mmc2-html.html

Article

Microglial CMPK2 promotes neuroinflammation
and brain injury after ischemic stroke

Xin Guan,

1,5

Sitong Zhu,

1,5

Jinqian Song,

Kui Liu,

Mei Liu,

Luyang Xie,

Yifang Wang,

Jin Wu,

Xiaojun Xu,

and Tao Pang

1,4,6,

State Key Laboratory of Natural Medicines, New Drug Screening Center, Key Laboratory of Drug Quality Control and Pharmacovigilance

(Ministry of Education), China Pharmaceutical University, Nanjing 210009, P.R. China

Department of Neurology, The Second Affiliated Hospital of Nanjing Medical University, Nanjing 210011, P.R. China

Department of Pharmacy, The Fourth Affiliated Hospital, Zhejiang University School of Medicine, Center for Innovative Traditional Chinese

Medicine Target and New Drug Research, International Institutes of Medicine, Zhejiang University, Yiwu, Zhejiang Province 322000, P.R.
China

State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University, Nanjing 210023, P.R. China

These authors contributed equally

Lead contact

*Correspondence:

wujin@njmu.edu.cn

(J.W.),

xiaojunxu@zju.edu.cn

(X.X.),

tpang@cpu.edu.cn

(T.P.)

https://doi.org/10.1016/j.xcrm.2024.101522

SUMMARY

Neuroinflammation plays a significant role in ischemic injury, which can be promoted by oxidized mitochon-
drial DNA (Ox-mtDNA). Cytidine/uridine monophosphate kinase 2 (CMPK2) regulates mtDNA replication, but
its role in neuroinflammation and ischemic injury remains unknown. Here, we report that CMPK2 expression
is upregulated in monocytes/macrophages and microglia post-stroke in humans and mice, respectively. Mi-
croglia/macrophage CMPK2 knockdown using the

Cre

recombination-dependent adeno-associated virus

suppresses the inflammatory responses in the brain, reduces infarcts, and improves neurological outcomes
in ischemic

CR1

Cre/ERT2

mice. Mechanistically, CMPK2 knockdown limits newly synthesized mtDNA and

Ox-mtDNA formation and subsequently blocks NLRP3 inflammasome activation in microglia/macrophages.
Nordihydroguaiaretic acid (NDGA), as a CMPK2 inhibitor, is discovered to reduce neuroinflammation and
ischemic injury in mice and prevent the inflammatory responses in primary human monocytes from ischemic
patients. Thus, these findings identify CMPK2 as a promising therapeutic target for ischemic stroke and other
brain disorders associated with neuroinflammation.

INTRODUCTION

Acute ischemic stroke is a devastating acute cerebrovascular
disorder,

manifesting as fast-onset muscle weakness, mental

confusion, loss of balance or coordination, sudden syncope, pa-
ralysis, and other severe symptoms.

Despite the advancements

in medical technology and surgical instruments, stroke remains
the second leading cause of death globally, accounting for
11.6% of all disease mortality,

and ranks as the third leading

cause of disability-adjusted life years (DALYs), accounting for
approximately 5.7% of all DALYs.

Currently, intravenous throm-

bolysis with recombinant tissue plasminogen activator (r-tPA) is
the only treatment strategy recognized as effective by all regula-
tory agencies.

Nevertheless, intravenous thrombolysis with

r-tPA has strict time window limitations of approximate 4.5 h
as well as medical restrictions,

so less than 3% of ischemic pa-

tients could benefit from it. Although numerous neuroprotective
agents have displayed significant efficacy in animal experiments
for decades, they have failed in clinical trials.

Consequently,

there is an urgent need to identify potential therapeutic targets
and interventions for ischemic stroke.

Multiple pathophysiological mechanisms participate in the

progression of cerebral ischemic stroke, including neuroinflam-
matory response, calcium overload, oxidative stress, and exci-
totoxicity.

Neuroinflammation mediated by microglia exhibits

crucial influences over ischemic injury and is highly correlated
with long-term cognitive and motor function after stroke.

Upon ischemia attack, neuronal death triggers secondary im-
mune responses characterized by microglia hyperactivation
and release of cytokines and chemokines.

Afterward, acti-

vated microglia accumulate around the lesioned area and
peri-infarct border zone, lasting from a few hours to several
weeks.

In microglial mitochondria, the reactive oxygen

species (ROSs) are extensively produced after stroke.

As mitochondrial DNA (mtDNA) possesses limited repair
capacity and is lacking the protection of histone proteins,

mtDNA is vulnerable and easily oxidized, primarily in the form
of 8-hydroxy-2-deoxyguanosine (8-OHdG).

The level of

8-OHdG has been validated to be highly correlated with the
severity of ischemic stroke and to be an independent prog-
nostic factor of the functional outcomes in patients with
stroke.

15–19

However, it is unclear whether inhibition of oxidized

Cell Reports Medicine

, 101522, May 21, 2024

2024 The Authors. Published by Elsevier Inc.

This is an open access article under the CC BY-NC-ND license (

http://creativecommons.org/licenses/by-nc-nd/4.0/

Please cite this article in press as: Guan et al., Microglial CMPK2 promotes neuroinflammation and brain injury after ischemic stroke, Cell Reports Med-
icine (2024), https://doi.org/10.1016/j.xcrm.2024.101522

OPEN ACCESS

mtDNA (Ox-mtDNA) production could be an effective way to
treat ischemic stroke.

Cytidine/uridine monophosphate kinase 2 (CMPK2) is a

rate-limiting enzyme for the synthesis of deoxynucleoside tri-
phosphates to regulate mtDNA replication.

Several reports

have shown that CMPK2 participates in the progression
of diseases, including acute respiratory distress syndrome
(ARDS), brain calcification, spinal cord injury, and coronaviruses
(CoVs).

20–24

mtDNA is an intramitochondrial closed double-

stranded DNA circle that functions as a damage-associated mo-
lecular marker to induce inflammation.

Under normal physi-

ological circumstances, on the one hand, transcription factor A
(TFAM) can promote the packaging and compaction of naked
mtDNA into nucleoids,

thereby minimizing susceptibility to

mitochondrial ROS (mtROS); on the other hand, timely termina-
tion of inflammation evoked by mtDNA can promote tissue repair
and help re-establish homeostasis in the host by clearing patho-
gens.

However, under pathological states, ROSs, pathogen-

associated molecular patterns, and damage-associated molec-
ular patterns (DAMPs) can prompt mitochondrial permeability
transition pore (mPTP) opening and voltage-dependent anion
channel (VDAC) oligomerization in mitochondrial membranes to
facilitate mtDNA release to the cytoplasm.

mtROS does

not directly induce inflammasome activation,

but mtDNA

oxidized by mtROS (Ox-mtDNA) can directly activate the
NLRP3 inflammasome to induce cleaved caspase-1 and mature
interleukin (IL)-1

secretion. The upregulation of CMPK2 expres-

sion could initiate new synthesis of naked and uncompacted
mtDNA in lipopolysaccharide (LPS)-primed immune cells to
trigger inflammasome-mediated inflammation.

Despite

considerable studies highlighting the pivotal roles of CMPK2 in
diseases including ARDS, brain calcification, spinal cord injury,
and CoVs,

20–24

little is yet known about the role of CMPK2 in

cerebral ischemic stroke.

In this study, we investigated the function of CMPK2

in ischemic stroke. We found that adeno-associated virus
(AAV)-mediated microglia/macrophage CMPK2 knockdown in

CR1

Cre/ERT2

mice, or a natural product, nordihydroguaiaretic

acid (NDGA), identified as a specific CMPK2 inhibitor, reduced
brain injury and presented better long-term neurological out-
comes after transient middle cerebral artery occlusion (tMCAO)
treatment. Moreover, microglia/macrophage CMPK2 knock-
down or NDGA treatment inhibited the expression of newly syn-
thesized mtDNA and Ox-mtDNA formation and subsequently
limited NLRP3 inflammasome activation. Our study laid the foun-
dation for CMPK2 as a potential therapeutic target for ischemic
stroke and other neurological disorders associated with
neuroinflammation.

RESULTS

Peripheral blood expression of CMPK2 is upregulated in
patients with stroke and correlated with infarct volume

Gene expression variability in peripheral whole blood between
healthy subjects and patients with ischemic stroke, which has
been recognized as an approach to identify valuable bio-
markers or predictors for the diagnosis or prognosis of
ischemic stroke,

has garnered investigators’ interest for

decades. Currently, the precise role of CMPK2 in the patholog-
ical processes of ischemic stroke remains to be elucidated. We
evaluated

CMPK2

gene expression in peripheral whole blood

and discovered that

CMPK2

gene expression was highly

elevated in patients with ischemic stroke compared with that
in the control subjects (

Figure 1

A). We subsequently analyzed

the relationship between

CMPK

gene expression and infarct

volume or National Institutes of Health Stroke Scale (NIHSS)
score. We observed a positive correlation between

CMPK2

gene expression and infarct volume or NIHSS score, indicating
that CMPK2 may be positively correlated with stroke severity in
patients (

Figures 1

B–1D). Afterward, we evaluated the changes

of mtDNA levels in the peripheral blood of patients with
stroke by examining the expressions of representative mtDNA
markers, including displacement loop (

D-LOOP

COX1

CYTB

, and nDNA (

TERT

The results revealed that the ex-

pressions of

D-LOOP

COX-1

, and

CYTB

were upregulated in

patients with stroke in comparison to control subjects (

Fig-

ure S1

A). Then, we examined the expression of 8-OHdG, the

biomarker of Ox-mtDNA, which is promoted by CMPK2 upre-
gulation.

Immunofluorescence

analyses

showed

that

8-OHdG expression and CMPK2 expression were increased
in peripheral whole blood after stroke onset relative to that of
the control subject (

Figures 1

E–1H). Moreover, we observed

that 8-OHdG expression was obviously increased in CD11b

cells (

Figures S1

B and S1C). These findings demonstrated

that elevated CMPK2 levels in peripheral whole blood corre-
lated positively with stroke severity in human stroke subjects.

CMPK2 is a microglia/macrophage-specific regulator
after ischemic stroke onset

Coincident with the finding in human peripheral blood, we
discovered that CMPK2 protein expression was significantly
elevated at 3 h, peaked at 24 h, and finally returned to baseline
at 7 days in the brain peri-infarct region in the mouse or rat
tMCAO model (

Figures 2

A, 2B,

A, and S2B). Subsequently,

we investigated which cell type would increase CMPK2
expression in the ischemic hemisphere. Immunofluorescence
staining results revealed that CMPK2 expression was obviously
increased in microglia/macrophages, but not astrocytes or neu-
rons, after stroke onset (

Figures 2

C, 2D,

C, and S2D). To

further determine the expression of CMPK2 in microglia/macro-
phages after ischemia attack, we isolated microglia/macro-
phages by magnetic bead sorting (

Figures S2

E and S2F). Also,

we observed a significant increase in CMPK2 expression in mi-
croglia/macrophages at 24 h post-stroke compared with the
sham group (

Figures 2

E and 2F).

Then, to simulate the

in vivo

cerebral ischemia, the oxygen-

glucose deprivation (OGD)-induced cell death model was estab-
lished in primary mouse astrocytes, microglia, and cortical
neurons

in vitro

. Consistent with the findings of

in vivo

experi-

ments, CMPK2 expression was markedly elevated in a time-
dependent manner and peaked at 6 h in primary mouse microglia
(

Figures 2

G and 2H). However, after OGD conditions, primary

neurons displayed slight upregulation in CMPK2 expression,
and primary astrocytes displayed no discernible change in
CMPK2 expression (

Figures 2

I, 2J, and

G–S2J). Overall, these

data demonstrated that CMPK2 expression was predominantly

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

increased in microglia/macrophages in the ischemic brain during
the acute phase of stroke.

Microglia/macrophage CMPK2 knockdown protects
against ischemic brain injury in the mouse tMCAO model

According to the previous literature,

CR1

Cre/ERT2

mice

were microinjected with microglia/macrophage-specific ad-
eno

associated virus (AAV)

CMPK2 short hairpin RNA (AAV-

U6-LoxP-CMV-mCherry-LoxP-CMPK2-shRNA-WPRE-poly(A)
[AAV-shCMPK2]) or the scrambled control AAV (AAV-U6-LoxP-
CMV-mCherry-LoxP-scramble-shRNA-WPRE-poly(A) [AAV-
shCon]) in the right cortex, followed by administration with
tamoxifen for 5 consecutive days (

Figure 3

A). We examined

the transfection

efficiency of mCherry-tagged

AAV and

observed that mCherry was primarily detected in microglia/
macrophages, with little colocalization with neurons or astro-
cytes (

Figures 3

B and

A–S3C). The immunohistochemical

staining and western blotting results showed that CMPK2
expression was obviously downregulated in microglia/macro-
phages in the cerebral cortex of AAV-shCMPK2

CR1

Cre/ERT2

mice (

Figures 3

C, 3D, and

D–S3G). Next, we explored

whether microglial proliferation after stroke onset affected
CMPK2 knockdown efficiency by western blotting. We isolated
primary mouse microglia cells from

CR1

Cre/ERT2

mice at

postnatal days 1–3 and then supplemented them with mouse
recombinant macrophage colony-stimulating factor (MCSF)
protein for 48 h to stimulate the microglial proliferation accord-
ing to previous literature.

The western blotting analysis

Figure 1. Peripheral blood expression of CMPK2 is elevated and positively correlative with infarct volume in human ischemic subjects

(A) Quantification for

CMPK2

gene expression (control = 21, stroke = 31).

(B) Representative magnetic resonance imaging (MRI) images (diffusion-weighted imaging) of patients with stroke with

CMPK2

gene expression (fold change < 10

and fold change > 10).
(C) The correlation analysis was performed between

CMPK2

transcript levels in blood and infarct volume in patients with stroke.

= 20.

(D) The correlation analysis was performed between

CMPK2

transcript levels in blood and NIHSS score in patients with stroke.

= 20.

(E and F) Representative immunofluorescence images and quantification of 8-OHdG expression levels in the peripheral blood of ischemic patients. Scale bar:
20

= 6.

(G and H) Representative immunofluorescence images and quantification of CMPK2 expression levels in the peripheral blood of patients with acute ischemic
stroke. Scale bar: 20

= 6.

Results are represented as means

SD. ***

< 0.001 vs. non-stroke control subjects.

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

indicated that transfection with AAV-shCMPK2 small interfering
RNAs (siRNAs) could effectively reduce the protein expression
of CMPK2 to approximately 50% compared to cells transfected
with AAV-shCon subjected to MCSF stimulation (

Figure S3

H).

Thus, supported by the above findings, we validated that
CMPK2 shRNA-specific targeting to microglia/macrophages
was established in

CR1

Cre/ERT2

mice. Subsequently, we

developed the mouse tMCAO model to seek out the potential
effects of microglial CMPK2 on cerebral ischemic damage.
We initially assessed the regional cerebral blood flow (rCBF)
to exclude the possibilities of variations in cerebral vessel
congestion and eventually verified that there existed no
apparent

differences

the

rCBF

baseline,

during

ischemia, or after reperfusion between the AAV-shCon and
AAV-shCMPK2

CR1

Cre/ERT2

mice (

Figures 3

E and 3F).

Then, after ischemia attack, AAV-shCMPK2

CR1

Cre/ERT2

Figure 2. CMPK2 expression is mainly upre-
gulated in microglia/macrophages of the
brain after ischemia onset

(A and B) Western blotting images and quantification
for CMPK2 protein expression in the peri-infarct re-
gion of sham-operated and ischemic mice at 3 h, 12
h, 1 day, 3 days, and 7 days following operation.

= 6.

(C and D) Representative immunofluorescence im-
ages and quantitative analysis of CMPK2

Iba1

cells

in the peri-infarct region of the ischemic mice. Scale
bar: 50

= 6.

(E and F) Western blotting images and quantification
for CMPK2 protein expression in microglia/macro-
phages isolated from the peri-infarct region with
CD11b

magnetic beads at 24 h following operation.

= 6.

(G and H) Western blotting images and quantification
for CMPK2 expression in primary mouse microglia
exposed to OGD conditions for 0, 2, 4, 6, 9, or 12 h.

= 4.

(I and J) Western blotting images and quantification
for CMPK2 expression in primary mouse microglia,
astrocytes, or cortical neurons exposed to OGD
conditions for 0 and 6 h.

= 6.

Results are represented as means

SD. **

< 0.01

and ***

< 0.001 vs. sham or 0 h group.

mice

displayed

markedly

ameliorated

ischemic volume as determined by 2,3,5-
triphenyltetrazolium chloride (TTC) staining
and

significantly

improved

short-term

neurological outcomes as indicated by
lower scores in the Longa test in compar-
ison to the AAV-shCon

CR1

Cre/ERT2

mice (

Figures 3

G and 3H). Moreover, we

also found that less ischemic volume and
improved neurological functions existed
in the AAV-shCMPK2

CR1

Cre/ERT2

fe-

male mice compared with AAV-shCon

CR1

Cre/ERT2

female

mice

following

ischemic

stroke

(

Figures

and

3J).

Based on the aforementioned finding, we

conclude that microglia/macrophage-specific CMPK2 knock-
down ameliorates ischemic injury in the mouse tMCAO model.

Microglia/macrophage CMPK2 knockdown improves
the long-term functional performances in the mouse
tMCAO model

The prognosis and quality of life of long-term stroke survivors are
correlated with motor function recovery post-ischemic stroke.

To further investigate the effects of CMPK2 on long-term func-
tional outcomes after ischemia onset, we conducted longitudinal
T2-weighted MRI scanning and a series of long-term functional
tests on the AAV-shCon and AAV-shCMPK2

CR1

Cre/ERT2

mice subjected to ischemic stroke (

Figure 4

A). MRI scanning re-

vealed obviously enlarged ischemic lesions in the AAV-shCon

CR1

Cre/ERT2

mice at 14 days after tMCAO, whereas the

AAV-shCMPK2 injection in the

CR1

Cre/ERT2

mice efficiently

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

reduced the infarct volume (

Figures 4

B and 4C). In addition, AAV-

shCMPK2 treatment significantly ameliorated long-term func-
tional deficits in the

CR1

Cre/ERT2

mice, as revealed by the

markedly reduced score in the modified neurological severity
score (mNSS) test, enhanced score in the wire-hanging test,
decreased percentage of forelimb faults in foot missteps, and
increased time in the rotarod test in contrast to the AAV-shCon

CR1

Cre/ERT2

mice 0, 3, 5, 7, 14, 21, and 28 days post-

ischemia (

Figure 4

D). Furthermore, AAV-shCMPK2 administra-

tion substantially enhanced the survival proportions during the
relatively delayed stage (28 days) compared to ischemic AAV-
shCon

CR1

Cre/ERT2

mice (

Figure 4

E). These results high-

lighted that CMPK2 in microglia/macrophages contributed, at
least partly, to ischemic injury and poor long-term functional per-
formances after ischemic stroke.

Microglia/macrophage CMPK2 knockdown promotes
microglial ramification and inhibits Ox-mtDNA-induced
NLRP3 inflammasome activation in the brain of ischemic
mice

Microglia morphology is highly plastic and rapidly changes to
closely monitor and regulate neuronal functions in response to
varying environmental situations, especially ischemic insult.

To seek out the potential mechanism underlying microglial
protection against ischemic stroke, we assessed the influences
of microglia/macrophage CMPK2 silencing on microglial
morphology. We observed that microglial morphology was
more ramified and homeostatic-like in the AAV-shCMPK2

CR1

Cre/ERT2

mice than in the AAV-shCon

CR1

Cre/ERT2

mice at 3 days after ischemia (

Figure 5

A). Particularly, the

morphometric analysis indicated that the AAV-shCMPK2 treat-
ment could increase the process lengths and the number of
branches compared to the AAV-shCon treatment (

Figure 5

B).

In addition, ramified microglial cells were accompanied by less
proliferation ability.

Then, to evaluate whether CMPK2 knock-

down influences microglia proliferation in the mouse tMCAO
model

in vivo

, we conducted immunofluorescence staining by

costaining Ki67 with Iba1, according to previous reports,

and found that no statistical differences were observed in the
numbers of Ki67

Iba1

cells between the AAV-shCon and

AAV-shCMPK2

CR1

Cre/ERT2

mice (

Figure S4

A). To evaluate

the effects of CMPK2 knockdown on microglia proliferation

in vitro

, we established an MCSF-induced proliferation model

in mouse primary microglia cells and applied EdU staining ac-
cording to previous literature.

The immunofluorescence results

of EdU staining indicated that the MCSF-induced microglia pro-
liferation was not influenced by CMPK2 knockdown (

Figure S4

B).

Thus, CMPK2 knockdown did not affect microglia proliferation

in vivo

and

in vitro

. Consequently, these data demonstrated

that CMPK2 knockdown inhibited microglial hyperactivation
following ischemic insult.

We then explored the effects of microglia/macrophage

CMPK2 silencing on 8-OHdG, which is the biomarker of Ox-
mtDNA. As illustrated by

Figures 5

C–5F, the immunofluores-

cence analysis results demonstrated that the 8-OHdG level
was significantly elevated in microglia in the AAV-shCon

CR1

Cre/ERT2

mice, but substantially decreased in the peri-

infarct area of AAV-shCMPK2

CR1

Cre/ERT2

mice, at 24 and

72 h after ischemia.

Ox-mtDNA mediated by CMPK2 could drive NLRP3 inflamma-

some activation,

which eventually promotes uncontrolled

inflammation and death in the neighboring cell.

Compared

to sham-operated

CR1

Cre/ERT2

mice, the protein expressions

of bioactive gasdermin D N-terminal fragment (N-GSDMD [p30]),
cleaved caspase-1 (caspase-1 p20), and mature IL-1

(IL-1

p17) were obviously upregulated in the peri-infarct region after
AAV-shCon administration, which was significantly reduced af-
ter AAV-shCMPK2 treatment in the peri-infarct region of

CR1

Cre/ERT2

mice at 3 days after ischemia (

Figure 5

G). Based

on the aforementioned findings, we conclude that CMPK2 could
promote mtDNA replication to stimulate Ox-mtDNA formation
and subsequently trigger NLRP3 inflammasome activation in mi-
croglia/macrophages after ischemic stroke.

CMPK2 facilitates Ox-mtDNA formation and NLRP3
inflammasome activation in primary mouse microglia
cells

To validate these findings from

in vivo

experiments, we investi-

gated the effects of CMPK2 on Ox-mtDNA formation in primary
mouse microglia cells

in vitro

. Priming with LPSs followed by ATP

stimulation is a well

established model to activate the NLRP3 in-

flammasome in primary mouse microglia cells

41–43

and has been

implicated in a plethora of literature exploring the pathological
mechanisms of microglia in the context of ischemic stroke.

First, we examined the gene and protein expressions of
CMPK2 in microglia stimulated with LPS. As depicted in the

Figures S5

A–S5C, the gene and protein expressions of CMPK2

were significantly increased from 3 to 12 h after LPS stimulation.
Afterward, we confirmed that CMPK2 siRNA potently reduced
CMPK2 protein expression in primary mouse microglia cells rela-
tive to the control group, with or without LPS stimulation
(

Figures S5

D–S5G). Then, we observed significant increases of

cleaved caspase-1, N-GSDMD, and mature IL-1

in the cell su-

pernatant of LPS-primed microglia cells followed by stimulation
with ATP, which were significantly suppressed by CMPK2
knockdown in primary mouse microglia (

Figures S5

H–S5J).

Lactate dehydrogenase (LDH) assay analysis showed that LPS
plus ATP stimulation induced LDH release in the supernatant

(D) Western blotting images for CMPK2 protein expression in microglia/macrophages isolated from the brain of AAV-shCon and AAV-shCMPK2

CR1

Cre/ERT2

mice with CD11b

magnetic beads.

= 6.

(E and F) Regional cerebral blood flow. Results are expressed as means

SD.

= 4.

(G and H) AAV-shCMPK2 administration reduced cerebral infarct volume and promoted neurological functions in

CR1

Cre/ERT2

male mice.

= 12.

(I and J) AAV-shCMPK2 administration toned down cerebral infarct volume and improved neurological function in the

CR1

Cre/ERT2

female mice.

= 8. Results

are represented as means

SD.

***

< 0.001 vs. sham plus AAV-shCon group and

###

< 0.001 vs. tMCAO plus AAV-shCon group.

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

of primary mouse microglia cells, which was remarkably attenu-
ated by CMPK2 knockdown, suggesting the inhibition of CMPK2
on pyroptosis (

Figure S5

K). To mimic cerebral ischemia condi-

tions

in vitro

, we further applied OGD conditions and found

that CMPK2 knockdown inhibited LDH release after OGD stimu-
lation (

Figure S5

L). Collectively, these results indicated that

CMPK2 knockdown inhibited NLRP3 inflammasome activation
and pyroptosis in microglia subjected to LPS plus ATP stimula-
tion or OGD conditions.

Furthermore, we examined the effects of CMPK2 on mtDNA

synthesis and Ox-mtDNA production. CMPK2 knockdown
decreased the expressions of

D-loop

Cox-1

, and

CYTB

, which

were upregulated in primary mouse microglia cells in response
to LPS (

Figures S6

A–S6C). In addition, CMPK2 silencing blocked

Figure

5. Microglia/macrophage

CMPK2

silencing promotes microglial ramification
and disrupts Ox-mtDNA-induced NLRP3 in-
flammasome activation in the brain of
ischemic mice

(A) Imaris-based 3D reconstruction images of mi-
croglia immunofluorescently stained with Iba1.
Scale bar: 15

(B) The process length and the number of branches
of microglia stained with Iba1 in AAV-shCon and
AAV-shCMPK2

CR1

Cre/ERT2

mice at 3 days

after ischemia.

= 10. Results are represented as

means

SD.

(C–F) Representative immunofluorescence images
and

quantification

microglia/macrophages

stained with 8-OHdG and quantitative analysis of
8-OHdG

Iba1

cells in the peri-infarct region of

CR1

Cre/ERT2

mice at 24 and 72 h after ischemia.

Scale bar: 50

= 6.

(G) Western blotting images and quantification for
cleaved caspase-1, mature IL-1

, and N-GSDMD

in the peri-infarct region of AAV-shCon and AAV-
shCMPK2

CR1

Cre/ERT2

mice

days

following ischemic stroke.

= 6. Results are rep-

resented as means

SD.

***

< 0.001 vs. sham plus AAV-shCon group and

###

< 0.001 vs. tMCAO plus AAV-shCon group.

the replication of newly synthesized
mtDNA in microglia subjected to LPS
stimulation or the OGD model (

Figures

D–S6G). Furthermore, we observed a

significant increase in the 8-OHdG level
following LPS plus ATP stimulation or
OGD treatment, which was significantly
suppressed by CMPK2 knockdown in mi-
croglia (

Figures S6

H–S6K).

Then, we assessed the potential effects

of CMPK2 knockdown on mitochondrial
functions,

including

bioenergetics,

biogenesis, and morphology. Firstly, we
evaluated

mitochondrial

bioenergetics

by measuring intracellular ATP levels ac-
cording to previous studies

and

discovered that CMPK2 knockdown, fol-

lowed by LPS plus ATP stimulation, did not influence the intracel-
lular ATP level in primary mouse microglia cells (

Figure S7

A).

Then, we measured the protein expression of PGC-1

and

NRF1, which are widely recognized as pivotal regulators of mito-
chondrial biogenesis.

Western blotting analysis showed

that the protein expressions of PGC-1

and NRF1 were signifi-

cantly reduced in the microglial cells after LPS plus ATP stimula-
tion, and they were also not influenced by CMPK2 knockdown
(

Figures S7

B and S7C). Afterward, we assessed the changes

of mitochondrial morphology outlined by TOMM20 antibodies
according to previous studies.

The immunofluorescence an-

alyses showed that, compared to the controls, the mitochondria
turned out to be obviously more fragmented after LPS plus ATP
stimulation, which could be mitigated by CMPK2 knockdown

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

(

Figures S7

D and S7E). Hence, CMPK2 knockdown, followed by

LPS plus ATP stimulation, could mitigate the mitochondrial frag-
mentation in primary mouse microglia cells.

Several studies have indicated that oxidative stress also en-

hances NLRP3 inflammasome activation

49–53

; thus, we investi-

gated whether the inhibitory effects of CMPK2 knockdown on
NLRP3 inflammasome activation were reliant on ROS genera-
tion. However, we found that CMPK2 knockdown did not change
the level of mtROS production in primary microglial cells after
LPS plus ATP stimulation (

Figures S8

A and S8B). The LPS-

primed primary mouse microglia cells were treated with H

for 6 h and then stimulated with ATP according to previous
studies,

and we found that H

administration failed to

restore NLRP3 inflammasome activation in CMPK2-silenced mi-
croglia (

Figures S8

C and S8D). Furthermore, CMPK2 knock-

down inhibited the colocalization of 8-OHdG with apoptosis-
associated speck-like protein containing a CARD (ASC) specks
in LPS-primed primary mouse microglia followed by ATP stimu-
lation (

Figures S8

E and S8F). Hence, these findings indicated

that CMPK2 knockdown may act by regulating Ox-mtDNA pro-
duction, rather than ROS generation, to inhibit NLRP3 inflamma-
some activation in microglial cells.

Discovery of NDGA as a valid CMPK2 inhibitor with
potent inhibitory effects on Ox-mtDNA formation and
NLRP3 inflammasome activation

in vitro

Based on above findings, we proposed that CMPK2 inhibitors
may serve as promising pharmacological treatments for
ischemic stroke. We screened compounds from compound li-
brary containing natural products and FDA-approved drugs
and discovered that a natural compound, NDGA, was an effec-
tive CMPK2 inhibitor with a half-maximum inhibitory concentra-
tion of 5.958

M (

Figure 6

A). NDGA is a plant lignan isolated

from creosote bush and

Larrea

tridentata, usually discovered

in North America and Mexico.

NDGA possesses pleiotropic

effects, including anticancer, antioxidant, anti-inflammation,
and analgesic activities.

55–57

Furthermore, as indicated by the

surface plasmon resonance (SPR) results, there was a strong
binding between NDGA and CMPK2 protein with a dissociation
constant (K

) value of 0.6

M (

Figure 6

B). Therefore, we inves-

tigated whether NDGA could regulate NLRP3 inflammasome
activation in microglia. As illustrated in

Figures 6

C and

treatment with NDGA significantly inhibited the release of
cleaved caspase-1, N-GSDMD, and mature IL-1

into cell su-

pernatant in LPS-primed microglia followed by ATP stimulation.
Then, we assessed the effects of NDGA on cell death in the LPS
plus ATP or OGD-induced microglia cell injury model and found
that NDGA could effectively tone down LDH release (

Figures 6

and 6E). All the aforementioned suggested that NDGA could
inhibit microglial pyroptosis

in vitro

. Inspired by these findings,

we continued to seek out whether the effects of NDGA were
associated with CMPK2 inhibition. As depicted by

Figure 6

NDGA treatment remarkably downregulated the expressions

D-loop

Cox-1

, and

CYTB

in primary mouse microglia after

LPS treatment. Also, the EdU staining results revealed that
NDGA treatment remarkably reduced the newly synthesized
mtDNA after LPS stimulation (

Figure 6

G and

B). Additionally,

the level of Ox-mtDNA was suppressed by NDGA treatment, as
determined by 8-OHdG immunofluorescence staining in LPS-
primed microglia followed by ATP stimulation (

Figures 6

H and

C). Moreover, NDGA treatment greatly inhibited the newly

synthesized mtDNA and Ox-mtDNA formation in the OGD-
induced microglia injury model (

Figures S9

D–S9G). Then, we

assessed the potential effects of NDGA treatment on mitochon-
drial functions, and the data showed that NDGA treatment did
not change the ATP contents and could not rescue the reduc-
tion of PGC-1

and NRF1 expressions but effectively mitigated

mitochondrial fragmentation in primary mouse microglia cells
followed by LPS plus ATP stimulation (

Figures S10

A–S10E)

Previous literature has shown that NDGA is a valid lipoxyge-

nase (LOX) inhibitor to elicit potent antioxidant actions.

55–57

determine whether 5-LOX was crucial for the inhibitory effects
of NDGA on the activation of the NLRP3 inflammasome, we
knocked down 5-LOX using 5-LOX siRNA in primary mouse mi-
croglia cells (

Figure S11

A). As depicted in

Figures S11

B and

S11C, 5-LOX knockdown did not impact the protein expression
of cleaved caspase-1, N-GSDMD, or mature IL-1

in the cell su-

pernatant of LPS-primed microglia followed by ATP stimulation.
In addition, the inhibitory effects of NDGA on the release of
cleaved caspase-1, N-GSDMD, and mature IL-1

were not

influenced by 5-LOX silencing in microglia after LPS plus ATP
stimulation (

Figures S11

B–S11C). In addition, NDGA treatment

in CMPK2-silenced microglia did not produce more noticeable
inhibitory effects on the release of cleaved caspase-1,
N-GSDMD, and mature IL-1

in the cell supernatant of LPS-

primed microglia followed by ATP stimulation (

Figures S11

and S11E). Furthermore, H

, which further augmented the

release of caspase-1, N-GSDMD, and mature IL-1

, cannot

restore NLRP3 inflammasome activation in NDGA-treated mi-
croglia followed by LPS plus ATP stimulation (

Figures S11

and S11G). These findings support the hypothesis that NDGA
could suppress NLRP3 inflammasome activation primarily by in-
hibiting CMPK2 but not 5-LOX.

CMPK2 inhibitor NDGA elicits potent neuroprotective
activity in the cerebral ischemic mice

Inspired by the above data, we examined the effects of NDGA in
the mouse tMCAO model (

Figures S12

A–S12C). The CBF

assessment indicated that there were no apparent differences
in the CBF at baseline, during ischemia, or after reperfusion
between the NDGA and vehicle groups (

Figures S12

D and

S12E). Then, we discovered that NDGA administration dose-
dependently ameliorated ischemic volume using TTC staining
and markedly reduced neurological scores in the Longa test
compared with the vehicle group after ischemia attack
(

Figures 7

A–7C).

(F) NDGA decreased the

Cox-1

CYTB

, and

loop

expression in primary mouse microglia cells.

= 3. Results are represented as means

SD. ***

< 0.001 vs. CN

group and

< 0.01 and

###

< 0.001 vs. LPS group.

(G) NDGA reduced the newly synthesized DNA, as visualized by EdU labeling in primary mouse microglia cells. Scale bar: 5

= 6.

(H) NDGA reduced 8-OHdG expression in LPS-primed microglia followed by ATP stimulation. Scale bar: 20

= 4.

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

To explore whether the neuroprotective effects of NDGA in

ischemic stroke were associated with CMPK2, we administrated
AAV-shCon and AAV-shCMPK2

CR1

Cre/ERT2

mice with an

effective dosage of NDGA (20 mg/kg). As shown in

Figures

S12

F–S12H

NDGA could not exhibit additional beneficial effects,

including decreased ischemic lesion and enhanced short-term
functional outcome, in AAV-shCMPK2

CR1

Cre/ERT2

mice.

Based on the above findings, we conclude that NDGA could
effectively ameliorate ischemic brain injury in tMCAO mice mainly
through CMPK2 inhibition.

In addition, the long-term functional performances were mark-

edly improved with NDGA treatment in mice subjected to
tMCAO, as indicated by the significantly decreased scores in
the mNSS test, upregulated scores in the wire-hanging test,
downregulated percentage of the forelimb faults in foot mis-

steps, and increased time in the rotarod test relative to the
vehicle group 0, 3, 5, 7, 14, 21, and 28 days post-ischemia (

Fig-

ure 7

D). Moreover, NDGA administration promoted the survival

proportions at the relatively delayed stage (28 days) compared
to the vehicle group in the mice post-ischemia (

Figure 7

E). These

results indicated that NDGA could ameliorate ischemic damage
and improve long-term neurological performances in mice sub-
jected to ischemic stroke.

To better understand the mechanism underlying then neuro-

protective effects of NDGA in ischemic stroke, we analyzed mi-
croglial morphological alterations using Iba1 immunofluorescent
staining followed by 3D reconstruction. Evidently, administration
of NDGA suppressed the hyperactivated state of microglia,
as evidenced by increased process lengths and branches in
comparison to the vehicle group following ischemic onset

Figure 7. The CMPK2 inhibitor NDGA reduces brain ischemic volume and improves neurological performances in the mouse tMCAO model

(A–C) NDGA reduced cerebral infarct volume and promoted short-term neurological functions in the male mice.

= 8.

(D) NDGA improved long-term functional performances, including mNSS test, wire-hanging test, forelimb foot fault test, and rotarod test in the male mice.

= 8.

(E) NDGA improved long-term survival proportions in the male mice.

= 12–15.

Results are represented as means

SD. *

< 0.05, **

< 0.01, and ***

< 0.001 vs. sham group and

< 0.05,

< 0.01, and

###

< 0.001 vs. tMCAO plus vehicle

group.

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

(

Figures S13

A–S13C). Then, we investigated the effects of

NDGA on Ox-mtDNA formation. Immunohistochemical staining
analysis

confirmed

that

NDGA

substantially

decreased

8-OHdG expression in microglia in the peri-infarct region of
ischemic mice at 72 h after tMCAO compared with the vehicle
group (

Figures S13

D and S13E). In addition, NDGA treatment

greatly reduced the levels of cleaved caspase-1, N-GSDMD,
and mature IL-1

in the peri-infarct region of ischemic mice at

after

tMCAO

compared

with

the

vehicle

group

(

Figures S13

F and S13G). Therefore, these data indicated that

the NDGA could attenuate Ox-mtDNA formation and NLRP3 in-
flammasome activation in mice subjected to tMCAO.

CMPK2 inhibitor NDGA suppresses Ox-mtDNA
production and NLPR3 inflammasome activation in
primary human monocytes from patients with cerebral
ischemic stroke

To corroborate the translational relevance of the findings from
the

in vivo

and

in vitro

models described above, we further iso-

lated the peripheral blood mononuclear cells (PBMCs) from pa-
tients with acute cerebral ischemic stroke and established the
NLRP3 inflammasome activation model. Upon priming with
LPS followed by ATP stimulation, mature IL-1

was substantially

elevated in the cell supernatant of PBMCs (

Figure S14

A).

Congruent with the above

in vitro

finding, NDGA treatment mark-

edly inhibited IL-1

release (

Figure S14

A). Moreover, a positive

correlation was observed between

CMPK2

gene expression

and IL-1

inhibition rate by NDGA in PBMCs from patients with

cerebral ischemic stroke (

Figure S14

B). We also examined the

effects of NDGA on PBMCs from healthy donors and found
that NDGA treatment could also inhibit mature IL-1

release after

LPS plus ATP stimulation (

Figure S14

C). Additionally, NDGA

treatment also mitigated Ox-mtDNA production in PBMCs, as
determined

8-OHdG

immunofluorescence

staining

(

Figures S14

D and S14E). The aforementioned data support

our hypothesis that NDGA, as a specific CMPK2 inhibitor, could
limit Ox-mtDNA formation and NLRP3 inflammasome activation
in primary human monocytes from patients with stroke and may
have better therapeutic potential in patients with stroke with
higher CMPK2 expression.

DISCUSSION

Currently, little is known about the potential role of CMPK2 in
neuroinflammation and brain injury during ischemic stroke. Our
study revealed that

Cre

recombination-dependent AAV-medi-

ated microglia/macrophage CMPK2 knockdown or pharmaco-
logical intervention with NDGA mitigated ischemic injury by in-
hibiting Ox-mtDNA-induced NLRP3 inflammasome activation,
indicating that microglial CMPK2 may play a crucial role in the
pathological processes of ischemic stroke.

In the peripheral circulation system, Ox-mtDNA can act as one

of the DAMPs to activate the NLRP3 inflammasome and then
prompt the formation of GSDMD pores in the membranes of pe-
ripheral cells including monocytes, macrophages, and neutro-
phils, which allows the circulatory release of more Ox-mtDNA
to initiate the inflammatory cascade.

Lorente et al. and Liu

et al. also confirmed that concentrations of 8-OHdG were higher

in the peripheral blood of ischemic patients and positively corre-
lated with the severity of ischemic injury and poor prognosis.

Our data indicated that

CMPK2

gene expression in human

stroke patients was positively correlated with infarct volume
and NIHSS score, supporting the clinical significance of
CMPK2 in ischemic stroke.

In peripheral system disorders, pharmacological interventions

or gene knockout/knockdown of CMPK2 have demonstrated
therapeutic effects on a series of inflammation-associated dis-
eases, including ARDS, atherosclerosis, oral ulcer healing, and
liver ischemia/reperfusion injury.

21–23

The sterile inflammation

mediated by DAMPs is a crucial component in the pathophysi-
ology after stroke; ATP, HMGB-1, MRP8 (S100A8), and MRP14
(S100A9) may be promising DAMPs to promote CMPK2 upregu-
lation in the brain after stroke, which is worthy of further investi-
gation. Nevertheless, the precise function of CMPK2 in neuroin-
flammation of central nervous system disorders remains elusive.
Our experiments revealed that CMPK2 expression was substan-
tially elevated in the peri-infarct region of ischemic rodents,
predominantly located in the microglia. The

in vitro

data also

supported the above finding that CMPK2 expression was
markedly enhanced in microglia following OGD treatment.
Consequently, our findings firmly suggested that microglia/
macrophage CMPK2 expression was markedly increased in
the peri-infarct region after cerebral ischemic attack.

In brain calcification, biallelic variants of CMPK2 lead to

reduced mtDNA copies and promote mitochondrial dysfunc-
tion.

In addition, cannabidiol negatively regulates the transcrip-

tional activity of PPAR

on CMPK2 and inhibits inflammasome

activation, thereby relieving oral ulcer healing.

Moreover,

myeloid-specific knockout of CMPK2 can mitigate pulmonary
injury in ARDS.

Furthermore, CMPK2 accelerates the synthesis

of deoxy-3

-didehydro-cytidine triphosphate, blocking virus

replication in CoVs.

However, the precise function of CMPK2

in microglia, particularly in the context of ischemic stroke, has
not been investigated. Hence, we adopted a well-established
strategy of AAV selective expression based on the

Cre

recombi-

nase system to knock down CMPK2 expression in microglia. Mi-
croglia/macrophage CMPK2 silencing elicited protective effects
in reducing ischemic volume and promoting short-term neuro-
logical outcome in ischemic mice at 3 days post-operation.
Moreover, AAV-shCMPK2 injection attenuated long-term neuro-
logical deficits in the

CR1

Cre/ERT2

mice, indicating that

CMPK2 might be a biomarker associated with poor prognosis
following stroke.

Microglia dynamically modulate the morphology between the

ramified state and ameboid morphology, the latter of which indi-
cates the activated state.

Although microglia of ameboid

morphology help remove necrotic cells, cellular and myelin
debris, and pathogens, prolonged exposure to hyperactivated
microglia leads to neuron damage and loss. Our data demon-
strated that CMPK2 knockdown maintained the homeostatic
state of microglia, as indicated by increases in process lengths
and branch number, which might enable microglia to react
appropriately to an ischemic insult and enhance neurological
performance. Han et al. also confirmed that enhanced microglial
ramification mediated by PGC-1

protected against ischemia-

induced brain injury in mice.

Cao et al. also highlighted the

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

essential role of microglia with a ramified morphology in surveil-
lance and scavenging functions.

Therefore, we speculated that

CMPK2 silencing would limit the overactivation of microglia, thus
reducing brain infarct injury in the ischemic mice.

Under normal physiological conditions, mtDNA exists as a

nucleoid with a compaction size of less than 5

m, which is pre-

cisely regulated by the packaging and compaction of TFAM.

Under conditions of microbial infections, non-sterile inflamma-
tion, and tumor cell stimuli, mtDNA replication is extensively
accelerated by CMPK2. Of note, the newly synthesized mtDNA
is naked and unstable and can easily secrete into the cytosol
through mPTP and VDAC1, thereby exacerbating tissue injury.

Especially, the D-loop bubble of mtDNA is vulnerable to oxida-
tive stress and then is oxidized by ROSs to produce Ox-
mtDNA, which can directly activate the NLRP3 inflammasome.
Furthermore, microglial Hv1 increases ROS generation through
NADPH oxidase regulation and then aggravates neuron death
and brain injury after neurological disorders,

61–65

which may syn-

ergize with CMPK2 to exacerbate neuroinflammation and neuron
death after ischemic stroke. Our current study showed that
CMPK2 knockdown suppressed new mtDNA synthesis, as
determined by EdU staining, inhibited levels of representative
mtDNA markers, including D-loop,

Cox1

, and

CYTB

, and disrup-

ted Ox-mtDNA formation, as determined by 8-OHdG immunoflu-
orescence staining in primary mouse microglia upon LPS prim-
ing followed by ATP stimulation or OGD conditions

in vitro

. The

in vivo

data also validated that CMPK2 knockdown inhibited

8-OHdG expression in microglia/macrophages in the ischemic
mice stroke model. Furthermore, CMPK2 silencing inhibited
cleaved caspase-1, mature IL-1

, and bioactive N-GSDMD in

the mouse tMCAO model

in vivo

and LPS-primed microglia fol-

lowed by ATP stimulation

in vitro

. Xian et al. also verified that

myeloid-specific CMPK2 ablation suppressed IL-1

production

and NLRP3 inflammasome assembly in LPS-induced ARDS.

Zhong et al. highlighted that mtDNA and 8-OHdG, rather than
mtROSs, directly enabled NLRP3 inflammasome activation.

Thus, we concluded that CMPK2 accelerated mtDNA replication
and Ox-mtDNA production and then promoted NLRP3 inflam-
masome activation in the microglia so as to aggravate the brain
injury after ischemic stroke.

Target identification and lead discovery are critical steps in

drug discovery for clinical use. Using the CMPK2 kinase assay
combined with the SPR assay, we identified NDGA as a potent
CMPK2 inhibitor. NDGA is the principal metabolite of creosote
bush and has been confirmed as a promising agent for treating
multiple diseases, including cardiovascular diseases, neurolog-
ical disorders, and cancers, as well as viral infections.

55–57

addition, a previous investigation reported that NDGA amelio-
rated cerebral infarction volume by inhibiting neuronal apoptosis
in a rat with ischemic stroke, but the cellular and molecular
mechanisms underlying the neuroprotective effects of NDGA
remain ambiguous under ischemia stimuli.

55–57

Therefore, a

deeper understanding of the precise mechanisms of NDGA in
microglia will help explain its therapeutic efficacy and lead to
the discovery of NDGA derivatives with enhanced therapeutic
potential for ischemic stroke. Our data revealed that NDGA treat-
ment blocked the new mtDNA synthesis and inhibited 8-OHdG
formation, corroborating the abovementioned findings that

CMPK2 knockdown could efficiently suppress mtDNA replica-
tion and Ox-mtDNA production. In addition, NDGA treatment
could potently reduce the levels of cleaved caspase-1, mature
IL-1

, and bioactive N-GSDMD in LPS-primed microglia fol-

lowed by ATP stimulation

in vitro

and the mouse tMCAO model

in vivo

. However, 5-LOX knockdown could not diminish the pro-

tein expressions of cleaved caspase-1, mature IL-1

, or bioac-

tive N-GSDMD, nor could it influence the inhibitory effects of
NDGA on NLRP3 inflammasome in LPS-primed microglia
following ATP stimulation. Moreover, NDGA treatment had no
further suppressive effects on NLRP3 inflammasome activation
in CMPK2-silenced microglia. These findings indicated that the
role of NDGA in NLRP3 inflammasome activation was indepen-
dent of 5-LOX but at least partially dependent on CMPK2.
Furthermore, we also observed that NDGA reduced ischemic
volume and short-term and long-term neurological deficits in
the mouse tMCAO model. Subsequently, we studied whether
the neuroprotective effects of NDGA were correlated with
CMPK2. Our data proved that NDGA cannot produce additional
therapeutic effects in AAV-shCMPK2 CX

CR1

Cre/ERT2

mice after

ischemic stroke, which validated that the neuroprotective effects
of NDGA in ischemic stroke were at least partially dependent on
CMPK2. To better ascertain the clinical relevance of CMPK2, we
evaluated the effects of CMPK2-specific inhibitor NDGA on pri-
mary human monocytes. We discovered that NDGA substan-
tially reduced mature IL-1

release and 8-OHdG expression in

LPS-primed PBMCs, followed by ATP stimulation. Of note, we
observed a positive correlation between

CMPK2

gene expres-

sion and IL-1

inhibition rate by NDGA, indicating that CMPK2

expression was highly associated with the inhibition rate of
NDGA on CMPK2. In conclusion, NDGA administration, a prom-
ising immunotherapy based on CMPK2 inhibition, promises to
be a mainstay of ischemic stroke management in future clinical
research.

Limitations of the study

There are several unsolved issues in the current investigation.
First, we confirmed that microglia CMPK2 knockdown amelio-
rates cerebral ischemic injury by inhibiting the Ox-mtDNA-
induced NLRP3 inflammasome activation. However, it cannot
be ruled out that other mechanisms of CMPK2 action besides
the NLRP3 inflammasome may contribute to the progression of
ischemic stroke. Specifically, it remains to be determined
whether the STING signaling pathway is involved in the function
of CMPK2 in cerebral ischemic injury. Second, the signaling
pathways or key mediators that promote the increase of
CMPK2 protein remain obscure, so further experimental and ro-
dent models (microglia CMPK2 transgenic or knockout mice) are
required. Third, neuroinflammation is implicated in a series of
disease processes, including Alzheimer’s disease, vascular de-
mentia, traumatic brain injury, and multiple sclerosis. We will
further continue to explore whether CMPK2 participates in other
neurologic disorders, such as multiple sclerosis and vascular de-
mentia. Fourth, due to the unavailability of human brain samples
and primary human microglia from patients with stroke, we ob-
tained primary human monocytes isolated from patients with
stroke to evaluate the inhibitory effects of CMPK2-specific inhib-
itor NDGA on NLRP3 inflammasome activation

in vitro

. Thus, our

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

understanding of CMPK2 function in patients with ischemic
stroke is less refined. Hence, human postmortem brain tissue
may be utilized in our ongoing research. Lastly, the brain tissue
of rodents and humans differs greatly; therefore, non-human pri-
mates would be an alternative model to evaluate the role of
CMPK2 in ischemic stroke through pharmacological modulation
and genetic intervention. Nevertheless, our findings in this study
would provide CMPK2 as a potential target for ischemic stroke
and also CMPK2 inhibitor NDGA as a valuable therapeutic agent
for the treatment of ischemic stroke.

STAR

METHODS

Detailed methods are provided in the online version of this paper and include
the following:

KEY RESOURCES TABLE

RESOURCE AVAILABILITY

Lead contact

Materials availability

Data and code availability

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Human sample collection and processing

Animals

Isolation of primary mouse cortical neurons, astrocytes and micro-
glia

METHOD DETAILS

Quantitative real-time PCR (qRT-PCR)

Western blotting analysis

CMPK2 protein and antibody production

Establishment of focal cerebral ischemia/reperfusion model

Assessment of cerebral infarct volume and neurological function

Immunofluorescence staining

Adult mouse microglial isolation and flow cytometry examination

Oxygen-glucose deprivation (OGD) treatment

Stereotaxic surgery

In vivo

mouse brain magnetic resonance imaging (MRI) scanning

Three-dimensional (3D) reconstruction

LPS-primed NLRP3 inflammasome model in primary mouse micro-
glia cells and human primary monocytes

Cell transfection

Analysis of lactate dehydrogenase (LDH) release

Edu staining experiment

Measurement of total mtDNA

Enzyme-linked immunosorbent assay (ELISA) for IL-1

cytokine

In vitro

CMPK2 kinase assay and compounds screening

Surface plasmon resonance (SPR) assay

QUANTIFICATION AND STATISTICAL ANALYSIS

SUPPLEMENTAL INFORMATION

Supplemental information can be found online at

https://doi.org/10.1016/j.

xcrm.2024.101522

ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China
(82174010, 81973512, and 81973500); The Open Fund of the State Key Labo-
ratory of Pharmaceutical Biotechnology, Nanjing University, China (grant no.
KF-202206), and the Double First-Class Project of China Pharmaceutical Uni-
versity (CPUQNJC22_02).

AUTHOR CONTRIBUTIONS

T.P., X.X., J.W., and X.G. conceived the project and designed the studies.
X.G., J.W., and T.P. wrote the paper. M.L. and Y.W. collected human samples.
S.Z. performed enzyme assay and compound screening. X.G., J.S., K.L., and
L.X. revised the manuscript. X.G., J.S., K.L., and L.X. conducted animal exper-
iments. X.G., J.S., and K.L. performed the molecular biology and cell biology
experiments.

DECLARATION OF INTERESTS

The authors declare no competing interests.

Received: October 29, 2023
Revised: February 8, 2024
Accepted: March 28, 2024
Published: April 26, 2024

REFERENCES

1. Owolabi, M.O., Thrift, A.G., Mahal, A., Ishida, M., Martins, S., Johnson,

W.D., Pandian, J., Abd-Allah, F., Yaria, J., Phan, H.T., et al. (2022). Primary
stroke prevention worldwide: translating evidence into action. Lancet Pub-
lic Health

, e74–e85.

https://doi.org/10.1016/S2468-2667(21)00230-9

2. Yu, H., Chen, X., Guo, X., Chen, D., Jiang, L., Qi, Y., Shao, J., Tao, L.,

Hang, J., Lu, G., et al. (2023). The clinical value of serum xanthine oxidase
levels in patients with acute ischemic stroke. Redox Biol.

, 102623.

https://doi.org/10.1016/j.redox.2023.102623

3. Barclay, R.E., Stevenson, T.J., Poluha, W., Semenko, B., and Schubert, J.

(2020). Mental practice for treating upper extremity deficits in individuals
with hemiparesis after stroke. Cochrane Database Syst. Rev.

CD005950.

https://doi.org/10.1002/14651858.CD005950.pub5

4. (2021). Global, regional, and national burden of stroke and its risk factors,

1990-2019: a systematic analysis for the Global Burden of Disease Study
2019. Lancet Neurol.

, 795–820.

https://doi.org/10.1016/S1474-

4422(21)00252-0

5. Wang, R., Zhu, Y., Liu, Z., Chang, L., Bai, X., Kang, L., Cao, Y., Yang, X.,

Yu, H., Shi, M.-J., et al. (2021). Neutrophil extracellular traps promote
tPA-induced brain hemorrhage via cGAS in mice with stroke. Blood

138

91–103.

https://doi.org/10.1182/blood.2020008913

6. Brenna, S., Altmeppen, H.C., Mohammadi, B., Rissiek, B., Schlink, F., Lu-

dewig, P., Krisp, C., Schl

€

uter, H., Failla, A.V., Schneider, C., et al. (2020).

Characterization of brain-derived extracellular vesicles reveals changes
in cellular origin after stroke and enrichment of the prion protein with a po-
tential role in cellular uptake. J. Extracell. Vesicles

, 1809065.

https://doi.

org/10.1080/20013078.2020.1809065

7. Loppi, S., Korhonen, P., Bouvy-Liivrand, M., Caligola, S., Turunen, T.A.,

Turunen, M.P., Hernandez de Sande, A., Ko

osowska, N., Scoyni, F.,

Rosell, A., et al. (2021). Peripheral inflammation preceeding ischemia im-
pairs neuronal survival through mechanisms involving miR-127 in aged an-
imals. Aging Cell

, e13287.

https://doi.org/10.1111/acel.13287

8. Guan, X., Wu, J., Geng, J., Ji, D., Wei, D., Ling, Y., Zhang, Y., Jiang, G.,

Pang, T., and Huang, Z. (2024). A Novel Hybrid of Telmisartan and Borneol
Ameliorates Neuroinflammation and White Matter Injury in Ischemic
Stroke Through ATF3/CH25H Axis. Transl. Stroke Res.

, 195–218.

https://doi.org/10.1007/s12975-022-01121-5

9. Paolicelli, R.C., Sierra, A., Stevens, B., Tremblay, M.-E., Aguzzi, A., Ajami,

B., Amit, I., Audinat, E., Bechmann, I., Bennett, M., et al. (2022). Microglia
states and nomenclature: A field at its crossroads. Neuron

110

, 3458–

3483.

https://doi.org/10.1016/j.neuron.2022.10.020

10. Kim, S., Lee, W., Jo, H., Sonn, S.-K., Jeong, S.-J., Seo, S., Suh, J., Jin, J.,

Kweon, H.Y., Kim, T.K., et al. (2022). The antioxidant enzyme
Peroxiredoxin-1 controls stroke-associated microglia against acute
ischemic stroke. Redox Biol.

, 102347.

https://doi.org/10.1016/j.

redox.2022.102347

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

11. Bernhart, E., Kogelnik, N., Prasch, J., Gottschalk, B., Goeritzer, M., De-

paoli, M.R., Reicher, H., Nusshold, C., Plastira, I., Hammer, A., et al.
(2018). 2-Chlorohexadecanoic acid induces ER stress and mitochondrial
dysfunction in brain microvascular endothelial cells. Redox Biol.

441–451.

https://doi.org/10.1016/j.redox.2018.01.003

12. Kotrys, A.V., Cysewski, D., Czarnomska, S.D., Pietras, Z., Borowski, L.S.,

Dziembowski, A., and Szczesny, R.J. (2019). Quantitative proteomics re-
vealed C6orf203/MTRES1 as a factor preventing stress-induced tran-
scription deficiency in human mitochondria. Nucleic Acids Res.

7502–7517.

https://doi.org/10.1093/nar/gkz542

13. Kosmider, B., Lin, C.-R., Karim, L., Tomar, D., Vlasenko, L., Marchetti, N.,

Bolla, S., Madesh, M., Criner, G.J., and Bahmed, K. (2019). Mitochondrial
dysfunction in human primary alveolar type II cells in emphysema. EBio-
Medicine

, 305–316.

https://doi.org/10.1016/j.ebiom.2019.07.063

14. Xian, H., and Karin, M. (2023). Oxidized mitochondrial DNA: a protective

signal gone awry. Trends Immunol.

, 188–200.

https://doi.org/10.

1016/j.it.2023.01.006

15. Hsieh, Y.-W., Lin, K.-C., Korivi, M., Lee, T.-H., Wu, C.-Y., and Wu, K.-Y.

(2014). The reliability and predictive ability of a biomarker of oxidative
DNA damage on functional outcomes after stroke rehabilitation. Int. J.
Mol. Sci.

, 6504–6516.

https://doi.org/10.3390/ijms15046504

16. Liu, Z., Liu, Y., Tu, X., Shen, H., Qiu, H., Chen, H., and He, J. (2017). High

Serum Levels of Malondialdehyde and 8-OHdG are both Associated with
Early Cognitive Impairment in Patients with Acute Ischaemic Stroke. Sci.
Rep.

, 9493.

https://doi.org/10.1038/s41598-017-09988-3

17. Nakajima, H., Unoda, K.-I., Ito, T., Kitaoka, H., Kimura, F., and Hanafusa,

T. (2012). The Relation of Urinary 8-OHdG, A Marker of Oxidative Stress to
DNA, and Clinical Outcomes for Ischemic Stroke. Open Neurol. J.

51–57.

https://doi.org/10.2174/1874205X01206010051

18. Liu, Z., Cai, Y., and He, J. (2018). High serum levels of 8-OHdG are an in-

dependent predictor of post-stroke depression in Chinese stroke survi-
vors. Neuropsychiatr. Dis. Treat.

, 587–596.

https://doi.org/10.2147/

NDT.S155144

19.

Mizukoshi, G., Katsura, K.-I., and Katayama, Y. (2005). Urinary 8-hydroxy-
2’-deoxyguanosine and serum S100beta in acute cardioembolic stroke
patients. Neurol. Res.

, 644–646.

20. Zhao, M., Su, H.-Z., Zeng, Y.-H., Sun, Y., Guo, X.-X., Li, Y.-L., Wang, C.,

Zhao, Z.-Y., Huang, X.-J., Lin, K.-J., et al. (2022). Loss of function of
CMPK2 causes mitochondria deficiency and brain calcification. Cell Dis-
cov.

, 128.

https://doi.org/10.1038/s41421-022-00475-2

21. Xian, H., Liu, Y., Rundberg Nilsson, A., Gatchalian, R., Crother, T.R., Tour-

tellotte, W.G., Zhang, Y., Aleman-Muench, G.R., Lewis, G., Chen, W., et al.
(2021). Metformin inhibition of mitochondrial ATP and DNA synthesis ab-
rogates NLRP3 inflammasome activation and pulmonary inflammation.
Immunity

, 1463–1477.e11.

https://doi.org/10.1016/j.immuni.2021.

05.004

22. Chen, Y., Wu, L., Shi, M., Zeng, D., Hu, R., Wu, X., Han, S., He, K., Xu, H.,

Shao, X., and Ma, R. (2022). Electroacupuncture Inhibits NLRP3 Activation
by Regulating CMPK2 After Spinal Cord Injury. Front. Immunol.

788556.

https://doi.org/10.3389/fimmu.2022.788556

23. Pawlak, J.B., Hsu, J.C.-C., Xia, H., Han, P., Suh, H.-W., Grove, T.L., Mor-

rison, J., Shi, P.-Y., Cresswell, P., and Laurent-Rolle, M. (2023). CMPK2
restricts Zika virus replication by inhibiting viral translation. PLoS Pathog.

, e1011286.

https://doi.org/10.1371/journal.ppat.1011286

24. Zhu, M., Lv, J., Wang, W., Guo, R., Zhong, C., Antia, A., Zeng, Q., Li, J., Liu,

Q., Zhou, J., et al. (2023). CMPK2 is a host restriction factor that inhibits
infection of multiple coronaviruses in a cell-intrinsic manner. PLoS Biol.

, e3002039.

https://doi.org/10.1371/journal.pbio.3002039

25. Newman, L.E., and Shadel, G.S. (2023). Mitochondrial DNA Release in

Innate Immune Signaling. Annu. Rev. Biochem.

, 299–332.

https://doi.

org/10.1146/annurev-biochem-032620-104401

26. Zhong, Z., Liang, S., Sanchez-Lopez, E., He, F., Shalapour, S., Lin, X.-J.,

Wong, J., Ding, S., Seki, E., Schnabl, B., et al. (2018). New mitochondrial

DNA synthesis enables NLRP3 inflammasome activation. Nature

560

198–203.

https://doi.org/10.1038/s41586-018-0372-z

27.

Tang, Y., Xu, H., Du, X., Lit, L., Walker, W., Lu, A., Ran, R., Gregg, J.P., Re-
illy, M., Pancioli, A., et al. (2006). Gene expression in blood changes rapidly
in neutrophils and monocytes after ischemic stroke in humans: a microar-
ray study. J. Cerebr. Blood Flow Metabol.

, 1089–1102.

28. Raman, K., O’Donnell, M.J., Czlonkowska, A., Duarte, Y.C., Lopez-Jara-

millo, P., Pen˜aherrera, E., Sharma, M., Shoamanesh, A., Skowronska,
M., Yusuf, S., and Pare´, G. (2016). Peripheral Blood MCEMP1 Gene
Expression as a Biomarker for Stroke Prognosis. Stroke

, 652–658.

https://doi.org/10.1161/STROKEAHA.115.011854

29. Zhang, J., Wang, C., Chen, X., Takada, M., Fan, C., Zheng, X., Wen, H.,

Liu, Y., Wang, C., Pestell, R.G., et al. (2015). EglN2 associates with the
NRF1-PGC1

complex and controls mitochondrial function in breast can-

cer. EMBO J.

, 2953–2970.

https://doi.org/10.15252/embj.201591437

30. West, A.P., Khoury-Hanold, W., Staron, M., Tal, M.C., Pineda, C.M., Lang,

S.M., Bestwick, M., Duguay, B.A., Raimundo, N., MacDuff, D.A., et al.
(2015). Mitochondrial DNA stress primes the antiviral innate immune
response. Nature

520

, 553–557.

https://doi.org/10.1038/nature14156

31. Cao, P., Chen, C., Liu, A., Shan, Q., Zhu, X., Jia, C., Peng, X., Zhang, M.,

Farzinpour, Z., Zhou, W., et al. (2021). Early-life inflammation promotes
depressive symptoms in adolescence via microglial engulfment of den-
dritic spines. Neuron

109

, 2573–2589.e9.

https://doi.org/10.1016/j.

neuron.2021.06.012

32. Harrington, J.S., Ryter, S.W., Plataki, M., Price, D.R., and Choi, A.M.K.

(2023). Mitochondria in health, disease, and aging. Physiol. Rev.

103

2349–2422.

https://doi.org/10.1152/physrev.00058.2021

33. Studer, B., Timm, A., Sahakian, B.J., Kalenscher, T., and Knecht, S. (2021).

A decision-neuroscientific intervention to improve cognitive recovery after
stroke. Brain

144

, 1764–1773.

https://doi.org/10.1093/brain/awab128

34. Han, B., Jiang, W., Cui, P., Zheng, K., Dang, C., Wang, J., Li, H., Chen, L.,

Zhang, R., Wang, Q.M., et al. (2021). Microglial PGC-1

protects against

ischemic brain injury by suppressing neuroinflammation. Genome Med.

, 47.

https://doi.org/10.1186/s13073-021-00863-5

35. Li, D., Lang, W., Zhou, C., Wu, C., Zhang, F., Liu, Q., Yang, S., and Hao, J.

(2018). Upregulation of Microglial ZEB1 Ameliorates Brain Damage after
Acute Ischemic Stroke. Cell Rep.

, 3574–3586.

https://doi.org/10.

1016/j.celrep.2018.03.011

36. Gu, N., Peng, J., Murugan, M., Wang, X., Eyo, U.B., Sun, D., Ren, Y., Di-

Cicco-Bloom, E., Young, W., Dong, H., and Wu, L.-J. (2016). Spinal Micro-
gliosis Due to Resident Microglial Proliferation Is Required for Pain Hyper-
sensitivity after Peripheral Nerve Injury. Cell Rep.

, 605–614.

https://doi.

org/10.1016/j.celrep.2016.06.018

37. Zhang, C.-J., Jiang, M., Zhou, H., Liu, W., Wang, C., Kang, Z., Han, B.,

Zhang, Q., Chen, X., Xiao, J., et al. (2018). TLR-stimulated IRAKM acti-
vates caspase-8 inflammasome in microglia and promotes neuroinflam-
mation.

Clin.

Invest.

128

5399–5412.

https://doi.org/10.1172/

JCI121901

38. McDonough, A., Noor, S., Lee, R.V., Dodge, R., Strosnider, J.S., Shen, J.,

Davidson, S., Mo¨ller, T., Garden, G.A., and Weinstein, J.R. (2020).
Ischemic preconditioning induces cortical microglial proliferation and a
transcriptomic program of robust cell cycle activation. Glia

, 76–94.

https://doi.org/10.1002/glia.23701

39. Walsh, J.G., Muruve, D.A., and Power, C. (2014). Inflammasomes in the

CNS. Nat. Rev. Neurosci.

, 84–97.

https://doi.org/10.1038/nrn3638

40. Chen, S., Mei, S., Luo, Y., Wu, H., Zhang, J., and Zhu, J. (2018). Gasdermin

Family: a Promising Therapeutic Target for Stroke. Transl. Stroke Res.

555–563.

https://doi.org/10.1007/s12975-018-0666-3

41. He, M., Chiang, H.-H., Luo, H., Zheng, Z., Qiao, Q., Wang, L., Tan, M., Oh-

kubo, R., Mu, W.-C., Zhao, S., et al. (2020). An Acetylation Switch of the
NLRP3 Inflammasome Regulates Aging-Associated Chronic Inflammation
and Insulin Resistance. Cell Metabol.

, 580–591.e5.

https://doi.org/10.

1016/j.cmet.2020.01.009

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

42. Han, X., Xu, T., Fang, Q., Zhang, H., Yue, L., Hu, G., and Sun, L. (2021).

Quercetin hinders microglial activation to alleviate neurotoxicity via the
interplay between NLRP3 inflammasome and mitophagy. Redox Biol.

, 102010.

https://doi.org/10.1016/j.redox.2021.102010

43. Li, S., Fang, Y., Zhang, Y., Song, M., Zhang, X., Ding, X., Yao, H., Chen, M.,

Sun, Y., Ding, J., et al. (2022). Microglial NLRP3 inflammasome activates
neurotoxic astrocytes in depression-like mice. Cell Rep.

, 111532.

https://doi.org/10.1016/j.celrep.2022.111532

44. Ding, L., Yang, X., Tian, H., Liang, J., Zhang, F., Wang, G., Wang, Y., Ding,

M., Shui, G., and Huang, X. (2018). Seipin regulates lipid homeostasis by
ensuring calcium-dependent mitochondrial metabolism. EMBO J.

e97572.

https://doi.org/10.15252/embj.201797572

45. Yu, B., Ma, J., Li, J., Wang, D., Wang, Z., and Wang, S. (2020). Mitochon-

drial phosphatase PGAM5 modulates cellular senescence by regulating
mitochondrial dynamics. Nat. Commun.

, 2549.

https://doi.org/10.

1038/s41467-020-16312-7

46. Huang, T., Lin, R., Su, Y., Sun, H., Zheng, X., Zhang, J., Lu, X., Zhao, B.,

Jiang, X., Huang, L., et al. (2023). Efficient intervention for pulmonary
fibrosis via mitochondrial transfer promoted by mitochondrial biogenesis.
Nat. Commun.

, 5781.

https://doi.org/10.1038/s41467-023-41529-7

47. Zhao, T., Zhang, J., Lei, H., Meng, Y., Cheng, H., Zhao, Y., Geng, G., Mu,

C., Chen, L., Liu, Q., et al. (2023). NRF1-mediated mitochondrial biogen-
esis antagonizes innate antiviral immunity. EMBO J.

, e113258.

https://doi.org/10.15252/embj.2022113258

48. Zeng, Z., You, M., Fan, C., Rong, R., Li, H., and Xia, X. (2023). Pathologi-

cally high intraocular pressure induces mitochondrial dysfunction through
Drp1 and leads to retinal ganglion cell PANoptosis in glaucoma. Redox
Biol.

, 102687.

https://doi.org/10.1016/j.redox.2023.102687

49. Chen, J., Sun, W., Zhang, H., Ma, J., Xu, P., Yu, Y., Fang, H., Zhou, L., Lv,

J., Xie, J., et al. (2020). Macrophages reprogrammed by lung cancer mi-
croparticles promote tumor development via release of IL-1

. Cell. Mol.

Immunol.

, 1233–1244.

https://doi.org/10.1038/s41423-019-0313-2

50. Mun˜oz-Wolf, N., Ward, R.W., Hearnden, C.H., Sharp, F.A., Geoghegan, J.,

O’Grady, K., McEntee, C.P., Shanahan, K.A., Guy, C., Bowie, A.G., et al.
(2023). Non-canonical inflammasome activation mediates the adjuvantic-
ity of nanoparticles. Cell Rep. Med.

, 100899.

https://doi.org/10.1016/j.

xcrm.2022.100899

51. Jin, X., Liu, D., Zhou, X., Luo, X., Huang, Q., and Huang, Y. (2023). Entrec-

tinib inhibits NLRP3 inflammasome and inflammatory diseases by directly
targeting NEK7. Cell Rep. Med.

, 101310.

https://doi.org/10.1016/j.xcrm.

2023.101310

52. Sidor, K., Jeznach, A., Hoser, G., and Skirecki, T. (2023). 1-Methylnicotina-

mide (1-MNA) inhibits the activation of the NLRP3 inflammasome in human
macrophages. Int. Immunopharm.

121

, 110445.

https://doi.org/10.1016/j.

intimp.2023.110445

53. Bai, H., Yang, B., Yu, W., Xiao, Y., Yu, D., and Zhang, Q. (2018). Cathepsin

B links oxidative stress to the activation of NLRP3 inflammasome. Exp.
Cell Res.

362

, 180–187.

https://doi.org/10.1016/j.yexcr.2017.11.015

54. Rajanbabu, V., Galam, L., Fukumoto, J., Enciso, J., Tadikonda, P., Lane,

T.N., Bandyopadhyay, S., Parthasarathy, P.T., Cho, Y., Cho, S.H., et al.
(2015). Genipin suppresses NLRP3 inflammasome activation through un-
coupling protein-2. Cell. Immunol.

297

, 40–45.

https://doi.org/10.1016/j.

cellimm.2015.06.002

55. Liu, Y., Wang, H., Zhu, Y., Chen, L., Qu, Y., and Zhu, Y. (2012). The protec-

tive effect of nordihydroguaiaretic acid on cerebral ischemia/reperfusion
injury is mediated by the JNK pathway. Brain Res.

1445

, 73–81.

https://

doi.org/10.1016/j.brainres.2012.01.031

56.

Shishido, Y., Furushiro, M., Hashimoto, S., and Yokokura, T. (2001). Effect
of nordihydroguaiaretic acid on behavioral impairment and neuronal cell
death after forebrain ischemia. Pharmacol. Biochem. Behav.

, 469–474.

57.

Aktan, S., Aykut, C., Yegen, B.C., Okar, I., Ozkutlu, U., and Ercan, S.
(1993). The effect of nordihydroguaiaretic acid on leukotriene C4 and pros-
taglandin E2 production following different reperfusion periods in rat brain

after forebrain ischemia correlated with morphological changes. Prosta-
glandins Leukot. Essent. Fatty Acids

, 633–641.

58. Lorente, L., Martı´n, M.M., Gonza´lez-Rivero, A.F., Pe´rez-Cejas, A., Abreu-

Gonza´lez, P., Ramos, L., Argueso, M., Ca´ceres, J.J., Sole´-Viola´n, J., Al-
varez-Castillo, A., et al. (2021). DNA and RNA oxidative damage are asso-
ciated to mortality in patients with cerebral infarction. Med. Intensiva

35–41.

https://doi.org/10.1016/j.medin.2019.07.008

59. Wu, W., Bao, W., Chen, X., Lu, Y., Fang, J., Liu, J., Peng, S., Pi, J., Tom-

linson, B., Chan, P., et al. (2023). Endothelial Gata6 deletion reduces
monocyte recruitment and proinflammatory macrophage formation and
attenuates atherosclerosis through Cmpk2-Nlrp3 pathways. Redox Biol.

, 102775.

https://doi.org/10.1016/j.redox.2023.102775

60. Qi, X., Lin, W., Wu, Y., Li, Q., Zhou, X., Li, H., Xiao, Q., Wang, Y., Shao, B.,

and Yuan, Q. (2022). CBD Promotes Oral Ulcer Healing via Inhibiting
CMPK2-Mediated Inflammasome. J. Dent. Res.

101

, 206–215.

https://

doi.org/10.1177/00220345211024528

61. Wu, L.-J. (2014). Microglial voltage-gated proton channel Hv1 in ischemic

stroke. Transl. Stroke Res.

, 99–108.

https://doi.org/10.1007/s12975-

013-0289-7

62. Wu, L.-J., Wu, G., Akhavan Sharif, M.R., Baker, A., Jia, Y., Fahey, F.H.,

Luo, H.R., Feener, E.P., and Clapham, D.E. (2012). The voltage-gated pro-
ton channel Hv1 enhances brain damage from ischemic stroke. Nat. Neu-
rosci.

, 565–573.

https://doi.org/10.1038/nn.3059

63. Liu, X., Lei, Z., Gilhooly, D., He, J., Li, Y., Ritzel, R.M., Li, H., Wu, L.-J., Liu,

S., and Wu, J. (2023). Traumatic brain injury-induced inflammatory
changes in the olfactory bulb disrupt neuronal networks leading to olfac-
tory dysfunction. Brain Behav. Immun.

114

, 22–45.

https://doi.org/10.

1016/j.bbi.2023.08.004

64. Haruwaka, K., Ying, Y., Liang, Y., Umpierre, A.D., Yi, M.-H., Kremen, V.,

Chen, T., Xie, T., Qi, F., Zhao, S., et al. (2024). Microglia enhance post-
anesthesia neuronal activity by shielding inhibitory synapses. Nat. Neuro-
sci.

, 449–461.

https://doi.org/10.1038/s41593-023-01537-8

65. Zhao, S., Umpierre, A.D., and Wu, L.-J. (2024). Tuning neural circuits and

behaviors by microglia in the adult brain. Trends Neurosci.

, 181–194.

https://doi.org/10.1016/j.tins.2023.12.003

66. He, T., Shang, J., Gao, C., Guan, X., Chen, Y., Zhu, L., Zhang, L., Zhang,

C., Zhang, J., and Pang, T. (2021). A novel SIRT6 activator ameliorates
neuroinflammation and ischemic brain injury via EZH2/FOXC1 axis. Acta
Pharm. Sin. B

, 708–726.

https://doi.org/10.1016/j.apsb.2020.11.002

67. Iiott, N.E., Heward, J.A., Roux, B., Tsitsiou, E., Fenwick, P.S., Lenzi, L.,

Goodhead, I., Hertz-Fowler, C., Heger, A., Hall, N., et al. (2014). Long
non-coding RNAs and enhancer RNAs regulate the lipopolysaccharide-
induced inflammatory response in human monocytes. Nat. Commun.

3979.

https://doi.org/10.1038/ncomms4979

68. Jin, M., Shiwaku, H., Tanaka, H., Obita, T., Ohuchi, S., Yoshioka, Y., Jin, X.,

Kondo, K., Fujita, K., Homma, H., et al. (2021). Tau activates microglia via
the PQBP1-cGAS-STING pathway to promote brain inflammation. Nat.
Commun.

, 6565.

https://doi.org/10.1038/s41467-021-26851-2

69. Hupp, S., Fo¨rtsch, C., Graber, F., Mitchell, T.J., and Iliev, A.I. (2022). Pneu-

molysin boosts the neuroinflammatory response to Streptococcus pneu-
moniae through enhanced endocytosis. Nat. Commun.

, 5032.

https://

doi.org/10.1038/s41467-022-32624-2

70. Guo, J., Wang, F., Hu, Y., Luo, Y., Wei, Y., Xu, K., Zhang, H., Liu, H., Bo, L.,

Lv, S., et al. (2023). Exosome-based bone-targeting drug delivery allevi-
ates impaired osteoblastic bone formation and bone loss in inflammatory
bowel diseases. Cell Rep. Med.

, 100881.

https://doi.org/10.1016/j.xcrm.

2022.100881

71. Chin, C., Ravichandran, R., Sanborn, K., Fleming, T., Wheatcroft, S.B.,

Kearney, M.T., Tokman, S., Walia, R., Smith, M.A., Flint, D.J., et al.
(2023). Loss of IGFBP2 mediates alveolar type 2 cell senescence and pro-
motes lung fibrosis. Cell Rep. Med.

, 100945.

https://doi.org/10.1016/j.

xcrm.2023.100945

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

72. Singh, A., D’Amico, D., Andreux, P.A., Fouassier, A.M., Blanco-Bose, W.,

Evans, M., Aebischer, P., Auwerx, J., and Rinsch, C. (2022). Urolithin A im-
proves muscle strength, exercise performance, and biomarkers of mito-
chondrial health in a randomized trial in middle-aged adults. Cell Rep.
Med.

, 100633.

https://doi.org/10.1016/j.xcrm.2022.100633

73. Wang, X., Yu, J., Liu, X., Luo, D., Li, Y., Song, L., Jiang, X., Yin, X., Wang,

Y., Chai, L., et al. (2022). PSMG2-controlled proteasome-autophagy bal-
ance mediates the tolerance for MEK-targeted therapy in triple-negative
breast cancer. Cell Rep. Med.

, 100741.

https://doi.org/10.1016/j.xcrm.

2022.100741

74. Luo, Z., Hu, X., Wu, C., Chan, J., Liu, Z., Guo, C., Zhu, R., Zhang, L., Zhang,

Y., Jin, S., and He, S. (2023). Plasma exosomes generated by ischaemic
preconditioning are cardioprotective in a rat heart failure model. Br. J.
Anaesth.

130

, 29–38.

https://doi.org/10.1016/j.bja.2022.08.040

75. Wang, Y., Guan, X., Gao, C.-L., Ruan, W., Zhao, S., Kai, G., Li, F., and

Pang, T. (2021). Medioresinol as a novel PGC-1

activator prevents pyrop-

tosis of endothelial cells in ischemic stroke through PPAR

-GOT1 axis.

Pharmacol.

Res.

169

105640.

https://doi.org/10.1016/j.phrs.2021.

105640

76. Gao, C., Xu, Y., Liang, Z., Wang, Y., Shang, Q., Zhang, S., Wang, C., Ni, M.,

Wu, D., Huang, Z., and Pang, T. (2021). A novel PGAM5 inhibitor LFHP-1c
protects blood-brain barrier integrity in ischemic stroke. Acta Pharm. Sin.
B

, 1867–1884.

https://doi.org/10.1016/j.apsb.2021.01.008

77. Wlodarczyk, A., Holtman, I.R., Krueger, M., Yogev, N., Bruttger, J., Khor-

ooshi, R., Benmamar-Badel, A., de Boer-Bergsma, J.J., Martin, N.A., Kar-
ram, K., et al. (2017). A novel microglial subset plays a key role in myelino-
genesis in developing brain. EMBO J.

, 3292–3308.

https://doi.org/10.

15252/embj.201696056

78. Kana, V., Desland, F.A., Casanova-Acebes, M., Ayata, P., Badimon, A.,

Nabel, E., Yamamuro, K., Sneeboer, M., Tan, I.L., Flanigan, M.E., et al.
(2019). CSF-1 controls cerebellar microglia and is required for motor func-
tion and social interaction. J. Exp. Med.

216

, 2265–2281.

https://doi.org/

10.1084/jem.20182037

79. Wan, P., Su, W., Zhang, Y., Li, Z., Deng, C., Li, J., Jiang, N., Huang, S.,

Long, E., and Zhuo, Y. (2020). LncRNA H19 initiates microglial pyroptosis
and neuronal death in retinal ischemia/reperfusion injury. Cell Death Differ.

, 176–191.

https://doi.org/10.1038/s41418-019-0351-4

80. Zhang, J., Rong, P., Zhang, L., He, H., Zhou, T., Fan, Y., Mo, L., Zhao, Q.,

Han, Y., Li, S., et al. (2021). IL4-driven microglia modulate stress resilience
through BDNF-dependent neurogenesis. Sci. Adv.

, eabb9888.

https://

doi.org/10.1126/sciadv.abb9888

81. Al Mamun, A., Chauhan, A., Qi, S., Ngwa, C., Xu, Y., Sharmeen, R., Hazen,

A.L., Li, J., Aronowski, J.A., McCullough, L.D., and Liu, F. (2020). Microglial
IRF5-IRF4 regulatory axis regulates neuroinflammation after cerebral
ischemia and impacts stroke outcomes. Proc. Natl. Acad. Sci. USA

117

1742–1752.

https://doi.org/10.1073/pnas.1914742117

82. Lin, R., Zhou, Y., Yan, T., Wang, R., Li, H., Wu, Z., Zhang, X., Zhou, X.,

Zhao, F., Zhang, L., et al. (2022). Directed evolution of adeno-associated
virus for efficient gene delivery to microglia. Nat. Methods

, 976–985.

https://doi.org/10.1038/s41592-022-01547-7

83. Chen, J., Jin, J., Zhang, X., Yu, H., Zhu, X., Yu, L., Chen, Y., Liu, P., Dong,

X., Cao, X., et al. (2021). Microglial lnc-U90926 facilitates neutrophil infil-
tration in ischemic stroke via MDH2/CXCL2 axis. Mol. Ther.

, 2873–

2885.

https://doi.org/10.1016/j.ymthe.2021.04.025

84. Hu, Y., Yao, Y., Qi, H., Yang, J., Zhang, C., Zhang, A., Liu, X., Zhang, C.,

Gan, G., and Zhu, X. (2023). Microglia sense and suppress epileptic
neuronal hyperexcitability. Pharmacol. Res.

195

, 106881.

https://doi.org/

10.1016/j.phrs.2023.106881

85. Seo, D.-O., O’Donnell, D., Jain, N., Ulrich, J.D., Herz, J., Li, Y., Lemieux,

M., Cheng, J., Hu, H., Serrano, J.R., et al. (2023). ApoE isoform- and micro-
biota-dependent progression of neurodegeneration in a mouse model of
tauopathy. Science

379

, eadd1236.

https://doi.org/10.1126/science.

add1236

86. Xu, Y., Johansson, M., and Karlsson, A. (2008). Human UMP-CMP kinase

2, a novel nucleoside monophosphate kinase localized in mitochondria.
J. Biol. Chem.

283

, 1563–1571.

https://doi.org/10.1074/jbc.M707997200

87. Renner, M., Flanagan, A., Dejnirattisai, W., Puttikhunt, C., Kasinrerk, W.,

Supasa, P., Wongwiwat, W., Chawansuntati, K., Duangchinda, T., Cow-
per, A., et al. (2018). Characterization of a potent and highly unusual mini-
mally enhancing antibody directed against dengue virus. Nat. Immunol.

, 1248–1256.

https://doi.org/10.1038/s41590-018-0227-7

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

STAR

METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Antibodies

Mouse anti-

-actin

Proteintech

Cat# 20536-1-AP; RRID: AB_10700003

Rabbit anti-Caspase-1

CST

Cat# 89332S; RRID: AB_2923067

Rabbit anti-GSDMD

Abcam

Cat# ab209845; RRID: AB_2783550

Rabbit anti-GSDMD

Abcam

Cat# ab219800; RRID: AB_2888940

Goat anti-IL-1

R&D Systems

Cat# AB-401-NA; RRID: AB_354347

Rabbit anti-IBA1

Wako

Cat# 01919741; RRID: AB_839504

Goat anti-IBA1

NOVUS

Cat# NB100-1028; RRID: AB_521594

Mouse anti-GFAP

CST

Cat# 3670; RRID:AB_561049

Mouse anti-NeuN

Proteintech

Cat# 66836-1-Ig; RRID: AB_2882179

Rabbit anti-8-OHdG

Bioss

Cat# bs-1278R; RRID: AB_10856120

Rat anti-Ki67

Invitrogen

Cat# 14-5698-82; RRID: AB_10854564

Goat anti-ASC

Abcam

Cat# Ab175449; RRID: AB_3096354

Rat anti-CD11b

Invitrogen

Cat# 14-0112-82; RRID: AB_467108

Rabbit anti-TOMM20

Proteintech

Cat# 11802-1-AP; RRID: AB_2207530

Alexa Flour

488 goat anti-rabbit IgG (H + L)

Invitrogen

Cat# A11008; RRID: AB_143165

Alexa Flour

633

goat anti-mouse IgG (H + L)

Invitrogen

Cat# A21050; RRID: AB_2535718

FITC–conjugated donkey Anti-rabbit IgG(H + L)

Proteintech

Cat# SA00003-8; RRID: AB_2890899

Cy3 donkey anti-goat IgG (H + L)

Servicebio

Cat# GB21404; RRID: AB_2868507

Cy3 goat anti-rat IgG (H + L)

Servicebio

Cat# GB21302; RRID: AB_2936331

Chemicals, peptides, and recombinant proteins

NDGA

Sigma

74540

LPS

Sigma

I4524

ATP

Sigma

A8937

Poly-L-lysine

Sigma

P2636

Tamoxifen

Sigma

T5648

DMEM

Invitrogen

11965092

glucose-free DMEM

Invitrogen

11966025

DMEM/F-12

Invitrogen

11330032

MitoSOX

Yeasen

40778ES50

MCSF

Novoprotein

CB34

Lipofectamine 2000

Invitrogen

11668019

Hoechst 33342

Beyotime

C1025

DMSO

Sigma

D8418

Sigma

88597

Critical commercial assays

LDH Cytotoxicity Assay Kit

Beyotime

C0017

FastPure Cell/Tissue DNA Isolation Mini Kit

Vazyme Biotech Co.,Ltd

DC112

BCA protein assay

Beyotime

P0012

Human IL-1 beta ELISA Kit

VAL101

CD11b (Microglia) MicroBead

Miltenyi

130-093-636

ATP assay kits

Beyotime

S0027

(

Continued on next page

)

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be direct to and will be fulfilled by the lead contact, Tao Pang
(

tpang@cpu.edu.cn

Materials availability

This study did not generate new unique reagents.

Data and code availability

All data reported in this paper will be shared by the

lead contact

upon request. This paper does not report original code.

Any additional information required to reanalyze the data reported in this paper is available from the

lead contact

upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Human sample collection and processing

This investigation recruited 31 patients with acute ischemic stroke and 21 age-matched healthy subjects from September 2021 to
September 2023. This study was approved by the Ethics Committee of the Second Affiliated Hospital of Nanjing Medical University
with written informed consent from ischemic stroke patients or age-matched healthy subjects (Ethics approval number:
2019-KY116). The investigation was accomplished according to the guidelines of the Helsinki Declaration. Firstly, the patients
were confirmed with symptoms of acute ischemic stroke based on clinical manifestations and then

diagnosed using magnetic

resonance imaging (MRI) with diffusion-weighted imaging (DWI). Then, the blood samples from all patients were collected before
any therapeutic medication. Approximately 4–5 mL of peripheral blood samples were obtained in EDTA-K

anticoagulant tubes,

lysed in TRIzol regents, and then stored in a

C refrigerator. Ischemic stroke patients and healthy control subjects with the

following characteristics were excluded from this study: (1) having other intracranial diseases; (2) having an autoimmune disorder,
acute or chronic infectious diseases, hematologic disorders, or a history of surgical surgery in the past year; (3) taking hormones
or immunosuppressants within the past 6 months. As previously reported,

peripheral blood mononuclear cells (PBMC) were

sorted using the PBMC isolation kit (Famacs, Nanjing, China) in line with the manufacturer’s instructions, the

purity

which was

above

% for monocytes as determined by flow cytometry. The patient parameters including age, gender, comorbidity, stroke loca-

tion, collection of blood after stroke (hours/days) were listed in

Table S2

Animals

CR1

Cre/ERT2

mice were obtained from the Shanghai Nanfang Research Center for Model Organisms. Wide-type (WT) C57BL/6J

mice were obtained from the

Slake Company

(Shanghai, China)

All the animal procedures were approved by Animal Ethics Commit-

tee of China Pharmaceutical University (No. 2021-01-020) and carried out in accordance with the requirements of the National Insti-
tutes of Health for the care and use of experimental animals.

Isolation of primary mouse cortical neurons, astrocytes and microglia

Primary mouse cortical neurons, astrocytes, and microglia were isolated according to previous literature.

Primary mouse cortical

neurons were isolated from

pregnant mice

at embryonic

day

(E) 15-E16.

In brief, the brain cortices tissue of fetal mice was removed

and placed in Hanks balanced salt

solution (HBSS

). Then, the brain tissue was minced into 2 mm-wide pieces and subsequently

immersed in 2 mg/mL papain solution at 37

C for 30 min. After digestion, the dissociated neurons were cultivated in Neurobasal me-

dium supplemented with 2%

B27

and 1% Gluta

MAX

in an incubator at 37

C with 5% CO

until the cells grew mature at 7–9 days.

Microglia and astrocytes were obtained from the

mixed glial

population with different attachment methods.

Briefly, the

neonatal

Continued

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Experimental models: Organisms/strains

CR1-Cre

ERT2

Shanghai Model Organisms Center

NM-KI-200157

Oligonucleotides

Primer used, see

Table S1

Sangon Biotech

N/A

Software and algorithms

ImageJ

NIH

N/A

Prism

GraphPad

N/A

IMARIS 9.6.2

Bitplane

N/A

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

mice

(postnatal day

) were

disinfected with 75%

alcohol, and then the cerebral cortex tissue was carefully removed. After gently

dissecting the attached

meninges and

blood

vessels, the minced cerebral tissue pieces were immersed in 0.25% trypsin solution

C for 6–10 min

before terminating digestion with

6 mL

DMEM-F12

medium supplemented with 10% FBS. Afterward, the disso-

ciated

mixed glial

population was cultivated in

DMEM-F12

supplemented with 10% FBS. When

mixed glial cells

were grown

to 90–

100

% cell

density at 10 days,

the microglia were harvested from the

T75

culture flask after shaking at

180

rpm

for 2

h. Next, the

astrocytes adherent to the T75

culture flask

were digested with 0.25% trypsin solution and cultured in

12-well

plates (Jet Biofil,

Guangzhou)

for

the subsequent experiments

METHOD DETAILS

Quantitative real-time PCR (qRT-PCR)

The total RNA was extracted with TRIzol reagents (Vazyme Biotech Co., Ltd, China). Then, one microgram of mRNA was reverse-
transcribed into cDNA according to the instruction of cDNA Synthesis Kit (Vazyme Biotech Co., Ltd, China). Subsequently, the first
strand cDNA was then applied as a template for PCR amplification with

gene-

pecific primers

using SYBR Green reagent (Vazyme

Biotech Co., Ltd, China) on ABI StepOnePlus sequence detection system. The final gene expression was normalized to 18S

RNA.

The primers applied in the study are displayed in

Table S1

Western blotting analysis

The Western blotting analysis was conducted as we previously described.

70–74

First, the total protein was extracted using RIRA buffer

and then the protein concentrations were determined using the BCA kit (Thermo). After mixing with Laemmli sample buffer, the pro-
tein samples were separated electrophoretically and then wet-transferred onto 0.22

m PVDF membranes (Millipore). After blocking

with 5% non-fat dry milk in the Tris-buffered saline (TBS) (pH 7.4) containing 0.1% Tween 20 at room temperature for 2 h, the protein
membranes were sequentially immersed with the primary antibody at 4

C overnight and secondary antibody at room temperature for

1 h. Finally, the membranes were quantified using Quantity One gel image analysis system with enhanced chemiluminescence detec-
tion reagents. All unprocessed Western blotting images were displayed in

Figures S7

and

CMPK2 protein and antibody production

Full length of CMPK2 cDNA was amplified and cloned into the Nco I and Hind III sites on a pET-28a(+) vector with 6

His tag. Sub-

cequently, the recombinant plasmids were transformed into BL21(DE3) bateria and the expression of CMPK2 protein was induced by
addition of 0.5 mmol/L IPTG. Then, CMPK2 protein purification was carried out with Ni-IDA-Sepharose CL-6B affinity chromatog-
raphy column (Sangon Biotech, Shanghai). To prepare CMPK2 antibody, aa 135–265 of CMPK2 was amplified and cloned into
pET-28a-SUMO vector. After expression and purification, CMPK2 aa135-265 (1.5 mg/mL) was injected into rabbits, following by
screening with ELISA and immunoblots. Lastly, the antisera were collected and affinity-purified for further experiments.

Establishment of focal cerebral ischemia/reperfusion model

The mouse tMCAO model is a well-established model to form focal cerebral ischemia, according to previous literature.

In short,

mice are anesthetized in an animal anesthesia chamber containing 3% isoflurane and then maintained under anesthesia at 2% iso-
flurane. During the surgery, the mouse body temperature was held at 37

0.5

C with the heating pad. First, after bluntly separating

the neck muscle, the common carotid artery (CCA), external carotid artery (ECA), and internal carotid artery (ICA) of the mouse were
gently isolated without excessive wounding. Then, the silicone-coated monofilament (1800A, Jialing Biotechnology, China) was
introduced from ECA and extended to the anterior cerebral artery to block the opening of MCA. Afterward, the arterial blood flow
was occluded to maintain the cerebral ischemia for about 1 h, and the silicone-coated monofilament was subsequently extracted
to restore cerebral blood flow. During the surgery, laser speckle flow imaging was used to monitor cerebral cortical blood flow.
When cerebral blood flow was dropped below 70%, as assessed by Laser Speckle Imaging analysis system (RWD Biotechnology,
China), the tMCAO model was considered to be successfully established. As for the Sham group, the mice were exposed to the same
operation without inserting silicone-coated monofilament for 1 h.

To explore the effects of NDGA in short-term neurological deficits in the mouse ischemic stroke model, a total of 60 mice were

randomly divided into four groups: sham group, tMCAO plus vehicle group, tMCAO plus NDGA (5 mg/kg or 20 mg/kg) group at 3
h, 6 h, 24 h, and 48 h. In the mice subjected to tMCAO model, 10 mice were excluded from the analysis for unsuccessful operation
(

= 1) or hemorrhagic transformation (

= 3), and 6 mice died intraoperatively (

= 2) or postoperatively (

= 4). To explore the effects of

NDGA and CMPK2 knockdown in short-term neurological deficits in the mouse ischemic stroke model, a total of 148

CR1

Cre/ERT2

mice were randomly divided into six groups: sham plus AAV-shCon, sham plus AAV-shCMPK2, tMCAO plus AAV-shCon, tMCAO
plus AAV-shCMPK2, tMCAO plus AAV-shCon plus NDGA (20 mg/kg), tMCAO plus AAV-shCMPK2 plus NDGA (20 mg/kg). In the

CR1

Cre/ERT2

mice subjected to tMCAO model, 24 mice were excluded from the analysis for unsuccessful operation (

= 3) or hem-

orrhagic transformation (

= 6), and 15 mice died intraoperatively (

= 3) or postoperatively (

= 12). To explore the effects of NDGA in

long-term neurological deficits in the mouse ischemic stroke model, a total of 84 mice were randomly divided into three groups: sham
group, tMCAO plus vehicle group, tMCAO plus NDGA (20 mg/kg) group at 3 h, 6 h, 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, and

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

7 days. In the mice subjected to tMCAO model, 15 mice were excluded from the analysis for unsuccessful operation (

= 3), or hem-

orrhagic transformation (

= 3), and 9 mice died intraoperatively (

= 2) or postoperatively overnight (

= 7).

Assessment of cerebral infarct volume and neurological function

2, 3, 5-triphenyltetrazolium chloride (TTC, TCI, Japan) staining was applied to measure cerebral infarct volume. In brief, the mice were
sacrificed under deep isoflurane anesthesia. Then, brains were gently dissociated and cut into five coronal brain slices (1–2 mm thick)
and subsequently incubated with 2% TTC solution at 37

C for 10 min. Finally, the coronal brain slices were photographed using cam-

era, and the infarct volumes were determined with ImageJ software. The cerebral infarct volume (with correction for brain oedema)
(%) was evaluated as follows: (the contralateral hemisphere minus the non-infarct area of the ipsilateral hemisphere)/(the contralat-
eral hemisphere

100%.

In this study, Longa’s test was used to determine short-term neurological deficits in the mouse ischemic stroke model.

More-

over, long-term neurological performances were evaluated in mice of each group by the modified neurological severity score (mNSS),
foot-fault test, wire-hanging test, and rotarod test at 1, 3, 5, 7, 14, 21 and 28 days after tMCAO as previously demonstrated.

All

the behavioral tests were conducted

in a double-blind

manner.

Immunofluorescence staining

The immunofluorescence staining was conducted as we previously described.

Briefly, the dissociated brains were fixed in 4% para-

formaldehyde solution for 24 h and then sequentially dehydrated in 15% and 30% sucrose solutions at 4

C for 2–3 days until sunk. Next,

the frozen mouse brain tissue was sliced into 16

m or 40

m coronal slices using the Leica cryostat microtome. As for immunofluo-

rescence staining, the frozen mouse coronal slices and cells were fixed with 4% formaldehyde and then washed 3 times with phos-
phate-buffered saline (PBS) (pH 7.4) solution for 5 min. Afterward, the slices were immersed in PBS solution with 10% normal goat
serum or donkey serum for 1 h at room temperature to avoid non-specific staining. Next, the slices were sequentially incubated with
the primary antibodies at 4

C overnight and secondary antibodies at room temperature for 1 h. And the cell nuclei were counterstained

with Hoechst 33342. Finally, the immunofluorescence imaging was acquired and visualized using FV3000 microscope (Olympus,
Japan). As for immunofluorescence staining in PBMCs, the experiments were performed at about 1–2 h after blood collection.

Adult mouse microglial isolation and flow cytometry examination

Mouse microglia from adult mouse brains were isolated using the CD11b (Microglia) MicroBeads Kit (Miltenyi Biotec) following the
manufacturer’s instructions.

Then, the isolated microglia were blocked with rat anti-mouse CD16/32 for 20 min to avoid non-spe-

cific staining. Afterward, the cells were sequentially incubated with FVS780 and fluorochrome-conjugated primary antibodies against
CD11b

or CD45

in the dark. Finally, the samples were examined with the CytoFLEX flow cytometer (Beckman, USA).

Oxygen-glucose deprivation (OGD) treatment

To mimic the ischemic state for microglia, astrocytes, and neurons, the OGD model was established

in vitro

as previously re-

ported.

Briefly, the primary mouse microglia, astrocytes, or cortical neurons were replaced with DMEM medium without glucose

and FBS (Gibco, USA) in the hypoxia chamber (Billups-Rothenberg Inc, USA) for the indicated time at 37

C with 95% N

and 5% CO

for the following experiments.

Stereotaxic surgery

For stereotaxic surgery, the head of mouse was placed in the stereotactic frame (RWD Life Science, Shenzhen, China) in a prone
position under 2% isoflurane. Then, 1

L of rAAV-CMV-DIO-mCherry-Loxp-shRNA (Scramble)-WPRE-polyA (5

viral ge-

nomes/

L) viruses or rAAV-CMV-DIO-mCherry-Loxp-shRNA (CMPK2)-WPRE-polyA (5

viral genomes/

L) viruses using

AAV-MG1.2 vectors were injected into the target site (0.5 mm anterior to the bregma, 2.0 mm lateral (right) to the sagittal suture,
and 1.0 mm from the surface of the skull) of

CR1

Cre/ERT2

at a rate of 0.05

L/min with a 30-gauge needle for 20 min using a micro-

syringe pump (RWD Life Science, Shenzhen, China).

80–82

After injecting the

Cre

-dependent virus, the microsyringe remained in

position for 10 min before being removed. Subsequently, the mice were held on the heating pad to recuperate from the anesthesia.
After 14 days post stereotaxic surgery, the

CR1

Cre/ERT2

mice were injected intraperitoneally with 75 mg/kg tamoxifen (Sigma,

USA) per day for consecutive 5 days to induce

Cre

recombinase expression.

The AAV transfection area accounted for approx-

imately 19.25

3.26% of the total infarct area and 43.68

5.13% of the cerebral cortex infarct area.

In vivo

mouse brain magnetic resonance imaging (MRI) scanning

To assess the degree of ischemic stroke in the mice, MRI scanning was conducted at day 14 with the Bruker BioSpec 7.0-T/20-cm
MRI scanner (Bruker, Germany).

First, the mice were anesthetized and placed on the experimental bench in a prone position, with

the head fixed at the center of radiofrequency (RF) body transmit coil. Then, the MRI images were obtained using the T2 Star
sequence setting: (repetition time (TR) = 3000 ms, echo time (TE) = 45 ms, slice thickness of 0.50 mm, flip angle (FA) = 135

20 mm

20 mm in-plane resolution, and 512

256 matrix). The cerebral infarct volume (with correction for brain oedema) (%)

was evaluated as follows: (the contralateral hemisphere minus the non-infarct area of the ipsilateral hemisphere)/(the contralateral
hemisphere

100%.

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

Three-dimensional (3D) reconstruction

The 40

m brain slices were sequentially incubated with anti-Iba1 primary antibody at 4

C overnight and Alexa Fluor 488-conjugated

secondary antibody at room temperature for 1 h. After staining the cell nuclei with Hoechst 33342, 1024

1024-pixel resolution im-

ages were captured using FV3000 with 1.0

m steps in the Z-direction, followed by 3D reconstructions with IMARIS 9.0 software

(Bitplane).

Finally, the process lengths and total branch points were semi-automatically quantified using Imaris software.

LPS-primed NLRP3 inflammasome model in primary mouse microglia cells and human primary monocytes

As previously described, priming the cells with LPS followed by ATP stimulation is an established model to induce the NLRP3 inflam-
masome activation.

41–43

After seeding on 24-well plates overnight, primary mouse microglia and human primary monocytes were

simultaneously treated with ultrapure LPS (100 ng/mL) for 6 h and NDGA at the indicated concentrations, and then stimulated
with ATP (3 mM) for 30 min. Eventually, cell lysates and supernatants were collected for further experiments.

Cell transfection

To investigate the effects of target CMPK2 or 5-LOX in cells, primary mouse microglia were transfected with

Cmpk2

siRNA (si-

Cmpk2

5-Lox

siRNA (si-

5-Lox

) or negative siRNA using Lipofectamine 2000 reagent (Invitrogen, USA) for 48 h following the guide-

lines of manufacturer. The sequences of si-

Cmpk2

: GGCUUCUGAAAUAGCUAAATT; si-

5-Lox

: CCCGAGAUAUCCAGUUUGATT.

Analysis of lactate dehydrogenase (LDH) release

The contents of LDH in cell culture supernatant after priming with LPS followed by ATP stimulation or OGD treatment were deter-
mined using

LDH

kit (Beyotime Biotechnology, China) under the manufacturer’s guidelines. All samples and standards were as-

sessed in triplicate.

Edu staining experiment

To detect newly synthesized mtDNA, EdU detection was conducted using Click-iT EdU Cell Proliferation Kit with Alexa Fluor 488
(Beyotime, China

）

according to the previous reports.

Finally, immunofluorescence imaging was acquired and visualized using

FV3000 microscope.

Measurement of total mtDNA

After washing the microglia twice with sterile PBS solution, the microglia were collected to extract the total mtDNA using the FastPure
Cell DNA Isolation Mini Kit (Vazyme Biotech Co.,Ltd, China) following the manufacturers’ advisement. Then, the mtDNA contents
were assessed by qPCR amplification with gene-specific primers.

The final mtDNA expression was normalized to nuclear DNA en-

coding

Tert

. The gene-specific primers for the mtDNA applied in the study are displayed in

Table S1

Enzyme-linked immunosorbent assay (ELISA) for IL-1

cytokine

The concentrations of IL-1

cytokine in cell culture supernatant were determined using the human or mouse IL-1

ELISA kit (RD, USA)

in accordance with the manufacturer’s guidelines. All samples and standards were assessed in triplicate.

In vitro

CMPK2 kinase assay and compounds screening

The CMPK2 kinase assay was performed with 3180 compounds. The compound libraries were obtained from Selleck (L1300, L3800,
L8000) and all compounds were dissolved in DMSO at the equivalent concentration. Inhibition of CMPK2 kinase activity was deter-
mined using an ADP-Glo Kinase Assay Kit (Promega, USA).

Briefly, recombinant human CMPK2 kinase (2

g/reaction) was mixed

with the test compounds in the kinase assay buffer (5 mM MgCl

, 25 mM HEPES pH 7.2, 50 mM Tris-HCl pH 7.6–8.0, 5

M ATP,

100

M CMP). To explore the effects of NDGA on CMPK2 enzymatic activity, different concentrations of NDGA were incubated

with the CMPK2 kinase in the kinase assay buffer. Lastly, the kinase detection regent was added to determine the luminescence un-
der the Paradigm detection platform (Beckman).

Surface plasmon resonance (SPR) assay

In order to investigate whether NDGA binds with CMPK2 protein directly, the SPR binding experiment was carried out using a Biacore
8K instrument (GE Healthcare).

Recombinant CMPK2 kinase (50

g) was immobilized on the CM5 chip, followed by injection of

NDGA at different concentrations at a flow rate of 30

L/min for 350 s. Finally, the binding curves and affinity data were recorded and

assessed using Biacore Insight evaluation software.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data are displayed as mean

SD and analyzed by using the software GraphPad Prism 7.0. Depending on the dataset, statistical

analysis was carried out using unpaired Student’s t test, non-parametric Mann-Whitney test, log rank (Mantel-Cox) test, One-way
ANOVA with Bonferroni’s corrections, or two-way ANOVA with Bonferroni’s corrections with or without repeated measurements,
respectively. Differences are recognized statistically significant at

< 0.05.

Cell Reports Medicine

, 101522, May 21, 2024

Article

OPEN ACCESS

Supplemental information

Microglial CMPK2 promotes neuroinflammation and brain

injury after ischemic stroke

Authors

Xin Guan,

1,5

Sitong Zhu,

1,5

Jinqian Song,

Kui Liu,

Mei Liu,

Luyang Xie,

Yifang Wang,

Jin

Wu,

2,*

Xiaojun Xu,

3,*

and Tao Pang

1,4,6,*

Affiliations

State Key Laboratory of Natural Medicines, New drug screening center, Key Laboratory of Drug

Quality Control and Pharmacovigilance (Ministry of Education), China Pharmaceutical University,

Nanjing 210009, P.R. China.

Department of Neurology, The Second Affiliated Hospital of Nanjing Medical University, Nanjing

210011, P.R. China.

Department of Pharmacy, The Fourth Affiliated Hospital, Zhejiang University School of Medicine,

Center for Innovative Traditional Chinese Medicine Target and New Drug Research, International

Institutes of Medicine, Zhejiang University, Yiwu, Zhejiang Province 322000, P.R. China.

State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University, Nanjing 210023, P.R.

China.

These authors contributed equally

Lead contact

Correspondence: tpang@cpu.edu.cn (T.P.); xiaojunxu@zju.edu.cn (X.X.); wujin@njmu.edu.cn (J.W.)

Supplemental Table, supplementary figure titles and legends

Table S1. Sequences of primers for PCR. Related to STAR Methods.

Gene

Sense (5

′–

′

)

Anti-sense (3

′–

′

)

Primers for mouse genes

CMPK2

GGCAATTATCTCGT GGCTTC

GTAGCTATGGCGTAGGTGGC

5-LOX

ACTACATCTACCTCAGCCTCATT

GGTGACATCGTAGGAGTCCAC

18S

RNA

CTTTGGTCGCTCGCTCCTC

CTGACCGGGTTGGTTTTGAT

D-loop

AATCTACCATCCTCCGTGAAACC

TCAGTTTAGCTACCCCCAAGTTTAA

CYTB

GCTTTCCACTTCATCTTACCATTTA

TGTTGGGTTGTTTGATCCTG

COX-1

GCCCCAGATATAGCATTCCC

GTTCATCCTGTTCCTGCTCC

Tert

CTAGCT CATGTGTCAAGACCCTCTT

GCCAGCACGTTTCTCTCGTT

Primers for human genes

CMPK2

CCAGGTTGTTGCCATCGAAG

CAAGAGGGTGGTGACTTTAAGAGGG

18S

RNA

CTTTGGTCGCTCGCTCCTC

CTGACCGGGTTGGTTTTGAT

COX-1

CCCAATCTCTACCAGCATC

GGCTCATAGTATAGCTGGAG

CYTB

GCC TGC CTG ATC CTC CAA AT

AAG GTA GCG GAT GAT TCA GCC

TERT

CTAGCTCATGTGTCAAGACCCTCTT

GCCAGC ACGTTTCTCTCGTT

D-LOOP

CTAAATAGCCCACACGTTCC

TAGGATGAGGCAGGAATCAA

Table S2. Patient characteristics. Related to STAR Methods.

Patient
no.

Age Gender Comorbidity

Stroke location

Collection
of

blood

after stroke

Infarct
size (cm

)

Hypertension

Right basal ganglia

5 h

3.18

Hypertension

Right

lateral

ventricles

8 h

5.06

Hypertension,
hepatic dysfunct
ion

Right basal ganglia

3 h

3.46

Hypertension,
aortic

valve

calcification

Right

lateral

ventricles

6 h

Fatty

liver,

hypertension,
thyroid nodule

Left

lateral

ventricles

4 h

1.66

Hyperlipidemia,
hypertension

Corpus

callosum,

left

lateral

ventricle

6 h

18.35

Left

internal

carotid

artery

occlusion,
vertebral

artery

Right

lateral

ventricles

11 h

55.75

Hypertension,
cerebral vessel
stenosis

Right

basal ganglia

7 h

33.93

Hypertension,
atherosclerosis

Left basal ganglia

5 h

10.56

Atrial
fibrillation,
hypertension

Right

basal ganglia,

right temporal frontal
parietal region

7 h

Continued

Atrial
fibrillation,
hypertension

Right basal ganglia

5 h

26.84

Atrial
fibrillation,
hypertension

Basal ganglia

6 h

Cardiac valvular
disease,

atrial

fibrillation

Right basal ganglia

9 h

194.08

Hypertension,
cardiac valvular
disease

Right basal ganglia,
parahippocampal
gyrus

2 h

10.77

Left

atrial

myxoma

Right

lateral

ventricle

2 h

4.70

Hypertension

Left frontal, temporal,
and occipital lobes,
basal ganglia

2 h

9.32

Type 2 diabetes
mellitus,
coronary

atheros

clerosis

Right basal ganglia

8 h

Hypertension,
type
2 diabetes

mellit

Left basal ganglia

7 h

Hypertension,
type

diabetes

mellit

Right

lateral

ventricle

6 h

Hypertension,
hepatic dysfunct
ion

Left basal ganglia

3 h

7.42

Coronary

athero

sclerosis

hypertension,
fatty liver

Right basal ganglia

2 h

2.63

Continued

Cardiac valvular
disease

Left

frontal lobe,

left

lateral parietal

3 h

7.34

Atrial fibrillatio
n

Basal ganglia

5 h

94.75

Atrial fibrillatio
n, hypertension

Left occipitoparietal lo
be

3 h

2.27

Hypertension

Basal ganglia

9 h

Type
2 diabetes

mellit

us,

fatty liver

Left

lateral parietal,

basal ganglia

6 h

Hypertension,
type
2 diabetes mellit
us

Left basal ganglia

6 h

Hypertension,
type
2

diabetes

mellit

Right

lateral

ventricle

7 h

Hypertension,
hyperlipidemia

Right

frontal

lobe

12 h

Hypertension,
arteriolosclerosi
s

Right basal ganglia

5 h

7.93

Aortic sclerosis,
cardiac valvular
disease

Right

frontal

lobe,

basal ganglia

7 h

117.79

Note:

M, male; F, female; /, the person cannot be calculated for the precise ischemic volume for only

having paper version of magnetic resonance imaging (MRI) which indicated symptoms of acute

ischemic stroke.

Quantification of CMPK2 protein expression levels in microglia/macrophages

isolated from the sham-

operated brains of AAV-shCon or AAV-shCMPK2

CR1

Cre/ERT2

with CD11b

magnetic beads.

= 6.

(E) Western blotting images and quantification of the CMPK2 protein expression levels in

microglia/macrophages

isolated from the ischemic brains of AAV-shCon or AAV-shCMPK2

CR1

Cre/ERT2

with CD11b

magnetic beads.

= 6. (F and G) Representative immunofluorescence

images of microglia stained with CMPK2 and quantitative analysis of CMPK2

Iba1

cells in the peri-

infarct region of the

CR1

Cre/ERT2

mice after AAV-shRNA (Scramble) or AAV-shRNA (CMPK2)

microinjection. Scale bar: 50 μm.

= 6. Results are represented as means ± SD.

***

< 0.001

vs.

sham

plus AAV-shCon group;

###

< 0.001

vs.

tMCAO plus AAV-shCon group.

(H) Western blotting images

and quantification were performed to detect CMPK2 protein expression in primary mouse microglia

cells from

CR1

Cre/ERT2

mice at postnatal day 1-3 with or without stimulation with MCSF (10 ng/mL)

for 48 h.

= 6. Results are represented as means ± SD.

***

< 0.001

vs.

AAV-shCon group;

###

< 0.001

vs.

MCSF plus AAV-shCon group.

after stimulation with LPS (100 ng/mL) for the indicated time.

= 4. (C) The qRT-PCR was performed

to detect

Cmpk2

gene expression in primary mouse microglia cells after stimulation with LPS (100

ng/mL) for the indicated time.

= 3. (D and E) The protein expression of CMPK2 in primary mouse

microglia transfected with NC siRNA or CMPK2 siRNA for 48 h. Results are expressed as means



SD. One-way ANOVA with Bonferroni’s corrections.

= 6.

***

< 0.001

vs.

the NC siRNA group. (F

and G) The protein expression of CMPK2 with or without LPS stimulation in primary mouse microglia

transfected with NC siRNA or CMPK2 siRNA.

= 4. (H, I) The protein expressions of cleaved

100

caspase-1, mature IL-1



, and N-GSDMD in the culture supernatant of microglia transfected with NC

101

siRNA or CMPK2 siRNA, and then primed with LPS (100 ng/mL) for 6 h, followed by stimulation

102

with ATP (3 mM) for 30 min.

= 4. (J) The level of IL-1



protein was assessed in the supernatant of

103

microglia transfected with NC siRNA or CMPK2 siRNA after LPS plus ATP stimulation as determined

104

by ELISA kit.

= 4. (K, L) LDH release was determined in the supernatant of microglia transfected

105

with NC siRNA or CMPK2 siRNA after LPS plus ATP stimulation or OGD model

. n

= 4 or 5. Results

106

are represented as means ± SD.

***

< 0.001

vs.

CN plus NC siRNA group;

###

< 0.001

vs.

LPS + ATP

107

plus NC siRNA group or OGD plus NC siRNA group.

108

109

cells

in vitro

. Related to Figure 5.

(A-C) The

Cox-1

CYTB

D-loop

expressions were detected by PCR

112

and normalized to

Tert

(nuclear DNA) in primary mouse microglia cells after stimulation with LPS

113

(100 ng/mL).

= 3. Results are represented as means ± SD. (D-G) Representative immunofluorescence

114

images and quantification of newly

synthesized

DNA visualized by EdU labeling in primary mouse

115

microglia cells subjected to LPS (100 ng/mL) stimulation or OGD condition for 6 h. Scale bar: 5 μm.

116

= 6. Results are represented as means ± SD.

***

< 0.001

vs.

CN plus NC siRNA group;

###

< 0.001

vs.

117

LPS or OGD plus NC siRNA group. (H-K) Representative immunofluorescence images and

118

quantification of microglia stained with 8-OHdG after subjected to LPS plus ATP stimulation or OGD

119

condition. Scale bar: 20 μm.

= 4. Results are represented as means ± SD.

***

< 0.001

vs.

CN plus

120

NC siRNA group;

###

< 0.001

vs.

LPS + ATP or OGD plus NC siRNA group.

121

immunofluorescence images and quantitative analysis of cellular ROS level visualized by Mito-SOX (2

136

μM) staining in primary mouse microglia cells subjected to LPS plus ATP condition for 6 h. Scale bar:

137

20 μm.

= 4. Results are represented as means ± SD.

***

< 0.001

vs.

CN plus NC siRNA group. (C

138

and D) The protein expressions of cleaved caspase-1, mature IL-1



, and N-GSDMD in the culture

139

supernatant of LPS-primed microglial cell that were pretreated with 1 mM H

for 6 h in the presence

140

of ATP.

= 4. Results are represented as means ± SD.

***

< 0.001

vs.

CN plus NC siRNA group;

###

141

< 0.001

vs.

LPS + ATP plus NC siRNA group,

$$$

< 0.001

vs.

LPS + ATP plus NC siRNA group,

&&&

142

< 0.001

vs.

LPS +ATP+H

plus NC siRNA group. (E and F) The microglial cells were

143

immunofluorescently stained with 8-OHdG and ASC after stimulation with LPS plus ATP condition

144

with or without CMPK2 knockdown. Scale bar: 5 μm.

= 4. Results are represented as means ± SD.

145

***

< 0.001

vs.

CN plus NC siRNA group;

###

< 0.001

vs.

LPS + ATP plus NC siRNA group.

146

147

caspase-1, N-GSDMD, and mature IL-1β in the culture supernatant in microglia transfected with NC

185

siRNA or 5-LOX siRNA, and then incubated with LPS with or without NDGA treatment, followed by

186

stimulation with ATP.

= 4. Results are represented as means ± SD.

***

< 0.001

vs.

CN plus NC

187

siRNA group;

###

< 0.001

vs.

LPS+ATP plus NC siRNA group;

$$$

< 0.001

vs.

LPS+ATP plus 5-LOX

188

siRNA group. (D and E) The protein expression of cleaved caspase-1, N-GSDMD, and mature IL-1β in

189

the culture supernatant in microglia transfected with NC siRNA or CMPK2 siRNA, and then incubated

190

with LPS with or without NDGA treatment, followed by stimulation with ATP.

= 4. Results are

191

represented as means ± SD.

***

< 0.001

vs.

LPS+ATP plus NC siRNA group. (F and G) The protein

192

expressions of cleaved caspase-1, mature IL-1



, and N-GSDMD in the culture supernatant of LPS-

193

primed microglial cells with or without NDGA administration that were pretreated with 1 mM H

for

194

6 h in the presence of ATP.

= 4. Results are represented as means ± SD.

***

< 0.001

vs.

CN group;

195

###

< 0.001

vs.

LPS + ATP group,

$$$

< 0.001

vs.

LPS + ATP group,

&&&

< 0.001

vs.

LPS + ATP +

196

group.

197

198

significantly reduced the release of IL-1



in PBMC isolated from stroke patients. The PBMC were

225

simultaneously incubated with 100 ng/mL LPS and 10 µM NDGA or the same volume of DMSO for 6

226

h followed by exposure to 3 mM ATP for 30 min.

= 17. Results are represented as means ± SD.

***

227

0.001

vs.

LPS plus ATP group. (B) The correlation

analysis was performed between

CMPK2

transcript

228

levels in blood and inhibition rate of NDGA on IL-1



in monocytes of ischemic stroke patients.

= 17.

229



in monocytes isolated from healthy subjects. The

230

monocytes were simultaneously incubated with 100 ng/mL LPS and 10 µM NDGA or the same volume

231

of DMSO for 6 h followed by exposure to 3 mM ATP for 30 min.

= 12. (D and E) NDGA reduced 8-

232

OHdG expression in PBMC isolated from stroke patients after priming with LPS followed by ATP

233

stimulation as determined by immunofluorescence staining. Scale bar: 20 μm.

= 6. Results are

234

represented as means ± SD.

***

< 0.001

vs.

CN group;

###

< 0.001

vs.

LPS plus ATP group.

235

236