[1] 20w Possible involvement of progranulin in the protective effect of elastase inhibitor on cerebral ischemic injuries of neuronal and glial cells

[1] 198w In a previous study, we demonstrated that neutrophil elastase is activated in the brain parenchyma after cerebral ischemia, which enzyme cleaves progranulin (PGRN), an anti-inflammatory factor. In that study, we also found that sivelestat, a selective neutrophil elastase inhibitor, attenuates ischemia-induced inflammatory responses. However, it was not clear whether this anti-inflammatory effect was due to the direct effect of sivelestat. In this study, we evaluated the effects of sivelestat or recombinant PGRN (rPGRN) on cell injuries in cultured neurons, astrocytes, and microglia under oxygen/glucose deprivation (OGD) conditions. We demonstrated that OGDinduced neuronal cell injury, astrocyte activation, and increased proinflammatory cytokines caused by microglial activation, were suppressed by rPGRN treatment, whereas sivelestat had no effect on any of these events. These results indicate that the anti-inflammatory responses after in vivo cerebral ischemia were not due to the direct action of sivelestat but due to the suppression of PGRN cleavage by inhibition of elastase activity. It was also suggested that the pleiotropic effect of rPGRN could be attributed to the differentiation of M1 microglia into anti-inflammatory type M2 microglia. Therefore, the inhibition of PGRN cleavage by sivelestat could contribute to the establishment of a new therapeutic approach for cerebral ischemia.

[1] 125w Cerebral ischemia leads to cell death due to a reduced supply of glucose and oxygen to brain tissue. Glial cells (astrocytes and microglia) support neuronal functions by supplying energy sources and removing of cellular debris. Furthermore, activated glial cells, formed when neurons are damaged during the acute ischemic phase, release reactive oxygen species, inflammatory cytokines, and chemokines, all of which are involved in the exacerbation of the inflammatory responses and tissue damages. Although glial cells in the subacute stage of stroke maintain the activity, various factors such as anti-inflammatory cytokines and growth factors, which protect and repair tissue damages, are also produced. Therefore, controlling the activity of post-ischemic glial cells is important for a therapeutic strategy at the acute and/or subacute stage of cerebral ischemia.

[2] 187w In our previous study, we demonstrated that the activity of neutrophil elastase is increased in the brain parenchyma after cerebral ischemia, cleaving progranulin (PGRN) to generate granulin (GRN), and that GRN is potentially involved in the inflammatory responses after cerebral ischemia (Horinokita et al., 2019). Sivelestat is a selective neutrophil elastase inhibitor and does not affect other proteases such as plasmin, thrombin, kallikrein, cathepsin B or collagenase I (Kawabata et al., 1991). In addition, sivelestat has been shown to improve paraplegics and to have a protective effect on motoneurons in a spinal cord ischemia model (Yamauchi et al., 2006). The expression of inflammatory cytokines (TNF-α and IL-6) in a model of hepatic ischemiareperfusion is also suppressed by treatment with sivelestat (Uchida et al., 2010). However, the mechanisms underlying the effect of sivelestat on inflammatory responses after cerebral ischemia are not fully understood. In this sense, we confirmed that sivelestat attenuates ischemia-induced inflammatory responses in the brain. However, it is not clear whether this anti-inflammatory effect was due to the direct effect of sivelestat or to the indirect effects of the PGRN, whose cleavage was inhibited by sivelestat.

[3] 28w In the present study, we evaluated the direct effects of sivelestat or recombinant PGRN (rPGRN) on cell injuries in cultured neurons, astrocytes, and microglia under the OGD conditions.

[1] 146w We first examined changes in the levels of cleaved caspase-3 and Bax proteins in the cerebral cortex on day 1 after microsphere-induced cerebral embolism (ME). The levels of cleaved caspase-3 (Fig. 1A) and Bax (Fig. 1B) proteins in ME-operated rats were significantly increased as compared with those of sham-operated rats. We next assessed the effects of an elastase inhibitor, sivelestat, on the levels of cleaved caspase-3 and Bax on day 1 after ME. The increased levels of cleaved caspase-3 and Bax after ME were reduced by the administration of sivelestat (Fig. 1A and B). In addition, the effect of administration of sivelestat on the level of Bcl-2 protein on day 1 after ME was also examined. The level of bcl-2 protein of ME-operated rats was decreased as compared with that of sham-operated rats (Fig. 1C), and this decrease was attenuated by administration of sivelestat (Fig. 1C).

[2] 116w We next examined the effect of sivelestat on the number of cleaved caspase-3-positive cells on day 1 after ME by immunohistochemical analysis. The number of cleaved caspase-3-positive cells after ME was significantly increased as compared with that of the sham-operated rats (Fig. 2A and B), and the administration of sivelestat reduced this increase after ME (Fig. 2A and B). To further examine the effect of sivelestat on cell injury, we stained cells with fluoro-jade C (FJC), a marker of degenerating neurons. The number of FJC-positive cells was significantly increased on day 1 after ME (Fig. 3A and B), and this increased number of cells after ME was inhibited by treatment with sivelestat (Fig. 3A and B).

[3] 135w We next examined the effect of sivelestat on cell viability by performing the XTT assay using cultured neurons. Treatment with sivelestat in the normoxia group did not affect cell viability at any of the concentrations tested (Fig. 5A). The cell viability was significantly decreased by OGD/R and sivelestat treatment was not preventive at the doses examined under this condition (Fig. 5A). Furthermore, the effects of sivelestat treatment on the levels of cleaved caspase-3, bax, and bcl-2 in cultured neurons under OGD/R were examined. The levels of cleaved caspase-3 and bax were significantly increased under the OGD/R condition compared to those under normoxia and these increased levels were not affected by treatment with sivelestat (Fig. 5B and C), and the decrease in the level of bcl-2 under OGD/R was not reversed by the treatment (Fig. 5D).

[4] 129w We next examined the effect of rPGRN on neuronal cell viability by using the XTT assay and on the levels of cleaved caspase-3, bax, and bcl-2. Although treatment with rPGRN in the normoxia group did not affect cell viability at any of the concentrations tested, the decreased cell viability induced by OGD/R was attenuated by treatment with rPGRN (Fig. 5E). We next examined the effects of rPGRN treatment on the levels of cleaved caspase-3, bax, and bcl-2 in the cultured neurons under OGD/ R. The increased levels of cleaved caspase-3 and bax under the OGD/R condition compared with those under normoxia were reversed by treatment with rPGRN (Fig. 5F and G). Furthermore, the decrease in the level of bcl-2 under OGD/R was lessened by treatment with it (Fig. 5H).

[5] 162w We next examined the number of Iba-1-positive microglial cells in primary cultures of glial cells after OGD treatment. The number of Iba-1positive microglial cells was not different between normoxia and the OGD condition after incubation with or without sivelestat (Fig. 7B). To determine the number of M1 microglia, we immuno-stained primary cultures of glial cells with anti-iNOS and anti-Iba-1 antibodies. Almost no iNOS-positive microglial cells were detected under the normoxic condition, whereas the number of iNOS-positive microglia was significantly increased under the OGD condition (Fig. 7A and C). There were no detectable changes in the number of iNOS-positive microglia with or without treatment with sivelestat under the OGD condition (Fig. 7C). Furthermore, the expression of pro-inflammatory cytokines TNF-α and IL-1β mRNAs were determined. The mRNA levels of these proinflammatory cytokines were significantly increased under the OGD condition compared with those under normoxia (Fig. 7D and E). Treatment with sivelestat did not affect these increases in the mRNA levels (Fig. 7D and E).

[6] 100w We next examined the effect of rPGRN on the number of iNOSpositive microglia. This number did not differ between normoxia and the OGD condition after incubation with or without rPGRN (Fig. 8B). Although almost no iNOS-positive microglial cells were detected among microglia under normoxia, the number of iNOS-positive microglia was significantly increased under the OGD condition (Fig. 8A and C). Treatment with rPGRN reduced the increase in the number of those cells under this condition (Fig. 8C). Furthermore, the increased levels of TNFα and IL-1β mRNAs under the OGD condition were attenuated by treatment with rPGRN (Fig. 8D and E).

[1] 93w In this study, based on the results obtained in our previous study, we examined whether the anti-inflammatory and anti-cytotoxic effects of sivelestat were due to its direct action. We demonstrated in the present study that administration of sivelestat reduced the expression of cleaved caspase-3 and Bax and increased the expression of Bcl-2 in the in vivo brain ischemic model used. In addition, the number of cleaved caspase-3 positive cells was also decreased by treatment with sivelestat. These results indicate that administration of sivelestat suppressed caspase-3dependent cytotoxicity at the early stage after cerebral ischemia.

[2] 153w Our results are consistent with the report that administration of sivelestat suppresses cell death by inhibiting caspase 3/7 activity (Matayoshi et al., 2009). In addition, it was also reported that administration of sivelestat suppresses caspase-3 activation in a rabbit spinal cord ischemia model (Yamauchi et al., 2006). These results support our findings of a sivelestat-elicited reduction in the number of TUNELpositive cells after ME (Horinokita et al., 2019). The increase in the number of FJC-positive cells, FJC being a marker of degenerative neurons, was also detected in the present study. FJC is known to stain all degenerative neurons regardless of the cause of degeneration or the mechanism of cell death (Schmued et al., 2005). Therefore, our results suggest that the inhibition of cytotoxicity by sivelestat administration was due to the inhibition of cell death by various mechanisms induced during the acute phase of cerebral ischemia in addition to the inhibition of caspase-3-dependent apoptosis.

[3] 221w Cleaved caspase-3-positive neurons and glial cells are found in the ischemic brain of rat MCAO model rats (Giffard and Swanson, 2005;Guegan and Sola, 2000;Velier et al., 1999). Therefore, the direct effects of sivelestat on cultured neurons and glial cells were next examined. The results revealed no protective effect of sivelestat on OGD-treated cultured neurons, whereas treatment with rPGRN exerted a protective effect against OGD-induced cell injury as well as had reducing effects on the amounts of cleaved caspase-3 and Bax proteins and an increasing effect on the amount of Bcl-2 protein after OGD. In HK-2 cells, the increase in cleaved caspase-3 and Bax protein levels and the decrease in Bcl-2 protein levels after antimycin A/2-deoxyglucose treatment are suppressed by rPGRN treatment (Zhou et al., 2015). Furthermore, PGRN administration in a rat subarachnoid hemorrhage model suppresses apoptosis via the phosphoinositide 3-kinase (PI3-K)/Akt pathway (Li et al., 2015). These results are consistent with the cytoprotective effect of PRGN obtained in the present study, and suggest that the PI3-K/Akt pathway is involved as one of the mechanisms underlying the protective effects of PGRN. Therefore, our findings suggested that the neuroprotective effect of sivelestat obtained in the in vivo cerebral ischemia model was based not on the direct effect of sivelestat on various types of cells but on the inhibition of PGRN degradation by elastase.

[4] 458w Next, sivelestat had no effect on astrocytic GFAP expression. Although it has been reported that the expression of water channel protein (AQP-4) is suppressed by sivelestat administration in a mouse neuromyelitis optica model (Saadoun et al., 2012), the detailed effect on astrocytes is not fully understood at present. In the present study, increased GFAP expression in cultured astrocytes was suppressed by rPGRN treatment, but not by sivelestat treatment. Astrocytes play important roles in maintaining brain homeostasis, such as providing a nutrient supply to nerve cells, ion homeostasis, and neurite outgrowth scaffolding. They also play an auxiliary role for neuronal activity, such as repairing lesions during brain injury. In addition, astrocytes may affect neurotransmission by releasing excitatory amino acids and cytokines/chemokines (Volterra and Meldolesi, 2005). In neurodegenerative diseases such as cerebral infarction and Alzheimer's disease, activation of astrocytes has been suggested to induce an inflammatory response in the brain, which promotes infarct formation and neuronal degeneration (Stoll et al., 1998;Tuppo and Arias, 2005). The expression of GFAP and nuclear factor-kappa B (NF-κB) in aspartate aminotransferasecontaining cells isolated from MCAO/R rats has been reported to be suppressed by PGRN treatment (Shu et al., 2018). Therefore, these results suggest that rPGRN has the ability to suppress NF-κB-dependent inflammation and cell injury induced by activated astrocytes, although further studies will be needed to determine the direct effect of PGRN on activated astrocyte-induced inflammation and cell injury. (merge, d, h, l and p) for Iba-1 (red, b, f, j and n) and inducible nitric oxide synthase (iNOS) (green,c,g,k and o),and with Hoechst 33342 (blue,a,e,i,and m) of cultured microglia under normoxia-vehicle group (a-d), normoxia-sivelestat (100 μM) group (e-h), OGD-vehicle group (i-l) and OGD-sivelestat (100 μM) group (m-p) at 24 h after OGD. The scale bar represents 50 μm. The numbers of Iba-1-(B) and iNOS-positive cells (C) in cultured microglia under normoxia (Nor) or OGD were counted. Results are expressed as the percentage of Iba-1-positive cells among the total number of Hoechst-positive cells (Iba-1/Hoechst) and of iNOS-positive cells among the Iba-1-positive cells (iNOS/Iba-1), and as the means ± SD (n = 6 independent experiments. The total number of Hoechst-positive cells counted was 3128 (vehicle), 3137 (sivelestat) for the normoxia group and 3066 (vehicle), 3113 (sivelestat) for the OGD group). Levels of TNF-α (D) and IL-1β (E) mRNAs of cultured microglia in the vehicle-treated (-) normoxia (white bars) and OGD (black bars) groups and sivelestat (100 μM)treated (+) normoxia (white bars) and OGD (black bars) groups at 24 h after OGD are seen. Results are expressed as the mean ratio of the Normoxia or OGD to the control group ± SD (n = 5 independent experiments). *Significant difference from the vehicle-treated normoxic group (P < 0.05). *Significant difference from the vehicle-treated sham group (P < 0.05).

[1] 430w In the present study, by using cultured neurons, astrocytes, and microglia we examined whether the protective effects of sivelestat against cerebral ischemic injury obtained in our previous study were due to direct effects. OGD-induced neuronal cell death, astrocyte activation, and increased levels of proinflammatory cytokines caused by microglial activation were suppressed by rPGRN treatment. On the other hand, sivelestat treatment had no effect on any of these events in vitro. These results indicate that the suppression of inflammatory response and cell death after in vivo cerebral ischemia was not due to the direct action of sivelestat but rather to the suppression of PGRN cleavage by the inhibition of elastase activity in the brain after cerebral ischemia. Treatment with rPGRN reduced the number of microglial-derived inflammatory cytokines and the number of iNOS-positive cells, an M1 microglial marker. Therefore, the multifaceted protective effect of rPGRN might be attributed to differentiation of M1 microglia, which was increased after OGD, into anti-inflammatory type M2 microglia and to autocrine and paracrine actions of anti-inflammatory cytokines and PGRN, which are secreted from differentiated M2 microglia. The results described in the present study suggest that sivelestat-mediated inhibition of PGRN cleavage, resulting in an increased level of PGRN, could contribute to the establishment of a new therapeutic approach for cerebral ischemia. (merge, d, h, l and p) for b,f,j,and n) and iNOS (green,c,g,k,and o),and with Hoechst 33342 (blue,a,e,i,and m) of cultured microglia under normoxia-vehicle group (a-d), normoxia-rPGRN (100 ng/ml) group (e-h), OGD-vehicle group (i-l) and OGD-rPGRN (100 ng/ml) group (m-p) at 24 h after OGD. The scale bar represents 50 μm. The numbers of Iba-1-(B) and iNOS-positive cells (C) in cultured microglia under normoxia (Nor) or OGD were counted. Results are expressed as the percentage of Iba-1-positive cells among the total number of Hoechst-positive cells (Iba-1/Hoechst) and of iNOS-positive cells among the Iba-1-positive cells (iNOS/Iba-1), and as the means ± SD (n = 6 independent experiments. The total number of Hoechst-positive cells counted was 3404 (vehicle), 3224 (rPGRN) for the normoxia group and 3140 (vehicle), 3096 (rPGRN) for the OGD group). Levels of TNF-α (D) and IL-1β (E) mRNAs of cultured microglia in the vehicle-treated (-) normoxia (white bars) and OGD (black bars) groups and rPGRN (100 ng/ml)-treated (+) normoxia (white bars) and OGD (black bars) groups at 24 h after OGD are given. Results are expressed as the mean ratio of the normoxia or OGD to the control group ± SD (n = 5 independent experiments). *Significant difference from the vehicle-treated normoxic group (P < 0.05). # Significant difference from the vehicle-treated OGD group (P < 0.05).

[1] 20w Elastin-Congo Red (Cat. No. E164) was purchased from Elastin Product (Owensville, MO, USA). Sivelestat was obtained from Nipro (Osaka, Japan).

[2] 70w Sivelestat, which is a selective inhibitor of neutrophil elastase, was dissolved in phosphate-buffered saline (PBS) and administered (50 mg/ kg) intravenously twice just after the surgery of ME and subcutaneously 8 h after ME. The dose of sivelestat and this type of drug administration were based on the reports of Ikegame et al. (2010) and Tonai et al. (2001), and on results of our previous study (Horinokita et al., 2019).

[3] 184w On day 1 after surgery, ME-and sham-operated rats with or without sivelestat treatment were perfused via the heart with 4% paraformaldehyde in 0.1 mol/L phosphate buffer. The brains were quickly removed and immersed in 30% sucrose in 0.1 mol/L phosphate buffer and then cut into 5-mm-thick coronal slabs, which were subsequently embedded in Neg50 (Richard-Allan Scientific, Kalamazoo, MI, USA) and cut into 10-μm sections by using a cryostat. For immunostaining, rabbit anti-cleaved caspase-3 (9661S, Cell Signaling Technology, Danvers, MA, USA) antibody was used with AlexaFluor 594-labeled goat anti-rabbit IgG (Molecular Probes Inc., Eugene, OR, USA) antibody as the secondary antibody. Fluorescence was detected by using an Olympus fluorescence microscope (IX-71; Olympus, Tokyo, Japan). Omission of primary antibodies served as a negative control. No immunostaining was detected in this group. Fluorescent images were loaded into the MetaMorph software program (Molecular Devices, Downingtown, PA, USA). Based on background fluorescence and the size of their nucleus, antibody-labeled cells of the cerebral cortex were observed by use of the MetaMorph software program (5 sections per animal), which areas corresponded to coronal coordinates of 8.06 to 9.70 from interaural.

[4] 62w Fluoro-Jade C (FJC)-positive cells were detected by using an FJC Ready-to-Dilute Staining Kit (TR-100-FJT; Biosensis Pty. Ltd., Thebarton, South Australia). FJC-positive cells in the cerebral cortex were observed with a fluorescence microscope (BZ-X800; KEYENCE, Tokyo, Japan) and counted by use of a BZ-X Analyzer (KEYENCE) in 5 sections per animal, which were corresponded to coronal coordinates of 8.06 to 9.70 from interaural.

[5] 176w On day 1 after surgery, ME-and sham-operated rats were sacrificed by decapitation. The right hemisphere was homogenized in ice-cold buffer containing 10% sucrose, 1 mM ethylenediaminetetraacetic acid (EDTA), and protease inhibitor cocktail (Roche Diagnostics GmbH, Germany) in 20 mM Tris-HCl (pH 7.4). Then the protein concentration was determined. Samples were heated at 95 • C for 5 min in 10% glycerol and 2% sodium dodecyl sulfate (SDS) in 62.5 mM Tris-HCl (pH 6.8). Cultured cells were harvested in sample buffer comprising 62.5 mM Tris-HCl (pH 6.8), 10% glycerol, 2% SDS, and 5% b-mercaptoethanol and heated for 5 min at 95 • C. Western blotting was performed according to standard protocols. The following primary antibodies were used: mouse anti-β-actin (a1978, Sigma-Aldrich Corp., St. Louis, MO, USA), rabbit anti-cleaved caspase-3 (9661S, Cell Signaling Technology), rabbit anti-Bax (#2772, Cell Signaling Technology), rabbit anti-Bcl2 (#2870S, Cell Signaling Technology), and mouse anti-GFAP (556327, BD Pharmingen, Franklin Lakes, NJ, USA) antibodies. Quantification was performed by using computerized densitometry (Luminograph II, ATTO Co., Tokyo, Japan) and an image analyzer (CS Analyzer, ATTO Co.).

[6] 148w Total RNAs were extracted from the right hemisphere of rats and cortical microglia in primary culture by using an RNA extraction kit, Isogen II (Nippon Gene, Tokyo, Japan) and quantified with a BioSpecnano (Shimazu Corp., Kyoto, Japan). cDNAs were synthesized from 500 ng of total RNAs by using ReverTra Ace® qPCR RT Master Mix with gDNA Remover (TOYOBO CO., LTD., Tokyo, Japan). qRT-PCR was performed by using THUNDERBIRD® SYBR qPCR Mix (TOYOBO CO., LTD.) in a CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA). Data were normalized to the 18S rRNA mRNA expression and analyzed by the 2-ΔΔCt method. Primers used in the present study were as follow: 18S rRNA-forward, 5 ′ -CGGACAGGATTGACAGATTG-3 ′ ; reverse, 5 ′ -CAAATCGCTCCACCAAC-TAA-3 ′ . IL-1β-forward, 5 ′ -AGCTGCACTGCAGGCTTCGAGATG-3 ′ ; reverse, 5 ′ -GAACTGTGCAGACTCAAACTCCAC-3 ′ . TNF-α-forward, 5 ′ -ACCACGCTCTTCTGTCTACTG-3 ′ ; reverse, 5 ′ -CTTGGTGGTTTGCTAC-GAC-3 ′ .

[7] 72w The inhibitory effect of sivelestat on serum elastase was determined by using Elastin Congo Red as the substrate. A rat serum sample was incubated with Elastin Congo Red, which was dissolved in an extraction buffer at 37 • C while being rotated for 16 h. The mixed solution was centrifuged to pellet any undigested and insoluble elastin. The absorbance of cleaved Elastin-Congo Red by the elastolytic activity was measured at 490 nm.

[8] 113w Primary cultures of microglia were fixed with 4% paraformaldehyde and blocked with 10% donkey serum, 1% bovine serum albumin in Triton X-100 in PBS. The primary antibodies used were goat polyclonal anti-Iba-1 (ab5076; Abcam, Cambridge, UK) and mouse polyclonal anti-iNOS (ab49999; Abcam) antibodies; and the secondary ones, Alexa Fluor 594-labeled donkey anti-goat IgG (A11058; Invitrogen) and Alexa Fluor 488-labeled goat anti-mouse IgG (A21202; Invitrogen) antibodies, respectively. Fluorescence was detected by using an Olympus fluorescence microscope (IX-71; Olympus). Fluorescent images were loaded into the MetaMorph software program (Molecular Devices). Six images (435 × 330 μm) per experiment were randomly taken from 5 independent experiments and were analyzed with the MetaMorph software program (Molecular Devices).

[9] 62w Statistical analyses were performed with GraphPad Prism (version 8, GraphPad Software, San Diego, CA, USA). Statistical analyses among multiple groups were performed by using factorial analysis of variance (ANOVA), followed by the Tukey test as a post hoc test. The results were expressed as the means ± standard deviation (SD). P values of less than 0.05 were considered to indicate statistical significance.

[1] 88w Male Wistar rats weighing between 220 and 250 g (Charles River Japan Inc., Tsukuba, Japan) were used in the present study. The rats had free access to food and water and were maintained according to the National Institute of Health Guide for the Care and Use of Laboratory Animals and the Guidance for Experimental Animal Care issued by the Prime Minister's Office of Japan. The study was approved by the Committee of Animal Care and Welfare of Tokyo University of Pharmacy and Life Sciences (P18-62, 19 April 2018).

[2] 138w Microsphere-induced cerebral embolism (ME) was performed by the method described previously (Horinokita et al., 2019;Kisoh et al., 2017). Anesthesia was induced with 5% isoflurane and maintained with 2.5% isoflurane. The right external carotid and pterygopalatine arteries were temporarily occluded with strings. Immediately thereafter, a needle connected to a polyethylene catheter (TORAY Feeding Tube, Chiba, Japan) was inserted into the right common carotid artery, and then 700 microspheres (45.0 μm in diameter; Polysciences Inc., Warrington, PA, USA), suspended in 20% dextran solution (150 μl), were injected into the right internal carotid artery through the cannula. After the injection, the needle was removed, and the puncture wound was then repaired with surgical glue. The rats that underwent a sham operation received the same volume of vehicle without microspheres. Non-operated rats were used as naïve control rats in the present study.

[3] 86w On day 1 after the surgery, neurological deficits of the operated rats were scored on the basis of paucity of movement, truncal curvature, and forced circling during locomotion according to the criteria described previously (Furlow Jr. and Bass, 1976;McGraw, 1977;Moriyama et al., 2013). The score of each neurological deficit was rated from 3 to 0 (3, very severe; 2, severe; 1, moderate; 0, little or none). The rats with a total score of 7-9 points on day 1 after cerebral embolism were used in this study.

[4] 60w The rats were maintained according to the National Institute of Health Guide for the Care and Use of Laboratory Animals and the Guideline for Experimental Animal Care issued by the Prime Minister's Office of Japan. All experimental procedures were approved by the Committee of Animal Care and Welfare of Tokyo University of Pharmacy and Life Sciences (P19-22, 21 May 2019).

[5] 132w Primary cultures of cortical neurons were prepared from cerebral cortices of Sprague-Dawley rats (embryonic day 16; SLC, Shizuoka, Japan) according to Yamada et al. (2018). Briefly, cerebral cortices were dissected and digested with 0.25% trypsin (Invitrogen, Carlsbad, CA, USA) in phosphate-buffered saline (PBS) for 25 min at 37 • C. After transfer to neurobasal medium (Invitrogen) containing 10% fetal bovine serum, trituration was performed by use of a fire-polished Pasteur pipet; and then the isolated cortical cells were suspended in neurobasal medium containing 1 mM glutamine, 2% B27 serum-free supplement (Invitrogen), and 1% penicillin-streptomycin (Wako, Osaka, Japan). These cortical cells were plated at a density of 200,000 cells/well in 24well plates (Falcon, Corning, NY, USA) coated with poly-D-lysine (Wako). One half of the medium was replaced with fresh medium every 3 days.

[6] 166w Mixed glial cultures from cerebral cortices of Wistar rats were prepared on postnatal day 3 (SLC, Shizuoka, Japan) according to the method of Giulian and Baker (1986) with minor modifications. The cortices were minced and digested by incubation for 20 min at 37 • C in PBS containing 0.25% trypsin (Invitrogen) and 1 mg/ml DNase I (Worthington Biochemical Co., Lakewood, NJ, USA). After transfer to DMEM/Ham's F12 medium (Thermo Fisher Scientific, Waltham, MA, USA) containing 20% heat-inactivated fetal bovine serum (FBS) and 1% penicillin-streptomycin (Wako), trituration was performed with a firepolished Pasteur pipet. The isolated cells were pelleted by centrifugation at 120 ×g and resuspended in fresh DMEM/F12 medium containing 10% FBS. Next, these cells were plated at a density of 200,000 cells/well in 24-well plates (Falcon), one cortex/75 cm 2 flask (Falcon) coated with poly-D-lysine (Wako). Cultures were maintained at 37 • C and 5% CO 2 in DMEM containing 10% FBS. One half of the medium was replaced with fresh medium every 3 days.

[7] 95w Rat astrocytes were obtained primary cultures of mixed glial cells prepared from neonatal rats. The mixed glia cells were maintained in flasks for 5-7 days, and then incubated at 37 • C with PBS containing 0.125% trypsin for 30 min. The astrocytes, which remained unattached, were aspirated, pelleted by centrifugation at 120 ×g, resuspended in fresh DMEM/F12 medium containing 10% FBS, and placed in 24-well plates. Cultures were maintained at 37 • C and 5% CO 2 in DMEM containing 10% FBS. One half of the medium was replaced with fresh medium every 3 days.

[8] 86w Rat microglial cells were harvested from primary cultures of mixed glial cells prepared from neonatal rats, as previously reported (Saura et al., 2003). Confluent cultures of mixed glia cells were maintained in flasks for 3 weeks, and then incubated at 37 • C with DMEM/F12 containing 0.0625% trypsin for 30 min. The floating astrocytes were removed by aspiration, leaving the microglial cells, which remained attached to the bottom of the well. The specificity and purity of the cultured microglial cells was confirmed by immunostaining for Iba-1.

[9] 165w Cultured neurons were rinsed twice with Neurobasal™-A medium without D-glucose and sodium pyruvate (A24775-01, Thermo Fisher Scientific), after which the medium was replaced with the same medium, containing 2% B27. Neurons were placed in a hypoxic chamber (1% O 2 , 94% N 2 , and 5% CO 2 ) for 30 min, followed by normal culture medium under normoxic condition for 16 h or 24 h. For the normoxic condition, neurons were cultured in 95% atmospheric air and 5% CO 2 for the same times as cells under the hypoxic condition at 37 • C. For primary cultures of astrocytes or microglia, cells were replaced in DMEM without D-glucose and sodium pyruvate (11966025, Thermo Fisher Scientific). Then astrocytes or microglia were placed in a hypoxic chamber (1% O 2 , 94% N 2 , and 5% CO 2 ) for 4 h or in a hypoxic chamber (5% O 2 , 90% N 2 , and 5% CO 2 ) for 24 h, respectively.

[10] 74w To determine the effect of sivelestat on the activity of serum elastase, we incubated serum with various concentrations of sivelestat (1-100 μM). The activity of serum elastase, which was examined by the method using Elastin-Congo red, was decreased significantly and dosedependently by such treatment (Fig. 4) when tested at the concentrations of 10, 30, and 100 μM sivelestat (Fig. 4). The dose used in the following in vitro experiments was based on these concentrations.

[11] 69w The effects of sivelestat and rPGRN on the level of GFAP were examined in cultured astrocytes under the OGD/R condition. This level was significantly increased compared with that under normoxia, and treatment with sivelestat had no effects on the increase in the level of GFAP (Fig. 6A). In contrast, treatment with rPGRN inhibited the increase in the level of GFAP in cultured astrocytes under the OGD/R condition (Fig. 6B).

[12] 168w Furthermore, when the effect of sivelestat on cultured microglia was examined, the increase in the number of iNOS-positive cells, a marker of M1 microglia, was not changed; whereas rPGRN treatment inhibited the OGD-induced increase in the number of iNOS-positive cells. In addition, although sivelestat treatment had no effect on increased inflammatory cytokine (TNF-α and IL-1β) mRNA levels, rPGRN treatment suppressed them in the present study. Interestingly, traumatic brain injury has been shown to increase the number of CD68-positive active microglia at the site of injury in PGRN-deficient mice compared to wild-type mice (Tanaka et al., 2013). Also, in microglia in primary cultures derived from PGRN KO mice, the mRNA level of the anti-inflammatory cytokine IL-10 is decreased under the OGD condition as compared with the wildtype (Kanazawa et al., 2015), indicating that PGRN has an antiinflammatory effect mediated by IL-10 in microglia. Therefore, sivelestat has the ability to exert an indirect protective effect against microglial injuries in vivo, which can be associated with the inhibition of PGRN degradation.

[13] 251w Thus, our results suggest that the inhibitory effects of sivelestat on inflammatory responses and cell death after cerebral infarction could be caused by the indirect effects of PGRN, whose cleavage was suppressed by inhibition of neutrophil elastase. We previously demonstrated that the number of PGRN-positive microglia was increased after ME and OGD treatment (Horinokita et al., 2019), suggesting that PGRN, which was secreted from activated microglia and having escaped from cleavage, may act in an autocrine or paracrine manner to reduce cell damage after cerebral ischemia. Furthermore, it is possible that PGRN may induce the differentiation of microglia into the M2 phenotype (Ma et al., 2017;Pickford et al., 2011), which cells produce and releases antiinflammatory cytokines such as TGF-β and IL-10 and antiinflammatory and neuroprotective factors such as brain-derived neurotrophic factor (Parkhurst et al., 2013;Ponomarev et al., 2007;Quirie et al., 2013). In recent studies, the administration of M2 microglia preconditioned by optimal OGD after the onset of ischemic stroke has the ability to promote angiogenesis and axonal outgrowth (Kanazawa et al., 2017). In this sense, we demonstrated that although treatment with rPGRN did not change the number of Iba-1-positive microglia, that of iNOS-positive cells, a marker of M1 microglia, was decreased by rPGRN treatment, suggesting that rPGRN may have induced the differentiation of microglia into a lamified or M2 microglia and played a part in the cytoprotective effect. Therefore, we suggest that the induction of M2 microglia may have contributed to the in vivo effects of sivelestat obtained in this study.