Suberoylanilide hydroxamic acid suppresses axonal damage and neurological dysfunction after subarachnoid hemorrhage via the HDAC1/HSP70/TDP-43 axis

Increased focus has been placed on the role of histone deacetylase inhibitors as crucial players in subarachnoid hemorrhage (SAH) progression. Therefore, this study was designed to expand the understanding of SAH by exploring the downstream mechanism of the histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA) in SAH. The expression of TDP-43 in patients with SAH and rat models of SAH was measured. Then, western blot analysis, immunofluorescence staining, and transmission electron microscope were used to investigate the in vitro effect of TDP-43 on a neuronal cell model of SAH established by oxyhemoglobin treatment. Immunofluorescence staining and coimmunoprecipitation assays were conducted to explore the relationship among histone deacetylase 1 (HDAC1), heat shock protein 70 (HSP70), and TDP-43. Furthermore, the in vivo effect of HDAC1 on SAH was investigated in rat models of SAH established by endovascular perforation. High expression of TDP-43 in the cerebrospinal fluid of patients with SAH and brain tissues of rat models of SAH was observed, and TDP-43 accumulation in the cytoplasm and the formation of inclusion bodies were responsible for axonal damage, abnormal nuclear membrane morphology, and apoptosis in neurons. TDP-43 degradation was promoted by the HDAC1 inhibitor SAHA via the acetylation of HSP70, alleviating SAH, and this effect was verified in vivo in rat models. In conclusion, SAHA relieved axonal damage and neurological dysfunction after SAH via the HSP70 acetylation-induced degradation of TDP-43, highlighting a novel therapeutic target for SAH.


INTRODUCTION
Subarachnoid hemorrhage (SAH) is a lethal and devastating intracranial hemorrhage that is frequently misdiagnosed in the initial stage and increases the risk of mortality and disability 1,2 . SAH, which is mainly induced by aneurysmal rupture and perimesencephalic bleeding, is a dangerous disease that is commonly characterized by acute headache, nausea/vomiting and neck pain, with a mortality rate within 24 h of 25% and an overall mortality rate of 50% 3 . Accumulating evidence has suggested that SAHinduced early brain injury is a key factor that affects the prognosis of SAH patients 4 . Although great achievements have been made in understanding the pathogenesis and pathophysiology of SAH, as well as in the treatment and management of SAH, its underlying molecular, cellular, and circulatory dynamics remain largely unclear, and its prognosis is still not satisfactory 5,6 . Thus, indepth exploration of SAH is needed to improve management.
Suberoylanilide hydroxamic acid (SAHA), also called vorinostat, belongs to the hydroxamic acid class of histone deacetylase (HDAC) inhibitors 7,8 . It was shown that SAHA exerted great clinical utility as a therapeutic intervention after intracerebral hemorrhage 9 . Moreover, HDACs are potent enzymes that posttranslationally modify both histone and nonhistone acetylation sites, thereby affecting many cellular processes, such as the cell cycle and apoptosis 10 . Notably, the exacerbating effects of HDAC4 on SAH can be inhibited by SAHA 11 . HDAC1 inactivation was shown to protect against neuronal death and brain injury 12 . Moreover, it was identified that heat shock protein 70 (HSP70) was a cytosolic substrate of HDAC5 and could be hyperacetylated by HDAC5, which was further associated with the proliferation of hypoxic tumor cells 13 . HSP70 is a well-known therapeutic target for SAH 14 . In addition, HSP70 was shown to inhibit the cytoplasmic accumulation of TDP-43, which is responsible for the formation of insoluble inclusion bodies and is a hallmark of neurodegenerative diseases 15,16 . Moreover, TDP-43 was identified as a prognostic biomarker for SAH 17 . Herein, SAHA is hypothesized to be involved in the development of SAH by interacting with the HDAC1/HSP70/TDP-43 axis.

MATERIALS AND METHODS Ethics statement
The current study was approved by the IRB of The Third Xiangya Hospital of Central South University and was performed in accordance with the Helsinki Declaration. Written informed consent was obtained from each participant. Animal experiments were approved by the Animal Ethics Committee of The Third Xiangya Hospital of Central South University and Central South University and were conducted according to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health.

Bioinformatics analysis
Aneurysmal SAH-related microarray data (GSE54083) were downloaded from the Gene Expression Omnibus (GEO) database. Because the occurrence of SAH was induced by rupture of an intracranial aneurysm, eight cases of intracranial aneurysm rupture samples and ten normal control samples in GSE54083 were selected, and the differential analysis was conducted using the "limma" package of R language. Differentially expressed genes were identified with a threshold of P < 0.05.

Clinical samples
A total of 15 patients with aneurysmal SAH and 10 patients without SAH (non-SAH patients) from The Third Xiangya Hospital of Central South University were selected for cerebrospinal fluid collection. Cerebrospinal fluid (5 mL) was collected every 6 h for 8 days and then immediately centrifuged in a 10 mL sterile centrifuge tube at 12,000 × g for 15 min at 4°C. The supernatant of the cerebrospinal fluid, which was free from blood cell components, was collected, subpackaged in 1.5 mL centrifuge tubes in equal amounts, and stored at −80°C for subsequent experiments. Control cerebrospinal fluid and blood samples were collected from age-matched patients who underwent elective surgery that was not associated with the central nervous system.

Enzyme-linked immunosorbent assay (ELISA)
ELISA was performed to measure the levels of TDP-43 in cerebrospinal fluid based on the instructions of the kit (R&D Systems, Minneapolis, MN, MAB77782, https://www.rndsystems.com/cn).

Construction of rat models of SAH
A Wistar rat endovascular puncture model of SAH was constructed according to previous research methods 18,19 . A similar operation was performed on control rats, except that the suture was inserted less than 8 mm to avoid ACA perforation. Body temperature was monitored during the operation. During the operation and within 2 h after the operation, heating plates were used to keep the animal's body temperature at 36.5-37.5°C. After being anesthetized with 3% sodium pentobarbital, SAH rats were fixed in a stereotaxic apparatus (RWD Life Science, Shenzhen, Guangdong, China), and 15 μL (μg/μL) of lentivirus [short hairpin RNA-negative control (sh-NC), sh-histone deacetylase 1 (sh-HDAC1), overexpression (oe)-NC, and oe-TDP-43] were injected into the rat lateral ventricle (1.0 mm after the bregma, 2.0 mm on the right side, 3.5 mm in-depth) within 10 min at a constant rate (5 μL/min). The needle was maintained for 10 min to prevent blood reflux and then slowly pulled out, after which the scalp was sutured. The success rate of the model was~85% (68/80, 12 died). Rats were subjected to sham operation or SAH modeling. After successful modeling, the rats were treated with oe-NC, sh-NC, oe-TDP-43, sh-HDAC1, dimethyl sulfoxide (DMSO, 5% in normal saline, injected every day), or SAHA (50 mg/kg, injected every day) or without any injection (n = 6 each). No rats died during viral infection or drug treatment.

Isolation and culture of primary neurons and the construction of cell models of SAH in vitro
Primary neurons were isolated from the brain tissues of fetal rats (embryos aged 16-18 d). The meninges and blood vessels were removed, and the brain tissues were digested with 0.25% trypsin (containing ethylenediaminetetraacetic acid (EDTA)) at 37°C for 5 min. Then, the tissues were washed three times with phosphate-buffered saline (PBS) to stop the trypsinization. Next, the cells were resuspended on neural basal medium supplemented with 2% B27, 2 mM L-glutamine, 50 U/mL penicillin, and 50 U/mL streptomycin (GIBCO BRL, Grand Island, NY, USA). Half of the medium was renewed every 2 days. An in vitro cell model of SAH was established by stimulating neurons with oxyhemoglobin (OxyHb). The neurons were incubated with OxyHb (20 μM) at 37°C with 5% CO 2 for 6 h. Next, the medium was removed, and the neurons were washed with PBS (three times) for subsequent experiments.

Cell culture and infection
Human embryonic kidney (293T) cells (American Type Culture Collection) were cultured in Dulbecco's Modified Eagles Medium (Gibco) with 10% (v/v) fetal bovine serum. Primary neurons were cultured in neural basal medium (Gibco) in a cell incubator at 37°C with 5% CO 2 .
293T cells were cotransfected with the packaging virus and the target vector with LV5-green fluorescent protein (GFP) (lentiviral gene overexpression vector) and pSIH (lentiviral gene silencing vector) for 48 h, followed by supernatant collection. After filtration and centrifugation, viral particles in the supernatant were obtained, and the viral titer was determined. Viruses in the logarithmic growth phase were collected. Cells were treated with OxyHb (treated with 20 μM OxyHb for 24 h), OxyHb + sh-NC (treated with 20 μM OxyHb for 24 h after being transfected with sh-NC for 48 h), OxyHb + sh-HDAC1 (treated with 20 μM OxyHb for 24 h after being transfected with sh-HDAC1 for 48 h), OxyHb + sh-TDP-43 (treated with 20 μM OxyHb for 24 h after being transfected with sh-HDAC1 for 48 h), OxyHb + DMSO (treated with 20 μM OxyHb for 24 h after being transfected with 5% DMSO for 48 h), and OxyHb + SAHA (treated with 20 μM OxyHb for 24 h after being transfected with SAHA for 48 h). sh-NC, sh-HDAC1, oe-NC, and oe-TDP-43 were all synthesized by GenePharma Co., Ltd. (Shanghai, China). Cells were prepared into a 4 × 10 5 cells/mL cell suspension, seeded in a six-well plate in 2 mL per well, and cultured overnight.

Garcia behavioral score
Neurological status was evaluated 72 h after SAH modeling. The modified Garcia test included vibrissae touch, trunk touch, spontaneous movement, spontaneous movement of limbs, forelimb extension, and climbing ability (maximum score = 18). Neurological evaluations were performed by an investigator who was blinded to the experimental conditions.

Rotarod experiment
This test was used to evaluate the exercise ability of rats by rotating a rod cylinder (ZH-300B, Zhenghua Biologic Apparatus Facilities Co., Ltd., Anhui, China) 20 . All rats were trained 3 d before modeling and were tested 1 d before SAH modeling and 1, 3, 7, and 14 d after SAH modeling.

Morris water maze
The Morris water maze test was carried out as previously described in a circular tank with a diameter of 180 cm and a depth of 50 cm 21 . Starting from the 18th d after modeling, all rats were trained for 4 d, three times per day. Each training session lasted for 1 min, with 5 min intervals between every two sessions. The test was performed from the 22nd to the 26th d after SAH modeling.

Brain water content analysis
Brain water content was measured by the wet/dry method 22 . In short, rats were deeply anesthetized with pentobarbital sodium 72 h after SAH modeling and euthanized. An electronic analytical balance (Sartorius BS 210 S, Göttingen, Germany) was used to weigh each part.

Immunohistochemical (IHC) staining
Brain section (4 μm in thickness) were prepared and treated with an EDTA buffer solution (0.05 mol/L Tris, 0.001 mol/L EDTA, pH 8.5), followed by antigen retrieval. Endogenous peroxidase activity was inactivated with 0.3% H 2 O 2 for 10 min. After being blocked with 5% bovine serum albumin (BSA) for 20 min, the sections were incubated with primary antibodies against amyloid precursor protein (APP, ab101492, 1:100, Abcam, Cambridge, UK) overnight at 4°C and biotinylated goat anti-mouse immunoglobulin G (IgG) secondary antibodies (ab190475, 1:500, Abcam) for 20 min. Finally, the sections were stained with 3,3-diaminobenzidine and counterstained with hematoxylin. Images were obtained with a microscope (Leica-DM2500, Wetzlar, Germany). Two independent pathologists evaluated the sections. The rate of positive cells for each sample was counted according to the APP staining intensity.
Immunofluorescence and terminal deoxyribonucleotidyl transferase-mediated 2'-deoxyuridine 5'-triphosphate nick end labeling (TUNEL) staining MA, USA), Alexa Fluor 488 donkey anti-mouse IgG (Thermo Fisher Scientific), and CY3 donkey anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) at 25°C for 1 h. Next, the sections were stained with 4',6-diamidino-2-phenylindole (10 μg/mL, Sigma-Aldrich, St. Louis, MO, USA) at 25°C for 15 min. Transferase-mediated TUNEL (Cell Death Detection Kit, Roche, Basel, Switzerland) costaining with rabbit anti-NeuN (1:200, Abcam) and neuronal apoptosis were assayed. The sections were observed with an Olympus BX51 fluorescence microscope (Olympus, Tokyo, Japan) or a laser scanning confocal microscope (FV500, Olympus). The image control parameters were set by the control group, and all other parameters remained the same for image capture. Five hematoma border regions were selected in each part for analysis, and four parts were selected for each animal (n = 6). ImageJ (National Institutes of Health, Bethesda, Maryland, USA) was used to count the total number of TUNEL and NeuN doublepositive cells in five areas near the injury area.

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
Cells were lysed using a TRIzol kit (Invitrogen, Carlsbad, CA, USA), and total RNA was extracted from cell and tissue samples. RNA quality and concentration were measured by UV-Vis spectrophotometry (ND-1000, Thermo Fisher Scientific). For mRNA analysis, an RT kit (RR047A, Takara, Tokyo, Japan) was used to perform RT to obtain cDNA. Subsequently, cDNA was used as a template, and the SYBR ® Premix Ex Taq TM II kit (Perfect Real Time, DRR081, Takara) was used to perform fluorescent qPCR. The samples were subjected to RT-qPCR in a real-time fluorescent qPCR instrument (ABI 7500, ABI, Foster City, CA, USA). The 2− ΔΔ CT method was used to quantify expression, and GAPDH was used as the internal reference. The primers are shown in Supplementary Table 1.

Nuclear and cytoplasmic separation experiment
A NE-PER™ Nuclear and Cytoplasm Extraction Kit (78833, Thermo Fisher Scientific, https://www.thermofisher.cn/) was used to separate the nucleus and cytoplasm in each group of cells, as well as to separate nuclear and cytoplasmic proteins. Then, western blot analysis was performed to measure the expression of TDP-43 (ab109535, 1:1000, Abcam). GAPDH (ab8245, 1:1000, Abcam) was used as the internal reference for cytoplasmic proteins, and Lamin A (ab133256, 1:1000, Abcam) was used for nucleoproteins. The ratio of the gray value of TDP-43 to the gray value of each control protein was used for quantitative analysis by ImageJ 1.48 u software (Bio-Rad, Hercules, CA, USA).

Determination of malondialdehyde (MDA) levels and superoxide dismutase (SOD) activity
After 72 h, the brain was perfused with PBS, and then the ipsilateral cortex was homogenized to measure MDA levels and the activity of SOD. The MDA level was measured by reacting MDA and thiobarbituric acid under acidic conditions (high temperature). SOD activity was measured by the WST-1 method. c The protein level of TDP-43 in the brain tissues of rats at 3 h, 6 h, 24 h, 72 h, 7 d, and 14 d after SAH modeling was determined by western blot analysis, *P < 0.05 compared to sham-operated rats. d TDP-43 expression and localization in different neurons was examined by immunofluorescence staining (neuronal marker: NeuN, astrocyte marker: GFAP, microglial marker: Iba1). *P < 0.05, n = 6. The experiment was repeated three times independently.

Fluoro-Jade C (FJC) staining
(Olympus) was used to observe FJC-positive cells, and the number was counted by technicians who were blinded to experimental conditions.

Dihydroethidium (DHE) staining
DHE staining was carried out to measure superoxide anions, which indicate the level of oxidative stress in tissues 23 . The ventral side of the left hemisphere was examined with the help of a laser scanning confocal microscope (A1 Si, Nikon, Tokyo, Japan). The representative image was obtained from a Section 2 mm posterior to the bregma.

Coimmunoprecipitation (Co-IP) assay
Equal amounts of cell lysates (1500 µg) were immunoprecipitated with 1 μg of HSP70, HDAC1, TDP-43, or acetyl-K antibodies and 50 μL protein A/G agarose. The immunoprecipitate was washed twice with 10 mM HEPES (pH 7.9), 1 mM EDTA, 150 mM NaCl, and 1% Nonidet P-40 and boiled for 10 min. Then, the precipitated proteins were eluted with 30 μL of sodium dodecyl sulfate (SDS)-PAGE buffer. The eluted proteins were separated by 8% SDS-PAGE, transferred to a nitrocellulose membrane, and examined with the corresponding antibodies.

Flow cytometry
After 48 h of transfection, the cells were detached with EDTA-free 0.25% trypsin (YB15050057, YuBo Biotech Co., Ltd., Shanghai, China), collected in a flow tube, and centrifuged, and the supernatant was discarded.

Measurement of neurite length
ImageJ software was used to quantify neurite length. Images of cells labeled with anti-GFP and anti-neurofilament antibodies were captured by an Axiovert camera (images were processed using Volocity software). The Neuron J ImageJ plug-in was used to measure the lengths of neurites. Approximately 150 cells under each experimental condition were analyzed by three independent researchers.

Transmission electron microscopy (TEM)
The grid was checked at 80 KV on a Philips CM120 electron microscope with a low-magnification image (X4, X800 or X7000) of the white matter area containing the axon tract, as well as a high-magnification image (X15000 or X20000) of axons and other subcellular elements that were captured. From the examined samples, axons with myelin sheaths and axon organelles conformed to the ultrastructural standards. TEM images were captured with a charge-coupled device camera (Gatan, Pleasanton, CA, USA) and processed with Digital Micrograph software (Gatan). i Axonal damage was observed by TEM. *P < 0.05 compared with SAH rats infected with oe-NC, n = 6. The experiment was repeated three times independently.
Statistical analysis SPSS 21.0 statistical software (IBM Corp., Armonk, NY, USA) was used for statistical analysis. Measurement data are displayed as the mean ± standard deviation. First, the normality and homogeneity of variance were tested. Data with a normal distribution and even variance between two groups were analyzed by unpaired t tests, while data comparisons among multiple groups were performed using one-way analysis of variance (ANOVA).
Comparisons among data at different time points were performed by repeated-measures ANOVA with Tukey's post-hoc test. A value of P < 0.05 was considered statistically significant.

RESULTS
TDP-43 was elevated in SAH patients and a rat model of SAH To explore the mechanism of TDP-43 in SAH, cerebrospinal fluid was collected from 10 non-SAH patients and 15 patients with aneurysmal SAH. We found that the expression of TDP-43 in the cerebrospinal fluid of SAH patients was much higher than that in the CSF of non-SAH patients (Fig. 1a, b).
Then, a rat model of SAH were constructed by the endovascular puncture. In model rats, we found severe cerebral hemorrhage, increased brain water content ( Supplementary Fig. 1a), severe neurological damage ( Supplementary Fig. 1b), impaired motor function ( Supplementary Fig. 1c), and learning and memory impairment ( Supplementary Fig. 1d). The number of DHE-positive cells increased dramatically, SOD activity decreased, and MDA levels increased in rats with SAH ( Supplementary Fig. 1e, f). TUNEL and FJC staining demonstrated that the number of apoptotic neuronal cells and neuronal cell degeneration notably increased in rats with SAH ( Supplementary Fig. 1g, h). IHC staining of APP showed that there was a large number of positive local accumulations in the white matter area of rats with SAH ( Supplementary Fig. 1i). TEM revealed that there were a variety of ultrastructural features of myelinated and unmyelinated axonal damage at the cerebral infarcts and hairy infarcts of rats with SAH, including axon enlargement, nerve fiber compaction, and organelle aggregation ( Supplementary Fig. 1j). Therefore, the rat model of SAH was successfully constructed and featured cognitive impairment and axonal injury.
The expression of TDP-43 in the brain tissues of SAH rats at 3 h, 6 h, 24 h, 72 h, 7 d, and 14 d after SAH modeling was examined, and the results showed that TDP-43 increased notably after SAH modeling, and the highest expression was observed at 72 h (Fig. 1c). Immunofluorescence analysis revealed that the fluorescence intensity of TDP-43 in neurons was much higher than that in microglia or astrocytes (Fig. 1d). In summary, TDP-43 was mainly located in neurons and was highly expressed in the cerebrospinal fluid of SAH patients and a rat model of SAH.

TDP-43 overexpression promoted cognitive impairment and axonal damage in rats with SAH
The lateral ventricle was injected with TDP-43-overexpressing lentivirus to determine the mechanism by which TDP-43 regulated cognitive impairment in SAH. The level of TDP-43 mRNA in the brain tissues of rats 72 h after SAH modeling was determined by RT-qPCR. The results demonstrated that the expression of TDP-43 in rats with SAH that were infected with oe-TDP-43 was sharply increased compared with that in rats with SAH that were injected with oe-NC (Fig. 2a). Three rats were randomly selected, and the level of TDP-43 protein in brain tissues was analyzed by western blotting. The expression of TDP-43 in rats after infection with oe-TDP-43 was notably increased (Fig. 2b). Cognitive tests revealed that oe-TDP-43 increased nerve function damage, reduced exercise capacity, impaired memory function, and increased total travel distance in the water maze in rats with SAH ( Fig. 2c-e). Rats that were treated with oe-TDP-43 exhibited an increased number of FJCpositive cells and NeuN + TUNEL + cells (Fig. 2f, g), indicating that neuronal injury was intensified. In addition, the number of APPpositive cells in rats infected with oe-TDP-43 was elevated (Fig. 2h). Compared with that in rats infected with oe-NC, the number of swollen and dystrophic axons in rats infected with oe-TDP-43 increased dramatically (Fig. 2i), indicating that axon damage was enhanced. In summary, TDP-43 overexpression further deteriorated cognitive dysfunction and axonal damage in rats with SAH.

OxyHb-induced axonal damage by promoting TDP-43 accumulation in the cytoplasm
To explore the underlying mechanism by which TDP-43 affects axonal damage, neurons were treated with 20 μM OxyHb for 6, 12, and 24 h. We showed that the expression of TDP-43 increased during the treatment time (Fig. 3a). The expression of axon damage marker proteins, including dynactin, neurofilament light (NFL) and apolipoprotein E (ApoE), was then measured, the protein expression of dynactin and NFL increased, while the level of ApoE decreased increasing OxyHb treatment time (Fig. 3a), indicating that OxyHb induced axonal damage. The position of TDP-43 in OxyHb-treated neurons was further observed by immunofluorescence staining, and the results showed that the localization of TDP-43 in the cytoplasm was notably increased (Fig. 3b). Nuclear and cytoplasmic separation experiments showed that the expression of TDP-43 in the cytoplasm increased notably (Fig. 3c).
The length of neurofilament-associated antigen (NAA) antibodylabeled neurites and enhanced green fluorescent protein (EGFP)-labeled neurites was measured by immunofluorescence staining, and the results demonstrated that OxyHb decreased total neurite length in neurons, while TDP-43 silencing reversed this outcome (Fig. 3d). In addition, OxyHb treatment induced irregular nuclear morphology and invaginations of the nuclear membrane, which was reversed by TDP-43 silencing (Fig. 3e). Flow cytometry showed that neuronal apoptosis was increased with OxyHb treatment time, and TDP-43 silencing inhibited OxyHb-induced neuronal apoptosis (Fig. 3f). In summary, OxyHb induced cytoplasmic accumulation of TDP-43 and resulted in axonal damage and abnormal nuclear membrane morphology, and further promoted neuronal apoptosis. TDP-43 silencing alleviated axonal damage and abnormal nuclear membrane morphology and inhibited neuronal apoptosis.

HDAC1/HSP70/TDP-43 triple complexes promoted cytoplasmic accumulation of TDP-43
The STRING website predicted that HDAC1-HSP70 (Hspa1b)-TARDBP (TDP-43) was a regulatory pathway (Fig. 4a). The colocalization of HDAC1 and HSP70 with TDP-43 was examined, and the results demonstrated that the positive rate of HDAC1, HSP70, and TDP-43 in the brain tissues of rats with SAH increased notably compared with that in sham-operated rats, and the pathological colocalization resulted in the formation of point aggregates (Fig. 4b). Then, 293T cells were transfected with HDAC1 or HSP70. the Co-IP results showed that acetyl-K levels were sharply reduced ( Supplementary Fig. 2a), indicating that HDAC1 mediated the deacetylation of HSP70. In addition, HDAC1 silencing notably enhanced HSP70 acetylation (Supplementary Fig. 2b). The expression of TDP-43 in the cytoplasm gradually increased with increasing HDAC1 concentrations, while TDP-43 expression in the nucleus remained unchanged (Fig. 4c). sh-HDAC1-treated neurons were treated with 50 μg/mL cycloheximide (CHX) for 0 h, 2 h, 4 h, and 8 h. TDP-43 protein levels gradually decreased with prolonged CHX treatment time (Fig. 4d). sh-HDAC1-treated neurons were treated with 5 μM of the proteasome inhibitor MG132. Western blot analysis revealed that MG132 inhibited the degradation of TDP-43 compared with that in neurons that were not treated with MG132 (Fig. 4e), indicating that HDAC1 silencing promoted TDP-43 degradation through the proteasome pathway. 293 T cells were transfected with Flag-HDAC1, HA-TDP-43, and Myc-HDAC1, and the Co-IP results revealed that these factors could bind with each other (Fig. 4f). In summary, HDAC1 could bind to HSP70 and TDP-43, promote HSP70 deacetylation and enhance TDP-43 accumulation in the cytoplasm while inhibiting protein degradation. In contrast, HDAC1 silencing promoted proteasomal degradation of TDP-43.

HDAC1 silencing inhibited TDP-43 expression to reduce OxyHb-induced axonal damage
The SAH-related microarray GSE54083 was obtained from the GEO database and further analyzed, and the results revealed that HDAC1 was highly expressed in SAH (Fig. 5a). KEGG enrichment analysis showed that HDAC1-related genes were mainly enriched in pathways such as "Metabolic pathways", "Neuroactive ligand-receptor interaction", "Cytokine-cytokine receptor interaction", "Herpes simplex virus 1 infection" and "Huntington disease" (Fig. 5b). RT-qPCR showed that OxyHb notably elevated the expression of HDAC1 in neurons, and sh-HDAC1 reduced the mRNA levels of HDAC1 in OxyHb-treated neurons (Fig. 5c). Based on the western blot results, OxyHb treatment promoted the expression of HDAC1 and TDP-43 in neurons and inhibited the level of acetyl-K, while sh-HDAC1 treatment reduced the levels of HDAC1 and TDP-43 but increased the acetyl-K level in OxyHb-treated neurons (Fig. 5d). Immunofluorescence staining indicated that OxyHb increased the level of TDP-43 in the cytoplasm, while sh-HDAC1 resulted in decreased TDP-43 in the cytoplasm (Fig. 5e). Moreover, we found reduced expression of TDP-43 in the cytoplasm but increased neurite length in neurons that were treated with OxyHb + sh-HDAC1 compared with those treated with OxyHb + sh-NC (Fig. 5f, g). TEM revealed that compared with that of neurons treated with OxyHb + sh-NC, the morphology of neurons treated with OxyHb + sh-HDAC1 was round and regular, and nuclear membrane morphology was normal (Fig. 5h). Flow cytometry showed that compared with neurons treated with OxyHb + sh-NC, neurons treated with OxyHb + sh-HDAC1 exhibited reduced apoptosis (Fig. 5i). In summary, HDAC1 silencing inhibited TDP-43 expression and alleviated OxyHb-induced axonal damage.
SAHA alleviated neuronal damage by promoting TDP-43 degradation by maintaining the acetylation level of HDAC1/ HSP70 complexes Neurons were treated with the HDAC1 inhibitor SAHA (1 μM) for 30 min, and the acetylation level of HSP70 was measured. SAHA treatment notably promoted HSP70 acetylation in 293T cells ( Supplementary Fig. 3a). According to the IP results, a notable increase in HSP70 levels was observed ( Supplementary Fig. 3b). SAHA-treated 293 T cells were further transfected with Flag-HDAC1, Myc-HSP70, and HA-TDP-43. The Co-IP results showed that the level of TDP-43 that immunoprecipitated with HDAC1 was markedly reduced (Fig. 6a), and the same results were observed in SAHAtreated neurons (Fig. 6a). CHX-treated neurons were further treated with SAHA, and the results showed that SAHA treatment markedly promoted the protein degradation of TDP-43 in neurons that were treated with CHX (Fig. 6b). The colocalization of TDP-43 and HDAC1 was examined by immunofluorescence staining, and the results showed that the accumulation of TDP-43 in the cytoplasm of neurons induced by OxyHb was reduced by SAHA treatment, and the ratio of nuclear-to-cytoplasmic TDP-43 was reduced (Fig. 6c).
In addition, compared with that of neurons treated with OxyHb + DMSO, neurite length was increased by TDP-43 (Fig. 6d). Flow cytometry showed that compared with neurons treated with OxyHb + DMSO, neurons treated with OxyHb + SAHA had notably reduced levels of apoptosis (Fig. 6e). Overall, SAHA, which is an HDAC1 inhibitor, promoted the degradation of the TDP-43 protein by maintaining the acetylation of HSP70 and inhibiting the accumulation of TDP in the cytoplasm, ultimately alleviating neurite damage.
SAHA relieved axonal damage and neurological dysfunction after SAH by inhibiting the HDAC1/HSP70/TDP-43 axis To further explore whether the deacetylase inhibitor SAHA is involved in the regulation of nerve damage in SAH by mediating the HDAC1/HSP70/TDP-43 axis, SAH model rats were injected with lentivirus-mediated sh-HDAC1 and SAHA in the lateral ventricle. RT-qPCR analysis showed that the level of HDAC1 in rats with SAH that were injected with sh-NC was notably increased relative to that in sham-operated rats. The level of HDAC1 was notably decreased in rats with SAH that were injected with sh-HDAC1 compared with that of rats with SAH that were injected with sh-NC. Compared with that of rats with SAH that were treated with DMSO, the level of HDAC1 showed no obvious change in rats with SAH that were treated with SAHA (Fig. 7a). In addition, the levels of HDAC1 and TDP-43 in rats with SAH that were injected with sh-HDAC1 were reduced, the level of acetyl-K was elevated, and the level of HSP70 remained unchanged relative to the effect of sh-NC treatment. Compared with rats with SAH that were treated with DMSO, acetyl-K levels increased markedly, HDAC1 and HSP70 levels remained unchanged, and TDP-43 levels decreased notably in rats with SAH that were treated with SAHA (Fig. 7b). Cognitive tests showed that compared with that of rats with SAH that were injected with sh-NC, the neurological damage score was reduced, and the exercise capacity and memory level were increased in rats with SAH that were injected with sh-HDAC1; the same behavioral phenotype was  observed after SAHA treatment (Fig. 7c-e). Then, FJC and TUNEL staining demonstrated that in comparison to that of rats with SAH that were injected with sh-NC, the number of FJC-positive cells and the number of NeuN + TUNEL + cells in rats with SAH that were injected with sh-HDAC1 were reduced, and the same trend was observed after SAHA treatment (Fig. 7f, g), indicating that neuronal damage was reduced. APP expression was then examined by IHC staining, and the results demonstrated that the number of APPpositive cells in rats with SAH that were injected with sh-HDAC1 was largely decreased compared with that of rats with SAH that were injected with sh-NC, and SAHA treatment inhibited the positive rate of APP in brain tissues (Fig. 7h). The number of swollen and dystrophic axons in rats with SAH that were injected with sh-NC was notably reduced compared with that in rats with SAH that were injected with sh-HDAC1, and SAHA treatment alleviated axonal damage (Fig. 7i). In conclusion, SAHA alleviated axonal damage and cognitive impairment by mediating the degradation of TDP-43 through the acetylation of HSP70 and the inhibition of HDAC1.

DISCUSSION
The average SAH patient is 55 years old, and this condition commonly leads to cognitive impairments in survivors because of axonal damage and white matter injury 24,25 . Much attention has been given to the function of various HDAC inhibitors in the progression of intracerebral hemorrhage, including SAH 26,27 . Here, we attempted to clarify the mechanisms of the HDAC inhibitor SAHA in SAH, and our results revealed that SAHA relieved neuronal and axonal damage by promoting TDP-43 degradation by enhancing the acetylation of HSP70 through HDAC1 inhibition, thereby attenuating the development of SAH. We discovered that TDP-43 was highly expressed in the cerebrospinal fluid of SAH patients and in the brain tissues of a rat model of SAH. Consistent with our findings, a previous study also showed upregulated expression of TDP-43 in the cerebrospinal fluid of patients with aneurysmal SAH and the brain tissues of a rat model of SAH 17 . Furthermore, our study revealed that TDP-43 overexpression enhanced cognitive impairment in rats with SAH in vivo and that TDP-43 accumulation in the cytoplasm in a neuronal cell model of SAH worsened axonal damage in vitro. Cytoplasmic mislocalization is a characteristic of TDP-43 pathology and is further recognized as a hallmark of neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS), frontotemporal lobar degeneration (FTLD), Parkinson's disease and Huntington's disease [28][29][30] Fig. 7 SAHA alleviates brain damage in SAH rats by suppressing HSP70/HDAC1/TDP-43. a HDAC1 mRNA levels in rats after different treatments were determined by RT-qPCR. b Protein levels of HDAC1, acetyl-K, HSP70, and TDP-43 in rats with SAH after different treatments was determined by western blot analysis. c Nerve damage in rats with SAH after different treatments was determined by the modified Garcia behavior score. d The movement of rats with SAH after different treatments was determined by the rotarod test. e Memory impairment in rats with SAH after different treatments was determined by the Morris water maze test. f The degeneration of cortical neurons in rats with SAH after different treatments was determined by FJC staining. g Apoptosis in neurons was determined by TUNEL and NeuN double staining. h APP expression in rats with SAH after different treatments was determined by IHC staining. i Axonal damage in the brain tissue of rats with SAH after different treatments was observed by TEM. *P < 0.05, n = 6. The experiment was repeated three times independently.  Fig. 8 The molecular mechanism by which SAHA alleviates early SAH. SAH induces HDAC1 binding to HSP70 and promotes TDP-43 accumulation in the cytoplasm to form inclusion bodies, ultimately promoting neuronal and axonal damage and inducing brain damage in early SAH. SAHA protects against early SAH by promoting the degradation of TDP-43 by enhancing HSP70 acetylation through the inhibition of HDAC1.