Neuroprotective Effects of Trilobatin, a Novel Naturally Occurring Sirt3 Agonist from Lithocarpus polystachyus Rehd., Mitigate Cerebral Ischemia/Reperfusion Injury: Involvement of TLR4/NF-κB and Nrf2/Keap-1 Signaling

Article abstract of DOI:10.1089/ars.2019.7825

Aims: Neuroinflammation and oxidative stress are deemed the prime causes of brain injury after cerebral ischemia/reperfusion (I/R). Since the silent mating-type information regulation 2 homologue 3 (Sirt3) pathway plays an imperative role in protecting against neuroinflammation and oxidative stress, it has been verified as a target to treat ischemia stroke. Therefore, we attempted to seek novel Sirt3 agonist and explore its underlying mechanism for stroke treatment both in vivo and in vitro. Results: Trilobatin (TLB) not only dramatically suppressed neuroinflammation and oxidative stress injury after middle cerebral artery occlusion in rats, but also effectively mitigated oxygen and glucose deprivation/reoxygenation injury in primary cultured astrocytes. These beneficial effects, along with the reduced proinflammatory cytokines via suppressing Toll-like receptor 4 (TLR4) signaling pathway, lessened oxidative injury via activating nuclear factor erythroid 2-related factor 2 (Nrf2) signaling pathways, in keeping with the findings in vivo. Intriguingly, the TLB-mediated neuroprotection on cerebral I/R injury was modulated by reciprocity between TLR4-mediated neuroinflammatory responses and Nrf2 antioxidant responses as evidenced by molecular docking and silencing TLR4 and Nrf2, respectively. Most importantly, TLB not only directly bonded to Sirt3 but also increased Sirt3 expression and activity, indicating that Sirt3 might be a promising therapeutic target of TLB. Innovation: TLB is a naturally occurring Sirt3 agonist with potent neuroprotective effects via regulation of TLR4/nuclear factor-kappa B and Nrf2/Kelch-like ECH-associated protein 1 (Keap-1) signaling pathways both in vivo and in vitro. Conclusion: Our findings indicate that TLB protects against cerebral I/R-induced neuroinflammation and oxidative injury through the regulation of neuroinflammatory and oxidative responses via TLR4, Nrf2, and Sirt3, suggesting that TLB might be a promising Sirt3 agonist against ischemic stroke.

Full text of DOI:10.1089/ars.2019.7825

Antioxidants and Redox Signaling
Mary Ann Liebert, Inc.
Page 1 of 60
©
DOI: 10.1089/ars.2019.7825
1
Neuroprotective effects of trilobatin, a novel naturally-occurring
Sirt3 agonist from Lithocarpus polystachyus Rehd., mitigates
cerebral ischemia/reperfusion injury: Involvement of TLR4/NF-κB
and Nrf2/ Keap-1 signaling
Abbreviated title：A Sirt3 agonist cerebral ischemia injury
1,2,3
1,2,3
1,2,3
1,2,3
2,3
1,2,3
Jianmei Gao , Nana Chen , Na Li , Fan Xu , Wei Wang ,Yaying Lei , Jingshan
2, 3
1,2,3*
Shi , Qihai Gong
1
Department of Clinical Pharmacotherapeutics, School of Pharmacy, Zunyi Medical
University, Zunyi 563000, China
2
Key Laboratory of Basic Pharmacology of Ministry of Education and Joint International
Research Laboratory of Ethnomedicine of Ministry of Education, Zunyi Medical University,
Zunyi 563000, China
3
Key Laboratory of Basic Pharmacology of Guizhou Province, Zunyi Medical University,
Zunyi, Guizhou, 563000
Corresponding author: Prof. Qihai Gong, Ph.D.
6
Xuefu West Road, Zunyi City, Guizhou Province, 563000, P. R. China
Tel: +86-851-286-435-75
Fax: +86-851-286-435-75
E-mail: gqh@zmu.edu.cn

Page 2 of 60
2
Abstract
Aims: Neuroinflammation and oxidative stress are deemed the prime causes of brain
injury after cerebral ischemia-reperfusion (I/R). Since Sirt3 pathway plays an imperative
role in protecting against neuroinflammation and oxidative stress and has been verified as
a target to treat ischemia stroke. Therefore, we attempted to seek novel Sirt3 agonist and
explore its underlying mechanism for stroke treatment both in vivo and in vitro.
Results: Trilobatin (TLB) not only dramatically suppressed neuroinflammation and
oxidative stress injury after middle cerebral artery occlusion (MCAO) in rats, but also
effectively mitigated oxygen and glucose deprivation/reoxygenation (OGD/R) injury in
primary cultured astrocytes. These beneficial effects along with the reduced pro-
inflammatory cytokines via suppressing TLR4 signaling pathway, as well as lessened
oxidative injury via activating Nrf2 signaling pathways, in keeping with the findings in vivo.
Intriguingly, the TLB-mediated neuroprotection on cerebral I/R injury was modulated by
reciprocity between TLR4-mediated neuroinflammatory responses and Nrf2 antioxidant
responses as evidenced by molecular docking and silencing TLR4 and Nrf2, respectively.
Most importantly, TLB not only directly bond to Sirt3, but also increased Sirt3 expression
and activity, indicating that Sirt3 might be a promising therapeutic target of TLB.
Innovation: TLB is a naturally-occurring Sirt3 agonist with potent neuroprotective effects
via regulation of TLR4/NF-κB and Nrf2/Keap-1 signaling pathways both in vivo and in vitro,.
Conclusion: Our findings indicate that TLB protects against cerebral I/R-induced
neuroinflammation and oxidative injury through the regulation of neuroinflammatory and
oxidative responses via TLR4, Nrf2 and Sirt3, suggesting that TLB might be a promising
Sirt3 agonist against ischemic stroke.
Keywords: trilobatin; cerebral ischemia; reperfusion; TLR4; Nrf2; Sirt3

Page 3 of 60
3
Introduction
Ischemic stroke is a major cause of death and disability worldwide. The outcomes of
ischemic stroke damage are far-reaching, generating immense burden to both the family
and society(5). To date, ideal drugs or strategies for ischemic stroke are still unavailable.
Present strategies for treatment of stroke include recanalization by means of
pharmacologic or mechanical thrombolysis and neuroprotective agents(2). So far,
recombinant tissue plasminogen activator (rtPA), known as a thrombolytic agent, is the
only FDA-approved medical therapy for ischemic stroke(40). However, clinical application
of rtPA is limited because of its rigid narrow therapeutic time window, and especially
latent ischemia/reperfusion (I/R) injury, which plays a crucial role in aggravating
succeeding ischemic brain lesion(35,42). Therefore, it is of significance to elucidate the
mechanisms of cerebral I/R injury and develop more effective strategies or agents to treat
cerebral I/R injury.
Emerging evidence suggests that cerebral I/R injury causes a sophisticated cascade of
pathophysiologic events, especially neuroinflammation and oxidative stress, which
ultimately lead to neuronal injury, even demise(12,17). Toll-like receptors (TLRs) are
known as a trans-membrane pattern-recognition receptor family with crucial roles in the
mediation of inflammatory responses(49). Up to now, thirteen TLRs have been identified in
mammals, of which TLR4 is expressed on the cell surface, and is identified as the most
important pattern recognition receptor involved in cerebral I/R injury. TLR4 is activated in
response to cerebral I/R injury, and subsequently directly promote its pivotal adapter
protein myeloid differentiation factor 88 (MyD88) recruitment, then results to the
activation of downstream nuclear factor-kappa B (NF-κB) and thereby releases
proinflammatory cytokines to aggravate cerebral I/R injury in turn(18,20). Nuclear factor
erythroid 2-related factor 2 (Nrf2) is a transcriptional factor involved in anti-oxidative
stress insults(37). Under quiescent condition, Nrf2 localizes to the cytoplasm and is
restrained by Kelch-like ECH-associated protein 1 (Keap-1)(27). When cells were subjected
to oxidative stress or any other deleterious insults such as cerebral I/R injury, Nrf2
liberates from Keap1 and translocates into the nucleus and then binds to antioxidant
response element (ARE), including hemeoxygenase-1(HO1), NAD(P)H-quinone

Page 4 of 60
4
oxidoreductase 1 (NQO1) and glutathione peroxidase (GSH-Px)(1,33). Intriguingly,
accumulating evidence indicates that the Nrf2 signaling pathway is activated during
inflammation, which induces the downstream genes of Nrf2 to restrain the inflammatory
response, and the Nrf2 signaling pathway cross talks with TLR4 and its downstream genes
(e.g NF-κB)(41); However, whether reciprocity between TLR4 and Nrf2 is involved in the
cerebral I/R injury remains still a mystery.
Trilobatin (TLB),a major active constituent of Lithocarpus polystachyus Rehd., is a folk
medicine which is used to prophylaxis and treatment of multiple diseases in China with a
long history(46). Of note, previous report demonstrated that TLB exerted attenuation
effect on LPS-induced inflammatory response through hindering the NF-κB signaling
pathway(6). Moreover, our previous study revealed that TLB also presented apparent
potential neuroprotective effect on hydrogen peroxide-induced oxidative injury in a
neuron-like PC12 cell via regulating Nrf2/silent mating-type information regulation 2
homolog 3 (Sirt3) signaling pathway(8). Whereas, whether TLB can protect against cerebral
I/R injury and its underlying mechanisms are associated with reciprocity between the Nrf2
and TLR4 pathways remains still unclear.
Consequently, the focus of this study was designed to investigate whether TLB
treatment can elicit neuroprotection against MCAO-induced cerebral I/R injury in rats and
oxygen and glucose deprivation/reoxygenation (OGD/R)-induced injury in primary cultured
rat astrocytes through mediating TLR4/Nrf2 signaling pathway.
Results
TLB suppressed MCAO-induced injury through decreasing neurologic deficits, infarct
volume and cerebral edema
The inhibitory effects of TLB on cerebral I/R-induced outcome 3 d after 2 h MCAO
were measured by neurologic deficits, cerebral edema and infarct volume in rats. Firstly,
regional cerebral blood flow (rCBF) was reduced to below 20% and recoverd to greater
than 80% of baseline, indicating that a successful MCAO model was accepted (Fig. 1A, B).
The results showed that TLB ameliorated cerebral I/R-induced neurological deficits (Fig.
1
C), cerebral edema (Fig. 1D) and reduced infarct volume (Fig. 1E, F), respectively. These

Page 5 of 60
5
findings suggested that TLB dose-dependently inhibited cerebral I/R-induced injury.
Furthermore, the time window for TLB treatment after MCAO was explored, due to a
appropriate time window for treating cerebral I/R injury was crucial for the therapeutic
effects of TLB in the clinic. TLB was administrated at 1, 2, 3, 4 and 6 h after MCAO and
neurological deficits, cerebral edema and infarct volume were also measured in rats,
respectively. The results showed that TLB (20 mg/kg) significantly reduced the neurologic
deficits, cerebral edema and infarct volume after MCAO within 4 h, while, these effects
were decreased at 6 h after MCAO (Fig. 2).
TLB restored long-term neurological functions at 28 days after MCAO in rats
To further explore the effect of TLB on long-term (28 days after MCAO) neurological
function recovery, an attery of sensorimotor and cognitive tests including neurological
scores, rotarod, adhesive tape removal, Y-Maze and NOR tests were performed at 28 days
after MCAO in rats. The results showed that the neurological scores of rats treated with
TLB were markedly lower than those of rats after MCAO at day 28 (Fig. 3A). TLB obviously
improved the performance in rotarod and adhesive tape removal from 28 days after
MCAO in rats (Fig. 3B, C). Additionally, TLB conspicuously increased percentage of correct
spontaneous alternations in Y maze test (Fig. 3D) and the discrimination index of novel
from familiar objects in the NOR test (Fig. 3E). Collectively, these findings indicated that
TLB effectively restored long-term neurological functions after MCAO in rats.
Microarray Data Analyses
Differentially expressed genes (DEGs) were determined in the condition of both P-
value < 0.05 and FC > 1.5. As shown in Fig. 4A, 1014 DEGs were identified in MCAO versus
sham groups, while 363 DEGs were identified in TLB versus MCAO groups. Furthermore, 52
out of the 415 DEGs responding to TLB treatment were associated with DEGs elicited by
cerebral I/R injury as evidenced by Venn diagram. Moreover, Hierarchical clustering
analysis showed that the expression profiles of the 415 DEGs in Sham and TLB groups were
significantly different to that of the MCAO group (Fig. 4B). Moreover, the KEGG pathway
and enrichment of GO terms for the selected 415 DEGs involved were depicted. We
revealed that the top three pathways enriched from KEGG database were positive

Page 6 of 60
6
regulation of MAPK cascade, negative regulation of endopeptidase activity and negative
regulation of inflammatory response (Fig. 4C). As shown in Fig. 4D-G, top 10 biological
process, cellular component and molecular functions for up-regulated or down-regulated
DEGs were listed in bubble plots, which showed that beneficial effect such as positive
regulation of MAPK cascade, positive regulation of angiogenesis and negative regulation of
apoptotic process were included. As present in Fig. 4H, protein-protein interactions in the
4
15 DEGs identified after TLB treatment were depicted. There were 742 interaction pairs
among in these DEGs-encoded proteins, the size of circles represents that the degree of
protein connection to others. Notably, TLR4 was involved in 19 interactions, suggesting
that TLR4 might be the dominant relevant signaling molecule. Furthermore, we also found
that TLR4 interacted with Nrf2 and Nrf2 interacted with Sirt3 using ClusterONE analysis. Of
note, to validate the microarray results, the expressions of TLR4, Nrf2 and Sirt3 were
determined by qRT-PCR. The results showed that TLB down-regulated TLR4, up-regulated
Nrf2 and Sirt3 than those of MCAO group in accordance with the array data (Fig. 4I-K).
TLB reduced MCAO-induced astrocyte and microglial activation
The effect of TLB on glial activation was measured by immunohistochemical (IHC)
staining using anti-glial fibrillary acidic protein (GFAP) antibody to mark astrocyte and anti-
Iba-1 antibody to mark microglia. The results showed that the number of GFAP-positive
cells and Iba-1-positive cells were markedly elevated after MCAO than those of sham
group; However, the increased number of activated astrocyte and microglia were
attenuated by TLB (Fig. 5).
TLB inhibited inflammatory cytokine and TLR4 signaling pathway after MCAO
Inflammatory cytokine levels and inducible nitric oxide synthase (iNOS), TLR4, MyD88,
TRAF6, NF-κBp65 in the brain tissues of rats after MCAO were detected using according
inflammatory cytokine ELISA kits and Western blot, respectively. The results indicated that
the levels of IL-1β, IL-6, TNF-α and iNOS expression were markedly augmented in MCAO
group than those of sham group, as well as expressions of TLR4, MyD88, TRAF6 and
phosphorylation level of NF-κBp65. Whereas, TLB reduced IL-1β, IL-6, TNF-α and iNOS

Page 7 of 60
7
expression (Fig. 6A-D), as well as expression of TLR4, MyD88, TRAF6 and
phosphorylation level of NF-κBp65 (Fig. 6E-I).
TLB attenuated oxidative injury and activated Nrf2/Sirt3 signaling pathway after MCAO
Reactive oxygen species (ROS) and malondialdehyde (MDA) levels, superoxide
dismutase (SOD) and glutathione peroxidase (GSH-Px) activities were detected using the
according kits, and Nrf2, Keap1, HO-1, NQO1 expressions were detected by Western blot.
The results indicated that ROS and MDA levels were elevated and SOD and GSH-Px
activities were reduced after MCAO; However, these change were suppressed by TLB (Fig.
7A-D). Furthermore, nuclear Nrf2 level was increased and cytosol Nrf2 level was decreased
after MCAO than those of sham group; However, TLB significantly upregulated Nrf2
expression of the nucleus and decreased it in the cytoplasm than those of MCAO group
(Fig. 7E-G). Notably, TLB not only downregulated the Keap-1 expression and upregulated
NQO-1 and HO-1 expressions than those of MCAO group (Fig. 7E, H-J), but also increased
Sirt3 expression (Fig. 5K).
Prediction of drug targets of TLB against MCAO-induced injury
Additionally, the results further exhibited the strong binding affinity between TLB and
Sirt3 with binding energy of -5.44 kcal/mol. To investigate the activity and structure
relationship of TLB, we synthesized an analog of TLB, which was named as Tr1. TLB and Tr1
have a similar structure, except for the glucose of TLB. Tr1 and Sirt3 with binding energy of
-4.09 kcal/mol, which indicated that Tr1 failed to bound with Sirt3. Thereafter, the
presumptive binding modes and the pocket of amino acid were tested by a molecular
docking, including LYS205, SER152, ARG365, PRO355, HIS354, LEU 173 and ASP172, which
further confirmed that TLB bound to the hydrophobic pocket of Sirt3, consisting with the
results both in vivo and in vitro (Fig. 8). These findings indicated that TLB might directly
bind to Sirt3 to exert its pharmacological activities.

Page 8 of 60
8
TLB protected against OGD/R-induced injury in astrocytes via inhibiting inflammatory
cytokine and TLR4 signaling pathway
To further explore the role of TLB during cerebral I/R, the OGD/R model in primary
cultured rat astrocytes was applied to mimic the cerebral I/R. We first detected the
cytotoxicity of TLB and Tr1, an analog of TLB (Fig. 9A) in astrocytes to determine suitable in
vitro treatment concentrations by a 3-(4,5-dimethythiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT) assay. TLB or Tr1 exerted no effect on astrocytes below 50 μM within 48 h
(Fig. 9B, C). Therefore, 50 μM or less of TLB and Tr1 was used in the following experiments.
Next, the cell viability and cytotoxicity were determined using MTT assay and lactate
dehydrogenase (LDH) assay, respectively. The results showed that TLB protected against
OGD/R-induced astrocyte injury in a concentration-dependent manner, however, 50 μM
Tr1 did not improve the cell viability (Fig. 9D). Moreover, TLB also reduced the amount of
LDH release in a concentration-dependent manner. However, 50 μM Tr1 did not alter the
LDH level (Fig. 9E). Additionally, OGD/R resulted in astrocytes shrink, depletion in cell
numbers, even death. Whereas, TLB reversed these change after OGD/R, as evidenced by
observation of light converted microscopy. However, 50 μM Tr1 did not alter these change
(Fig. 9F). Furthermore, inflammatory cytokine levels and iNOS, TLR4, MyD88, TRAF6, NF-
κBp65 in the astrocytes after OGD/R were detected using according inflammatory cytokine
ELISA kits and Western blot, respectively. The results demonstrated that the levels of IL-1β,
IL-6, TNF-α and iNOS expressions were markedly enhanced in OGD/R group than those of
control group, as well as expressions of TLR4, MyD88, TRAF6 and phosphorylation level of
NF-κBp65. Whereas, TLB mitigated IL-1β, IL-6, TNF-α and iNOS expression (Fig. 9G-J), as
well as expressions of TLR4, MyD88, TRAF6 and phosphorylation level of NF-κBp65 (Fig. 9K-
M).
TLB attenuated oxidative injury and activated Nrf2/Sirt3 signaling pathway after OGD/R
insult in astrocytes
MDA level, SOD, GSH-Px activities, intracellular ROS and mitochondrial superoxide
•−
anions (O
2
) generation were detected using according kits and mito-SOX staining,
respectively, and NRF2, Keap1, HO-1, NQO1 expressions were evaluated by Western blot.

Page 9 of 60
9
The results indicated that the red fluorescence was enhanced after OGD/R than that of
•−
control group, which suggested that OGD/R accelerated O
2
generation in mitochondria.
Moreover, the generation of intracellular ROS and MDA were also elevated and SOD and
GSH-Px activities were reduced after OGD/R; However, these change were restrained by
TLB (Fig. 10A-E). Furthermore, nuclear Nrf2 level was increased and cytosol Nrf2 level was
decreased after OGD/R than that of control group; However, TLB upregulated Nrf2
expression of the nucleus and downregulated it of the cytoplasm (Fig. 10F-H). Notably, TLB
not only downregulated the Keap-1 expression and upregulated NQO-1 and HO-1
expressions than those of OGD/R group (Fig. 10I-K), but also increased Sirt3 expression and
its activity (Fig. 10L, M). Whereas, 50 μM Tr1 did not alter Sirt3 expression and its activity
(Fig. 11A-C).
TLB protected against OGD/R-induced injury in primary cortical neurons
The effect of TLB on OGD/R-induced injury in primary cortical neurons was also
investigated. TLB or Tr1, an analog of TLB, exerted no effect on neurons below 25 μM
within 24 h (Fig. 12A, B). Therefore, 25 μM or less of TLB and Tr1 was used in the following
experiments. Next, the cell viability and cytotoxicity were determined using MTT assay and
lactate dehydrogenase (LDH) assay, respectively. The results showed that TLB
concentration-dependently protected against OGD/R-induced neuronal injury, however,
2
5 μM Tr1 did not improve the cell viability (Fig.12C). Moreover, TLB also reduced the
amount of LDH release in a concentration-dependent manner. However, 25 μM Tr1 did
not alter the LDH level (Fig. 12D). In addition, OGD/R lead to neurons shrink, reduction in
cell numbers, even death. Whereas, TLB reversed these change after OGD/R, as evidenced
by observation of light converted microscopy. However, 25 μM Tr1 did not alter these
change (Fig. 12E).
TLB attenuated oxidative injury and activated Nrf2/Sirt3 signaling pathway after OGD/R
insult in primary cortical neurons
MDA level, SOD, GSH-Px activities, intracellular ROS and mitochondrial superoxide
•−
anions (O
2
) generation were determined using according kits and mito-SOX staining,
respectively, and NRF2, Keap1, HO-1, NQO1 expressions were detected by Western blot.

Page 10 of 60
10
The red fluorescence was augmented after OGD/R than that of control, which suggested
•
−
that OGD/R accelerated O
2
accumulation in mitochondria. Moreover, the generation of
intracellular ROS and MDA were also increased and SOD and GSH-Px activities were
decreased after OGD/R; However, these change were reversed by TLB (Fig. 13A-E).
Furthermore, nuclear Nrf2 level was increased and cytosol Nrf2 level was decreased after
OGD/R than that of control group; However, TLB upregulated Nrf2 expression of the
nucleus and downregulated it of the cytoplasm (Fig. 13F, G). Notably, TLB not only
downregulated the Keap-1 expression and upregulated NQO-1 and HO-1 expressions than
those of OGD/R group (Fig. 13F, H), but also increased Sirt3 expression and its activity (Fig.
3I, J). Whereas, 25 μM Tr1 did not alter Sirt3 expression and its activity (Fig. 11D-F).
1
The effect of TLB in Sirt3-knockout (Sirt3-KO) rats and Sirt3-KO astrocytes or neurons
Sirt3-KO rats and Sirt3-KO astrocytes or neurons were generated using the
CRISPR/Cas9 system resulted in > 80% loss in Sirt3 mRNA levels relative to sham group (Fig.
S1A) or control cells (Fig. S1B). To further confirm that TLB works via binding to and
activating Sirt3, we conditionally deleted Sirt3 both in rats and in neurons and astrocytes
using CRISPR-Cas9 Sirt3 lentivirus. The results showed that Sirt3 KO rats showed more
severe injury after MCAO insulted than that of WT rats, which was keeping with the
previous report(39). Whereas, TLB treatment partly lost its ability to reduction of MCAO-
induced injury in Sirt3-KO rats than those of wild type (WT) rats (Fig. 14A-C). Interestingly,
consistent with the results in vivo, Sirt3-KO neurons or astrocytes also showed more
severe injury after OGD/R insult than that of WT neurons or astrocytes. However, the
protective effects of TLB on OGD/R-induced injury were partially occluded in Sirt3-KO
astrocytes (Fig. 14D, E), and Sirt3-KO neurons (Fig. 14F, G) than those of WT astrocytes or
neurons. These findings highlighted that TLB-mediated protection is, at least partly,
dependent on the the presence of Sirt3, and Sirt3 might be a potential target of TLB, which
was consistent with the findings of molecular docking mentioned above.

Page 11 of 60
11
Reciprocity between TLR4 and Nrf2 was involved in the beneficial effects of TLB on
OGD/R-induced injury
Furthermore, siRNA was applied to verify whether TLB regulated TLR4/Nrf2 signaling
pathway to attenuate OGD/R-induced injury in astrocytes. Firstly, the levels of TLR4 and
Nrf2 in TLR4 and Nrf2 siRNA treated group dramastically decreased than those of
scrambled siRNA transfected group (Fig. S2). Without any OGD/R stimulus, silencing of
TLR4 or Nrf2 exerted no effect on cell viability and cytotoxicity. Whereas, upon OGD/R
stimulus, the cell viability and cytotoxicity of TLR4 siRNA transfected cells were
significantly increased and decreased, respectively, than those of scrambled siRNA
transfected cells; However, the beneficial effect of TLB was markedly promoted by the
TLR4 siRNA. Meanwhile, the cell viability and cytotoxicity of Nrf2 siRNA transfected cells
were significantly reduced and augmented, respectively, than those of scrambled siRNA
transfected cells; However, the beneficial effect of TLB was significantly abolished by the
Nrf2 siRNA (Fig. S3). Next, the effect of TLB on inflammatory cytokine in primary rat
astrocytes were also further examined by measuring the levels of IL-1β, IL-6, TNF-α and
iNOS in the serum of rats with siTLR4 and siNrf2 by ELISA. The results indicated that the
inhibitory effects of TLB in inflammatory cytokines were enhanced by TLR4 siRNA and
partially abolished by Nrf2 siRNA (Fig. S4A-D). Moreover, the inhibitory effects of TLB in
ROS and MDA levels were increased by TLR4 siRNA and almost abolished by Nrf2 siRNA;
while, the elevation effects of TLB in antioxidant enzymes including SOD and GSH-Px were
elevated by TLR4 siRNA and almost abolished by Nrf2 siRNA, respectively (Fig. S4E-H).
Of note, the results further demonstrated that knockdown of Nrf2 enhanced TLR4
expression in response to OGD/R, while, the attenuative effects of TLB on OGD/R-induced
TLR4 increase were almost abolished by Nrf2 siRNA (Fig. 15A, B). Moreover, knockdown of
TLR4 significantly increased Nrf2 expression in the nucleus and decreased it in the
cytoplasm after OGD/R, while, the promotion effects of TLB were also elevated by TLR4
siRNA (Fig. 15C, D). Furthermore, the docking of Nrf2 and TLR4 was carried out by ZDOCK
and RDOCK method as previous study. The results showed that there were strong
interactions between Nrf2 and TLR4 as evidenced by the output values of score of
E_RDOCK (Tab.1). Of note, the results further demonstrated that hydrogen bond and

Page 12 of 60
12
charge interactions were generated through the amino acid residues including HIS3,
ASN409, HIS5, PHE263, PHE429, GLN430, SER9 and HIS7; and Pi interactions with the
residues such as TRP332 and PHE330, TRP332 and PHE313, TRP332 and PHE330, TRP332
and PHE313, PHE408 and HIS431 (Fig. 15E-G). Notably, we further verified the direct
interaction between Nrf2 and TLR4 by surface plasmon resonance (SPR). The results
demonstrated that Nrf2 directly bound to TLR4 in a concentration-dependent manner with
-
9
a KD value of 2.139e M (Fig. 15H). These findings suggested that there existed potent
affinity between Nrf2 with TLR4.
Discussion
The present study, for the first time, discovered that: (1) TLB, a naturally-occurring
Sirt3 agonist, derived from herbal Lithocarpus polystachyus Rehd. exerted neuroprotection
against MCAO-induced injury in rats and OGD/R-induced injury in primary cultured
astrocytes or neurons in vitro; (2) The inhibitory effects of TLB were due to inhibition of
neuroinflammation through suppressing TLR4 signaling pathway, as well as reduction of
oxidative stress injury via activating Nrf2 signaling pathway; (3) Reciprocity between TLR4
and Nrf2 was involved in the neuroprotection of TLB against cerebral I/R injury (Fig.16).
On account of lacking effective prophylaxis and treatment strategies for ischemic
stroke, a mass of pharmacological neuroprotectant have been researched and developed,
unfortunately, though with limitation of clinical success or failure of preclinical due to a
complicated pathological process with manifold mechanisms during cerebral I/R
injury(19,26). Thus, exploring effective neuroprotectant is an extreme clinical demand. In
the current study, we explored the neuroprotection of TLB, a natural small molecule
monomer, against cerebral I/R injury and endeavored to elucidate its underlying
mechanisms. Firstly, our findings demonstrated that TLB effectively protected against
outcomes of cerebral I/R injury as evidenced by reduced MCAO-induced neurological
deficits, cerebral edema and infarction volume and in vivo. Next, of note, TLR4, Nrf2 and
Sirt3 were identified as DEGs as proved by Microarray Data Analyses (Venn diagram,
Hierarchical clustering analysis, GeneMANIA analysis, GO terms) and TLB also reduced
TLR4 mRNA level, increased Nrf2 and Sirt3 mRNA levels in line with the array data as

Page 13 of 60
13
affirmed by qRT-PCR. Therefore, it is reasonable to speculate TLR4-mediating
neuroinflammation and Nrf2/Sirt3-mediating oxidative stress were involved in TLB-
induced neuroprotection against cerebral I/R injury. As noted, neuroinflammation is a
crucial step and a secondary damage mechanism in the cerebral I/R injury(34). When
neuronal cell are challenged by cerebral I/R, neurogliocyte (astrocyte and microglia) are
activated and release pro-inflammatory factors including IL-1β, IL-6, TNF-α, and iNOS
thereby exacerbating brain injury(28). As we expected, MCAO-induced activation of
astrocyte and microglia, and also promoted release of inflammatory cytokine (IL-1β, IL-6,
TNF-α, and iNOS), which are the premier triggers of activate astrocytes in cerebral I/R
injury; whereas, TLB reversed these change after MCAO, suggesting that TLB-mediated
neuroprotection against cerebral I/R injury, at least partially, through inactivation of
neurogliocyte and inhibition of pro-inflammatory factors release. Furthermore, due to
TLR4 is expressed by neurogliocyte and contributes to cerebral I/R-induced inflammatory
injuries(13), the role of TLR4 signaling pathway during the beneficial effects of TLB on
cerebral I/R injury were determined. The results revealed that the expressions of TLR4,
MyD88 and TRAF6 significantly increased after MCAO, which suggested that TLR4-
regulated signaling pathway was activated after cerebral I/R and heightened inflammatory
responses and further aggravated brain injury, in keeping with previous study; However,
TLB inhibited MCAO-elicited these change. Moreover, TLR4 signaling was stimulated by
cerebral I/R injury results in activating the transcription factors NF-κB which are involved in
activation of pro-inflammatory genes and cytokines(30). Our findings also demonstrated
that phosphorylation level of NF-κBp65 was up-regulated by MCAO, while, TLB obviously
repressed phosphorylation level of NF-κBp65 after MCAO. These findings suggested that
the attenuative neuroinflammation effects of TLB on cerebral I/R injury was mainly due to
inactivation of TLR4/MyD88/TRAF6/NF-κB pathway.
Additionally, mounting evidence suggests that oxidative stress, termed as an
imbalance between the antioxidase system and the generation of reactive oxygen species
(ROS), is known as an early event during cerebral I/R injury for that the brain is very
suggestible to oxidative stress(22). Moreover, accumulation of ROS is primary in activating
TLR4-regulated signaling pathways and induced pro-inflammatory factors release in

Page 14 of 60
14
cerebral I/R injury(23). Our results manifested that MCAO-induced increase in ROS
generation, level of MDA, which is derived from lipid peroxidation and considered as a
biomarker of oxidative stress, consistent with theory that cerebral I/R injury can attract a
remarkable amount of MDA generation in the ischemic brain hemisphere(25); also, MCAO-
induced decrease in SOD and GSH-Px activities, which were the anti-oxidant enzymes to
maintain redox homeostasis and affect the inflammatory response(9). Conversely, TLB
reversed these change by MCAO insulted, inferring that TLB-mediated neuroprotection
against cerebral I/R injury, partially, was via suppressing oxidative lesions. Thereafter, our
results further clarified that MCAO promoted Nrf2 translocate from cytoplasm into
nucleus; however, TLB significantly facilitated nuclear translocation of Nrf2. Whereas, the
change of Nrf2 protein expression after MCAO was not consistent with the the results of
the mRNA level of Nrf2, which did not change. The reason might be due to that there is a
more intrinsic and complex dependence between mRNA and protein, such as translation
and transcription of gene expression existed time-space span, and there would be post-
transcriptional processing, degradation of transcriptional products, translation, post-
translational processing and modification after transcription. Furthermore, MCAO also
decreased ARE including HO-1 and NQO1 expressions, while, TLB increased HO-1 and
NQO1 expressions after MCAO, indicating that TLB prompted Nrf2 dissociated from Keap1
and then translocated into the nucleus, thereby activating ARE to protected against
oxidative stress in cerebral I/R injury. Thus, we believed herein, TLB alleviated
neuroinflammatory responses and oxidative stress during cerebral I/R injury. Nevertheless,
what is the potential target of TLB and whether reciprocity between TLR4 and Nrf2 was
involved in the TLB-mediated neuroprotection against on cerebral I/R injury were still
unclear. Therefore, we further explore the uncovered problem in neurons and astrocytes
in vitro. Notably, astrocytes are the most plentiful non-neuronal cell type in the central
nervous system and have been indicated to have intensive resistance to cerebral I/R injury
compared to neurons(10,45); while, susceptibility of neurons to cerebral I/R will be
augmented once astrocyte dysfunction(14). Most importantly, astrocytes were not only
expressed TLR4, but also enriched transcription factor Nrf2(21). Moreover, since in
ischemic stroke, a dramatically decrease in regional cerebral blood flow elicits deprivation
of oxygen and glucose and eventually leads to cerebral injury. Thus, we used OGD/R-

Page 15 of 60
15
induced astrocyte and neuron injury to mimic cerebral I/R injury in vitro to further
investigate the mechanism of TLB-mediated neuroprotection. The results showed that TLB
effectively increased cell viability and decreased cytotoxicity after OGD/R insult both in
neurons and astrocytes, respectively. Furthermore, TLB also restrained pro-inflammatory
factors release as well as inactivated TLR4/MyD88/TRAF6/NF-κB pathway after OGD/R
insult in astrocytes. Meanwhile, some oxygen free radicals and their derivatives are
•
−
accumulated after cerebral I/R injury, including O , which generated in mitochondria. The
2
results indicated that both intracellular and mitochondrial ROS generation exhibited
significant decrease by TLB after OGD/R, as well as activated Nrf2 signaling pathway
accompanied with elevated ARE genes and enzymes both in neurons and astrocytes. Our
findings further verified that TLB effectively attenuated cerebral I/R injury through
inactivation of TLR4-mediated pathway and activation of Nrf2-mediatied pathway to
suppress neuroinflammation and oxidative stress, in keeping with the findings in vivo.
Although multiple studies demonstrated that TLR4 signaling pathway and Nrf2
signaling are all activated during neuroinflammation(11), little is known about whether
reciprocity between TLR4 and Nrf2 was involved in cerebral I/R injury. Interestingly, our
results revealed that knockdown of TLR4 with siRNA evidently partly strengthened the
protective roles of TLB in promoting astrocyte survival, decreasing proinflammatory
cytokines production and ROS generation as well as facilitating Nrf 2 nuclear translocation
in astrocytes after OGD/R insult. Moreover, knockdown of Nrf 2 with siRNA apparently
partly abolished the protective roles of TLB in facilitating astrocyte survival, decreasing
inflammatory cytokines and ROS generation as well as reducing TLR4 protein expression in
astrocytes under OGD/R condition. Thus, it is speculated that there might be a direct
reciprocity between TLR4 and Nrf2. Of particular interest was that, TLR4 directly bound to
Nrf2 as confirmed by ZDOCK and RDOCK as well as SPR. These findings disclosed the effect
of TLB against cerebral I/R injury via reciprocity between the Nrf2 and TLR4 signaling
pathways. Of note, Sirt3 is a mitochondrial nicotinamide adenine dinucleotide-dependent
protein, which localized to the mitochondrial matrix(38). Recently, Sirt3 has been deemed
to not only mediated oxidative stress and mtROS homeostasis, but also orchestrate
suppression of inflammation(15,29). Noteworthily, the results in this study demonstrated

Page 16 of 60
16
that Sirt3 was apparently decreased after MCAO and OGD/R, in line with the theory of
Sirt3 deficiency impairs neurovascular recovery in ischemic stroke(4); whereas, TLB
increased Sirt3 expression both in vivo and in vitro consistent with our previous study(8).
Moreover, TLB also significantly enhanced Sirt3 activity after OGD/R insult. Intriguingly, in
fact, a direct interaction between TLB and Sirt3 was evidenced by molecular docking.
Furthermore, the beneficial effects of TLB on cerebral I/R injury were paritially abolished in
Sirt3 deficiency rats although Sirt3-KO rats showed normal under common conditions,
which indicated that Sirt3 plays a vital role under stress conditions, and further verified
that TLB might be a naturally-occurring Sirt3 agonist. Thus, it is plausible to suppose that
Sirt3 was likely to be a promising therapeutic target of TLB against cerebral I/R injury.
Emerging evidence demonstrates that a mass of neuroprotective agents exhibited valid
effects in vivo experiments, but did not display significant effects in clinical trial due to
these narrow therapeutic window. Therefore, it is necessary to explore the therapeutic
window of neuroprotective agents in vivo experiments. Our findings indicated that TLB
exerted significantly therapeutic effects on cerebral I/R injury in rats within 4 h, although
its beneficial effects partially lost after 6 h, it still effectively attenuated the injury after
MCAO in rats. Thus, whether TLB can induce extended window of ischemic tolerance in the
rat brain will be investigated in our next story. It should be noted that the present study
has evaluated the neuroprotection of TLB against cerebral I/R injury and its preliminary
mechanisms. Whether TLB can cross blood-brain barrier and mitigate cerebral I/R-induced
astrocyte injury further reinforce the resistance of neurons to neuroinflammation and
oxidative injury still need to be in-depth explored. In fact, these problems could be solved
by means pharmacokinetics, BBB-3D model, co-cultured astrocytes and neurons in trans-
well in our next story.
Collectively, our findings afford new insight into that TLB inhibited cerebral I/R-
induced neuroinflammation and oxidative injury via TLR4/Nrf2/Sirt3 signaling pathway.
These findings highlight the feasibility that TLB might be a promising Sirt3 agonist against
ischemic stroke.

Page 17 of 60
17
Innovation
The present study, for the first time, discovers that TLB, a novel naturally-occurring
Sirt3 agonist from Lithocarpus polystachyus Rehd., mitigates cerebral ischemia/reperfusion
injury through regulation of TLR4/NF-κB and Nrf2/ Keap-1 signaling. Hence TLB is a novel
lead toward the development of a neuroprotective agent.
Materials and Methods
Chemicals and Reagents
Trilobatin was from Guangdong Kedi Medical Technology Corporation (purity ≥ 98%).
The syntheses of Tr1 was implemented as described in Fig. 9A. The NMR spectra and MS
data of the TLB and Tr1 were showed in Supplementary Fig.S5-7. 2,3,5-
Triphenyltetrazolium chloride (TTC) and MTT were purchased from Sigma-Aldrich),
Neurobasal™-A Medium, B-27™ Supplement (50X) and MitoSOX Red were obtained from
Invitrogen (Eugene, OR, USA). The lactate dehydrogenase (LDH) Cytotoxicity Assay Kit, IL-
1
β, IL-6, TNF-α, ROS, MDA, SOD, GSH-Px assay kits were obtained from Shanghai Renjie
Bioengineering Institute (Shanghai, China). The primary antibodies used in present study,
including GFAP, Iba-1, iNOS, TLR4, MyD88, TRAF, NF-κBp65, Nrf2, Keap1, HO-1, NQO1,
Sirt3 and Sirt3 activity assay kit (Fluorometric) were purchased from Abcam (Cambridge,
UK). TLR4 siRNA (r) and Nrf2 siRNA (r) were purchased from Santa Cruz Biotech (Santa
Cruz, CA, USA). Lipofectamine™ RNAiMAX transfection reagent and scrambled siRNA were
obtained from Invitrogen (Eugene, OR, USA).
Animal
Male adult Sprague-Dawley (SD) rats (8-10 weeks old, 250-280 g), were purchased
from the Experimental Animal Center of Daping Hospital (Certificate No. SCXK 2014-0011).
All rats were housed in five to six per cage with a 12 h light/dark cycles and provided free
access to standard rodent diet and tap water and kept under temperature (23ꢀ±ꢀ1ꢀ°C) and
humidity (55ꢀ±ꢀ5%)-controlled environment. Randomization was used to allocated animals
to various experimental groups and the data analysis were performed by a blinded
investigator. All animal experimental protocols in the present study were operated

Page 18 of 60
18
according to the Guide for the Care and Use of Laboratory Animals published by the US
National Institutes of Health (National Institutes of Health Publication 85-23, revised
1
996), and were approved by the Experimental Animal Ethics Committee of the Zunyi
Medical University (Guizhou, China).
Induction of focal cerebral ischemia and drug treatments
Focal cerebral ischemia was induced by MCAO as described in previous study(43). In
brief, the rats were anaesthetized with 1% sodium pentobarbital (45 mg/kg, i.p.) and the
right common carotid artery, external carotid artery and internal carotid artery were
separated carefully. Then, a piece of 5/0 monofilament nylon suture with diameter 0.36
mm was inserted into the internal carotid artery through the external carotid artery stump
to occlude the origin of middle cerebral artery. After 2 h, the filament was withdrawn to
establish reperfusion and the rat body temperature was kept at 37 °C during surgery. Laser
Doppler flowmetry (Moor Instruments, UK) was applied to monitor MCAO severity rCBF
during the surgical procedure and at the reperfusion as described previously. A successful
MCAO model was accepted when rCBF lowered to below 20% and recovered to higher
than 80% of baseline. In the first study of TLB treatment on cerebral I/R, the rats were
randomly divided into five groups: sham group, sham + TLB (20 mg/kg) group, MCAO
group, MCAO + TLB (5 mg/kg) group, MCAO + TLB (10 mg/kg) group, MCAO + TLB (20
mg/kg) group. Sham group underwent the same operation as mentioned above except
MCAO. Animals were treated TLB by gavage at doses of 5, 10 and 20 mg/kg at the onset
reperfusion twice a day for 3 days, and the rats of sham and model groups were given
volume-matched saline, instead. The second group was designed to evaluate the time
window of TLB for the treatment of cerebral I/R, TLB was administered at the doses of 20
mg/kg at 1, 2, 3, 4 and 6 h after MCAO. The sham group and MCAO group of rats were
received volume-matched saline. In the third group, to discover the effect of TLB on
functional recovery after MCAO, TLB was administered at the dosage of 5, 10 and 20
mg/kg at the onset reperfusion twice daily for 28 days after MCAO. The sham group and
MCAO group of rats were administrated volume-matched saline. In the final group, to
determine the effects of TLB on the MCAO model, rats were divided into the following four
groups 3 weeks after CRISPR-Cas9 Sirt3 lentivirus microinjection: Wild type (WT) + sham,

Page 19 of 60
19
WT + sham + TLB (20 mg/kg), WT + MCAO, WT + MCAO + TLB (20 mg/kg), Sirt3-knockout
(Sirt3-KO) + sham, Sirt3-KO + sham +TLB (20 mg/kg), Sirt3-KO + MCAO, Sirt3-KO + MCAO +
TLB (20 mg/kg).
Determination of Neurological deficit scores
Neurological injury after MCAO was determined by a neurobehavioral test that was
scored on a five-point scale: grade 0, observable deficit; grade 1, failure to fully extend left
forepaw; grade 2, circling to the left; falling 3, falling to the left; 4, inable to walk
spontaneously accompanied with a depressed level of consciousness. Neurobehavioral
test was carried out 3 days after MACO in a double-blind manner as described
previously(16).
Assessment for long term functional recovery after cerebral I/R injury in rats
Neurological deficit scores were determined after day 28 as mentioned above. In
addition, rats were trained in the sensorimotor tests for successive 3 days before MCAO.
Rotorod and adhesive tape removal tests were performed as described in previous
reports(31,48). The rotarod was applied to determined general fitness and motor co-
ordination. In brief, rats were placed on an accelerating rotating beam (acceleration from 4
rpm to 40 rpm within 5 min) and a stop-clock was started. The latency for each rat to fall
off the rotarod onto the sensing platform below was registered. Each rat performed three
trials. Moreover, the sensorimotor impariment-induced by MCAO were detected by
adhesive tape removal tests. In brief, a piece of adhesive tape (3 × 1 cm) were applied as
bilateral tactile stimuli occupying the distal-radial area on the wrist of each forelimb in
rats. Then, the rats were placed bace to their cages. The mean latency to remove adhesive
tapes from the forepaws was registered for the lesioned forepaws. The times to detect
and remove the tapes were recorded with a maximum limition of 120 s. Furthermore, the
spatial memory and learning were detected using Y-Maze test as previous study(44). In
brief, three same arms of the maze apparatus (30 cm × 10 cm× 20 cm) were randomly
assigned as the begin arm, novel arm, or other arm. Firstly, the rats were subjected to
investigate the begining arm and the other arm for 8ꢀmin. Thereafter, rats were returned
back in the maze in the same begining arm and allowed to investigate for 5ꢀmin with

Page 20 of 60
20
access to all three arms at liberty. The number of entries and percentage of time spent in
the novel arm that rats detected were recorded. Spontaneous alternation (%) = [(number
of actual alternations)/(number of total arm entries-2)] × 100, was applied as an index to
evaluate spatial memory. Moreover, NOR tasks were use to determine learning and
memory after MCAO as described in previous report(42). In Brief, rats were accustomed to
an open-field box (50 × 50 × 50 cm) at 28 days after MCAO. During the first phase (10 min),
two of the same objects (A1 and A2, green cubes) were placed symmetrically from the
wall. During familiarization phase (10 min), two dissimilar objects (object A2 was replaced
with a novel object B1, a brown box) were placed in the same box for 1 h after the first
trial. The total time of an rat spent detecting each object during two phases was recorded.
The objects discrimination index was caculated using a discrimination index (DI) as
described in previous report(42).
Measurement of infarct volume
After neurological test, the rats were anaesthetized using 1% sodium pentobarbital
and decapitated under anesthesia. Thereafter, the rat brains were promptly removed and
frozen at -20 °C for 15 min. Then, five coronal brain sections of 2 mm thickness were
stained with TTC at 37 °C for 15 min in darkness, and fixed with 4% formaldehyde for 24 h
in the dark. the red color area and the pale gray color area of the slices were evaluated by
Image J software in a double-blind manner as described previously(14).
Microarray Processing and Data Analysis
The total RNA was extracted from the brain tissues of rats in Sham, MCAO + MCAO
group and MCAO + TLB (20 mg/kg) groups using TRIzol buffer on the basis of experimental
protocol, and then quantified them using an Agilent 2100 (Agilent Technologies Co. Ltd.,
Palo Alto, CA, United States) analysis meter. Qualified RNA transcriptome was sequenced
on the platform of BGISEQ-500RS RNA-Seq supported by Beijing Genomics Institute
(Shenzhen, China). The gene expression was calculated using FPKM value of each sample,
gene expression with a fold change (FC) greater than 1.5 and P-value less than 0.05 were
identified as DEGs in this study. OmicShare tools (www.omicshare.com/tools) was used to
3
plotted volcano plot and advanced bubble diagram, Venny 2.1.0 was applied to draw

Page 21 of 60
21
Venn diagram. Protein-protein interactions (PPIs) network was constructed by Cytoscape
3
.6.0 software. DAVID web-based tools (www.david.ncifcrf.gov) was employed to carried
out the functional enrichment analysis, the enrichment of pathways and functional
processing were formed based on KEGG and annotations of Gene Ontology (GO),
respectively. Furthermore, STRING web-based tools (www.string-db.org) was adopted to
analyzed PPIs.
Immunohistochemical (IHC) staining
IHC was used to determined the activation of astrocyte and microglia using GFAP
staining and Iba-1 staining as described previously(3). In brief, the brain sections with
paraffin-embedded (4 μm) were rinsed with 3% H
2
O for 10 min to block endogenous
2
peroxide activity and incubated with 10% goat serum albumin for 30 min to block
nonspecific binding. Then, the sections were incubated with the primary antibodies anti-
GFAP (1:100) and anti-Iba1 (1:200) overnight at 4 °C, after incubation, sections were
incubated with corresponding secondary antibody (anti-goat IgG-HRP, 1:200).
Immunoreactions were visualized using 3,30-diaminobenzidine tetrahydrochloride.
Counterstaining was performed using hematoxylin. Negative control sections were stained
only with secondary antibody to control for possible on specific staining. For the
semiquantitative analysis of the immunohistochemical results, three sections from each
brain, with each section containing three microscopic fields from the ischemic boundary
zone (penumbra) in the cerebral cortex, were digitized under a 40 × objective. The
immunoreactivity of the target proteins was quantified based on the integrated optical
density of immunostaining per field using Image Pro Plus 6.0 software.
Primary rat astrocytes and cortical neurons culture and drug treatment
Primary rat astrocytes culture were obtained as described in previous study(3). In
briefly, the astrocytes were collected and identified using anti-GFAP antibody. The
astrocytes were cultured on 96-well plates or 6-well plates were washed with HBSS, the
culture medium was shifted with a glucose-free DMEM/F12, and then the astrocytes were
cultured for 1.5 h in oxygen-free N
2
/CO (95%/5%) gas. The 1.5-h time point was chosen
2
because greater than 50% cell death occurred. Then, the OGD/R group medium was

Page 22 of 60
22
replaced with standard culture medium, or treated with various concentrations of TLB
(12.5, 25, 50 μM) for another 48 h.
Primary rat cortical neurons were obtained from new born SD rats as mentioned in
previous study(47). In brief, the primary cortical neurons were cultured in 96-well plates or
6
-well plates with poly L-lysine-coated and resuspended in neurobasal medium containing
0% FBS and 2% B27 supplement, then the cells were cultured in a humidified incubator
1
with 5% CO at 37 °C for successive 10 days. Cultures contain > 95% neurons were
2
identified by neuron-specific enolase. Thereafter, neurons were insulted by OGD for 2 h to
mimic ischemic damage in vitro as mentioned in previous report(32). Brief, neurons were
incubated for 2 h in oxygen-free N
2
/CO (95%/5%) gas, and the neurons of control group
2
was cultured in EBSS with 10 mM glucose. Then the OGD/R group medium was replaced
with standard culture medium, or treated with various concentrations of TLB (6.25, 12.5,
2
5, 50 μM) for another 24 h.
Determination of cell viability and neurotoxicity
The primary rat astrocytes or cortical neurons were treated as mentioned above. In
briefly, at the end point of the treatment, each well was cultured with MTT (5 mg/ml) for
another 4 h. Then the medium was discarded and DMSO (150 ml) was added to dissolve
the formazan. whereafter, the absorbance of formazan formation was evaluated by a
microplate reader at 490 nm wavelength. Additionally, in parallel, neurotoxicity was also
detected by a LDH kit as described previously(7). In briefly, at the end point of the
treatment as described above, the supernatants were obtained with centrifuged for 5 min
at 400 × g. The amount of LDH released from astrocytes were lysed in 1% Triton X-100
following to the manufacturer’s introduction. Moreover, cellular morphologic change were
observed by a phase contrast microscopy.
Determination of inflammatory factors
The astrocytes were treated as mentioned above. Briefly, the tissues or cells were
collected and homogenized using 0.1 M PBS (pH 7.4). Thereafter, the tissue homogenates
were centrifuged at 3000 × g for 20 min at 4 °C. Then, levels of inflammatory factors

Page 23 of 60
23
including IL-1β, IL-6 and TNF-α were detected according ELISA kits complied with the
manufacturer's protocol.
Determination of ROS level
The astrocytes or cortical neurons were treated as mentioned above. In brief, the
tissues or cells were collected and homogenized using 0.1 M PBS (pH 7.4). Thereafter, the
tissue homogenates were centrifuged at 3000 × g for 20 min at 4 °C. Then, levels of ROS
were determined using related ELISA kits according to the manufacturer's protocol.
•−
Additionally, the mitochondrial O
2
generation was measured by MitoSOX Red staining, a
•−
peculiar fluorescent probe for detecting mitochondrial O
2
generation (7). In brief, after
astrocytes or cortical neurons were treated as mentioned above and then were washed
with balanced salt solution and stained with 5 μM MitoSOX Red in the dark at 37 °C for 20
min. Then, the astrocytes were observed using fluorescence microscopy (Olympus IX73;
Olympus, Tokyo, Japan) with excitation/emission (510/580 nm) filters. ROS level was
quantified by the Image Pro Plus software.
Measurement of SOD and GSH-Px activities
The astrocytes or cortical neurons were treated as mentioned above. Briefly, the
tissues or cells were collected and homogenized using 0.1 M PBS (pH 7.4). Thereafter, the
tissue homogenates were centrifuged at 3000 × g for 20 min at 4 °C. Then, activities of SOD
and GSH-Px were determined using related ELISA kits according to the manufacturer's
protocol.
Quantitative real-time PCR (qRT-PCR)
Total RNA was obtained with the Trizol Reagent, which was reverse transcribed to
cDNA with the PrimeScript™ RT Reagent Kit. The CFX96 real-time PCR detection system
(Bio-Rad Laboratories Ltd, Hertfordshire, UK) was used to perform qRT-PCR. The specific
primers and according sequences were displayed as follows: GAPDH, forward 5’-
AACGACCCCTTCATTGACCT-3’ and reverse 5’- CCCCATTTGATGTTAGCGGG -3’; Nrf2,
forward 5’- GTTCAGTCGGTGCTTTGACA -3’ and reverse 5’- CTCTGATGTGCGTCTCTCCA -3’;
TLR4, forward 5’- CTGGGTGAGAAAGCTGGTAA -3’ and reverse 5’-

Page 24 of 60
24
AGCCTTCCTGGATGATGTTGG -3’; Sirt3, forward 5’- TACTTCCTTCGGCTGCTTCA -3’ and
reverse 5’- AAGGCGAAATCAGCCACA -3’. In the reaction, 1 μl cDNA of each sample was
mixed with SYBR®GREEN PCR Master Mix according to the manufacture’s protocol. The
PCR conditions were listed as follow: 30 s at 95 °C, then 40 cycles at 95 °C for 5 s, followed
by 56 °C for 30 s. Results were normalized to GAPDH mRNA level and presented as the fold
–ΔΔCt
change (2
).
Western blot analysis
In vivo, at 3 day after reperfusion, the rats were sacrificed 1% sodium pentobarbital
after treatment with TLB, and then the ischemic penumbra was collected as described
previously(24). In vitro, primary rat astrocytes or cortical neurons were washed three
times with ice-cold PBS and then for following protein extraction after treatment with TLB
as above-mentioned. Protein extracts from nuclear and cytosolic fraction were extracted
by a nuclear extraction kit according to the manufacturer introduction, and the protein
concentration was measured using BCA assay. The brain tissues and primary rat astrocytes
or cortical neurons were homogenized with RIPA buffer. Thereafter, the lysates were
normalized to equal amounts of protein, and 10 μg protein from tissue lysates or 20 μg
protein of cell lysates were divided in sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (10%) and transferred to a nitrocellulose membrane, with blocked with 5%
nonfat milk in TBST for 1 h at room temperature. Then, the membranes were incubated
with primary antibodies including iNOS (1:1000), TLR4 (1:1000), MyD88 (1:1000), TRAF6
(1:1000), p-NF-κBp65 (1:1000), NF-κBp65 (1:1000), Nrf2 (1:1000), Keap1 (1:1000), HO-1
(1:1000), NQO1 (1:1000) and Sirt3 (1:1000) overnight at 4°C. whereafter, proteins were
determined with according species-specific HRP-conjugated secondary antibodies for 2 h
at room temperature. Representative bands then were visualized by ECL Western blot
detection reagents and Image J software was applied to quantify the band optical
intensity.
Transient silencing by small interfering RNAs (siRNAs)
Transfection was implemented when the primary rat astrocytes were achieved to 70–
8
0% confluence in 96-well or six-well plates. The Nrf2-targeted siRNA or TLR4-targeted

Page 25 of 60
25
siRNA was diluted with Opti-MEM and balanced for 15 min at room temperature. Then
astrocytes were transfected with Nrf2 siRNA, TLR4 siRNA or scrambled siRNA by
lipofectamine™ RNAiMAX transfection reagent following the manufacturer’s protocol. The
knockdown of endogenous Nrf2 or TLR4 with siRNA was verified using qRT-PCR and
Western blot, respectively. After transfected for 24 h, the transfected astrocytes were
exposed to OGD/R and treated with TLB as described above. Thereafter, cell viability, LDH
release, inflammatory factors, ROS, MDA, anti-oxidant enzymes, expression of TLR4, levels
of cytoplasmic Nrf2 and nuclear Nrf2 were also detected.
Molecular docking analysis
Molecular docking analysis between TLB and Sirt3 were performed using Autodock
4
.2 and Autodock Tools (ADT). The human X-ray crystal structure of Sirt3 (PDB ID: 3GLS)
was obtained from the Protein Data Bank (PDB) archives and used as target for molecular
docking. The molecular docking results were written as a pose viewer file, and the protein-
ligand complex interactions were studied using the PyMOL molecular graphics system. The
interactions between TLR4 (PDB ID: 2Z63) and Nrf2 (PDB ID: 2LZL) were detected using
ZDOCK and RDOCK, which was an widespread admissive method to implement detailed
prediction of protein-protein docking. Subsequently, the poses of predicted protein were
scored from ZDOCK, then RDOCK was used to refine and re-rank the scores. Thereafter,
the binding affinity of TLR4 and Nrf2 was predicted by E_RDOCK, which was used as the
default scoring function. At last, the 18 poses of docked conformations that had lower
E_RDOCK were chose for subsequent analysis.
SPR
Biacore X100 instrument (GE Healthcare, Uppsala, Sweden) with Biacore X100 and
sensor chip CM5 (GE, BR-1003-99) was performed to further confirm the interaction
between TLR4 and Nrf2(36). Briefly, TLR4 (Abcam, ab233665) was dissolved in 10 mM
sodium acetate (PH 5.0) at a concentration of 20 μg/mL, and absorbed as immobilization.
Nrf2 (Abcam, ab132356) was diluted at 10, 20, 40, 80 and 160 μM in HBS-EP buffer. The
protein interaction time was set to 120 s and 300 s for dissociation. Glycine-HCl (PH 2.0)
(GE Healthcare, BR-1003-55) was applied for regeneration. The Biacore evaluation 3.1

Page 26 of 60
26
analysis software (GE Healthcare) was used to analyzed the data for evaluating binding
affinity between TLR4 and Nrf2. Equilibrium dissociation constants (K ) was determined by
D
global fitting of the kinetic data from various concentrations of Nrf2 with a 1:1 Langmuir
binding model.
Measurement of Sirt3 activity
The SIRT3 activity was determined using Sirt3 activity assay kit. Briefly, the astrocytes
or neurons were treated with TLB or Tr1 as described above. Then the samples were
added double-distilled water, Sirt3 assay buffer, flour-substrate peptide and NAD, then
developer to each well of the microtiter plate and mix well and initiate reactions by adding
5
µL of enzyme sample or buffer of enzyme sample or recombinant Sirt3 to each well and
mixing thoroughly at room temperature according to the manufacturer's introduction.
Thereafter, Sirt3 activity was determined by kinetic measurements at 2 min intervals by a
microtiter plate fluorometer with excitation at 340-360 nm and emission at 440-460 nm
for 30 min.
Generation of Sirt3-KO rats and Sirt3-KO primary rat astrocytes or cortical neurons
Sirt3-KO rats and Sirt3-KO primary rat astrocytes or cortical neurons were generated
by a CRISPR-Cas9 system. The lentivirus-based CRISPR/Cas9 KO plasmid, pHBLV-U6-gRNA-
EF1-CAS9-PURO, with the Sirt3 gRNA sequences 5’-TGGTAGTCATGCGTGTTGGG-3’
(forward) and 5’-CTCAGAACCCAGAAGGTGTG-3’ (reverse) primers. Promoter U6 drived
gRNA and CMV drived Cas9. In brief, the lateral ventricles of 6-week-old SD rats were
microinjected with either a lentiviral-packed CRISPR-Cas9 Sirt3 (5 μl, MOI of 100) (Hanbio
Biotechnology Co., Ltd., Shanghai, China) at a constant rate (1 μl/min) using a syringe
pump system at the subsequent microinjection coordinates: 1.0 mm behind the bregma
and 1.0 mm lateral from the sagittal midline, at a depth of 3.5 mm from the skull surface. 3
weeks after lentivirus microinjection, the rats were treated with or without TLB for 3 days
after MCAO. Thereafter, the neurological function and infarct volume were determined
using five-point scale and TTC staining, respectively. Moreover, primary rat astrocytes or
cortical neurons were transfected with lentiviral-packed CRISPR-Cas9 Sirt3 at an MOI of
1
00 for 48 h. Thereafter, the knockout of endogenous Sirt3 by CRISPR-Cas9 Sirt3 lentivirus

Page 27 of 60
27
were confirmed by RT-PCR and Western blot, respectively. Then, the transfected cells were
cultured and treated with or without TLB after OGD/R insults for further analysis.
Statistical analysis
Data are expressed as mean ± SD, and ‘n’ represents the number of independent
experiments and not replicates. GraphPad Prism 5 (GraphPad Software Inc., San Diego, CA,
USA) was used to analyze all data. Some data were normalized to control for unwanted
sources of variation as follows. Firstly, the data was normalized to bring all of the variation
into proportion with one another after deleting any outliers in various groups. Then, the
coefficients associated with each variable will scale properly to adjust for the disparity in
the variable sizes. The cell viability of control group was deemed to be 100%, and the cell
viability of TLB-treated group was displayed as a percentage of the control group. The
iNOS, TLR4, MyD88, TRAF6, Keap1, HO-1 and NQO1expression were normalized to β-actin,
and Sirt3 expression was normalized to GAPDH. Additionally, cytoplasmic Nrf2 level was
normalized to cytoplasmic β-actin and nuclear Nrf2 level was normalized to nuclear PCNA.
The protein expressions or levels of TLB-treated group were presented as fold change of
the sham or control group, which expression was set to 1. Two or multiple groups were
compared using Student's unpaired t-test or one-way ANOVA followed by Bonferroni post
hoc test with F at P < 0.05 and no significant variance inhomogeneity. P < 0.05 was
considered statistically significant.
Author Disclosure Statement
The authors declare no conflict of interest.
Acknowledgments
This work was supported by Natural Science Foundation of China (Grant No.
8
1760727), National key R & D plan for Research on modernization of Traditional Chinese
Medicine (Grant No. 2017YFC1702005), Post subsidy project of State key R & D plan in
social development field (Grant No. SQ2017YFC170204-05), the "hundred" level of high-
level innovative talents in Guizhou Province (Grant No. QKHRCPT 20165684), Program for
Changjiang Scholars and Innovative Research Team in University, China (Grant No.

Page 28 of 60
28
IRT_17R113), and Program for Outstanding Youth of Zunyi Medical University (Grant No.
5zy-002).
1
List of Abbreviations
ARE, antioxidant response element; DCF-DA, 2’,7’-dichlorofluorescin diacetate; DMEM,
Dulbecco’s Modified Eagle Medium; DEGs, differentially expressed genes; GSH-Px,
glutathione peroxidase; GFAP, glial fibrillary acidic protein; HO-1, heme oxygenase-1; IHC,
immunohistochemical; I/R, ischemia-reperfusion; IL-1β, interleukin-1β; IL-6, interleukin-6;
Keap-1, Kelch-like ECH-associated protein 1; LDH, lactate dehydrogenase; MCAO, Middle
cerebral artery occlusion; MDA, malondialdehyde; MTT, 3-(4,5-22dimethylthiazol-2-yl)-2,5-
diphenyltetrazolium bromide; MyD88, myeloid differentiation factor 88; Nrf2, nuclear
factor erythroid 2-related factor 2; NQO1, NAD(P)H-quinone oxidoreductase 1; NF-κB,
nuclear factor-kappa B; OGD/R, oxygen–glucose deprivation followed by reperfusion; qRT-
PCR, quantitative real-time reverse transcriptase-polymerase chain reaction; rtPA,
recombinant tissue plasminogen activator; ROS, reactive oxygen species; SOD, superoxide
dismutase; SD, Sprague-Dawley; Sirt3, silent mating-type information regulation 2
homolog 3; TLB, trilobatin; TTC, 2,3,5-triphenyltetrazolium chloride; TLR4, Toll-like
receptor4; TNF-α, tumor necrosis factor α.

Page 29 of 60
29
References
1
.
.
.
Alfieri A, Srivastava S, Siow RCM, Cash D, Modo M, Duchen MR, Fraser PA, Williams
SCR, Mann GE. Sulforaphane preconditioning of the Nrf2/HO-1 defense pathway
protects the cerebral vasculature against blood-brain barrier disruption and
neurological deficits in stroke. Free Radic Biol Med 65: 1012-1022, 2013.
2
3
Alim I, Caulfield JT, Chen Y, Swarup V, Geschwind DH, Ivanova E, Seravalli J, Ai Y,
Sansing LH, Ste Marie EJ, Hondal RJ, Mukherjee S, Cave JW, Sagdullaev BT,
Karuppagounder SS, Ratan RR. Selenium Drives a Transcriptional Adaptive Program
to Block Ferroptosis and Treat Stroke. Cell 177: 1262-1279 e25, 2019.
Bachstetter AD, Webster SJ, Goulding DS, Morton JE, Watterson DM, Van Eldik LJ.
Attenuation of traumatic brain injury-induced cognitive impairment in mice by
targeting increased cytokine levels with a small molecule experimental therapeutic.
J Neuroinflammation 12: 69, 2015.
4.
5.
6.
7.
Chen T, Dai SH, Li X, Luo P, Zhu J, Wang YH, Fei Z, Jiang XF. Sirt1-Sirt3 axis regulates
human blood-brain barrier permeability in response to ischemia. Redox Biol 14:
229-236, 2018.
Collaborators GBDCoD. Global, regional, and national age-sex-specific mortality for
82 causes of death in 195 countries and territories, 1980-2017: a systematic
2
analysis for the Global Burden of Disease Study 2017. Lancet 392: 1736-1788, 2018.
Fan X, Zhang Y, Dong H, Wang B, Ji H, Liu X. Trilobatin attenuates the LPS-mediated
inflammatory response by suppressing the NF-kappaB signaling pathway. Food
Chem 166: 609-15, 2015.
Gao J, Deng Y, Yin C, Liu Y, Zhang W, Shi J, Gong Q. Icariside II, a novel
phosphodiesterase 5 inhibitor, protects against H
2
O -induced PC12 cells death by
2
inhibiting mitochondria-mediated autophagy. J Cell Mol Med 21: 375-386, 2017.

Page 30 of 60
30
8.
Gao J, Liu S, Xu F, Liu Y, Lv C, Deng Y, Shi J, Gong Q. Trilobatin Protects Against
Oxidative Injury in Neuronal PC12 Cells Through Regulating Mitochondrial ROS
Homeostasis Mediated by AMPK/Nrf2/Sirt3 Signaling Pathway. Front Mol Neurosci
1: 267, 2018.
1
9
1
1
.
Gupta S, Gupta YK. Combination of Zizyphus jujuba and silymarin showed better
neuroprotective effect as compared to single agent in MCAo-induced focal cerebral
ischemia in rats. J Ethnopharmacol 197: 118-127, 2017.
0.
1.
Hind WH, Tufarelli C, Neophytou M, Anderson SI, England TJ, O'Sullivan SE.
Endocannabinoids modulate human blood-brain barrier permeability in vitro. Br J
Pharmacol 172: 3015-27, 2015.
Ikram M, Muhammad T, Rehman SU, Khan A, Jo MG, Ali T, Kim MO. Hesperetin
Confers Neuroprotection by Regulating Nrf2/TLR4/NF-kappaB Signaling in an Abeta
Mouse Model. Mol Neurobiol, 2019.
1
2.
3.
Kahles T, Brandes RP. Which NADPH oxidase isoform is relevant for ischemic
stroke? The case for nox 2. Antioxid Redox Signal 18: 1400-17, 2013.
1
Kuang X, Wang LF, Yu L, Li YJ, Wang YN, He Q, Chen C, Du JR. Ligustilide ameliorates
neuroinflammation and brain injury in focal cerebral ischemia/reperfusion rats:
involvement of inhibition of TLR4/peroxiredoxin 6 signaling. Free Radic Biol Med
7
1: 165-75, 2014.
1
4.
5.
Liu L, Vollmer MK, Ahmad AS, Fernandez VM, Kim H, Dore S. Pretreatment with
Korean red ginseng or dimethyl fumarate attenuates reactive gliosis and confers
sustained neuroprotection against cerebral hypoxic-ischemic damage by an Nrf2-
dependent mechanism. Free Radic Biol Med 131: 98-114, 2019.
1
Liu T, Ma X, Ouyang T, Chen H, Xiao Y, Huang Y, Liu J, Xu M. Efficacy of 5-
aminolevulinic acid-based photodynamic therapy against keloid compromised by
downregulation of SIRT1-SIRT3-SOD2-mROS dependent autophagy pathway. Redox
Biol 20: 195-203, 2019.