Pharmacological Activation of RXR-α Promotes Hematoma Absorption via a PPAR-γ-dependent Pathway After Intracerebral Hemorrhage

Abstract

Endogenously eliminating the hematoma is a favorable strategy in addressing intracerebral hemorrhage (ICH). This study sought to determine the role of retinoid X receptor-α (RXR-α) in the context of hematoma absorption after ICH. Our results showed that pharmacologically activating RXR-α with bexarotene significantly accelerated hematoma clearance and alleviated neurological dysfunction after ICH. RXR-α was expressed in microglia/macrophages, neurons, and astrocytes. Mechanistically, bexarotene promoted the nuclear translocation of RXR-α and PPAR-γ, as well as reducing neuroinflammation by modulating microglia/macrophage reprograming from the M1 into the M2 phenotype. Furthermore, all the beneficial effects of RXR-α in ICH were reversed by the PPAR-γ inhibitor GW9662. In conclusion, the pharmacological activation of RXR-α confers robust neuroprotection against ICH by accelerating hematoma clearance and repolarizing microglia/macrophages towards the M2 phenotype through PPAR-γ-related mechanisms. Our data support the notion that RXR-α might be a promising therapeutic target for ICH.

Introduction

Intracerebral hemorrhage (ICH) is one of the most fatal cerebrovascular diseases and is linked to high morbidity and mortality rates [1, 2]. Primary brain injury caused by ICH usually occurs within the first few hours after onset and the hematoma mechanically damages adjacent tissue. Subsequently, hematoma-induced secondary brain injuries, including inflammatory responses, microglial activation, oxidative stress, and neuronal apoptosis and necrosis, occurring after the primary injury can result in severe neurological deficits and death [1, 3]. A growing body of studies indicates that hematoma clearance may be a unique target in ICH treatment [4]. Although early removal of the hematoma using surgical procedures yields promising results, it has unsatisfactory effects on neurological recovery and fails to address the secondary brain damage [5–7]. Therefore, new therapeutic strategies focusing on endogenous hematoma clearance for ICH are urgently needed.

In the central nervous system (CNS), microglia/macrophages serve as the main phagocytes involved in the defense against brain damage, including ICH, through the phagocytosis of red blood cells and the removal of tissue debris [4, 8, 9]. Microglia/macrophages can be activated by hematoma components such as thrombin and neurotoxins after tissue damage occurs [10]. It has been found that microglia/macrophages in the brain can be dynamically activated into two polarized states, termed the classical M1-like phenotype and the alternative M2-like phenotype [11]. Alternatively activated M2-like microglia/macrophages of the brain are considered to be anti-inflammatory and are generally involved in phagocytosis and tissue repair [3]. The evidence above suggests that promoting the resolution of inflammation and increasing phagocytic activity by inhibiting the M1 phenotype and promoting the M2 phenotype may be an effective therapeutic strategy for ICH.

Retinoid X receptors (RXRs) are members of the nuclear receptor superfamily [12, 13]. Currently, there are three RXR isoforms (α, β, and γ). RXR-α is strongly expressed in all human and rodent macrophage-type cells (brain microglia, liver Kupffer cells, and bone osteoclasts). Once activated, RXR-α is translocated into the nucleus and regulates the transcription of target genes, and is involved in multiple cellular processes, such as monocyte/macrophage differentiation, phagocytosis, and metabolism [13–16]. Current studies have reported that activation of RXR-α, which forms a heterodimer with peroxisome proliferator-activated receptor-γ (PPAR-γ), may be a promising treatment for Alzheimer’s disease [16, 17]. Notably, a recent study has demonstrated that the activation of RXR-α promotes the nuclear accumulation and transcriptional activity of PPAR-γ, therefore modulating microglia polarization in traumatic brain injury [18]. However, the role of RXR-α in the pathological processes following ICH is not yet fully understood.

Based on the evidence above, we aimed to determine the role of RXR-α in modulating the polarization of microglia/macrophages and promoting hematoma clearance and neurological function via the PPAR-γ pathway.

Materials and Methods

Animals

All procedures were approved by the Ethics Committee of Zhejiang University and followed the National Institutes of Health guidelines for the Care and Use of Laboratory Animals. Adult male C57BL/6J mice (aged 8 weeks–10 weeks, weight 22 g–25 g, n = 409) were obtained from the SLAC Laboratory Animal Co., Ltd. (Shanghai, China). The mice were kept in a humidity-controlled room (25°C ± 1°C, 12-h light/dark cycle) and were raised with free access to food and water.

ICH Model

The autologous blood injection model of ICH was created as previously described [19]. Briefly, each mouse was anesthetized and maintained under 1% pentobarbital sodium, after which 25 μL of autologous blood was injected 2.5 mm to the right and 3 mm below bregma at a 5° angle toward the midline. Mice in the sham group underwent the same procedure (anesthesia and needle insertion), except for injection.

Experimental Design

All mice were randomly assigned to the experiments described below (Fig. 1). A total of 409 mice (including those that died) were used in this study.

Fig. 1.

Experimental design and animal groups. Part of the figure was created using BioRender (www.biorender.com).

Experiment I

Bexarotene is a highly-selective RXR-α agonist approved by the The United States Food and Drug Administration as an antineoplastic agent for the treatment of cutaneous T-cell lymphoma, with high blood-brain barrier permeability and a good safety profile [20]. We used bexarotene to assess the effects of RXR-α activation on hematoma clearance and neurological function after ICH. Neurological function was assessed by the cylinder test, corner turn test, and forelimb placement test after ICH. Mice were assigned into three groups: sham, ICH + vehicle [10% dimethyl sulfoxide (DMSO) in saline], and ICH + bexarotene (5 mg/kg) (n = 9 mice per group). The sham group received the same volume of vehicle intraperitoneally at the same time points after ICH induction. A T2* weighted magnetic resonance imaging (MRI) scan was used to measure the hematoma volume at 1, 3, 7, 14, and 28 days after ICH. Mice were randomly divided into two groups: ICH + vehicle (10% DMSO in saline), and ICH + bexarotene (5 mg/kg) (n = 6 mice per group).

Experiment II

To assess the expression patterns of RXR-α and PPAR-γ after ICH, mice were assigned to five groups: sham; ICH 1 day; ICH 3 days; ICH 7 days; and ICH 14 day (n = 6 mice per group). Whole-cell lysates Western blots (n = 6 mice per group) and cytoplasmic and nuclear protein Western blots (n = 6 mice per group) were performed at different time-points. Next, the cellular location of RXR-α was assessed by double immunofluorescence staining in the ICH 3 days group. To assess the nuclear translocation of RXR-α, mice were divided into three groups: sham; ICH + vehicle (10% DMSO in saline); and ICH + bexarotene (n = 6 mice per group). Whole-cell lysate Western blots and cytoplasmic and nuclear protein Western blots in these groups were performed at 3 days after ICH. To determine whether PPAR-γ plays a role in RXR-α activation after ICH, we used the RXR-α agonist bexarotene and the PPAR-γ antagonist GW9662. To assess the nuclear translocation of PPAR-γ, mice were divided into four groups: sham; ICH + vehicle (10% DMSO in saline); ICH + bexarotene + vehicle; and ICH + bexarotene + GW9662 (4 mg/kg) (n = 6 mice per group). Whole-cell lysate Western blots and cytoplasmic and nuclear protein Western blots in these groups were performed at 3 days after ICH.

Experiment III

To investigate the mechanism underlying the role of RXR-α activation in ICH, mice were randomly assigned to four groups: sham; ICH + vehicle (10% DMSO in saline); ICH + bexarotene + vehicle (10% DMSO in saline); and ICH + bexarotene (5 mg/kg) + GW9662 (4 mg/kg). Immunofluorescence staining (n = 4 mice per group), Western blots (n = 6 mice per group), and enzyme-linked immunosorbent assays (ELISAs) (n = 6 mice per group) were used to assess the polarization of microglia/macrophages and PPAR-γ transcription activity. To determine whether microglia play a major role, mice were divided into four groups: ICH + Clophosome (5 µL) + vehicle (10% DMSO in saline); ICH + Clophosome (5 µL) + bexarotene (5 mg/kg); ICH + control liposome (5 µL) + vehicle (10% DMSO in saline); and ICH + control liposome (5 µL) + bexarotene (5 mg/kg) (n = 6 mice per group). Before that, a total of 6 mice per group were used to determine the efficiency of microglia/macrophage depletion [ICH + control liposome (5 µL); ICH + Clophosome (5 µL)].

Experiment IV

Next, we assessed whether PPAR-γ plays a role in the protective effects linked to RXR-α activation. To this end, the cylinder test, corner turn test, and forelimb placement test were used after ICH. Mice were randomly divided into three groups: ICH + vehicle (10% DMSO in saline); ICH + bexarotene + vehicle (5 mg/kg); and ICH + bexarotene (5 mg/kg) + GW9662 (4 mg/kg) (n = 6 mice per group). The sham group received the same volume of vehicle intraperitoneally at the same time points after ICH induction. An MRI scan was used to measure the hematoma volume at days 1, 3, 7, and 14 after ICH. Mice were randomly divided into three groups: ICH + vehicle (10% DMSO in saline); ICH + bexarotene + vehicle (5 mg/kg); and ICH + bexarotene (5 mg/kg) + GW9662 (4 mg/kg) (n = 6 mice per group).

Drug Administration

Bexarotene (MedChem Express, Monmouth Junction, NJ, USA) was dissolved in 10% DMSO as previously described [21]. The selective PPAR-γ antagonist GW9662 (MedChem Express) was diluted in 10% DMSO. Bexarotene (5 mg/kg) or an equal volume of vehicle, or bexarotene (5 mg/kg) + GW9662 (4 mg/kg) was first administered intraperitoneally 1 h after ICH, followed by daily injections until sacrifice. To deplete microglia and macrophages, mice were injected with 25 μL of blood mixed with 5 μL of the anionic forms of Clophosome or control liposomes (FormuMax Inc., Sunnyvale, CA, USA). The dosage and time points of bexarotene, GW9662, and Clophosome were based on previous studies [18, 22, 23].

Calculation of Hematoma Volume

Mice were anesthetized with 1% pentobarbital sodium for MRI scanning. An MRI was performed on days 1, 3, 7, 14, and 28 after ICH in a 3.0-T MRI scanner. The MRI included a T2* sequence. The scanning parameters for T2* weighted imaging were: TR/TE = 2200/103.8 ms, number of averages = 10, acquisition matrix = 208 × 208, voxel size = 0.12 × 0.12 × 1 mm³, flip angle = 130°, slices = 5. MRI image datasets were obtained in DICOM (Digital Imaging and Communications in Medicine) format. The data were transformed into the NIfTI (.nii) format and then assessed with 3D Slicer. The 3D-Slicer method is one of the software methods serving to measure the volume of a hematoma (http://www.slicer.org/). Hematomas were manually identified pixel by pixel in each slice. Next, a 3D model was established and the hematoma volume was calculated by summing the volumes of the pixels. Alternatively, hematoma hemoglobin content was used to quantify the hematoma volume [24]. Four 1-mm coronal slices in the area of bleeding were homogenized and extracted ultrasonically with 300 μL distilled water. Different volumes of autologous blood (0 μL, 1.0 μL, 2.0 μL, 4.0 μL, and 8.0 μL) were added to 300 μL of normal brain tissue lysate to generate a standard curve. After centrifugation at 12,000g for 30 min, the supernatant was collected and 80 μL of Drabkin’s reagent (Sigma-Aldrich, St Louis, MO, USA) was added to 20 μL of supernatant and incubated in a 96-well plate at room temperature for 15 min. The absorbance of the solution was measured at a wavelength of 540 nm.

Behavioral Tests

Neurobehavioral functions were evaluated by a forelimb placing test, a forelimb use asymmetry (cylinder) test, and a corner turn test, as previously reported [20]. Baseline data were recorded to reduce variability and identify the preferred side. The neurological scores were evaluated by a blinded observer.

Forelimb placement can be assessed by stimulating a mouse's vibrissae to trigger a response. To test the function of the forelimbs, the researchers held the animal's torso, allowing the forelimbs to hang freely, while brushing its vibrissae on the corner edge of a table. Intact animals usually responded by placing the forelimb on the table on the same side, while ICH mice had impaired paw placement. This test was scored by counting the percentage of placements.

In the forelimb-use asymmetry test (cylinder test), the mouse was placed in a transparent cylinder and the numbers of independent wall contacts were counted. The behavioral score was recorded as the number of contacts of the ipsilateral (unimpaired) forelimb (I), contralateral (impaired) forelimb (C), and both forelimbs (B). A single overall limb use asymmetry score was calculated as follows: score = [I/(I + C + B)] − [C/(I + C + B)].

In the corner test, a mouse was placed between boards facing a 30° corner. As the mouse approached the corner, the vibrissae on both sides were simultaneously stimulated causing the animal to rear and turn 180°; the number of right turns was counted.

Immunofluorescence Double Labeling

Coronal sections were blocked with 5% bovine serum albumin and 0.3% Triton X-100 and then incubated with primary antibodies overnight at 4°C. The primary antibodies used were mouse anti–NeuN (1:500, ab-104224, Abcam, Cambridge, UK), goat anti-Iba-1 (1:500, ab-5076, Abcam), mouse anti–GFAP protein (1:500, ab10062, Abcam), rabbit anti-RXR-alpha (1:250, ab125001, Abcam), rabbit anti-Arg1 (1:500, 16001-1-AP, Proteintech, Hubei, China), rabbit anti-nitric oxide synthase (iNOS) (18985-1-AP,1:500, Proteintech), and rabbit anti-PPAR-γ (1:250, ab178860, Abcam). The sections were incubated with secondary antibodies at room temperature for 2 h. Finally, the sections were examined and analyzed using a fluorescence microscope (Olympus, Tokyo, Japan). Photomicrographs were saved and merged using Image-Pro Plus software. To assess nuclear fluorescence ratio, ImageJ software was used to determine the nuclear fluorescence and nuclear + cytoplasmic fluorescence as previously reported [25]. Briefly, the integrated density measurement of pixel numbers was made on a region of interest consisting of the total nucleus and the whole cell (nucleus and cytoplasm). In each group, 6 mice were evaluated, and each sample had 3 brain sections. Three random fields of vision in each section were examined to acquire cells (48–60 cells per group were counted).

Enzyme-linked Immunosorbent Assay

Brain samples were homogenized in chilled lysis buffer containing protease and phosphatase inhibitor cocktails (P1005, Beyotime, Shanghai, China). An ELISA kit for mouse TNF-α (EK0527; Boster, Wuhan, China) was used to assess the levels of TNF-α in the brain. The total protein content of each sample was determined by bicinchoninic acid (BCA) assay (ThermoFisher, Waltham, MA, USA). An equal amount of protein brain homogenate was diluted 1:10 with the sample diluent provided with the kit, and all procedures were performed per the manufacturer’s instructions. The PPAR-γ activity was measured using a PPAR-γ Transcription Factor Assay (ab133101, Abcam). The protein concentrations were equal. The ELISA was performed following the manufacturer’s instructions.

Western Blot Analysis

Western blot analysis was performed as previously described [26]. Briefly, the basal ganglia were homogenized and centrifuged for 15 min (13,000 g, 4°C). For the whole-cell lysates, tissue proteins from the basal ganglia were lysed using RIPA lysis buffer. The cytoplasmic and nuclear proteins were extracted using a Cytoplasmic and Nuclear Protein Extraction Kit (P0027, Beyotime). Proteins were assessed using a BCA Protein Assay Kit (Thermo Fisher Scientific). An equal amount of protein (40 μg) was suspended in loading buffer (denatured at 95°C for 5 min), loaded on SDS-PAGE, and transferred to nitrocellulose membranes. Next, the membranes were blocked with non-fat dry milk buffer for 1 h and incubated overnight with the primary antibody. The membranes were then incubated with the secondary antibody for 1 h at room temperature. Bands were visualized using the ECL Plus chemiluminescence reagent kit (Amersham Bioscience, Arlington Heights, IL). The band densities were quantified using ImageJ software. The primary antibodies used in this study were rabbit anti-PPAR-γ (1:1000, ab45036, Abcam), rabbit anti-RXR-α (1:1000, ab125001, Abcam), rabbit anti-Arg1 (1:1000, 16001-1-AP, Proteintech), mouse anti-GAPDH (1:5000, ab8245, Abcam), mouse anti-β actin (1:5000, ab8227, Abcam), and rabbit anti-histone H3 (1:2000, #9715, Cell Signaling Technology, MA, USA).

Statistical Analysis

All data are expressed as the mean ± standard error of the mean (SEM). For data that met a normal distribution and homogeneity of variance, differences among groups were analyzed using one-way analysis of variance followed by a Tukey’s multiple comparison test. The Kruskal-Wallis test with Bonferroni correction was used for non-normally distributed data. All statistical analyses were performed in SPSS (version 22.0). A P value <0.05 was considered statistically significant.

Results

Animal Mortality Rate

A total of 409 mice were used, of which 80 underwent a sham procedure and 329 underwent ICH induction. None of the sham mice died, and the mortality rate in the ICH group was 9.7% (32/329). There were no significant differences in mortality rate across the model groups.

Pharmacological Activation of RXR-α Alleviates Neurological Dysfunction and Promotes Hematoma Clearance After ICH

We introduced Bexarotene, an FDA-approved antineoplastic agent that functions by selectively activating RXR-α, to determine the effects of RXR-α under ICH conditions. And we used three behavioral tests (the cylinder, corner-turn, and forelimb placement tests) to assess the effects of RXR-α on behavioral recovery at days 1, 3, 7, 14, and 28 after ICH. There were no significant differences in the behavioral tests scores among the four groups at baseline (pre-ICH). Mice in the model groups suffered from a serious neurological deficit at 24 h after ICH when compared with sham-operated mice (Fig. 2A–C). The activation of RXR-α significantly improved the neurological function in all tests on days 3 and 7 after ICH, compared to the vehicle group. A T2*-weighted MRI scan (coronal sections) was used to measure the hematoma volume in the parenchyma and at days 1, 3, 7, 14, and 28 after ICH. No significant differences in hematoma volume were noted at 24 h post-ICH, suggesting that the models yield consistent hematoma volumes. Compared with the vehicle group, the activation of RXR-α with bexarotene significantly decreased the hematoma volume from days 3 to 7 after ICH (Fig. 2D).

Fig. 2.

Pharmacological activation of RXR-α significantly alleviates neurological dysfunction and promotes hematoma clearance after ICH. A–C Quantification of neurological function with the forelimb placement test (A), corner turn test (B), and cylinder test (C) at days 1, 3, 7, 14, and 28 (n = 9/group; data are represented as the mean ± SEM; ^#P <0.05 vs ICH + vehicle group). D Quantitative analysis of hematoma volumes (n = 6/group; data are represented as mean ± SEM; ^#P <0.05 vs ICH + vehicle group, ^##P <0.001 vs ICH + vehicle group). E T2*-weighted MRI scans (coronal sections) showing the sham group image and the changes in hematoma volumes over time after ICH. The red dotted lines denote the hematoma.

Bexarotene Promotes Nuclear Translocation of RXR-α in Microglia/Macrophages

Dual-label immunofluorescence staining showed that RXR-α was expressed in microglia/macrophages (Iba-1), neurons (NeuN), and astrocytes (GFAP) in the sham group and in the mice basal ganglia region (Fig. 3A). Then we noticed that the microglia/macrophages at day 3 after ICH had a stronger nuclear signal and higher nucleoplasm ratio compared to the sham group (Fig. 3B). However, no significant changes in the ratio of nucleus:cytoplasm and were found in neurons and astrocytes (Fig. 3C, D). To determine the temporal profile of RXR-α after ICH in mice, we examined their total and subcellular expression patterns by Western blotting on days 1, 3, 7, and 14 after ICH. The whole-cell expression patterns RXR-α in the ipsilateral hemisphere did not show any significant differences across these groups (Fig. 3E). The cytoplasmic expression of RXR-α was significantly reduced on days 3 and 7 after ICH (Fig. 3F), while the nuclear level showed an increase on the same days in ICH-injured mice as in the sham group (Fig. 3G). To further explore the effect of bexarotene on the nuclear-cytoplasmic shuttling of RXR-α, its expression in the ipsilateral basal ganglia at day 3 after ICH was assessed by Western blotting. The total expression of RXR-α did not show any significant differences across the three groups (Fig. 3H). The RXR-α expression level in the bexarotene group was higher in the nucleus and lower in the cytoplasm than in the vehicle group (Fig. 3I, J). In sum, the nuclear translocation of RXR-α was induced after ICH, especially in microglia/macrophages, and the administration of bexarotene further enhanced this translocation.

Fig. 3.

Bexarotene promotes the nuclear translocation of RXR-α in microglia/macrophages. A Representative microphotographs of immunofluorescence double staining showing RXR-α (green) with NeuN, GFAP, and Iba-1 (red) in the sham group and ICH 72 h group (scale bar, 5 μm). B Ratio of nuclear to whole-cell fluorescence of microglia/macrophages expressing RXR-α (n = 6/group; data are represented as the mean ± SEM; *P <0.05 vs sham). C, D Ratios of nuclear to whole-cell fluorescence of neurons (C) and astrocytes (D) expressing RXR-α (n = 6/group; mean ± SEM). E–G Representative Western blotting images and quantitative analyses of RXR-α in whole cells (E), cytoplasm (F), and nuclei (G) of the ipsilateral basal ganglia after ICH by Western blot (n = 6/group; mean ± SEM; *P <0.05 vs sham. H–J Representative Western blotting images and quantitative analyses of RXR-α in the whole cells (H), cytoplasm (I), and nuclei (J) of the ipsilateral basal ganglia after ICH (n = 6/group; mean ± SEM; *P <0.05 vs sham, ^#P <0.05 vs ICH + vehicle group).

Depletion of Microglia/Macrophages Inhibits Bexarotene-mediated Hematoma Clearance

To determine whether microglia/macrophages play a major role in the bexarotene-mediated hematoma resolution, Clodronate liposome treatment was used to deplete microglia/macrophages. To verify the validity of depletion, the number of microglia/macrophages and astrocytes in the peri-hematoma area at day 3 after ICH were counted. The number of Iba1-positive cells was less in the Clophosome treatment group, reduced by 85.5% of the control group (Fig. 4A, B). But this treatment did not deplete astrocytes (Fig. 4C). Besides, we examined the effect of bexarotene on the hematoma resolution after depletion of microglia/macrophages. Quantification of hemoglobin showed that Clophosome treatment slowed the absorption of the hematoma, and there was no statistical difference between bexarotene treatment and vehicle groups. Consistent with our previous results, the hemoglobin quantification of the group treated with control liposomes showed that bexarotene accelerated hematoma clearance at day 3 after ICH (Fig. 4D, E).

Fig. 4.

Depletion of microglia/macrophages inhibits bexarotene-mediated hematoma clearance. A Representative microphotographs of Iba1-positive and GFAP-positive cells around the hematoma (scale bar, 100 μm; dotted lines mark the hematoma boundary). B, C Quantitative analysis of Iba1-positive (B) and GFAP-positive (C) cells (mean ± SEM; *P <0.05 vs control group). D Representative brain coronal sections showing the hematoma after treatment with Clophosome + vehicle, Clophosome + bexarotene, control + vehicle, and control + bexarotene at day 3. (scale bar, 5 mm). E Quantification of hematoma volume (n = 6/group; mean ± SEM; *P <0.05 vs Clophosome + vehicle, ^#P <0.05 vs control + vehicle group).

Activation of RXR-α with Bexarotene Enhances the Nuclear Translocation and Transcription Activity of PPAR-γ

To determine the nuclear-cytoplasmic translocation of PPAR-γ after ICH, immunofluorescence staining was performed in the perihematomal region from the sham group and the third-day post-ICH group (Fig. 5A). We noted that nuclear translocation of PPAR-γ was prominent in microglia/macrophages after ICH compared to the sham group (Fig. 5B–D). Besides, we tested the total and subcellular expression of PPAR-γ by Western blotting at days 1, 3, 7, and 14 after ICH. The endogenous expression of PPAR-γ in the whole cells increased on day 3 and remained significantly higher on day 14 (Fig. 5E). The expression of PPAR-γ in the cytoplasm was reduced on day 7, and its nuclear expression increased on days 3 and 7 after ICH (Fig. 5F, G). To explore whether PPAR-γ plays a role in the process of RXR-α activation, the nuclear-cytoplasmic translocation of PPAR-γ was assessed at day 3 after ICH in the four experimental groups. The total expression of PPAR-γ did not show any significant differences between the vehicle, bexarotene, and bexarotene + GW9662 groups (Fig. 5H). The PPAR-γ expression in the bexarotene group was remarkably reduced in the cytoplasm and was increased in the nucleus compared with the vehicle group, while the PPAR-γ antagonist GW9662, in contrast, decreased the PPAR-γ expression in the nucleus and increase it in the cytoplasm (Fig. 5I, J). To examine the effect of bexarotene on PPAR-γ transcription activity, we performed a PPAR-γ transcriptional activity assay. The transcriptional activity of PPAR-γ was increased after ICH, and bexarotene further increased it. However, the effect was abolished by GW9662 treatment (Fig. 5K).

Fig. 5.

Activation of RXR-α with bexarotene enhances the nuclear translocation and transcription activity of PPAR-γ. A Representative microphotographs of immunofluorescence double staining showing PPAR-γ (green) with NeuN, GFAP, and Iba-1 (red) in the sham group and ICH 72 h group (scale bar, 5 μm). B–D Ratios of the nuclear to whole-cell fluorescence of microglia/macrophages (B), neurons (C), and astrocytes (D) expressing PPAR-γ (n = 6/group; mean ± SEM; *P <0.05 vs sham). E–G Representative Western blotting images and quantitative analyses of PPAR-γ in whole cells (E), cytoplasm (F), and nuclei (G) of the ipsilateral basal ganglia after ICH (n = 6/group; mean ± SEM; *P <0.05 vs sham). H–J Representative Western blotting images and quantitative analyses of PPAR-γ in whole cells (H), cytoplasm (I), and nuclei (J) of the ipsilateral basal ganglia after ICH (n = 6/group; mean ± SEM; *P <0.05 vs sham, ^#P <0.05 vs ICH + vehicle group, ^@P <0.05 vs ICH + bexarotene group). K Quantitative analyses of PPAR-γ transcription activity at day 3 after ICH (n = 6/group; mean ± SEM; *P <0.05 vs sham, ^#P <0.05 vs ICH + vehicle group, ^@P <0.05 vs ICH + bexarotene group).

RXR-α Activation Regulates Microglia/Macrophage Polarization Through PPAR-γ

To determine whether PPAR-γ plays a role in the polarization of microglia/macrophages, we examined microglia/macrophage markers of polarization iNOS [a pro-inflammatory state (M1) marker] and Arg1 [an anti-inflammatory state (M2) marker] on day 3 after ICH by immunofluorescence. We found that the expression of iNOS and Arg1 increased after ICH compared to the sham group (Fig. 6A, B). Furthermore, the expression of iNOS and Arg1 was high in Iba1+ cells (microglia/macrophages) in the perihematomal area of mouse brain sections. The co-expression of iNOS and Iba1 was less pronounced in the presence of bexarotene, while GW9662 reversed this effect (Fig. 6A). In contrast, bexarotene led to a significant increase in the number of M2-like polarized microglia/macrophages, and this was reversed by GW9662 administration (Fig. 6B). ELISA revealed that the TNF-α levels were increased after ICH (Fig. 6C). Bexarotene reduced TNF-α, and this effect was attenuated by GW9662 (Fig. 6C). When the expression of the M2-like marker Arg1 was measured by Western blot, bexarotene stimulated the M2-like marker while GW9662 reversed this reduction (Fig. 6D).

Fig. 6.

RXR-α activation regulates microglia/macrophage polarization through PPAR-γ. A Representative microphotographs and quantitative analysis of iNOS-positive (green) and Iba1-positive (red) cells (scale bar, 50 μm; dotted lines mark the hematoma boundary; n = 4/group. mean ± SEM; *P <0.05 vs sham, ^#P <0.05 vs ICH + vehicle group, ^@P <0.05 vs ICH + bexarotene group). B Representative microphotographs and quantitative analysis of ARG1-positive (green) and Iba1-positive (red) cells (scale bar, 50 μm, dotted lines mark the hematoma boundary; n = 4/group; mean ± SEM; *P <0.05 vs sham, ^#P <0.05 vs ICH + vehicle group, ^@P <0.05 vs ICH + bexarotene group). C Quantitative analyses of TNF-α at day 3 after ICH (n = 6/group; mean ± SEM; *P <0.05 vs sham, ^#P <0.05 vs ICH + vehicle group. ^@P <0.05 vs ICH + bexarotene group). D Representative Western blotting images and quantitative analyses of ARG1 at day 3 after ICH (n = 6/group; mean ± SEM; ^#P <0.05 vs ICH + vehicle group, ^@P <0.05 vs ICH + bexarotene group).

Role of PPAR-γ in RXR-α-mediated Neuroprotection Against ICH and Hematoma Absorption

In the next part of this study, the PPAR-γ antagonist GW9662 was introduced to further determine the role of PPAR-γ in RXR-α-mediated protective effects under ICH conditions. Three behavioral tests were used to examine the role of PPAR-γ in behavioral recovery on days 1, 3, 7, 14, and 28 after ICH. The activation of RXR-α significantly improved neurological function, while the protective effects of bexarotene treatment were reversed by PPAR-γ inhibition (Fig. 7A–C). A T2*-weighted MRI scan (coronal sections) was used to measure the hematoma volume on days 1, 3, 7, and 14 after ICH. Similarly, RXR-α activation significantly promoted hematoma absorption on days 3 and 7 after ICH (Fig. 7D, E). However, the blockade of PPAR-γ with GW9662 reversed the bexarotene-mediated hematoma absorption (Fig. 7D, E).

Fig. 7.

Role of PPAR-γ in RXR-α-mediated neuroprotection against ICH and hematoma absorption. A–C Quantification of neurological function in the forelimb placement test (A), corner turn test (B), and cylinder test (C) at days 1, 3, 7, and 14 (n = 9/group; mean ± SEM; ^#P <0.05 vs ICH + vehicle group, ^@P <0.05 vs ICH + bexarotene group, ^@@P <0.01 vs ICH + bexarotene group). D Quantitative analysis of hematoma volumes (n = 6/group; mean ± SEM; ^#P <0.05 vs ICH + vehicle group, ^@P <0.05 vs ICH + bexarotene group). E Representative T2*-weighted MRI scans (coronal sections) showing the changes in hematoma volume over time after ICH. The red dotted lines denote the hematoma.

Discussion

In this study, we investigated the role of RXR-α in the pathological process following ICH and assessed the relevant underlying mechanisms. Our findings are as follows: (1) the activation of RXR-α alleviated ICH-induced neurological deficits and promoted hematoma absorption; (2) nuclear translocation of RXR-α occurred after ICH, and this effect was further enhanced by the pharmacological activation of RXR-α; (3) the hematoma clearance effect mediated by RXR-α activation was attenuated after microglia/macrophage depletion; (4) a similar nuclear translocation of PPAR-γ also occurred following injury. Moreover, RXR-α activation enhanced PPAR-γ translocation, while the inhibition of PPAR-γ reversed this effect; (5) RXR-α activation tended to polarize microglia/macrophages into M2-like rather than M1-like cells through PPAR-γ signaling; and (6) the PPAR-γ antagonist GW9662 abolished the protective effect of RXR-α activation after ICH. Based on the evidence above, the pharmacological activation of RXR-α significantly alleviates ICH-induced neurological deficits and promotes hematoma absorption. The protective effect of RXR-α activation was at least partly mediated by modulating M1/M2 macrophage polarization through PPAR-γ signaling.

ICH is a cerebrovascular disease linked to high mortality and morbidity rates [27]. A hematoma within the brain parenchyma leads to secondary injuries and severe neurological deficits [28]. To date, the efficient removal of the hematoma is a clinically promising intervention for ICH [29, 30]. Successful elimination of deposited blood and apoptotic cells by phagocytosis is essential for the resolution of inflammation [4]. RXR is a member of the superfamily of nuclear hormone receptors and consists of three distinct functional domains: an amino-terminal domain, involved in its ligand-independent basal transcriptional activity, a DNA-binding domain (DBD), and a ligand-binding domain (LBD). The LBD contains the ligand-binding regions that coordinate the formation of dimers and the recruitment of transcriptional co-regulators [31]. RXR mediates their ligand-dependent transcriptional activity via the LBD and DBD. It regulates metabolic and immune responses, including the regulation of macrophage immune phenotypes and the clearance of apoptotic cells [13, 32]. In addition, several studies have highlighted that RXR indeed plays an important role in optimizing the phagocytic function in the CNS [33, 34]. Although the roles of RXR-α have been studied for years, little is known about its role in the pathological processes following ICH.

Therefore, in the first part of this study, we assessed the effects of RXR-α on the pathological process of ICH. The RXR-α agonist bexarotene significantly attenuated neurological function, which is negatively related to the volume of the hematoma after ICH in humans and rodents [35, 36]. In addition, bexarotene significantly promoted hematoma resolution in the experimental mouse model, consistent with the results of Chang et al. [37]. All of the above data suggest that the activation of RXR-α promotes hematoma clearance while having a neuroprotective effect against ICH. However, the mechanisms underlying the protective effects of RXR-α remain unclear.

Microglia are vital to maintaining brain homeostasis, acting as phagocytes to scavenge for debris in the CNS. Targeting microglia may be an ideal therapeutic strategy for stroke [10, 38]. Besides, immunofluorescence staining showed that RXR-α is expressed in microglia/macrophages, astrocytes, and neurons. Notably, recent studies have demonstrated that PPAR-γ in microglia/macrophages increases phagocytosis and modulates inflammation by downregulating the expression of pro-inflammatory mediators [30, 39]. Furthermore, to activate transcription, PPAR-γ needs to heterodimerize with RXR-α, which is an important signaling hub in nuclear receptor-controlled transcription [40, 41]. More importantly, recent studies have confirmed that the PPAR-γ signaling pathway is upregulated after exposure to bexarotene, suggesting an eminent role for PPAR-γ in the transcriptional regulation of RXR-α [21, 34]. The mechanism underlying RXR transcriptional activity is related to its nucleoplasmic sub-localization to the nuclear splicing factor compartment [42]. Yasmin et al. reported that enhanced RXR nuclear translocation is enabled by importin-β after RXR activation [43]. Given the essential role of RXR-α in regulating PPAR-γ transcription and the expression of RXR-α in microglia/macrophages after ICH, we hypothesized that RXR-α acts as an important regulator mediating the nuclear translocation of PPAR-γ.

Therefore, we then investigated the nuclear translocation of RXR-α and PPAR-γ after ICH. In the next experiment, we used the specific and irreversible inhibitor of PPAR-γ, GW9662, which inactivates PPAR-γ through covalent modification and does not change RXR-α activity [44]. We noted the enhanced translocation to the nucleus of RXR-α and PPAR-γ after injury. Our findings are consistent with a previous study that reported enhanced translocation of RXR-α and PPAR-γ after traumatic brain injury [18]. Several studies on intracellular RXR shuttling have been performed, showing that the nuclear translocation of RXR is triggered by inflammatory cytokines such as IL-1β, IL-6, and TNF-α [45, 46]. The local inflammatory response following injury may, therefore, contribute to the activation and subcellular translocation of the retinoid X receptor, which has been shown to have anti-inflammatory properties [46]. Even though the translocation of RXR-α and PPAR-γ into the nucleus increased after ICH, this may not be sufficient to reduce the neuroinflammation after ICH. The pharmacological activation of RXR-α may further potentiate protective pathways against neuroinflammation after ICH. We found that bexarotene significantly increased the nuclear translocation of RXR-α and PPAR-γ, while the PPAR-γ inhibitor GW9662 reversed this effect. Notably, RXR-α and PPAR-γ showed a similar trend towards nuclear translocation, and pharmacologically activating RXR-α further promoted their nuclear translocation, suggesting that RXR-α acts synergistically with PPAR-γ through their interaction.

Various mechanisms are involved in the regulation of tissue healing, hematoma removal, and inflammation resolution. One of the essential mechanisms, especially concerning the neuroinflammation that is incurred by post-hemorrhagic secondary brain injury, is associated with the conversion of the classical activation phenotype (M1) to the alternative activation phenotype (M2) [10]. In the CNS, M2-like microglia/macrophages are efficient phagocytes [11]. Nuclear receptors play a key role in balancing M1/M2 polarization and controlling phagocytosis by regulating the transcription of functional genes [40]. It has been demonstrated that PPAR-γ is an important transcription factor that plays a crucial role in switching on M1-like or M2-like marker genes, thereby mediating the priming of microglia/macrophages towards M2 polarization [47]. In addition, PPAR-γ activation mediates the transcription of downstream genes such as ABCA1 and Cd36, which are also considered to be biomarkers of scavenger receptors that contribute to the phagocytic ability of microglia/macrophages [3, 34]. A previous study has shown that activation of the RXR/PPAR heterodimer exerts neuroprotective effects by modulating microglial polarization [18]. Given the involvement of RXR-α and PPAR-γ in microglia/macrophage-type cell polarization and differentiation, we assessed the expression of M1/M2 markers in microglia/macrophages. M1 marker proteins (iNOS and TNF-α) and an M2 marker protein (ARG1) were used, revealing that microglia/macrophages were significantly polarized into the classical activation phenotype (M1) after ICH. This finding is consistent with a previous study [18], as our data indicated that the activation of RXR-α inhibited M1 activation and drove microglia/macrophages towards the M2 phenotype, while the PPAR-γ inhibitor GW9662 reversed these effects. In the rest of this study, we further explored the role of PPAR-γ in neurological function and hematoma volume in response to RXR-α activation after ICH. Consistent with our previous results, RXR-α agonists significantly alleviated the neurological dysfunction and promoted hematoma resolution. However, all these beneficial effects were abolished by the PPAR-γ inhibitor GW9662. Taken together, we deduced that the activation of RXR-α promotes microglia/macrophage polarization towards the M2 phenotype through PPAR-γ after ICH, therefore facilitating hematoma clearance and improving neurological function.

However, our research has some undeniable limitations. First, as noted above, RXR-α was also expressed in neurons and astrocytes, suggesting it may play a potential role in other cells. Second, in addition to forming heterodimers with PPAR-γ, recent findings have indicated that a separate RXR-α homodimer can also affect the function of the immune system and participate in the regulation of innate immunity [32]. We solely focused on the RXR-α/PPAR-γ signaling pathway in the current study. Lastly, resident microglia and peripheral microglia may both contribute to neuroinflammation; however, which of these mediate the beneficial effects has not yet been determined [48]. Therefore, further efforts are required to obtain more details with regard to other possible mechanisms mediated by RXR-α in ICH.

In summary, our data showed that pharmacological activation of RXR-α promoted hematoma clearance, improved neurological function, and polarized microglia/macrophages towards the M2 phenotype via regulation of the nuclear translocation of PPAR-γ after ICH. Altogether, the current study supports the hypothesis that targeting RXR-α might be a novel and promising therapeutic strategy for ICH.

Availability of Data and Materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflict of interest

The authors declare that they have no conflicts of interest.