Introduction

Intracerebral haemorrhage (ICH) accounts for 12%–20% of all types of stroke with over 2 million individuals worldwide being afflicted annually.1 Even with the development of minimally invasive surgical processes, the prognosis of ICH is catastrophic.2 The latter is attributed to inflammation-mediated secondary brain injury and limited tissue regeneration.3 4 CNS-resident microglia and infiltrating monocyte-derived macrophages are at the nexus of inflammation and regeneration.5 From postmortem studies, the density of microglia/macrophages (M/M) elevates in the perihaematomal region from the first day of ICH and increases thereafter.6 The transcriptome of haematoma-associated myeloid cells may shift temporally from proinflammatory to phagocytic, anti-inflammatory and antioxidative.7 Preclinical data show that inhibiting microglia activation or macrophage infiltration in the early phases after experimental ICH reduces secondary brain injury,8–10 while longterm inhibition impairs tissue regeneration and recovery, presumably due to insufficient beneficial M/M at later stages.5 11 12 Thus, it would be ideal to inhibit the proinflammatory/injurious properties of M/M and promote their regulatory and reparative properties following ICH. One such candidate could be the elevation of Liver X receptor (LXR) signalling.

LXRα or β nuclear receptors form heterodimers with retinoid X receptor (RXR), where the complex then binds to a specific DNA region containing repetitive AGGTCA sequence named LXR response element.13 14 When ligands are absent, the LXR-RXR complex binds to a corepressor such as histone deacetylase 3, silencing mediator for retinoid and thyroid receptors, or nuclear receptor corepressor to remain inactivated, thereby preventing the transcription of downstream genes including apolipoprotein E (ApoE), ATP-binding cassette transporter subfamily A and G (ABCA, ABCG), and inducible degrader of LDLR.15–18 When endogenous (eg, cholesterol derivatives or cholic acid) or synthetic ligands (eg, GW3965 and T0901317) bind LXRs, the LXR-RXR heterodimer changes conformation and releases corepressors while recruiting coactivators such as histone acetyltransferase p300 and activating signal co-integrator 2.19–25 The latter sequentially induces the expression of target genes to facilitate cholesterol efflux and lipid metabolism.26

Because the cholesterol within the adult CNS mainly relies on de novo synthesis in physiological condition,27 neural cells including neural stem cells (NSCs) and oligodendrocyte progenitor cells can only use recycled cholesterol after an acute injury due to insufficient synthesis, and thus may be vulnerable to cholesterol deficiency. As the main processors of cholesterol-rich myelin and cellular components, M/M could degrade these debris and recycle the cholesterol for repair. If so, upregulated LXR signalling could promote cholesterol efflux from phagocytes to facilitate tissue regeneration.28–30 Additionally, ligand bonded LXRs can be SUMOlyated to interact with nuclear factor kappa B, activator protein 1 (AP-1) and signal transducer and activator of transcription 1, leading to the transrepression of their downstream proinflammatory genes.31–33 Moreover, analyses by transposase-accessible chromatin sequencing and chromatin immunoprecipitation sequencing show that activated LXRs could bind inflammatory gene enhancers to induce chromatin closing, resulting in cis-repression of these genes in lipopolysaccharide-treated bone marrow derived macrophages.34

Given the requirement for cholesterol recycling for lipid synthesis in regenerating oligodendrocytes and myelin,29 30 and the importance of preventing excessive neuroinflammation, stimulating the LXR pathway may have therapeutic benefits in ICH through neuroprotection in the acute period, and in regenerative processes in the longer term. In this manuscript, we evaluated the therapeutic effects and potential mechanisms of LXR stimulation following ICH for neuroprotection and reparative oligodendrocyte repopulation.

Methods

The data that support the findings of this study are available from the corresponding authors on reasonable request.

Animals

Animals were sacrificed when they showed any signs of distress: inability to properly drink, eat, groom themselves, unresponsiveness, isolation, seizures or when there was unexpected loss of body weight (higher than 20%, relative to the starting weight of the animal and controls). Male C57Bl/6J mice acquired from Charles River (Montreal) were used at 8–12 weeks of age for in vivo experiments and were housed in individual ventilated cages with 22°C–24°C temperature, 50%–60% humidity, environmental enrichment and free access to food and water. Eighty-eight C57Bl/6J mice were randomly assigned to 6 groups including sham (n=9), ICH (n=15), sham+DMSO (n=6), sham+GW3965 (n=6), ICH+DMSO (n=24) and ICH+GW3965 (n=24). Two mice that died immediately after surgery and 2 other mice that did not have anticipated behavioural deterioration (less than 20% drop 1 day after surgery compared with baseline of either behavioural test) after experimental ICH were excluded according to preset criteria.

CX3CR1CreER (JAX 021160) mice and Rosa26iDTR (JAX 007900) mice from The Jackson Laboratory were bred in the single mouse barrier unit at the University of Calgary to produce CX3CR1CreER:Rosa26iDTR mice. On tamoxifen treatment to induce the diphtheria toxin (DT) receptor on CX3CR1-expressing cells (mostly M/M), the subsequent administration of DT would kill M/M. Male mice were used in the M/M-conditional depletion experiments at the age of 12–20 weeks. Tamoxifen (75 mg/kg, T5648, Sigma) dissolved in corn oil (Sigma, C8267) was given intraperitoneally (IP) once a day for five consecutive days. Three days after the last injection, and corresponding to the day of ICH surgery, DT (50 µg/kg) or saline was injected IP, followed by daily DT or saline for the next 6 days to ensure the depletion of M/M. Moreover, 1 hour after ICH, randomly assigned mice were initiated daily IP of GW3965 or vehicle (50% DMSO). In this manner, 24 CX3CR1CreER:Rosa26iDTR mice that were subjected to ICH were in four groups of saline+DMSO (n=4), saline+GW3965 (n=4), DT+DMSO (n=8) and DT+GW3965 (n=7, one mice died 6 days after surgery).

Collagenase-induced ICH

The procedure for collagenase-induced ICH in mice is described in previous studies,35 and was approved by Animal Care Committee of the University of Calgary. Briefly, mice were anaesthetised with IP injection of ketamine (100 mg/kg) and xylazine (10 mg/kg). Fur was removed from the surgical site using hair clippers. Ophthalmic gel was applied to both eyes to keep them moist throughout the surgery and recovery period. A heat lamp was used over the stereotaxic frame to provide heat support to the animal during the procedure. The mouse was placed in a stereotaxic frame using modified ear bars fitted with blunt rubber ends designed for mice. The surgical site was cleaned using betadine and a skin incision was made to reveal the skull. An area of the skull above the right striatum was located according to the following coordinates (2.0 mm lateral and 0.8 mm anterior to the bregma), and a 0.5 mm cranial burr hole was made with a microdrill. A Hamilton syringe, containing 0.05 U of collagenase type VII dissolved in 0.5 µL of saline was lowered through the hole until 3.2 mm beneath the skull, ending in the striatum. The injection of collagenase lasted for 5 min per mouse and was controlled by a UMP3 UltraMicroPump. Following injection, the needle was maintained inside the brain for another 5 min to prevent reflux. The needle was then slowly withdrawn. Finally, the surgical site was sutured and disinfected. Animals were transferred into a warm incubator chamber at approximately 32°C until full recovery. Mice were then returned to their cages, given free access to food and water and monitored daily. For sham animals, the same volume of saline (0.5 µL) without collagenase was injected following the same procedure.

GW3965 treatment

For LXR activation in vivo, mice were injected once daily, IP with sterile 50% DMSO (Sigma) or 50% DMSO containing the LXR agonist GW3965 (Sigma) at 10 mg/kg.36 37 Treatment started 1 hour after intracerebral collagenase injection, and then once daily until day 7 (7 doses in total). Mice were randomly numbered before surgical procedure and odd numbers were assigned to DMSO control group while even numbers were assigned to GW3965 treatment group. Injection was conducted alternatively between the two groups, and each cage included mice from both groups.

Brain tissue isolation

Mice were euthanised with a ketamine and xylazine overdose at 1, 3, 7 or 14 days postsurgery. A total of 15 mL of PBS was then perfused via cardiac puncture, followed by perfusion with 15 mL of 4% paraformaldehyde (PFA) in PBS. The whole brain was then dissected from the skull of the mouse and left in 4% PFA in PBS for fixation. After overnight fixation in 4% PFA at 4°C, brain samples were transferred to 30% sucrose solution for dehydration. After at least 48 hours of incubation in 30% sucrose solution, the cerebellum was removed and the remaining brain tissue was frozen in FSC 22 frozen section media (Leica). Using a cryostat (ThermoFisher Scientific), brain tissue was cut coronally into 20 µm sections, collected onto Superfrost Plus microscope slides (VWR) and stored at −20°C before analysis.

Antibodies and fluorescent dyes

The following primary antibodies were used for immunofluorescence microscopy (stock concentrations were not informed by some suppliers): chicken anti-mouse glial fibrillar acidic protein (GFAP) (1:1000; BioLegend, 829401); rabbit anti-human/mouse Iba1 (1:1000 of 0.6 mg/mL stock; Wako, 019-19741); rabbit anti-human/mouse NeuN (1:500 of 0.82 mg/mL stock; Abcam, ab177487); rat anti-mouse CD68 (1:500 of 0.5 mg/mL stock; BioLegend, 137002); goat anti-human/mouse Olig2 (1:200 of 0.2 mg/mL stock; R&D Systems, AF2418); chicken anti-human/mouse myelin basic protein (MBP) (1:2000; ThermoFisher, PA1-10008); rabbit anti-human/mouse neurofilament heavy chain (NF-H) (1:1000 of 1 mg/mL stock; Encor Biotechnology, RPCA-NF-H); rat anti-mouse ATP Binding Cassette Subfamily A Member 1 (ABCA1) (1:200; ThermoFisher, MA1-16936); rabbit anti-human/mouse ApoE (1:500 of 0.5 mg/mL stock; ThermoFisher, 701241); goat anti-mouse interleukin (IL)-1β (1:50 of 0.2 mg/mL stock; R&D Systems, AF-401-NA); rat anti-mouse inducible nitric oxide synthase (iNOS) (1:100 of 0.5 mg/mL stock; ThermoFisher, 14-5920-82); rabbit anti-mouse arginase (Arg)1 (1:200 of 18 µg/mL stock; Cell Signaling, 93 668S); goat anti-mouse CD206 (1:200 of 0.2 mg/mL stock; ThermoFisher, PA5-46994); rabbit anti-human/ mouse Olig2 (1:200; Millipore, AB9610); goat anti-mouse platelet-derived growth factor receptor (PDGFR)α (1:200 of 0.2 mg/mL stock; R&D Systems, AF1062); mouse anti-human/mouse adenomatous polyposis coli (APC), clone CC-1 (1:200 of 0.1 mg/mL stock; Millipore, OP80-100UG); rabbit anti-human/mouse amyloid precursor protein (APP) (1:500 of 0.25 mg/mL stock; ThermoFisher, 36-6900); rat anti-human/mouse SOX2 (1:200 of 0.5 mg/mL stock; ThermoFisher, 14-9811-82); chicken anti-human/mouse nestin (1:1000 of 0.3 mg/mL stock; Novusbio, NB100-1604); rabbit anti-human/mouse Ki67 (1:500 of 0.8 mg/mL stock; Abcam, ab15580); rabbit anti-mouse doublecortin (DCX) (1:800 of 130 µg/mL; Cell Signaling, 4604) and rat anti-mouse CD16/CD32 Fc blocking antibody (1:100 of 0.5 mg/mL stock; BD Biosciences, 553142).

The following secondary antibodies from Jackson ImmunoResearch were used at 1:400 dilution of 0.625 mg/mL stock: Alexa Fluor 488 donkey anti-mouse IgG, Alexa Fluor 488 donkey anti-rat IgG, Alexa Fluor 488 donkey anti-chicken IgY, cyanine Cy3 donkey anti-goat IgG, cyanine Cy3 donkey anti-chicken IgY, Alexa Fluor 594 donkey anti-chicken IgY, and Alexa Fluor 647 donkey anti-rabbit IgG.

These fluorescent dyes were used to counterstain with antibodies for immunofluorescence microscopy: DAPI (1:1000 of 1 mg/mL stock; Sigma, D9542-10MG); BODIPY 493/503 (1:1500 of 1 mg/mL stock; ThermoFisher, D3922); and FluoroMyelin Green Fluorescent Myelin Stain (1:300; ThermoFisher, F34651).

Immunofluorescence staining

Microscope slides containing mouse brain samples were warmed to room temperature (RT) for 30 min. For MBP staining, slides were also delipidated by successive washes of 50%, 70%, 90%, 95%, 100%, 100%, 95%, 90%, 70% and 50% ethanol. Sections were then rehydrated with PBS for 10 min, and permeabilised with 0.2% Triton X-100 in PBS 10 min. Horse blocking solution (PBS, 10% horse serum, 1% BSA, 0.1% cold water fish skin gelatin, 0.1% Triton X-100, 0.05% Tween-20) was used to block the sample for 1 hour at RT. Alternatively, for staining using the CC1 antibody in mouse sections, purified rat anti-mouse CD16/CD32 Fc blocking antibody (1:100 of 0.5 mg/mL stock; BD Biosciences, 553142) was added to the blocking buffer. After blocking, samples were incubated with primary antibodies resuspended in antibody dilution buffer (PBS, 1% BSA, 0.1% cold water fish skin gelatin, 0.1% Triton X-100) overnight at 4°C. Following this, samples were washed three times, 5 min each, using PBS with 0.2% Tween-20, and then incubated with secondary antibodies (and FluoroMyelin or Bodipy for specific staining) and DAPI resuspended in the antibody dilution buffer for 1 hour at RT. The samples were then washed three more times using PBS with 0.2% Tween-20, 5 min each, and the coverslips were mounted onto the slides using Fluoromount-G solution (SouthernBiotech).

Confocal immunofluorescence microscopy

Laser confocal immunofluorescence images were acquired at RT using the Leica TCS Sp8 laser-scanning confocal microscope, using a×25/0.5 NA water objective for all samples and experiments. The 405 nm, 488 nm, 552 nm and 640 nm lasers were used to excite the fluorophores from antibodies and fluorescent dyes bound to samples and detected by two low-dark-current Hamamatsu PMT detectors and two high-sensitivity hybrid detectors on the Sp8. Multiple-FOV images of perihaematomal regions were acquired using the following parameters: in 8 bits, in a z-stack with one-directional scanning, 1 airy unit pinhole, ×0.75 zoom, 0.57 µm per optical section, and 2048×2048 pixels x–y resolution. For reflection microscopy, reflection of crystals was captured by placing a PMT detector directly over the 552 nm laser channel for reflection light capture, and fluorescence was simultaneously captured by the same confocal imaging techniques.30 38–40 For whole-lesion tile scan, zoom factor was adjusted to 1.3 to diminish shades between images, and all other parameters were kept the same. Equal laser, gain and offset settings to maximise contrast and minimise saturation were consistently used for all samples within each set of experiments. A sample slide stained with only the secondary antibodies and DAPI was always included for each experiment as a control to account for non-specific secondary immunofluorescence. Tile scan of the whole lesion or images from four FOV of the perihaematomal region of two lesion epicentres were acquired and averaged per biological replicate for analysis. Leica Application Suite X was used for image acquisition, ImageJ was used for image threshold and particle analysis, and Imaris was used for three-dimensional (3D) rendering.

The overview of brain section at different time points after surgery was imaged using Olympus vs120 slide scanner by a 20×objective with same laser and exposure settings. Images were extracted using OlyVIA (Olympus), and only the brightness and contrast were adjusted for better display.

Confocal image analysis

The z-stacks of confocal images were analysed using ImageJ. For each image z-stack, maximum-intensity projections were created with pseudocolor for each channel/marker from .lif file and analysed under eight-bit format. Then, an ROI was drawn around the lesion or perihaematomal region according to GFAP, Iba1, CD68 or nestin staining, and the area outside the ROI was excluded for analysis. To determine the positive signal, the threshold was set consistently using a value determined using the negative secondary antibody-stained control, contralateral and sham control. The analyse particles function was then used to create a mask and to quantify the positive signals in each ROI. The same brightness threshold values for setting the positive signal, as well as the size and circularity settings for particle analysis, were used consistently across all samples for each experimental set. For the representative images shown, the maximum-intensity projection of each channel/marker in a z-stack was merged and converted to RGB format by ImageJ. Only the brightness and contrast settings were adjusted, and consistently among samples, to better display the images. A 3D rendering of confocal image z-stacks was created in Imaris (Oxford Instrument) using the surface creation tool with background subtraction and machine learning-based signal filtration.

Quantitative real-time PCR

Total RNA was isolated from perihaematomal regions using RNeasy Mini kit (Qiagen) according to the manufacturer’s instructions and stored at −80°C. The concentration and purity of RNA were determined by measuring absorbance at 260/280 nm using a nanodrop spectrophotometer (Thermo Fisher Scientific). First strand cDNA synthesis was performed with 1 µg total RNA using miScript II RT Kit (Qiagen) for mRNA expression analyses according to the manufacturer’s instructions. Real-time qPCR was performed using QuantiFast SYBR Green master mix (Qiagen) and Quantitect Primers: ABCA1 (Qiagen, QT00165690), ApoE (Qiagen, QT01043889) and Actb (Qiagen, QT00095242) with the following cycling conditions: 95°C for 5 min, 40 cycles of denaturation at 95°C for 30 s, annealing at 60°C for 30 s and extension at 72°C for 30 s. Beta-actin housekeeping genes were used to normalise mRNA expression. The relative expression levels were analysed by 2-ΔΔct method.

Western blotting

Tissue extracts of perihaematomal regions were generated by RIPA Buffer (Sigma, R0278) supplemented with protease inhibitor cocktail (Sigma, P2714). Protein concentrations were determined using a bicinchoninic acid (BCA) assay kit (Sigma) with BSA as standard. Equal amounts of total protein and 10 µL of PageRuler Plus Prestained Protein Ladder (Thermo Fisher) were then run on NuPAGE Bis-Tris Protein Gels (Thermo Fisher) with Mops SDS Running Buffer (Thermo Fisher). After transferring to Hybond polyvinylidene difluoride (Amersham) membrane using a wet transfer system (Bio-Rad), membranes were blocked with 5% BSA and then probed overnight with the following primary antibodies: rabbit anti-human/mouse ABCA1 (1:1000; Cell Signaling, 96292), rabbit anti-mouse ApoE (1:1000; Cell Signaling, 49285), mouse anti-human/mouse IL-1β (1:1000; Cell Signaling, 12242), rabbit anti-mouse iNOS (1:1000; Cell Signaling, 13120), rabbit anti-human/mouse CD206 (1:1000; Cell Signaling, 24595), rabbit anti-human/mouse Arginase 1 (1:1000; Cell Signaling, 93668) and rabbit anti-mouse β-Actin (1:1000; Cell Signaling, 24595). The following day, membranes were washed with 0.1% 1×tris-buffered saline with Tween 20, incubated with the goat anti-rabbit or goat anti-mouse horseradish peroxidase–linked secondary antibody (1:20000; Thermo Fisher, 31460 or 31430); then developed using the SuperSignal West Femto Maximum Sensitivity Substrate (ThermoFisher), and imaged with the ChemiDoc Touch Imaging System (Bio-Rad).

Mosue brain histology

Eriochrome cyanine and neutral red (EC) staining were used to visualise lesions for histology in CX3CR1CreER: Rosa26iDTR strain. Briefly, brain cryosections were air dried then treated successively with CitriSolv (Fisher Scientific), isopropanol, and 100%, 95%, 90%, 70% and 50% ethanol, for 1 min each. Samples were then washed with distilled water for 1 min, followed by incubation with EC staining solution containing10% FeCl3 for 15 min, followed by washing with distilled water for 1 min, then immediately washed with 0.5% NH4OH for 10 s, followed by another 1 min of washing with distilled water. Samples were then incubated with 1% neutral red solution for 2 min, and washed again with distilled water for 1 min. Finally, samples were dehydrated with successive 1 min washes of 50%, 70%, 90%, 95% and 100% ethanol, 2 min wash of isopropanol, 4 min of CitriSolv, and coverslips were mounted with Acrytol mounting medium (Electron Microscopy Science). Brightfield images were then acquired using a×20/0.75 NA air objective using the Olympus vs120 Slidescanner. ROIs for lesions were drawn based on lower EC staining surrounded by higher neutral red staining in the region of basal ganglia, and the total areas were analysed using CellSens (Olympus).

Behavioural tests

Locomotor functions were tested before and 1, 3, 7, 14 days after the surgical procedure using rotarod and grip strength. Mice were brought to the testing room 30 min in advance to accommodate before each session. These are details of the methods:

Rotarod

Motor coordination and balance were assessed using a rotarod test. Each mouse was placed on top of the already revolving cylinder (4 rpm) facing away from the experimenter, in the orientation opposite to that of the rod movement so that forward locomotion was necessary to avoid a fall. The mice were then exposed to three test trials in which the rotation accelerated from 4 to 30 rpm for 5 min. Latency to fall was measured automatically and was averaged over the three trials, with increased latency indicating greater motor coordination and balance. A 3-day training session using a steady rotation speed was conducted before each experiment to diminish the learning effects.

Grip strength

This task measures the forelimb strength. The mouse was allowed to grasp onto a bar with their forelimbs and then gently pulled away by the experimenter. The force at the point when the mouse lost grip was automatically recorded. The test was repeated three times and averaged for each session.

MRI

MRI data were acquired using a 9.4T Bruker BioSpec equipped with a Bruker cryoprobe and operated with ParaVision V.5.1. Mice were anaesthetised initially with 2%–3% isoflurane and a tail vein cannulation was made before imaging. Mice were then head-fixed in an animal carrier using tooth and ear bars. Anaesthesia was maintained by 1.5% isoflurane. Respiration, and body temperature, were non-invasively monitored using a small animal monitoring system (Small Animal Instruments). Initially, a shimming procedure was conducted to correct the distortions and optimise the magnetic field homogeneity follow by a scout scan using FLASH sequence to determine the lesion epicentre and scan range. Subsequently, a high-resolution, T2-weighted sequence (T2WI) was acquired with 16 slices and 0.5 mm thickness using the following parameters: repetition time=4000 ms, echo time=48 ms, acquisition time=512 s, flip angle=180°, averages=2, FOV=19.2×19.2 mm2, resolution=256×256. Diffusion tensor imaging (DTI) at the same location was implemented using a DtiEpi SpinEcho sequence and the following parameters: repetition time=4000 ms, echo time=19.5 ms, spin echo time=0.3584 ms, acquisition time=1120 s, flip angle=90°, averages=2, FOV=19.2×19.2 mm2, matrix=256×256, segments=4, B0 images=5, gradient directions=30, gradient duration=3.8 ms, diffusion time=8 ms, and b value=2104.8 s/mm2. The dynamic contrast-enhanced MRI (DCE-MRI) was then performed using 3D gradient-echo with flow compensation T1-weighted imaging by using the following parameters: repetition time=20 ms, echo time=3.2 ms, acquisition time=245 s, flip angle=15°, averages=1, FOV=19.2×19.2×1.6 mm2, matrix=192×192×64. One image series was captured as baseline control before the injection of gadopentetate dimeglumine (0.4 mmol/kg, Bayer), followed by five consecutive scans starting immediately after contrast agent was delivered through the catheter.

Lesion volume, haematoma volume and ventricle volume were calculated by the product of slice (0.5 mm) thickness and manually drawn ROI at different slices of T2WI in ImageJ. The hypointense rim on T2WI was selected as lesion, while the hyperintensity within the rim was assigned as haematoma.41 42 Parameters of DCE-MRI were generated by algorithms in Jim9 (Xinapse Systems) according to modified Tofts model43 44: Ct(t) = Ktrans∫0t Cp(τ) exp(-Ktrans(t-τ) / ve) dτ + Cp(t)vp. The diffusion data were quantified using restricted diffusion imaging and reconstructed using generalised q-sampling imaging (GQI)45 46 in DSI studio with a diffusion sampling length ratio of 0.6. The region of right basal ganglia was manually drawn as ROI and the DTI parameters in ROI were exported. Fibre tracking was based on quantitative anisotropy (QA) and tractography was generated by fibres passing the ROI.

Statistical analyses

Data were collated using Microsoft Excel. Graph generation and statistical hypothesis testing were conducted using GraphPad Prism V.9. Data shown are the individual data point and mean±SD, whereby each point on a bar graph represents a biological replicate. No statistical methods were used to predetermine sample sizes, but our sample sizes were similar to those reported in previous publications.47 48 Two mice, each from GW3965 or DMSO group were excluded for all experiments and data collection due to absence of anticipated behavioural deterioration after experimental ICH. Animals were randomised when assigned to various experimental groups and conditions. Data collection and analysis of C57BL/6 wildtype mice were not performed blind to the conditions of the experiment, but all image and data analysis were completed with the same acquisition conditions and analysis thresholds. Tissue sections from CX3CR1CreER: Rosa26iDTR strain were blinded for microscopy and analyses. Normal distribution of data was tested using the D’Agostino-Pearson (n≥8) or Shapiro-Wilk (n＜8) normality test to verify whether statistical tests were appropriate. The non-parametric Mann-Whitney U test was employed when data did not pass the normality test. Two-way analysis of variance with Bonferroni’s multiple comparison’s test was used to analyse statistically significant differences among the means of multiple groups with two categorical variables. For all other data that passed the normality test, significance was determined by two-tailed, unpaired t-test or two-tailed, unpaired t-test with Welch’s correction when the homogeneity of variance is not fulfilled. Asterisks indicate significance where *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

ResultsThe LXR agonist GW3965 reduces post-ICH brain injury

To evaluate whether the synthetic LXR agonist GW3965 has overall therapeutic effects when administered after ICH, T2WI was applied to quantify the total lesion volume, haematoma volume and ventricle enlargement following ICH and whether this was affected by GW3965 treatment (figure 1A). The total lesion volume characterised by T2-hypointense rim was reduced in GW3965-treated mice at days 7 and 14 (figure 1B). The haematoma volume represented by T2-hyperintensity at day 7 was decreased by treatment (figure 1C), while the haematoma to lesion ratio was also lower in treated group (figure 1D), suggesting enhanced haematoma clearance. On day 14, measurement of ventricular enlargement as an indicator of brain tissue loss shows that this was lowered by GW3965 (figure 1E), indicating alleviated tissue loss and/or better regeneration by the LXR agonist treatment.

Figure 1

GW3965 accelerates haematoma clearance and alleviates brain tissue loss (A), Representative serial images of T2WI from DMSO-treated and GW3965-treated mice 7 or 14 days after collagenase induced ICH. Scale bar, 5 mm. (B–E), Bar graphs comparing data from T2WI between treatment and control groups: total lesion volume at days 7 and day 14 (B), haematoma volume and haematoma to lesion ratio at day 7 (C, D) and ipsilateral ventricle volume (E); the sample size was n=8 per group per time point. (F). Representative images of DCE-MRI from DMSO-treated and GW3965-treated mice 7 days after injury. Scale bar, 2 mm. (G, H), Bar graphs comparing the constant Ktrans (G) and Vp (H) between two groups at day 7 to evaluate the permeability of blood–brain barrier; the sample size was n=8 in GW3965 group but n=6 in DMSO group due to failed parameter generation by algorithm from raw data for two mice. DPI, days postinjury; ns, not significant. Significance is indicated as *p<0.05, **p<0.01, ***p<0.001; two-tailed, unpaired t-test (B–D, H) or two-tailed, unpaired t-test with Welch’s correction (E, G). Bar graphs show individual data points and represent mean±SD. DCE, dynamic contrast-enhanced; DPI, days postinjury; ICH, intracerebral haemorrhage; NS, not significant; T2WI, T2-weighted image.

DCE-MRI was applied 7 days after injury to test the permeability of blood–brain barrier (BBB) (figure 1F). There was a tendency of reduced Ktrans in treatment group (figure 1G), suggesting potentially ameliorated BBB breakdown by GW3965, but no statistical significance was found. There was no difference about Vp between two groups, indicating that blood flow did not contribute to the potential difference of Ktrans (figure 1H). Collectively, results from MRI suggest GW3965 has favourable effects on collagenase induced preclinical ICH, which prompted us to explore potential mechanisms.

GW3965 elevates LXR downstream gene expression

To address mechanisms by which GW3965 alleviates brain injury following ICH, we determined LXR regulated pathways (ABCA1 and ApoE expression) by tile scans of the whole lesion area of day 7 brain sections using confocal immunofluorescence microscopy (figure 2). While sham mice had barely detectable ABCA1 signal using identical confocal settings (online supplemental figure S1A), the expression of ABCA1 was elevated after ICH in the whole lesion area (figure 2A). Some of the ABCA1 profiles overlapped with Iba1+ M/M and minimally with GFAP+ astrocytes. The 3D reconstruction and confocal images at perihaematomal region exhibited close spatial association between ABCA1 and Iba1+ cells (figure 2B,C). The ABCA1 expression was then quantified (see Methods): total ABCA1 expression, the proportion of ABCA1+Iba1+ and ABCA1+GFAP+ area were increased by GW3965 compared with DMSO control (figure 2D–F).

Figure 2

GW3965 upregulates LXR downstream-gene expression. (A) Representative images of confocal tile scan showing whole day 7 lesion labelled with DAPI (blue), GFAP (Cyan), ABCA1 (green) and Iba1 (red). Scale bar, 200 µm. (B) Representative image of a three-dimensional (3D) reconstruction of a perihaematomal region in the treatment group using Imaris. Scale bar, 3 µm. (C) Representative confocal images of perihaematomal region labelled with DAPI (blue), GFAP (cyan), ABCA1 (green) and Iba1 (red). The left corner within dotted lines depicts lesion centre. Scale bar, 100 µm. (D–F), Bar graphs comparing the ABCA1+ percentage (D), Iba1+ABCA1+ over Iba1+ proportions (E) and GFAP+ABCA1+ over GFAP+ proportions (F) between GW3965 treatment and DMSO control within the lesion region at day 7; the sample size was n=10 per group. (G) Representative images of confocal tile scan showing whole day 7 lesion labelled with DAPI (blue), Iba1 (cyan), GFAP (green) and ApoE (red). Scale bar, 200 µm. Scale bars are not equal. (H), Representative image of a 3D reconstruction of a perihaematomal region in the treatment group using Imaris. Scale bar, 7 µm. (I), Representative confocal images of perihaematomal region labelled with DAPI (blue), Iba1 (cyan), GFAP (green) and ApoE (red). The left corner within dotted lines is the lesion centre. Scale bar, 100 µm. (J–L), Bar graphs comparing the ApoE+ percentage (J), GFAP+ApoE+ over GFAP+ proportions (K) and Iba1+ApoE+ over Iba1+ proportions (L) between GW3965 treatment and DMSO control within the lesion region at day 7; the sample size was n=10 per group. Significance is indicated as *p<0.05, **p<0.01, ****p<0.0001; two-tailed, unpaired t-test (D, F, K); or two-tailed, unpaired t-test with Welch’s correction (E); or Mann-Whitney U test (J, L). Bar graphs show individual data points and represent mean±SD. ApoE, apolipoprotein E; LXR, liver X receptor.

Similarly, ApoE was increased mainly at the outer layer of the perihaematomal area where it primarily overlapped with GFAP+ astrocytes (figure 2G, online supplemental figure S1B). The 3D reconstruction and confocal images at perihaematomal region exhibited close spatial association between ApoE and GFAP+ cells (figure 2H,I). Quantification showed upregulated ApoE expression and the increased ratio of ApoE+GFAP+ area in GW3965 treatment group (figure 2J,K). In contrast, the fractional ApoE+Iba1+ area was not significant between groups (figure 2L). Overall, the elevated ABCA1 in phagocytes may promote cholesterol and lipid efflux and recycling, while the increased ApoE would provide increased carrier capacity for lipid transport.

We also evaluated the expression of ABCA1 and ApoE at perimhaematomal regions using real-time qPCR (online supplemental figure S2A,B) and western blot (online supplemental figure S2C–E). Quantification showed that the transcripts and protein content of these genes were elevated after 7-day GW3965 treatment (online supplemental figure S2A–E), which is similar to the results from confocal data.

LXR agonist appears to promote phenotype change of M/M

We evaluated IL1β expression as it is a target gene of LXR trans-repression.31 33 IL1β was distributed predominantly in the same area as Iba1 positive M/M at the perihaematomal region (figure 3A–C). Iba1 and IL1β expression was downregulated by GW3965 in the lesion area (figure 3D,E), and the proportion of Iba1+IL1β+ M/M was also reduced (figure 3F).

Figure 3

Enhanced LXR activation promotes M/M phenotype shifting. (A) Representative images of confocal tile scan showing whole day 7 lesion labelled with DAPI (blue), IL1β (green) and Iba1 (red). Scale bar, 200 µm. The size of images (note the longer scale bar in the GW3965 group) is adjusted to present the whole lesion within each frame. (B) Representative image of a three-dimensional (3D) reconstruction of a perihaematomal region in the vehicle group using Imaris. Scale bar, 5 µm. (C) Representative confocal images of perihaematomal region labelled with DAPI (blue), IL1β (green) and Iba1 (red), the left corner within dot lines orients the lesion centre. Scale bar, 100 µm. (D–F) Bar graphs comparing the Iba1+ percentage (D) IL1β+ percentage (E) Iba1+IL1β+ over Iba1+ proportions (F) between GW3965 treatment and DMSO control within the lesion region at day 7; the sample size was n=10 per group. (G) Representative images of confocal tile scan showing whole day 7 lesion labelled with DAPI (blue), CD68 (green), CD206 (red) and Arg1 (magenta). Scale bar, 200 µm. Scale bars are not equal. (H) Representative image of a 3D reconstruction of a perihaematomal region in the treatment group using Imaris. Scale bar, 10 µm. (I) Representative confocal images of perihaematomal region labelled with DAPI (blue), CD68 (green), CD206 (red) and Arg1 (magenta); the left corner within dotted lines denotes the lesion centre. Scale bar, 100 µm. (J–N) Bar graphs comparing the CD68+ percentage (J), CD206+ percentage (K), Arg1+ percentage (L) CD68+CD206+ over CD68+ proportions (M) and CD68+Arg1+ over CD68+ proportions (N) between GW3965 treatment and DMSO control within the lesion region at day 7; the sample size was n=10 per group. Significance is indicated as *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001; two-tailed, unpaired t-test (D, J–L) or two-tailed, unpaired t-test with Welch’s correction (E, F, M, N). Bar graphs show individual data points and represent mean±SD. LXR, liver X receptor; M/M, microglia/macrophage.

With the reduced IL1β+ M/M, and as IL1β+ is an indicator of the proinflammatory nature of myeloid cells, we evaluated if GW3965 affected Arg1 and CD206, which are often used as regulatory markers of myeloid cells.5 49 Figure 4G–I displays representative staining in tissue from GW3965 or DMSO-treated animals. Consistent with lowered Iba1 percentage, CD68+ area was reduced in GW3965 compared with DMSO control sections (figure 3J). Conversely, the expression of regulatory markers in the lesion was upregulated (figure 3K,L), and the fraction of CD68+Arg1+ and CD68+CD206+ in total M/M was increased after 7-day GW3965 treatment (figure 3M,N).

Figure 4

Enhanced LXR activation mitigates lipid accumulation in phagocytes. (A) Representative images of confocal tile scan showing whole day 7 lesion labelled with DAPI (blue), Bodipy (green) and Iba1 (red). Scale bar, 200 µm. The size of images images (note the longer scale bar in the GW3965 group) is adjusted to present the whole lesion within each frame. (B) Representative image of a three-dimensioal (3D) reconstruction of a day 7 perihaematomal region in the vehicle group using Imaris, showing Iba1+ phagocytes enveloping Bodipy+ material. Scale bar, 15 µm. (C) Representative confocal images of perihaematomal region labelled with DAPI (blue), Bodipy (green) and Iba1 (red); the left corner within dotted lines depicts the lesion centre. Scale bar, 100 µm. (D, E), Bar graphs comparing the Bodipy+ percentage (D) and Iba1+Bodipy+ over Iba1+ proportions (E) between GW3965 treatment and DMSO control within the lesion region at day 7; the sample size was n=10 per group. (F) Representative confocal images of day 7 perihaematomal region labelled with DAPI (blue), crystal (grey), Fluoromyelin (green)and Iba1 (red); the left corner within dotted lines is the lesion centre. Scale bar, 100 µm. (G) Representative image of a 3D reconstruction of a day 7 perihaematomal region in the vehicle group using Imaris, showing Iba1+ phagocytes enveloping crystal and Fluoromyelin+ material. Scale bar, 10 µm. (H–K) Bar graphs comparing the crystal+ percentage (H) Iba1+crystal+ over Iba1+ proportions (I) Fluoromyelin+ percentage (J) and Iba1+Fluoromyelin+ over Iba1+ proportions (K) between GW3965 treatment and DMSO control within the perihaematomal region at day 7; the sample size was n=10 per group. Significance is indicated as *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001; two-tailed, unpaired t-test (D, E, H, I) or two-tailed, unpaired t-test with Welch’s correction (J, K). Bar graphs show individual data points and represent mean±SD. LXR, liver X receptor.

To corroborate the results from confocal microscopy, western blot was used to evaluate the expression of M/M markers in day 7 perihaematomal regions (online supplemental figure S2F). The expression of proinflammatory IL-1β and iNOS was inhibited by GW3965 treatment (online supplemental figure S2G,H), while that of regulatory/repair markers including CD206 and Arg1 was upregulated (online supplemental figure S2I,J).

LXR agonist alleviates lipid burden of phagocytes

To corroborate that GW3965 promoted lipid/c