SerpinA3N attenuates ischemic stroke injury by reducing apoptosis and neuroinflammation

Abstract Objective To assess the effect of serine protein inhibitor A3N (serpinA3N) in ischemic stroke and to explore its mechanism of action. Methods Mouse ischemic stroke model was induced by transient middle cerebral artery occlusion followed by reperfusion. The expression pattern of serpinA3N was assessed using immunofluorescence, Western blot analysis, and real‐time quantitative PCR. An adeno‐associated virus (AAV) and recombinant serpinA3N were administered. Additionally, co‐immunoprecipitation‐mass spectrometry and immunofluorescence co‐staining were used to identify protein interactions. Results SerpinA3N was upregulated in astrocytes and neurons within the ischemic penumbra after stroke in the acute phase. The expression of serpinA3N gradually increased 6 h after reperfusion, peaked on the day 2–3, and then decreased by day 7. Overexpression of serpinA3N by AAV significantly reduced the infarct size and improved motor function, associated with alleviated inflammation and oxidative stress. SerpinA3N treatment also reduced apoptosis both in vivo and in vitro. Co‐immunoprecipitation‐mass spectrometry and Western blotting revealed that clusterin interacts with serpinA3N, and Akt‐mTOR pathway members were upregulated by serpinA3N both in vivo and in vitro. Conclusions SerpinA3N is expressed in astrocytes and penumbra neurons after stroke in mice. It reduces brain damage possibly via interacting with clusterin and inhibiting neuronal apoptosis and neuroinflammation.

SerpinA3N, a murine orthologue of human α-1-antichymotrypsin, is a member of the serpin superfamily of protease inhibitors. 2 Its folding is highly conserved, consisting of 8~9 α-helices, 3 β-sheets, and a solvent-exposed stretch termed the reactive center loop (RCL), which interacts with the protease active site to promote protease activity. 3,4 A structure analysis revealed two features of serpinA3N: (1) the residues of the RCL are partially inserted into the A β-sheet, a structure motif that is associated with ligand-dependent activation in other serpins similar to non-heparin-activated antithrombin; and (2) two positively charged patches that might be associated with the binding of negatively charged entities such as DNA or glycosaminoglycans. 2 Several proteases have been identified as substrates for serpinA3N, including antichymotrypsin, cathepsin G, 2 leukocyte elastase, 2,5 granzyme B, 6 and matrix metalloprotein 9 (MMP9). 7 As an acute phase protein, serpinA3N is secreted in response to inflammation 7-10 and glucocorticoids 11 as well as in various pathological conditions, such as an aortic aneurysm 12 and colitis. 13 In the central nervous system (CNS), serpinA3N is regarded as a potential marker of reactive astrogliosis. 1,14 Its function in the CNS is controversial. On the one hand, it induces neuroprotection, attenuates neuropathic pain, 5 and reduces the severity of multiple sclerosis 6 by inhibiting proteases. However, on the other hand, overexpression of serpinA3N in mouse hippocampus abolishes the protective effects of melatonin on trimethyltin chloride-induced neuroinflammation and neurotoxicity. 15 In ischemic stroke, its role remains unclear.
In the present study, we investigated serpinA3N expression levels and described its temporospatial distribution pattern in a mouse ischemic stroke model. We also evaluated whether it is neuroprotective against neuronal ischemic injury. Finally, the molecular mechanisms underlying its protective effects were explored. reported following the ARRIVE guidelines. 16

| Transient middle cerebral artery occlusion
A transient middle cerebral artery occlusion (MCAO) model was used to induce a focal cerebral ischemic stroke as previously described with modification. 17 Briefly, mice were anesthetized with intraperitoneal injection of chloral hydrate (400 mg/kg). Cerebral focal ischemia was established by intraluminal occlusion of the right middle cerebral artery using a silicone rubber-coated nylon monofilament (Guangzhou Jialing biotech Co., Ltd.), which was inserted and advanced through the carotid artery. Occlusion was verified by laser Doppler flowmetry (Moor Instruments, Inc.) with >70% reduction in regional blood flow perfusion. One hour after occlusion, the filament was withdrawn to allow for reperfusion and the general carotid artery was ligated. The skin was sutured, and the animal was allowed to recover. In sham-operated mice, the same surgical procedure was performed, except that the monofilament was not inserted.
Four weeks prior to MCAO, mice were anesthetized and placed in a stereotaxic apparatus (Narishige). 2 μl of AAV suspension (2 × 10 9 genome copies/mouse) was injected through a 36-gauge glass cannula connected to a 2μl Hamilton syringe mounted on a microinjection pump (Univentor). The stereotaxic injection coordinates for the striatum were 2 mm posterior to bregma, 1.5 mm right of the midline, and 2.5 mm below the pia. The needle was kept in place for another 5 min before the cannula was slowly withdrawn to prevent reflux. The skin incision was then sutured, and the animal was kept warm with a heat blanket before being returned to the cage.

| Brain infarct volume
The mice were deeply anesthetized by chloral hydrate (600 mg/ kg), and the brains were removed from the skull and were frozen at −20°C for 10 min. Frozen brains were then cut into five sections in the coronal plane and stained with 2,3,5-triphenyltetrazolium chloride solution (TTC, Sigma-Aldrich) at 37°C for 30 min before fixed in 4% formaldehyde for 10 min. The infarct areas were then measured, and infarct volumes were calculated using Image J software (NIH) by a laboratory assistant who was blinded to the study groups.

| Mitochondria extraction
A mitochondria extraction kit (tissues) (Beyotime) was used to extract mitochondria according to the manufacturer's instructions. In brief, 80 mg of brain tissue was cut into small pieces and washed in PBS for 3 times. 640 µl mitochondria extraction buffer A was added, and tissue was ground for 20 s. After centrifugation at 600 g for 5 min, the supernatant was collected and centrifuged again at 11,000 g for 10 min. The supernatant containing cytoplasm was discarded, and pallet containing mitochondria was collected for further analysis. Gibco, Thermo Fisher) supplemented with 1% penicillin/streptomycin (Gibco). Cells were maintained at 37°C with 5% CO 2 for 10 days until an astrocyte monolayer was formed. 20 Primary microglia were shaken off at 180 rpm for 2h at 37°C and sub-cultured in D10 media.

Systems) for 24 h.
Cortical neurons were dissected from E18 mouse embryos. After digestion with trypsin, neuronal cells were suspended in high glucose DMEM (Gibco) containing 10% (v/v) equine serum and 25 μM L-glutamine. Cells were seeded at a density of 5 × 10 4 /well in 6-well tissue culture plates coated with 0.5 mg/ml poly-L-lysine (Gibco).

| Cell viability assay
Neuron viability was determined using the Cell Counting Kit-8 (CCK-8, Dojindo) following the manufacturer's instructions. In brief, 50 μl of CCK-8 reagent was added to each well for another 4 h at 37°C.
Absorbance at 450 nm was measured using a microplate reader (BioTek).

| Quantitative real-time polymerase chain reaction
Quantitative real-time polymerase chain reaction (qRT-PCR) was performed as previously described. 17 Total RNA was extracted from tissues or cultures using RNAfast200 kit (Fastagen Biotech) according to the manufacturer's instructions. PCR was then performed

| Immunohistochemistry
A modified immunofluorescence protocol was used based on previous reports. 21,22 Briefly, the mice were sacrificed and perfused through the aorta with a 0.9% NaCl solution. The brains were dissected, fixed in 4% paraformaldehyde, and then dehydrated with 25% sucrose. After rapidly freezing, floating slices were prepared by cutting in the coronal plane (20 μm in thickness, Leica cryostat). After washing with PBS, brain slices were incubated in antiserum solution (10% normal bovine serum, 0.2% Triton X-100, 0.4% sodium azide in 0.01 mol/L PBS pH 7.2) for 30 min, followed by sequential incubation with primary antibodies (overnight at 4°C, Table 3) and Cy3 or FITC conjugated secondary antibodies (1:400, 2 h at room temperature, Jackson ImmunoResearch Labs). Images were taken with a Nikon digital camera DXM1200 (Nikon) attached to a Nikon Eclipse E600 microscope (Nikon) or with a confocal microscope (Zeiss LSM 710).
In this study, neurons, reactive astrocytes, macrophage/microglia, and oligodendrocyte lineage cells were, respectively, indicated by NeuN + , 23 S100B + , 24 iba1 + , 25 and olig2 + cells. 26 SerpinA3N + cells and double-positive cells were manually counted by two laboratory assistants who were blinded to the study groups.

| Western blot and co-immunoprecipitation
Tissue lysates were prepared with RIPA buffer (30 mM HEPES [PH Western blotting was performed as described previously. 17 Protein fractions were separated by 10% or 12% SDS-PAGE and transferred onto nitrocellulose membranes. After blocking with 5% bovine serum albumin in 0.1% (v/v) Tween-20 in tris-buffered saline (TBS), membranes were sequentially incubated with primary antibodies (Table 3) and HRP-conjugated secondary antibodies.
Protein bands were developed with enhanced chemiluminescence (ECL) substrate solution (Beyotime) and visualized using a BIO-RAD Molecular Imager (Bio-Rad laboratories Inc).
Co-immunoprecipitation (Co-IP) was performed as previously described. 27 Briefly, 300-500 μl of tissue lysates was incubated with 0.5-2 μg of the corresponding antibodies ( in sample-loading buffer, then subjected to SDS-PAGE and Western blotting as described above.

| Liquid chromatography-tandem mass spectrometry
The proteins pulled-down by IP were subjected to liquid chromatography-tandem mass spectrometry (LC-MS-MS) analysis performed by Omics Space, Shanghai, China. Shotgun proteomics were used allowing for powerful separation of liquid chromatography in combination with highly sensitive and selective mass analysis.
Normality was tested using Shapiro-Wilk test. Data of normal distribution are expressed as the mean ± SEM and evaluated using an unpaired t test or ANOVA followed by the Tukey's post hoc test.
Data of non-normal distribution are expressed as median [quartile] and evaluated using Mann-Whitney U test. Significance was set at p < 0.05.

| Temporospatial distribution patterns of serpinA3N after stroke
To examine the temporal expression pattern of serpinA3N after stroke, serpinA3N mRNA and protein expression levels were detected at different time points. We found that serpinA3N mRNA expression increased 24 h after reperfusion, peaked at 2 days, and then gradually declined ( Figure 1A). SerpinA3N protein expression showed a similar pattern, which increased at 24 h, peaked at 3 days, a bit delayed compared to mRNA levels, and then gradually decreased ( Figure 1B).
The spatial distribution pattern of serpinA3N was examined by immunofluorescence analysis 24 h after reperfusion. Infarct core, penumbra, and non-infarct zone were identified depending on the intensity of NeuN signals and the morphology of Iba1 + microglia ( Figure 1C & Figure S1A). Increased serpinA3N expression was observed in the penumbra and non-infarction zone ( Figure 1C).
In the penumbra, most of the serpinA3N expressing cells were NeuN + ( Figure 1D). Interestingly in the non-infarction zone, ser-pinA3N expressing cells were NeuN − and significantly smaller in size ( Figure 1E). Further analysis showed those smaller serpinA3N + cells were actually S100b + (Figure 1F), indicating reactive astrocytes. 24

| Overexpression of serpinA3N improves neurologic function and reduces infarct volume following stroke
SerpinA3N-overexpressing AAV was injected intracranially into the striatum. Three weeks after, AAV was widely distributed in the striatum ( Figure S1A), predominantly in the penumbra area ( Figure S1B).
As for the cellular distribution, we observed significantly colocalization of ZsGreen with NeuN ( Figure S1C), suggesting neuronal infection of the AAV.
Overexpression effects were verified 3 weeks after AAV injection by qRT-PCR (Figure 2A), Western blot ( Figure 2B), and immunostaining ( Figure 2C). Neurologic functions in mice 30 h after MCAO were evaluated with Bederson's and Clark's systems, while no difference was found in the Bederson's system ( Figure 2D). SerpinA3N overexpression resulted in lower scores in the Clask's system ( Figure 2E).
In accordance with the neurologic tests, serpinA3N overexpression significantly reduced infarct volumes ( Figure 2F).

| SerpinA3N inhibits the pro-inflammatory and oxidative responses after stroke
To evaluate the effects of serpinA3N on pro-inflammatory responses after stroke, we measured the expression of inflammatory cytokines interleukin (IL)-6 and tumor necrosis factor (TNF)α and found that IL-6 and TNFα were significantly lower in serpinA3Noverexpressed mice at 24 h poststroke ( Figure 3A,B). We also evaluated the oxidative stress molecules including p22 phox , p47 phox , and p67 phox , critical components for superoxide generation through the NAPDH oxidase system, and cyclooxygenase (cox)-2 and nitric oxide synthase 2 (encoding inducible nitric oxide synthase [iNOS]), two critical enzymes in the synthesis of reactive oxygen species and nitric oxide. We found that the mRNA expression of p22 phox subunit and cox2 ( Figure 3C) and the protein level of iNOS ( Figure 3D) were significantly decreased by serpinA3N overexpression.
Microglia are brain resident immune cells and are a critical source of neuroinflammation and oxidative stress. [28][29][30] To determine whether serpinA3N directly affects microglial phenotype, primary microglial cultures were treated with LPS to induce M1 polarization in the presence of serpinA3N. We found that serpinA3N failed to alter M1 and M2 markers 31-33 ( Figure S2) after LPS stimulation.
These results indicate that serpinA3N reduces inflammation and oxidative stress indirectly, rather than directly activating microglia or changing their polarizing status.

| SerpinA3N decreases apoptosis both in vitro and in vivo
Apoptosis serves as a major mechanism responsible for neuronal loss after ischemic stroke. We therefore detected the effect of serpinA3N on apoptosis both in vitro and in vivo. In primary neuronal cultures, a dose-response experiment of recombinant serpinA3N was performed and founded that serpinA3N treatment promoted cell survival measured 4 hours after OGD/R with the dose of 50 ng/ml ( Figure 4A, Figure S3). Both phospho-p38 (p-p38) and neuronal nitric oxide synthase (nNOS) are associated with neuronal apoptosis, 34,35 and both were found to be downregulated by serpinA3N ( Figure 4B). As expected, apoptosis indi-

| Identification of clusterin as a SerpinA3Ninteracting protein
We then try to identify the molecular mechanisms of serpinA3N's protective effects by detecting serpinA3N-interacting proteins. from each brain are shown in Table 4. A total of 47 unique proteins were identified, among which 3 of them (SerpinA3K, Clusterin, and Rps25) were significantly different (1.5 folds and p < 0.05) between the two groups ( Figure S4). Combining what we found with the data from the interaction network of serpinA3N (http://www.strin g-db.org/, Figure S4), we focused on clusterin, a member of small heat shock protein family and protein chaperone associated with apoptosis. 36 Co-IP/MS results indicated that most of the serpinA3Ninteracting proteins are associated with protein synthesis and oxidation respiratory chain enzymes, which are located in mitochondria. Western blotting with brain tissue lysates proved the presence of serpinA3N ( Figure S5A), which was predominantly in the cytosol while barely detectible in the mitochondria ( Figure   S5B).
To further confirm the interaction between serpinA3N and clusterin, we performed a Co-IP assay. The brain lysates were IPed by anti-clusterin α subunit (clusterinα), and serpinA3N was blotted ( Figure 5A), and then, IP-ed by anti-serpinA3N with clusterinα blotted ( Figure 5B). Both consistently confirmed the binding of serpinA3N with clusterinα. As expected, immunofluorescence double-staining results confirmed that serpinA3N and clusterinα are colocalized in cortical neurons after stroke ( Figure 5C).
We then detected the effect of serpinA3N on clusterinα expression. Western blot analysis showed that serpinA3N overexpression reduced clusterinα in mouse brain 30h after MCAO ( Figure 5D), and in cultured neuron after OGD/R ( Figure 5E). These findings suggest that serpinA3N reduced clusterinα, probably through their interaction.

| SerpinA3N increases activation of Akt signaling pathway
Previous reports demonstrated that clusterin enhances cell survival  Figure 6A-B). Similar findings were also observed in mouse brain lysates after MCAO ( Figure 6C). These results suggest that serpi-nA3N protects the neurons against apoptosis associated with activation of the Akt signaling pathway.

| DISCUSS ION
In the present study, we confirmed the upregulation of serpinA3N after ischemic stroke and described its temporospatial distribution poststroke. We then reported that overexpression of serpinA3N in vivo elicited neuroprotection associated with suppressed neuroinflammation, oxidative stress, and neuronal apoptosis. Finally, we proved the molecular interaction between serpinA3N and clusterin, a molecular chaperone, as well as the participation of Akt-mTOR pathway, which may serve as the underlying mechanism for serpi-nA3N's antiapoptotic effects.
Consistent with previous study 1 , we found that serpinA3N expression was upregulated at 6-12 h after reperfusion and lasted for 3 days. Notably, although the serpinA3N expression levels began to drop after 3 d, they remained significantly higher even 7 d after In the present study, serpinA3N was barely detectable in mitochondrial extracts from whole brain tissue. This is in concert with previous reports, which showed that serpinA3N was a secreted In the present study, we also reported that serpinA3N de-  A drawback of the present study is that we did not perform ser-pinA3N knockout or knockdown, given that serpinA3N itself is upregulated post-MCAO. Nevertheless, we have proved that additional upregulation of serpinA3N expression elicited beneficial effects against stroke, which has some translational value. Further study should use knockout or knockdown approaches to prove the indispensable role of serpinA3N in neuroprotection. Another drawback is that only male animals were included in the present study. Sex dimorphism is well-known in microglial inflammatory response. 40,69 Therefore, the results from this study may not be able to be generalized into female individuals.

| CON CLUS IONS
SerpinA3N is expressed in astrocytes and penumbral neurons after stroke in mice and reduces damage possibly via interacting with clusterin and inhibiting neuronal apoptosis.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

AUTH O R CO NTR I B UTI O N S
Yu Zhang made equal contribution in roles/writing-original draft and methodology. Qianbo Chen involved in roles/writing-original draft and data curation. Dashuang Chen involved in formal analysis. Wenqi Zhao involved in data curation. Haowei Wang designed the methodology.
Mei Yang involved in data curation and formal analysis. Zhenghua Xiang conceptualized the study and designed the methodology. Hongbin Yuan involved in conceptualization and writing-review and editing.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.