Mesenchymal stromal cells attenuate post-stroke infection by preventing caspase-1-dependent splenic marginal zone B cell death

Dear Editor, Spontaneous infection is one of the most common complications of acute ischemic stroke (AIS) and leads to increased morbidity and mortality in stroke patients. However, current interventions fail to improve clinical outcomes in these patients. Since stroke-induced immunodeficiency is thought to contribute to the risk of infection, we hypothesized that mesenchymal stromal cell (MSC) therapy, which harnesses potent immunomodulatory capacities, could be a potential solution. Here, we show that intravenously administered MSCs preferentially migrate to the marginal zone (MZ) of the spleen, preserve the injured MZ B cells and ameliorate post-stroke infection and mortality in the mouse middle cerebral artery occlusion (MCAO) model of AIS. To address the therapeutic effects of MSCs in post-stroke infections, male C57BL/6 mice were intravenously treated with MSCs at 6 h after reperfusion. To rule out the general benefits of cell therapy, we applied the treatment of human dermal fibroblasts (hDF) as control. Impressively, MSC-treated mice exhibited higher survival rate and smaller infarct volume compared to PBS-treated group (Fig. 1a and Supplementary Fig. 1). Furthermore, MSC therapy significantly reduced bacterial loads and rescued spleen shrinkage at 5 days after MCAO (Fig. 1b–d). Based on the protective effects of MSCs in the spleen, we examined the bio-distribution of MSCs after transplantation through transducing the cells with lentiviral vector encoding tdTomato. The infused MSCs were transiently observed in the lung but rapidly disappeared from this region after 1 day, while numerous alive MSCs were detected in the spleen during the first 5 days after transplantation (Supplementary Fig. 2a–c). More specifically, the transplanted MSCs mainly distributed in the MZ of the white pulp (Fig. 1e). The MZ is a highly transited area that facilitates filtration of blood-borne pathogens and initiates immune responses. The lymphocytes found in this area are principally MZ B cells, serving as the gatekeeper against infections. As expected, the infused MSCs were found near MZ B cells (Fig. 1f), indicating the possible interaction between MSCs and MZ B cells. We further evaluated the immunomodulatory impacts of MSCs on splenocytes and observed striking reductions in T and B cells at 5 days post-MCAO, but MSC administration rescued this B cells loss (Supplementary Fig. 3). FACS analysis showed the number of B220CD21CD23 MZ B cells and B220CD21CD23 follicular B cells were increased in the MSC group compared with the PBS group, with a greater difference seen among MZ B cells (Fig. 1g, h). Moreover, we demonstrated that MSC treatment effectively rescued the loss of immunoreactivity for these MZ B cell markers and increased circulating IgM levels (Fig. 1i, j). Given that MSCs in the MZ mainly distributed near MZ B cells and innate-like MZ B cells play a critical role in rapid anti-bacterial defense, we sought to determine the potential mechanisms of MZ B cells loss and how MSC administration preserved this subset of B cells. Lymphocyte apoptosis is one of the main factors contributing to the post-stroke splenic atrophy. Here, a large percentage of CD1dMZ B cells co-localized with TUNEL labeling in splenic sections from PBS-treated mice at 5 days post-MCAO, and MSC treatment decreased the population of CD1dTUNELMZ B cells (Supplementary Fig. 4a, b). However, TUNEL labeling of fragmented DNA has been observed not only in classical apoptotic cells, but also in caspase-1dependent pyroptosis. Therefore, we applied fluorescently labeled inhibitor of caspases (FLICA) probes to determine the distribution of active caspase-3 and caspase-1. Surprisingly, CD1dMZ B cells were found to be co-localized with active caspase-1, but not active caspase3. Meanwhile, a large fraction of the MZ B cells exhibited caspase-1 activation at 5 days post-MCAO, which was reduced by MSC treatment (Supplementary Fig. 4c–e). Furthermore, we revealed that experimental stroke-induced caspase-1 activation and GSDMDmediated cell lysis in MZ B cells, and that MSC administration decreased this inflammatory form of programmed cell death (Supplementary Fig. 4f). A series of cytosolic receptors known as the NOD-like receptors (NLRs) have been found involved in inflammasome formation and mediate caspase-1 activation. So, we measured the mRNA levels of various NLRs in MZ B cells. In comparison to NLRP3, the expression levels of NLRP1 and NLRC4 were markedly lower. Additionally, MCC950 (an inhibitor of NLRP3) treatment reduced caspase-1 activation and subsequent GSDMD-mediated MZ B cells lysis (Supplementary Fig. 4g, h). These results indicate that the NLRP3caspase-1 pathway plays an essential role in stroke-induced MZ B cell death and MSC infusion reduces caspase-1 activity in these MZ B cells, potentially through the inhibition of NLRP3 activation. Increasing evidence suggests that damaged/dysfunctional mitochondria may release mitochondrial reactive oxidative species (mtROS), which could bind to and activate the NLRP3 inflammasome. Here, we exposed experimental animals to a mitochondriatargeted antioxidant (mito-TEMPLE) for 5 days after MCAO. The protein levels of NLRP3, cleaved caspase-1, and cleaved GSDMD were decreased in mice treated with mito-TEMPLE. As expected, the number of MZ B cells was also rescued by administration of mitoTEMPLE (Supplementary Fig. 5a, b). Besides, the excessive mtROS generation and disrupted mitochondrial membrane potential (MMP) in the MZ B cells after MCAO were alleviated by MSC treatment (Supplementary Fig. 5c–f). Considering the injected MSCs were distributed near MZ B cells, this might reflect the occurrence of direct mitochondrial transfer. To test this hypothesis, we labeled MSCs with mito-tracker prior to cell transplantation. Mitochondria from MSCs (mito-tracker) were detected in CD1d MZ B cells at 5 days after MCAO, indicating that mitochondria were transferred from transplanted MSCs to injured MZ B cells (Supplementary Fig. 5g). Several previous studies have described the formation of microtubule for mitochondria transfer between MSCs and injured cells. Accordingly, we detected that mitochondria from MSCs were transferred via


Animals and experimental stroke model
Male C57BL/6 mice aged 8 to 10 weeks were purchased from Guangdong Medical Laboratory Animal Center and housed in individually ventilated cages under specific-pathogen-free (SPF) conditions. A standard 12 h light/dark cycle was set, and the animals were given free access to food and water. All animal experimental procedures were approved by the Ethical Committee of Sun Yat-Sen University.
To generate the MCAO model, mice were anesthetized with 1.5% isoflurane in a 30% O 2 /69% N 2 O mixture and randomly assigned to the sham group or MCAO group. All surgical instruments were sterilized before operation. As previously described, 1 a 10-mm incision was made on the right side of neck and the carotid artery, external carotid artery (ECA), and internal carotid artery (ICA) were exposed.
A silicone-coated suture was inserted into the ECA and then through the ICA to block the MCA. After 40 min, the suture was removed for the purpose of reperfusion, and the incision was carefully sutured. Mice of the sham group underwent the same isoflurane anesthesia and surgical procedures as the MCAO group, except there was no insertion of an intraluminal filament. The body temperature of all experimental mice was maintained at 37°C throughout the surgery and recovery using a heating pad. The inclusion criteria for tMCAO were: 1) regional cerebral blood flow decreases ＞70% during occlusion as detected by laser Doppler flowmetry; 2) neurological deficits 3 hour after MCAO with neurological score between 2~3 points (determined as described). 2 6 hours later, the experimental mice were randomly assigned to PBS and MSC-treatment groups for further investigations.

Isolation, characterization, and transplantation of MSC/hDF
Heparin-treated bone marrow was obtained from healthy donors along with their informed consent. Each bone marrow aspirate was diluted 1:1 with human MSC culture medium (low-glucose DMEM with 10% fetal bovine serum, HyClone). MSCs were separated using density gradient centrifugation in Ficoll-Paque (1.077 g/mL, Amersham Biosciences) and seeded at 1x10 5 cells/cm 2 in the MSC culture medium as passage 0. Nonadherent cells were removed by a complete change of the culture medium, and when the remaining cells reached 80%-90% confluence, they were continuously passaged using 0.25% trypsin-EDTA. For MSC administration, 1x10 6 cells (passages 4-8) were suspended in 0.1 ml PBS and transplanted via the caudal vein at 6 h after MCAO. Besides, hDFs were isolated from the foreskin of adult healthy donor with informed consent and cultured as previously described. 3 These fibroblasts were adopted as the cell control in this study.

Bacteriological analysis
Mice were euthanized at the indicated time points after MCAO and washed with 75% ethanol. Blood samples were obtained via cardiac puncture, and lung tissues were mechanically separated, minced, and homogenized under sterile conditions. To measure bacteria in terms of colony-forming units (C.F.U.), 10 μl blood or lung homogenate was serially diluted with PBS, plated on blood agar plates, and incubated at 37°C. After 24 h, the bacterial colonies were counted by researchers blinded to the treatment assignment.

Flow cytometry
Experimental mice were euthanized at 1, 3, and 5 days after MCAO and perfused with ice-cold saline, and the spleen, lung, and ipsilateral hemisphere tissues were isolated. Single cells from these tissues were prepared for immunostaining and FACS analysis. Spleen samples were mechanically homogenized and filtered through 40-μm cell strainers. 1X Red blood cell (RBC) lysis buffer (BioGems) was used to remove RBCs from the splenic cell suspensions. Lung tissues were collected, dissected, homogenized, digested in 0.25% trypsin at 37°C for 30 min, and passed through 40-μm filters. For brains, the ipsilateral hemispheres were digested with 0.25% trypsin at 37°C for 30 min and pressed through 40-μm filters, and then the myelin debris was removed from brain cells using gradient centrifugation in 30%/70% Percoll solution (Sigma-Aldrich). ImmunoChemistry Technologies), anti-human COX4 antibody (1:100, Santa Cruz). Nuclei were visualized by DAPI (Fluka) staining for 10 min. Immunofluorescent images were acquired using an LSM800 confocal microscope (Zeiss), an LSM880 confocal microscope (Zeiss), and Dragonfly high-speed confocal microscopy (ANDOR, Oxford Instruments). The supplementary movie was edited using the Imaris microscopy image analysis software (Oxford Instruments).

Enzyme-linked immunosorbent assay (ELISA)
A commercially available ELISA kit for anti-mouse IgM (RayBiotech) was used as recommended by the manufacturer to determine the IgM levels in the sera of experimental mice.

Mitochondrial membrane potential and ROS assessment
Mitochondrial membrane potential was detected using the cell-permeant dye, tetramethylrhodamine (TMRM, Thermo Fisher Scientific), which accumulates in mitochondria with intact membranes. To assess the levels of mitochondrial ROS, MZ B cells were labeled with fluorescent probe, MitoSOX Red (Thermo Fisher Scientific). Cells were incubated with the indicated dyes at 37°C for 30 min and analyzed by flow cytometry.

PCR)
Total RNA was extracted from isolated MZ B cells using the RNeasy mini kit (QIAGEN), according to the manufacturer's instructions. RNA (1 μg per sample) was subjected to reverse transcription using a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). qRT-PCR was performed in duplicate using the SYBR Green qRT-PCR Super Mix (Roche) and detected by a Light Cycler 480 Detection System (Roche) as previously described. 4 The primers used for qRT-PCR analysis were:

Drug administration
MCC950 (Millipore) was dissolved in sterile PBS and intraperitoneally (i.p.) administered (10 mg/kg/d) to mice for 5 consecutive days starting at 6 h post-reperfusion. Mito-TEMPO (Sigma-Aldrich) was given i.p. at the same time points as MCC950, at 1 mg/kg/d. Mice in the sham and vehicle groups received equal volumes of saline (i.p.) and served as control groups. All animals were sacrificed at 5 days after MCAO for further analysis.

Statistical analysis
All results are presented as the mean ± S.E.M. of at least three independent experiments. The sample size of each experiment is indicated in the corresponding figure legend. One-way analysis of variance (ANOVA) was used to compare means among multiple groups. Survival curves were compared using the log-rank test. A P-value less than 0.05 was considered statistically significant. 5