Pan-HDAC Inhibitors Promote Tau Aggregation by Increasing the Level of Acetylated Tau

Epigenetic remodeling via histone acetylation has become a popular therapeutic strategy to treat Alzheimer’s disease (AD). In particular, histone deacetylase (HDAC) inhibitors including M344 and SAHA have been elucidated to be new drug candidates for AD, improving cognitive abilities impaired in AD mouse models. Although emerged as a promising target for AD, most of the HDAC inhibitors are poorly selective and could cause unwanted side effects. Here we show that tau is one of the cytosolic substrates of HDAC and the treatment of HDAC inhibitors such as Scriptaid, M344, BML281, and SAHA could increase the level of acetylated tau, resulting in the activation of tau pathology.


Introduction
Alzheimer's disease (AD) is a chronic neurodegenerative disorder that characterized by extracellular deposits of amyloid plaques and neuronal deposits of tau aggregates composed of hyperphosphorylated tau [1,2]. Over the last few decades, a number of compounds, designed to reduce the formation of amyloid plaques or to enhance their clearance, have failed in clinical trials. Since the causes of AD are still unknown, diverse targets are being applied for anti-Alzheimer's drug discovery. Among the diverse, epigenetic regulation has been proposed to be a new promising therapeutic strategy for neurological disorders, particularly for AD [3]. In aged animal models, decreased levels of histone acetylation have been observed in the hippocampus and cerebral cortex [4,5]. Such changes could contribute to the development of neurodegeneration by down-regulating genes, which are critical for learning and memory. A number of recent studies have showed that histone deacetylase (HDAC) inhibitors exhibit neuro-protective properties, rescuing learning and memory abilities impaired in AD animal models [4,[6][7][8][9][10][11]. Accordingly, HDAC inhibition has emerged as an alternative therapeutic strategy in AD treatment.
Histone deacetylases are divided into four classes. Class I, II, and IV contain the classic HDAC enzymes, and Class III contains the sirtuin enzymes, which require NAD+ as a cofactor [12,13]. In 2009, Francis et al. proposed that epigenetic alteration by HDAC inhibition could be a therapeutic target to prevent AD progression [7]. In their study, trichostatin A, a HDAC inhibitor, rescued fear memory
the Tau-BiFC Null group, BML210 and PhenylbutyrateNa slightly increased histone acetylation by showing 2.5-and 2.3-fold increases. The results indicate that Scriptaid, M344, BML281, and SAHA are pan-HDAC inhibitors, which strongly inhibit both cytoplasmic and nuclear HDACs. As a cytosolic substrate of HDACs, tau was also strongly acetylated by pan-HDAC inhibitors. Similar to the increased level of acetylated tubulin, Tau(K280) acetylation increased almost 3-fold by the treatment of the pan-HDAC inhibitors. Different from acetylated tubulin, acetylated tau seems accumulated in the cells, increasing the amount of total tau.   Quantification of phosphorylated, acetylated, and total tau in total cell lysates (D) and total tau in GFP-trap fractions (E). The relative amounts of phosphorylated, acetylated and total tau were quantified by Image J. Data represent the mean ± S.D. of replicate experiments. * p < 0.05. ** p < 0.01, *** p < 0.001.
In the Tau-BiFC Null group, BML210 and PhenylbutyrateNa slightly increased histone acetylation by showing 2.5-and 2.3-fold increases. The results indicate that Scriptaid, M344, BML281, and SAHA are pan-HDAC inhibitors, which strongly inhibit both cytoplasmic and nuclear HDACs. As a cytosolic substrate of HDACs, tau was also strongly acetylated by pan-HDAC inhibitors. Similar to the increased level of acetylated tubulin, Tau(K280) acetylation increased almost 3-fold by the treatment of the pan-HDAC inhibitors. Different from acetylated tubulin, acetylated tau seems accumulated in the cells, increasing the amount of total tau.
For immunoblot analysis, tau-BiFC cells were treated with each compound for 36 h and cell lysates were prepared. S199 and S396 are the representative phosphorylation sites of tau phosphorylated by GSK-3β [44]. While the levels of phospho-Ser396 (pS396) did not change, the levels of phospho-Ser199 (pS199) increased slightly upon the treatment of the selected HDAC inhibitors, by showing 1.7~1.9-fold increases ( Figure 2B). Compare to the phosphorylation levels, tau acetylation (Ac-K280) increased more strikingly by showing 2.7~3.2-fold increases upon the treatment of Scriptaid, M344, BML281, or SAHA. Total tau levels were also elevated over 2.7-up to 3.1-fold. (Figure 2B,D). The elevated levels of total tau closely matched with the levels of acetylated tau than that of phosphorylated tau. To evaluated the level of tau aggregation, matured tau-BiFC complexes in each cell lysates were enriched by using GFP-trap ® ( Figure 2C) [45]. GFP-trap ® captures only paired BiFC complexes, not N-or C-terminal fragments of Venus. GFP-trap ® -captured tau-BiFC complexes directly indicate the level of tau aggregation from soluble dimers to insoluble aggregates. Upon the treatment of Scriptaid, M344, BML281, or SAHA, GFP-trap ® -captured tau-BiFC complexes were significantly increased by showing 4.0~4.9-fold increases ( Figure 2C,E). Our results correspond to the previous studies, suggesting that tau acetylation slows tau turnover by inhibiting proteasomal degradation and leads to the accumulation of tau [20,33,34]. Especially, tau acetylation at K280/K281 is known to be critical in fibrillar tau aggregation [19,46]. Taken these together, our results suggest that tau acetylation increased by pan-HDAC inhibitors could lead pathological tau accumulation and aggregation.

Increase of Tau Acetylation by the Inhibition of Cytoplasmic HDAC6
Scriptaid, M344, BML281, and SAHA are pan-HDAC inhibitors that are known to inhibit both nuclear and cytoplasmic HDACs [6,[47][48][49][50] (Table 1). Depending on the subcellular localization, HDAC enzymes can be divided into three classes [51]. Class I (HDAC 1, 2, 3, and 8) are primarily located in nucleus. Class IIa (HDAC 4, 5, 7, and 9) and class IV (HDAC 11) shuttle between nucleus and cytoplasm. Class IIb (HDAC 6 and 10) are primarily found in cytosol ( Figure 3A). To investigate HDAC subclasses that affect tau acetylation, we prepared siRNAs for HDAC3, HDAC5, and HDAC6, representing each subclass. Tau-BiFC cells were transfected with siRNAs against HDAC3, HDAC5, or HDAC6. After 72 h, mRNA expression of each of HDACs was evaluated ( Figure 3B). By siRNA transfection, mRNA expression reduced 90% in HDAC3, 68% in HDAC5, and 61% in HDAC6, compared to that of scrambled control (siControl). As expected, HDAC6 knockdown led to 1.8-fold increase in tau-BiFC fluorescence, while HDAC3 knock-down did not affect tau-BiFC response. (Figure 3C,D). HDAC5 knock-down increased tau-BiFC response slightly, but the increase was not significant. Immunoblot analysis was followed to investigate tau acetylation levels upon HDAC6 knock-down ( Figure 3E). Comparable to the reduced mRNA levels, HDAC6 protein expression decreased by 60% by siRNA transfection. In result, acetylated α-tubulin increased 1.9-fold and acetylated tau increased 2-fold. Similar to the treatment of the pan-HDAC inhibitors, HDAC6 knockdown increased total tau, but not total tubulin ( Figure 3E,F). Although the role of HDAC6 in tau pathology is still controversial [20,52,53], our result shows that tau is constantly deacetylated by cytoplasmic HDACs. Therefore, inhibition or knock-down of cytoplasmic HDACs increases tau acetylation, leading to pathological tau accumulation.

Discussion
Histone acetylation play a critical role in memory formation and synaptic elasticities in hippocampus, and alterations of histone acetylation were observed in AD mouse models and AD patients [4,6,7,[54][55][56]. Accordingly, histone deacetylases (HDACs) emerged as a new therapeutic target for AD. Accumulating studies also showed the therapeutic potential of HDAC inhibitors (M344, SAHA, and Trichostatin A), of which administration rescued learning and memory abilities impaired in an AD mouse model [4,6,7]. However, HDACs acetylate not only histone, but also a variety of non-histone proteins in nuclei and cytosols [17,18]. Since most HDAC inhibitors are poorly selective, the treatment would cause unwanted side effects. Here, we show that tau is one of the cytosolic substrates of HDACs and the treatment of pan-HDAC inhibitors (Scriptaid, M344, BML281, and SAHA) significantly increase tau acetylation and aggregation.
For screening of HDACs, SIRTs, and HATs modulators, tau-BiFC cells were plated on µ-clear 384-well plates. The next day, the cells were treated with 34 library compounds at 3 µM concentration. After 48 h, nuclei were counterstained with 2 µg/mL Hoechst (Invitrogen, Waltham, MA, USA). The entire 384-well plate was automatically imaged by using Operetta ® (PerkinElmer, Waltham, MA, USA). High resolution images were acquired by using a Nikon Eclipse inverted microscope (Ti, Nikon, Tokyo, Japan) at 200× magnification. SCREEN-WELL ® Epigenetics library was purchased from Enzo Life Sciences Inc. (Farmingdale, NY, USA). Forskolin (Sigma, St. Louis, MO, USA) was used as a positive control.

BiFC-Image Analysis
BiFC fluorescence images were acquired using Operetta and analyzed using Harmony 3.1 software (PerkinElmer, Waltham, MA, USA). All experiments were performed in triplicate. The means and standard deviations (S.D.) of tau-BiFC fluorescence intensities were plotted by using Prism7 software (GraphPad, San Diego, CA, USA). Quantification data was analyzed by Student's t-test.

siRNA Transfection and Analysis
For siRNA transfection, HDAC3, HDAC5, and HDAC6 siRNAs were purchased from OriGene Technologies, Inc. (Rockville, MD, USA). Tau-BiFC cells were plated on µ-clear 96-well plate or 12-well culture plate in Opti-MEM (Invitrogen, Waltham, MA, USA) and transfected with each of siRNAs by using Lipofectamine 2000 (Invitrogen). Scrambled siRNA was used as a control. Twelve hours after transfection, the medium was replaced with fresh growth medium (DMEM containing 10% FBS with antibiotics). Seventy-two hours of transfection, siRNA-transfected tau-BiFC cells were used for BiFC-image analysis and immunoblot analysis.