Abstract
Alzheimer's disease (AD) is the most common neurodegenerative disease,
characterized by memory loss, speech and motor defects, personality
changes, and psychological disorders. The exact cause of AD remains
unclear. Current treatments focus on maintaining neurotransmitter
levels or targeting β-amyloid (Aβ) protein, but these only alleviate
symptoms and do not reverse the disease. Developing new drugs is
time-consuming, costly, and has a high failure rate. Utilizing
multi-omics for drug repositioning has emerged as a new strategy. Based
on transcriptomic perturbation data of over 40,000 drugs in human cells
from the LINCS-L1000 database, our study employed the Jaccard index and
hypergeometric distribution test for reverse transcriptional feature
matching analysis, identifying Givinostat as a potential treatment for
AD. Our research found that Givinostat improved cognitive behavior and
brain pathology in models and enhanced hippocampal synaptic plasticity.
Transcriptome sequencing revealed increased expression of mitochondrial
respiratory chain complex proteins in the brains of APP/PS1 mice after
Givinostat treatment. Functionally, Givinostat restored mitochondrial
membrane potential, reduced reactive oxygen species, and increased ATP
content in Aβ-induced HT22 cells. Additionally, it improved
mitochondrial morphology and quantity in the hippocampus of APP/PS1
mice and enhanced brain glucose metabolic activity. These effects are
linked to Givinostat promoting mitochondrial biogenesis and improving
mitochondrial function. In summary, Givinostat offers a promising new
strategy for AD treatment by targeting mitochondrial dysfunction.
Keywords: Alzheimer's disease, Drug repositioning, Givinostat, APP/PS1
mice, Mitochondrial dysfunction
1. Introduction
Alzheimer's disease (AD) is a neurodegenerative disorder and the most
common type of dementia. According to the latest data from the World
Health Organization, there were approximately 55 million dementia
patients in 2019, and it is estimated to increase to 139 million by
2050 [[39]1]. The clinical features of AD primarily involve cognitive
impairments, accompanied by typical phenotypes such as aphasia,
apraxia, agnosia, and mood disturbances, severely impacting patients'
quality of life and even endangering life in the late stages of the
disease [[40]2]. Amyloid plaques formed by Aβ and neurofibrillary
tangles formed by excessive phosphorylation of Tau protein are the main
pathological hallmarks and diagnostic gold standards of AD [[41]3].
Currently, the etiology of Alzheimer's disease remains unclear, and
drugs for its treatment mainly focus on restoring or maintaining
neurotransmitters in damaged neurons. However, these methods only
provide symptomatic relief and cannot reverse the progression of the
disease, even with long-term use [[42]4]. The latest drugs targeting Aβ
protein through immunotherapy have yielded disappointing results in
phase 3 clinical trials [[43]5]. Therefore, it is essential to search
for new prevention and treatment strategies.
Mitochondrial dysfunction is a common intracellular functional
disturbance in neurodegenerative diseases. Mitochondria, organelles
found in eukaryotic cells, are responsible for various biological
functions such as energy production, metabolic regulation, cellular
signal transduction, and calcium buffering [[44]6]. Maintaining
synaptic function by mitochondria plays a crucial role in the
development of neurons [[45]7]. The brain is an organ with high oxygen
consumption, and the process of synaptic transmission in neurons relies
on the energy provided by mitochondria in the form of ATP (adenosine
triphosphate) [[46]8]. ATP, derived from the mitochondrial electron
transport chain, flows from complex I to complex IV, simultaneously
generating an electrochemical proton gradient across the mitochondrial
inner membrane. Complex V (ATP synthase) utilizes this gradient to
produce ATP [[47]9]. Disruption of energy metabolism and mitochondrial
dysfunction can lead to ATP deficiency and excessive ROS (reactive
oxygen species) production, further resulting in abnormal neuronal
function or even apoptosis, directly associated with the pathology of
AD [[48]10]. Past studies have indicated that mitochondrial dysfunction
can cause pathological Aβ production, while the accumulation of
pathological Aβ protein can, in turn, lead to mitochondrial
dysfunction, forming a vicious cycle of positive feedback
[[49]11,[50]12]. It has been reported that mitochondrial dysfunction
occurs prior to the appearance of AD pathological features, further
highlighting the significant role of mitochondrial dysfunction in the
progression of AD [[51]13]. Therefore, mitochondrial dysfunction is a
key factor in the progression of AD.
The development of new drugs is characterized by high cost investment,
high failure rates, and high risks [[52]14]. In recent years, with the
rapid development of omics and medical informatics, drug research and
development technologies have progressed rapidly. Drug repositioning is
a strategy that utilizes big data-related techniques to re-screen,
combine, or modify existing drugs to discover their unknown new uses
[[53]15]. This approach has many advantages over traditional drug
development, including lower risk, shorter development cycles, and
lower investment. It can also enter clinical trials more quickly,
making it an effective way to accelerate new drug development [[54]16].
By integrating data such as disease and drug transcriptomes and using
correlation algorithms, we have predicted the drug Givinostat for
Alzheimer's disease.
Givinostat is an effective inhibitor of class I/II histone deacetylases
(HDACi), and HDACi is associated with cognitive function, memory and
learning, stress response, synaptic plasticity, and DNA damage repair
[[55][17], [56][18], [57][19]]. Duchenne muscular dystrophy (DMD) is a
degenerative muscle disease. Givinostat administration can restore the
physiological epigenetic characteristics of the PGC-1α promoter,
improve mitochondrial damage in DMD, enhance mitochondrial biogenesis,
and slow down muscle degeneration in the mdx mouse model [[58]20]. In
another in vitro study, Givinostat alleviated Aβ-induced neurotoxicity
while reducing ROS production, exerting neuroprotective effects
[[59]21]. DMD is a degenerative muscle disease. Hayward GC et al. were
the first to report a shift in memory-specific brain regions towards
amyloid protein production in a DMD model mouse, suggesting that the
pathogenic mechanism of DMD is similar to that of AD [[60]22].
Furthermore, some studies have indicated that Givinostat can
effectively cross the blood-brain barrier (BBB) [[61]23]. The above
evidence suggests that Givinostat may have the potential to treat AD.
However, the effect of Givinostat on AD animal models and its
systematic mechanism have not been reported yet. In this study, we used
behavioral tests to examine the effect of chronic intraperitoneal
injection of Givinostat on improving learning and memory in APP/PS1
mice, and further explored its changes and mechanisms of action on
brain pathology.
2. Materials and methods
2.1. Data set selection and analysis
The AD transcriptome dataset was retrieved from the Gene Expression
Omnibus (GEO) database by screening for (1) species Homo sapiens, (2)
tissue origin of the brain hippocampus, and (3) attainment of at least
three biological replicates. The [62]GSE173955 dataset was finally
selected. The dataset was provided by Mizuno et al. [[63]24], which
contains hippocampal transcriptome sequencing data from 10 health
controls and 8 patients with AD. Differentially expressed genes (DEGs)
were identified using R-packet edgeR with |logFC| > 1 and P < 0.05 as
the threshold, and functional enrichment analysis was performed using
Metascape [[64]25].
2.2. Silico-based drug repurposing
Large-scale drug or small molecule perturbation data were obtained from
the Library of Integrated Network-Based Cellular Signatures (LINCS)
L1000 database [[65]26]. LINCS aims to collect perturbation responses
to human cells caused by compounds, genetic states, and diseases
worldwide, with the L1000 database containing more than 40,000
perturbation responses to human cells from all types of substances
[[66]27]. In this study, the DEGs identified and the drug perturbation
response data were analyzed for reverse transcriptional feature
matching, and the jaccard index was first used to assess the degree of
association between drugs and diseases [[67]28], followed by the
hypergeometric distribution test to compute the statistical P-value
[[68]29], and Fisher's combined probability test to obtain the final
P-value [[69]30].The drug repurposing process is shown in [70]Fig.
1.The formula algorithm is shown in [71]Table 1.
Fig. 1.
[72]Fig. 1
[73]Open in a new tab
Research flowchart for this section.
Table 1.
Silico -based drug repurposing strategies in this study.
Algorithmic metrics Formulas
Jaccard similarity coefficient
[MATH: Jaccardindex=(|Dru<
mi>gup∩DEGsd<
mi>own||Dru<
mi>gup∪DEGsd<
mi>own|+|Dru<
mi>gdown∩DEGsup||Dru<
mi>gdown∪DEGsup|)/2 :MATH]
Hypergeometric test
[MATH:
Pup(X≥kup↓)=1−Fup
(kup<
/mrow>−1;N,Kdown,nup) :MATH]
[MATH:
Pdown(X≥kdown↑)=1−Fdo
wn(kdo<
mi>wn−1;N,Kup,ndown) :MATH]
Fisher's combined probability test
[MATH: χ2=−2(ln(Pup
)+ln(Pdown)) :MATH]
[74]Open in a new tab
2.3. Experimental mice and pharmacological treatment
In this study, male APP/PS1 double transgenic mice (n = 32) and
age-matched wild-type C57 male mice (n = 32) were used. All mice were
purchased from Beijing Hua FuKang Biological Technology Co., Ltd.
(Beijing, China). The APP/PS1 mice overexpress human amyloid precursor
protein with the Swedish (K595 N/M596L) mutation and human PSEN1 gene
with exon 9 deletion (PS1dE9) (strain B6C3-Tg (APPswe, PSEN1dE9)
85Dbo/J, stock no. #004462 from the Jackson Laboratory mouse database
[[75]31]. They were housed in a barrier facility at the Laboratory of
Physiology, Shanxi Medical University, with a temperature of 22 ± 2 °C
and a 12-h light/dark cycle, and had ad libitum access to food and
water. All animal experiments were conducted in accordance with the
guidelines for the care and use of laboratory animals at Shanxi Medical
University, under an approved protocol by the Experimental Animal
Ethics Committee of Shanxi Medical University (Taiyuan, China) (License
No.: SYDL2023032). The 8-month-old mice were randomly divided into four
groups: WT + Saline, WT + Givinostat, APP/PS1+Saline, and
APP/PS1+Givinostat, with 16 mice in each group. Each group randomly
selected 10 animals for behavioral experiments.
Givinostat (ITF-2357, Invivo Chem) was dissolved in sterile 100 %
dimethyl sulfoxide (DMSO), then further diluted in 0.9 % saline to a
final concentration of 1 g/L. Referencing previously studied drug
dosages, Givinostat (10 mg/kg) or an equivalent vehicle was
administered via intraperitoneal injection (i.p.) for 30 consecutive
days prior to the behavioral experiments. The injections were continued
during the behavioral testing period and maintained for an additional 2
weeks after the experiments [[76]32,[77]33].
2.4. Open field test
The purpose of the open field experiment is to eliminate animal
movement barriers and assess the animals' spontaneous activity and
exploratory ability. The open field is a square box with dimensions of
40 cm in length, width, and height. The central quarter of the bottom
surface is designated as the central zone, while the rest is the
peripheral zone. Each mouse is placed into the open field from the same
central position and allowed to freely explore for 5 min.
Simultaneously, their walking trajectory is recorded using a camera,
and the Smart 3.0 software is used to analyze the total distance
traveled by the mice in the open field and the ratio of time spent in
the central zone.
2.5. Y-maze
The Y-maze is used to assess the short-term working memory of mice. The
maze consists of three interconnected Y-shaped arms with angles of 120°
each. Each arm has dimensions of 30 cm in length, 8 cm in width, and
15 cm in height. The central area of the maze forms an equilateral
triangle. Mice are placed into the center of the triangle and allowed
to explore without interference for 8 min. Cameras and Smart 3.0
software record their activity trajectory and the sequence of arm
entries. Each time a mouse enters an arm different from the previous
two entries, it is counted as a correct entry. During the interchange
between two mice, the maze is cleaned with 75 % alcohol. Finally, the
total number of arm entries and the spontaneous alternation correct
rate are calculated. The spontaneous alternation correct rate is
calculated as follows: [number of correct entries/(total number of arm
entries - 2)] × 100 %.
2.6. Morris water maze
The Morris water maze is used to assess long-term spatial learning and
memory capabilities. The maze consists of a cylindrical water tank with
a radius of 60 cm and a height of 50 cm. The tank is filled with water
at 22 ± 1 °C and maintained at a constant temperature. The circular
water surface is evenly divided into four equal-sized sectors,
representing four quadrants, with different shapes marked on the walls
of the tank in each quadrant to help mice reference memory locations
and find the platform. A platform is placed at the center of the fourth
quadrant, with a circular shape and a radius of 6 cm. To prevent mice
from directly seeing the platform, a sufficient amount of titanium
dioxide is mixed into the water in the maze, and the water level should
be about 1 cm higher than the platform. During the five days of the
spatial acquisition phase, mice are gently placed into the water facing
the maze wall from a random quadrant. Due to stress, mice will swim
toward the platform for escape. If a mouse reaches the platform within
60 s of entering the water, it is allowed to stay on the platform for
5 s. If the platform is not found within 60 s, the mouse is guided to
the platform manually and allowed to stay for 15 s to memorize its
location. Each mouse is trained four times per day. The escape latency
for mice to reach the platform is recorded. On the sixth day, the
spatial probe trial is conducted by removing the platform from the
water. Mice are placed into the water from two randomly selected
quadrants excluding the target quadrant, and their swimming time is
recorded for 60 s. Their swimming trajectories and swimming times in
each of the four quadrants are recorded. Finally, the visible platform
trial is conducted to rule out the influence of visual acuity on mice's
swimming performance. The entire process, including trajectories and
time, is recorded using cameras and EthoVision XT 15 software (Noldus
Information Technology).
2.7. In vivo electrophysiology
After completing the behavioral experiments, in vivo hippocampal LTP
recording experiments were conducted. Mice were anesthetized by
intraperitoneal injection of pentobarbital sodium (40 mg/kg) and their
heads were securely fixed on a stereotaxic instrument. The anterior
fontanelle was located on the skull, and a mark was made 2.0 mm
posterior to the anterior fontanelle and 1.5 mm lateral to the sagittal
suture. A dental drill was used to create a circular hole with a
diameter of 3 mm. According to the mouse brain atlas, stimulating and
recording electrodes were positioned in the Schaffer collateral pathway
and the stratum radiatum, respectively, in the CA1 region of the
hippocampus, until the maximum field excitatory postsynaptic potential
(fEPSP) was recorded. Subsequently, the current intensity was increased
from 0 in steps of 0.002 mA–0.2 mA, and the corresponding fEPSP was
recorded to construct an input/output (I/O) curve. The current
intensity corresponding to 1/3–1/2 of the maximum fEPSP amplitude was
selected as the test stimulation intensity. This intensity was applied
60 times with an interval of 30 s, and the fEPSP was stably recorded
for 30 min. Then, three paired-pulse stimuli with an interval of 50 ms
were given, with each pair separated by 30 s, inducing paired-pulse
facilitation (PPF), which reflects presynaptic mechanisms. Following
PPF, three trains of high-frequency stimulation (HFS) consisting of 20
pulses at 200 Hz were applied, inducing long-term potentiation (LTP).
The stable recording lasted for 60 min using a multi-channel biological
signal acquisition/processing system (RM6240C, Chengdu Instruments,
China). Finally, the percentage change in fEPSP slope before and after
high-frequency stimulation was calculated for each group.
2.8. Tissue preparation and antibodies
After electrophysiological experiments, randomly selected groups of
mice underwent immunofluorescence staining and Western blotting (WB).
They were deeply anesthetized with pentobarbital sodium (40 mg/kg,
intraperitoneal injection). For immunofluorescence staining, mice were
perfused with 0.01 M phosphate-buffered saline (PBS) followed by 4 %
paraformaldehyde (PFA). Subsequently, the brains were removed and
sequentially immersed in 4 % PFA for 24 h, followed by 15 % sucrose for
24 h and 30 % sucrose for 48 h. Coronal brain sections (30 μm thick)
were cut using a cryostat (CM1950, Leica) and mounted on glass slides.
For Western blotting, mice were perfused with pre-chilled physiological
saline (0.9 %) through the heart, and the hippocampal tissue was
rapidly dissected and stored at −80 °C. Details of the antibodies used
in this study can be found in Supplementary document 1 [78]Table S1.
2.9. Immunofluorescence and thioflavin S (ThioS) staining
Wash brain slices with PBS three times, 5 min each time. Permeabilize
cells with 0.5 % Triton X-100 at room temperature for 15 min, followed
by washing brain slices with PBS three times, 5 min each time. Block
with 5 % BSA at room temperature for 60 min, then discard the blocking
solution and directly incubate with primary antibody at 4 °C overnight.
The next day, discard the primary antibody, and wash with PBS five
times, 5 min each time. Incubate with fluorescent secondary antibody,
avoiding light, at 37 °C for 1 h, then wash with PBS five times, 5 min
each time. Stain with Sulforhodamine S staining solution at room
temperature, avoiding light, for 8 min, then wash twice with 80 %
ethanol, 10 s each time. After washing with PBS three times (5 min each
time), incubate with DAPI staining solution (AR1176, BOSTER) at room
temperature for 10 min, then wash with PBS 3–5 times, 5 min each time.
Mount with antifade reagent (AR1109, BOSTER), observe under a
fluorescence microscope, take pictures, record the results, and analyze
using Image J software. Import the merged image into Image J, split the
channels, select the positive signal areas in each channel, and use the
ROI Manager to identify signals common to multiple channels as the
colocalization regions, followed by statistical analysis.
2.10. Western blot
First, protein extraction was performed by adding RIPA buffer (AR 0102,
BOSTER), PMSF (AR1179, BOSTER), and protein phosphatase inhibitor
(AR1183, BOSTER) to the hippocampal tissue samples, followed by
homogenization using sonication. After centrifugation at 4 °C for
30 min (16,000g, 5430 R, Eppendorf, Germany), the protein concentration
in the supernatant was quantified using the BCA protein assay kit
(AR0197, BOSTER). The protein was diluted in loading buffer (AR 0131,
BOSTER, China) and denatured at 95 °C for 5 min. Equal amounts of total
protein from each sample (AR0138, BOSTER) were separated by SDS-PAGE
and transferred to PVDF membranes (0.45 μm or 0.22 μm, Millipore). The
membranes were then blocked at room temperature for 15 min (AR0041,
BOSTER), followed by overnight incubation with primary antibody at
4 °C, washing of the membranes, and subsequent incubation with
HRP-conjugated secondary antibody at room temperature for 2 h. The
protein bands were visualized using the ECL protein blotting detection
kit (P0018 FS, Beyotime, China) and imaged using the Azure c300
chemiluminescent protein blot imaging system (Azure Biosystems, USA).
Densitometry analysis was performed using Image J software.
2.11. Enzyme-linked immunosorbent assay
The levels of Aβ40 and Aβ42 were detected according to the instructions
of ELISA kits (E-EL-H0542, Elabscience; E-EL-H0543, Elabscience). In
brief, fresh mouse brain tissue homogenates were prepared, standards
were made, samples were added, enzymes were added, incubation was
performed, washing was conducted, color development was initiated,
termination was carried out, and the levels of Aβ40 and Aβ42 were
analyzed and calculated using an ELISA reader (SpectraMax 190,
Molecular Devices, USA).
2.12. mRNA sequencing (mRNA-Seq)
The mRNA-Seq experiment was conducted by Novogene Bioinformatics
Institute in Beijing, China. In brief, the experiment was carried out
according to detailed instructions from the manufacturer. Total RNA was
successfully extracted from each group of hippocampal samples (each
group containing 3 samples) using TRIzol reagent (Invitrogen, USA).
Subsequently, RNA purity was carefully checked using a NanoPhotometer®
spectrophotometer (IMPLEN, USA), and RNA integrity was evaluated using
the Bioanalyzer 2100 system from Agilent Technologies with the RNA Nano
6000 assay kit. During the preparation of mRNA-seq libraries required
for sequencing, we followed standard Illumina experimental protocols.
Subsequently, RNA-seq sequencing was performed using the Illumina
NovaSeq 6000 platform, generating paired-end reads of 150 bp in length.
During the analysis stage of sequencing data, base calling was
conducted using CASAVA software, and reads were accurately aligned to
the genome using the HISAT 2 (v2.0.5) spliced-read aligner with default
parameters. Finally, Novogene Corporation was responsible for
conducting mRNA quantification and differential gene expression
analysis, ensuring the accuracy and reliability of the experimental
results.
2.13. RT-PCR
Total RNA was extracted using the Total RNA Extraction Reagent
(MF034-01, Mei5 Biotechnology), and RNA was reverse transcribed into
cDNA using the M5 Super Plus qPCR RT Kit (MF166-Plius-01, Mei5
Biotechnology). Subsequently, each cDNA sample underwent real-time PCR
reactions in triplicate using the M5 Hiper SYBR Premix EsTaq (MF787-01,
Mei5 Biotechnology). Quantification was performed using the comparative
CT method (2−ΔΔCT), and the expression of each mRNA was calculated
relative to β-actin. The gene-specific primer sequences used are
provided in Supplementary document 1 [79]Table S2.
2.14. Transmission electron microscopy (TEM) analysis
Mouse hippocampal CA1 tissue sections were fixed at 4 °C in 2.5 %
glutaraldehyde and 1 % osmium tetroxide, followed by graded dehydration
in acetone solutions (30 % → 50 % → 70 % → 80 % → 90 % → 95 % → 100 %).
The tissue was then embedded in epoxy resin, and 70 nm ultrathin
sections were prepared using an ultramicrotome (Leica EM UC7). The
sections were stained with uranyl acetate and lead citrate. Finally,
transmission electron microscopy (JEM-1400PLUS, JEOL) was used to
capture images at the same magnification, and three random fields of
view per sample were selected to quantify the number of mitochondria
within neurons.
2.15. MicroPET-CT
Using a micro-PET/CT scanner (BioSpec70/20USR, Bruke. Co, GER), a
detailed scan of the mouse heads was performed to observe the metabolic
distribution of glucose in the mouse brain. The sterile injection
solution of 18F-FDG used in the experiment was provided by the Imaging
Department of the First Hospital of Shanxi Medical University. Prior to
the experiment, the mice were fasted to ensure the accuracy of the
experimental results. Subsequently, the mice were successfully
anesthetized by inhalation of isoflurane. After disinfection of the
body parts, 220 μCi (7.4 MBq) of sterile solution of 18F-FDG was
administered to the mice via intraperitoneal injection. The mice were
allowed to absorb the injection for 60 min. The micro-PET/CT scanner
scanned the mouse heads to observe the distribution of 18F-FDG.
Anesthesia was maintained throughout the entire experiment.
Three-dimensional models were used to collect data, and image filtering
and back projection were used to reconstruct cross-sectional CT images
of the mice. The uptake rates of the mouse hippocampus and cortical
brain regions were analyzed using PMOD software (PMOD Technologies,
Zurich, Switzerland).
2.16. Cell culture and Aβ Oligomer preparation
This study employed the mouse hippocampal neuron cell line HT22 (Cell
Resource Center, Peking Union Medical College, China). HT22 cells were
cultured in high-glucose DMEM medium (BOSTER, PYG0073) supplemented
with 10 % fetal bovine serum (Excell Bio, FSP500) and 50 U/ml
penicillin-streptomycin mixture (MeilunBio, MA0347) at 37 °C with 5 %
CO2. The culture medium was refreshed every day. When the cells reached
95 % confluency at the bottom of the culture flask, they were
dissociated using trypsin (MeilunBio, MA0232) at 37 °C for 5 min to
terminate digestion, followed by complete digestion of the culture
medium. Aβ1-42 was purchased from QYAOBIO (China). Following previously
established methods, Aβ1-42 was dissolved in hexafluoroisopropanol to a
concentration of 1 mM, aliquoted, dried under nitrogen, and stored at
−80 °C. Prior to use, it was dissolved and incubated at 4 °C for 24 h
to complete oligomerization[[80]34,[81]35].
2.17. Cell counting kit-8
The CCK-8 assay kit was used to assess cell viability and determine the
optimal concentration of Givinostat for antagonizing Aβ1-42. After
trypsin digestion, the cells were resuspended in culture medium and
seeded into a 96-well plate. The cells were divided into control groups
(without Aβ intervention), and intervention solutions containing
20 μM Aβ1-42 and different concentrations of Givinostat (0 nM, 50 nM,
100 nM, 150 nM, 200 nM) were added. After 24 h of incubation, 10 μL of
CCK-8 solution was added to each well and incubated for 1 h. Absorbance
was measured at 450 nm using a microplate reader (SpectraMax M2,
Molecular Devices).
2.18. Measurement of mitochondrial membrane potential (Δψm)
Except for the CCK-8 assay, all cell experiments were divided into
three groups: control group (no Aβ intervention), Aβ group
(20 μM Aβ1-42), and Aβ+Givinostat group (20 μM Aβ1-42 + 100 nM
Givinostat), with drug treatment for 24 h. The mitochondrial membrane
potential was assessed using the JC-1 Mitochondrial Membrane Potential
Detection Kit (C2003S, Beyotime). The culture medium was removed, and
the cells were washed with PBS. Then, 1 ml of cell culture medium and
1 ml of JC-1 staining working solution were added and mixed thoroughly.
The cells were incubated in a 37 °C cell incubator for 20 min. After
removing the supernatant, the cells were washed with JC-1 staining
buffer, and 2 ml of cell culture medium was added for observation under
a fluorescence microscope.
2.19. ROS production assays
The reactive oxygen species (ROS) assay kit (S0033S, Beyotime) was
utilized to detect ROS levels. DCFH-DA was diluted in serum-free
culture medium at a ratio of 1:1000. After removing the cell culture
medium, an appropriate volume of diluted DCFH-DA was added to the
cells. The cells were then incubated at 37 °C in a cell culture
incubator for 20 min. Afterward, the cells were washed three times with
serum-free cell culture medium and observed under a fluorescence
microscope.
2.20. Measurements of MDA, GSH, and GSSG
Following the provided guidelines, the levels of MDA and GSH/GSSG were
measured using the corresponding kits (S0131 M, Beyotime; S0053,
Beyotime). The absorbance was determined using a microplate reader
(SpectraMax M2, Molecular Devices).
2.21. ATP concentration assays
The ATP assay kit (S0026, Beyotime) was employed to detect ATP levels.
The culture medium was aspirated, and the cells were lysed. After
centrifugation at 12,000g for 5 min at 4 °C, the supernatant was
collected for subsequent analysis. A standard curve was prepared, and
ATP detection working solution was added to the detection wells. Then,
samples and standard solutions were added, mixed rapidly, and the ATP
content was measured using a multi-mode microplate reader (SpectraMax
M2, Molecular Devices).
2.22. Statistical analysis
Statistical analysis and graphing of the data were performed using
GraphPad Prism 9 software. All experimental data are presented as
mean ± standard error of the mean (SEM). The escape latency and
swimming speed in the Morris water maze experiment were analyzed using
repeated measures multivariate analysis of variance, while other data
were analyzed for group differences using either two-way ANOVA or
one-way ANOVA, followed by Tukey's multiple comparisons test. The
comparison of Aβ level between the two groups was performed using an
unpaired t-test. P < 0.05 was considered statistically significant.
3. Results
3.1. Identification of Differentially Expressed Genes in Alzheimer's Disease
and Drug Repositioning
After screening the [82]GSE173955 dataset, 1596 differentially
expressed genes (DEGs) were identified ([83]Fig. 2A), among which 795
genes were upregulated and 801 genes were downregulated. The results of
Gene Ontology (GO) functional enrichment analysis revealed that these
DEGs were primarily enriched in cellular components such as the
postsynaptic membrane, plasma membrane, and neuronal projection. These
components are involved in biological processes like regulation of
postsynaptic membrane potential, potassium ion transport, and nervous
system processes ([84]Fig. 2B). Kyoto Encyclopedia of Genes and Genomes
(KEGG) pathway enrichment analysis showed that the DEGs were mainly
enriched in signaling pathways such as neuroactive ligand-receptor
interaction, glutamatergic synapse, cholinergic synapse, and long-term
potentiation ([85]Fig. 2C).
Fig. 2.
[86]Fig. 2
[87]Open in a new tab
Identification of Differentially Expressed Genes in Alzheimer's Disease
and Drug Repositioning. (A) Volcano map of DEGs screened on
[88]GSE173955 dataset. (B) GO enrichment of DEGs for the [89]GSE173955
dataset. (C) KEGG enrichment of DEGs for the [90]GSE173955 dataset. (D)
GO enrichment of DEGs for the givinostat targets. (E) KEGG enrichment
of DEGs for the givinostat targets.
Reverse transcription profiling analysis was performed on the obtained
DEGs and the L1000 database, which resulted in the selection of 14
drugs. These drugs exhibited significant reverse similarity in
perturbation profiles with the transcriptional signature of AD
(P < 0.05 and Jaccard index >0.2), as shown in [91]Table 2.
Additionally, we annotated the blood-brain barrier (BBB) permeability
of these drugs. Based on the Jaccard index and BBB permeability, we
selected the drug Givinostat. Approved by the FDA for the treatment of
DMD, Givinostat reduces inflammation and muscle loss, protecting muscle
function, with certain safety and effectiveness. There are no reported
studies on the effects of Givinostat on AD models. Enrichment analysis
of Givinostat perturbation profiles revealed mitochondrial-related
biological processes such as regulation of outer mitochondrial membrane
permeability, negative regulation of mitochondrial depolarization,
release of cytochrome c from mitochondria, and oxidative
phosphorylation uncoupler activity. It also involves signaling pathways
like MAPK signaling pathway, glycolysis, and P53 signaling pathway,
which aligns with our expectations for the screening ([92]Fig. 2D–E).
Below, we will explore its potential role in AD animal models.
Table 2.
Potential therapeutic drugs for AD obtained based on reverse
transcriptional features.
Compound name Jaccard index P value BBB permeability
Trichostatin-a 0.229 <0.001 BBB+
Cyclosporin-a 0.229 <0.001 BBB-
Doxorubicin 0.228 <0.001 BBB-
Vorinostat 0.219 <0.001 BBB+
Daunorubicin 0.215 0.003 BBB-
Givinostat 0.213 <0.001 BBB+
Ouabain 0.211 <0.001 BBB-
Parthenolide 0.210 0.003 BBB+
Wortmannin 0.204 <0.001 BBB+
Piceatannol 0.203 <0.001 BBB+
Geldanamycin 0.202 0.007 BBB-
Dorsomorphin 0.202 <0.001 BBB-
Vincristine 0.201 0.015 BBB-
Indirubin 0.201 <0.001 BBB+
[93]Open in a new tab
3.2. Givinostat improves cognitive abilities in APP/PS1 mice
In the open field test, there was no statistically significant
difference in the total distance traveled by mice in the WT + Saline
group, WT + Givinostat group, APP/PS1 + Saline group, and
APP/PS1 + Givinostat group (P > 0.05), indicating that Givinostat does
not affect the locomotor activity of mice ([94]Fig. 3C). However, the
percentage of time spent in the central zone by mice in the
APP/PS1 + Saline group was significantly lower than that in the
WT + Saline group (P < 0.05). After Givinostat treatment, the
percentage of time spent in the central zone by APP/PS1 mice increased
(P < 0.05) ([95]Fig. 3B). In summary, we conclude that 9-month-old
APP/PS1 mice exhibit behaviors resembling depression, with reduced
spontaneous exploration and investigatory behavior in unfamiliar
environments. Treatment with Givinostat enhances the exploratory
behavior of APP/PS1 mice in unfamiliar environments, improves their
adaptation to new environments, and suggests that Givinostat can
improve the autonomous exploratory behavior of APP/PS1 mice.
Fig. 3.
[96]Fig. 3
[97]Open in a new tab
Givinostat treatment improved cognitive abilities in APP/PS1 mice. (A)
Behavioral Experiment Flowchart. (B) Percentage of time spent in the
central area of the open field by mice in each group. (C) Histogram
showing the total distance traveled by mice in the open field. (D)
Representative movement trajectories of mice in the open field
experiment for each group. (E) Schematic diagram of the Y-maze
experiment apparatus. (F) Histogram showing the spontaneous alternation
rate of mice in the Y-maze. (G) Histogram showing the total number of
arm entries in the Y-maze for each group. (H) Line graph showing the
escape latency during the place navigation test. ∗ indicates
APP/PS1 + Saline vs WT + Saline, # indicates APP/PS1 + Givinostat vs
APP/PS1 + Saline. (I) Histogram showing the number of platform
crossings during the spatial exploration test for each group. (J)
Histogram showing the percentage of time spent in the target quadrant
during the spatial exploration test for each group. (K) The time of
arrival of each group of mice at the platform in the visual platform
test. (L) Line graph showing the change in average swimming speed of
mice during the Morris water maze experiment. (M) Representative
swimming trajectories of mice during the spatial exploration test for
each group. n = 10, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. ^#P < 0.05,
^##P < 0.001.
The Y-maze test is used to assess the short-term working memory of
mice. The experimental results show that there is no difference in the
average number of arm entries within 8 min among the four groups of
mice (P > 0.05) ([98]Fig. 3F). However, the spontaneous alternation
accuracy of APP/PS1 mice is significantly lower than WT mice
(P < 0.001), and Givinostat significantly increases the spontaneous
alternation accuracy of APP/PS1 mice (P < 0.05) ([99]Fig. 3G). This
indicates that Givinostat can improve the impairment of short-term
working memory in APP/PS1 mice without affecting the spontaneous
movement of mice.
The Morris water maze is used to assess the long-term spatial learning
and memory abilities of mice. Over a period of 6 days, there were no
significant differences in swimming speed among the four groups of
mice, indicating no difference in motor ability. During the 1–5 days of
the spatial acquisition training period, the escape latency of the
APP/PS1 mice treated with saline was significantly longer on the 4th
day (P < 0.05) and 5th day (P < 0.001), while Givinostat significantly
shortened the escape latency of APP/PS1 mice on the 4th day (P < 0.05)
and 5th day (P < 0.01), suggesting that Givinostat reversed the spatial
learning impairment in APP/PS1 mice ([100]Fig. 3H). During the spatial
probe test on the 6th day, compared to the WT + Saline group, the
percentage of time spent in the target quadrant and the number of
platform crossings were significantly reduced in the APP/PS1 + Saline
group, while Givinostat reversed this situation (P < 0.05) ([101]Fig.
3I–J), indicating that Givinostat improved the spatial memory
impairment in APP/PS1 mice. During the testing of the visible platform,
the mice directly reached the platform, and there was no difference in
the time taken to arrive, indicating that the differences among the
mice during the experiment were not influenced by their vision
([102]Fig. 3K).
3.3. Givinostat rescues synaptic plasticity damage in APP/PS1 mice
Synaptic plasticity is closely associated with learning and memory, and
synaptic plasticity impairment is involved in various neurodegenerative
diseases, including AD. Synaptic plasticity includes both functional
and structural plasticity. Synaptic plasticity is measured by changes
in synaptic strength. LTP is an activity-dependent form of synaptic
plasticity that leads to long-lasting enhancement of synaptic
transmission. Hippocampal LTP is an important indicator of synaptic
plasticity. We investigated LTP in APP/PS1 mice, as shown in [103]Fig.
4. There were no differences in the I–V relationships and the ratio of
fEPSP 2/fEPSP 1 among the four groups of mice, indicating no
differences in baseline synaptic transmission capacity among the four
groups, and neither the APP/PS1 gene mutation nor Givinostat treatment
affected the release of presynaptic neurotransmitters. Figure [104]Fig.
4B shows changes in fEPSP slope in the four groups of mice before and
after HFS. The fEPSP slope increased in all four groups after HFS, with
a subsequent decline over time. Compared to the WT + Saline group, the
decline was more pronounced in the APP/PS1+Saline group, with
significant differences at 30 min (P < 0.001) and 60 min (P < 0.001).
In contrast, the APP/PS1+Givinostat group showed better maintenance
(30 min: P < 0.001; 60 min: P < 0.01) ([105]Fig. 4C). The LTP
experimental results provide electrophysiological evidence for the
neuroprotective effect of Givinostat in behavior. Proteins in the
presynaptic membrane and postsynaptic membrane are the basis of
synaptic structural plasticity. We measured the protein levels of
synaptic-related proteins PSD-95 and SYN ([106]Fig. 4G–I). SYN and
PSD95 are respectively characteristic presynaptic and postsynaptic
proteins, mainly involved in regulating synaptic activity and
plasticity. The results showed that the protein levels of PSD-95
(P < 0.05) and SYN (P < 0.01) in the APP/PS1+Saline group were
significantly lower than those in the WT + Saline group, while the
protein levels of PSD-95 (P < 0.05) and SYN (P < 0.01) were restored
after Givinostat treatment. These findings indicate that Givinostat
enhances hippocampal synaptic plasticity in APP/PS1 mice, accompanied
by an increase in the levels of synaptic-related proteins PSD-95 and
SYN.
Fig. 4.
[107]Fig. 4
[108]Open in a new tab
Givinostat rescues hippocampal synaptic plasticity in APP/PS1 mice. (A)
Schematic representation of the positions of the stimulating and
recording electrodes in the hippocampal brain region during in vivo
field potential recordings. SE denotes the stimulating electrode, RE
denotes the recording electrode, and SC denotes the Schaffer collateral
pathway. (B) Time course plot showing the percentage change in
hippocampal fEPSP slope before and after HFS in each group of mice. (C)
Histogram showing the percentage change in hippocampal fEPSP slope at
different time points before and after high-frequency stimulation for
each group of mice. (D) Histogram showing the paired-pulse facilitation
(PPF) in each group of mice. (E) Input-output (I–O) curves depicting
the trend of fEPSP amplitude changes with increasing stimulation
current intensity in each group of mice. (F) Representative fEPSP
waveforms before (solid line) and 60 min after (dashed line)
high-frequency stimulation in each group of mice. n = 6, ∗∗P < 0.01,
∗∗∗P < 0.001. (G) Representative immunoblot bands of hippocampal PSD95
and SYN for each group of mice. (H) Histogram showing the expression
levels of PSD95. (I) Histogram showing the expression levels of SYN.
n = 5, ∗P < 0.05, ∗∗P < 0.01.
3.4. Givinostat reduces Aβ levels in the hippocampus of APP/PS1 mice
Aβ plaques are important pathological markers in AD. We reflected the
levels of hippocampal Aβ in mice through Thio S staining and Aβ (6E10)
immunostaining. As shown in [109]Fig. 5A, in the immunofluorescence
experiment, Thio S staining and 6E10 immunostaining were not detected
in WT mice, indicating the absence of gene mutations in WT mice.
However, both groups of APP/PS1 mice exhibited positive Thio S staining
and 6E10 immunostaining, with a good overlap between the two.
Importantly, the number and area of Thio S-positive and 6E10
immunostained plaques were significantly lower in APP/PS1+Givinostat
group than in APP/PS1+Saline group (Thio S: P < 0.01, 6E10: P < 0.05)
([110]Fig. 5B–C).The co-localization results of Thio S and 6E10 are
close to those of 6E10 (Fig. 5D). Furthermore, we also detected the
levels of Aβ40 and Aβ42 subtypes in the hippocampus using ELISA
([111]Fig. 5E).Consistent with the former, Givinostat significantly
reduced the levels of Aβ40 (P < 0.05) and Aβ42 (P < 0.05) in APP/PS1
mice. This indicates that Givinostat can alleviate the brain pathology
in APP/PS1 mice.
Fig. 5.
[112]Fig. 5
[113]Open in a new tab
Givinostat treatment reduces Aβ levels in the hippocampal tissue of
APP/PS1 mice. (A) Representative immunofluorescence images of
hippocampal Aβ plaques in the four groups of mice. The first row shows
cell nuclei stained with DAPI (blue), the second row shows Aβ plaques
stained with Thioflavin S (Thio S) (green), the third row shows
positive Aβ plaques stained with 6E10 antibody (red), and the fourth
row shows merged images of the previous three rows. Scale bar: 300 μm.
(B) Histogram showing the number and area of Thio S-positive stained
plaques in the hippocampal region of each group of mice. (C) Histogram
showing the number and area of Aβ plaques positive for 6E10 antibody in
the hippocampal region of each group of mice. (D) Histograms showing
the number and area of co-localized positive plaques in the hippocampus
of each group of mice. (E) ELISA detection of Aβ40 and Aβ42 levels in
the hippocampal tissue of each group of mice. n = 4, ∗P < 0.05,
∗∗P < 0.01.
3.5. Givinostat treatment increased the expression of mitochondrial
respiratory chain complex-related proteins in the hippocampal tissue of
APP/PS1 mice
In the previous section, we observed that Givinostat improved cognitive
behavior, enhanced hippocampal synaptic plasticity, and ameliorated
brain pathology in APP/PS1 mice. To further investigate the underlying
mechanisms, we used RNA-Seq to analyze the transcriptomic changes in
the hippocampus of Givinostat-treated APP/PS1 mice. We found 945
differentially expressed genes between the WT + Saline group and the
APP/PS1+Saline group, and 571 differentially expressed genes between
the APP/PS1+Givinostat group and the APP/PS1+Saline group. There were
401 overlapping genes, indicating that 42.4 % of the genes were
reversed after Givinostat treatment. We performed KEGG pathway
enrichment analysis on these differentially expressed genes to explore
the pathways potentially regulated by Givinostat treatment, as shown in
[114]Fig. 6B. The results were primarily enriched in neurodegenerative
diseases such as Huntington's disease, Parkinson's disease, Alzheimer's
disease, and oxidative phosphorylation. According to the enrichment
results (Supplementary document 2), the AD pathway and the oxidative
phosphorylation pathway shared common genes, including respiratory
chain complex-related genes Atp5k, Cox6b2, Uqcr11, Cox5b, Ndufa3, and
Cox6a2. We then used RT-PCR to assess the expression of Atp5k, Cox6b2,
Uqcr11, and Cox5b, and found that the expression of these genes was
reduced in APP/PS1 mice. However, Givinostat treatment reversed the
reduction of Atp5k, Cox6b2, and Cox5b, while Uqcr11 was only partially
reversed([115]Fig. 6D–G). Furthermore, we validated the protein
expression of these genes, and as shown in [116]Fig. 6H–L, both mRNA
and protein levels were consistent with the RNA-seq results. We
hypothesize that the improvement in learning and memory abilities and
the neuroprotective effects of Givinostat in APP/PS1 mice are
associated with mitochondrial function.
Fig. 6.
[117]Fig. 6
[118]Open in a new tab
Givinostat treatment increased the expression of mitochondrial
respiratory chain complex-related proteins in the hippocampus of
APP/PS1 mice. (A) Differential gene Venn diagrams for WT + Saline vs
APP/PS1 + Saline and APP/PS1 + Givinostat vs APP/PS1 + Saline. (B) KEGG
enrichment analysis of DEGs (C) Heatmap of differential genes encoding
mitochondrial respiratory chain complex proteins, n = 3. (D–G)
Statistical histograms of mRNA expression levels of Cox5b, Cox6b2,
Atp5k, and Uqcr11 in the hippocampus of mice in each group, n = 5.
(H–L) Statistical histograms of protein expression levels of Cox5b,
Cox6b2, Atp5k, and Uqcr11 in the hippocampus of mice in each group,
n = 5. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.
3.6. Givinostat prevents Aβ-induced mitochondrial dysfunction in HT22 cells
Mitochondrial dysfunction is an early event in AD pathogenesis, as
indicated in the preceding sections, Givinostat may act through
mitochondria. We observed the morphology of hippocampal mitochondria in
APP/PS1 mice using TEM. As shown in [119]Fig. 7A, compared to WT mice,
hippocampal mitochondria in APP/PS1 mice exhibited severe damage,
characterized by morphological swelling, vacuolization, and cristae
disruption, along with a significant decrease in quantity (P < 0.05).
However, after Givinostat treatment, mitochondria appeared with clearer
cristae, reduced structural damage, and increased normal mitochondrial
numbers (P < 0.05). This suggests that Givinostat treatment can
ameliorate the morphological abnormalities of mitochondria in the
hippocampal region of APP/PS1 mice. To determine the effect of
Givinostat on mitochondrial function induced by Aβ in HT22 cells, we
used the CCK8 method to determine the optimal concentration of
Givinostat (Supplementary document 1 [120]Fig. S1). Givinostat showed a
trend of inhibiting Aβ toxicity at concentrations of 50 nM, 100 nM,
150 nM, and 200 nM, with the cell viability of the 100 nM group being
superior to that of the 0 nM group (P < 0.05). Therefore, 100 nM was
determined as the most appropriate drug concentration. We measured
mitochondrial membrane potential using JC-1 dye [121]Fig. 7C. When the
mitochondrial membrane potential is normal, JC-1 enters mitochondria
and forms red fluorescent aggregates; when the membrane potential
decreases, JC-1 is released in the form of green fluorescent monomers.
[122]Fig. 7D shows the ratio of JC-1 aggregates to monomers, indicating
that Aβ-induced HT22 cell membrane potential decreased (P < 0.001),
while Givinostat rescued the Aβ-induced decrease in HT22 cell membrane
potential (P < 0.01). Additionally, we investigated whether Givinostat
could inhibit Aβ-induced oxidative damage in HT22 cells, quantifying
ROS levels using the ROS-sensitive fluorescent indicator DCFH-DA
([123]Fig. 7E). Aβ-induced HT22 cells showed a significant increase in
ROS (P < 0.01), while Givinostat significantly reversed intracellular
ROS levels (P < 0.05). Similarly, we found that MDA levels in
Aβ-induced HT22 cells significantly increased (P < 0.001) and were
markedly reduced after Givinostat treatment (P < 0.01). The GSH/GSSG
ratio in Aβ-induced HT22 cells was significantly decreased (P < 0.001),
but it significantly increased following Givinostat treatment
(P < 0.01) ([124]Fig. 7F–H). In summary, these results indicate that
Givinostat rescues mitochondrial structural and functional damage as
well as oxidative stress in AD.
Fig. 7.
[125]Fig. 7
[126]Open in a new tab
Givinostat rescued mitochondrial function. (A) Transmission electron
microscope images of mitochondria (red arrows) in the hippocampal
region of mice from each group. (B) Quantitative histogram of
mitochondrial counts in equally sized regions, with a scale of 1 μm,
n = 5. (C–D) Representative images of JC-1 staining in HT22 cells from
different treatment groups, with a scale of 100 μm; histograms
quantifying red/green fluorescence intensity, n = 5. (E–F)
Representative images of ROS (DCFH-DA) staining in HT22 cells from
different treatment groups, with a scale of 00 μm; histograms
quantifying fluorescence intensity, n = 4.(G-H)The levels of MDA,
GSH/GSSG in each group, n = 4. ∗P < 0.05. ∗∗P < 0.01, ∗∗∗P < 0.001.
3.7. Givinostat enhances cerebral glucose metabolism in APP/PS1 mice and
energy metabolism in Aβ-pre-treated HT22 cells
To assess whether Givinostat treatment can activate metabolism, we used
microPET to study the glucose metabolism levels in the brains of
APP/PS1 mice ([127]Fig. 8A). As expected, compared to the WT group, the
hippocampal and cortical FDG uptake rates (SUV) were significantly
decreased in APP/PS1 mice (hippocampus P < 0.01, cortex P < 0.05), and
significantly increased after Givinostat treatment (hippocampus
P < 0.05, cortex P < 0.05) ([128]Fig. 8B–C). This indicates that
Givinostat can increase cerebral glucose metabolism in APP/PS1 mice,
which is directly related to short-term working memory and spatial
learning memory. Glucose is one of the most direct sources of energy in
organisms, and ATP is the direct carrier of energy. Meanwhile, we
measured the levels of ATP at the cellular level, and found that the
ATP levels significantly decreased in Aβ-induced HT22 cells
(P < 0.001), while they were restored after Givinostat treatment
(P < 0.01) ([129]Fig. 8D). Therefore, Givinostat can improve glucose
metabolism and ATP levels in AD.
Fig. 8.
[130]Fig. 8
[131]Open in a new tab
Givinostat enhanced glucose metabolism in the brains of APP/PS1 mice
and energy metabolism in Aβ-preconditioned HT22 cells. (A) Distribution
of 18F-FDG uptake in the brain tissues of mice from each group. (B–C)
Statistical histograms of SUV values of 18F-FDG in the hippocampal and
cortical regions of the brains of mice from each group. n = 5,
∗P < 0.05, ∗∗P < 0.01. (D) Histogram of ATP levels in different
treatment groups of HT22 cells. n = 3, ∗∗P < 0.01, ∗∗∗P < 0.001.
3.8. Givinostat enhances mitochondrial biogenesis in the hippocampus of
APP/PS1 mice
The PGC-1α/NRF-1-2/TFAM signaling pathway is a primary regulator of
mitochondrial biogenesis [[132]36]. NRF transcription factor plays a
role in the transcription of some mitochondrial genes, especially those
encoding subunits of mitochondrial respiratory chain complexes
[[133]37]. To determine whether the beneficial effects of Givinostat on
mitochondria are related to mitochondrial biogenesis, relevant
indicators were analyzed by immunoblotting. As expected, compared to
the WT group, levels of PGC-1α, NRF1, NRF2, and TFAM were reduced in
APP/PS1 mice, and Givinostat treatment could restore their levels to
varying degrees, indicating that Givinostat stimulates mitochondrial
biogenesis by inducing the PGC-1α signaling pathway ([134]Fig. 9A–E).
Fig. 9.
[135]Fig. 9
[136]Open in a new tab
Givinostat enhances mitochondrial biogenesis in the hippocampus of
APP/PS1 mice. (A) Representative immunoblot bands of PGC-1α, Nrf2,
Nrf1, and TFAM in the hippocampal tissues of mice from each group.
(B–E) Statistical histograms of expression levels of PGC-1α, Nrf2,
Nrf1, and TFAM in the hippocampus of mice from each group. n = 6,
∗P < 0.05, ∗∗P < 0.01.
3.9. Givinostat improved mitochondrial dynamics in the hippocampus of APP/PS1
mice
Mitochondrial dynamics primarily involve mitochondrial fusion and
fission [[137]38], which are crucial processes regulating mitochondrial
ultrastructure, quality, and function [[138]39]. Mitochondrial dynamics
are closely associated with mitochondrial function, and neurons are
particularly sensitive to disturbances in mitochondrial dynamics
[[139]40]. It has been reported that levels of mitochondrial fusion
proteins are significantly reduced in the brains of AD patients, while
levels of mitochondrial fission proteins are increased [[140]41]. We
found that compared to the WT group, levels of mitochondrial fission
proteins DRP1 and FIS1 were elevated in the hippocampus of APP/PS1
mice, while levels of mitochondrial fusion proteins MFN1 and MFN2 were
decreased. After Givinostat treatment, mitochondrial dynamics were
restored, with decreased levels of DRP1 and FIS1 and increased levels
of MFN1 and MFN2 in the hippocampus of APP/PS1 mice. In conclusion,
Givinostat improved mitochondrial dynamics in the hippocampus of
APP/PS1 transgenic mice ([141]Fig. 10A–E).
Fig. 10.
[142]Fig. 10
[143]Open in a new tab
Givinostat improved mitochondrial dynamics in the hippocampus of
APP/PS1 transgenic mice. (A) Representative immunoblot bands of Mfn1,
Mfn2, Drp1, and Fis1 in the hippocampal tissues of mice from each
group. (B–E) Statistical histograms of expression levels of Mfn1, Mfn2,
Drp1, and Fis1 in the hippocampus of mice from each group. n = 6,
∗P < 0.05, ∗∗P < 0.01.
4. Discussion
The understanding of AD pathogenesis remains limited. Three types of
FDA-approved medications for AD include cholinesterase inhibitors,
N-methyl-d-aspartate (NMDA) receptor antagonists, and Aβ-directed
antibodies [[144]42,[145]43]. The first two can only mildly alleviate
symptoms without providing a complete cure and are associated with
severe long-term side effects. The latter, represented by Aducanumab
and Leqembi, have been approved by the FDA in the past two years but
are currently only being used in patients with mild cognitive
impairment due to AD. The effectiveness and widespread applicability of
these treatments are still under investigation, leaving a lack of
effective drugs for AD [[146]5,[147]44]. Drug repurposing has shown
numerous successful examples. For instance, Sildenafil was initially
developed for hypertension but was unexpectedly found to be effective
in treating erectile dysfunction [[148]45]. Azidothymidine, initially
designed as a chemotherapeutic agent, failed in clinical trials but
eventually became the first FDA-approved antiretroviral drug for AIDS
[[149]46]). In this study, we focused on AD and used big data analysis
to investigate similarities in molecular mechanisms across different
diseases, as well as potential associations between drugs and diseases.
Through a scoring and ranking system for candidate drugs, we identified
Givinostat and applied it for the first time in AD animal experiments.
Disruption of the HAT/HDAC balance in the neuronal epigenome and a
reduction in histone acetylation have been observed in the hippocampus
of AD patients and AD mouse models, which leads to severe cognitive
impairment [[150]47]. Priyalakshmi Panikker also demonstrated that
elevated HDAC2 mRNA and protein levels occur in the early stages of
neurodegeneration, long before the formation of amyloid plaques and the
individual's death [[151]48]. Based on their sequence homology, HDACs
can be classified into four categories: I, II, III, and IV [[152]49].
Class I HDACi can improve learning and memory deficits and synaptic
damage in AD mice [[153]50], while thioacetamide-based Class II HDACi
can reduce Aβ levels in AD mice and enhance learning and memory
[[154]51]. In fact, simultaneously inhibiting different targets may
yield more effective results compared to using specific inhibitors
alone [[155]52]. Concurrently targeting several effective enzyme
activities represents a novel inhibition mode that can reduce side
effects, and this synergistic effect may contribute to better safety of
HDACIs in chronic treatments [[156]53]. Givinostat is a potent
broad-spectrum HDACi classified as Class I/II, which has garnered
significant attention due to its diverse applicability, efficacy, and
safety in humans [[157][54], [158][55], [159][56]]. Clinical trials
have demonstrated favorable tolerability and safety profiles for
Givinostat, leading to its approval by the U.S. FDA in 2024 as the
first non-steroidal drug for patients with all DMD gene mutations
[[160][57], [161][58], [162][59]]. Furthermore, research has indicated
that Givinostat treatment enhances overall functional and neurological
recovery following traumatic brain injury. Notably, administration of a
single dose of Givinostat even 24 h post-injury continues to confer
lasting benefits [[163]33]. Currently, there are no relevant studies on
Givinostat in AD animal models. Our study found that Givinostat
improves short-term working memory and long-term spatial learning
ability in APP/PS1 mice by restoring mitochondrial dysfunction, while
also alleviating their brain pathology and enhancing synaptic
plasticity in hippocampal neurons. Given Givinostat's neuroprotective
advantages, we believe it is a promising candidate for Alzheimer's
disease therapeutics.
The typical clinical feature of AD is cognitive impairment. In this
study, we utilized the APP/PS1 transgenic mouse model, which develops
Aβ deposition and cognitive impairment at 6 months of age [[164][60],
[165][61], [166][62]], widely employed in AD research. In our
investigation, we initially employed behavioral tests such as the open
field, Y-maze, and Morris water maze to assess exploratory activity,
short-term working memory, and long-term spatial learning memory in
mice. At 9 months of age, APP/PS1 mice exhibited significantly lower
learning and memory abilities compared to wild-type (WT) mice,
consistent with previous reports [[167]63]. However, after Givinostat
treatment, the overall locomotor activity in the open field test
remained unaffected in APP/PS1 mice, while exploratory behavior
markedly increased. Results from the Y-maze and Morris water maze
demonstrated that Givinostat improved short-term working memory and
long-term spatial learning memory in APP/PS1 mice. In behavioral
experiments, there were no significant differences in various
indicators between the Givinostat-treated and saline-treated groups of
WT mice, indicating that Givinostat did not significantly affect mouse
behavior within a certain concentration range, thus demonstrating its
safety.
Previous studies have shown a negative correlation between Aβ burden in
the brains of APP/PS1 mice and their learning and memory abilities
[[168]64,[169]65]. Aβ plaques, a hallmark pathological feature of AD
[[170]66], are neurotoxic and exacerbate the progression of AD
[[171]67,[172]68]. Aβ40 and Aβ42 are the most common subtypes produced
through the amyloidogenic pathway. In this study, we utilized dual
fluorescence labeling (6E10 and Thioflavin-S) to mark hippocampal Aβ
plaques, revealing a significant reduction in insoluble Aβ plaques in
the hippocampus of Givinostat-treated APP/PS1 mice. We further measured
soluble Aβ40 and Aβ42 levels in the hippocampus using ELISA, and
consistent with the earlier immunofluorescence results, both levels
were reduced following Givinostat treatment. Previous studies have
shown that mitochondrial energy deficits precede Aβ pathology, and
restoring mitochondrial function can effectively lower Aβ levels
[[173]69]. Additionally, reversing mitochondrial dysfunction and
increasing PGC-1α levels in the brains of AD mice have been shown to
reduce β-secretase processing, thereby lowering soluble Aβ levels
[[174]70]. In this study, we also found that Givinostat improves
mitochondrial dysfunction, and the restoration of mitochondrial
function is closely related to the metabolism of APP, mitophagy, or the
phagocytic function of microglia [[175][71], [176][72], [177][73],
[178][74]]. These Aβ clearance pathways may contribute to the reduction
in Aβ levels following Givinostat treatment; however, the precise
underlying mechanisms remain unclear and warrant further systematic
investigation.
How does Givinostat improve cognitive behavior in APP/PS1 mice? The
hippocampus is a crucial brain region for information processing and
learning and memory [[179][75], [180][76], [181][77]]. LTP results
showed that Givinostat treatment didon't affect PPF and IO curves in
mice, indicating that Givinostat treatment does not influence
neurotransmitter release and basic synaptic transmission. However,
following HFS, APP/PS1 mice exhibited LTP inhibition and impaired
synaptic plasticity, which was partially restored in the Givinostat
treatment group. This finding aligns with our behavioral results,
suggesting that Givinostat's improvement of learning and memory in
APP/PS1 mice is associated with enhanced hippocampal synaptic
plasticity. SYN and PSD-95 are essential for synaptic plasticity and
synaptic transmission [[182]78]. SYN is a synaptic vesicle protein
involved in Ca^2+-mediated neurotransmitter release [[183]79], while
PSD-95 is localized to the postsynaptic density of brain neurons,
organizing synaptic protein composition and structure [[184]79]. As
expected, our results demonstrated that levels of SYN and PSD-95
proteins were significantly reduced in APP/PS1 mice compared to the WT
group, and Givinostat treatment reversed this effect. In summary,
Givinostat treatment can reverse LTP suppression in APP/PS1 mice,
increase the expression of synaptic-related proteins, and alleviate
memory impairment in these mice.
The brain is the body's largest energy consumer, with the majority of
its energy devoted to synaptic transmission and memory formation
[[185]80]. Mitochondria serve as the energy factories of the brain, and
any energy impairment can lead to insufficient energy supply [[186]81].
Additionally, mitochondria provide critical metabolites for
biosynthesis and signaling molecules, which perceive and respond to the
environment, regulating cellular functions [[187]82]. Mitochondrial
dysfunction is an early event in AD, occurring before Aβ plaque
deposition, and is believed to play a crucial role in AD
[[188]83,[189]84]. Further exploring the molecular mechanisms
underlying the action of Givinostat, we conducted high-throughput mRNA
gene sequencing and found that Givinostat can increase the levels of
mitochondrial respiratory chain complex-related proteins in the
hippocampal region of APP/PS1 mice. This is consistent with previous
studies by Adav, Sunil S et al. and Rui Biet al., which demonstrated a
general decrease in mitochondrial electron transport chain and ATP
synthase protein abundance in brain tissues of AD patients compared to
healthy elderly individuals [[190]85,[191]86]. Additionally, related
COX genes were universally decreased in AD models [[192]86]. These
assembly defects and structural alterations in mitochondrial
respiratory chain complexes subsequently lead to a cascade of
deleterious effects, including reduced ATP production and, more
critically, electron leakage, accumulation of toxic reactive oxygen
species, and ultimately cell death and degeneration of affected tissues
[[193]87]. Our study found that the levels of mitochondrial respiratory
chain complex V Atp5k, complex IV Cox5b and Cox6b2, and complex III
subunit Uqcr11 were reduced in APP/PS1 mice, while Givinostat
significantly increased the expression levels of these genes. This
suggests that Givinostat restores the steady-state levels of
mitochondrial respiratory chain complex proteins in APP/PS1 mice,
indicating its potential role in restoring mitochondrial function.
Mitochondrial dysfunction in AD includes morphological damage, reduced
membrane potential, increased reactive oxygen species, and decreased
ATP levels [[194][88], [195][89], [196][90]]. In the AD brain,
decreased glucose metabolism is mainly attributed to reduced energy
metabolism caused by oxidative phosphorylation, suggesting that
mitochondrial dysfunction may play a significant role in AD development
[[197]91]. Yin J et al.'s study suggests that enhancing mitochondrial
function may alleviate AD pathology [[198]92]. Through electron
microscopy, we observed that Givinostat restores mitochondrial
morphology and quantity in APP/PS1 mice. Given the gene sequencing and
electron microscopy observations, we further explored the effects of
Givinostat on Aβ-induced HT22 cells, demonstrating that Givinostat
antagonizes the reduction of mitochondrial membrane potential induced
by Aβ, counters the increase in reactive oxygen species induced by Aβ,
and restores ATP energy loss induced by Aβ in HT22 cells. We assessed
glucose metabolism levels in mouse brains and found that Givinostat
restores glucose metabolism in APP/PS1 mice. Additionally, studies have
shown that neuronal energy deficiency severely affects synaptic
plasticity, leading to impaired information transmission and subsequent
loss of learning and memory [[199]93]. This also indicates that the
reversal of LTP suppression in APP/PS1 mice by Givinostat in LTP
experiments is directly related to the restoration of mitochondrial
function and improvement of energy supply.
Mitochondrial dynamics refer to the balance between fusion and fission
within the mitochondrial network to maintain its shape, distribution,
and size, thus ensuring the energy supply required by neurons
[[200]40,[201]94]. Studies have shown that in neurons of AD patients
and animal models, there is an increase in mitochondrial fission and a
decrease in fusion, leading to mitochondrial fragmentation, which is
crucial for mitochondrial health and function [[202]95,[203]96].
Mitochondrial fission is mediated by Drp1 and Fis1, while mitochondrial
fusion is mediated by Mfn1and Mfn2and optic atrophy 1 (Opa1) [[204]97].
Mitochondrial fusion promotes the mixing of mitochondrial matrix and
outer and inner membrane proteins to facilitate substance exchange and
ATP production [[205]98]. Excessive mitochondrial fission reportedly
affects energy production by affecting cristae integrity and oxidative
phosphorylation complex assembly [[206]99]. Partial inhibition of Drp1
can prevent the toxic effects of Aβ and tau, stabilize mitochondrial
dynamics, and increase mitochondrial biogenesis and synaptic activity
[[207]100]. Meanwhile, cells lacking mitochondrial fusion exhibit
impaired respiration and heterogeneity, highlighting the importance of
fusion in maintaining healthy and homogeneous mitochondrial populations
[[208]101]. Our research found a decrease in mitochondrial quantity,
incomplete morphology, and imbalanced dynamics-related proteins in
APP/PS1 mice. However, after receiving Givinostat treatment, these
conditions were restored, suggesting a direct correlation with the
restoration of mitochondrial dynamics.
Numerous studies indicate that PGC-1α is a primary regulator of
mitochondrial biogenesis, and its expression level is directly
correlated with mitochondrial biogenesis [[209]102,[210]103].
Furthermore, the expression of PGC-1α in the brains of Alzheimer's
disease (AD) patients decreases with the severity of dementia
[[211]104]. PGC-1α can activate the expression of numerous
mitochondrial-related nuclear genes, such as nrf1, nrf2, TFAM, and
nuclear-encoded subunits of mitochondrial respiratory chain complexes
I–V, which encode almost all proteins required for mitochondrial
biogenesis [[212]105,[213]106]. Matteo Giovarelli et al. found that
givinostat positively alters the epigenetic characteristics of the
PGC-1α promoter in a DMD mouse model, maintaining mitochondrial
biogenesis and oxidative fibers, effectively promoting mitochondrial
biogenesis in atrophic muscles [[214]20]. Consistent with our
hypothesis, mitochondrial biogenesis is depleted in APP/PS1 mice, and
Givinostat reverses this situation by restoring the protein levels of
PGC-1α, Nrf1, Nrf2, and TFAM involved in mitochondrial biogenesis. This
also explains our earlier findings of increased synthesis of subunits
of respiratory chain complexes after Givinostat treatment in APP/PS1
mice.
In this study, we innovatively employed a drug repositioning approach
to predict and experimentally validate potential therapeutic agents for
AD. This paradigm leveraged efficient computational methods to analyze
molecular properties, drug-target binding affinities, and
compound-protein interactions, significantly reducing the time and cost
associated with traditional drug development. Notably, we have, for the
first time, validated in AD animal models the neuroprotective effects
of the predicted drug, Givinostat, in APP/PS1 mice. These effects were
primarily attributed to the restoration of neuronal mitochondrial
function. In conclusion, drug repositioning emerges as a promising
strategy for drug discovery, and Givinostat holds great potential as a
candidate for the prevention and treatment of AD.
CRediT authorship contribution statement
Qi-Chao Gao: Writing – original draft, Visualization, Investigation.
Ge-Liang Liu: Software, Methodology, Data curation. Qi Wang: Software,
Methodology, Data curation. Sheng-Xiao Zhang: Methodology,
Conceptualization. Zhi-Lin Ji: Investigation. Zhao-Jun Wang: Writing –
review & editing, Supervision, Methodology. Mei-Na Wu: Writing – review
& editing, Methodology. Qi Yu: Writing – review & editing, Supervision,
Project administration, Funding acquisition. Pei-Feng He: Writing –
review & editing, Supervision, Methodology, Data curation,
Conceptualization.
5. Funding information
This work was supported by the National Social Science Fund of China
(21BTQ050), Key R&D Project of Shanxi Province (202102130501003),
Shanxi Key Laboratory of Big Data for Clinical Decision Research
(2021D100012021515245001135236), State Natural Science Fund project
(72474125), Science and Technology Innovation Teams of Shanxi Province
(202304051001017), and Shanxi Province Higher Education “Billion
Project” Science and Technology Guidance Project (BYJL003).
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to
influence the work reported in this paper.
Footnotes
^Appendix A
Supplementary data to this article can be found online at
[215]https://doi.org/10.1016/j.redox.2024.103420.
Contributor Information
Qi Yu, Email: yuqi@sxmu.edu.cn.
Pei-Feng He, Email: hepeifeng2006@126.com.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
Multimedia component 1
[216]mmc1.docx^ (3.9MB, docx)
Multimedia component 2
[217]mmc2.xls^ (30.3KB, xls)
Data availability
Data will be made available on request.
References