High Salt Diet Down Regulates TREM2 Expression and Blunts Efferocytosis of Macrophage After Acute Ischemic Stroke

Mengyan Hu Third A liated Hospital of Sun Yat-Sen University Yinyao Lin Third A liated Hospital of Sun Yat-Sen University Xuejiao Men Third A liated Hospital of Sun Yat-Sen University Qiang Zhu Third A liated Hospital of Sun Yat-Sen University Danli Lu Third A liated Hospital of Sun Yat-Sen University Sanxin Liu Third A liated Hospital of Sun Yat-Sen University Bingjun Zhang Third A liated Hospital of Sun Yat-Sen University Wei Cai Third A liated Hospital of Sun Yat-Sen University Zhengqi Lu (  luzhq@mail.sysu.edu.cn ) Third A liated Hospital of Sun Yat-Sen University https://orcid.org/0000-0002-2118-0368

fed with HSD suffered exacerbated neural in ammation as higher level of in ammatory mediators and immune cells in ltration were documented. Polarization shift towards pro-in ammatory phenotype impaired efferocytosis of in ltrated macrophages within stroke lesion in HSD-fed-mice were detected. As was uncovered by PCR array, macrophage expression of triggering receptor expressed on myeloid cells 2 (TREM2), a receptor relevant with phagocytosis, was down regulated in high salt environment. Enhancing TREM2 signaling restored the efferocytosis capacity and cellular in ammatory resolution of macrophages in high salinity environment. In AIS patients, high concentration of urine sodium was correlated with lower expression of TREM2 and detrimental stroke outcomes.
Conclusions: HSD blunted efferocytic capacity of macrophages through down regulating the expression of TREM2, thus impeded in ammatory resolution after ischemic stroke. Enhancing TREM2 signaling in monocyte/macrophage could be a promising therapeutic strategy to enhance efferocytosis and promote post-stroke in ammatory resolution.

Background
High salt intake is highly associated with blood pressure, blood lipid concentration, the level of circulating alarmins and other factors affecting stroke prognosis, and considered as a risk factor of acute ischemic stroke (AIS) (1)(2)(3). Therefore, salt restriction is widely accepted as a vital step in e cient lifestyle intervention to prevent new vascular event, especially in AIS. Nevertheless, given that a large number of patients fail to convert diet habit before the arrival of acute vascular event, there is an unmet need to develop therapeutic strategy to tackle the already-exist high salinity and the associated pathophysiology.
As a pivotal part of innate immunity, macrophages characterize multiple roles in stroke lesion (4,5).
Whether they are involved in in ammatory resolution or serve as an in ammation ampli er depend on their speci c phenotype and microenvironment (6,7). Recent research has elucidated that surplus dietary salt directed macrophages/microglia towards the classical activated pro-in ammatory phenotype, which is often referred to M1 subtype (8), indicating that excessive salt intake breaks the balance of macrophages and further aggravates the in ammatory response. In vivo, the pro-in ammatory property of macrophages in high salt diet (HSD) fed mice contributed to blood-brain barrier (BBB) disruption after stroke, and thereafter exacerbating stroke outcomes (9).
Orchestrating macrophage activities and enhancing the in ammatory resolution make it possible to develop a promising therapeutic strategy in AIS. The post stroke in ammatory resolution heavily depends on the efferocytosis function of macrophage, which eliminates dead cells or debris that enhance sterile immune reactions in stroke lesion (6). Nevertheless, the impact of high salt diet on the phagocytic activity and the subsequent anti-in ammatory functions of macrophages remains elusive.
The current study investigated the impact of excessive salt intake on the in ammatory resolution of macrophages in post stroke neural in ammation. Of particular interest, efferocytic capacity of macrophages was evaluated under this harmful condition. Meanwhile, corresponding countermeasure to ne-tune macrophage activities in high salinity environment was purposed.

Ethical statement
The clinical and the animal experimental studies were approved by the Medical Ethics Committee of the  (20)(21)(22). MRI scans of patients were assessed by experienced neurologist Zhengqi Lu, who was blinded to the patients' clinical features.
All images were interpreted with the same window settings, same types of monitors and lighting conditions.

PBMC isolation
Anti-coagulated blood (3mL) was collected, and then diluted 2-fold with PBS, pipetted into centrifuge tube pre lled with Ficoll lymphocyte separation solution (TBDscience), followed by centrifuged at 2000rpm for 25 minutes at room temperature without deceleration. PBMCs from the buffy coat were washed twice with PBS, then stored at -80°C until further analysis.
Animals C57/Bl6 wild-type mice (8 weeks old, weight 18-25 g) were purchased from Guangdong Medical Laboratory Animal Center (Guangzhou, China) and housed in a humidity-and temperature-controlled animal facility in Sun Yat-sen University with a 12-h light-dark cycle. Mice received normal chow (0.5% NaCl) and tap water ad libitum (normal diet) or sodium-rich chow (8% NaCl) and tap water containing 1% NaCl ad libitum (HSD) for 4 weeks according to the experiment.

Model of acute ischemic stroke
Mice were subjected to focal acute ischemic stroke induced with transient middle cerebral artery occlusion (tMCAO). Procedures of tMCAO were described previously (19). Brie y, mice were anesthetized with 1.5-2.0% iso urane under conditions of spontaneous breathing. A lament was inserted into the external carotid artery (ECA) and was directed to the middle cerebral artery (MCA) through the internal carotid artery (ICA). Filament insertion into the ICA was maintained for 60min followed by reperfusion with maintenance of core body temperatures. Cerebral blood ow (CBF) during surgery was measured by laser Doppler ow cytometry. Mice with more than 70% reduction of blood ow in the ischemic core were included in the study and mice that died during surgery were excluded. Survival of mice were recorded.

Infarct volume analysis
For immunologic staining of NeuN, six equally spaced coronal brain sections encompassing the MCA territory were stained with NeuN antibodies. Infarct volume was analyzed with NIH Image J software on NeuN-stained sections. The infarct area was determined as the difference between the NeuN positive area of contralateral hemisphere and ipsilateral hemisphere. Brain infarct was determined by multiplying the mean area of tissue loss by the distances between the two adjacent stained brain slices.
Primary microglia culture Primary mouse microglia were obtained from BLUEFBIO company, and was cultured in culture medium (DMEM-HG + 10%FBS) until treatment.

Primary cortical neurons culture and OGD
Primary cortical neuronal cultures were prepared from E16-18 embryos of C57/Bl6 mice as previously described (23).

Phagocytosis assay
For evaluation of efferocytic capacity, apoptotic neurons were labelled with the dead cell marker Propidium iodide (PI) in PBS (1ug/ml, 37℃, 15min) and treated to macrophages, with a ratio of dead neurons : macrophages = 5:1, for indicated time periods. For in vitro immunol staining experiments, macrophages were pre-grown on poly-l-lysine coated cover slips. The cover slips of macrophage were washed for 2 times to remove unengulfed neurons and xed with 4% paraformaldehyde. The cover slips were then subjected to immunol staining and removed from wells using tweezers and mounted to the slides. F-actin of macrophage was then stained with Alexa Fluor488 phalloidin (A12379, 1:500 in PBS; Invitrogen) at room temperature in the dark for 30min. For ow cytometry experiment, macrophages were pre-seeded on 24-well plates and treated with the same ratio of dead neurons for indicated time periods. Macrophages were washed with PBS and detached from wells with trypsin and were subjected to ow cytometric analysis. Percentage of efferocytic macrophages (PI + ) was calculated with ow cytometric analysis.

Lentiviral infection of macrophage
Lenti virus was constructed and packaged by FenghBio (Changsha, China). The macrophage culture was infected for 3d with Lenti-TREM2 or the control vectors. The overexpression of TREM2 was con rmed by western blot and ow cytometry.

Flow cytometric analysis
Brain tissue was homogenized and prepared as single-cell suspensions for ow cytometric analysis (FACS). Brie y, brains were dissected, and ipsilateral hemispheres were collected. Each hemisphere was subjected to digestion with 0.25% trypsin-EDTA (Thermo Fisher, Carlsbad, CA, USA) at 37 °C for 25 min.
Brain tissue was then pressed through a cell strainer (70 μm). Brain cells were separated from myelin debris by centrifugation in 30%/70% Percoll solution (GE Healthcare Biosciences AB, Uppsala, Sweden).
Brain cells at the interface were collected, washed with HBSS, and subjected to further staining. The following antibodies were used: CD45-PE-Texas Red ( Immuno uorescence staining and cell quanti cation Animals were euthanized and perfused with PBS followed by 4% paraformaldehyde. After su cient perfusion, brains were removed and then cut into 25μm frozen cryo-sections using a microtome. Brain Confocal microscopy images were acquired using a Leica SP confocal microscope and Leica confocal software. Immunopositive cell quanti cation and area analysis were performed with the software of ImageJ (National Institutes of Health) by an investigator who was blinded to the experimental design. In quanti cation of cell in stroke penumbra, the stroke core was identi ed as the region in which the majority of DAPI-stained nuclei were shrunken, and the stroke penumbra was de ned as the region of generally morphologically normal cells, approximately 450-500μm wide, surrounding the stroke core.  Table 2. Double delta CT were calculated, and the data presented as fold change normalized to PBS-treated contralateral brain, PBS-treated macrophage, or negative control lentivirustreated macrophage. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a normalizer housekeeping gene. In data analysis in Figure 2, Figure 5, and Supplementary Figure 1B, the mRNA expression level was visualized with heat map and clustered with the software of R using the "pheatmap" package.

Western blot
Protein isolation was performed as previously described (24). Western blots were performed using the standard SDS-polyacrylamide gel electrophoresis method and enhanced chemiluminescence detection reagents ( Immunoreactivity was semi-quantitatively measured by gel densitometric scanning and analyzed by the MCID image analysis system (Imaging Research, Inc.).

Statistical analysis
All results were presented as mean ± standard error of the mean (SEM). The differences in the means among multiple groups were analyzed using one-or two-way analysis of variance (ANOVA). When ANOVA showed signi cant differences, pair-wise comparisons between means were tested by Dunnett's test. The Student's t test was used for two-group comparisons. The software used for statistical analysis was R v3.6.3. In all analysis, P < 0.05 was considered statistically signi cant.

Results
Excess salt intake exacerbates disease outcomes of ischemic stroke Healthy wild type (WT) C57/Bl6 male mice were fed with high salt diet (HSD) or normal diet (ND) for 28d.
Mice were then subjected to 60min of transient middle cerebral artery occlusion (tMCAO) and sacri ced at 3d or 7d after cerebral ischemia (Fig. 1A). In consistent with previous study (9,10), mice fed with high salt diet displayed increased lesion volume (Fig. 1B-C), detrimental neurological de cit (Fig. 1G) and poor survival rate (Fig. 1H). As assessed with immunol staining, we recorded accumulated dead neurons (NeuN + TUNEL + ) in stroke penumbra (Fig. 1D-E). Strikingly, at 7d after tMCAO, the number of dead neurons had a 67% reduction (vs. 3d) in mice fed with high salt diet, which was less than that of mice fed with normal diet (80%) (Fig. 1F). The results indicated that neurons in high salinity environment suffered processive injury and/or the injured neurons in mice fed with high salt diet were not eliminated in time after stroke.
Surplus salt intake ampli es post stroke neural in ammation To examine the neural in ammatory status in mice fed with high salt diet after tMCAO, in ltration of immune cells in stroke lesion was analyzed with ow cytometry (Fig. 2A). We found that the percentages of T cells (CD45 + CD3 + ), B cells (CD45 + CD19 + ), neutrophils (CD45 hi CD11b + Ly6G + ) and macrophages (CD45 hi CD11b + Ly6G -) among singlets increased in the ipsilateral hemisphere with ischemic stroke of HSD mice ( Fig. 2A), while composition of neutrophils (CD45 hi CD11b + Ly6G + ) and monocytes/macrophages (CD45 hi CD11b + Ly6G -) in peripheral blood and spleen remained to be comparable (Supplementary Figure 1A). With RT-PCR, we recorded that the level of multiple proin ammatory cytokines and chemokines elevated in the ipsilateral brain of HSD mice (Ccl1, Cxcl1, Cxcl2, Cxcl9, Il1a and Il6), while the anti-in ammatory markers, including Il4 and Arg1, decreased at the meantime ( Fig. 2B and Supplementary Figure 1B). The results illustrated that post stroke neural in ammation was ampli ed in mice fed with high salt diet.
Expression of in ammatory resolution associated molecules are down regulated in high salinity environment To testify the macrophages' role in the drastic neural in ammation of HSD mice, we evaluated the in ammatory resolution of these cells. Immunol staining revealed that the in ammatory resolution associated marker CD206 was down regulated in Iba1 + microglia/macrophages in the lesion of HSD mice at 3d after tMCAO (Fig. 3A). In contrast, the number of CD16 expressing Iba1 + microglia/macrophages were up regulated (Fig. 3A). To explore the impact of high salinity environment on macrophage, bone marrow derived primary cultured macrophages were treated with 40mM of NaCl overnight in the presence of LPS (100ng/ml) or IL-4 (20ng/ml). We recorded that high concentration of NaCl alone reduced the expression of in ammatory resolution marker of CD206 as assessed with RT-PCR (Fig. 3B) and ow cytometry (Fig. 3C), and addition of LPS in the culture system further down regulated the expression (Fig. 3B-C). As was reported, IL-4 increased the expression of Arg1 and CD206 in primary cultured macrophages. Nevertheless, macrophages failed to response to the IL-4 signaling in high salinity environment ( Fig. 3B-C). Macrophages pre-treated with NaCl, with or without the presence of IL-4, displayed high expression of TNFα ( Fig. 3B-C). Our data indicated that high salinity environment undermined the anti-in ammatory or in ammatory resolution property of macrophages.

Efferocytosis of macrophages is impaired in high salinity environment
Efferocytosis represents an important biological process for in ammatory resolution mediated by macrophages. Therefore, we evaluated the impact of high salt environment on the phagocytic activities of macrophages. Clearance of dead/dying neurons was determined by detecting the neuronal marker NeuN within Iba1 + microglia/macrophages in stroke penumbra with confocal microscopy (Fig. 4A-B).
Under the premise of similar amount of Iba1 + cells (Fig. 4B), the number of Iba1 + NeuN + cells, which indicated the microglia/macrophages that had engulfed neurons, was reduced in HSD mice at 3d after tMCAO compared to ND mice. Triple staining of Iba1/TUNEL/NeuN further revealed dampened phagocytosis of dead/dying neurons by microglia/macrophages as the engulfed dead neurons (Iba1 + NeuN + TUNEL + ) decreased while the un-engulfed dead neurons (Iba1 -NeuN + TUNEL + ) increased in HSD mice. Very few Iba1 + NeuN + TUNELcells were observed in stroke penumbra in both HSD and ND mice (Fig. 4A). Consistently, the phagocytic index, which was calculated as the proportion of dead/dying neurons engulfed by microglia/macrophages, was lower in HSD mice (Fig. 4B). We further evaluated the impact of high NaCl concentration on the efferocytic activity of macrophages upon encountering dead/dying neurons in vitro. Primary cortical neurons were exposed to 90 min of oxygen-glucose deprivation (OGD), an in vitro model simulating ischemic injury. Propidium iodide (PI) was added into neurons 24-hour after OGD (before cell xation) to label dead/dying cells. Macrophages pre-treated with 40mM of NaCl or equal volume of PBS were exposed to PI-labeled neurons at a ratio of 1:5. Efferocytic capacity of macrophages were evaluated over time with immunol staining and ow cytometry.
Macrophages that pre-exposed to high salinity environment displayed reduced efferocytic capacity, as the engulfed dead/dying neurons per macrophage (Fig. 4D) or the proportion of phagocytic macrophages (PI + F4/80 + ) in high salinity environment (Fig. 4E) were lower than those in control group from 0.5-4h after the onset of co-cultured though no difference of cell viability between the two groups was observed ( Fig   4C). To estimate the capacity of cellular in ammatory resolution, mRNA level of the pro-in ammatory cytokine Tnfα and in ammatory resolving molecule Arg1 was assessed at 6h after the onset of efferocytosis. Macrophages pre-treated with high NaCl displayed increased expression of Tnfα and reduced expression of Arg1 compared with those treated with PBS ( Fig 4F). Our results revealed that efferocytosis and the subsequent cellular in ammatory resolution of macrophages were impaired in high salinity environment.

Excess salt down regulates TREM2 expression in macrophages and impairs in ammatory resolution
We went on to look into the mechanism of how excess salt suppressed efferocytosis of macrophages. Expression of phagocytosis-related receptors (PRRs) in macrophages treated with 40 mM of NaCl or equal volume of PBS was assessed with PCR array. We recorded that high salt environment down regulated the mRNA level Trem2, while expression of other PRRs, including Trem1 and Tim4, remained stable (Fig. 5A). Moreover, molecules of downstream signaling of TREM2 were down regulated in high salt concentration including Arp2, Vav3, and Rac (11,12) (Supplementary Figure 2). We con rmed the down regulation of TREM2 in macrophages exposed to excess salt in vitro on basis of western blot (Fig.  5B) and ow cytometric analysis (FACS) (Fig. 5C). We then examined TREM2 expression in vivo and recorded that the mRNA (Fig. 5D) and protein level (Fig. 5E-F) of TREM2 in the ipsilateral hemisphere of HSD mice was lower than that in ND mice at 3d after tMCAO. Nevertheless, the level of TREM1 did not showed signi cant alteration in high salinity environment (Supplementary Figure 3A-B). When discovering the relationship of TREM2 expression and in ammatory phenotype of macrophages with FACS, we found that macrophages with high TREM2 expression (CD45 + F4/80 + TREM2 hi ) displayed antiin ammatory phenotype with higher CD206-MFI than those with low TREM2 expression (CD45 + F4/80 + TREM lo ), while CD16-MFI showed no difference between macrophages with high and low TREM2 expression in both HSD and ND mice (Fig. 5G).
Decreased TREM2 expression is correlated with pro-in ammatory property of circulating monocytes and detrimental stroke outcomes in AIS patients We then tested the TREM2 level in monocytes of AIS patients and evaluated the relationship between TREM2 expression and stroke outcomes. Dietary salt intake of AIS patients was measured with 24-hour urine sodium with a normal limit of 170mmol (13). Thereafter, we found that patients with high urine sodium concentration had larger infarct scale (Fig. 6A) and higher NIHSS scores (Fig. 6B) than those with normal urine sodium concentration. To assess the impact of excessive salt on phenotypic shift of circulating monocyte in AIS patient during acute phase (0-3d after disease onset), expression of the proin ammatory marker CD80 and the anti-in ammatory marker CD206 (14)(15)(16) in monocyte (CD11b + CD14 + ) of their peripheral blood was analyzed with FACS. Detailed gating strategy is displayed in Supplementary Figure 4. As expected, monocyte from stroke patients with high urine sodium concentration expressed less CD206 compared with normal diet stroke patients, while no different expression of CD80 was recorded (Fig. 6C, D and Supplementary Figure 5). We documented that TREM2 expression in monocytes was down regulated in stroke patients with high urine sodium concentration compared with those with normal diet using FACS (Fig. 6C, E and Supplementary Figure 5). Moreover, we found that the TREM2 mRNA level decreased in the peripheral blood mononuclear cells (PBMC) of patients with high urine sodium concentration (Fig. 6F), while expression of other PRRs remained to be stable (Supplementary Figure 6). Since PRRs are mainly expressed in monocytes in PBMC (17), our data indicated that high salinity environment speci cally down regulated TREM2 expression in monocytes of AIS patients. Through spearman correlation analysis, we recorded that CD206 MFI of peripheral blood monocyte showed signi cant positive correlation with TREM2 MFI in stroke patients, while CD80 MFI showed negative correlation with TREM2 MFI (Supplementary Figure 7), which was in consistent with our data in animal models. Interestingly, we found that TREM2 expression in the circulating monocyte of AIS patients was negatively correlated with the 24-hour urine excretion ( Fig. 6G and Supplementary Figure 7), while decreased TREM2 level of macrophages was associated with increased NIHSS scores (Fig 6G and  Supplementary Figure 7). The results indicated that TREM2 expression in monocytes/macrophages favored efferocytosis and the subsequent in ammatory resolution after ischemic stroke.
Enhancing TREM2 signaling restores the efferocytic capacity and cellular in ammatory resolution of macrophages in high salinity environment TREM2 is a vital functional molecule implicated in the phagocytosis activity of macrophages. The efferocytosis capacity of macrophages plays a decisive role in in ammatory resolution after stroke and affects the disease outcomes. Therefore, we hypothesized that enhancing TREM2 signaling in macrophages could restore their efferocytic capacity and promote in ammatory resolution. Macrophages were infected with lent viral vectors carrying TREM2 cDNA or empty vector for 2d with or without addition of NaCl (40mM) and incubated with PI-labeled post-OGD neurons. E cacy of transfection was con rmed with ow cytometry (Fig. 7A) and western blot (Fig. 7B). Gratifyingly, overexpressing TREM2 in macrophages exposed to excess salt restored the efferocytic capacity as the engulfed dead/dying neurons per macrophage (Fig. 7C) or the proportion of phagocytic macrophages (PI + F4/80 + ) (Fig. 7D) recovered to that of PBS treated macrophages at 1h after co-culture. Moreover, at 24h after co-culture, protein level of CD206 and Arg1 in TREM2 over-expressed macrophages treated with high concentration of NaCl resembled that treated with PBS (Fig. 7E). Our data revealed that enhancing TREM2 signaling could restore the efferocytic capacity and cellular in ammatory resolution of macrophages which were damaged by surplus salt concentration in the microenvironment.

Discussion
The current study documents that microenvironment with excess salt concentration could impair the in ammatory resolution property of macrophages after ischemic stroke. Mechanistically, surplus salt down regulates TREM2 expression in macrophages, which is associated with decreased efferocytic capacity, excessive neural in ammation and exacerbated stroke outcomes.
It has been reported that high salt diet could promote BBB injury after ischemic stroke (9). Consistently, we recorded that the increased in ltration of multiple leukocytes, including macrophage, neutrophil, T lymphocyte and B lymphocyte, in the stroke lesion of HSD mice at 3d after stroke, could be attributed to the exacerbated BBB damage. It was found that surplus dietary salt directed macrophages/microglia towards the classical activated "M1" phenotype, which further exacerbated stroke outcomes (10). In accordance, our data indicated that the in ammatory resolution property of macrophages were down regulated by excess salt, which led to postponed recovery of stroke lesion.
Efferocytosis represents a key process of in ammatory resolution. Elimination of the dead or injured components within stroke lesion arrests ampli cation of neural in ammation. We demonstrated that the efferocytic capacity, together with the subsequent cellular in ammatory resolution of macrophages, were impaired in high salinity environment, which could be the reason for accumulated dead cells in the stroke penumbra. It has been demonstrated that the function of TREM2 is indispensable for phagocytic activities of microglia and macrophages (18). Our data indicated that TREM2 was down regulated in macrophages by the high salinity environment. Decreased TREM2 expression was correlated with robust post-stroke neural in ammation and exacerbated stroke outcomes, which indicated that inhibition of TREM2 signaling in macrophages was the potential mechanism involved in the detrimental impact of high salt microenvironment.
It has been recognized that high salt diet is a key risk factor for ischemic stroke. Restriction of dietary salt intake serves as an e cient prevention of new vascular events. Nevertheless, no niched therapy that targets the already impaired in ammatory resolution property of macrophages in the high salt environment has been reported. Our study revealed that overexpression of TREM2 could restore the efferocytic capacity and cellular in ammatory resolution of macrophages in high salinity environment. The data appealed further research on the therapeutic potential of enhancing TREM2 signaling in patents of ischemic stroke, especially those with high salt intake.

Conclusions
Conclusively, HSD aggravated ischemic stroke outcomes by exacerbated neural in ammation, which was associated with the impaired in ammatory resolution property of macrophages. TREM2 expression in macrophages was down regulated by high salt environment, and enhancing TREM2 signaling could restore the efferocytic capacity and cellular in ammatory resolution of macrophages. Further study on the value of TREM2 signaling as a therapeutic target in AIS is warranted. Ccl2 chemokine (C-C motif) ligand 2 Cx3cr1 C-X3-C motif chemokine receptor 1

Cxcl5
C-X-C motif chemokine ligand 5 Cxcl7 C-X-C motif chemokine ligand 7 Cxcl9 C-X-C motif chemokine ligand 9 Cxcl10 C-X-C motif chemokine ligand 10 Cxcl11 C-X-C motif chemokine ligand 11 Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.