correlated (Fig. 4A) (r = 0.53, p < 0.001), as were HDAC2 and HDAC4 (Fig. 4B) (r = 0.40, p = 0.005) and HDAC4 and HDAC6 (Fig. 4C) (r = 0.50, p < 0.001) across clinical groups (Table 5). Given evidence that HDACs are involved in the aberrant expression of genes involved in learning, memory, and synaptic plasticity, we probed the association between cortical HDAC levels and cognitive measurements only in the RROS cases. After adjusting for multiple comparisons, HDAC1 and SIRT1 were the only proteins that significantly correlated with cognitive measures (Table 6). Levels of HDAC1 were negatively correlated with perceptual speed z-scores (Fig. 5A; r = –0.45, p = 0.004), whereas SIRT1 was positively correlated with perceptual speed z-scores (Fig. 5B; r = 0.61, p < 0.001). SIRT1 levels were moderately correlated with GCS (Fig. 5C; r = 0.40, p = 0.01), episodic memory scores (Fig. 5D; r = 0.36, p < 0.05) and MMSE status (Fig. 5E; r = 0.37, p = 0.02). Since SIRT1 and HDAC6 have been linked to tau accumulation and deacetylation [23 , 73], respectively, we examined the association between altered levels of HDACs and NFT counts in the frontal cortex between the NCI, MCI, and mAD groups. HDAC1 (Supplementary Figure 1A) and SIRT1 (Supplementary Figure 1B) showed weak correlations with frontal NFT counts (Table 7, r = 0.37, p = 0.02; r = –0.31, p = 0.05; respectively), none of the other HDACs correlated with NFT counts (Supplementary Figure 2A-D). Comparisons between low (0-III) and high (IV-VI) Braak stages revealed no significant differences between groups (p = 0.64) (Supplementary Table 2). No significant differences in protein levels were noted for any HDAC protein or SIRT1 in cases with low or high Braak stages (Supplementary Table 3). No significant differences were found in NCI protein levels between high and low Braak groups, indicating that in controls Braak stage also did not have a significant effect on HDAC levels (Supplementary Table 4). In addition, levels of HDAC proteins were not correlated with neuritic plaques, diffuse plaques, or neuritic plaques + diffuse plaques (Supplementary Figures 3, 4, 5A-F). The demographic variables (age at death, education, PMI, brain weight at autopsy, time between last clinical assessment and autopsy) did not correlate with any of the proteins examined.

Table 5

Summary of interprotein correlations

	HDAC1	HDAC2	HDAC3	HDAC4	HDAC6	SIRT1
HDAC1		0.22	0.53 ^b	0.09	–0.14	–0.07
HDAC2	0.22		–0.106	0.40 ^a	0.25	–0.01
HDAC3	0.53 ^b	–0.11		–0.14	–0.20	0.02
HDAC4	0.09	0.40 ^a	–0.14		0.50 ^b	0.14
HDAC6	–0.14	0.25	–0.20	0.50 ^b		0.01
SIRT1	–0.07	–0.01	0.02	0.14	0.01

^ap = 0.005, ^bp < 0.001.

Table 6

Correlations between neuropsychological tests and protein levels

	MMSE	GCS	Episodic Memory	Semantic Memory	Working Memory	Visuospatial Ability	Perceptual Speed
HDAC1	0.04	–0.27	–0.26	–0.003	0.02	–0.03	–0.45 ^a
HDAC2	0.26	–0.07	–0.13	0.11	–0.13	0.06	–0.12
HDAC3	–0.08	–0.21	–0.22	–0.04	–0.13	0.13	–0.25
HDAC4	0.09	–0.11	–0.19	0.07	–0.21	–0.11	–0.04
HDAC6	0.003	–0.09	–0.12	–0.07	0.04	0.01	–0.03
SIRT1	0.37 ^a	0.40 ^a	0.36 ^a	0.24	–0.004	0.19	0.61 ^b

^ap < 0.05, ^bp < 0.001. GCS, Global Cognitive Score; MMSE, Mini-Mental State Examination.

Fig.4

A) Frontal cortex HDAC3 levels correlated positively with HDAC1 levels (r = 0.53, p < 0.001). B) Significant positive correlations were found between HDAC4 and HDAC2 (r = 0.40, p = 0.005), and (C) between HDAC6 and HDAC4 (r = 0.50, p < 0.001). NCI, open circles; MCI, open squares; mAD, open triangles; sAD, open diamonds.

Fig.5

A) Frontal cortex levels of HDAC1 correlated negatively with perceptual speed z-score (r = –0.45, p < 0.05), while SIRT1 levels correlated positively with perceptual speed z-score (B) (r = 0.61, p < 0.001), Global Cognitive score (C) (r = 0.40, p < 0.05), episodic memory z-score (D) (r = 0.36, p < 0.05), and MMSE (E) (r = 0.37, p < 0.05). NCI, open circles; MCI, open squares; mAD, open triangles. MMSE, Mini-Mental State examination.

Table 7

Correlations between proteins and neuropathological variables

	HDAC1	HDAC2	HDAC3	HDAC4	HDAC6	SIRT1
Diffuse plaques	0.12	–0.11	–0.03	–0.05	0.02	–0.06
Neuritic plaques	0.01	–0.29	–0.01	–0.04	–0.06	–0.15
Diffuse + Neuritic plaques	0.16	–0.20	–0.02	–0.07	–0.04	–0.15
Mid-frontal NFT counts	0.37^a	–0.01	0.16	0.07	0.14	–0.31^a

^ap≤0.05, all other correlations not statistically significant.

DISCUSSION

HDACs are modulators of chromatin plasticity that interact with other epigenetic regulators to influence genomic and cellular stability. These proteins are associated with the promoter regions of memory-associated genes, including immediate early genes [74, 75], brain-derived neurotrophic factor [76, 77], and synaptic proteins [7, 11] in corticolimbic regions. Increasing evidence from animal studies suggest that perturbations to the balance of HDAC proteins and acetylation/deacetylation levels in the frontal cortex play a negative role in learning and memory [13 , 79]. Whether alterations in cortical HDAC protein levels are neuroprotective or neurotoxic during AD progression is controversial. These proteins may have different functions dependent on disease stage, cellular milieu, and environmental factors. Our data demonstrate that frontal cortex HDAC proteins are differentially dysregulated depending upon antemortem clinical classification and selectively correlate with cognitive tests and NFT pathology during the onset of AD.

Here, we found that frontal cortex levels of HDAC1 and HDAC3 were significantly increased in MCI and mAD followed by a significant decrease in sAD compared to NCI subjects. These protein levels were significantly correlated with each other across disease progression. Functionally, HDAC1 exhibits dual roles in the regulation of neuronal activity [80 –84], subserving neuroprotection and neurotoxicity based on its interaction with either SIRT1 or HDAC3 [80 , 83]. HDAC1 promotes cell survival via interaction with DNA damage response and repair pathways [85, 86], while SIRT1 deacetylates HDAC1 to promote non-homologous end joining in the face of DNA damage [81], which has been reported in the AD brain [87]. Previous reports show that the interaction between HDAC1 and HDAC3 is increased in neurons primed to die compared with healthy neurons [80]. A recent study demonstrated a reduction of HDAC1 expression in AD frontal cortex compared to age-matched controls using mass spectrometry [30]. Whether increases in the cortical levels of HDAC1 and HDAC3 during prodromal AD reflect the activation of neuronal death pathways or a concerted effort to maintain cellular homeostasis is unknown. However, we have previously proposed that an upregulation in choline acetyltransferase activity in the MCI frontal cortex is indicative of a reorganizational compensatory response [88]; therefore, it is possible that the increase in HDAC1 and HDAC3 protein levels reported here in MCI may also be an example of neuroplasticity occurring in the prodromal stage of AD. A possible downstream consequence of epigenomic plasticity may be the suppression of gene transcription. Increased cortical levels of HDAC1 were negatively correlated with perceptual speed z-scores across disease progression, indicating an association between HDAC1 and poorer performance on perceptual speed tasks. Functionally, the frontal cortex plays a critical role in perceptual decision-making [89 –91], and binds perceptual and executive control information to guide goal-driven behavior [92]. Thus, increased HDAC1 levels in the frontal cortex may affect downstream molecular pathways associated with perceptual speed domains during AD progression. In addition, levels of HDAC1 were positively correlated with frontal cortex NFT counts, indicating that both increased HDAC1 levels and NFT burden may contribute to AD neuropathogenesis; however, to our knowledge, there are no neurobiological reports directly associating HDAC1 and NFT formation.

The present study did not observe changes in HDAC2 levels in the frontal cortex. By contrast two immunohistochemical studies have demonstrated both increases and decreases in HDAC2 in the entorhinal cortex and hippocampus [11, 29]. This may reflect regional-specific changes related to specific cellular populations. Interestingly, a case report of monozygotic twins discordant for AD reported increased HDAC2 gene expression in peripheral blood cells of the AD twin [21], suggesting that global levels of HDAC2 expression may increase in AD, but depending on the region evaluated, may not be translated or posttranslationally modified to yield a functional protein.

Intracellular trafficking of HDAC4 has been implicated in neuronal survival and transcriptional repression. HDAC4 is localized in the cytoplasm under normal conditions; in response to low potassium or excitotoxic glutamate conditions, it rapidly accumulates in neuronal nuclei where it suppresses myocyte enhancer factor-2 (MEF-2) and cAMP response element-binding protein (CREB)-dependent transcription [93, 94]. HDAC4 is dephosphorylated and translocated to the nucleus in neurodegenerative diseases including AD [25], where it selectively accumulates in frontal cortex layer III nuclei [25]. Although we did not find increases in total cortical HDAC4 levels in mAD and sAD, it is possible that using other HDAC4 antibodies directed against its dephosphorylated form would reveal differences across clinical groups. By contrast, we found a significant increase in HDAC4 levels in MCI compared to NCI, which may reflect a response to an initial disease insult indicating yet another epigenetic plasticity response.

HDAC6 levels were increased in MCI and sAD compared to NCI and mAD in the frontal cortex, suggesting that similar to HDAC1 and HDAC3, HDAC6 levels may increase in response to disease onset, plateau in mild/moderate AD before an increase in severe AD. HDAC6 is involved in both the ubiquitin-proteasome system and autophagic pathways [95 –97], where it can regulate the formation and transportation of aggresomes, as well as promote the proteasome or lysosome-mediated degradation of misfolded proteins. Lysosomal-endosomal and autophagic upregulation occur in cortical, hippocampal, and basal forebrain neurons in MCI and AD even before plaque and tangle pathology [98 –100]. Perhaps increased HDAC6 levels in the frontal cortex in MCI reflect augmented protein clearance and degradation. HDAC6 has also been implicated in tubulin acetylation [22] and acetylated α-tubulin, a marker for microtubule stability, and is reduced in AD brain [22]. By contrast, HDAC6 levels have been reported to be increased in the cortex and hippocampus, which is paralleled by a decrease in acetylation of α-tubulin in AD compared to control brains [22]. Perhaps the increase in frontal cortex HDAC6 levels destabilize microtubules impairing intracellular transport [31 , 102]. A few reports have demonstrated a reduction in acetylated alpha-tubulin immunoreactivity in NFT bearing neurons in the human CA1 region of the hippocampus, suggesting that HDAC6 is also involved in tau microtubule destabilization in AD [35 , 73]. Although these findings suggest an interaction between HDAC6 and tau, we did not find a correlation between HDAC6, NFT counts, or Braak stage. We did observe a significant positive correlation between HDAC4 and HDAC6 levels, indicating that these cytoplasmic HDACs may interact at the cellular level.

In the present study, frontal cortex SIRT1 levels steadily decreased across AD stages. Decreased expression of SIRT1 in Braak stages I-II has been previously shown by Bossers et al. in the AD prefrontal cortex, suggesting that loss of SIRT1-mediated neuroprotection may be associated with early NFT formation [103]. Decreases in SIRT1 mRNA and protein have been demonstrated in the parietal cortex in AD [23], but were not significantly decreased in MCI in comparison to our data which showed a significant decrease in SIRT1 levels in MCI compared to NCI. This earlier study also found a significant positive correlation between parietal cortex SIRT1 levels and GCS in AD [23], similar to the findings reported here for the frontal cortex. We also found significant correlations between SIRT1 and perceptual speed, episodic memory z-scores, as well as MMSE, indicating that a reduction of frontal cortex SIRT1 protein levels may affect multiple cognitive domains during AD progression. SIRT1 knockout mice exhibit a significant deficit in both short and long-term hippocampal-dependent memory, while mice expressing high levels of SIRT1 in the hippocampus show normal long term potentiation and memory function, suggesting that SIRT1 may be critical for hippocampal-dependent memory tasks [31, 104]. What role cortical SIRT1 plays in episodic memory is unknown. However, it is possible that a concomitant change in hippocampal and cortical levels of SIRT1 may indicate a double hit to multiple memory circuits. Given the functional disconnection between the frontal cortex and hippocampus in AD [105], alterations in cortical levels of SIRT1 may exacerbate hippocampal dysfunction.

SIRT1 directly modulates synaptic plasticity and memory formation [106]. Although we have shown in tissue from the RROS cohort that drebrin, a postsynaptic dendritic spine marker is decreased in the AD hippocampus [107], whether decreases in SIRT1 and drebrin levels are related to alterations in synaptic plasticity has not been established. In cell culture models, SIRT1 is suggested to be protective against cellular stress and inflammation by inhibiting NF-kappaB signaling to mediate protection against microglia-dependent Aβ toxicity [108]. However, we did not find any association between SIRT1 levels and CERAD or amyloid density counts in the frontal cortex. We did observe a moderate negative correlation between SIRT1 levels and frontal cortex NFT counts, indicating as others have suggested that a reduction in SIRT1 is associated with tau accumulation in the AD cortex [23].

Interestingly, when we divided our NCI cases into high and low Braak pathological staging no difference in HDAC and SIRT1 protein levels was found between these NFT defined groups. In this regard, many studies have shown that NCI cases display both high Braak NFT and amyloid plaque pathology [39 , 109–113] but do not differ in the level of different molecular and cellular markers [114]. The concepts of brain and cognitive reserve and “super agers” have been used to explain these findings [110 , 115]. Despite these concepts, a clear biological explanation remains to be defined. Aged people with Braak stage III, IV and V display significant neuronal damage to the medial temporal lobe (MTL) memory circuit consisting of the transentorhinal and entorhinal cortex and hippocampus [4 –6], which may disconnect multimodal neocortical information from reaching the MTL. However, despite an apparent MTL disconnection and extensive amyloid plaque pathology, aged people remain cognitively intact [110, 111]. These observations suggest that NFT and amyloid lesions may not be a necessary precondition for cognitive impairment in the elderly or AD. Therefore, the biological factors that maintain cognitive stability in the face of extensive AD lesions in cognitively intact elders remain to be defined. Perhaps dysregulation of epigenetic factors plays a more pivotal role in cellular alterations during the course of AD.

A caveat of this or any other human tissue study is what role end of life agonal state has upon HDAC protein levels in the brain. The effects of agonal conditions on RNA and DNA are well documented, an average PMI of 3–4 hours is considered suitable for gene expression analysis [116]. On the other hand, studies indicate that proteins such as actin and GAPDH, often used as internal loading controls for western blot studies, remain intact even at longer PMIs (i.e., over 48 h) [116]. Our PMIs were ∼5 h, which is slightly higher than what would be considered appropriate for gene expression studies, but well within time limits for protein assay usage.

Figure 6 summarizes our findings of HDAC alterations in the frontal cortex during the clinical progression of AD. Whether HDAC profiles can be monitored as biomarkers for disease progression remains to be determined. The advent of HDAC PET tracers ([¹¹C]Martinostat) [117 –119] make it possible to image neuroepigenetic regulation in the human brain. Combining this technology with peripheral blood cell [21], and serum HDAC analyses [24] will aid in the development of a more accurate bioassay for determining the transition from NCI to MCI. Mapping epigenetic alterations in different brain regions and uncovering the molecular and cellular mechanisms related to disease pathogenesis may provide the basis for novel therapeutic platforms for the treatment of dementia and perhaps personalized medicine. In this regard, HDAC inhibitors are used for treating cancer [120 –124] and possibly AD [125 –127].

Fig.6

Summary diagram showing changes in histone deacetylase levels in the frontal cortex during the progression of Alzheimer’s disease. NCI, no cognitive impairment; MCI, mild cognitive impairment; mAD, mild/moderate Alzheimer’s disease; sAD, severe Alzheimer’s disease.

Footnotes

ACKNOWLEDGMENTS

We are indebted to the nuns, priests, and lay brothers who participated in the Rush Religious Orders Study and to the members of the Rush ADC. This study was supported by grants P01AG014449, P30AG019610, P30AG010161, R01AG043375, and R01AG042146 from the National Institute on Aging, National Institutes of Health and Barrow Neurological Institute Barrow and Beyond.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Wimo

, Guerchet

, Ali

, Wu

, Prina

, Winblad

, Jonsson

, Liu

, Prince

(2017) The worldwide costs of dementia 2015 and comparisons with 2010. Alzheimers Dement 13, 1–7.

Salat

, Kaye

, Janowsky

(2001) Selective preservation and degeneration within the prefrontal cortex in aging and Alzheimer disease. Arch Neurol 58, 1403–1408.

Levine

, Lu

, Bennett

, Horvath

(2015) Epigenetic age of the pre-frontal cortex is associated with neuritic plaques, amyloid load, and Alzheimer’s disease related cognitive functioning. AGING-US 7, 1198–1211.

Buckner

, Snyder

, Shannon

, LaRossa

, Sachs

, Fotenos

, Sheline

, Klunk

, Mathis

, Morris

, Mintun

(2005) Molecular, structural, and functional characterization of Alzheimer’s disease: Evidence for a relationship between default activity, amyloid, and memory. J Neurosci 25, 7709–7717.

Buckner

, Sepulcre

, Talukdar

, Krienen

, Liu

, Hedden

, Andrews-Hanna

, Sperling

, Johnson

(2009) Cortical hubs revealed by intrinsic functional connectivity: Mapping, assessment of stability, and relation to Alzheimer’s disease. J Neurosci 29, 1860–1873.

Simic

, Babic

, Borovecki

, Hof

(2014) Early failure of the default-mode network and the pathogenesis of Alzheimer’s disease. CNS Neurosci Therap 20, 692–698.

Kouzarides

(2007) Chromatin modifications and their function. Cell 128, 693–705.

Graff

, Tsai

(2013) Histone acetylation: Molecular mnemonics on the chromatin. Nat Rev Neurosci 14, 97–111.

Volmar

, Wahlestedt

(2015) Histone deacetylases (HDACs) and brain function. Neuroepigenetics 1, 20–27.

10.

Bredy

, Wu

, Crego

, Zellhoefer

, Sun

, Barad

(2007) Histone modifications around individual BDNF gene promoters in prefrontal cortex are associated with extinction of conditioned fear. Learn Mem 14, 268–276.

11.

Graff

, Rei

, Guan

, Wang

, Seo

, Hennig

, Nieland

, Fass

, Kao

, Kahn

, Su

, Samiei

, Joseph

, Haggarty

, Delalle

, Tsai

(2012) An epigenetic blockade of cognitive functions in the neurodegenerating brain. Nature 483, 222–226.

12.

Kosik

, Rapp

, Raz

, Small

, Sweatt

, Tsai

(2012) Mechanisms of age-related cognitive change and targets for intervention: Epigenetics. J Gerontology 67, 741–746.

13.

Bredy

, Wu

, Crego

, Zellhoefer

, Sun

, Barad

(2007) Histone modifications around individual BDNF gene promoters in prefrontal cortex are associated with extinction of conditioned fear. Learn Mem 14, 268–276.

14.

Dawson

, Kouzarides

(2012) Cancer epigenetics: From mechanism to therapy. Cell 150, 12–27.

15.

Mastroeni

, Grover

, Delvaux

, Whiteside

, Coleman

, Rogers

(2011) Epigenetic mechanisms in Alzheimer’s disease. Neurobiol Aging 32, 1161–1180.

16.

Bahari-Javan

, Maddalena

, Kerimoglu

, Wittnam

, Held

, Bahr

, Burkhardt

, Delalle

, Kugler

, Fischer

, Sananbenesi

(2012) HDAC1 regulates fear extinction in mice. J Neurosci 32, 5062–5073.

17.

Guan

, Haggarty

, Giacometti

, Dannenberg

, Joseph

, Gao

, Nieland

, Zhou

, Wang

, Mazitschek

, Bradner

, DePinho

, Jaenisch

, Tsai

(2009) HDAC2 negatively regulates memory formation and synaptic plasticity. Nature 459, 55–60.

18.

Gupta-Agarwal

, Franklin

, Deramus

, Wheelock

, Davis

, McMahon

, Lubin

(2012) G9a/GLP histone lysine dimethyltransferase complex activity in the hippocampus and the entorhinal cortex is required for gene activation and silencing during memory consolidation. J Neurosci 32, 5440–5453.

19.

Kim

, Akhtar

, Adachi

, Mahgoub

, Bassel-Duby

, Kavalali

, Olson

, Monteggia

(2012) An essential role for histone deacetylase 4 in synaptic plasticity and memory formation. J Neurosci 32, 10879–10886.

20.

McQuown

, Barrett

, Matheos

, Post

, Rogge

, Alenghat

, Mullican

, Jones

, Rusche

, Lazar

, Wood

(2011) HDAC3 is a critical negative regulator of long-term memory formation. J Neurosci 31, 764–774.

21.

D’Addario

, Candia

, Arosio

, Di Bartolomeo

, Abbate

, Case

, Candeletti

, Romualdi

, Damanti

, Maccarrone

, Bergamaschini

, Mari

(2017) Transcriptional and epigenetic phenomena in peripheral blood cells of monozygotic twins discordant for alzheimer’s disease, a case report. J Neurol Sci 372, 211–216.

22.

Ding

, Dolan

, Johnson

(2008) Histone deacetylase 6 interacts with the microtubule-associated protein tau. J Neurochem 106, 2119–2130.

23.

Julien

, Tremblay

, Emond

, Lebbadi

, Salem

, Bennett

, Calon

(2009) Sirtuin 1 reduction parallels the accumulation of tau in Alzheimer disease. J Neuropathol Exp Neurol 68, 48–58.

24.

Kumar

, Chaterjee

, Sharma

, Singh

, Gupta

, Gill

, Tripathi

, Dey

(2013) Sirtuin1: A promising serum protein marker for early detection of Alzheimer’s disease. PLoS One 8, e61560.

25.

Shen

, Chen

, Li

, Kofler

, Herrup

(2016) Neurons in vulnerable regions of the Alzheimer’s disease brain display reduced ATM signaling. eNeuro 3, pii: ENEURO.0124-15.2016.

26.

Mastroeni

, McKee

, Grover

, Rogers

, Coleman

(2009) Epigenetic differences in cortical neurons from a pair of monozygotic twins discordant for Alzheimer’s disease. PLoS One 4, e6617.

27.

Poulsen

, Esteller

, Vaag

, Fraga

(2007) The epigenetic basis of twin discordance in age-related diseases. Pediatr Res 61, 38R–42R.

28.

Braak

, Braak

(1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol 82, 239–259.

29.

Mastroeni

, Grover

, Delvaux

, Whiteside

, Coleman

, Rogers

(2010) Epigenetic changes in Alzheimer’s disease: Decrements in DNA methylation. Neurobiol Aging 31, 2025–2037.

30.

Anderson

, Chen

, Wang

, Mast

, Pikuleva

, Turko

(2015) Quantification of histone deacetylase isoforms in human frontal cortex, human retina, and mouse brain. PLoS One 10, e0126592.

31.

Gao

, Wang

, Mao

, Graff

, Guan

, Pan

, Mak

, Kim

, Su

, Tsai

(2010) A novel pathway regulates memory and plasticity via SIRT1 and miR-134. Nature 466, 1105–1109.

32.

Hild

, Pandey

, Nedelsky

, Schwartz

, McCray

, Yao

, Batlevi

, Nie

, Schuldiner

, Garza

, Ritson

, Padmanabhan

, DiProspero

, Hubbert

, Baehrecke

, Knight

, Taylor

, Berry

(2007) HDAC6 rescues neurodegeneration and provides an essential link between autophagy and the UPS. Nature 447, 860–864.

33.

Pandey

, Taylor

, Baehrecke

, Batlevi

(2007) HDAC6 at the intersection of autophagy, the ubiquitin-proteasome system, and neurodegeneration. Autophagy 3, 643–645.

34.

Chen

, Owens

, Makarenkova

, Edelman

(2010) HDAC6 regulates mitochondrial transport in hippocampal neurons. PLoS One 5, e10848.

35.

Perez

, Santa-Maria

, Gomez de Barreda

, Zhu

, Cuadros

, Cabrero

, Sanchez-Madrid

, Dawson

, Vitek

, Perry

, Smith

, Avila

(2009) Tau–an inhibitor of deacetylase HDAC6 function. J Neurochem 109, 1756–1766.

36.

Petersen

(2004) Mild cognitive impairment as a diagnostic entity. J Intern Med 256, 183–194.

37.

Bennett

, Schneider

, Bienias

, Evans

, Wilson

(2005) Mild cognitive impairment is related to Alzheimer disease pathology and cerebral infarctions. Neurology 64, 834–841.

38.

Bennett

, Schneider

, Arvanitakis

, Kelly

, Aggarwal

, Shah

, Wilson

(2006) Neuropathology of older persons without cognitive impairment from two community-based studies. Neurology 66, 1837–1844.

39.

Mufson

, Chen

, Cochran

, Beckett

, Bennett

, Kordower

(1999) Entorhinal cortex beta-amyloid load in individuals with mild cognitive impairment. Exp Neurol 158, 469–490.

40.

Mufson

, He

, Nadeem

, Perez

, Counts

, Leurgans

, Fritz

, Lah

, Ginsberg

, Wuu

, Scheff

(2012) Hippocampal proNGF signaling pathways and beta-amyloid levels in mild cognitive impairment and Alzheimer disease. J Neuropathol Exp Neurol 71, 1018–1029.

41.

Perez

, Getova

, He

, Counts

, Geula

, Desire

, Coutadeur

, Peillon

, Ginsberg

, Mufson

(2012) Rac1b increases with progressive tau pathology within cholinergic nucleus basalis neurons in Alzheimer’s disease. Am J Pathol 180, 526–540.

42.

Wilson

, Beckett

, Barnes

, Schneider

, Bach

, Evans

, Bennett

(2002) Individual differences in rates of change in cognitive abilities of older persons. Psychol Aging 17, 179–193.

43.

Bennett

, Wilson

, Schneider

, Evans

, Beckett

, Aggarwal

, Barnes

, Fox

, Bach

(2002) Natural history of mild cognitive impairment in older persons. Neurology 59, 198–205.

44.

Newell

, Hyman

, Growdon

, Hedley-Whyte

(1999) Application of the National Institute on Aging (NIA)-Reagan Institute criteria for the neuropathological diagnosis of Alzheimer disease. J Neuropathol Exp Neurol 58, 1147–1155.

45.

Mirra

(1997) The CERAD neuropathology protocol and consensus recommendations for the postmortem diagnosis of Alzheimer’s disease: A commentary. Neurobiol Aging 18, 91–S94.

46.

Schneider

, Aggarwal

, Barnes

, Boyle

, Bennett

(2009) The neuropathology of older persons with and without dementia from community versus clinic cohorts. J Alzheimers Dis 18, 691–701.

47.

Bennett

, Schneider

, Wilson

, Bienias

, Arnold

(2004) Neurofibrillary tangles mediate the association of amyloid load with clinical Alzheimer disease and level of cognitive function. Arch Neurology 61, 378–384.

48.

Bicknell

, Walker

, Klingseisen

, Stiff

, Leitch

, Kerzendorfer

, Martin

, Yeyati

, Al Sanna

, Bober

, Johnson

, Wise

, Jackson

, O’Driscoll

, Jeggo

(2011) Mutations in ORC1, encoding the largest subunit of the origin recognition complex, cause microcephalic primordial dwarfism resembling Meier-Gorlin syndrome. Nat Genet 43, 350–355.

49.

Lutz

, Fitzner

, Ahrens

, Geissler

, Makowiec

, Hopt

, Bogatyreva

, Hauschke

, Werner

, Lassmann

(2016) Histone modifiers and marks define heterogeneous groups of colorectal carcinomas and affect responses to HDAC inhibitors in vitro. Am J Cancer Res 6, 664–676.

50.

Poleshko

, Palagin

, Zhang

, Boimel

, Castagna

, Adams

, Skalka

, Katz

(2008) Identification of cellular proteins that maintain retroviral epigenetic silencing: Evidence for an antiviral response. J Virol 82, 2313–2323.

51.

Rashid

, Pilecka

, Torun

, Olchowik

, Bielinska

, Miaczynska

(2009) Endosomal adaptor proteins APPL1 and APPL2 are novel activators of beta-catenin/TCF-mediated transcription. J Biol Chem 284, 18115–18128.

52.

Slack

, Jones

, Bhatia

, Thompson

(2011) Autodeimination of protein arginine deiminase 4 alters protein-protein interactions but not activity. Biochemistry 50, 3997–4010.

53.

Burd

, Kinyamu

, Miller

, Archer

(2008) UV radiation regulates Mi-2 through protein translation and stability. J Biol Chem 283, 34976–34982.

54.

Fritah

, Col

, Boyault

, Govin

, Sadoul

, Chiocca

, Christians

, Khochbin

, Jolly

, Vourc’h

(2009) Heat-shock factor 1 controls genome-wide acetylation in heat-shocked cells. Mol Biol Cell 20, 4976–4984.

55.

Lee

, Murthy

SRK

, Cawley

, Dhanvantari

, Hewitt

, Lou

, Lau

, Ma

, Huynh

, Wesley

, Ng

, Pacak

, Poon

, Loh

(2011) An N-terminal truncated carboxypeptidase E splice isoform induces tumor growth and is a biomarker for predicting future metastasis in human cancers. J Clin Invest 121, 880.

56.

Saito

, Nagai

, Ladanyi

(2006) SYT-SSX1 and SYT-SSX2 interfere with repression of E-cadherin by snail and slug: A potential mechanism for aberrant mesenchymal to epithelial transition in human synovial sarcoma. Cancer Res 66, 6919–6927.

57.

Chuang

, Wu

, Tang

, Lai

, Hsu

, Sun

, Chang

, Yu

, Hsiao

, Huang

, Cha

(2013) The HDAC inhibitor LBH589 induces ERK-dependent prometaphase arrest in prostate cancer via HDAC6 inactivation and down-regulation. PLoS One 8, e73401.

58.

Huang

, Langdon

, Tse

, Mullen

, Um

, Faratian

, Harrison

(2016) The role of HDAC2 in chromatin remodelling and response to chemotherapy in ovarian cancer. Oncotarget 7, 4695–4711.

59.

Gupta

, Samant

, Smith

, Shroff

(2008) HDAC4 and PCAF bind to cardiac sarcomeres and play a role in regulating myofilament contractile activity. J Biol Chem 283, 10135–10146.

60.

Ronn

, Volkov

, Davegardh

, Dayeh

, Hall

, Olsson

, Nilsson

, Tornberg

, Dekker Nitert

, Eriksson

, Jones

, Groop

, Ling

(2013) A six months exercise intervention influences the genome-wide DNA methylation pattern in human adipose tissue. PLoS Genet 9, e1003572.

61.

Pernet

, Faure

, Gilquin

, Dufour-Guerin

, Khochbin

, Vourc’h

(2014) HDAC6-ubiquitin interaction controls the duration of HSF1 activation after heat shock. Mol Biol Cell 25, 4187–4194.

62.

Pujol Lopez

, Kenis

, Stettinger

, Neumeier

, de Jonge

, Steinbusch

, Zill

, van den Hove

, Myint

(2016) Effects of prenatal Poly I:C exposure on global histone deacetylase (HDAC) and DNA methyltransferase (DNMT) activity in the mouse brain. Mol Biol Rep 43, 711–717.

63.

Ratti

, Ramond

, Moncollin

, Simonet

, Milan

, Méjat

, Thomas

, Streichenberger

, Gilquin

, Matthias

, Khochbin

, Sandri

, Schaeffer

(2015) Histone deacetylase 6 is a FoxO transcription factor-dependent effector in skeletal muscle atrophy. J Biol Chem 290, 4215–4224.

64.

Bao

, Zheng

, Zhang

, Li

, Zhang

, Li

, Bai

, Zhang

, Huo

, Zhao

, Shang

, Wang

, Zhang

, Ji

(2016) Deacetylation of TFEB promotes fibrillar Abeta degradation by upregulating lysosomal biogenesis in microglia. Protein Cell 7, 417–433.

65.

van Gent

, Di Sanza

, van den Broek

, Fleskens

, Veenstra

, Stout

, Brenkman

(2014) SIRT1 mediates FOXA2 breakdown by deacetylation in a nutrient-dependent manner. PLoS One 9, e98438.

66.

Ringholm

, Bienso

, Kiilerich

, Guadalupe-Grau

, Aachmann-Andersen

, Saltin

, Plomgaard

, Lundby

, Wojtaszewski

, Calbet

, Pilegaard

(2011) Bed rest reduces metabolic protein content and abolishes exercise-induced mRNA responses in human skeletal muscle. Am J Physiol 301, 649–E658.

67.

Counts

, Nadeem

, Wuu

, Ginsberg

, Saragovi

, Mufson

(2004) Reduction of cortical TrkA but not p75(NTR) protein in early-stage Alzheimer’s disease. Ann Neurol 56, 520.

68.

Benjamini

, Hochberg

(1995) Controlling the false discovery rate: A practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol 57, 289–300.

69.

Glickman

, Rao

, Schultz

(2014) False discovery rate control is a recommended alternative to Bonferroni-type adjustments in health studies. J Clin Epidemiol 67, 850–857.

70.

Buckner

, Sepulcre

, Talukdar

, Krienen

, Liu

, Hedden

, Andrews-Hanna

, Sperling

, Johnson

(2009) Cortical hubs revealed by intrinsic functional connectivity: Mapping, assessment of stability, and relation to Alzheimer’s disease. J Neurosci 29, 1860–1873.

71.

Simic

, Babic

, Borovecki

, Hof

(2014) Early failure of the default-mode network and the pathogenesis of Alzheimer’s disease. CNS Neurosci Ther 20, 692–698.

72.

Cook

, Gendron

, Scheffel

, Carlomagno

, Dunmore

, DeTure

, Petrucelli

(2012) Loss of HDAC6, a novel CHIP substrate, alleviates abnormal tau accumulation. Hum Mol Genet 21, 2936–2945.

73.

Hempen

, Brion

(1996) Reduction of acetylated alpha-tubulin immunoreactivity in neurofibrillary tangle-bearing neurons in Alzheimer’s disease. J Neuropathol Exp Neurol 55, 964–972.

74.

French

, O’Connor

, Jones

, Davis

, Errington

, Voss

, Truchet

, Wotjak

, Stean

, Doyere

, Maroun

, Laroche

, Bliss

TVP

(2001) Subfield-specific immediate early gene expression associated with hippocampal long-term potentiation in vivo. Eur J Neurosci 13, 968–976.

75.

Guzowski

, Lyford

, Stevenson

, Houston

, McGaugh

, Worley

, Barnes

(2000) Inhibition of activity dependent arc protein expression in the rat hippocampus impairs the maintenance of long-term potentiation and the consolidation of long-term memory. J Neurosci 20, 3993–4001.

76.

Hall

, Thomas

, Everitt

(2000) Rapid and selective induction of BDNF expression in the hippocampus during contextual learning. Nat Neurosci 3, 533–535.

77.

Linnarsson

, Bjorklund

, Ernfors

(1997) Learning deficit in BDNF mutant mice. Eur J Neurosci 9, 2581–2587.

78.

Jakovcevski

, Akbarian

(2012) Epigenetic mechanisms in neurological disease. Nat Med 18, 1194–1204.

79.

Morris

, Mahgoub

, Na

, Pranav

, Monteggia

(2013) Loss of histone deacetylase 2 improves working memory and accelerates extinction learning. J Neurosci 33, 6401–6411.

80.

Bardai

, Price

, Zaayman

, Wang

, D’Mello

(2012) Histone deacetylase-1 (HDAC1) is a molecular switch between neuronal survival and death. J Biol Chem 287, 35444–35453.

81.

Dobbin

, Madabhushi

, Pan

, Chen

, Kim

, Gao

, Ahanonu

, Pao

, Qiu

, Zhao

, Tsai

(2013) SIRT1 collaborates with ATM and HDAC1 to maintain genomic stability in neurons. Nat Neurosci 16, 1008–1015.

82.

Jeong

, Then

, Melia

, Mazzulli

, Cui

, Savas

, Voisine

, Paganetti

, Tanese

, Hart

, Yamamoto

, Krainc

(2009) Acetylation targets mutant huntingtin to autophagosomes for degradation. Cell 137, 60–72.

83.

Kim

, Frank

, Dobbin

, Tsunemoto

, Tu

, Peng

, Guan

, Lee

, Moy

, Giusti

, Broodie

, Mazitschek

, Delalle

, Haggarty

, Neve

, Lu

, Tsai

(2008) Deregulation of HDAC1 by p25/Cdk5 in neurotoxicity. Neuron 60, 803–817.

84.

Reynolds

, Casaccia

, Dietz

, Shen

, Howell

, Kim

, He

(2010) HDAC1 nuclear export induced by pathological conditions is essential for the onset of axonal damage. Nat Neurosci 13, 180–189.

85.

Miller

, Tjeertes

, Coates

, Legube

, Polo

, Britton

, Jackson

(2010) Human HDAC1 and HDAC2 function in the DNA-damage response to promote DNA nonhomologous end-joining. Nat Struct Mol Biol 17, 1144–1151.

86.

Wang

, Pan

, Su

, Quinn

, Sasaki

, Jimenez

, Mackenzie

, Huang

, Tsai

(2013) Interaction of FUS and HDAC1 regulates DNA damage response and repair in neurons. Nat Neurosci 16, 1383–1391.

87.

Adamec

, Vonsattel

, Nixon

(1999) DNA strand breaks in Alzheimer’s disease. Brain Res 849, 67–77.

88.

DeKosky

, Ikonomovic

, Styren

, Beckett

, Wisniewski

, Bennett

, Cochran

, Kordower

, Mufson

(2002) Upregulation of choline acetyltransferase activity in hippocampus and frontal cortex of elderly subjects with mild cognitive impairment. Ann Neurol 51, 145–155.

89.

Rahnev

, Lau

, de Lange

(2011) Prior expectation modulates the interaction between sensory and prefrontal regions in the human brain. J Neurosci 31, 10741–10748.

90.

Rahnev

, Nee

, Riddle

, Larson

, D’Esposito

(2016) Causal evidence for frontal cortex organization for perceptual decision making. Proc Nat Acad Sci U S A 113, 6059–6064.

91.

Fleming

, Huijgen

, Dolan

(2012) Prefrontal contributions to metacognition in perceptual decision making. J Neurosci 32, 6117–6125.

92.

Opris

, Santos

, Gerhardt

, Song

, Berger

, Hampson

(2013) Prefrontal cortical microcircuits bind perception to executive control. Sci Rep 3, 2285.

93.

Bolger

, Yao

(2005) Intracellular trafficking of histone deacetylase 4 regulates neuronal cell death. J Neurosci 25, 9544–9553.

94.

Herrup

, Li

, Chen

(2013) The role of ATM and DNA damage in neurons: Upstream and downstream connections. DNA Repair 12, 600–604.

95.

Boyault

, Gilquin

, Zhang

, Rybin

, Garman

, Meyer-Klaucke

, Matthias

, Muller

, Khochbin

(2006) HDAC6-p97/VCP controlled polyubiquitin chain turnover. EMBO 25, 3357–3366.

96.

Hildmann

, Riester

, Schwienhorst

(2007) Histone deacetylases–an important class of cellular regulators with a variety of functions. Appl Microbiol Biotechnol 75, 487–497.

97.

Kopito

(2000) Aggresomes, inclusion bodies and protein aggregation. Trends Cell Biol 10, 524–530.

98.

Nixon

, Yang

(2011) Autophagy failure in Alzheimer’s disease—locating the primary defect. Neurobiol Dis 43, 38–45.

99.

Orr

, Oddo

(2013) The role of autophagy in Alzheimer’s disease. Alzheimers Dement 9, P189.

100.

Perez

, He

, Nadeem

, Wuu

, Ginsberg

, Ikonomovic

, Mufson

(2015) Hippocampal endosomal, lysosomal, and autophagic dysregulation in mild cognitive impairment: Correlation with abeta and tau pathology. J Neuropathol Exp Neurol 74, 345–358.

101.

Chen

, Zhou

, Mueller-Steiner

, Chen

, Kwon

, Yi

, Mucke

, Gan

(2005) SIRT1 protects against microglia-dependent amyloid-beta toxicity through inhibiting NF-kappaB signaling. J Biol Chem 280, 40364.

102.

Dompierre

, Godin

, Charrin

, Cordelieres

, King

, Humbert

, Saudou

(2007) Histone deacetylase 6 inhibition compensates for the transport deficit in Huntington’s disease by increasing tubulin acetylation. J Neurosci 27, 3571–3583.

103.

Bossers

, Wirz

KTS

, Meerhoff

, Essing

AHW

, Van Dongen

, Houba

, Kruse

, Verhaagen

, Swaab

(2010) Concerted changes in transcripts in the prefrontal cortex precede neuropathology in Alzheimer’s disease. Brain 133, 3699–3723.

104.

Michán

, Li

, Chou

, Parrella

, Long

, Allard

, Lewis

, Miller

, Xu

, Mervis

, Chen

, Guerin

, Smith

LEH

, Mcburney

, Sinclair

, Baudry

, Cabo

, Longo

(2010) SIRT1 is essential for normal cognitive function and synaptic plasticity. J Neurosci 30, 9695–9707.

105.

Grady

, Furey

, Pietrini

, Horwitz

, Rapoport

(2001) Altered brain functional connectivity and impaired short-term memory in Alzheimer’s disease. Brain 124, 739–756.

106.

Kim

, Pan

, Gao

, Wang

, Tsai

, Su

, Guan

, Mao

, Gräff

, Mak

(2010) A novel pathway regulates memory and plasticity via SIRT1 and miR-134. Nature 466, 1105–1109.

107.

Counts

, He

, Nadeem

, Wuu

, Scheff

, Mufson

(2012) Hippocampal drebrin loss in mild cognitive impairment. Neurodegener Dis 10, 216–219.

108.

Chen

, Zhou

, Mueller-Steiner

, Chen

, Kwon

, Yi

, Mucke

, Gan

(2005) SIRT1 protects against microglia-dependent amyloid-beta toxicity through inhibiting NF-kappaB signaling. J Biol Chem 280, 40364.

109.

Erten-Lyons

, Woltjer

, Dodge

, Nixon

, Vorobik

, Calvert

, Leahy

, Montine

, Kaye

(2009) Factors associated with resistance to dementia despite high Alzheimer disease pathology. Neurology 72, 354–360.

110.

Mufson

, Malek-Ahmadi

, Perez

, Chen

(2016) Braak staging, plaque pathology, and APOE status in elderly persons without cognitive impairment. Neurobiol Aging 37, 147–153.

111.

Mufson

, Malek-Ahmadi

, Snyder

, Ausdemore

, Chen

, Perez

(2016b) Braak stage and trajectory of cognitive decline in noncognitively impaired elders. Neurobiol Aging 43, 101–110.

112.

Price

, Morris

(1999) Tangles and plaques in nondemented aging and “preclinical” Alzheimer’s disease. Ann Neurol 45, 358–368.

113.

Wegiel

, Flory

, Kuchna

, Nowicki

, Yong Ma

, Wegiel

, Badmaev

, Silverman

, de Leon

, Reisberg

, Wisniewski

(2017) Multiregional age-associated reduction of brain neuronal reserve without association with neurofibrillary degeneration or β-amyloidosis. J Neuropathol Exp Neurol 76, 439–457.

114.

Perez

, Nadeem

, He

, Miguel

, Malek-Ahmadi

, Chen

, Mufson

(2017) Neocortical and hippocampal TREM2 protein levels during the progression of Alzheimer’s disease. Neurobiol Aging 54, 133–143.

115.

Arnold

, Louneva

, Cao

, Wang

, Han

, Wolk

, Negash

, Leurgans

, Schneider

, Buchman

, Wilson

, Bennett

(2013) Cellular, synaptic, and biochemical features of resilient cognition in Alzheimer’s disease. Neurobiol Aging 34, 157–168.

116.

Blair

, Wang

, Hernandez

, Siedlak

, Rodgers

, Achar

, Fahmy

, Torres

, Petersen

, Zhu

, Casadesus

, Lee

(2016) Individual case analysis of postmortem interval time on brain tissue preservation. PLoS One 11, e0151615.

117.

Schroeder

, Wang

, Van de Bittner

, Neelamegam

, Takakura

, Karunakaran

, Wey

, Reis

, Gale

, Zhang

, Holson

, Haggarty

, Hooker

(2014) PET imaging demonstrates histone deacetylase target engagement and clarifies brain penetrance of known and novel small molecule inhibitors in rat. ACS Chem Neurosci 5, 1055–1062.

118.

Wang

, Schroeder

, Wey

, Borra

, Wagner

, Reis

, Kim

, Holson

, Haggarty

, Hooker

(2014) In vivo imaging of histone deacetylases (HDACs) in the central nervous system and major peripheral organs. J Med Chem 57, 7999–8009.

119.

Wey

, Gilbert

, Zürcher

, She

, Bhanot

, Taillon

, Schroeder

, Wang

, Haggarty

, Hooker

(2016) Insights into neuroepigenetics through human histone deacetylase PET imaging. Sci Transl Med 8, 351ra106.

120.

Arrighetti

, Corno

, Gatti

(2015) Drug combinations with HDAC inhibitors in antitumor therapy. Crit Rev Oncog 20, 83–117.

121.

Behera

, Jayprakash

, Sinha

(2015) Histone deacetylase inhibitors: A review on class-I specific inhibition. Mini Rev Med Chem 15, 731–750.

122.

Ceccacci

, Minucci

(2016) Inhibition of histone deacetylases in cancer therapy: Lessons from leukaemia. Br J Cancer 114, 605–611.

123.

Gallinari

, Di Marco

, Jones

, Pallaoro

, Steinkuhler

(2007) HDACs, histone deacetylation and gene transcription: From molecular biology to cancer therapeutics. Cell Res 17, 195–211.

124.

Marks

, Xu

(2009) Histone deacetylase inhibitors: Potential in cancer therapy. J Cell Biochem 107, 600–608.

125.

Cacabelos

, Torrellas

(2015) Epigenetics of aging and Alzheimer’s disease: Implications for pharmacogenomics and drug response. Int J Mol Sci 16, 30483–30543.

126.

Delcuve

, Khan

, Davie

(2012) Roles of histone deacetylases in epigenetic regulation: Emerging paradigms from studies with inhibitors. Clin Epigenetics 4, 5.

127.

, Dai

, Huang

, Jiang

(2011) Targeting HDACs: A promising therapy for Alzheimer’s disease. Oxid Med Cell Longev 2011, 143269.