Efferocytosis is restricted by axon guidance molecule EphA4 via ERK/Stat6/Mertk signaling following brain injury

Background Efferocytosis is a process that removes apoptotic cells and cellular debris. Clearance of these cells alleviates neuroinflammation and prevents the release of inflammatory molecules and promotes the production of anti-inflammatory cytokines to help maintain tissue homeostasis. The underlying mechanisms by which this occurs in the brain after injury remains ill-defined. Methods We demonstrate using GFP bone marrow chimeric knockout (KO) mice, that the axon guidance molecule EphA4 receptor tyrosine kinase is involved in suppressing Mertk signaling in the brain to restrict the function of efferocytosis on resident microglia and peripheral-derived monocyte/macrophages. Results Single-cell RNAseq identified Mertk expression, the primary receptor involved in efferocytosis, on monocytes, microglia, and a subset of astrocytes in the damaged cortex following brain injury. Loss of EphA4 on infiltrating GFP-expressing immune cells improved functional outcome concomitant with enhanced efferocytosis, and overall protein expression of p-Mertk, p-ERK, and p-Stat6. The percentage of GFP+ monocyte/macrophages and resident microglia engulfing NeuN+ or TUNEL+ cells was significantly higher in KO chimeric mice. Importantly, mRNA expression of Mertk and its cognate ligand Gas6 was significantly elevated in these mice compared to wild-type. Analysis of cell-specific expression showed that p-ERK and p-Stat6 co-localized with Mertk-expressing GFP + cells in the peri-lesional area of the cortex following brain injury. Using an in vitro efferocytosis assay, co-culturing pHrodo-labeled apoptotic Jurkat cells and bone marrow (BM)-derived macrophages, we demonstrate that efferocytosis efficiency and mRNA expression of Mertk and Gas6 was enhanced in the absence of EphA4. Select inhibitors of ERK and Stat6 attenuated this effect confirming that EphA4 suppresses monocyte/macrophage efferocytosis via inhibition of the ERK/Stat6 pathway. Conclusions Our findings implicate the Mertk/ERK/Stat6 axis as a novel regulator of apoptotic debris clearance in brain injury that is restricted by peripheral myeloid-derived EphA4 to prevent the resolution of inflammation.

Generation of bone marrow chimeric mice. Bone marrow ablation was induced by exposing EphA4 f/f recipient mice (6-8 weeks) to X-ray irradiation (two doses of 550 rad at six h apart). Within 24 h of irradiation, recipient mice were intravenously injected with 3-4 million bone marrow cells (BMCs) isolated from femur and tibia of donor Rosa26 mtmg /Tie2-Cre/EphA4 +/+ or Rosa26 mtmg /Tie2-Cre/EphA4 f/f mice as previously described [10,12]. Recipient mice were placed on gentamycin sulfate water (1 mg/ml) for three days before irradiation and two weeks after adoptive transfer. Controlled cortical impact (CCI) injury for adoptive transfer WT (WT + WTBMCs ) and KO (WT + KOBMCs ) was performed at four weeks post-chimera generation.
Controlled cortical impact injury. CCI injury was induced as previously described [10,11]. Brie y, mice were anesthetized with subcutaneous ketamine (100 mg/kg) and xylazine (10 mg/kg) and securely positioned in a stereotaxic frame, maintaining a body temperature of 37°C. A midline incision was made on the sanitized and shaved scalp to expose the skull. Then, a craniectomy (4 mm diameter) was performed over the right parietal-temporal cortex (− 2.5 mm A/P and 2.0 mm lateral from the bregma) using a portable drill. The injury was induced with a velocity of 5 m/s and a depth of 2 mm. The incision was sutured, and animals were monitored until recovery.
Perfusion xation and brain serial sectioning. At 1-or 3-days post-CCI injury (dpi), mice were euthanized by subcutaneously (s.c.) injecting a combination of 150 mg/kg ketamine and 20 mg/kg xylazine and perfusion xation was performed as previously described [11]. Five minutes before perfusion, heparin (2000 units/kg, s.c.) and sodium nitroprusside (0.75 mg/kg, s.c.) were injected. After con rming the loss of pedal re ex, cardiac perfusion of heparin (20 units/ml) in phosphate-buffered saline (PBS) was performed to clear the circulation of blood, followed by perfusion with ice-cold 4% paraformaldehyde (PFA) in PBS. Fixed brains were cryoprotected with gradient sucrose solutions, snap-frozen, embedded in OCT compound with 30% sucrose, and stored at -80°C until sectioning. Serial coronal sections (30 µm) were obtained using a cryostat (− 1.1 to − 2.6 mm posterior from bregma), mounted on positively charged slides (with ve sections spaced 450 µm apart), and stored at -80°C for further analysis.
Single-cell sequencing, library generation, Gene Ontology enrichment, and computational analyses. At 1day post-CCI injury, CD1 mice were euthanized using iso urane and 4 mm x 4mm ipsilateral (injured) cortical tissues were harvested and dissociated using papain digest neural dissociation kit (Miltenyi Biotech). Cells were then cryopreserved in 1mL of CryoStore® CS10 media (Stem cell technologies, Seattle, WA, USA) and sent to Medgenome for scRNAseq (Foster City, California, USA). ScRNA-seq libraries were generated using the Chromium Next GEM Single Cell 5' v2 chemistry (10x Genomics) and sequenced on a NovaSeq 6000 (Illumina). An alignment of the libraries and read counts were performed using mouse RNA-Seq Database consisting of expression values of 358 bulk Mouse RNA-seq samples of sorted cell populations (Cell Ranger V7.0, 10X Genomics). Quality control was performed excluding the genes that are not expressed in at least 3 cells and the cells that do not express > 200 genes (Seurat, V 4.1.0 Read10X function). Doublet Finder package, v2.0.3 was used to lter for doublets, LogNormalize was used for global-scaling normalization, and Seurat was used for clustering. Unbiased cell type recognition from scRNAseq data was performed using Cellenics. The following markers were used for the clustering of endothelial cells (Cd31, Tie2, Cdh5, Glut1), microglia (Tmem119, Ccr2-, cx3cr1), astrocytes (GFAP, Aldhl1, sox9), and monocytes/macrophages (Ccr2, Mmp8) as well as program automated clusters by brain tissue type. We applied modularity optimization by Louvain algorithm to iteratively group cells together and visualize the data using UMAP.
Quantitative real-time PCR. Total RNA was freshly isolated from cultured bone marrow-derived macrophages, ipsilateral (4x4 mm), or contralateral (4x4 mm) cortical tissue using TRIzol® reagent (Ambion) according to manufacturer's instructions. RNA concentration was measured using a spectrophotometer ND-1000 (NanoDrop). DNAse I treatment (for a 1000 ng RNA) was performed to degrade the genomic DNA using an ampli cation grade Kit (Sigma Aldrich, St. Lois, MO). To synthesize cDNA, reverse transcription was performed using iScript™ cDNA synthesis kit (Biorad, Hercules, CA). For qRT-PCR, iTaq™ Universal SYBR® Green Supermix (Biorad, Hercules, CA) and speci c primers (Table 1) were used to amplify cDNA (50 ng) following the manufacturer's instructions. mRNA expression (fold change) was calculated relative to GAPDH as an internal control using 2 -ΔΔCT method.
Stereological cell counts. Cell quanti cation was assessed by a blinded investigator using the optical fractionator probe function of Stereoinvestigator (MicroBrightField, Williston, VT, USA) on an upright Olympus BX51TRF motorized microscope (Olympus America, Center Valley, PA, USA) with a grid size set at 450 × 450 mm and 150 × 150 mm counting frame for cortex as previously described [16,17].
TUNEL staining/imaging and counting. TUNEL staining of serial coronal sections of perfused xed brains was performed using the Click-iT Plus TUNEL Assay; 647 (Thermo Fisher Scienti c) according to the manufacturer's instructions. Serial coronal sections (300 µm) were permeabilized with 0.2% Triton X-100 in phosphate-buffered saline (PBS) and incubated with the TUNEL reaction mixture containing terminal deoxynucleotidyl transferase enzyme followed by Alexa Fluor 647-conjugated nucleotides. Following the TUNEL reaction, the sections were blocked in 2% cold water sh gelatin (Sigma, Inc.) in 0.2% triton for 1 h, then incubated with Rb anti-IBA1 (Wako) or Rb anti-NeuN (cell signaling) antibody (1:200 in blocking solution) overnight. Slides were then washed with 1×PBS, treated with AlexFluor donkey anti-rabbit-555 (1:250 in block) for 1 hour, further washed in 1×PBS, and then mounted in media with DAPI counterstain (SouthernBiotech). Z-stack images were acquired using a Zeiss 880 confocal microscope (Carl-Zeiss, Oberkochen, Germany).
Protein bands were imaged with LI-COR Odyssey system and quanti ed using Fiji Image-J software.
Cerebral blood ow analysis. A laser speckle contrast imaging system (RFLSI III Laser Speckle Imaging System, RWD Life Science, Dover, DE, USA) was used to scan cerebral blood ow over the right parietal cortex after performing a craniectomy (4 mm diameter). Cerebral blood ow was recorded in the region of interest (ROI, 2.5 mm diameter) for 30 seconds and 4 readings for each recording were taken before injury (baseline), at 10 min, and 3 dpi. The average of the 4 readings and the percentage from the baseline were calculated for each mouse. Laser speckle contrast (LSC) and bright eld images were captured at every recording.
Blood-brain barrier permeability (IgG deposition). Serial coronal sections of perfused xed brains were blocked in 2% cold water sh gelatin (Sigma, Inc.) in 0.2% triton for 1 hour, incubated with AlexFluor donkey anti-mouse-594 (1:250 in block) for 1 hour, washed in 1×PBS, and then mounted in media with DAPI counterstain (SouthernBiotech). The volume of IgG deposition (mm 3 ) was measured using the Cavalieri Estimator from StereoInvestigator software.
Statistical Analysis. Data are presented as mean ± standard error of the mean (SEM), graphed, and analyzed using the GraphPad Prism program, version 9 (GraphPad Software, Inc., San Diego, CA). The intergroup variations were analyzed using Student's unpaired two-tailed t-test (to compare between two experimental groups), one-way analysis of variance (ANOVA) followed by Bonferroni's multiple comparisons test (to compare multiple groups with one independent variable), or two-way ANOVA followed by Šídák's multiple comparisons test (to compare multiple groups with two independent variables). The variations were considered signi cant at P < 0.05.

Results
Efferocytosis-related gene expression across single cells in the damaged cortex. Efferocytosis has not been fully characterized in brain injury as a result of trauma. To identify speci c cell types across the damaged cortex that expression efferocytosis-related genes, we conducted scRNA-seq expression analysis on cells dissociated from the ipsilateral cortex at 1-day post-CCI injury. A total of 5000 cells were included in our analysis. Uniform Manifold Approximation and Projection (UMAP) plot shows different clusters of microglia, astrocytes, endothelial cells, pericytes, neutrophils, and monocytes/macrophages isolated from the injured cortex based on RNA gene expression (Fig. 1A). The analysis of differential expression of efferocytosis signals in different cell clusters revealed the expression of efferocytic receptor, Mertk, is present predominately in microglia and astrocytes, as well as monocyte/macrophages and endothelial cells. Its cognate ligands, Gas6 and Pros1, predominate in microglia and endothelial cells (Fig. 1B). Feature UMAPs highlight the spatial location of these genes in the different clusters con rming their enriched expression (Fig. 1C-1F).
Upregulation of Mertk in the ipsilateral cortex following CCI injury. To investigate the effect of CCI injury on efferocytosis-related genes, we measured temporal mRNA expression of receptors involved in apoptotic cell recognition and engulfment in the ipsilateral cortex at 1, 3, and 7dpi. We observed a signi cant increase in recognition receptors (S1pr1 and Cx3cr1), engulfment receptors (Mertk), and bridging molecules (Gas6 and Pros1) at 1, 3, and 7dpi ( Fig. 2A-2E). To determine if Mertk protein is upregulated in Cx3cr1-expressing cells following CCI injury, we used Cx3cr1 CreER mice, which express EYFP in Cx3cr1-positive cells. Immunohistochemical analysis using Iba1 (a marker for microglia and PDM) revealed increased Mertk expression on IBA1+/Cx3cr1 EYFP+ cells located in the peri-lesion at 3 dpi compared to the contralateral cortex ( Fig. 2G-2O). To distinguish between microglia and PDMs, we used GFP bone marrow chimeric wild type mice. Quanti cation of Mertk was measured in serial coronal sections immuno-stained for IBA1 at 1 and 3 dpi. The estimated number of Mertk+/GFP+/IBA1+, PDMs and Mertk+/GFP-/IBA + microglia was measured in the ipsilateral cortex using non-biased stereology, optical fractionator probe, Stereoinvestigator. Consistent with scRNAseq data, Mertk expression was present on microglia and PDMs, however more microglia where positive for Mertk at 1dpi. Microglia expression remained consistent at 3dpi, however the estimated number of PDMs expressing Mertk was much higher. This could re ect the increased tra cking of monocytes to the brain at this time point (Fig. 2P-2Y). We show a distinct temporal change in expression of Mertk across two distinct myeloid populations following brain injury.
Initiation of efferocytosis by Cx3cr1 + microglia/macrophages following CCI injury. To determine if the change in Mertk expression in microglia and PDMs is correlated with the initiation of apoptotic neuron engulfment, coronal sections of CCI-injured Cx3cr1 CreER mice were stained with TUNEL and anti-NeuN antibody (Fig. 3). Z-stacked confocal images show Cx3cr1 EYFP+ efferocytes containing DAPI+/TUNEL + and/or DAPI+/NeuN + nuclei along with DAPI+/NeuN-/TUNEL-nuclei in the peri-lesion cortex at 3 dpi ( Fig. 3F-3J). This data provides the rst evidence for the engulfment of dead/dying neurons by microglia and PDMs in traumatic brain injured mice. or TUNEL + cells (cells contain at least one TUNEL + and one TUNEL-nuclei; Fig. 4I-4K) was counted in the core and/or peri-lesion of the injured cortex. We found that the percentage of PDMs and microglia containing 2 + nuclei is signi cantly higher in the lesion core (Fig. 4C) and peri-lesion (Fig. 4D) of WT + KOBMCs than WT + WTBMCs . The percentage of PDMs engul ng NeuN + cells in the lesion core (Fig. 4G) and the peri-lesion (Fig. 4H), as well as TUNEL + cells in the lesion core (Fig. 4K), is higher in WT + KOBMCs than WT + WTBMCs . Microglia engul ng NeuN + cells are signi cantly higher in the peri-lesion of WT + KOBMCs compared to WT + WTBMCs . No signi cant difference was observed in microglia engul ng NeuN + or TUNEL + cells in the lesion core. This data suggests that EphA4 de ciency in BM-derived immune cells enhances the efferocytosis capacity of PDMs and peri-lesional microglia.
De ciency of peripheral-derived EphA4 enhances ERK/Stat6/Mertk signaling in the CCI-injured cortex. Given that myeloid Mertk-mediated efferocytosis signaling may predominant in the brain following CCI injury and is enhanced in the absence of EphA4 (Figs. 1 & 2), we sought to investigate the mRNA and protein phosphorylation status of key players in this pathway. We found that mRNA expression of Mertk and Gas6 was signi cantly higher in the ipsilateral cortex of WT + KOBMCs at 3dpi when compared to WT + WTBMCs (Fig. 5A, 5B). No signi cant difference was observed in Pros1 transcript (Fig. 5C). Further, Western blot analysis showed increased phosphorylation of Mertk (P-Mertk), ERK 1/2 (P-ERK), and Stat6 (P-Stat6) in the ipsilateral cortex of WT + KOBMCs at 3 dpi compared to WT + WTBMCs (Fig. 5D & Supp. Figure 1). To speci cally address whether ERK and Stat6 are present on Mertk expressing cells, coronal sections were immune-stained with anti-Mertk and anti-P-ERK (Fig. 5E & 5F) or anti-Mertk and anti-P-Stat6 (Fig. 5G &  5H). Confocal images show colocalization of P-ERK and P-Stat6 with MERK in the ipsilateral cortex of WT + KOBMCs mice and that Stat6 is increased on GFP expressing cells in KO mice compared to WT. This indicates the co-expression of these key signaling molecules on in ltrating efferocytes.
Enhanced efferocytosis in EphA4-de cient bone marrow-derived macrophages (BMDMSs) is mediated by ERK/Stat6 pathway. To con rm that EphA4 de ciency in bone marrow-derived macrophages enhances the clearance of apoptotic debris, an in vitro efferocytosis assay was performed. GFP + BMDMSs from Epha4 +/+ /ROSA mTmG /Tie2-Cre (WT) or Epha4 f/f / ROSA mTmG /Tie2-Cre (KO) mice were co-cultured with pHrodo-stained live or apoptotic Jurkat cells. Apoptosis was induced by treating Jurkat cells with staurosporine. After 3 hours of treatment, 90% of Jurkat cells were in the early phase of apoptosis (Annexin V + and PI − , Supp. Figure 2). No engulfment of live Jurkat cells was observed in EphA4 WT or EphA4 KO macrophages (Fig. 6A-6H). More engulfment of apoptotic Jurkat cells was observed in untreated, LPS-treated, and HMGB1-treated EphA4 KO macrophages when compared to WT (Fig. 6I).
Importantly, treating macrophages with EphA4-FC clusters (to activate the reverse Ephrin signaling) did not reduce the efferocytosis e ciency of EphA4 KO macrophages con rming the enhanced efferocytosis in KO macrophages is mediated by the blockade of EphA4 forward signals (Fig. 6J). In addition, Mertk and Gas6 expression is signi cantly upregulated in EphA4 KO macrophages in the absence of apoptotic Jurkat cells compared to WT macrophages. After engul ng apoptotic Jurkat cells, Mertk and Gas6 expression increased in WT and KO macrophages; however, the expression is signi cantly higher in KO than in WT macrophages (Fig. 6K, 6L). Pros1 expression increased only in EphA4 KO macrophages after apoptotic Jurkat cell engulfment (Fig. 6M). To determine if enhanced efferocytosis in EphA4 KO macrophages is mediated by Mertk, ERK, and Stat6 activation, selective inhibitors were used. Mertk inhibitor (UNC2250) reduced efferocytosis e ciency of both WT and KO macrophages; however, ERK inhibitor (FR18020R) and Stat6 inhibitor (AS1517499) selectively reduced efferocytosis e ciency of EphA4 KO macrophages (Fig. 6N). Interestingly, ERK inhibition reduced Mertk expression in EphA4 KO macrophages engul ng apoptotic cells, and Stat6 inhibition reduced Mertk and Gas6 expression. A nonsigni cant difference was observed in Pros1 expression in the presence of ERK or Stat6 inhibitors (Fig. 6O). Data suggests that EphA4 forward signaling reduces macrophage efferocytosis by inhibiting ERK and Stat6 activation, which in turn reduces the expression of efferocytosis receptor (Mertk) and its ligand (Gas6) and restricts the efferocytosis process (Fig. 6P).
Peripheral EphA4 de ciency reduced apoptosis and improved functional outcomes following CCI injury.
To determine if improved efferocytosis in bone marrow chimeric EphA4 KO mice is associated with reduced apoptotic cell accumulation, the number of apoptotic cells was counted in serial coronal sections stained with TUNEL at 3 dpi using the optical fractionator probe function on Stereoinvestigator (Fig. 7A).
The total number of apoptotic cells (total TUNEL+) and apoptotic peripheral-derived immune cells (GFP + TUNEL+) were signi cantly reduced in WT + KOBMCs compared to WT + WTBMCs mice (Fig. 7B-7H). This effect was correlated with improved cerebral blood ow and reduced blood-brain barrier permeability. Cerebral blood ow was measured at 10 minutes and 3dpi using Laser Speckle Contrast Imaging (LSCI) and calculated as percentage ow from the baseline (pre-injury). WT + KOBMCs showed a signi cant increase in cerebral blood ow at 3dpi compared to WT + WTBMCs (Fig. 7I, 7J). Blood-brain barrier permeability was measured at 3 dpi by estimating the volume of IgG deposition in serial coronal section using the Cavalieri Estimator from StereoInvestigator software. Signi cant reduction in IgG deposition was observed in WT + KOBMCs compared to WT + WTBMCs indicating improved blood-brain barrier function and neuroprotection (Fig. 7K). We nd greater improvements in functional outcome correlate with efferocytosis enhancement by gene deletion of EphA4 on in ltrating innate immune cells.

Discussion
Apoptosis is a prominent mode of cell death that mediates tissue damage and may contribute to the propagation neuroin ammation following brain injury. In cases of excessive apoptotic cell death, an overwhelmed clearance mechanism may result in the accumulation of apoptotic cells and cellular debris, prolonging the exposure to pro-in ammatory signals and activated immune cells. Therefore, a greater understanding of the efferocytotic process may aid in the development of strategies targeting the timely removal of dying cells in the brain. Our results demonstrate an upregulation of " nd-me" signal receptors (S1pr1 and Cx3cr1), engulfment receptor (Mertk), and bridging molecules (Gas6 and Pros1) in the damaged cortex, indicating the onset of efferocytosis acutely following trauma. We identi ed that resident (microglia) and peripheral-derived (monocyte/macrophages; PDMs) myeloid cells show distinct temporal expression patterns of Mertk. Notably, we nd that ephrin type-A receptor 4 (EphA4) de ciency on in ltrating immune cells enhanced the capacity of PDMs and peri-lesional microglia for efferocytosis, as well as enhanced cortical expression of Mertk, and Gas6 transcripts. Our investigation into the molecular mechanisms revealed increased phosphorylated (p) levels of p-Mertk, p-ERK and p-Stat6 in the injured cortex of chimeric EphA4 KO mice. In addition, inhibition of ERK and Stat6 attenuated the enhanced efferocytosis and Mertk expression in EphA4-de cient BMDMSs in vitro. This demonstrates that EphA4 suppresses efferocytosis by inhibiting the ERK/Stat6 pathway. These ndings highlight a new and novel role for the axon guidance molecule EphA4 in regulating the coordinated process of efferocytosis, which may contribute to the overall neuroin ammatory milieu after brain injury.
EphA4 has been identi ed as a regulator of neuroin ammation and secondary injury following brain trauma. Our previous studies utilizing mouse models and bone marrow chimeric approaches have demonstrated that the absence or inhibition of EphA4 in peripheral-derived monocytes/macrophages results in neuroprotection, reduced cortical in ltration of monocytes/macrophages, and a shift in their gene pro le from a proin ammatory to an anti-in ammatory state [10,12]. These ndings underscore the pivotal role of EphA4 in mediating the pro-in ammatory phenotypic state of PDM, thereby prompting further investigation into whether EphA4 hampers the adequate clearance of apoptotic debris by modulating the phenotypic state of efferocytes following CCI injury. Interestingly, deletion of EphA4 on in ltrating immune cells fosters the augmentation of efferocytosis in both PDMs and resident microglia. This may be due to the fact that EphA4-de ciency in monocytes enhances expression of Gas6, which may fuel the neighboring microglia and promote their efferocytotic abilities. This also suggests that monocytes may have key properties that may allow them to communicate with and regulate microglial function.
Eph receptors and ephrin ligands facilitate intercellular communication to regulate diverse processes, including adhesion, repulsion, migration, survival, proliferation, remodeling, and differentiation. Upon binding, the clustering of Eph receptors and ligands triggers a cascade of signaling events, in both the cells expressing the receptor and those bearing the ligand [19]. EphA4-null BMDMCs, treated with clustered EphA4-FC, showed no effect on efferocytosis e ciency con rming forward EphA4 signaling on monocyte/macrophages mediates the suppression of ERK/Stat6 signaling and subsequent enhancement of efferocytosis. Whether the genetic deletion of PDM EphA4 may impair reverse ephrin signaling on apoptotic cells and its relevance in the current study remains unknown and requires additional investigation. Forward Eph signaling exhibits a remarkable ability to attenuate the RAS-ERK pathway, overriding its activation by other receptor tyrosine kinases [19,20]. This inhibition has been observed in diverse scenarios, including the modulation of growth cone motility in neurons [21] and the suppression of tumorigenicity in cancer cells [22,23]. The mechanism underlying Eph receptor-dependent ERK inhibition often involves the activation of the RAS GTPase, p120 RAS GAP, which leads to the subsequent inactivation of H-RAS [24]. Conversely, Eph receptors can activate the RAS-ERK pathway in speci c contexts, promoting cellular processes such as proliferation, early gene transcription, cell migration, or repulsion [25,26]. Eph receptor signaling exerts a complex and context-dependent in uence on the RAS-ERK pathway, however, our novel ndings indicate that EphA4 suppresses this signaling to limit efferocytosis in the brain following injury. ERK1/2 signaling modulates multiple aspects of efferocytosis and in ammation resolution, including the regulation of phagocytic receptor expression, cytoskeletal rearrangement, and the production of anti-in ammatory molecules [27,28]. Activation of the ERK1/2 signaling pathway can enhance the expression of phagocytic receptors, including Mertk, Gas6, MFG-E8, and integrins (such as αvβ3) [29][30][31]. In addition, ERK1/2 signaling is implicated in promoting actin polymerization and cytoskeletal rearrangement, thereby facilitating the process of efferocytosis [32]. Importantly, ERK1/2 activation also regulates the production of anti-in ammatory mediators during efferocytosis, promoting the polarization of macrophages toward a pro-resolving phenotype, which contributes to sustained efferocytosis and the resolution of in ammation [33,34]. One crucial transcription factor in uenced by ERK1/2 activation to regulate macrophage polarization is Stat6 [35,36]. The phosphorylation of Stat6 promotes the transcription of genes involved in macrophage pro-resolving and efferocytosis response, such as Gas6 and Mertk [37,38]. In the present study, we observed an activation of Mertk, ERK1/2, and Stat6 in the injured cortex of chimeric EphA4 KO mice. Furthermore, selective inhibition of ERK and Stat6 reduced efferocytosis and Mertk expression, speci cally in EphA4-de cient BMDMS in vitro. These ndings suggest that EphA4 impedes efferocytosis by inhibiting the P-ERK/P-Stat6/Mertk signaling pathway.

Conclusions
Efferocytosis occurs acutely in the brain following trauma in a limited capacity. Blockade of EphA4 receptor function improves this activity, which coincided with tissue protection, restoration of cerebral blood ow and BBB stability. Unraveling the molecular mechanisms underlying efferocytosis in TBI will led to new therapeutic avenues promoting this process and in order to mitigate the deleterious consequences of apoptotic death and facilitating optimal recovery following brain trauma. Declarations 30. Kurihara, Y., T. Nakahara, and M. Furue, alphaVbeta3-integrin expression through ERK activation mediates cell attachment and is necessary for production of tumor necrosis factor alpha in monocytic THP-        images. I, J) Cerebral blood ow is improved in WT +KO BMCs compared to WT +WT BMCs . Blood ow was measured using laser speckle contrast imaging (LSCI) at 10 minutes and 3 days post-injury and presented as a percentage of the baseline. K) Blood-brain barrier permeability is improved in WT +KO BMCs compared to WT +WT BMCs . The volume of IgG deposition was measured at 3 dpi in serial coronal sections