Experimental Animals

All experiments involving animals were approved by the Ethics Committee of the Third Military Medical University (AMUWEC2020793). All procedures were performed in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. All the experimental mice were housed in a humidity- (50–60% relative humidity) and temperature-controlled (24 ± 1 °C) incubator with access to food and drinking water ad libitum under a 12-h light/dark cycle. Each cage usually contained five mice depending on the experimental group. The different groups of animals were not housed in the same cage. All the experimental mice received subcutaneous buprenorphine (0.02 mg/kg) within 24 h after surgery to reduce their distress and discomfort. PDGFRβret/ret mice are an important experimental models for studying pericyte function. Compared with wild-type mice, they had significantly fewer pericytes in their brain capillaries (Moura et al. 2017). To clearly determine the relationship between TIMP-3 and pericyte, we used PDGFRβret/ret mice as experimental subjects.

Experimental DesignExperiment I

To investigate the relationship between TIMP-3 and pericytes, we performed immunofluorescence staining of frozen brain tissue samples from six wild-type mice and six PDGFRβret/ret mice. The colocalization of TIMP-3 and PDGFRβ was also analyzed. The relationship between TIMP-3 and pericytes was observed and clarified by the above methods (Fig. 1Experiment I).

Fig. 1

Schematic diagram of the experimental design

Experiment II

To investigate the effect of SAH on the number of pericytes and the level of TIMP-3, we randomly divided 24 PDGFRβret/ret mice (n = 6) into 4 groups, namely, the sham group, the 1-day post-SAH group, the 3-day post-SAH group, and the 7-day post-SAH group. We repeated the grouping method with 24 wild-type mice. ELISA was performed to detect the content of TIMP-3 in the cerebrospinal fluid (CSF) of the experimental mice. Immunofluorescence staining and counting of TIMP-3-positive cells per mm2 were performed. The number of pericytes in the brain tissue of each wild-type mouse was determined. These experimental methods were applied to confirm the effect of SAH on pericytes and TIMP-3 (Fig. 1Experiment II).

Experiment III

To investigate the effect of aberrant expression of TIMP-3 on repair after SAH-induced white matter injury, we performed in vivo and in vitro experiments (Fig. 1Experiment III).

In Vivo

We randomly divided the 30 wild-type mice into 5 groups (n = 6): the sham group, the sham + vehicle group, SAH + vehicle group, SAH + TIMP-3 RNAi group, and SAH + TIMP-3 RNAi + TIMP-3 group. The behavioral tests included the beam balance test, determination of the modified Garcia score and the open field test. These experiments were performed at 24 h after SAH. These tests were used to assess the neurological function of the experimental mice. Changes in the myelin sheath and the content of MBP at that time were assessed by immunofluorescence staining, determination of the mean gray values, and transmission electron microscopy. Immunofluorescence staining and counting of PDGFRa-positive cells per mm2 were subsequently performed (PDGFRa is a marker of OPCs). The number of OPCs in the brain tissue of each experimental mouse 24 h after SAH was determined.

In Vitro

We used mouse primary pericytes and primary OLs as the study subjects. OxyHb and TIMP-3 overexpression and pericyte culture supernatant were used as intervention factors. ELISA, immunofluorescence staining and quantitative analysis of the number of positive cells per mm2 were used to detect the content of TIMP-3, as well as the number of pericytes and OLs, at 24 h after treatment.

These methods were applied to clarify the effect of TIMP-3 expression on the repair after white matter injury.

Mouse SAH Model

We used endovascular perforation to construct a mouse SAH model(Hao et al. 2022). Briefly, all the experimental mice were anesthetized with halothane (70% N2O and 30% O2; 4% for induction, 2% for maintenance, RWD 21,081,501). A midline incision in the neck was made to expose the left external carotid artery. The vessel was ligated and severed at the distal end, leaving a 2-mm stump and blocking blood flow from the left internal carotid artery. A 5–0 sharpened monofilament nylon suture was inserted into the left internal carotid artery until slight resistance was encountered (approximately 5–12 mm). Another 1–2 mm was then inserted to penetrate the anterior and middle cerebral bifurcations. After being held in place for 10 s, the nylon suture was removed, and the left external carotid artery stump was ligated. Finally, we restored internal carotid artery blood flow (Supplemental Fig. 1a). The experimental mice in the sham group underwent the same surgical procedure as those described above, except for the punctured cerebral vessels.

CSF Collection

We extracted CSF from the cisterna magna of the experimental mice(Siler et al. 2014). After the experimental mice were anesthetized and fixed on a stereotaxic frame, the atlas-occipital membrane was exposed through a median incision. A 10 μL Hamilton syringe (Microliter 701; Hamilton Company) was inserted (approximately 1–1.5 mm) into the cisterna magna, after which 10 μL of CSF was extracted.

Intracerebroventricular Injection

The intracerebroventricular injection steps were performed in experimental mice as previously described (Zhou et al. 2023). Briefly, after anesthesia, the left side of the skull was drilled (coordinates: 0.6 mm posterior to the fontanelle and 1.5 mm lateral to the sagittal suture of the skull). The needle of a 10 μL Hamilton syringe was subsequently inserted 1.7 mm into the left ventricle of the experimental mice. The needle was held in place for 10 min and then slowly removed. Two types of lentiviral vectors that regulate TIMP-3 expression were used for in vivo experiments (GeneChem, Shanghai, China):

1.

The Ubi-MCS-3FLAG-SV40-IREs-Puromycin lentiviral vector was used for TIMP-3 overexpression (abbreviated as the TIMP-3 IREs Group);

2.

The hU6-MCS-CMV-Puromycin lentiviral vector was used for TIMP-3 interference (abbreviated as the TIMP-3 RNAi group);

All lentiviral vectors were stored at −80 °C before use. The effects of the lentiviral vectors were verified by in vitro experiments (Supplement Fig. 1b). The lentiviral vector was injected into the left lateral ventricle at 1 μL/min. One week later, the experimental mice were used to construct the SAH model. For the SAH + TIMP-3 RNAi + TIMP-3 group, mice previously injected with TIMP-3 RNAi were injected with recombinant TIMP-3 (2 µl, Novus, 973-TM-010, 1 µg/µl) through the lateral ventricle 24 h before SAH.

Neurological Scores

At 24 h after SAH, the neurological function of the experimental mice was assessed by a double-blind method. The test included determination of the modified Garcia score and execution of the beam balance and open field tests (Qu et al. 2018; Chung et al. 2021; Zuo et al. 2017). The modified Garcia score is mainly used as a measure of spontaneous activity, symmetry of limb movement, forepaw extension, climbing, body proprioception, and response to vibrissal touch in experimental mice and ranges from 3 to 18. In the beam balance test, the experimental mice were placed in the center of a wooden beam. Then, walking distance within 1 min was assessed and scored from 0–4 points. In the open field test, the experimental mice were removed from the feeding cage and quickly placed in the central area of the experimental box. The activity of the experimental mice in the experimental box was recorded using animal behavioral analysis software for 15 min.

Transmission Electron Microscopy

Transmission electron microscopy was used to observe the myelin sheaths of the experimental mice 24 h after SAH (Zhou et al. 2023). The experimental mice were anesthetized and sacrificed by transcardial perfusion with cold 0.9% saline. The brain tissue was removed and fixed with 4% glutaraldehyde containing 100 mmol/mL sodium cacodylate (pH = 7.3) at 4 °C overnight. Then, the brain tissue was minced into 1 mm3 pieces and placed in Epon (Agar 100 resin, Agar Science, Essex, UK). Afterward, the myelin sheaths of the experimental mice were observed and imaged using a transmission electron microscope (JEM-1200 EX, JEOL, Tokyo, Japan).

Fluorescence Immunohistochemistry and Immunocytochemistry

Fluorescence immunohistochemistry was performed as previously reported (Zhou et al. 2023). In this study, experimental mice were sacrificed via transcardiac perfusion of cold saline on days 1, 3, and 7 after SAH. Then, the brain tissues were harvested, fixed, dehydrated, and sectioned using a cryomicrotome (CM3050S-3-1-1; Leica Biosystems, New York, USA). The brain tissue sections were first blocked with 0.3% Triton X-100 and 5% goat serum for 60 min at room temperature. Afterward, the brain tissue sections were incubated with diluted primary antibody overnight at 4 °C (the antibodies used were as follows: anti-PDGFRbeta antibody, Abcam #ab69506; anti-TIMP-3 antibody [20HCLC], Abcam # ab277794; mouse aminopeptidase N/CD13 antibody, R&D Systems # AF2335; and MBP-probe antibody (R29.6), Santa Cruz Biotechnology # sc-13564). The next day, the brain tissue sections were incubated with the corresponding fluorescent secondary antibody for 2 h at room temperature. Later, the nuclei were stained with DAPI for 5 min. The expression and distribution of different protein markers were observed via laser confocal microscopy (LSM880, Carl Zeiss, Oberkochen, Germany). Representative brain tissue sections from each experimental group were obtained for imaging. Immunofluorescence images were analyzed using ImageJ V1.53 (NIH) to determine the fluorescence intensity and number of positive cells.

Immunocytochemistry was also performed as previously reported (Zhou et al. 2023). The cultured experimental cells were first fixed with 4% paraformaldehyde for 15 min. Afterward, the impurities and paraformaldehyde around the cells were removed by washing with PBS. Primary antibodies (anti-TIMP-3 antibody [20HCLC], Abcam # ab277794; mouse aminopeptidase N/CD13 antibody, R&D Systems # AF2335; MBP-probe antibody [R29.6], Santa Cruz Biotechnology # sc-13564); secondary antibodies; and DAPI staining were used. The following steps were performed for fluorescence immunohistochemistry as described above.

Cell Culture and Treatments

Primary pericytes were cultured following previously reported procedures (Zhou et al. 2023). In brief, the steps were as follows: 1. Brain tissue was obtained from C57B6J mice (1 to 3 days old). 2. The cells were digested with MEM, HEPES (A501612, Sangon Biotech, Shanghai, China) and 40 μg/mL DNase I for 70 min. An additional 30% bovine serum albumin (A8010, Solarbio, Beijing, China) was added to the PBS for the final digestion step (BSA:PBS = 1.7:1). 3. The samples were then centrifuged at 1360 × g for 10 min. The pellet was carefully collected and incubated in EGM-2MV BulletKit medium (CC-3202, Lonza Bioscience, Basel, Switzerland). The cells were subsequently resuspended. The samples were subsequently centrifuged at 250 × g for 5 min. 4. Primary pericytes were subsequently seeded in culture dishes (C8061; Solarbio, Beijing, China) at a density of 5 × 105/cm2 and cultured in an incubator. 5. The culture medium was changed after 20 h and then every 3 days thereafter. When the primary pericytes formed a confluent layer, three consecutive cell passages were performed at a 1:4 ratio. 6. After the third generation, the primary pericytes were cultured in pericyte medium (e76127787; ScienCell Research Laboratories, California, USA). Primary pericytes were subsequently subjected to in vitro experiments.

Primary OLs were cultured following previously reported procedures (Li et al. 2022a). In brief, the steps were as follows: 1. Brain tissue was obtained from C57B6J mice (1 to 3 days old). 2. The cerebral cortex was separated and cut into 2–3 mm2 pieces. 3. Then, brain tissue digestion was performed with papain and DNase (Sigma‒Aldrich, Darmstadt, Germany) in a 37 °C water bath for 10 min with occasional inversion. 4. The obtained cells were centrifuged at 1000 × g for 5 min and resuspended in growth medium supplemented with 90% DMEM/F12 (Gibco, Massachusetts, USA), 10% fetal bovine serum (Gibco, Massachusetts, USA) and 1% penicillin streptomycin (Solarbio, Beijing, China). 5. The isolated cells were maintained in T-75 culture flasks (Solarbio, Beijing, China). 6. The culture medium was changed once every 3 days. On day 6, the culture medium was changed to OL differentiation medium. 7. The OL differentiation medium was changed every 3 days. On day 9, OLs were obtained and purified for use in in vitro experiments.

For in vitro experiments, primary pericytes were incubated with OxyHb (Sangon Biotech, Shanghai, China) for 24 h to mimic SAH. The Ubi-MCS-3FLAG-SV40-puromycin lentiviral vector was used to upregulate TIMP-3 expression in pericytes (TIMP-3 RNAi group). The pericytes were infected (MOI = 10) for 24 h, transferred back to conventional medium and cultured for 72 h before the experiments were performed.

ImageJ Was Used for Image Analysis

Immunofluorescence images of each group taken under the same magnification and exposure conditions were collected. The immunofluorescence images taken were imported into ImageJ software. To reduce the impact of background noise, background subtraction was applied to each image. By selecting the average intensity of the undyed area in the image, this intensity value was subtracted from the entire image. ImageJ’s threshold tool was used to select fluorescently labeled areas in the image. We subsequently adjusted the threshold to include all positive fluorescence signals while excluding nonspecific signals. For each image, we manually mapped the area of interest, ensuring that only specific parts of the cell or tissue were analyzed for fluorescence intensity. Afterward, using ImageJ’s “Measure” function, the average fluorescence intensity within each ROI was calculated, as was the corresponding region size. The fluorescence intensity data and region sizes of all ROIs were exported into spreadsheet software for further analysis. All the imaging data in each experimental group were statistically analyzed. The mean fluorescence intensity and standard deviation were determined. To ensure the accuracy of the analysis, all the quantitative steps were performed independently by two researchers. The results were compared and validated. In our Results section, we reported the mean fluorescence intensities and their standard deviations for all the experimental groups. Afterward, we compared the results using appropriate statistical methods.

Statistical Analysis

All the data were statistically analyzed and plotted using Prism 9 software. The required sample size for each experimental group is determined by ‘http://powerandsamplesize.com/’. Calculated using online tools and used as the basis for research design, Shapiro‒Wilk tests were performed on the data from each experimental group. P values greater than 0.05 were considered to not significantly violate the normality hypothesis. To determine the homogeneity of the variance between the experimental groups, we performed an F test. A P value greater than 0.05 indicated that there was no variance heterogeneity. These tests were all conducted using Prism 9 (GraphPad). The results of the tests determined the statistical methods used for subsequent data analysis. All the data are presented as the means ± standard deviations and were analyzed using Prism 9 (GraphPad) software. Groups were compared using one-way ANOVA + Tukey’s multiple comparison test. Neurological scores were analyzed by using the χ2 test. Two-way repeated-measures ANOVA was used to compare the behavioral data of TIMP-3 levels between different time points and different groups. Differences were considered statistically significant at P < 0.05.