Fig 2: Phagocytosis of cellular debris in reactive astrocytes is mediated by AXL. (A–D) Representative TEM images of the astrocytes in the ipsilateral cortex (A, blue), those in the 3 dpi group were characterized by prominent intermediate filament bundles (B, pink area) and abundant glycogen granules (B, asterisks), the cytoplasm often contained small pieces of cellular debris (arrow). (E) The ingested debris was completely included in the astrocytic cytoplasm (arrow), and even situated in a cytoplasmic vacuole within astrocytes. (F) Analysis of phagocytic astrocytes participating in debris clearance after TBI. *p < 0.05, **p < 0.01 by one-way ANOVA). (G) rmGas6 increased the phagocytic astrocytes at 3 dpi. *p < 0.05 by two-tailed Student’s t-test. (H) Dot plots of debris density (number/area [µm2]) in astrocytes in the sham and 3 or 7 dpi ipsilateral cortex. **p < 0.01 versus sham, #p < 0.05 versus 3 dpi by one-way ANOVA. (I) Astrocytic phagocytosis enhanced in rmGas6 treated ipsilateral cortex at 3 dpi. **p < 0.01 by two-tailed Student’s t-test

Fig 3: AXL influences TBI-induced neuroinflammation and neurological impairments. (A) Inflammation cytokines are determined at 3 dpi under drug treatment. IL-6/TNFa/IL-1ß were upregulated at 3 dpi and rmGas6 treatment attenuated the inflammation upregulation, whereas R428 treatment intensified neuroinflammation. **p < 0.01 versus sham, #p < 0.05, ##p < 0.01 versus vehicle by one-way ANOVA. (B) Drug treatment mediates TBI-induced body weight changes. The relative change of body weight was calculated by comparing with 0 dpi. *p < 0.01 versus sham, #p < 0.05, ##p < 0.01 versus vehicle group by two-way ANOVA. (C) Motor coordination was reduced by TBI, and the disorder was enhanced by R428 treatment while rmGas6 attenuated this change. **p < 0.01, *p < 0.05 versus sham, ##p < 0.01 versus vehicle by two-way ANOVA. (D) Temporal neurological function assessment of mice suffered TBI at different time points by using modified neurological severity score test (mNSS). **p < 0.01 versus sham, ##p < 0.01 versus 1 dpi by one-way ANOVA. (E) mNSS at 3 dpi under drug treatment. **p < 0.01 by one-way ANOVA. (F) Depressive-like behavior was determined after TBI with/ without drug administration by time spent immobile in the TST at 7 dpi (left) and 14 dpi (right). All data are presented as mean ± SD; (n = 6). **p < 0.01 versus sham, ##p < 0.01 versus vehicle by one-way ANOVA

Fig 4: AXL mediates the phagocytic transformation of astrocytes in vitro. (A) Flow cytometry indicated the temporal phagocytosis ratio change of astrocytes. n = 3 per time point. (B) Representative images of PKH26 labeled apoptotic neurons (red) engulfed by astrocytes (green). The phagocytic astrocytes had a phagocytic cup structure (arrowhead) that contained PKH26+ debris or cells. (C) Brief process of in-vitro astrocytic phagocytosis analysis. (D) Immunofluorescence staining of astrocytic phagocytosis and quantification analysis of phagocytosis ratio with rmGas6 or R428 treatment after 24 h co-incubated. n = 3 per group. *p < 0.05, **p < 0.01 by one-way ANOVA. (E) The astrocytic ingestion was assessed by flow cytometry after 24 h co-culture. Astrocytes were pretreated with R428 or rmGas6 for 1 h prior to the addition of PKH+ neurons. n = 3 per group. **p < 0.01 by one-way ANOVA

Fig 5: The schematic diagram demonstrating that AXL which initially activated by microglia derived GAS6 modulates transformation of astrocytes from resting to phagocytic phenotype through STAT1-ABCA1 signaling pathway to alleviate the inflammation and ROS injury after TBI