Growth arrest specific protein 6 alleviated white matter injury after experimental ischemic stroke

Abstract

Ischemic white matter injury leads to long-term neurological deficits and lacks effective medication. Growth arrest specific protein 6 (Gas6) clears myelin debris, which is hypothesized to promote white matter integrity in experimental stroke models. By the middle cerebral artery occlusion (MCAO) stroke model, we observed that Gas6 reduced infarcted volume and behavior deficits 4 weeks after MCAO. Compared with control mice, Gas6-treatment mice represented higher FA values in the ipsilateral external capsules by MRI DTI scan. The SMI32/MBP ratio of the ipsilateral cortex and striatum was profoundly alleviated by Gas6 administration. Gas6-treatment group manifested thicker myelin sheaths than the control group by electron microscopy. We observed that Gas6 mainly promoted OPC maturation, which was closely related to microglia. Mechanically, Gas6 accelerated microglia-mediated myelin debris clearance and cholesterol transport protein expression (abca1, abcg1, apoc1, apoe) in vivo and in vitro, accordingly less myelin debris and lipid deposited in Gas6 treated stroke mice. HX531 (RXR inhibitor) administration mitigated the functions of Gas6 in speeding up debris clearance and cholesterol transport protein expression. Generally, we concluded that Gas6 cleared myelin debris and promoted cholesterol transportation protein expression through activating RXR, which could be one critical mechanism contributing to white matter repair after stroke.

Introduction

Ischemic white matter injury either secondary to cerebral small vessel disease or large vessel occlusion contributes to long-term sensory-motor dysfunction and cognitive impairment.^{1 –3} However, there is a lack of effective therapeutic strategies in the clinic currently. Failure of remyelination is mainly owed to the death and insufficient maturation of oligodendrocyte progenitor cells (OPCs) after ischemic white matter injury. Stroke elicits abundant myelin debris or other cell debris, which overwhelms the processing capability of cerebral phagocytes. ⁴ Myelin debris contains inhibitors of OPC differentiation in the injured central nervous system, and therefore its clearance by microglial phagocytosis is essential to creating an optimal milieu for remyelination.^5,6

Growth arrest specific protein 6 (Gas6) is one of the ligands for TAM receptors (Tyro3, Axl, Mer) with the highest affinity for Axl. ⁷ Gas6 is widely expressed in the central nervous system by microglia, astrocytes, endothelial cells, and neurons.^8,9 Upon cerebral ischemia, microglia/macrophages increase the expression of Gas6 and Axl. ⁹ Microglia are the professional resident phagocytes in the brain. The importance of Gas6 in debris engulfment has been well documented in the cuprizone-induced demyelination model.^{10 –13} Therefore, Gas6 plays a critical role in maintaining white matter integrity and regulating cerebral inflammation in experimental multiple sclerosis research. Emerging evidence demonstrated that Gas6 reduces infarct volume and improves neurological outcomes after stroke by suppressing acute inflammatory response and reserving blood-brain barrier (BBB) integrity.^9,14 However, the therapeutic effects of Gas6 on white matter integrity in the rehabilitation stage after stroke have not been investigated. Additionally, whether microglia-mediated myelin clearance is involved in the protective process of Gas6 in ischemic stroke remains unknown.

The clearance of cholesterol-rich myelin debris by microglia could switch the lesion circumstance from pro-inflammation to anti-inflammation status.^5,15 Meanwhile, cerebral cholesterol is provided by de novo syntheses or self-recycling because peripheral cholesterol cannot cross BBB. Therefore, cholesterol recycling in myelin-phagocytosing microglia/macrophages is also essential for endogenous remyelination. ¹⁶ Gas6 and Axl double-knockout mice present pronounced pathological lipid droplets deposit, ¹⁰ which raises an intriguing question that Gas6 might be not only responsible for debris clearance but also participate in defective cholesterol transportation. The retinoid X receptors (RXRs) belong to the nuclear receptor superfamily and play potential specific roles in cholesterol metabolism and inflammatory response through forming functional heterodimers with liver X receptors (LXR) and peroxisome proliferator-activated receptors (PPAR).^17,18 Additionally, phagocytes lacking RXR/LXR accumulate cholesteryl ester and are prone to become foam cells.^19,20 We hypothesized that Gas6 might modulate RXR activation, which promotes myelin clearance as well as facilitates cholesterol transportation.

In this study, by using the middle cerebral artery occlusion (MCAO) mouse model, we sought to study the therapeutic potential of Gas6 for ischemic white matter injury four weeks after stroke. Furthermore, we aimed to identify the necessity of microglia-mediated myelin clearance and the expression of the cholesterol-transfer protein in ischemic white matter repair by Gas6 administration, which provided a more theoretical basis for clinical applications.

Methods

Experimental animals

All animal experimental procedures were approved by the Animal Use and Care Committee of Affiliated Drum Tower Hospital, Medical School of Nanjing University (permit no. 2020AE01122) and performed according to the Institutional Guide for the Care and Use of Laboratory Animals. All animal experiments were followed with the ARRIVE 2.0 guidelines. ²¹ Eight-week-year-old C57BL/6J male mice (20–25 g) were purchased from the Animal Model Center of Nanjing Medical University (Nanjing, Jiangsu, China). The mice were bred and housed in an air-conditioned, temperature-controlled, and humidity-controlled room under a 12-hour light/dark cycle. Drugs and animal strains were arranged and labeled by an independent researcher according to the randomization plan.

Transient middle cerebral artery occlusion model mice and treatment administration

We used the intraluminal filament technique to establish a transient middle cerebral artery occlusion stroke model which has a good survival rate and applies a longer ischemic period. Firstly, the mouse was anesthetized with 2.5% Avertin, and the body temperature was maintained at 37 ± 0.5 °C with a heating pad. After a small midline neck incision was made under a dissecting microscope, the right common carotid artery (CCA) was isolated from the external carotid artery (ECA) and ligated using a 4-0 silk suture. Then we made a wedge-shaped incision and introduced a filament (Doccol Corporation, MA, USA) into the ECA and further inserted it to obstruct the middle cerebral artery. A small transverse incision was then made between the eye and ear to test cerebral blood flow (CBF). Mice were included if the Doppler laser reading CBF was dropped more than 20% of baseline. After one hour of occlusion, the filament was withdrawn for reperfusion. Sham-treated mice were subjected to the same procedure without introducing the filament. The mice were treated with Gas6 or PBS after occlusion by nasal administration and repeated once every day until the 28^th day after MCAO. PBS or Gas6 dissolved in PBS (20 mg/kg, 0.5 μg/μl, Abbexa, UK) were delivered alternately into the bilateral nares of the mouse as slowly as possible.

Behavioral tests

Behavioral tests were performed by an observer blinded to experimental groups. The rotarod test was performed to assess motor coordination function after the stroke. Mice were forced to run on a five-lane rotarod device (IITC Life Science) at 20 rpm, 30 rpm, and 40 rpm for 5 minutes separately. Training the mice two trials per day for 3 days before surgery with an interval of 15 min. On the test day, the mice were placed on the rod in turn, and the average time of latency to fall for three rounds of experiments was recorded. The adhesive removal test was used to evaluate sensory-motor impairments after stroke. Mice were trained with one trial per day for 3 days. Two 2 × 3 mm adhesive tapes were applied to the forepaws of each mouse and then placed in a transparent box. Tactile responses were measured by recording the time to remove the adhesive tape, with a maximum observation period of 2 minutes. The Morris water maze test was used to test the cognitive function of mice on the 28^th-day post-MCAO. During the learning test (23^rd-day to 27^th-day post-stroke), allow a mouse to find the submerged platform with a maximum period of 1 minute from four locations to place the mice into the pool. Training the mice to stay on the platform for 30 s after each trial whether they found the platform or not. For the probe test on the 28^th-day post-stroke, the mice were placed into the pool from two of the four locations which are on the diagonal, and allowed to swim freely for 60 s without the platform. All the data were recorded and analyzed by the ANY-maze system (Stoelting, USA).

Immunofluorescence staining and confocal microscopy analysis

Mice were anesthetized and then transcardially perfused with PBS followed by 4% paraformaldehyde (PFA). Brains were dissected and fixed in 4% PFA, dehydrated in the 15% sucrose solution on the second day, and the 30% sucrose solution on the third day. To achieve the slices, brains were dried and frozen at -80°C in advance. After being fixed on the metal base with optimal cutting temperature compound, brains were sectioned at 20 μm and attached to the glass slides. For immunofluorescence staining, cells were fixed in 4% PFA for 15 min at room temperature in advance. Brain slices or cells were permeabilized with 0.25% Triton X-100 for 20 min followed by blocking with 2% bovine serum albumin (BSA) for 120 min and then incubated with primary antibodies for 24 hours at 4 °C. Following primary antibodies were used: Anti-APC (CC-1), 1:100, Biochem, #OP80; Anti-CD140a (pdgfrα), 1:500, BD Biosciences, #558774; Anti-dMBP, 1:500, Millipore, #ab5864; Anti-Iba1, 1:500, Wako, #019-19741; Anti-MBP, 1:500, Abcam, #ab7349; Anti-NeuN, 1:500, Abcam, #ab177487; Anti-NF-H Antibody (SMI32), 1:500, BioLegend, #801702; Anti-Olig2, 1:500, Millipore Sigma, #ABN899; Anti-TMEM119, 1:500, Synaptic Systems, #400011. The brain slices or the cells were incubated with secondary antibodies for 2 h at room temperature in the dark the next day. FluoroMyelin (Invitrogen, #F34651) was diluting 300-fold into PBS and stained for 20 min at room temperature before staining DAPI (5 mg/ml, Bioworld Biotechnology) for 15 min, which can label the nuclei. Confocal fluorescence microscopy (Olympus FV3000, Japan) was used to capture images. All images were captured using the same microscope settings and processed with the same adjustments and parameters. The exported images were loaded into ImageJ (NIH) and were quantified by 2 independent observers blinded to grouping.

Cerebral infarction volume measurement

Following immunofluorescence staining of brain tissue sections with NeuN, ischemic lesions were identified on each section in a blinded manner, and the total infarct volume was determined using ImageJ software. The brain slices were sectioned from Bregma 1.18 mm to Bregma - 0.34 mm with an intermediate interval of about 0.5 mm. The % of infarct volume was calculated using the following formula: % of infarct volume = (contralateral hemisphere area - non-infracted ipsilateral hemisphere area)/contralateral hemisphere area × 100%.

Transmission electron microscopy (TEM)

Targeted fresh tissues were selected to minimize mechanical damage in 3 min. The size of the tissue was about 1 mm³ and the tissue was fixed in an EP tube with fresh TEM fixative (Servicebio, Wuhan, China). After being washed in 0.1 M PB (pH 7.4) for 15 min 3 times, tissues were fixed with 1% OsO4 in 0.1 M PB (pH 7.4) for 2 h at room temperature in the dark. After removing OsO4, the tissues were rinsed in 0.1 M PB (pH 7.4) 3 times and dehydrated in ethanol gradient at room temperature as followed: 30% ethanol for 20 min, 50% ethanol for 20 min, 70% ethanol for 20 min, 80% ethanol for 20 min, 95% ethanol for 20 min, two changes of 100% ethanol for 20 min, finally two changes of acetone for 15 min. Then the tissues were resin penetrated and embedded as followed: Acetone: EMBed 812 = 1:1 for 2–4 h at 37 °C, Acetone: EMBed 812 = 1:2 overnight at 37 °C, pure EMBed 812 for 5–8 h at 37 °C. Pour the pure EMBed 812 into the embedding models and insert the tissues in a 37 °C oven overnight. The embedding models with resin and samples were moved into a 65 °C oven to polymerize for more than 48 h. The resin blocks were cut into 60–80 nm slices on the ultramicrotome (Leica UC7, Germany) and fished out onto the 150 meshes cuprum grids with formvar film, then stained in 2% uranium acetate saturated alcohol solution for 8 min in the dark, rinsed in 70% ethanol for 3 times and rinsed in ultra-pure water for 3 times. Then 2.6% lead citrate avoid CO₂ staining for 8 min and rinsed with ultra-pure water 3 times. After drying with the filer paper, the cuprum grids were put into the grids board and dried overnight at room temperature. The cuprum grids were observed under TEM (HT7800/HT7700, HITACHI, Japan) and taken images. The exported images were loaded into ImageJ and were quantified by one independent observer blinded to grouping.

Magnetic resonance imaging

The magnetic resonance imaging (MRI) scans were conducted on pre-MCAO and post-MCAO mice on a 9.4 T Bruker MR system (BioSpec 94/20 USR, Bruker) with a 440-mT/m gradient, an 86-mm volume transit RF coil, and a single channel surface head coil. Mice were anesthetized using isoflurane inhalation (2.5–3%) and monitored to maintain constant physiological parameters. Tooth bars and ear bars were used to restrain mice on a mouse holder for data acquisition. Diffusion tensor images (DTI) were acquired with spin-echo echo-planar imaging (SE-EPI) sequence with the following parameters: Two b-values (b = 0 and 1000 s/mm²) along with 30 non-collinear directions, δ = 4.1 ms, Δ = 10.3 ms; TR: 1500 ms, TE: 23.27 ms, FOV: 20 mm × 20 mm, matrix: 128 × 128, and 22 adjacent slices of 0.7 mm slice thickness. Imaging data were converted into NIFTI format using MRIcron. Diffusion data were post-processing using the FSL (v.5.0.9) pipeline including corrections for eddy currents and movement artifacts (eddy_correct), rotations of gradient directions according to eddy currents corrections (fdt_rotate_bvecs), brain mask extractions based on b0 images (bet) and FA maps calculations by fitting a diffusion tensor model at each voxel (dtifit). External capsule (EC) and internal capsule (IC) areas were drawn using the itk-SNAP to extract the FA values.

EdU injections and staining

To label proliferating cells, animals were intraperitoneally injected with the 5-Ethynyl-2'-deoxyuridine (EdU, 7 mg/kg, Invitrogen) once a day for 7 consecutive days, beginning at 7 days after MCAO. EdU staining was performed with a Click-iT® EdU Imaging Kit (#10338, Invitrogen) according to the manufacturer’s protocol after being permeabilized with 0.5% Triton X-100 for 20 min. At room temperature, the Click-iT® reaction cocktail, containing 1X Click-iT® reaction buffer 1 ml, CuSO₄ 40 µl, Alexa Fluor® 647 azide 2.5 µl, and Reaction buffer additive 100 µl, was used to incubate the sections for 30 min in the dark. Then the brain slices were washed with PBS and incubated with other primary antibodies as mentioned before.

Myelin isolation and purification

Myelin was isolated from 6-week-old C57BL/6J mice brains by sequential centrifugation on discontinuous sucrose gradient according to a protocol previously described. ²² In brief, brain tissues were homogenized in 0.32 M sucrose solution with a sterile hand-held rotary homogenizer. Then the homogenized brain solution was gently added to the top of the 0.83 M sucrose solution without mixing the two layers and making the volume of 0.32 M to 0.32 M sucrose solution 3:4. The solution was centrifuged at 100,000 × g for 45 min at 4 °C with the minimum acceleration and deceleration speed to achieve the crude myelin debris from the interface of the two sucrose densities. After being homogenized, the crude myelin debris was centrifuged at 100,000 × g for 45 min at 4 °C with the maximum acceleration and deceleration speed. After discarding the supernatant and re-suspending the pellets in Tris·Cl (pH 7.4) buffer solution, third-times centrifugation was given repeatedly to get the pure myelin debris. The aseptic operation should be noticed during the whole process. The purified myelin pellet was sub-packaged in a concentration of 20 mg/mL with PBS and stored at −20°C.

Primary microglia cultures and treatment

Primary microglia were isolated from 1-to-2-day-old C57BL/6J mice brains. The isolated cells from brain cortices were seeded into 75 cm² T-flasks and cultured for 10–13 days in 90% DMEM (Invitrogen, Frederick, MD, USA), 10% fetal bovine serum (FBS, Hyclone, Logan, UT, USA) at 37 °C with a humidified atmosphere of 5% CO₂. After a slight shaking with hands for 8–10 minutes, the loosely adherent microglia were floated and replanted into new plates for the following experiments. To stimulate primary microglia, 1 mg/ml of myelin debris was added to the microglia media, along with R428 (100 nM, Selleck, China), or Gas6 (50 nM), or HX531 (10 μM, Sigma-Aldrich, America) for 6 h, and washed out and collected the microglia on 6 h, 24 h, and 48 h to test the genes of lipid metabolism.

Primary OPCs and oligodendrocyte cultures and treatment

Primary OPCs were isolated from 1-to-2-day-old C57BL/6J mice brains. The isolated cells from brain cortices were plated onto poly-D-lysine-coated plates with culture media OPCs and maintained for 5–7 days in a serum-free OPC proliferation medium (DMEM/F12, 2% B27, 0.5% 100 × Pen/Strep, 0.03% 30 ng/mL hFGF-AA and 0.01% 10 ng/mL bFGF) at 37 °C with a humidified atmosphere of 5% CO₂. For oligodendrocyte induction, OPCs were exchanged medium into OPC differentiation medium containing T3 (30 ng/mL) and CNTF (20 ng/mL) for 2–3 days. To stimulate pre-mature OLs, 1 mg/ml of myelin debris, or Gas6 (50 nM), or 1 × 10⁵/cm² primary microglia (10% of OPCs) were added on the 8^th day of cell culture for 48 h. After exchanging medium, the pre-mature OLs cultured for 2–3 days until they became mature OLs for the following experiments.

Fluorescence-activated cell sorting (FACS) for acute isolation of microglia

Mice were killed to dissect the right hemisphere of the MCAO brains or whole sham brains after sham or MCAO surgery, excluding brain stem and olfactory bulbs, and then cut into pieces with tissue scissors. Tissues were put into a glass homogenizer for homogenizing in DMEM media. Single-cell suspensions were filtered using 70 μm cell filters. After centrifuging and discarding the supernatant, the cell compounds were resuspended in 30% isotonic Percoll (GE Healthcare, Shanghai, China), overlaid with 70% isotonic Percoll to the bottom of the solution, and centrifuged at 2000 rpm at 4 °C for 25 min with slow acceleration and no brake. The microglia-enriched cell population was isolated from the 30–70% interphase and diluted 1:5 in ice-cold PBS, then centrifugated at 1500 rpm at 4 °C for 5 min. The obtained cells were incubated with a mixture of antibodies against CD11b-APC (1:300, eBioscience, #17-0112-82), CD45-FITC (1:300, eBioscience, #56-0451-82) at 4 °C for 30 min in the dark, and CD11b⁺CD45^int cells were sorted by BD FACSAria™ III (San Jose, CA, USA) as microglia for the subsequent experiments.

Total RNA extraction and quantitative real-time PCR analysis

Total RNA was extracted from the cultured cells or cells sorted by FACS using the TRIzol reagent (Invitrogen) according to the manufacturer’s protocol, then reverse-transcribed into cDNA with a PrimeScript RT Reagent Kit (Takara). Quantitative real-time PCR (qPCR) of the cDNA was performed using a Step One Plus PCR system (Applied Biosystems, Foster City, CA, USA) with an SYBR Green Kit (Applied Biosystems). Relative quantified levels were compared using the 2^−ΔΔCT method normalized to the endogenous control GAPDH. The primer sequences were as follows:

• Abca1 Forward CTGTTTCCCCCAACTTCTG

• Reverse TCTGCTCCATCTCTGCTTTC

• Abcg1 Forward TCTTTGATGAGCCCACCAGT

• Reverse GGGCCAGTCCTTTCATCA

• Apoc1 Forward TCCTGTCCTGATTGTGGTCGT

• Reverse CCAAAGTGTTCCCAAACTCCTT

• Apoe Forward CTGACAGGATGCCTAGCCG

• Reverse CCAGCTCCTTTTTGTAAGCCTTT

• Mag Forward CTGCCGCTGTTTTGGATAATGA

• Reverse CATCGGGGAAGTCGAAACGG

• Mbp Forward GACCATCCAAGAAGACCCCAC

• Reverse GCCATAATGGGTAGTTCTCGTGT

• Plp1 Forward CCAGAATGTATGGTGTTCTCCC

• Reverse GGCCCATGAGTTTAAGGACG

• Gapdh Forward AGGTCGGTGTGAACGGATTTG

• Reverse TGTAGACCATGTAGTTGAGGTCA

Phagocytosis assay

Myelin debris was first with DID (Beyotime, China) dye for 15 min, and washed with PBS three times. The primary microglia were dyed DIO (Beyotime, China) for 15 min, then the media were removed. Cells were washed with PBS three times to eliminate free dye in the media. After being treated with R428 (100 nM, Selleck, China), Gas6 (50 nM), or HX531 (10 μM, Sigma-Aldrich, America) for 2 h, the DID-myelin debris (0.1 mg/ml) was added. After stimulation for 30 min, cells were washed three times to remove unphagocytosed floating myelin debris and can be used for follow-up experiments. Use FITC and APC channels of the flow cytometer (BD Accuri® C6) to analyze the uptake of DID-myelin debris by DIO-microglia. DIO-myelin was obtained in the same method.

Statistical analysis

All analysis was performed using GraphPad Prism 8 software. Values were shown as the mean ± standard deviation (SD) for data that were normally distributed or median and interquartile range for data that were not normally distributed for continuous variables. The Shapiro-Wilk test was performed to inspect the normality and homogeneity of variance of all the data. Differences between the 2 groups were compared by Student’s t-test for the data was normally distributed. For multiple comparisons, one-way or two-way analysis of variance (ANOVA) followed by Tukey’s Honestly Significant Difference Test was used. For continuous variables with non-normal distributions, the two-tailed Mann-Whitney U test was used. For all comparisons, p < 0.05 was considered statistically significant.

Results

Gas6 decreased infarct size and behavior deficits in stroke model

To determine whether Gas6 plays a role after MCAO, we first quantified the effects of Gas6 treatment on brain infarct size after stroke by NeuN staining. We found a significant reduction of infarct volume in MCAO + Gas6 mice compared to MCAO + Vehicle mice until 28 days after MCAO (Figure 1(a) and (b)). There were no differences in the survival curves between these two groups (Figure 1(c)). The Morris water maze test was performed to estimate the long-term memory functions post-stroke. Mice treated with Gas6 performed better during the training phase, spending less time finding the platform (Figure 1(d) and (e)). On the probe test day, mice in the MCAO + Gas6 group swam more times across the platform and spent more time in the goal quadrant than mice in the MCAO + Vehicle group (Figure 1(f) and (g)), with no differences in the swimming speed (Figure 1(h)). We also applied the adhesive removal test to estimate the sensory coordination function (Figure 1(i)) and the rotarod test to estimate the motor function (Figure 1(j)) of mice on the 28^th day after MCAO. Mice receiving Gas6 presented better sensory and motor rehabilitation than the MCAO +Vehicle mice group. These data suggested that Gas6 may promote the recovery of memory and sensorimotor functions after MCAO.

Figure 1.

Gas6 decreased infarct size and behavior deficits in stroke model. (a) Representative NeuN-stained coronal sections in MCAO + Vehicle and MCAO + Gas6 mice on the 28^th day after MCAO. The infarct areas are highlighted with white dotted lines. (b) Quantification of infarct volume by NeuN staining (28 days after stroke). N = 7 for each group. (c) The survival curve for the mice in each group after MCAO. (d) The training phase of the Morris water maze test to estimate the spatial learning ability of mice from Day 23 to Day 27 after MCAO. N ≥ 12 for each group. (e) Representative images showed the swim paths in the probe test. (f and g) The probe test of the Morris water maze to estimate the memory of mice by the number of entering the platform (f) and the time spent in goal quadrant (g). N ≥ 12 for each group. (h) The swimming speed of mice in Morris water maze test. N ≥ 12 for each group. (i and j) Adhesive removal test to estimate the sensory coordination function (i) and Rotarod test to estimate the motor function (j) of mice on the 28^th day after MCAO. N ≥ 12 for each group. *p < 0.05, **p < 0.01, ***p < 0.001, and ns not significant. Values were mean ± SD.

Gas6 alleviated white matter injury after ischemic stroke

White matter injury is closely related to the memory and sensorimotor functions of ischemic stroke, so we wondered whether the Gas6 treatment affected the white matter recovery after stroke. We performed the MRI scan of the mice before and on the 28^th day after the MCAO. The fractional anisotropy (FA) values serve as an indicator of the extent of anisotropic diffusion of water molecules within white matter fibers, which reflects white matter integrity.^23,24 We first analyzed the FA value in ipsilateral and contralateral sides of each MCAO mice. The FA values in ipsilateral were significantly lower than that in contralateral sides in internal capsule as well as external capsule areas (Supplemental Figure 1 A&B). Then, we used the FA ratio of the ipsilateral side to the contralateral side to represent the white matter injury, which is considered a robust FA metrics to indicate diminished structural integrity after stroke.^{25 –27} The FA ratio of mice treated with Gas6 was higher than that of control mice in the EC area but not in the IC area, indicating better white matter repair after Gas6 treatment especially in the EC area (Figure 2(a) to (c)). We also used the SMI32, which is a monoclonal antibody that specifically targets non-phosphorylated neurofilaments, to show the damaged neurons/axons. The utilization of the SMI32/MBP protein ratio is a widely accepted approach for assessing demyelination. ²⁸ Whether in the cortex or striatum, the mice receiving Gas6 had the lower SMI32/MBP ratio, meaning the milder white matter damage on the 28^th day after MCAO (Figure 2(d) to (f). Electron microscopy was performed to investigate the myelin sheath of nerve fibers in the striatum on the 28^th-day post-stroke. It can be observed that myelinated axons were more in the MCAO + Gas6 group than that in the MCAO + Vehicle group (Figure 2(g)). The g-Ratio of the MCAO + Gas6 group was obviously decreased in contrast to the MCAO + Vehicle group (p < 0.001), suggesting the thicker myelin sheath (Figure 2(h)). These findings suggested that the Gas6 treatment alleviated white matter injury after ischemic stroke.

Figure 2.

Gas6 alleviated white matter injury after ischemic stroke. (a) Representative DTI axial views of the same mouse brain pre-MCAO and on the 28d post-MCAO. Green arrowheads pointed to EC and yellow arrowheads pointed to IC. (b and c) Quantification of FA ratio of ipsilateral values to the contralateral values in EC (b) and IC (c) area. N = 10 for each group. (d) SMI32 (red) and MBP (green) immunofluorescence staining in the striatum and cortex on the 28d post-MCAO. Scale bars = 100 μm. (e and f) The ratios of SMI32 to MBP staining intensity in the infarcted striatum and cortex. N ≥ 6 for each group. (g) The representative electron microscopy pictures and the scatterplots of g-Ratio (h) on the 28d post-MCAO. Scale bar = 2 μm. *p < 0.05, **p < 0.01, ***p < 0.001, and ns not significant. Values were mean ± SD.

Gas6 potentiated OPC maturation closely related to microglia

To further study whether Gas6 promoted OPC maturation after stroke, we labeled newborn cells by injecting EdU from the 7th day to the 13th day after MCAO. EdU is a detectable thymidine analogue that can be injected into mice and bind to DNA during replication to act as a marker to identify newborn cells. ²⁹ The adenomatous polyposis coli (APC) tumor suppressor protein, anti-APC (CC-1) antibody, is used to label mature myelinating oligodendrocytes. The number of EdU-positive cells in the two groups made no differences in the striatum area (Figure 3(a) and (b)), while the number of EdU and APC double-positive cells was profoundly increased in the MCAO + Gas6 group (Figure 3(a) and (c)). We also used Olig2 and Pdgfrα to label the whole oligodendrocyte cell line and pre-oligodendrocytes (pre-OLs) respectively. Quantitative analysis showed that the number of EdU and Olig2 double-positive cells increased in MCAO + Gas6 group (Supplemental Figure 2 A&B). However, the number of EdU⁺Olig2⁺Pdgfrα⁺ cells in the two groups made no differences in the striatum area (Supplemental Figure 2 A&C), which indicated that Gas6 promoted the maturation of oligodendrocytes after stroke. Next, we investigated the direct effect of Gas6 on OPCs by incubating primary cultured OPCs with myelin debris and Gas6 separately or together for 48 h. We detected the expression of the pre-OL gene (plp1) or mature OL genes (mbp and mag). The results of qPCR reflected that myelin debris stimulation reduced the expression of pre-OLs and mature OLs genes, while Gas6 treatment didn’t increase these genes’ expression, indicating that Gas6 might promote white matter repair not directly through OPCs (Figure 3(d) to (f)). Given the note that Axl is the highest affinity receptor of Gas6, which is mainly expressed in microglia, we further added microglia in OPCs and co-cultured them for 48 h with or without myelin debris and Gas6 stimulation. Then we stained the MBP and Olig2 to show the maturity situation of OPCs (Figure 3(g) to (i)). In line with the result of qPCR, Olig2⁺MBP⁺ cells and MBP fluorescence intensity decreased under the myelin debris treatment. The microglia partially increased the MBP fluorescence intensity compared to the myelin debris group. However, the Gas6 + microglia + myelin debris group demonstrated more Olig2⁺MBP⁺ cells and higher MBP fluorescence intensity compared to the myelin debris processing alone group (Figure 3(h) and (i)). These data suggested that Gas6 stimulation in this co-culture system can promote OPCs maturation and rescue the injury caused by myelin debris, which was closely related to microglia.

Figure 3.

Gas6 potentiated OPC maturation closely related to microglia. (a) Immunofluorescence staining of EdU (green) and APC (red) on the striatum area on the 28d post-MCAO on the ipsilateral side. Scale bar = 50 μm. The number of EdU single-positive cells (b), as well as EdU and APC double-positive cells per field on the ipsilateral side (c). N = 4 per group. *p < 0.05, **p < 0.01, ***p < 0.001, and ns not significant. (d-f) The qPCR results of plp1, mbp, and mag for primary microglia under the stimulation with myelin debris and Gas6 separately or together, and the control group was treated with PBS. N = 4 repeats. *p < 0.05, and ns not significant. (g) The number of Olig2 and MBP double-positive cells per field. N = 4. (h) The relative mean fluorescence intensity of MBP in each treatment group. N = 4. (i) The representative images of Olig2 (green), MBP (red), and DAPI (blue) of the primary OPCs under the different treatments. Scale bar = 20 μm. All values were mean ± SD, **p < 0.01, and ***p < 0.001 compared to the negative control group. ^##p < 0.01, and ^###p < 0.001 compared to myelin debris group.

Gas6 promoted microglia to engulf myelin debris

Then, we detected the capability of clearing myelin debris after Gas6 stimulation in primary cultured microglia. We observed that microglia were round and filled with myelin fragments when cultured with Gas6 + Myelin for 24 hours, while microglia presented spindle and contained fewer myelin fragments in the myelin group under the microscope (Figure 4(a)). Then, we labeled microglia with CD68 and DID-Myelin which showed that primary microglia did phagocytize myelin debris in our Supplemental Figure 3 A. The flow cytometry analysis was applied to measure DID-signals inside microglia and indicated that the primary microglia engulfed more myelin debris after the Gas6 treatment compared with the control group (Figure 4(b) and (c)). We used R428, an Axl selective inhibitor, to treat microglia and detect the phagocytosis capability. Flow cytometry analysis showed that R428 suppressed the microglial phagocytosis of myelin debris and completely abolished the potential of Gas6 on phagocytosis (Figure 4(b) and (c)). We then used the DIO-myelin debris to stimulate OPCs with or without microglia and Gas6 treatment. We found that the DIO-myelin debris spread in cultured OPCs diffusely when there were no microglia existed. However, in the Gas6 + Myelin + microglia group, we observed microglia did swallow the myelin debris showing more Iba1-positive cells were in contact with myelin debris (Figure 4(d) and (e)). To verify this observation in MCAO models, the immunofluorescence staining of TMEM119 and dMBP was applied to label microglia and myelin debris respectively. The results showed that more microglia in the MCAO + Gas6 mice wrapped degraded myelin debris in three-dimensional reconstruction on the 28^th day after stroke in the ipsilateral striatum (Figure 4(f) and (g)).

Figure 4.

Gas6 promoted microglia to engulf myelin debris. (a) The light microscopy images of the myelin (control group) and Gas6 + myelin (Gas6 group) treated microglia for 24h. Scale bar = 20 μm. (b) The relative mean fluorescence intensity of DID-myelin swallowed by microglia was detected by FACS. N = 5 repeats. **p < 0.01, and ***p < 0.001 compared to the control group. (c) The representative flow cytometric images for the control group, Gas6 group, R428 group, and Gas6 + R428 group. (d) The number of microglia that are positive with myelin debris per mm² in microglia/OPCs/myelin debris cocultured system with or without Gas6. N = 5. **p < 0.01. (e) The representative immunofluorescence images of MBP (red), Iba1 (gray), DAPI (blue), and myelin debris (green) in different treatments. Scale bar = 20 μm. (f) Number of dMBP positive microglia per field. (g) Representative images of dMBP (red), TMEM119 (green), and DAPI (blue) immunostaining and the three-dimension reconstruction of microglia engulfed myelin debris on the 28^th Day after MCAO. Scale bar = 20 μm. All values were mean ± SD. For D and F **p < 0.01.

Gas6 promoted cholesterol transportation protein expression in microglia

The efflux of cholesterol from swallowed cholesterol-rich myelin debris is critical for oligodendrocytes to synthesize new myelin ¹⁶ . Thus, we speculated that the Gas6/Axl signaling pathway may affect cholesterol transportation in the process of myelin phagocytosis. In vitro, we incubated microglia with myelin debris, R428, Gas6, and R428 + Gas6 for 6 hours. Then, the medium was changed, and we detected cholesterol transportation genes at 6 h, 24 h, and 48 h by qPCR. Generally, the expressions of abca1, abcg1, apoc1, and apoe all appeared a relatively similar trend. The expression of these genes increased under Gas6 administration and decreased under R428 treatment (Figure 5(a) to (d)). However, Gas6 could not provoke these genes^, expression when microglia were pretreated with R428 for 2 h. The changes of these genes, especially abca1 and abcg1 that mediate cholesterol efflux were more obvious at the 6 h and 24 h (Figure 5(a) to (d)). To verify the cholesterol transportation protein expression of microglia in vivo, we sorted out microglia of the mice’s cerebral ipsilateral hemisphere in the sham group, MCAO + Vehicle group, and MCAO + Gas6 group on the 28^th day after MCAO. There were no differences in the number of microglia among these groups (Figure 5(e) and (f)). Consistent with the results in vitro, the expression of abca1, abcg1, apoc1, and apoe increased after MCAO surgery, and the expression in the MCAO + Gas6 group was even higher than that in the MCAO + Vehicle group (Figure 5(g) to (j)).

Figure 5.

Gas6 promoted cholesterol transportation protein expression in microglia. (a-d) The qPCR results of abca1, abcg1, apoc1, and apoe for different treatments at different time points. N = 4 repeats. At each time point, *p < 0.05, **p < 0.01, and ***p < 0.001 compared to the all negative group, ^#p < 0.05, ^##p < 0.01, and ^###p < 0.001 compared to the myelin debris group. ^&p < 0.05, ^&&p < 0.01, and ^&&&p < 0.001 compared to myelin debris + Gas6 group. (e) Gating strategy of the CD45^midCD11b^high microglia and the percentage of microglia in all single cell suspensions. (f) The number of microglia per ipsilateral hemisphere in sham, MCAO + Vehicle, and MCAO + Gas6 groups. N = 3. (g-j) The ipsilateral microglial mRNA expression of abca1, abcg1, apoc1, and apoe in sham, MCAO + Vehicle, and MCAO + Gas6 group by qPCR. N = 3. *p < 0.05, **p < 0.01, and ns for not significant. All values were mean ± SD.

Gas6 alleviated myelin debris and lipid accumulation in vivo

Furthermore, we used Fluoromyelin and dMBP to label myelin sheaths and myelin debris respectively in stroke mice. We found that the fluorescence intensity of dMBP decreased significantly in the striatum of mice with Gas6 administration (Figure 6(a) and (b)), and the fluorescence intensity of Fluoromyelin increased coherently, suggesting that MCAO + Gas6 mice manifested less myelin debris and better myelin sheaths during the recovery period of stroke (Figure 6(a) to (c)). Oil red O staining was used to display the lipid accumulation in the brain. It was found that the Gas6 administration group accumulated fewer lipid droplets four weeks after stroke, which further suggested that the Gas6/Axl pathway may regulate the lipid metabolism after phagocytosis of myelin debris (Figure 6(d) and (e)). These results suggested that the Gas6/Axl signaling pathway may improve white matter damage after stroke, which was related to microglial phagocytosis and lipid metabolism.

Figure 6.

Gas6 alleviated myelin debris and lipid accumulation in vivo. (a) Fluoromyelin (green) and dMBP (red) immunofluorescence staining of the striatum on the 28d post-MCAO. Scale bars = 50 μm. (b and c) The relative mean fluorescence intensity of Fluoromyelin and dMBP in the MCAO + Vehicle and MCAO + Gas6 group. N = 6. (d) The quantitative analysis of Oil Red O staining. N = 6. (e) The representative images of Oil Red O staining on the 28d post-MCAO. Scale bar = 20 μm. All values were mean ± SD, *p < 0.05, **p < 0.01, and ***p < 0.001.

Gas6 accelerated myelin debris clearance and cholesterol transportation protein expression through the RXR pathway

RXR represents a key nuclear receptor responsible for phagocytic functions and cholesterol efflux through forming a heterodimer with LXR.^30,31 Therefore, we used HX531 (the RXR inhibitor) to explore whether RXR participated in the functions of Gas6. The flow cytometry analysis showed that HX531 restrained microglia from engulfing myelin debris and completely blocked the potential of Gas6 on phagocytosis (Figure 7(a)). The qPCR results of different time points after adding myelin debris into microglia with Gas6, HX531, or HX531 + Gas6 stimulation reflected that abca1, abcg1, apoc1, and apoe were significantly repressed under the HX531 treatment. Meanwhile, Gas6 could not enhance the expression of abca1, abcg1, apoc1, and apoe with the presence of HX531 (Figure 7(b) to (e)). These results partly indicated that Gas6 accelerated myelin debris clearance and cholesterol transportation protein expression through the RXR pathway.

Figure 7.

Gas6 accelerated myelin debris clearance and cholesterol transportation protein expression through the RXR pathway. (a) The relative mean fluorescence intensity of DID-myelin swallowed by microglia after HX531 with or without Gas6 stimulation in cultured microglia. N = 5. Values were mean ± SD, **p < 0.01, ***p < 0.001, and ns for not significant. The representative flow cytometric images for these groups. (b-e) The qPCR results of abca1, abcg1, apoc1, and apoe for myelin, HX531 + myelin, Gas6 + myelin, and HX531 + Gas6 + myelin treatment at different time points. N = 3 repeats. Values were mean ± SD. At each time point, *p < 0.05, **p < 0.01, ***p < 0.001, and ns for not significant compared to the myelin group, ^#p < 0.05, and ^###p < 0.001 compared to Gas6 + myelin treatment group.

Discussions

To summarize our findings, Gas6 reduced brain tissue loss and behavior deficits 28 days after MCAO. Gas6 promoted white matter integrity by promoting OPC maturation. Gas6 accelerated microglia-mediated myelin debris clearance and cholesterol transportation protein expression in vivo and in vitro, accordingly less lipid and myelin debris were deposited in Gas6-treated stroke mice. RXR pathway could be one critical target of Gas6 in speeding up myelin debris clearance and cholesterol transportation.

Gas6 has been studied in cerebrovascular disease recently. Intranasal administration of Gas6 could reduce brain infarct volume and hemorrhage transformations 3 days after MCAO. ¹⁴ Gas6 treatment repressed acute inflammatory cytokines including IL-1β, IL-6, and TNF-α after ischemic stroke through inhibiting Axl/Toll-like receptor (TLR)/Tumor necrosis factor receptor-associated factor (TRAF)/NF-kB signaling pathways. ⁹ In subarachnoid hemorrhage, Gas6 promoted microglial engulfment of apoptotic cells and alleviated inflammation, which is partly dependent on Axl and Rac1 activation. ³² In our study, we treated MCAO mice with intranasal Gas6 once daily until 28 days after the stroke. We once again confirmed the effectiveness of Gas6 in ischemic stroke. In this study, we extended the efficacy and safety to the rehabilitation stage of stroke and found that Gas6 could improve cognitive and sensorimotor functions 28 days after ischemic stroke. To be summarized, Gas6 is a promising drug for stroke treatment and has a promising clinical application in the future.

Myelin debris retard oligodendrocyte lineage cells to proliferate and mature. Therefore, myelin debris is an endogenous remyelination inhibitor if it is not adequately cleared by phagocytes. ³³ Sufficient uptake of myelin debris by microglia could be protective in ischemic stroke. ³⁴ In our in vivo study, we observed that Gas6 increased newborn mature oligodendrocytes in the infarcted area. In our primary culture oligodendrocytes, we found that myelin debris directly inhibited the mature oligodendrocyte markers (mbp, mag, and plp1). However, in the absence of microglia coculture, Gas6 demonstrated incompetence to rescue the OPC abnormal maturation mediated by myelin debris. Multiple studies suggested that the protective effect of Gas6 on white matter largely depends on its immune regulation of glial cells.^12,35 As a result, our research on the protective mechanism of Gas6 mainly focuses on microglia.

Professional phagocytes in the central nervous system mainly include microglia and macrophages. ⁴ Phagocytosis of excessive cholesterol-rich myelin debris may cause intracellular cholesterol overload. To keep cholesterol hemostasis, supererogatory cholesterol flows out through ABC transporters, especially ATP-binding cassette transporter A1 (ABCA1) and ATP-binding cassette transporter G1 (ABCG1). Cholesterol is insoluble in water, whose transportation requires apolipoproteins including apolipoprotein E (APOE) and apolipoprotein C1 (APOC1).^20,36 Myelin clearance by phagocytes might lead to a compensatory increase in the expression of lipid transporters through cholesterol derivates, which could be beneficial for cholesterol recycling and remyelination.^16,37 If the phagocyte is not accompanied by a corresponding increase in cholesterol transporters, lipid droplet aggregation will result in the formation of foam cells and an inflammatory response.^38,39 Recent studies demonstrated that dysfunction of phagocytic receptors may also lead to abnormal cholesterol transportation. For instance, triggering receptor expressed on myeloid cells 2 (TREM2, a phagocytic receptor) deficient microglia fails to clear myelin cholesterol, leading to cholesteryl ester accumulation. ⁴⁰ Gas6 and Axl double-knockout mice presented pronounced pathological lipid droplets deposit and prolonged inflammation following cuprizone exposure. ¹⁰ In this regard, we tested both myelin clearance and cholesterol transportation protein levels after the Gas6 treatment. We observed that Gas6 provoked the engulfment of myelin debris and the expression of abca1, abcg1, apoc1, and apoe from the data of the cultured microglia and ipsilateral microglia from the infarcted area in the brain. R428, the Axl inhibitor, or HX531, the RXR inhibitor, obligated this biological effect of Gas6, indicating the necessity of the Axl receptor and RXR pathway for myelin clearance and lipid transport protein expression. RXR forms a heterodimer with LXR and interacts with sequence-specific DNA elements, which directly upregulates the transcription of lipid transport genes. ¹⁷ RXR represents a key pathway for phagocytic functions and brain recovery after stroke. ³⁰ In this study, we did not apply HX531 in vivo to further verify the action of Gas6, for HX531 has a high fatality rate (data did not show) and lacks evidence to cross the blood-brain barrier by peripheral injection. Nevertheless, we surmise that Gas6 promotes myelin clearance at least partially through Axl receptors. Subsequently, cholesterol metabolic derivates from myelin debris can stimulate RXR/LXR transcriptional activation, which promoted cholesterol transport and remyelination.

This research has some limitations. Initially, in order to minimize the number of experimental study variables, we solely examined the protective impact of Gas6 on white matter injury in male MCAO mice. However, the potential therapeutic efficacy of Gas6 on female MCAO mice remains unexplored. Multiple studies addressed the molecular sexual difference after ischemic stroke.^41,42 For instance, compared with male-ε4 carriers, female ApoE ε4 carriers exhibited a precipitous age-related cerebral perfusion decline and expressed lower neuroinflammatory related genes.^43,44 Besides, like other studies, we used young mice to make MCAO model which ignored the effects of aging and comorbidities.^45,46 Additionally, although our data strongly suggested the protective role of Gas6 in improving post-stroke white matter injury, we believed that Gas6 promoted white matter repair in multiple ways regarding the mechanisms, including but not limited to microglia-mediated myelin debris clearance and cholesterol transportation protein expression. For one thing, Axl is expressed in several cells, including astrocytes, ⁴⁷ neurons, and microglia. ⁹ Although, microglia showed higher Axl expression after stroke. ⁹ Nevertheless, we could not rule out the effect of Gas6 on other cells covalently affecting the prognosis of stroke. For another, in addition to Axl, the TAM receptor family also includes Tyro3, which is expressed on oligodendrocytes and plays an important role in the development of myelin sheath.^48,49 Although Gas6, as a ligand, has the strongest affinity for Axl, it may still act on Tyro3 and play a direct role in white matter tracts. Furthermore, the RXR inhibitor was non-cell specificity in vivo and the animal experiments were impeded by high mortality rate. We expected further works in Axl or RXR-conditioned microglia-knockout mice to better illustrate these questions.

Our results clearly suggest that Gas6 plays a critical role in myelin engulfment and cholesterol processing in microglia, which could be a promising medication for clinical application in white matter injury of ischemic stroke.